IAEA-TECDOC-872

Progress in design, research

and development and testing of
safety systems for advanced
water cooled reactors
Proceedings of a Technical Committee meeting
held in Piacenza, Italy, 16-19 May 1995

INTERNATIONAL ATOMIC ENERGY AGENCY

The originating Section of this publication in the IAEA was:

Nuclear Power Technology Development Section
International Atomic Energy Agency
Wagramerstrasse 5
P.O. Box 100
A-1400 Vienna, Austria

PROGRESS IN DESIGN, RESEARCH AND DEVELOPMENT

AND TESTING OF SAFETY SYSTEMS FOR ADVANCED
WATER COOLED REACTORS
IAEA, VIENNA, 1996
IAEA-TECDOC-872
ISSN 1011-4289
IAEA, 1996

Printed by the IAEA in Austria

April 1996

FOREWORD

The Technical Committee Meeting on Progress in Design, Research & Development

and Testing of Safety Systems for Advanced Water Cooled Reactors was held at the ENEL
Training Centre, Piacenza, Italy from 16 to 19 May 1995. The meeting was convened by
the International Atomic Energy Agency (IAEA) within the frame of activities of the IAEA's
International Working Group on Advanced Technologies for Water Cooled Reactors and was
a joint activity of the IAEA's Nuclear Power Technology Development Section of the
Division of Nuclear Power and the Engineering Safety Section of the Division of Nuclear
Safety. IAEA activities in advanced technology for water cooled reactors serve to provide
an international source of objective information on advanced water cooled reactors and to
provide an international forum for discussion and review of technical information on research
and development activities.
Heat removal systems for evolutionary designs (advanced reactor designs which
achieve improvements over existing designs through small to moderate modifications) require
at most only engineering and confirmatory testing. Heat removal systems for developmental
designs (advanced reactor designs which range from moderate modification of existing
designs to entirely new design concepts) in general require more extensive testing and
demonstration to verify component and system performance. Key issues are scaling effects
for simulated plant configurations, component and system reliabilities, aging, and interactions
among different systems. Furthermore, a key activity is validation of the computer codes
used for design and safety analyses of advanced water cooled reactors by comparison with
experimental test data.
The meeting covered the following topics:

Developments in design of safety-related heat removal components and systems for

advanced water cooled reactors.
Status of test programmes on heat removal components and systems of new designs.
Range of validity and extrapolation of test results for the qualification of
design/licensing computer models and codes for advanced water cooled reactors.
Future needs and trends in testing of safety systems for advanced water cooled
reactors.

Tests of heat removal safety systems have been conducted by various groups
supporting the design, testing and certification of advanced water cooled reactors. The
Technical Committee concluded that the reported test results generally confirm the predicted
performance features of the advanced designs.

EDITORIAL NOTE
In preparing this publication for press, staff of the IAEA have made up the pages from the
original manuscripts as submitted by the authors. The views expressed do not necessarily reflect those
of the governments of the nominating Member States or of the nominating organizations.
Throughout the text names of Member States are retained as they were when the text was
compiled.
The use of particular designations of countries or territories does not imply any judgement by
the publisher, the IAEA, as to the legal status of such countries or territories, of their authorities and
institutions or of the delimitation of their boundaries.
The mention of names of specific companies or products (whether or not indicated as registered)
does not imply any intention to infringe proprietary rights, nor should it be construed as an
endorsement or recommendation on the part of the IAEA.
The authors are responsible for having obtained the necessary permission for the IAEA to
reproduce, translate or use material from sources already protected by copyrights.

PLEASE BE AWARE THAT

ALL OF THE MISSING PAGES IN THIS DOCUMENT
WERE ORIGINALLY BLANK

CONTENTS

SUMMARY OF THE TECHNICAL COMMITTEE MEETING

OPENING ADDRESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
L. Noviello

19

STATUS AND PLANS OF DEVELOPMENT AND TESTING

PROGRAMMES (Session I)

The programme of advanced light water reactors in Spain . . . . . . . . . . . . . . . .

M. Malave
Progress in design, research and development and testing of safety systems for the
Korean next generation reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Young Sang Choi, Byong Sup Kirn
The status of the ALPHA-Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
G. Yadigaroglu, G. Varadi, J. Dreier, F. de Cachard, B. Smith, S. Giintay,
Th. Bandurski

23
27
37

PREDICTED PERFORMANCE AND ANALYSIS OF ADVANCED

WATER COOLED REACTOR DESIGNS (Session II)
Evaluation of the design options for future power plants: Identification of the safety
related criteria and evaluation of the decay heat removal options . . . . . . . . . .
G.L. Fiorini
Postulated small break LOCA simulation in a CANDU type reactor with ECC
injection under natural circulation conditions . . . . . . . . . . . . . . . . . . . . . .

55
67

G. Bedrossian, S. Gersberg
A risk-based margins approach for passive system performance reliability analysis .
N.T. Saltos, A. El-Bassioni, C.P. Tzanos

79

Feasibility and efficiency studies of future PWR safety systems . . . . . . . . . . . . .

87

P. Aujollet
Risk reduction potential of jet condensers . . . . . . . . . . . . . . . . . . . . . . . . . .
A.W. Reinsch, T.G. Hook, K.I. Soplenkov, V.G. Selivanov,
V.V. Bredikhin, I.I. Shmal, Y. Filimonov, Y.N. Pavlenko
Diversified emergency core cooling in CANDU . . . . . . . . . . . . . . . . . . . . . .
P.J. Alien, N.J. Spinks
Mitigation of total loss of feedwater event by using safety depressurization system .
Y.M. Kwon, J.H. Song, S.Y. Lee, S.K. Lee

95

105
113

DESIGN AND ANALYSIS OF ADVANCED WATER COOLED REACTOR

SAFETY COMPONENTS AND SYSTEMS (Session ffl)

European pressurized water reactor configuration, functional requirements and

efficiency of the safety injection system . . . . . . . . . . . . . . . . . . . . . . . . .
F. Curca-Tivig, J.L. Gandrille
SWR-1000: the dimensioning of emergency condenser and passive
pressure pulse transmitters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
C. Palavecino

125
137

Design, fabrication and testing of full scale prototype for passive

cooling applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
P. Giribaldi, F. Magris, M. Orsini, Fl. Rizzo

147

RESULTS OF SAFETY-RELATED COMPONENTS/SYSTEMS TESTS

(Session IV) (Part 1)

Investigation on passive decay heat removal in advanced water cooled reactors . . . 159
F.J. Erbacher, X. Cheng, H.J. Neitzel
SPES-2, AP600 integral systems test results . . . . . . . . . . . . . . . . . . . . . . . . 173
L.E. Conway, R. Hundal
Experimental investigation on an inherently actuated passive injection and
depressurization system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
L. Mansani, L. Barucca, G.P. Gaspari
Design and testing of passive heat removal system with ejector-condenser . . . . . . 197
K.I. Soplenkov, V.G. Selivanov, Yu.N. Filimontsev, B.I. Nigmatulin,
V.V. Bredikhin, E.I. Trubkin, E.Z. Emeljanenko, A.W. Reinsch
An experimental study on the behaviour of a passive containment cooling system
using a small scale model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
D. Saha, N.K. Maheshwari, D.K. Chandraker, V. Venkat Raj, A. Kakodkar
RESULTS OF SAFETY RELATED COMPONENTS/SYSTEMS TESTS
(Session IV) (Part 2)
Steam injector development for ALWR's application . . . . . . . . . . . . . . . . . . .

221

G. Cattadori, L. Galbiati, L. Mazzocchi, P. Vanini, V. Cavicchia

Tests on full-scale prototypical passive condensers for SBWR application . . . . . . . 233
P. Masoni, A. Achilli, P.P. Billig, S. Botti, G. Cattadori, R. Silverii
Testing status of the Westinghouse AP600 . . . . . . . . . . . . . . . . . . . . . . . . . 245
E.J. Piplica, J.C. Butler
Testing for the AP-600 automatic depresurization system . . . . . . . . . . . . . . . . 263
T. Bueter, L. Conway, P. Incalcaterra, C. Kropp
The study of the effectiveness of the emergency condenser of the BWR 600/1000
in the NOKO test facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
E.F. Hicken, H. Jaegers, A. Schaffrath
Experimental study of isolation condenser performances by piper-one apparatus . . . 279
R. Bovalini, F. D'Auria, G.M. Galassi, M. Mazzini
COMPUTER MODEL DEVELOPMENT AND VALIDATION (Session V)

Analysis of PACTEL passive safety injection tests with RELAP5 code . . . . . . . .

R. Munther, J. Vihavainen, J. Kouhia
Heat transfer to an in-containment heat exchanger in natural convection flow:
validation of the AEA Technology computational fluid dynamics code

297

CFDS-FLOW3D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
R. O'Mahoney, J.N. Lillington
ATHLET model improvement for the determination of heat transfer coefficients
during condensation of vapor in horizontal tubes . . . . . . . . . . . . . . . . . . . . 323
E.F. Hicken, H. Jaegers, A. Schaffrath
RELAP5/MOD3 pre-test predictions for the SPES-2 1" C.L. break test S01613 ... 331

A. Alemberti, C. Frepoli, G. Graziosi

Development and initial validation of fast-running simulator of PWRs: TRAP-2 . . .

E. Brega, C. Lombardi, M. Ricotti, R. Sordi
Heat and mass transfer phenomena in innovative light water reactors . . . . . . . . .

W. Ambrosini, F. Oriolo, G. Fruttuoso, A. Manfredini,

F. Parozzi, M. Valisi
Application of the UMAE uncertainty method in assessing the design and the
safety of new generation reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
F. D'Auria, G. Fruttuoso, G.M. Galassi, S. Galeazzi, F. Oriolo, L. Bella,
V. Cavicchia, E. Fiorino
LIST OF PARTICIPANTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

341

353

365

381

SUMMARY OF THE TECHNICAL COMMITTEE MEETING

The progress in design, R&D and testing of safety systems for advanced water cooled
reactors was addressed in a TCM convened in Piacenza, Italy in May 1995 and organized in
cooperation with ENEL SpA (ENEL), Ente per le Nuove Technologic, 1'Energia e 1'ambiente
(ENEA) and Societa Informazioni Esperienze Termoidrauliche (SIET). The meeting was
attended by 71 participants from 12 countries and two international organizations as follows:
Canada (1), Finland (2), France (8), Germany (5), India (1), Italy (37), Republic of Korea
(3), Russia (5), Spain (1), Switzerland (1), United Kingdom (2), United States of America
(12), World Association of Nuclear Plant Operators, WANO (1), IAEA (2). Thirty four
papers were presented and future needs and trends in the field were discussed.
Background

A main focus of reactor design and development efforts in the industrialized countries
is on large size water cooled reactor units, with power outputs well above 1000 MW(e),
typically aiming at achieving certain improvements over existing designs. The alterations and
modifications to a specific design are generally kept as small as possible taking maximum
advantage of successful proven design features and components while taking into account
feedback of experience from licensing, construction, commissioning and operation of the
water cooled reactor plants currently in operation. The design improvements span a wide
range but still enhanced safety, increased reliability, more user-friendly systems, and better
economics represent rather common denominators for the new designs.
New designs - designs that have not yet been built or operated - are generally called
advanced designs. If they do not deviate too much from their predecessors, they are
designated evolutionary designs; the designers have retained many proven features of today's
plant design, and the evolutionary plant designs therefore require at most engineering, and,
possibly, confirmatory testing of some components and systems prior to commercial
deployment.
Some evolutionary large size LWR designs are currently under construction, while
others are in varying phases of design and regulatory review. The N4 model, a 1400 MW(e)
PWR, which is under construction in France, derives directly from the standardized P4 series
of 1300 MW(e). The British Sizewell-B, a 1250 MW(e) PWR, is another example of an
evolutionary design - an evolution of the WPSS design of Westinghouse. Other examples
of large advanced, evolutionary designs are the Westinghouse-Mitsubishi advanced pressurized
water reactor APWR (a 1350 MW(e) PWR), the ABB Combustion Engineering System
80+ (a 1300 MW(e) PWR), the Russian V-392 (a 1000 MW(e) PWR of the WWER-1000
type), and the General Electric-Hitachi-Toshiba Advanced Boiling Water Reactor ABWR (a
1350 MW(e) BWR). In this context, it may be noted that two ABWR units are under
construction in Japan - at Kashiwazaki Kariwa (No. 6/7) - with grid connections scheduled
for 1996 and 1997, and that construction of two APWR units is planned at the Tsuruga site
in Japan.

Framatome and Siemens have established a joint company, Nuclear Power

International, which is developing a new advanced large PWR of 1500 MW(e) (gross) with

enhanced safety features, and they intend to have it reviewed jointly by the French and

German safety authorities. This procedure will provide strong motivation for the practical
harmonization of the safety requirements of two major countries, which could later be
enlarged to a broader basis. Siemens is also, together with German utilities, engaged in the
development of an advanced BWR design, the SWR-1000, which will incorporate a number
of passive safety features, for initiation of safety functions, for residual heat removal and for
containment heat removal. In Sweden, ABB Atom, with involvement of the utility
Teollisuuden Voima Oy (TVO) of Finland, is developing the BWR 90 as an upgraded version
of the BWRs of the BWR 75 version operating in both countries.
Smaller advanced LWRs are also being developed in a number of countries, in most
cases with a great emphasis on utilization of passive safety systems and inherent safety
features. Two typical examples in this context are the Advanced Passive PWR (AP-600) of
Westinghouse and the Simplified BWR of General Electric - the mid-size passive units (units
in the 600 MW(e) range) of the US ALWR programme. The development of both these
designs has been governed by a key guideline: new features should need no more than
engineering and confirmatory testing before commercial deployment; they must not lead to
a requirement of building and operating a prototype or demonstration plant.

An important programme in the development of advanced light water reactors was

initiated in 1984 by the Electric Power Research Institute (EPRI), an organization of US
utilities, with financial support from the US Department of Energy, and participation of US
nuclear plant designers. Several foreign utilities have also participated in, and contributed
funding to, the programme. As a part of the programme, to guide the ALWR design and
development, utility requirements were established for large BWRs and PWRs having power
ratings of about 1200 MW(e), and for mid-sized BWRs and PWRs having power ratings of
about 600 MW(e). In an important related development, a major step forward in licensing
of future plants was reached with the ratification of the Energy Policy Act of 1992. The new
licensing process allows nuclear plant designers to submit their designs to the US Nuclear
Regulatory Commission (US NRC) for design certification. Once a design is certified, the
standardized units will be commercially offered, and a utility can order a plant confident that
generic design and safety issues have been resolved. The licensing process will allow the
power company to request a combined license to build and operate a new plant, and as long
as the plant is built to pre-approved specifications, the company can start up the plant when
construction is complete, assuming no new safety issues have emerged. In 1994, the two
large evolutionary plants which have resulted from the U S programme, ABB-Combustion
Engineering's System 80+ and General Electric's (GE's) ABWR, received Final Design
Approval, the last step before design certification, from the US NRC. The two smaller 600
MW(e) plants, Westinghouse's AP-600 and GE's Simplified BWR, are expected to receive
Final Design Approval in 1996 and 1997 respectively. Today, completion of design
certification for advanced light water reactors to assure availability of ALWRs by the end of
the decade is a top priority of the Advanced Reactor Programme of the US Department of
Energy.
An adaption of the AP600 and SBWR designs in order to meet the European Utility
Requirements and to meet the need for a higher net electric power is under progress in
Europe. The two programmes are named respectively EPP and ESBWR.

In the Russian Federation, design work on the evolutionary V-392, an upgraded

version of the WWER-1000, has been started, and another design version is being developed
in cooperation with Finland. The Russian Federation is also developing an evolutionary

10

WWER-640 (V-407) design along lines similar to those for the AP-600, as well as a more
innovative, integral design, the VPBER-600.

In Japan, the Ministry of International Trade and Industry is conducting an 'LWR

Technology Sophistication' programme focusing on development of future LWRs and
including requirements and design objectives. The Japan Atomic Energy Research Institute
(JAERI) has been investigating conceptual designs of advanced water cooled reactors with
emphasis on passive systems. These are the JAERI Passive Safety Reactor (JPSR) and the
System-Integrated PWR (SPWR). Development programmes for a Japanese simplified BWR
(JSBWR) and for a Japanese simplified PWR (JSPWR) are in progress jointly involving
vendors and utilities.
In Canada, two advanced CANDU heavy water moderated and cooled reactor designs,
the CANDU-3 and CANDU-9, are under development. The CANDU-3 is a 450 MW(e)
design which features modular design and construction methods. The 1050 MW(e) CANDU9 design is based on the proven Darlington and Bruce B large size plants with improvements
derived from the design activities of CANDU-3.

In India, development efforts are conducted on an advanced pressure tube AHWR

which incorporates passive features including containment cooling and decay heat removal by
natural convection.
National and international development efforts presented at TCM:

While current NPPs resulted largely from national design and development activities,
advanced LWR programmes are strongly based on international cooperation.
The basic design of the European Pressurized Water Reactor (EPR), an evolutionary
PWR with innovative features, has been started in February 1995 under the leadership of
Nuclear Power International. The EPR derives from experience gained by EdF, German
utilities, Framatome and Siemens during design, licensing construction and operation of
current nuclear power plants. The main functions and functional requirements of the EPR
safety injection system have been established. The safety injection system design consists of
four independent trains, each including an accumulator, a medium head safety injection pump

and a low head safety injection pump, both of which inject water from a refuelling water
storage tank located in the containment into the reactor coolant system. Analyses have shown
that the safety injection system ensures mitigation of all LOG A scenarios and all relevant nonLOCA scenarios.
In the Republic of Korea an effort has been under way since 1992 to develop an
advanced design termed the Korean Next Generation Reactor (KNGR) which is an
evolutionary large 4000 MW(th) PWR design. The goal is to complete a detailed standard
design by the year 2000. KEPCO is the lead organization. KAERI is responsible for NSSS
design, with Korean Power Engineering Company (KOPEC) and Korea Heavy Industry
Company (KHIC) responsible for architectural engineering and component design
respectively. In the first phase, completed in 1994, the design requirements and the design
concept were established. In the current phase, the basic design is being developed by
KEPCO and the Korean nuclear industry, with financial support by the government, in
parallel with licensing review. Phase 3, scheduled to begin in 1997, will establish the detailed
standard design.

11

Many confirmatory component and systems performance tests for mid-sized ALWRs
have been completed while others will be completed in 1996. These mid-sized advanced
plant designs are the result of an intensive international effort including more than a dozen
countries which participate in the US ALWR programme. A key characteristic of these plants
is the use of simple passive systems to respond to design basis events. Many tests of these
systems are required to satisfy regulatory review leading to design certification. Development
of the ALWR designs is supported by both private (including international utility
organizations) and government funding, and the programme is carried out in parallel with the
design certification review by the US NRC. Test programmes to confirm the predicted safety
performance of the Westinghouse AP-600 and the General Electric SBWR are close to
completion. Especially in the past two years extensive progress both in testing programmes
and in the regulatory review has been made in the US ALWR programme toward the
objective of deploying an advanced plant in the United States.
A technical tour to the SIET facilities in Piacenza provided a detailed understanding
of the SPES-2 full height, full pressure integral heat transport systems test for the AP-600,
and the testing at SIET's PANTHERS test loop of full-scale prototype condensers for both
the Passive Containment Cooling System (PCCS) and the Isolation Condenser System (ICS)
of the GE-SB WR. A detailed description of the PANDA facility at the Paul Scherrer Institute
(PSI) for SBWR heat removal systems tests was provided. These programmes at SIET and
PSI support the current design certification activities for AP-600 and the GE-SBWR.
Siemens is developing the SWR-1000, characterized by its use of passive safety
systems with the aim of achieving system simplification and reduced capital costs. Passive
systems which are incorporated into SWR-1000 include emergency condensers, containment
cooling condensers, passive pressure pulse transmitters and gravity driven core flooding lines.
The testing programme for the emergency condenser is currently underway in a cooperative
effort at the Juelich Research Center (KFA) Germany with funding from the federal
government, German utilities, Siemens and KFA.

The French Atomic Energy Commission (CEA) is conducting evaluations of innovative

safety systems for ALWRs, in terms of efficiency and reliability in selected accidental
sequences. Three kinds of systems are currently evaluated: direct depressurization system
designed to control PWR primary pressure to prevent high pressure core melt sequences,
passive residual heat removal systems and steam injector systems. A systematic integrated
approach for identifying safety objectives and establishing criteria for selecting design options
is being developed.
Passive containment cooling by natural convection of air has been proposed for
ALWRs. To investigate the phenomena involved, separate effects tests are conducted at KfK,
Germany in the PASCO facility. Further, a scaled test facility to investigate integral
containment coolability limits by air under coupled convection and thermal radiation
conditions is being designed.
Spain's advanced technology for water cooled reactors includes activities to improve
performance of its existing reactors, participation in the design of the four ALWRs under
development in the US, and involvement in European activities to define user requirements
for advanced reactors. The effort is aimed at maintaining and increasing Spain's domestic
nuclear industrial capabilities and providing a technical base for choices among electricity
supply options in Spain's next National Energy Plan.

12

In Russia, experimental and analytical efforts are on-going to develop a new type of
passive heat removal system which is applicable to all nuclear plants with steam cycle power
conversion, i.e. PWRs and BWRs including advanced designs, and for heat removal from the
containment as well. The activities are conducted at the Research Institute for Nuclear Power
Plant Operation, RINPO (Moscow, Russia) in cooperation with other research centers in
Russia, Ukraine and USA. The concept is based on passive heat removal by steam
condensation (PAHRSEC) in an ejector-condenser recirculating the condensate to the nuclear
heat source. Tests of a full size PAHRSEC system proposed for retrofit for the currently
operating WWER-440 plants at Novovoronezh are underway. Probabilistic safety analyses
carried out in the USA have indicated that retrofit of the San Onofre nuclear plant with a
PAHRSEC system would reduce the probability of core damage from internal events by a
factor of about 2.

In Switzerland, in 1991 at the Paul Scherrer Institute, the ALPHA project was initiated
with the goal of performing experimental and analytical investigations of long-term decay heat
removal from the containment of next generation passive ALWRs. The project is conducted
in collaboration with a large international team including design and research organizations
from USA, the Netherlands, Japan, Mexico and Italy. Because results of certain tests will
be submitted to the USNRC as part of the GE-SBWR design certification process, they are
conducted under formal quality assurance procedures. Three major facilities have been
constructed including a large scale integral facility (PANDA) to simulate the GE-SBWR
containment, a facility (LINX-2) for investigating condensation and buoyancy driven mixing
in the containment, and a facility to investigate aerosol transport in the containment. Testing
at these facilities is underway.
In Italy, an extensive development programme for safety components and systems for
advanced water cooled reactors is conducted in a framework of international cooperation.
A comprehensive test and analysis effort to support design confirmation of the safety features
of the Westinghouse AP-600 is being completed at the test facilities VAPORE (ENEA
Casaccia - full-scale Automatic Depressurization System tests) and SPES-2 (SIET Piacenza integral systems tests) and involves cooperation between Westinghouse, Ansaldo, ENEA,
ENEL and University of Pisa. As a part of General Electric's design certification effort for
the SBWR, ANSALDO, ENEA and ENEL in cooperation with GE are conducting a test
programme of full scale prototypical condensers for the SBWR passive containment cooling
system and the SBWR isolation condenser system in SIET's PANTHERS test facilities. Also,
an innovative passive depressurization and safety injection system is under development by
ANSALDO and the viability of this system has been demonstrated by testing at SIET.
ENEL is testing an innovative passive containment cooling system from a concrete
containment at CISE. ENEL and CISE have developed a high performance steam injector and
its operation has been demonstrated by testing at SIET. ENEL is also testing in an
international framework the behaviour under severe accidents of the personnel airlock of the

abandoned Alto Lazio plant. At University of Pisa, the PIPER-ONE facility, simulating a
BWR, was modified to include an out-of-scale Isolation Condenser component. Tests have
been performed that demonstrated the facility to be very useful for code assessment.
In Canada, a passive moderator heat rejection system for advanced CANDU reactors
is being developed as a diverse emergency heat rejection system. In CANDU reactors, low
pressure heavy water moderator surrounds the fuel channels and is available as a heat sink

to maintain channel integrity and avoid fuel melting in the unlikely event of a loss of coolant
13

accident with loss of emergency coolant injection. Existing CANDU reactors use pumps to
circulate the moderator and the cooling water. The passive moderator heat rejection system
involves a natural circulation of heavy water driven by flashing to steam as the heavy water
flows to an elevated heat exchanger. The heat exchanger is cooled in turn by a natural
circulation flow of light water to a large reservoir. The overall concept has been verified
both by analysis and by operation of a full elevation, light water, l/60th volume scaled test
of the natural circulation heavy water loop by AECL.
In Argentina, the CANDU core refill performance of the high pressure accumulator
emergency coolant injection system under natural circulation conditions has been investigated
for the Embalse nuclear plant for conditions resulting from a small LOCA. Acceptably safe
fuel temperatures are predicted even for cases assuming manual initiation of the ECCS.
In India, testing of a passive containment cooling system which is proposed for India's
Advanced HWR has been conducted at the Bhabha Atomic Research Centre. The system is
designed to provide containment cooling following a LOCA by removing energy through
isolation condensers immersed in a pool of water at the top of the containment. The test
programme is examining the efficiency of heat transfer and of the removal of noncondensables to a vapour suppression pool.
In summary, arriving at the most advanced point in the realization of the passive safety
plants has required extensive component and integral systems effects tests. These tests have

been performed by the various groups supporting the design, testing and certification of
advanced PWRs and BWRs. The papers presented at the TCM showed results generally
confirming the expected performance of the passive safety systems of advanced designs.
Brief description of major components and systems being developed:

Ansaldo Passive Injection and Depressurization System (PIPS)

One of the means to mitigate the consequences of a loss of coolant accident is to inject
cold water into the reactor coolant system at low pressure. This requires depressurization of
the primary system. A concept of a passive injection and depressurization system which
functions without making use of any active component or actuation logic is being developed
by Ansaldo (Italy). The system performs the depressurization function by mixing (borated)
cold water with the steam present in the reactor cooling system (RCS). The cold water
injection is actuated by low RCS water inventory, while the depressurization (down to the
containment pressure) is performed by a valve passively actuated by low RCS pressure.
Because of the innovative nature of this concept, before proceeding to the actual
system design, an experimental investigation of the physical phenomena has been carried out
at SIET and has demonstrated concept viability.

The next steps will consist of:

separate effect tests aimed to gain information on the attainable depressurization rates
and on the associated phenomena;
integral scoping tests aimed at assessing the PIDS interaction with the other NPP
systems.
14

Steam injector

A steam injector is a device in which steam is used as the energy source to pump cold
water from a pressure lower than the steam to a pressure higher than the steam. Heat
available from steam condensation can be partly converted into mechanical work useful for
pumping the liquid. The steam injector consists of a steam nozzle to partially convert steam
enthalpy into kinetic energy, a water nozzle to distribute inlet liquid around the steam, a
mixing section for heat, mass and momentum transfer from the steam to the water resulting
in condensation at higher pressure than the inlet steam, and a diffuser. Steam injectors are
claimed to be useful in advanced light water reactors for high pressure makeup water supply
without introduction of any rotating machinery. In particular, steam injectors could be used
for high pressure safety injection in BWRs or for emergency feedwater in the secondary side
of PWRs. An optimization study has been conducted by CISE in cooperation with ENEL,
and steam injectors for advanced water cooled reactor application were tested in 1994 at
SIET.
BWR passive containment cooling systems (GE-SBWR and Siemens SWR-1000)

Passive containment cooling systems for BWR application consist of condensers

connected to the upper space of the drywell region of the BWR containment. During a lossof-coolant accident, steam in the drywell is driven into the containment cooling condenser by
the pressure difference between the drywell and wetwell in combination with the vacuum
produced by condensation in the condenser. The condensate flows down into the gravitydriven cooling system pool in the drywell which provides makeup to the reactor. Noncondensable gases, such as containment nitrogen, are separated in the passive containment
condenser and vented to the wetwell. In the General Electric SBWR concept the condensers
are located in water pools outside the containment, whereas in the Siemens SWR-1000
concept the condensers are inside the containment. In the GE-SBWR the containment
"atmosphere" flows inside the condenser tubes in an expanded containment; in the Siemens
SWR-1000 the containment atmosphere is on the outside of the condenser tubes through which
water from a pool outside containment flows in natural circulation.
Passive containment cooling system (AP-600)

The passive containment cooling system developed for AP-600 consists of a steel
containment shell, inside a shield building, cooled by natural circulation of air, transferring
heat from the containment to the environment (as the ultimate heat sink). In order to enhance
the heat transfer capability, the system has been provided with large, elevated water storage
tanks; when needed valves are opened and water is spread onto the containment shell yielding
enhanced heat transfer by water evaporation.
Core make-up tanks, in-containment refuelling water storage tanks, and accumulators

Core make-up tanks and in-containment refuelling water storage tanks contain large
volumes of borated water which can be injected into the primary system of PWRs by gravity.
Accumulators are pressurized water storage tanks which passively inject borated water
into the primary system of PWRs when the primary system pressure falls below the design
selected value.

15

Isolation condenser (GE-SBWR)

The main purpose of the isolation condenser is to limit the pressure in the reactor
system to a value below the set-point of the safety relief valves, in the event of a main steam
line isolation. The condensers are submerged in a pool of water located in the reactor
building above the reactor containment. The primary side of the three isolation condensers
is connected by piping to the reactor containment. Closed valves in each condensate return
line prevent flow through the condenser during normal power operation of the plant. When
operation of the isolation condenser system is required, the valves are opened, the steam
flows directly from the reactor to the condensers and the condensate is returned to the reactor
vessel by gravity. The rate of flow is determined by natural circulation. Vent lines are
provided to remove non-condensable gases (radiolytic hydrogen and oxygen) which may
reduce heat transfer rates during extended periods of operation.

Emergency condenser (Siemens SWR-1000)

Emergency condensers are heat exchangers consisting of a parallel arrangement of

horizontal U-tubes between two common heads. The top header is connected via piping to
the reactor vessel steam space, while the lower header is connected to the reactor vessel
below the reactor vessel water level. The heat exchangers are located in a pool filled with
cold water. The emergency condensers and the reactor vessel thus form a system of
communicating pipes. At normal reactor water level, the emergency condensers are flooded
with cold, non-flowing water. If there is a drop in the reactor water level, the heat
exchanging surfaces are gradually uncovered and the incoming steam condenses on the cold
surfaces. The cold condensate is returned to the reactor vessel. The condenser pool is
located at a rather low level - with the return lines to the reactor vessel connected about 2m
above the top of the core. System function does not require opening of valves, but requires
a drop in the reactor water level.
Ejector-condenser

The principle of the ejector-condenser is based on the dynamic form of natural

convection utilizing inertial forces instead of gravity for fluid circulation. The process
develops in a loop combining an ejector specifically designed for dynamic natural convection
and a heat exchanger for condensation. Since the motive power does not depend on gravity,
heat can be rejected from a high elevation to a lower level. The condensate is recirculated
to the nuclear steam supply system.
Passive pressure pulse transmitters (Siemens SWR-1000)
Passive pressure pulse transmitters function in a similar manner as the emergency
condensers. The pressure generated in a heat exchanger secondary circuit is used to actuate
pilot or main valves.
(In-vessel) residual heat removal system

The in-vessel residual heat removal system consists of in-reactor vessel heat
exchangers located above the core, removing heat by natural convection to an external heat
sink. To integrate the heat exchangers the size of the reactor pressure vessel has to be
increased with respect to current PWR reactor vessel sizes.
16

Secondary condensing system

The secondary condensing system consists of a heat exchanger in a condensing pool

located above the steam generator and connected to the steam generator by piping, with heat
removal to the pool by natural convection. Condensate flow is returned to the steam
generator by gravity. It might be useful to add that the development of operating procedures
is interesting for transient conditions such as start-up and shutdown and is best pursued using
the larger facilities.
Computer model development and validation:

A number of papers dealt with the computer code modelling of components, separate
systems and integral systems and generally it appeared that after careful "tuning" the codes
have been successfully applied to the thermalhydraulics of passive systems - especially
problematic at low pressure. A point that should be kept in mind in these circumstances is
that phenomena and components should be modeled individually, and that these models should
be combined into the systems code rather than modifying the code by "tuning" or "biasing"
to obtain agreement between code predictions and experimental results.
Future needs in development and testing of safety systems for advanced water cooled
reactors:

Because of the intensive international cooperation in ALWR development and

certification several suggestions for future cooperative activities resulted.
The issues discussed at the meeting ranged from the fundamental choice of advanced
approaches to safety through the analysis and testing of selected approaches, to the results of
tests in direct support of the safety systems proposed in the designs of those reactors in the
most advanced stages of certification. As for future information exchange meetings in this
area, it would seem appropriate to focus on individual areas. Topics could include: national
advanced reactor programmes, methodology in selecting and testing advanced safety systems,
initiation and reliability of passive systems including analysis approaches to examine passive
system reliability, testing and analysis of component and system performance, quantification
of uncertainties in computer codes, testing to address new thermohydraulic phenomena which
are being incorporated into these codes and the results of test programmes in support of
advanced reactor designs undergoing regulatory review.

Another topic of interest would be accident management especially for accidents

beyond design basis including in-vessel retention of corium in the event of a core melt
accident by flooding the reactor cavity.
It was the general consensus that testing of heat removal safety systems at large scale
integral test facilities should continue to provide an extensive experience base in system
behaviour and data for validation of computer codes. Importantly, these integral facilities
could also be used to develop operating procedures for ALWRs.
While proof of predicted performance to satisfy safety requirements in support of
design certification is the major goal of such testing, more complete understanding of the
basic heat transfer phenomena, including the influence of non-condensable gases, would be
very worthwhile for thorough understanding of the phenomena. This is especially important
for passive heat transport systems which rely on small driving forces at low pressure thereby

17

requiring comprehensive testing to assure that conditions resulting in system initiation and
conditions affecting system reliability are thoroughly understood.
The question of extrapolability of the test results to larger sized plants which rely on
the same phenomena should be addressed, as such larger plants could bring economic
advantages. Existing facilities with some modification may be useful in examining
performance and systems interaction (a concern for extrapolation) for such larger sized plants.
Development of operating procedures is important for transient conditions such as start-up and
shutdown and is best pursued using the larger facilities. PRA could be used to examine the
need for additional testing with regard to potential systems interactions and the sequences to
be examined.

Regarding code validation, predicting the performance of passive components and

systems represents a new challenge to existing codes. New models are being developed in
the following areas in order to qualify the codes:
nonequilibrium mixtures involving subcooled and saturated water, saturated and
superheated steam
low velocity natural circulation
initiation of passive systems
effects of non-condensables on steam condensation
water circulation in pools
rapid condensation caused by interfacing steam and subcooled water.

Code benchmarking activities on an international level could be useful to assure proper

modelling of passive components and systems.

Peer reviews by international experts would be worthwhile on topics such as, for
example
ejector condenser and other passive safety components mentioned above
review of test results - both integral systems and safety components
computer code qualification and approaches for uncertainty analysis.
In summary, the meeting provided a timely forum for review of design, R&D and
testing of safety systems for advanced water cooled reactors, and for identification of key
activities for future international cooperation in advanced technology for water cooled
reactors.

18

OPENING ADDRESS
L. Noviello

ENEL-ATN, Rome, Italy

Three decades have passed from the inception of the peaceful use of nuclear energy. If I had
to characterize those three decades I would say that:
-The 1960s were the years during which most of the safety and the industrial rules were
established.
-The seventies were the years during which those rules were tested and a lot of experience
feedback was obtained.
-The eighties were the years during which many corrective actions were put in place.

In the sixties the safety authorities had to develop a lot of regulations to obtain the safety level
in the plants that they wanted.

In the seventies the industry, especially the utilities, had the heavy task of managing the
nuclear power plant projects in this complex licensing environment.

The first years of the 1980s were the years of a joint reflection: nuclear energy would have
not survived if the industrial and licensing environment did not stabilize and become more
predictable.
The successful French experience became the example to follow. The concepts of
construction in series, efficient project management and extensive pre-agreements with the
safety authorities became the prerequisites to any new initiative.

As a consequence in the first years of the 1980s the government research bodies, the utilities,
the vendors and the safety authorities jointly coordinated their roles and found new
agreements.
The 1980s were also the years of a large research effort to address the safety problems posed
by the TMI accident. The results of those researches have provided the basis for many
modifications to the operating plants and have caused an ample debate about the design basis
and the environmental impact objectives for the next generation of nuclear power plants.
The utilities are playing a major role in this redefinition. In the USA, in 1986, EPRI
launched a program to develop the Utilities Requirement Document (URD). At the same
time, in strong cooperation with DOE and the vendors, development was initiated on four
new plant designs.
A few years later, actually in a more difficult environment, a similar initiative has been
launched by a group of European Utilities, which have already issued the Revision A of the
European Utilities Requirements (EUR). In this case too, the development of some plant
designs is foreseen. The program for a large evolutionary design (EPR) is well underway,
a program for a passive PWR (EPP) has just started, while programs for passive BWRs are
under discussion.

19

Similar approaches have been taken in other countries, as for instance the Republic
of Korea.
For the first time all the conceivable accident scenarios have been considered in the
designs from their inception while, at the same time, the radioactivity release limit objectives
have been decreased. This evolution of the overall safety objectives of the next generation
of nuclear power plants (NPPs) has posed to the new designs very difficult, but also very
fascinating problems: the new plants had to have a reduced core melt probability and to
warrant the leak tightness of the containment for all the conceivable accident scenarios.
In a short time the easiest way was identified. Today, at the middle of the 1990s, we
can say that the next generation of NPPs will be characterized by the following features:
Reduced core melt probability, based on improved safety systems;
Lower sensitivity to the operator's actions, based on advanced man-machine interface;
Elimination by design of the most challenging sequences by the adoption of a primary
circuit depressurization system, of advanced reactor cavity design and of advanced
hydrogen control systems;
Elimination of the slow overpressurization of containment by the introduction of ad
hoc heat removal systems.

The development of many new systems has caused extensive test programs. During
this meeting we will have the opportunity to discuss the results of those programs and I am
sure that this TCM will help to improve the confidence on the new systems, in particular on
those based on the natural circulation.
The above are the areas where the improvements are already well established. There
are other areas where additional work is still needed to reach a consensus. I refer, in
particular to the in-vessel coolability of , but perhaps intend to improve it further or intend
to go from a design to a family of designs in terms of plant size.
the corium. Unfortunately not very much will be presented in this TCM on the effectiveness
of heat transfer through die vessel wall and on the related design features. I hope that the
next opportunity, that is the TCM on severe accidents to be held in Vienna in October, will
allow to look at those areas.

The last remark I want to make is on the possibility to extrapolate the developmental
work already done to different conditions and to different plant sizes. The investigation of
this subjecit during this TCM would be very beneficial to those who intend to maintain the
basic phlosophy of a specific design

20

STATUS AND PLANS OF DEVELOPMENT AND

TESTING PROGRAMMES

(Session I)
Chairman
L. NOV1ELLO
Italy

THE PROGRAMME OF ADVANCED LIGHT WATER REACTORS IN SPAIN

M. MALAVE
Consejo de Seguridad Nuclear,
C/Justo Dorado, Madrid,
Spain
Abstract

Spain's programme in advanced technologies for water cooled reactors has the objectives to
(1) improve performance of currently operating plants, (2) participate in establishment of user
requirements, and (3) be informed of advances to support future governmental choices.
Spain has nine water reactors in operation, such as two of them are of "3rd
generation" reactors: a Westinghouse PWR at Vandellos II, and a KWU PWR at Trillo I.
In addition, Spain has two modem BWR's at Valdecaballeros (GE BWR-6's) and two
modem PWR's at Lemoniz which are partially completed, but which were halted in
construction by the Energy Plan's moratorium of 1983 such as have been confirmed
recently by a new bill on December 31, 1994 that will make sweeping changes to the
country's national electricity supply sector (Ley de Ordenamiento del Sistema Electrico).
At the beginning of the nuclear programme in Spain, a major effort was made to invest
in domestic capabilities, including new machinery, new quality assurance practices,
and training of engineers. As a result, domestic participation in the civil work and
manufacture has risen from 24% in the first generation (1969 with Zorita start up) to 7880% in the third generation (1988 with Vandellos II). Spanish utilities intend to maintain
and increase this participation in the future.
Spain has a very active applying of advanced water reactor technologies aimed at
achieving performance from their operating reactors. Their R&D programmes are
supported by the 0.3% fund of electric utilities income, and include:

Improvements in control room, man machine interface and post accident

monitoring systems
Probabilistic safety assessments
Steam generator and secondary side improvements
BWR integrated corrosion programme
Life extension
Advanced fuel design and fabrication

- Source term.
Spain actually also participates in multinational programmes such as PHEBUS-FP,
HALDEN, RASPLAV, CAMP, Post-ACE, ISA, MCAP, Several code validations and
benchmarkings, CSAR Programme, IV Framework Programme of the EU, Assistance to
Eastern countries and others contributions.
Spain has an effort underway in the participation to design and build Evolutionary
and Passives LWR's plants as AP-600 of Westinghouse, ABWR and SBWR of GE
and SYSTEM 80+ of ABB-Combustion. Nothing at longer term of concerning Inherent
Safety Reactors is being made.
23

Spain is also participating in projects to define objectives and requirements of

advanced water reactors to be built in the near future in the European Union (EU). We
can extract from the 1991 Spanish Energy Plan as pattern of the institutional interest on
it the following:

"...Spain must participate in an active way in the consecutions of EU goals, with

the aim of reach the harmonization of rules in the nuclear field, giving place to the
opening of exportation of technology, equipments and engineering services and to
cause the colaboration in the EU for the development of advanced reactors clearly
european".

However, Spain is not only following closely the development of others advanced
water reactor designs in other countries and intends to be in a position to apply these,
but it is actively participating in the EPRI's ALWR Programme and others, both within
Spain and for possible export, should circumstances permit. On this regard, the
anticipated demand growth within Spain should call for the commissioning of new
facilities as from approximately year 2000. Two alternatives being considered to meet
this demand growth are: new gas thermal plants and developing and installing
advanced light water reactors. When a decision is made among these alternatives, it will
be included in the pertinent Energy Plan to be submitted by the Spanish Government for
approval by the Legislative Assembly. To explain the position of Spain I choosed seven
different topics to be analized:
Underlying Rationale For Advancements

Spain has five reasons for pursuing advanced light water reactors:
- It would be extremely difficult to maintain Spain's existing technological capacity in
the nuclear field without new nuclear power projects. Monitoring international
efforts to develop advanced light water reactors will assure that Spain's
technological capacity is up-to-date and available to benefit the safety and
operation of existing reactors as well as those to be built in the future.

- Improved water reactors would result into improved availability factors, decreased
operating costs, improved nuclear safety and radiological protection.

- Positive impact on Spanish R&D activities in areas such as robotics, expert

systems, welding, etc.
- Obtaining a high degree of technology transfer from multinational firms well
recognised in the nuclear field, with the ultimate objective of enabling the future
independent development in Spain of competitive designs of nuclear power plants.
- The eventual prospect of exporting the design and equipment of advanced nuclear
power plants to developing countries. This might include fuel cycle services as
well.
The Spain's Electricity Supply Industry has decided to pursue an Advanced Water
Reactor Programme. The Nuclear Safety Council (NSC) is getting involved in the

24

revision "C" of the European Utility Requirements -in whose redaction the Electricity

Sector is participating- and until present moment NSC has presented to a document
study of state-of-art of basically the most important designs of Future Reactors, enhace-

ments of design objectives, nuclear trends, new NRC's Licensing, European

Requirements and Conclusions.
Relevance to Operating Water Reactors

Should Spain decide to go forward with a programmme of Advanced Water Reactors,

initially using imported technology and later becoming more self-sufficient, they would
not see any conflict with continuing operation and reliance on existing operating light
water reactors, as indeed, the completion of four units is presently cancelled. There is a
very active programme of research and applying advanced technology to operating
plants that continue to assure these plants operate safely, reliably and economically.
The fact that advanced water reactors have even greater capability in the areas of

safety, reliability and aconomics, does not in any way detract from the conclusion that
those in operation are completely satisfactory in each one of these areas.

Approach to Obtaining Advancements

The most probable systems of future reactors are considered to be the Advanced
Pressurised Water Reactor (APWR), the Advanced Boiling Water Reactor (ABWR),
with a potential capacity of about 1000 MWe and Passives Plants (AP-600 and SBWR).
This is largely due to the vast experience obtained by the Spanish nuclear industry in
the design and construction of reactors from which these types have originated.
Selection of Plant Size

Should the National Energy Plan call for new nuclear power plants, Spain is interested
in advanced nuclear power plants. The size will be based on utility-specific
considerations such as demand growth, grid size and economics.
Prospects for Success

The main ingredient influencing the prospects for success of advanced water reactors in
Spain is the decisions made on electricity supply choices to be included in the National
Energy Plan. In the medium term, the possible further lowering of the price of crude oil,

and the dollar would favour continued reliance on gas fuel plants in Spain.
A further consideration in choosing among electricity supply choices is the existence of
various current nuclear projects (e.g. Valdecaballeros I and II and Lemoniz I and II)
authorized for construction and in different stages of progress, affected by the nuclear
moratorium of 1983 and finally cancelled.

In the event that Spain does undertake on a programme for future advanced water
reactors in Spain, the following are important ingredients for success:

25

1) Cooperation agreements between Spanish firms, or groups of firms and foreign

companies that are leaders in their field, with a view towards assimilating and
transferring technology.
2) The decisive support of the administration to these agreements, either directly or
through the official bodies in R&D and regulation.
3) Support of investing community.
Prospects for Cooperative Efforts
Spanish organisations are participating in various multinational nuclear research
programmes such as PHEBUS-FP, HALDEN, RASPLAV, CAMP, MCAP, CSAR
Programme, EPRI-ALWR Programme, etc, as we've seem before. With respect to
advanced reactors, there are some groups of companies all of them integrated in a
common company, called Agrupacion Electrica para el Desarrollo Tecnologico Nuclear
(DTN), that are participating in the development of advanced reactors, for possible
future application in Spain and also in the redaction of the European Utility
Requirements (EUR) and in the European Pressurized Passive Reactor (EPP Project)
wich main objective is to assess compliance with the EUR by the AP-600 concept and to
present alternatives. Utilities have a great interest in having any design of advanced
water reactors pre-licensed, so that licensing effort needed to construct the plant in
Spain would be limited to site specific aspects only.
Economics of Advanced Water Reactors
Spain indicates that economics will be a very important consideration in choosing
among electric supply options in Spain's next National Energy Plan. Emphasis for
nuclear power would be a reduction of construction periods, reduction of capital costs,
and stabilization of the licensing process.

26

PROGRESS W DESIGN, RESEARCH AND DEVELOPMENT

AND TESTING OF SAFETY SYSTEMS FOR THE KOREAN
NEXT GENERATION REACTOR

YOUNG SANG CHOI, BYONG SUP KIM

KEPCO Research Center, Taejeon,
Republic of Korea
Abstract

Korean Next Generation Reactor Development Project has been launched

to develop the advanced design which has the potential to become a safe,
economical and environmentally sound energy source in Korea.

The target of the project is to develop the whole plant design to

the sufficient level of detail to be built in the early 2000's.
This paper provides development program and design concept overview of
the KNGR
1. Introduction

Korea has launched next generation reactor development project, so called

the KNGR project, in 1992 to complete a detailed standard design by 2,000.

This project aims at developing the advanced design which has the
potential to become a safe, economical and environmentally sound energy
source for the early 2000's in Korea.
To develope the KNGR, an integrated project team has been setup to
incorporte the specialized institutes and companies of Korea nuclear
industry. KEPCO is the leading organization in this project and
Korea Atomic Energy Research Institute (KAERD is responsible for the
NSSS design, Korea Power Engineering Company (KOPEC) is for
architecture engineering and Korea Heavy Industry Company (KJflC) is for
component design.

27

For the basic research, the Center for Advanced Reactor Research (CARR),
an university association, is participating in this project and for the early
interaction of licensing issues, Korea Institute of Nuclear Safety (KINS),
regulatory agency is also participating in the project

This project consists of three phase in developmental activities from 1992

to 2000.
The phase I which has already been finished was a two year program from
1992 to 1994 and its major activities were the development of top tier
design requirements .and design concept for the next generation nuclear
plant The phase IE is 3 year program from 1995 to 1998 and its major
activities are the development of basic design for the licensing review. The
phase HI will start from 1997 to 2000 and its main activities are the
development of detailed standard design to the level which will enable the
accurate cost estimation and investment assurance.

2. Phase I Program Summary

The main focus of the first phase was to establish design concept of the
Korean next generation reactor.

Several studies have been performed to establish the design concept such
as the top-tier requirement development, comparative study on evolutionary
versus passive type plant and case studies for adapting hybrid design
concept, etc.

The top-tier requirements have been developed to specify the utility goal of
safety and economics that will give a guidence in the second phase design
activities.
The comparative study has been performed to select the reactor type
suitable and desirable for Korea by comprehensively reviewing the available
design options.
In parallel with the comparative study, case studies have been performed

to examine the possibility of uprating of passive type plants and the

feasibility for adapting passive design features in evolutionary plants.
28

In this series of studies, the KNGR team outlined design requirements and
the KNGR design concept for the future PWR in Korea.

3. Top-tier Design Requirement

As the primary design requirements, an evolutionary type pressurized water
reactor (PWR) with a capacity of 4000MWth and 60 year design life have
been selected Major considerations for this selection were the level of
maturity of technology, the economy of scale and the availability of new
nuclear sites, etc.
As the safety goal of the KNGR, the core damage and containment failure
frequencies are required to be less than 10~5/RY and 10~6/RY respectively,

which are, to our thought, in the world-wide consensus as seen in the

EPRI URD, EUR and IAEA recomendations. In addition to the above safety
goals, the limits to
lived Csisy at the
technical soundness
in the future and

the whole body dose and to the concentration of long

site boundary have been placed in order to provide
for possible reduction of emergency planning boundary
to prevent land contamination in case of postulated

accidents.
The economic goal of the KNGR is to accomplish cost advantage over the
coal by at least 20%. To achieve this goal, adoption of new technology as
well as the introduction of advanced management systems will be
necessary. Standardization policy is emphasized again and the improvement
of regulatory systems is continuously pursued The 48 months of
constuction period and maximization of modular construction are the subset
of the requirements, and also the reliability assurance program and
configuration management system are required to be applied systematically.

The design of the KNGR shall be standardized, and so site parameters

were determined to bound all potential nuclear sites in Korea. The seismic
value of 0.3g is specified as the safe shutdown earthquake(SSE).

29

4. Design Concept of KNGR

Nuclear Si

The KNGR core design consists of 241 fuel assemblies and utilizes
uranium dioxide fuel which is operated at a power density (95.9 KM/ Z)

lower than conventional large PWRs. The uranium-235 enrichment of the

fuel has been selected to achive 18 month or 24 month operation cycle.
The KNGR Reactor Coolant System (RCS) is configured in two separate
loops with the core power density of 95.9 KW/.C and water inventory of
322.1 i /MWe which is a 3.5% increase over existing reactors.
The primary coolant flow is set to give a conservative thermal margin
of approximately 15% to improve response to a variety of performance
transients.
The reactor vessel is fabricated by steel forging with controlled copper,
nickel, phosphorus, and sulfur content. By using these materials, the
change nilductility transition temperature (RT NDT) can be minimized
resulting in the longer life of a reactor vessel
Two U-tube, vertical steam generators provide the means to transfer heat
from the primary RCS to the secondary feed water. The resistance stress
corrosion cracking in the S/G tubes will be improved by using In-690
tube material.
A tube plugging margin of 10% will be required to provides a longer life
of steam generators.
The volume of pressurizer, which provide a means of maintaining the RCS
pressure by compensating volumetric fluctuation in RCS, will be increased
to 2400 ft3.

KNGR safety features are designed to follow the defence in-depth

approach.

The one feature of the defence-in-depth design is the reactor core

protection with the protection systems.
Protection systems include four independent trains in the Safety Injection
30

System, Shutdown Cooling System which can be used to backup the

containment spray pumps, and a four-train dedicated Emergency
Feedwater System.
Four independent trains of safety injection system discharge borated

water directly into the reactor vessel This direct vessel injection method
eliminates complicated interconnection pipes and ensures more water to
reach the core. The IRWST is employed for the suction to the safety
injection system which simplifies operation and increases reliability. The
safety injection function of shutdown cooling system has been removed
for simplicity. In stead, shutdown cooling pumps can be used for
cotainment spray when spray pumps fail. The design pressure of the SCS
is increased to minimize intersystem LOCA.

The other features of the defence-in-depth is the mitigation of the

consequences of a severe accident
This level of defense is accomplished through design features such as the
Safety Depressurization System, the Hydrogen Mitigation System, the
Cavity Flooding System, and a reactor cavity designed to mitigate the
consequences of the steam explosions and molten core-concrete
interaction.
The containment is of particular importance within the safety concept of
accident consequences. It is the final barrier in the defence-in-depth
concept
A double, cylinderical, and concrete containment design has been selected
for the KNGR taking into account the severe accident It consists of a
prestressed concrete inner wall designed to withstand post accident
pressure buildup with significant margin and a reinforced concrete outer
wall designed to withstand external hazards. The annulus between the
inner containment and the outer wall is maintained at a pressure below
ambient atmospheric pressure in order to collect any leak through the
inner containment of penetrations and filter them before release to the
environment

31

In addition, we have studied feasibility to apply passive design features

as a supplemental function to enhance the safety. Currently we are
implementing the conceptual design to apply the following passive
features to KNGR, :

Fluidic device in safety injection tank

Passive Cavity Flooding System
Secondary Condensing System
Catalytic hydrogen igniter

Man-Marhine Interface System (MMIS)

The KNGR MMIS employs digital I&C and advanced man-machine

interface technology to improve the operational reliability, plant safety, and
economy. The KNGR control room is implemented through the systematic
human factor engineering process and provides the primary interaction
with the plant via computer based interfaces, such as CRT, ELD. Large
wall mimic is employed in the MCR to promote the plant situation

awareness of operators. Functional information such as SPDS and task

information such as procedures are extensively provided to address the
information necessary for operators. A mockup of control room with plant
simulation models will be used to verifiy the suitability of the design.
Open architecture digital I&C technology such as PLC, mutiplexing, fiber
optic comunication is employed for safety and non-safety systems.
Diverse digital technologies are used for safety system and non-safety
system to prevent the common mode failure of digital I&C. Commercial
I&C equipment and software are evaluated and adopted according to
commercial item dedication criteria which will be approved by regulation.

General Arrangement
General arrangement of the KNGR has been developed based on the twin
unit concept and slide-along arrangement with common facilities such as
radwaste building. The schematic for twin units is depicted in figure 1.
The auxiliary and fuel buildings which accomodate the safety related
systems and components are located adjacent to the reactor building and
32

3 I

RAOWASTE TUNNEL
1 I

1 . CONTAINMENT BUILDING
2. A U X I L I A R Y BUILDING
3. TURBINE BUILDING
4. RADWASTE BUILDING
5. ACCESS CONTROL B U I L D I N G
6. FUEL HANDLING AREA
7. DIESEL GENERATOR
8- SWITCHGEAR BUILDING

Fig.

1 : Dual Unit Arrangement

they are seated on a common basemat with the reactor building in lieu of

separate basemats. common basemat will improve the resistance against

seismic event It also allows to reduction of the number of walls between
buildings so that rebar and form work cost can be reduced compared to
the separate basemat of current plant
As mentioned earlier, the KNGR adopts four train safety injection system.
The safety injection pumps are located in the aux. building near the
containment structure and each pump is located in each of four quadrants
surrounding containment This arrangement maximizes physical separation

of the pumps to provide protection against damage due to fire, sabotage,

and internal flooding. The vertical plan view is depicted in figure 2.
33

170'-0"

4 CONTAINMENT BUILDING
T
180'-0"
ACCESS
CONTROL
FACILITY

D/G BLDG.

-TRUCK BAY

3" ISOLATION
JOINT

ELECT. __
ROOM
^

TURB.
BLDG.

HOT UACHNE SHOP

EOUPKENT OCCOM AREA
PERSONNEL DCCCN AREA

D/G BLDG.

Fig. 2 : Vertical View of the KNGR

5. Phase n Program Outline

The objectives of the Phase n program are to complete engineering for

the KNGR in sufficient detail to assess safety, economic and
contructability and prepare for construction of KNGR from the 2000.

To achieve these objectives, the scope of the KNGR phase n program

will encompass the work necessary to provide design detail, along with
the work necessary to .support the success of the design activities.
The major activities in Phase n program can be categorized such as basic
studies, detailed URD development, regulatory research, information
magement system development and basic design activities.
34

The scope of the phase in design development is to complete the basic

design. It should provide the total power block information necessary to
make a decision in building the 1st unit This will accomplish the design
details comparable to the Standard Safety Analysis Report as in the U.S.A
for the design certification.

The degree of design finality is difficult to quantify but the KNGR team
outline that the basic design will meet around 20% of all the necessary
works to complete the plant design.
The KNGR design work will be conducted to assure the design of the
highest quality and to meets all applicable requirements through the project
quality program and implementation of periodic evaluation of the design.
This evaluation will be performed three times in the second phase.

Licensing stabilization is one of the main concerns to the development of

new design.
In the KNGR development, regulatory agency is directly involved in the
KNGR team to establish the licensing requirement as early as possible
to guide the designers.
Licensing interactions with the Korea Institute of Nuclear Safety (KINS) on
the issues identified in the preapplication review will be continued
As required by the top-tier requirement, the KNGR design team will
develop and operate Information Management System (IMS) in the Phase IL
For the design process, the IMS will be used as a single central, logical
data base of storing, securing and providing access to the wide variety of
information necessity associated plant design and to utilize configuration
management

In addition to the above activities, some supportive research and

development will be conducted to enhance in-house engineering capability,
design alternatives study and extension of top-tier requirement to detailed
requirements to support system, structure and component design in the
second phase.
Also, major advanced and passive safety features such as IRWST, Fluidic
Device, Secondary Condensing System will be tested to verify it's function
and the design.

35

6. Conclusion Remark

With the KNGR concept outlined in this paper, KEPCO and the Korean
nuclear industry with the support of government will jointly develop basic
and detailed design.
The KNGR design will meet the enhanced safety reo^iirements and
economic goal for future nuclear power plants in Korea, in particular, for
the protection of investment and prevention of severe accidents.

Nuclear energy will remain as essential energy source for the 21st century
in Korea.
Accordingly, KNGR will play animponant role by satisfying the
requirements of enhanced safety, improved economics and environmental
protection.
Finally, it is anticipated that the Korean nuclear industry improves
dramatically its basic nuclear technology through the completion of the
KNGR project

36

THE STATUS OF THE ALPHA-PROJECT

G. YADIGAROGLU, G. VARADI, J. DREIER,

F. DE CACHARD, B. SMITH S. GUNTAY,
Th. BANDURSKI
Paul Scherrer Institute Wurenlingen & Villigen,
Villigen , Switzerland
Abstract

A review of the ALPHA project is presented, including a summary of progress and current
status. The project comprises the experimental and analytical investigation of the long-term
decay heat removal phenomena from the containment of the next generation of "passive" Advanced Light Water Reactors. The effects of aerosols that may result from hypothetical severe
accidents are also considered. The construction of the major ALPHA experimental facilities,
PANDA, LINX-2 andAIDA, has been completed. First steady-state tests have been performed
on PANDA. The other facilities are now in their commissioning phases. Scaling studies have
guided the design of the experimental facilities. Several small-scale experiments and studies
have already produced valuable results which can be used to direct the experimental work, as
well as the design of the passive ALWRs.

1. INTRODUCTION
In 1991, PSI initiated the major new project ALPHA [1]. The central goal of this project is
the experimental and analytical investigation of the long-term decay heat removal from the
containment of the next generation of "passive" Advanced Light Water Reactors (ALWR). The
dynamic containment response of such systems, as well as containment phenomena, are
investigated. Two such passive ALWR systems are presently being designed and engineered in
the United States. The projects are led by the US Department of Energy (DOE) and the Electric
Power Research Institute (EPRI). Both projects are for reactors in the 600 MWe range. The
research and development (R&D) efforts will lead to the generic "certification" of the reactor
designs by the US regulatory authorities (the US Nuclear Regulatory Commission, US NRC).
These efforts are conducted with broad international cooperation and the participation of institutions from several American, European and Far-Eastern countries.
The first passive ALWR concept is a Pressurized Water Reactor (PWR), the AP-600,
while the second is the Simplified Boiling Water Reactor (SBWR) shown in Fig. 1. Both make
use of large passive systems for the transfer of decay heat following an assumed depressurization of the primary system, from the containment building to either evaporating water pools, or
to convectively air-cooled structures. These systems should be able to remove the decay heat
for at least three days without any intervention. Simple measures, such as the refilling of an
open water pool, may be needed after this period. Similar passive systems are also under consideration in other countries including Germany, Italy and Japan.
In all passive ALWRs, the energy removal from the reactor containment involves the
condensation of the steam produced by the evaporation of water in the core. This takes place in
the presence of some of the non-condensable gases that initially constituted the containment
atmosphere. It is the efficiency of this condensation process and the distribution of the gases in
the various containment volumes that determine overall containment behavior. These are the
main containment phenomena of interest for the ALPHA project. The project has been, so far,
37

ReactorPressureVessel
RPV

IMSL

Fig. 1: Schematic Representation of the SBWR Showing the Various Containment

Components.

directed to the investigation of the SBWR Passive Containment Cooling System (PCCS) and
related phenomena.
The ALPHA Project at PSI is conducted in collaboration with a large, truly international
R&D team. The main PSI partners in this team have been so far EPRI, the General Electric
Company and the University of California-Berkeley in the US, KEMA and ECN in the Netherlands, Toshiba in Japan, the DDE in Mexico, and ENEL, ENEA SET and Ansaldo in Italy.
The project includes four major items:
- PANDA is a large-scale, integral system test facility, presently configured to simulate
the containment of the SBWR on a volumetric scale of 1:25 and a height scale of 1:1.
- LINX is an experimental and analytical investigation of condensation and buoyancy-

driven mixing and stratification phenomena of importance to containment performance.

LENX-2 is the major facility related to this part of the project.
- AIDA is a program set up to investigate aerosol transport and behavior in the SBWR
Passive Containment Cooling System (PCCS) during core destructive accidents.

- Analytical model development, code validation, system analyses, and the extension and
application of three-dimensional Computational Fluid Dynamics (CFD) tools to large
scale mixing problems complement the experimental studies and complete this
program.
38

The purpose of this paper is to review progress to date and to outline certain recent
achievements.
Although the ALPHA results were initially expected to bring only "confirmatory" information for the generic "certification" of the SBWR, recent developments have made the first
series of experiments to be conducted in the PANDA facility the essential experimental element in the certification process. Thus, the results of the first series of PANDA experiments
will be formally submitted to the US NRC as part of the SBWR Design Certification Process.
As a consequence of the "formalization" of these tests, they are performed according to the US
NRC Quality Assurance procedure NQA-1.
The PANDA data will be used for the development and validation of models for computer
codes such as TRACG [2], RELAP5 [3], etc.
Other elements of the international program closely linked to ALPHA and the SBWR are:
- Single-tube condensation experiments at the University of California-Berkeley [4] and
at MIT [5].
- The smaller scale (1/400) containment integral test facility GIRAFFE, located at
Toshiba in Japan [6].
- The full-scale PCCS condenser qualification experiments (PANTHERS) performed by
SIETinItaly[7].
In addition to the work which is focused directly on the PCCS for the SBWR, there is a
great deal of other work already in progress at both small and large scale: for example, towards
understanding the long-term behavior of the SBWR Pressure Suppression Pool, including its
possible thermal stratification, and potential heat-up and pressurization of the Wetwell. Certain
phenomena of generic interest to all passive containment concepts are also being investigated,
and tests directly related to the development of alternative concepts are also being conducted.
2. STATE OF THE ALPHA PROJECT AT PSI

The major project activities within the period 1991-94 have concentrated on the design,
scaling and construction of the three major experimental facilities, namely PANDA, LINX-2
and AIDA. The detailed conceptual design of all three was completed in 1992, together with
ordering of the major components. Significant work has also been carried out on the development of the required instrumentation and data acquisition systems. Construction of all the
facilities is now complete, and commissioning is underway or completed. First steady-state
tests have been performed on PANDA. Tests in the other major facilities will start in 1995.

In addition, small-scale experiments and numerous analyses were conducted within the
LINX framework (LINX-1 and LBSFX-1.5) to better understand basic phenomena, and to provide preliminary experimental data for the development of computational models for thermal
mixing in open pools: specifically, plume development, mixing and stratification [8,9,10]. To
understand SBWR system behavior and to support both facility design and the definition of
instrumentation, extensive computer calculations have been performed using the TRACG,
FLOW3D and ASTEC codes [11,12].
3. THE PANDA FACILITY
The PANDA general experimental philosophy, facility design, scaling, and measurement
concepts were defined in early 1991 [13]. On-site construction began in 1993 with the delivery
of the major PANDA components. The facility is now complete and fully instrumented. The

39

instrumentation includes numerous temperature sensors (including "floating thermocouples"

capable of measuring the temperature of the surface of water pools), pressure as well as differential pressure and water level difference sensors, flow rate measurements (over wide
ranges), non-condensable fraction sensors (oxygen probes), phase (liquid or gas) detectors,
electric power measurement, and sensors detecting the positions of valves; a total of some 620
channels. The computer-based data acquisition system (DAS) is capable of sampling all channels continuously with a frequency of 0.5 Hz. For short periods of time, sampling with a "burst"
frequency of 5 Hz is also available. The DAS displays the data in a variety of "screens" representing schematics of various parts of the facility [14]. The time histories (trends) of selected
channels can be displayed at will. The facility is operated and controlled remotely and interactively by a computer-screen-based system.
The very first series of PANDA experiments have been steady-state tests of PCCS condenser performance. These have been counterpart tests to similar tests conducted at the
PANTHERS facility in Italy [7]. Extensive facility characterization tests will follow: the heat
losses from the facility, as well as the actual pressure-drop-flowrate characteristics of the various lines will be obtained. These are needed for the accurate description of the facility by
computer codes. The actual transient system behavior tests will follow.
In relation to the SBWR certification effort, the PANDA facility was primarily designed
to examine system response during the long-term containment cooling period. The overall
objectives of the transient PANDA tests are to demonstrate that:
- The containment long-term cooling performance is similar in a larger scale system
to that previously demonstrated with the GIRAFFE tests.
- Any non-uniform spatial distributions in the Drywell or Wetwell do not create
significant adverse effects on the performance of the PCCS.
- There are no adverse effects associated with multi-unit PCCS operation and
interactions with other reactor systems.
- The tests will also extend the database available for computer code qualification.
The initial test series at PANDA will include two Main Steam Line Break (MSLB) tests.
One test will be similar to a GIRAFFE MSLB test with uniform Drywell conditions, while a
second is planned in a manner maximizing the influence of Drywell asymmetries on the operation of the PCCS condensers. Uniform and asymmetric Drywell conditions can be created in
PANDA via the capability to vary the fraction of steam flow which is injected into each of the
interconnected Drywell vessels, Fig. 2. The steam condensing capacity directly connected to
each vessel (i.e., the number of condenser units) can also be varied. Such tests and several
systematic variations of the experimental conditions will help identify any scale and multidimensionality effects that may not have been present in the smaller scale facilities. The test
matrix for further tests is dictated by the needs of certification and code assessment and is still
evolving. Steam and/or non-condensables can be injected at various locations hi the Drywell
vessels to study mixing phenomena and to provide envelope information for the corresponding
SBWR conditions.
3.1. Scaling Considerations
A rigorous scaling study [15] covering all the SBWR related tests describes the scaling
rationale and details of the PANDA facility. Other documents cover particular aspects of scaling [10,16,17,18].

40

The SBWR containment is particularly complex and thermal-hydraulically coupled to the

primary system. In addition to the complexity of the usual BWR pressure suppression system
and its various components, such as the main vents, one must now also consider the operation
of the particular PCCS system and its components. Both "top-down" and "bottom-up" scaling
considerations [19] and criteria were developed.
According to the "top-down" approach, general scaling criteria are derived by considering
the processes controlling the state of classes of containment sub-systems (e.g., containment
volumes, pipes, etc.). Close examination of the specific phenomena governing the operation of
certain system components (e.g., vents immersed in the Pressure Suppression Pool of the SBWR) leads to "bottom-up" scaling rules.

3.1.1. Top-Down Scaling

Generic scaling criteria for thermal-hydraulic facilities, such as those proposed by ISHn
and KATAOKA [20], are not specific to the combined thermodynamic and thermal-hydraulic
phenomena taking place inside containments. Thus, specific scaling criteria for the design of
facilities simulating the dynamic operation of BWR containments such as PANDA were derived [15]. For example, mass and energy transfers take place between containment volumes
through their junctions. Heat may also be exchanged between volumes by conduction through
the structures connecting them. These exchanges lead to changes in the thermodynamic condition of the various volumes; this, in particular, leads to changes of the volume pressures. The
junction flows (flows between volumes) are driven by the pressure differences between
volumes. Thus, the thermodynamic behavior of the system (essentially, its pressure history) is
linked to its thermal-hydraulic behavior (the flows of mass and energy between volumes).

Fig. 2: Schematic Representation

of the PANDA Facility.

41

The "top-down", generic scaling criteria were derived by considering generic processes,
including:
(1)
The effects of the addition of heat and mass to a gas or liquid volume (namely, the
resulting rates of change of the pressure).
(2)
The rates of phase change at interfaces such as pool surfaces.
(3)
Hows of mass between volumes.
Prototypical fluids, i.e. those pertaining to the actual reactor system, under prototypical
thermodynamic conditions, are used in PANDA. The fact that the fluids are expected to be in
similar thermodynamic states, and have similar composition in the prototype and the model,
can be used to simplify the analysis and scaling of the facility. The first two processes listed
above (1 and 2) confirm the validity of the (familiar) scaling of all the following variables with
the "system scale", R:
(power)R = (volume)R = (horizontal area in volume)R= (mass flow rate)R = R
where the subscript R denotes the ratio between the corresponding scales of prototype and
model. Although other choices are also possible, the system scale R can be defined as the ratio
of prototype to test facility power input. For PANDA, R = 25. A time scale of 1:1 between
prototype and model has been adopted for the presently planned PANDA tests; however, this
is not a necessity. Under certain conditions, the choice of a scale for the volumes different from
the system scale R will lead to accelerated (or decelerated) tests in time.
Process (3) leads to the determination of the pressure drops and of the hydraulic characteristics of the junctions between volumes. In the BWRs, certain pressure drops and the
corresponding junction flows are controlled by the submergence depth of vents in the Pressure
Suppression Pool. The analyses of these processes justify the choice of 1:1 scaling for the
vertical heights in general and for the submergence depths in particular [15].
The pressure evolution resulting from the thermodynamics of the system and the pressure
drops between volumes must clearly scale in an identical fashion. Considering the fact that
prototypical fluids are used, this requirement links the properties of the fluid (in particular the
latent heat and the specific volumes of water and steam) to the pressure differences between
volumes (and to the submergence depths of vents), resulting in 1:1 scaling for pressure drops.
Thus, the above considerations result in:
1:1 scaling for pressure differences, elevations and submergences.
This scaling rule determines the pipe diameters, lengths and hydraulic resistances, and the
transit times between volumes. These transit times should, in principle, have the same (1:1)
time scale as the inherent time constants of the system considered in the analysis of process (1).
This matching cannot be perfect, but is shown not to be important [15].
The criteria derived are combined to arrive at general scaling laws for the PANDA model
of the prototype SBWR. Several non-dimensional numbers, three time scales and certain geometric ratios must be matched.
The three time scales produced by the analysis are the scales for the rates of volume fill, of
inertial effects, and of pipe transfers. Clearly, the systems considered here are made of large
volumes connected by piping of much lesser volumetric capacity. The pressure drops between
these volumes are not expected to be dominated by inertial effects. Thus, the inertia and transit
times, which are of the same order of magnitude, are much smaller than the volume fill times.
Consequently, the important time scale that must be considered in scaling is the volume fill
time scale. The other two time scales are clearly of lesser importance. In relation to this con42

elusion, the lengths of piping connecting containment volumes and the velocities in these pipes
do not have to be scaled exactly. Usually (and fortunately), the total pressure drops in the piping are dominated by local losses, so that the total pressure drops hi the scaled facilities end up
being somewhat smaller. They can therefore be matched by introducing additional losses by
local orificing.
3.1.2. Scaling of Specific Phenomena - Bottom-Up Approach

Bottom-up scaling for phenomena that were selected as being of particular importance by
a Phenomena Identification and Ranking Table (PIRT) exercise were examined in detail to
arrive at then: proper simulation in the PANDA facility [15]. Several documents cover in detail
such particular aspects of scaling [10, 16,17,18,21], as already noted.
The scaling of thermal plumes, mixing and stratification phenomena in the pool, as well as
in the Drywell volume, heat and mass transfers at liquid-gas interfaces, the heat capacity of
containment structures, and heat losses were examined in detail in [15]. Of particular importance is the scaling of the various vents, discharging mixtures of steam and non-condensable
gases into the Pressure Suppression Pool. This is important in relation to the possibility of
steam "bypassing" this pool and entering directly into the Wetwell gas space. Heat and mass
transfer in the PCCS condensers must also be properly scaled, considering both condensation
inside the tubes and heat transfer on the secondary pool side [21]. The latter may be affected by
any induced natural circulation.

4. THE LINX PROGRAM

The LINX Program (Large-Scale Investigation of Natural Circulation, Condensation and
Mixing) aims at a better understanding of the most important physical processes taking place
In passive ALWR containments. Small-scale experiments (LINX-1 and LINX-1.5) addressing
certain issues encountered hi the SBWR have already been conducted [8,9,10]. The main effort
in this program includes medium-scale, highly instrumented experiments in the LINX-2 facility which are described below. These will look at natural circulation, mixing and condensation
phenomena in pressure suppression pools and containment volumes in the presence of noncondensable gases. This work will also support the application of Computational Fluid Dynamics (CFD) tools for single- and multi-phase flow to mixing and natural circulation
problems.
4.1. The LINX-2 Facility

The LINX-2 facility has been designed to study condensation and mixing phenomena encountered hi passive ALWRs, such as direct-contact and surface condensation in the presence
of non-condensables and pool thermal mixing induced by single and two-phase plumes. The
experiments are to be conducted at a fairly large scale and under prototypical pressure and
temperature conditions. Some issues relevant to the SBWR and to a European version of the
AP-600 are being addressed hi the present phase of this program.
The facility, Fig. 3, consists of a pressure vessel (rated at 10 bar and 250C), steam and
nitrogen (or air) supply lines, and a water conditioning loop. The vessel is very carefully insulated (35 cm insulation thickness) to minimize heat losses and allow the performance of
accurate heat balances. Heat losses are monitored using thermocouples on the vessel wall and
penetrations, and within the insulation itself. The vessel pressure is regulated. Constant (regu43

lated) steam and nitrogen flow rates can be injected from either the bottom or the top of the
vessel. The steam and nitrogen flow-rate ranges can be adjusted from 10 to 120 kg/h ( 1.3%
of the Measured Value) and 0.1 to 75 kg/h ( 3% MV), respectively. The gas injection temperature (up to 180C) is also regulated. The water conditioning loop is a multi-purpose heating
and cooling system. Outside the vessel, either heating or cooling power is provided by steam
or cold water, respectively. Inside the vessel, water can be circulated at constant (regulated)

Pressure regulation vent

Fig. 3: Schematic of the

LINX-2 Facility.

Pressure Vessel
r

Windows for
visualisation

Heating and
Cooling Exchangers

it

hxi-

Gas heater

flow rate in a closed loop, and can be brought to the required temperature. This flow rate may
be set between 0 and 10 m3/h ( 40 1/h) and the water temperature between 15 and 180C
(1C). The maximal heating and cooling powers are about 140 and 120 kW, respectively.
In the SBWR, following primary system depressurization, a steam/air mixture flows into
the Pressure Suppression Pool. The efficiency of the condensation and mixing processes there
affects the Wetwell, and consequently the whole containment pressure level. The LINX-2 facility will be used to study PCCS venting in the Pressure Suppression Pools and the threedimensional velocity and temperature fields created in the pool by the two-phase flow plume
emerging from the vent. The experiments will give a better understanding of the physical phenomena and allow the development of ad-hoc models. Three-dimensional, multi-phase,
multi-component CFD models, as well as simpler models intended for the system codes, will
be proposed and assessed against the results and/or further developed.
44

The three-dimensional temperature field within the vessel may be investigated, without
significant disturbance, using up to 273 thermocouples attached to 1 mm wires by miniature
pinch-screw clips. A number of wires are mounted on a swinging arm, which allows radial
scanning of the temperature field during an experiment. Six high-precision sensors are also
available inside the vessel. Impeller-type anemometers will be used for velocity measurements,
and double optical probes for two-phase flow investigations. A customized data acquisition
system, together with a PC-based, multi-tasking software package, will provide the data acquisition needs.
A cooperation agreement was concluded in 1994 between ENEL (the Italian national
electric utility) and PSI on ALWR thermal-hydraulics; a first common project has been defined
and is underway. This concerns a PCCS proposed as a European alternative to the Westinghouse AP-600 passive containment cooling design. The purpose is to substitute a double
concrete containment for the original AP-600 metallic envelope. Due to the high thermal resistance of concrete, the proposed PCCS needs a steam condensing heat exchanger placed
inside the containment and an intermediate loop extracting the heat from the containment and
taking it to an external atmospheric heat exchanger. The intermediate loop is a water-steam
thermosiphon. The internal heat exchanger is a compact, finned tube bundle; non-condensable
gases and steam circulate by natural draft. The external heat exchange takes place in a hybrid
cooling tower, combining a water pool with a natural draft cooling tower.
Internal heat exchanger and intermediate loop mock-ups of this system will be tested in the
LINX-2 facility. The use of finned tubes in a condensing heat exchanger allows the device size
to be reduced. Accumulation of non-condensable gases and condensate in the spaces between
fins could, however, take place, thus degrading heat transfer. Predictive computational models
of this system are being developed and the experimental results will be used to assess and
improve them.
For the passive PWR-related tests, in addition to the components depicted in Fig. 3, the
facility is equipped with a closed gas (steam/air) recirculating loop. The recirculated gas flow
rate (up to 0.5 kg/s) and its humidity are measured.
5. THE AIDA PROGRAM
The AIDA program examines the behavior of the PCCS system when aerosols are present
in the containment, following a hypothetical severe accident. Under such conditions, aerosols
present in the Drywell will be entrained into the PCCS condensers. This may degrade condenser performance. The condensers may, however, also act as scrubbers and help reduce the
aerosol concentration hi the Dryweil.

The possible formation of an aerosol layer at the condenser tube entrance (reduction of
tube-inlet free flow area) and inside the tubes (reduction of the tube cross-section) may cause
a new flow distribution among the tubes; it will also affect their heat transfer performance.
Flow redistribution among the tubes may change the heat removal characteristics of the entire
PCCS system; such changes may appear as a result of a reduction in the number of tubes that
are properly active, leading to a situation in which some tubes continuously receive more steam
than they can condense. Thus, it is clear that, under hypothetical severe accident conditions, the
long-term pressure history of the SBWR containment depends on the behavior of the PCCS
units in the presence of aerosols. Consequently, the goals of the AIDA program are to:
- Experimentally determine the degree of PCC condensation degradation in the presence
of aerosols.
45

Investigate aerosol behavior in the upper header of the condenser units.

Investigate aerosol behavior under strong condensation in condenser tubes.
Investigate the aerosol retention capability of the condenser units.
Provide the basis for the development of a physical model for aerosol behavior in the
condensers and its effects on the thermal-hydraulics within the PCC units.

5.1. The ATOA FacUity

A versatile, multiple-purpose aerosol testing facility was constructed at PSI [22] and is
also being used for the ADDA tests. Two plasma torches, two reaction chambers, a mixing tank,
and steam and non-condensable supply systems are the main components of the facility. The
system is computer controlled and can produce aerosol particles at a desired steady mass flow
rate and concentration. The particles are entrained by a carrier gas, composed of steam and
non-condensable gas at a desired composition. The plasma torches used for aerosol generation
produce aerosol mixtures of up to three components (Csl, CsOH and MnO or SnOi) with a
maximum concentration of 20 g/m3. Experiments can be performed with the following boundary conditions:
- steam fraction ranging from 0 to 95%,
- steam flow rate up to 250 kg/hr,
- non-condensable gas flow rate up to 280 kg/hr,
- system pressure up to 5 bar..
For the AIDA experiments, a slice of the SBWR PCCS condenser unit containing eight
full-height tubes and connected to full-diameter lower and upper headers, was constructed.
Both glass and steel tubes can be tested. The glass tubes are intended mainly for visualization
of the phenomena. The tubes are heavily instrumented with thermocouples to measure the gas
and wall temperatures, and estimate the heat flux across the tube wall. The secondary coolant
channel surrounding the tubes is made of glass to facilitate visualization of the aerosol deposition and transport phenomena within the glass tubes, Fig. 4. The secondary cooling water,
flowing upward at a desired small velocity and at a predefined temperature (up to 80C), can
properly simulate the heat transfer conditions expected in the prototype.
The condensed water is collected in a Condensate Tank simulating the Gravity-Driven
Cooling System (GDCS) pool, Fig. 5. The non-condensable gas and uncondensed steam flow
into a Scrubber Tank that condenses the steam and Scrubs the aerosol particles carried out of
the condenser unit. The condensate that is produced in the Scrubbing Tank is collected in a
second Collection Tank. The Scrubbing and Collection Tanks simulate the behavior of the
Wetwell. The pressures in the Condensate and Collection Tanks are regulated to obtain the
pressures expected in the Drywell and Wetwell.
The facility is instrumented with several devices to provide information on: a) energy
transfer and steam mass balance related to steam condensation in the condenser, as well as in
other parts of the experiment, and b) aerosol mass balances. The instruments provide on-line
data that is displayed via a special data acquisition system on a computer screen to continuously monitor the system response. The aerosol instrumentation comprises: a) off-line devices
like filters, impactors and deposition coupons, and b) on-line devices like photometers and ion
detectors.

46

5.2. Aerosol Particle Trajectories

Certain knowledge about the behavior of the aerosols in parts of the PCCS system can be
obtained by CFD calculations. Such information can later be integrated in aerosol behavior
models, including their deposition and re-entrainment behavior. As a first step in this direction,
aerosol tracking calculations with the ASTEC and FLOW3D codes were performed to examine
aerosol behavior in the upper header of the ADDA condenser [12].
An aerosol-tracking model has been specially written for this application and interfaced
with the ASTEC and FLOW3D codes. The calculations were conducted for the expected typical aerosol and flow conditions; they considered the deposition of the aerosols on a deflector
plate placed below the steam-inlet tube in the upper header of the condenser unit, Fig. 6. The
purpose of this deflector plate is to distribute the flow to the condenser tubes as well as possible; both the SBWR prototype condenser and the AIDA mockup have such plates installed.
The first calculations conducted assumed, in a simplistic manner, that all particles reaching a surface will be deposited and remain on the surface. With such a deposition "model", the
results show that 98% of the aerosols which enter the drum with the inlet steam jet are deposited on the deflector plate. The remainder circulate in the upper part of the drum and finally
deposit on its inner surface. To go a step further in the simulation and consider re-entrainment
of the aerosols from the impact plate, calculations were performed assuming that particles
could "spill-over" from the edges of the impact plate; this constitutes a very crude reentrainment model. It was found that any particles that are subsequently dislodged from the
plate have a 15% probability of entering one of the condenser tubes. The rest are deposited,
more or less uniformly, in the upper drum. Clearly, the first steps taken in this direction must
be supplemented and completed with more realistic aerosol deposition and re-entrainment
models. The possibility of re-entrainment of aerosol agglomerations will also have to be
addressed.
6. PANDA PRE-TEST CALCULATIONS
As part of the SBWR certification process, calculations to predict the outcomes of the
experiments that will be conducted later in the PANDA facility are being performed and formally submitted to the US NRC. These "blind" pre-test calculations are done in collaboration
with other SBWR international partners using the official SBWR safety analysis tool (the
TRACG code). The pre-test calculations will be used as part of the validation data base for the
application of TRACG to the SBWR. Following the experiments, post-test calculations will be
performed, as needed, to resolve any outstanding issues.
Pre-test calculations for the steady-state PANDA PCCS condenser tests have been completed and submitted akeady. During 1994, the TRACG input model of PANDA was updated
in a formal manner to satisfy the Quality Assurance (QA) requirements needed for such formal
submissions. This included an independent design review of the model, an independent verification of all the numbers used against the facility data base (the as-built drawings), formal
exchanges of comments and replies, and formal documentation of the entire procedure.
Two fully verified models were produced: The first is a partial model of the PCCS system
used to predict the steady-state PANDA condenser performance tests; the second is a model of
the entire PANDA facility, with initial and boundary conditions applicable to the first PANDA
system test M3. The initial and boundary conditions for PANDA test M3 are based upon
TRACG calculations of the state of the SBWR containment one hour after reactor scram. A
counterpart GIRAFFE test will be carried through using the same initial boundary conditions.
47

Steam + Gas
from Containment
"Drywell"
Aerosol Generation Facility

Condenser
"PCC"
"PCC-Pool"
Water Cooling

Suppression Pool Collection

"Wetwell"
Scrubber Tank

Tank

"GDCS-Pool"
Condensate Tank

Fig. 4: The AIDA PCCS Mockup.

Particle growth
due to
steam condensation,
hygroscopicity

Steam +
Gas +
Particles

Steam
condensation

Cooling water
outlet

Cooling water

Particle
deposition

X.

^ Condensate

Gas

Fig. 5: The AID A Facility Schematic.

48

(Water)

Deflector Plate

Steam +
Gas +
Particles

Particle
trajectories

Fig. 6: Calculated Paths of Aerosol Particles in the Upper Header of the AJDDA Condenser.
Top: most of the aerosols entering the header impact the deflector plate.
Bottom: aerosols are allowed to "spill over" from deflector plate.
Facility characterization tests will be performed, e.g. heat loss and pressure drop tests, and the
models will be further improved using the information obtained from these tests.
The partners of this international collaboration have so far included PSI, which had the
technical lead and produced the TRACG PANDA model and transient calculations, the IIE in
Mexico, which provided model verification; KEMA in the Netherlands, which performed the
condenser performance calculations, and the General Electric Company (GE) that provided
overall coordination and the "Design Record File".
7. OTHER ACHIEVEMENTS
In addition to the main project activities outlined above, several scientific achievements
that helped better understanding of system behavior and containment phenomena also took
place. These include:
ALPHA staff members have made significant contributions to the understanding of the
modes of passive operation of the SBWR. Contributions include, for example, detailed modeling of the condensation of steam in the presence of non-condensables in the PCCS units used
to remove the decay heat from the containment [21].
An example of sophisticated but small-scale experiments and analytical modelling are the
study of plumes in small-scale water pools (LINX-1) [8,9] and of two-phase flows of mixtures
of air and steam bubbles in water (LINX-1.5) [10].
49

Development and testing of instrumentation has also taken place, triggered by the needs of
the PANDA and LINX experiments (e.g., floating thermocouples for water surface temperature measurement, testing of non-condensable fraction sensors [23], etc.).

8. CONCLUSIONS
In several countries, there is a general move towards the introduction of more passive
systems for emergency core cooling and containment decay heat removal in future reactors. In
the US, this trend has materialized with the certification effort related to the AP-600 and the
SBWR. In Europe, certain recent concepts for new BWRs and PWRs also include long-term
passive decay heat removal systems. In Japan, Canada, and other countries there is interest in
adding such passive systems to either existing reactor designs or to new ones. The ALPHA
project is situated in this international framework. Its long-term objectives are to contribute at
the forefront of this research area worldwide, but in Europe in particular.
As stated, future reactor systems are likely to include some form of passive containment
cooling systems. Although the designs of such systems may vary from one reactor concept to
another, there is a need to provide basic scientific understanding of their performance under
fairly large-scale prototypical conditions. The PANDA and LINX-2 facilities provide the ideal
environment for long-term international collaboration in this area. For example, the LINX-2
facility will be used in the near future to test, in collaboration with the Italian ENEL, the design
of an advanced PWR containment building condenser.
Although the present PANDA experiments constituted initially "confirmatory" research,
the data that they will deliver has now become an essential part of the "certification" process
for the SBWR.

ACKNOWLEDGMENTS
All the members of the ALPHA team in the Thermal-Hydraulics Laboratory and of the
ABDA team at the Safety and Accident Research Laboratory have contributed to the progress
of the ALPHA project and, therefore, indirectly to this article. Their particular contributions
are too numerous to mention here but are gratefully acknowledged by the authors.
The ALPHA project receives financial support from the Nuclear Power Committee of the
Swiss Utilities (UAK), the (former) Swiss National Energy Research Foundation (NEFF) and
the General Electric Company. Particular contributions from international collaborations are
mentioned in the text.
The authors wish to acknowledge the Toshiba Corporation for permission to use data from
the GIRAFFE facility.
REFERENCES

[1] CODDINGTON P., HUGGENBERGER M., GUNTAY S., DREIER J., FISCHER O.,
VARADIG. AND YADIGAROGLU G., "ALPHA: The Long-Term Decay Heat
Removal and Aerosol Retention Programme", 5th International Topical Meeting on
Nuclear Reactor Thermal Hydraulics (NURETH-5), pp 203-211, Salt Lake City, USA,
Sept. 1992.
[2] ANDERSEN J.G.M., ET AL., (1993b), "TRACG Qualification", Licensing Topical
Report, NEDE-32177P, Class 3 (February 1993).
50

[3] CARLSON K.E., ET AL., "RELAP5/MOD3 Code Manual. Vol. I: Code Structure,
System Models, and Solution Methods", NUREG/CR-5535, EGG-2596 (June 1990).
[4] VIEROW K.M. AND SCHROCK V. E.,"Condensation in a Natural Circulation Loop with
Non-Condensable Gases, Part I Heat Transfer", Proc. Int. Conf. Multiphase Flows,
Tsukuba, Japan, September 1991.
[5] SEDDIQUE M., GOLAY M.W. AND KAZMI M.S., "Local Heat Transfer Coefficients
for Forced Convection Condensation of Steam in a Vertical Tube in the Presence of a
Non-condensable Gas", Nucl. Technol. 102 (1993) 386.
[6] YOKOBORIS., NAGASAKA H., TOBIMATSU T., "System Response Test of Isolation
Condenser Applied as a Passive Containment Cooling System", 1st JSME/ASEM Joint
International Conference on Nuclear Engineering (ICONE-I) Nov. 1991 Tokyo.
[7] BOTTI S. ET AL., "Tests on Full Scale Prototypical Condensers for SBWR Application",
European Two-Phase Flow Group Meeting, SET, June 6-8, 1994.
[8] SMITH B.L., DURY T.V., HUGGENBERGER M. AND NOTHIGER H., "Analysis of
Single-Phase Mixing Experiments in Open Pools", ASME Winter Annual Meeting, HTDVol. 209, pp 101-109, Anaheim, CA, USA, Nov. 1992.
[9] HUGGENBERGER M., NOTHIGER H., SMITH B.L. AND DURY T.V., "Single-Phase
Mixing in Open Pools", NURETH-5, pp 547-555, Salt Lake City, USA, Sept. 1992.
[10] CODDINGTON P. AND ANDREANI M. "SBWR PCCS Vent Phenomena and
Suppression Pool Mixing", paper submitted to NURETH-7, 10-15 September 1995,
Saratoga Springs, NY, USA.
[11] CODDINGTON P., "A TRACG Investigation of the Proposed Long-Term Decay Heat
Removal Facility PANDA at the Paul Scherrer Institute", NURETH-5, pp 192-202, Salt
Lake City, USA Sept. 1992.
[12] DURY T.V., "Pre-Test Thermal-Hydraulic Analysis in Support of the AIDA Testing
Design Using the ASTEC Code", PSI internal report TM-42-94-10, ALPHA-409.
[13] HUGGENBERGER M., "PANDA Experimental Facility Conceptual Design", PSI
internal report AN-42-91-09, ALPHA-105.
[14] DREffiR J., "PANDA-Versuchsanlage; Pflichtenheft fur Messung, Steuerung und
Regelung", PSI internal report TM-42-92-21, ALPHA- 217.
[15] YADIGAROGLU G., "Scaling of the SBWR Related Tests", GE Nuclear Energy report
NEDC-32288 (July 1994).
[16] CODDINGTON P., "A Procedure for Calculating Two-Phase Plume Entrainment and
Temperature Rise as Applied to LINX and the SBWR", PSI internal report TM-42-94-01,
ALPHA-401.
[17] CODDINGTON P., "A Review of the SBWR PCCS Venting Phenomena", PSI internal
report TM-42-94-02, ALPHA-402.
[18] ANDREANI M., "Study of the Horizontal Spreading of Rising Two-Phase Plumes and its
Effects on Pool Mixing", PSI internal report TM-42-94-05, ALPHA-404.
[19] ZUBER N., (1991), "Hierarchical, Two-Tiered Scaling Analysis" Appendix D to "An
Integrated Structure and Scaling Methodology for Severe Accident Technical Issue
Resolution", Nuclear Regulatory Commission Report NUREG/CR-5809, EGG-2659
(November 1991).
[20] ISHII M. and KATAOKA I., (1983), "Similarity Analysis and Scaling Criteria for
LWR's Under Single-Phase and Two-Phase Natural Circulation," NUREG/CR-3267
(ANL-83-32).

51

[21] MEffiR M, "SBWR-PCCS Numerical Integration Programme and Test Results", PSI in
ternal report TM-42-94-08, ALPHA-406.
[22] S. GUNTAY, G. VARADI, J. DREIER, "ALPHA- The Long-Term Passive Decay Heat
Removal and Aerosol Retention Program", IAEA Advisery Group Meeting on the
Technical Feasibility and Reliability of Passive Safety Systems, November 21-24,1994,
Jiilich, Germany.
[23] LOMPERSKI S., "High Temperature and Pressure Humidity Measurement Using an
Oxygen Sensor", PSI internal report TM-42-94-03, ALPHA-403.

52

PREDICTED PERFORMANCE AND ANALYSIS OF ADVANCED

WATER COOLED REACTOR DESIGNS
(Session II)
Chairman
G.D. MCPHERSON
United States of America

EVALUATION OF THE DESIGN OPTIONS FOR FUTURE

POWER PLANTS: IDENTIFICATION OF THE SAFETY RELATED
CRITERIA AND EVALUATION OF THE DECAY HEAT
REMOVAL OPTIONS

G.L. FIORINI
Centre d'dtudes de Cadarache,
France
Abstract

The Nuclear Reactor Division PRN) of CEA is in charge of the evaluation of the design options for the future power plant.
The objective of this report is the identification of the safety related concerns (criteria) that must be used to fulfil this task.

Starting from the main levels of Defence in Depth, and taking into account the recommendations for future Nuclear Power
Plants, these criteria are obtained using a functional approach (OWhat for ; What is to be doneo). After the identification of
these criteria, the exercise is performed, as a matter of example, for three options among those suggested for the Decay Heal
Removal (DHR).
L
INTRODUCTION
The design options for future fission plant (systems, design features, materials) must be evaluated by the Nuclear Reactor
Division (DRN) of CEA on the basis of several types of concerns: operation, safety, fuel cycle, economics, etc.

A task of which the Innovative Reactor Concept Service (CEA/DRN/DER/SIS) is in charge is the contribution to the evaluation
of the design options on safety level and licensing. The final goal is the assessment of the coherence between the design
choices and the Defence in Depth principles. The first objective of the task is to help formulate the top level interim safety
criteria essentials to perform this evaluation. This is why, within the frame of this task, it is requested to develop a standard
ad-hoc methodology to identify such a criteria (safety concerns).All the design safety related objectives already formulated
by the safety authorities and/or by the international advisory groups for the future power plants must be taken into account. The
need for a systematic integrated approach, useful to select the design options, is justified because of number and complexity of
the issues involved.
The main goal of the report is to suggest such a pragmatic approach for the identification of these safety related objectives and
consequently for the selection of the criteria needed for the final evaluation.

2.

RECALL ON THE SAFETY APPROACH FOR NUCLEAR PLANTS

2.1

Safety objectives and approach

An high safety level is advocated for future power plants.

To fulfil this objective the key recommendation is to design future fusion plants implementing the strategy of Defence in Depth
(D. in D.). This approach, can be summarized as follows :
1.
2.
3.

Priority to the prevention efforts to avoid incidents and accidents.

Design effort for an easier management of abnormal situations (protection).
Design effort to take into account and to mitigate the consequences of major accidents (mitigation).

Moreover, to cope with general requirements for future plants and still within the frame of a correct D. in D. implementation, it
is essential to design the safety systems and their architecture in order to achieve:
1.
2.
3.

an extended defence with an effort to prevent systematically potential accident initiators and take severe accidents into
account; the aim is to tentatively reduce, at the prevention level, their potential consequences,
a balanced defence to avoid singularities among the different accident families contributing to the plant degradation,
a gradual defence to avoid short sequences and to allow the operator back-up at intermediate accident stages.

The interest of the proposed design options (and so their evaluation) must be judged on the basis of their coherence when
compared to all above objectives.
2.2

Technical guide-lines

Following the reference /!/ the technical guidelines (principles) to concretize the above objectives cover the following items:
"^General technical principlesi^Specific principles: 'Siting; 'Design; Manufacturing and construction; 'Commissioning;

Operation; 'Accident management; 'Emergency preparedness

55

The reference 121 recognizes the previous items as mandatory, and recommend the adoption of some complementary generic
principles
o
The concept of plant design should be extended to include the operating and maintenance procedures required for it
=>
Design should avoid complexity
=>
Plants should be designed to be "user friendly"
<>
Design should further reduce dependence on early operator action

The design of the system provided to ensure confinement of radioactive materials after a postulated accident should take

into account the values of pressure and temperature encountered in severe accident analysis
Accidents that would be large contributors to risk should be designed out or should be reduced in probability and/or
consequences

o
o
o

The plant should be adequately protected by design against sabotage and conventional armed attack
Design features should reduce the uncertainty in the results of probabilistic safety analysis
Consideration should be given to passive safety features

As a complement of these international guide-lines, French safety authorities ask explicitly for the reduction of common mode
failures /3/ The implementation of functional redundancies capabilities (two or more systems to realize the same safety
function) is recommended to fulfil this objective Those recommendations have been confirmed within the frame of the
E'lropean Pressurized Reactor (EPR) activities A common French and German safety authorities report /4/ precogmze the
evolutive appproach for the future EPR. The role of the defence in depth is stressed and a significant reduction of radioactive

releases due to all conceivable accidents -including core melt accidents- is exphcitely requested. Among the suggested technical
principles it is interesting to recall the following
c?

Quality of design, manufacturing, construction and operation to point out the importance for the inspectabthty and

(he testability of equipments

Reduction of frequency of initiating events to reduce the frequency of occurrence of accidents (including core melt

o
<>

<^
=>

accidents)
Improved plant transient behaviour to avoid unnecessary safety systems actions
Redundancy and diversity to be consistent with the general objective of reducing the probabilities of occurrence of
accidents
Active and passive systems in order to identify the advantages and the disadvantages of passive systems
Integrity of primary circuit as well as the integrity of the other safety related high energy components and piping
within the containment systems to reduce the potential containments loads

=>
o

Man-machine interface to take advantage of the human abilities, while mmiminng the possibilities for human erors
and making the plant less sensitive to these errorsi
Qualification of computerized systems to obtain the necessary high reliability for instrumentation and control
systems

The use of probabilistic safety assessment is suggested to support the design options, a well balanced safety concept and the

valuation of expected deviations from present French and German safety practices
The reference /4/ also recommend that accident situations which would lead to large early releases have to be "practically

eliminated" (when they cannot be considered as physically unpossble, design provisions have to be taken to design them out)
reactivity accidents, high pressure core melt situations, global Hydrogen detonation Low pressure core melt accidents have
to be "deal with", so that the associated conceivable releases would necessitate only very limited protective measures in area
and in tune The objective of significant reduction of the radioactive releases implies a substantial improvement of the
containment function To do this, among others, the residual heat must be removed from the containment without venting
device For this function, a last-resort heat removal system must be installed this system should be preferably passive with
respect to its primary circuit inside the containment
2.3

Defence in Depth (D. in D.) implementation

As stated above, the design effort must be coherent with the D in D approach through the implementation of three genenc
goals prevention, protection and mitigation These goals can be expanded to obtain five levels of D mD
f I^
Prevention {
I

conservative design,
quality assurance,
safety culture

Protection

f 2n<l control of abnormal operations and detection of failures

{
l3rc* protection and safeguard systems

Mitigation

(4th
{
[5th

major accident management including confmment protection

off-site emergency response (consequences mitigation)

MOTE
It is essential to note that through the fourth and fifth levels, the Defence in Depth approach requires an ultimate
demonstration of the plant safety, taking into account, as a matter of routine, the plant degradation (severe accident) The
reasons of this last requirement can be interpreted as follows

56

a.
to cover the eventual lack of exhaustivity of the selected deterministic sequences,
b.
to demonstrate the potential of the concept for mitigating severe accidents,
c.
to demonstrate the avoidance, by design, of any cliff edge effect.
(The cliff edge is a discontinuity in the relationship between the frequencies and the consequences that defines the risk : Risk =
frequency x consequences).

3.

SAFETY CONCERNS

The section 2.2 presents some generic recommendations and the technical guide-line that must be taken into account for future
nuclear plants. Starting from the Defence in Depth levels (see section 2.3) all these indications are integrated and developed
following a functional analysis approach. This approach, currently used for the value analysis, details generic specifications
(what for) suggesting more and more detailed technical solutions (what is to be done).
The methodology allows to identify:

the generical criteria for the evaluation of the main plant design options,
the specific safety concerns (or specific criteria) for the evaluation of the safely function related system, subsystems or
components.

The first steps are presented on Table 1 (left hand). After the generical criteria (level 1 (O)) applicable to the future reactors,
the identified safety concerns (or design safety related objectives - level 2 ()) are still generic and must be applied to all safety
related design options. A further step is presented still on Table I (left hand). As a trial measure, the methodology is applied
to the decay heat removal identifying the design specifications (safety function related objectives - level 3 (*)) proper to the
corresponding systems, subsystems or components.
4.

OPTIONS FOR THE DECAY HEAT REMOVAL

4.1 Options

Several options for the decay heat removal coolant have been proposed for the implementation on future NPP. Three among
them are choosen to present and discuss the suggested methodology.
OPTION

System

PRHR

RRP

SCS

Comments

Passive Decay Heat Removal system implemented on the primary circuit of

the AP600 concept.
Studies have been performed at the CEA to implement similar system on a
900 MW PWR (see diagram).
Passive/active Decay Heat Removal system.
Heat exchangers installed within the primary vessel
Studies have been performed at the CEA to design the system for a 900
MW PWR (see diagram; 151 for the performances).
Passive Decay Heat Removal system implemented on the secondary circuit
of the SIR concept
Studies have been performed at the CEA to implement similar system on a
900 MW PWR (see diagram; /5/ for the performances).

4.3.

Qualitative evaluation of the heat removal

The different options are evaluated using the safety concerns identified and listed on table 1. The appreciation is expressed in
term of Favourable ft; unfavourable &; unaffected O. The preliminary evaluation results arc tentatively (and not compleily)
summarized on tables 1 (right hand). This allows to identify the pro and cons of each option. Those results are essential to

identify, motivate and prioritize the R & D efforts that support the system design activities.
The details of these results are not commented here. The main objective of the report is to present the methodology. The
design objectives presented on table 1 are open for discussion to improve their coherence versus the claimed goals. The system
qualitative evaluations are also preliminary and must be discussed in detail.
5.

CONCLUSIONS

The discussion of the recommendations already available for future nuclear plants provide clear guidelines directly applicable

for the design. The Defence in Depth approach remains the reference. Its correct implementation leads to take care of all the
levels (prevention, protection, mitigation) and provide an extended, gradual and well balanced defence.
The development of these levels following a functional approach (what for <- what is to be done) allows to identify a series of
technical generic design objectives for future nuclear plant The development can be pursued to define the objectives/criteria
needed for the evaluation of the systems, subsystems or components, related to a safety function. The report provides a first
proposal for the methodology and identifies, as a matter of example, the evaluation criteria useful for a comparative qualitative
evaluation among different design options for Decay Heat Removal. Similar approach can easily be applied for the evaluation of
the design options in charge of the other important safety functions e.g. Reactivity Control and Fission Products Containment.

57

It is important to point out that the results of this approach are essential to identify, motivate and prioritize the R & D efforts
that support the system design activities.

TABLE 1
FUTUREPWR
DECAY HEAT REMOVAL (DHR) FUNCTION - COMPARATIVE EVALUATION AMONG DIFFERENT DHR SYSTEMS
Implementation of a PRHR system, or a RRP system, or a SCS system for the Decay Heat removal of a future standard PWR.

Identification of the favourable (tt), unfavourable (-0 ), Indifferent (*) contributions linked to the systems implementation
PRHR system; or RRP system, or SCS system for the Decay Heat Removal of a future standard PWR.
Identification of the contribution linked to the implementation of each system
_____

Fvonrbk fr: unfavourable 0-; iodlfltrctrt *>

SYSTEMS PRHR

RRP

Comments

SCS

1st level: PREVENTION

conservative design, quality assurance, safety culture
*^ Elaborate a simplified design
Elaborate a simplified neutronic design
Elaborate a simplified thermohydrauuc design
Simplify the vessel internals

- IneKucd cotopknty withm the veud

Simplify the thermohydrauuc for the normal DHR

Simplify the thermohydrauuc for the safeguard DHR

-The DHR. ojncooa efficiency B ctiDjumnled

Separelhc normal operating DHR function from the safeguard DHR

Increase the range covered by normal DHR systems (forced conv.. natural conv )
Reduce the number of components per system
Standardize the components among normal operating DHR and safeguard DHR
Elaborate a simplified thermomecamc design

System Car nomul mslec ategata DHR 7

Simplify the vessel internals

WU> IncRued axnploaly withm the veud

:-^mn< be defined ina iccaac dcn^i)

Reduce the number of systems connected to the primary circuit

Reduce the impact of transients

Minimise the thermomechanical loads (pressure versus geometry. AP)

Reduce the number of components per system____________________
Satisfy the design rules
Elaborate a design consistent with all the plausible situations
Take info account the Passive Single Failure criterion for the short term

PKJ1R nptcaxnted f MI eaeraxKi of die P C

0-tf

PRHW07).SCS D MhtbetmjJvivc

The mk faf ovacootmg m eatte of complete system

ilu Million mutt b

PRHR- Pmmr LOCA.RRP note of inland beu

Qualify the materials (mechanical, electrical, etc.)

Qualify the materials for the plannedJunction (performances)

Qualify the materials versus the requested reliability
Qualify the materials versus the requested availability
Qualify the materials for the espected environmental conditions
Plan the possibility for representative tests
Standardize the components among systems (improve the feedback experience)
o

RW - Ad boc m

Sunpur> tbe reactor operations and the maiotenaoce procedures for normal

conditions (buman Cuter for operauon and shut down)

Improve the quality of the information (operational data)
Implement adequate control on systems behaviour and status

W - CXfficuftKt far DOC MBumsnttton?

(xtyto- itduood ndb)

tmf^ow the man-machine interface

Systow cay to opente

Sin.pury and automatize the procedures for the operation

SyxUBS c*xy to openlc

Simplify and automatize the procedures for the inspection

RW - OdEcuoci lor IHX nspcctwo

Simpufv and automatize the procedures for the maintenance and preventive repair

RJt? - Dd&cuhiel for IHX mimlcnmce wi iqwr

58

TABLE 1 (cent.)
PRHR system; or RRP system; or SCS system for the Decay Heat Removal of * future standard PWR.
Identification of the conlnbuUon linked to the implementation of etch system
Favourable ft; unfavourable ft; Indifferent *
SYSTEMS

PRHR

Comjncnls

RRP

lit level :PRJEVEMTION

conservative design, quality assurance, safety culture (follows)

<> Integrate the principles of the defence in depth: balanced, gradual and extended

M the systems cm be considered u oufepovknl LOD

defence

Tfcw comet onpfenwnttfion kids lo improve the

Take care to the balanced character of the implemented defence

* Implement an homogeneous number of Lines of Defence (LOD) for each

operating condition (OC: PIE applied to an Initial plant status)

Take care to the gradual durartcr of the implemented defence

* Implement functional redundancies: mdipendent LOD
Take care to the extensive character of the implemented defence
* For each operating condition, implement an number of Lines of Defence (LOD)
coherent with the probabilistic objectives-^fta-t-b)
Minimize the personnel exposure during normal operation

Minimize the contact dose

Reduce the corrosion phenomena and the radioactive products transport

ft?

ft'

PRHR, RKJ> lnaeolru>&oet of the seoood Inner

Limit the length of circuits which carry activated fluid

Ncgfapble rrau the SG?

Reduce the portions of circuits that comes primary coolant

Minimize the maintenance times for normal conditions

PRKR KBplematied *s i extauwn of the P C

Improve the accessibility

K n^Aeaienled n in cxlcnAon of the P C

Foresee equipments and robots

RJU> - Dd&ajtoa f MX >cceuildity

Minimize radioactive waste during normal operation

Ootefiy ndiffenrt

Simplify the chemistry of the primary circuit

Reduce the self -generation of radioactive waste
Reduce the corrosion phenomenon

PRHR, RW InereMcd surfaces of Ac second buncr

Ensure the good materials behaviour under irradiation

>ma (he SCT

Minimize the frequency for the Postulated Initiating Events

(PIE - abnormal situations; during normal operation and shut down) :
Control rod withdrawal
Uncontrolled boron dilution
LOFA - Sequences initiated by toss of primary coolant Oow

Loss of charge / turbine trip

Loss of normal fecdwater

Loss of external electrical power supply

LOCA - Sequences initiated by a leakage of primary coolant
Reduce the number of chipping on primary loops
Minimise the length of the loops which carry primary fluid
Minimise the fluids internal energy (primary pressure)

PRHR. KX? Inctemal turfed of be icconl httnet

R&P ta4 0 externl lowfa (pressure outink) ut

bnxntfc

Reduce the corrosion phenomenon

Sequences initiated by toss of secondary coolant or heat sink
Minimise the length of the loops vhich carry secondary fluid

RRP. SCS ocrtMdfavhtofttcecidoy loo(a

<RW* -

e of loop noxut)

Minimise the fluids internal energy (secondary pressure)

Reduce the corrosion phenomenon
Steam Generator tubes rupture
Minimise the fluids internal energy (AP primary/secondary pressure)

RRP - DOC 4es oKraknt to fte SC ones,

nevertfadenlhe cxleinal to*<k (prexcw outoene)

Reduce the corrosion phenomenon

59

TABLE 1 (cont.)

PRHR system; or RRJ> system; or SCS system for the Decay Heat Removal of a future standard PWR.
Identification of the contribution linked to the implementation of each system
__

_________Favourable ft; unfavourable 4; hwmferent

SYS1EMS

PRHR

RRP

SCS

Comments

Irt level PREVENTION

conservative design, quality assurance, safety culture (follows)

Minimize the potential for Common Modes (Initiators)

Separate and diversify the systems

Divernjy rte components

Keep segregate the single loops

Minimize the potential for flooding

Put out of voter the compnents important for safety

Minimize the potential for fires
Implement incombustible materials
Qualify the material for the earthquake

UU> - tMScuk todmntfr UK IHX

TV ettlhquafce fcspotte imat be carefij&y uutysed

Minimize the inherent potential consequences for the PIE (opcr. and shut down):
Control rod withdrawal
Uncontrolled boron dilution
LOFA Sequences initiated by loss of primary coolant flow

foresee an adequate pump inertia

Foresee the natural convection behaviour

ft

ft

foresee the natural convection behaviour

Loss of normal feedwaler

ft

o-

Foresee the natural convection behaviour

Loss of external electrical power supply
Foresee the natural convection behaviour

ft

Loss of charge / turbine trip

ft

LOCA Sequences initiated by a leakage of primary coolant

Minimise the primary depressunsation effects (on the three barriers)

Ensure the DHR with reduced primary voter inventory
Sequences initiated by loss of secondary coolant or heat sink
Minimise the secondary depressunsation effects (on the three barriers)
Steam Otaerator tubes rupture

ft

RRP - Tbe DHR &nctxn efficiency B

Confine the secondary discharges

Avoid by design (prevent) the sequences thai can leads to unacceptable

consequences and early releues. Reject the risk for the cliff edge effect
Avoid by design the reactivity excursions
Avoid by design the core melting under high primary pressure conditions
*

Participate efficiently to the primary circuit depressunsation

Oft
PRHR 0 - SO dxnmjy tsofcJM: pomny Croat Kill

Avoid by design the core metting concomitant to the loss of the containment
(bypass).

Set up within the containment all the hops that carry primary coolant

Foresee the isolation of all intermediate loops at the containment level

Avoid by design the risks for the steam explosions

Avoid by design the risks for the Hydrogen detonation

60

UK fydemi pounlol for DHit m OK oTicvtn

Foresee an ultimate passive DHR system within the containment

Oft

Oft

RR? W MHU! betl occtunga

SCS 0 Win man! pool

TABLE 1 (cont.)

PRHR system; or RRP syston; or SCS system for the Decay Heat Removal of a future standard PWR.
Identification of the contribution linked to the implementation of each system
___

Favourable ; unfavourable -0-; Indifferent O

SYSTEMS PRHR

RRP

Commote

SCS

* 2nd level :CONTROL

control of abnormal operations and detection of failures
<*

Detect the Postulated Initialing Events (PIE-

Control rod withdrawal

Uncontrolled boroo dilution

LOFA Sequences initialed by loss of primary coolant flow

Loss of charge / turbine trip

0
0

operation and shut down) -

GoUDy mcfenerent

Loss of normal feedwatcr

Loss of external electrical power supply

LOG A Sequences mrtialed by a leakage in primary coolant

Sequences initiated by loss of secondary coolant or heat link

Steam Generator tubes rupture

Minimize the uncertainties about the plant conditions

Implement an adequate instrumentation (for automatic and manual devices)

Foresee an instrumentation able to identify without ambiguity the systems

Passive systems reduced need for mstnancntbon ind

configurations and the consequences on the control parameters

Implement a design that simplify the abnormal sequences

Dispose of the natural convection within the primary circuit
Dispose of the natural convection wifhm the secondary loops

These advantages aose from the passive chancier of the

system

e Simplify the plant inherent behaviour under abnormal coodWocis (operation and
shutdown)
0

Control rod withdrawal

Uncontrolled boron dilution

LOFA-Sequences initialed by loss of primary coolant flow

Foresee an adequate pump inertia
Foresee the natural convection behaviour
Loss of charge / turbine trip
Implement passive and efficient functional redundancy
Loss of normal fccdwatcr

Euy lU&ml eonvecbon on the pnmuy orcut

o-

"

Implement passive and efficient functional redundancy

Loss of external electrical power supply

Implement passive and efficient functional redundancy

LOCA- Sequences initiated by a leakage in primary coolant
Foresee the passive cooling of the primary coolant soil within the vessel
Foresee the passive coolant injection to guarantee the primary coolant inventory
Foresee the passive cooling of the primary coolant rejected in the containment
(Control the containment temperature and pressiure)
Sequences initiated by loss of secondary coolant or heal sink
Foresee the passive coohng of the primary coolant

o-

Foresee the passive cooling of the secondary coolant rejected in the containmen
(Control the containment temperature and presssure)
Steam Generator tubes rupture
Avoid the activation of the secondary overflow valves
Confine the secondary discharges

tf

RRP - The DHR fcncban efficiency ts stiD guannled

wflh km ptmwry water inventory

0'

PRHR - SO bazuDy BObtal

e>

SCS - eOoenl ifte secondiry Orach B oolatod

Tbc systems cm coobbute to deprcssimzr the ccinuiy

61

TABLE 1 (cont.)

PRHR system, or RKP system, or SCS system for the Decay Heat Removal of * future standard PWR.

Identification of the contribution linked to the implementation of each system

______________Favourable it. unfavourable0-; indifferent O
SYSTEMS

RRP

Comments

SCS

* 2nd level :CONTROL

control of abnormal oper and detection of failures (follows)

Elaborate a forgiving design

Ensure appropriate physical margins

Improve the system efficiency

Increase the common range covered by complementary systems

IRP ncDHRfioKfton efficiency B s

wh low pcmtty wafer mvtntory

te mfc (orovcrcoobnKtncue of compfcte system

Bitoven&on must be evihuled

Ensure appropriate grace period

Increase the process internal inertia

Provide the passive access to adequate external inertia

Lirmt the intervention of safeguard functions (uraeoessary safety systems actions)

Foresee control and limitation devices

otcntufly Ervmnjbte The rak or die systems must be

doily octned___________________

<* Take mto account aggravating sttuauons (coherently with the PIE category)

Take mto account the corrective functions unavailability for maintenance

Foresee an internal redundancy

Take mto account the PIE with cumulative system failures

*
*

Take into account the Operating Conditions (OC) with the Loss ofOffsite Power
Take into account the OC considering the Single Failure Criterion
Take into account the OC with internal and external hazards

Sepuitt loops for the Syrians

RKRiMSCSinpatups moce tenable to the

exlgnuHunrett must be vqified

< SinipufytbereactcTorKraucasandmaimeaanceproceAiresund^abrionral

conditions (human factor for operation and shut down)

Improve the quality of (he information (operation data)

Simplify and automatize the procedures for the plant opcr under abnormal cond
Improve the man machine interface
Limit the interactions among systems that perform the same function
Implement safety system automatisation

Simplify and automatize the procedures for the plant inspection. mainL and repair

Improve die accessibility

Foresee equipments and robots___________________________

Muunuze the personnel eiqxksure under abnormal conditions (oper and shut down)
Strengthen the first barrier

Strengthen the second barrier

Reduce the portions of circuits that carries primary coolant

Reduce the lane for the intervention under abnormal conditions

Improve the accessibility

'

Forget equipments and robots___________________________

Dedicated systems

PusMsysbans

PRHX.KRf cUEctikUxocco SCS 0 if fee pod a

PRHK.RRJ> tUEeuilto*

SCS 0 if OK pool c

Minimize the radioactive waste under abnormal conditions (oper and shut down)
Strengthen the first barrier
Strengthen the second barrier

Conceive the circuits connected to the primary

- permanently

-* installed within the containment

RKP wit dometfatfe bat exchanger wuhai die

- temporarily
-+ eventually outside but isolable
Conceive the circuits connected to the secondary
-* designed to the maximum injection pressure
-* in order to confine the discharge within the containment
Strengthen the third barrier
Limit the number of containment penetrations

PRHR II bclheAHOOcaituni>an
RRP 0 4 mtenul bnl cxctunga

SCS 0 rflhc pool 13 mafe the conUguncrt

62

TABLE 1 (cont.)

PRHR system; or RRP system; or SCS *ystem for the Decay Heat Removal of a future standard PWR.
Identification of the contribution linked to the implementation of each system
________________Favonrabk ft; unfavourable &; indifferent O

SYSTEMS PRHR

RRP

Comments

SCS

*3rd level:
PROTECTION
safety systems and protection systems

Minimize (he uncertainties about the plant conditions under accidental conditions

(operation and shut down)

Implement an adequate instrumentation (for automatic and manual devices)
Foresee an instrumentation able to identify without ambiguity the systems
configurations and the consequences on the control parameters
Implement a design that simplify the accidental sequences
Reduce the number of possible systems response configurations
Dispose of a primary circuit natural convection configuration
Dispose of a primary circuit natural convection configuration
O Simplify the reactor management under accidental conditions (oper. and shut down)
Control rod withdrawal
0
Uncontrolled boron dilution
0
LOFA - Sequences initiated by loss of primary coolant flow
<> Loss of charge /turbine trip
0
Loss of normal fecdwater
0
Loss of external electrical power supply
0
LOCA - Sequences initiated by a leakage in primary coolant

*usfK systems (educed need for mstiusKnulioa uvi

control

rhese tdvantagei *nsc from the passive ctuneter of the

systems

Due to the degraded otuUMn. che systems fevomblc

conatHAon a due to Ac pusnc duncter of dar

nlecvxntion

ft

ft
ft

ft
ft
ft

ft
ft

ft
ft

ft

ft

- Ibc DHft Amcaon effioency is safl gvunnccd

wifl tow pcmuy wuer eiventoty

0

Sequences initiated by loss of secondary coolant or beat sink

Steam Generator tubes rupture

o Take into account aggravating situations (coherently with the PIE category)
Take into account (he safeguard functions unavailability for maintenance
Foresee a functional redundancy
Take iao account (he PIE with cumulative safeguard system failures (multiple
failure situations)
Take into account the Operating Conditions (OC) with the Loss ofOffsite Pover
Take into account the OC considering the Single Failure Criterion
Take into account the OC with internal and eaemal hazards

Take atfo account the loss of the redundant systems (complementary situations)
Foresee an ultimate passive DHR within the containment

Minimize the potential for Common Modes (mutual agressioos, internal or external

ft

ft

ft

li-

ft

ft

SCS c&aeni onty if the secondaly breach o ootacd

St
fr
PRHR md SCS we pcdiaps nMicscnsiUc lathe
exteaul hxzuds oust be venfied
A systems omliilKiUjnciapcdaps be mfuagcj

hazards)

Separate and diversify the safety systcmcs

Avoid any physical interaction between the systems in case of failure

TV coMDbudon on be rfbc

edoojy

Simplify the accidental intervention procedures under accidental conditions

(human fetor for operation and shut down)
Ensure ao adequate information (abnormal situation)
Simplify and automatize the procedures for the accident management
Improve the man-machine interface
Limit the interactions among systems that perform the same function
lmp*cm&t safety system automatization
Stn.plify end automatize the procedures for the plant inspection, and repair
Improve tfte accessibility
Foresee equipments and robots

Pauiw systems

PRHR. RW .ddSa* la MCCSS. SCS 0 if the pod n

ouewic t

63

TABLE 1 (cont.)

PRHR system; or RRP system; or SCS system lor the Decay Heal Removal of a future standard PWR.

Identification of the contribution linked to the implementation of each system

Favourable ft; unfavourable & ; Indifferent *

RRP

SCS

ft
ft

ft

ft

ft

ft*

SYSTEMS PRHR

Comments

*3rd level : PROTECTION

safety systems and protection systems

Reduce the core melt frequency

Improve the availability of the safeguard systems
Implement an irirqinlr functional redundancy for the important safety (unctions

ft

Tbae gcncnc lavounbk contrfbutKXu mutt be

ft

* Minimize the of&ilc accidental release (without core melt)

Conceive the plant in order to guarantee no necessity for protective measures

Strengthen the third homer

PRHR t* for tbc AP600 oonfipnbon

RRPU wh mxantl beMtxttuaga

SCS 0 if the pool tt mode the contaunmou

PRHR system; or RRP system; or SCS system for the Decay Heat Removal of a future standard PWR.
Identification of the contribution linked to the implementation of each system
Favourable ft; unfavourable 4; Indifferent O

SYSTEMS PRHR

* 4th level :

RRP

SCS

ft

ft?

Comments

MAJOR ACCIDENT MANAGEMENT

accident management including the confinement protection

& Take toto account the severe accident configurations

Ensure the safety function accomplishment under severe accident conditions

Foresee the DHR with severe accident configurations:
Conum within the primary vessel

ft'

Utf-Tbceffiocncy vcreus Ox D1(R function o rtifl

[uaimtod witti low faaary WIICT nvoittcy

For d* PRHR >nd SCS Die cOocncy mua be

iaoaamiai

Conum within the containment (core catcher)

Protect the material against the potential hazards (steam explosion. H2 defiag., etc.)

Protect the material against the fallowing hazards: conum. Hy deflagration,

ft?

ft?

ft?

ft

ft

temperature, pressure, etc

RXP - Mocc icra&tc to OK potential vi-veud

gfCKKOl

Minimize (he uncertainties about the plant conditions under severe accidental

conditions (operation and shut down)

Implement an ade<piale instrumentation

Elaborate a design that simplifies the inherent plant accidental scenarii

Simplify the rcacUx operations procedures under severe accident conditions

ft?

ft?

ft?
tbe fytiems efficiency must be derooRflrunj

Improve the grace delay

Implement an ultimate passive DHR for the conum cooling

ft?

ft?

ft?

Intenul eonbmmotf njiunl convection eonJSgmMns

the syxau efficiency must be demofutnled

=> Minimize the oftshe accidental release (low pressure core roeh situations)
Conceive in order to need only very limited protective measures in area and in time
* Qualify the third barrier to the configurations with the low pressure core melt'
guarantee ta cooling

64

*?

*?
nurt be evaluated

TABLE 1 (cont.)

PRHR system; or RRP system; or SCS system for the Decay Heat Removal of a future standard PWR.
Identification of the contributioa linked to the implementation of each system
Favourable *; unfavourable Os Indifferent e
SYSIEMS PRHR

RRP

Comments

SCS

* 4th level : MAJOR ACCIDENT MANAGEMENT

accident management including the confinement protection

Avoid by design (prevent) the sequences that can leads to unacceptable
consequences and early releases. Reject the risk for the cliff edge effect
(For recall - must be realized at (be prevention level)

Avoid by design the reactivity excursions

Avoid by design the core melting under high primary pressure conditions
Participate efficiently to the primary ctrcutt depressitnsation

ft

Thesyrtem*<anc&vdy p*itqplr totficpiinmy

oxcutt dcpcBamsa&on.
PttiR B-SGlbcflOlfixtfolrted pnmuy orcuit jtill

pnuuoxod

Avoid by design the core melung concomitant to the loss of the contain, (bypass).
Foresee an ultimate passive DHR system within the containment

Set up wrthin the containment all the loops that carry primary coolant

Foresee the isolation of all intermediate loops at the containment level

Avoid by design the rides for the steam explosions
Avoid by design (he risks for the Hydrogen detonation

0
ft

ft

The xytfont pofcnfMl for DHR tt cue of severe

cofcnt a wx daily aabtehcd

SCS OWnkMOMl pool

PRHR system; or RRP system; or SCS system for the Decay Heat Removal of a future standard PWR.
Identification of the contribution linked to the implementation of each system
FavooraMe tf; nnfavoorable & ; Indifferent
Oouuucnts
SYSTEMS PRHR RRP
SCS

* 5th ievd : CONSEQUENCES MITIGATION

offsite emergency response
=> Delay tSe offisite release
o

Minimize the oflsite radioactive release

65

Passive Residual Heat Removal

Heat exchanger

RRP In reactor
vessel

Secondary Condensing
System

RRP diagram

ECHANGEUR INTEGRE EN :..'.-:

DECAY HEAT REMOVAL SYSTEMS DIAGRAMS

REFERENCES
III

Nuclear Safety Advisory Groups: Basic Safety Principles for nuclear Power Plants
IAEA Safety Serie n75-INSAG3

111

Nuclear Safety Advisory Groups: The Safety of Nuclear Power

IAEA Safety Serie n75-INSAG5

/3/

Letter from the DSIN Director (M Laverie - Direction de la Surete des Installations Nucleaires) to the CEA General
Manager on the Pressurized Water Reactors for the future; DSIN 984/91

/4/

Letter from the DSIN Director (M. Lacoste - Direction de la Suretd des Installations Nucleaires) to the EDF General
Manager on the Pressurized Water Reactors for the future; DSIN 1394/93
+ the joined report: GPR/RSK Proposal for a common Safety Approach for Future Pressurized Water Reactors; adopted
during the GPR/RSK meeting on May 25, 1993

/5/

66

P. Aujollet CEA/DRN/DER/SIS - Efficiency Studies of Future PWR Safety Systems

Presentation to the same conference

POSTULATED SMALL BREAK LOCA SIMULATION

IN A CANDU TYPE REACTOR WITH ECC INJECTION

UNDER NATURAL CIRCULATION CONDITIONS

G. BEDROSSIAN, S. GERSBERG
Comision Nacional de Energia Atomica-Avda,
Buenos Aires, Argentina
Abstract

This paper presents a thennalhydraulic analysis of the simulation of a postulated 55 kg/s inlet
header break (very small LOCA) in the primary heat transport system of Embaise Nuclear
Power Plant, coincident with a loss of Class IV Power Supply, on the assumption that the
Emergency Core Cooling System action is not automatic but it must be performed manually.
The main objective of this work was to evaluate the refill performance of a high-pressure
accumulator tank emergency coolant injection system under natural circulation conditions.
It is proposed a multiple channel model approach to study the response of the broken loop on
the basis that it is expected that the behaviour will not be the same for all the channels in the
critical core pass. The intact loop was represented by an average channel model.
The emergency coolant was injected in all the headers of the primary system and in the inlet
headers only. The system was activated at 630 sec. and 1000 sec. after the beginning of the
event.
Stagnation and flow reversal phenomena were observed in the broken loop, before the
emergency coolant injection or because of it. This condition is known as bidirectional flow, that
is, flow in some channels in the same pass continues in the normal direction while flow in the
other parallel channels is in the reverse direction. This may be attributed to changes in the
hydrostatic forces in the loop.
In every case, single phase flow was established by the end of the transient, associated with
low fuel temperature.
However, it was seen that core cooling is improved whenever the injection is directed to the
inlet headers only, since nominal direction flow is reestablished in all the channels.
On the other hand, even though the results for the later injection showed a fuel temperature
decrease, it is recommended not to delay the initiation of the emergency coolant injection
beyond 600 sec. because it was seen that void would reach the critical pass inlet header,
denoting an insufficient heat removal by the boilers.
1. INTRODUCTION

This paper presents a thennalhydraulic analysis of the simulation of a posrulated 55 kg/'s inlet
header break (very small break) in the primary heat transport system of Embaise Nuclear
Power Plant coincident with a loss of Class IV Power Supply, on the assumption that the
Emergency Core Cooling System action is not automatic but h must be performed manually.
The main objective of this work was to evaluate the refill performance of a high-pressure
accumulator tank emergency coolant injection system under natural circulation conditions.
A multiple channel representation was proposed to study the broken loop response since
previous simulations demonstrated that emergency coolant injection under natural circulation
conditions causes low flow periods, even temporary stagnation and flow reversal. [1]
In these cases, it is thought that density gradients are the driving forces of coolant flow. So, it is
expected that all the channels in the critical core pass will not have the same behaviour, as they
differ in power, elevation and hydraulic resistance.
67

Therefore, a multiple channel representation for the critical pass of the broken loop would
account for these spatial phenomena.
The accident scenario included a loss of Class IV Power because a reactor trip can result in a
sudden loss of generation which in some cases results in a grid disturbance that may lead to a
loss of Class IV Power. [2]

2. MODELS AND ASSUMPTIONS

The study postulated a very small inlet header break (6.41 1CH m2) in the Primary Heat
Transport System of Embalse Nuclear Power Plant (fig. 1) which produces an initial flow
dicharge of 55 kg/s. The area corresponds to 0.32% of twice the cross-sectional area of the inlet
header.
For this break size it is assumed that:
. the Emergency Core Cooling System action is not automatic. [2]
. the Reactivity Control System is fast enough and has enough negative reactivity to
compensate for the increased reactivity. Therefore the reactor power stays constant up to the
trip. [3]

High reactor building pressure trip was credited; this would take place -according to the Safety
Report- around 25 seconds after the event initiation.
Forty seconds after reactor trip, Class IV Power Supply was supposed to be lost, and
reestablishment of Class HI Power was not credited.
The major consequence of a loss of electrical power is the loss of power to the heat transport
system pumps. Without these pumps the effectiveness of the steam generators as a heat sink
and the speed with which the Emergency Core Cooling System can refill fuel channels is quite
different than for a loss of coolant with the pumps running.
.After a loss of Class IV power, other functions are also lost. These include the steam generators
feedwater pumps, the medium and low pressure emergency coolant injection pump, the
moderator circulation pump, the heat transport feed pumps and the heavy water recovery
pumps.

Both circuits of the Heat Transport Primary System were modelled, which isolate when
pressure falls below 56.2 kg/cm2 in two of the three instrumented reactor headers, interrupting
coolant flow from the intact loop to the broken one.
The system representation [4] considered the 380 core channels as shown in Figure 1.
For the intact loop: two identical passes were modelled, each one representing 95 channels
considered as a single average channel (fig. 2)

For the broken loop: the model considered a critical pass (downstream the break) and a noncritical pass (upstream the break) (fig. 3 ). The former was represented by 5 channel groups
with the following features:

68

Group Number of
channels

1
2
3
4
5

24
24
16
16
15

Channel
power
[MW]
6.29
6.33
3.97
4.39
4.03

Initial
channel flow
[kg/s]
26.4
26.2
17.2
19.9
17.5

Initial
group flow

Elevation

[kg/s]
633.7
628.1

[m]
5.66
7.63

275.8
317.7
262.1

4.27
6.46
8.79

The non-critical pass was modelled identical to both passes of the intact loop, that is, as a
single average channel representing 95 channels.
The power of each one of these channels is 5.23 MW and their flow 22.3 kg/s.
The transient thermohydraulic computer program FIREBIRD EG, Mod. 1.0 was used to
calculate the response of the system. [5]
3. RESULTS AND DISCUSSION

The system was allowed to operate ginnnlating no corrective actions to mitigate the accident
effects.
It was observed that the break produces system depressurization and loss of inventory, which
would result in reactor building pressure increase that causes a reactor trip and then a loss of
Class IV Power, as it was postulated.
The sequence of events is the following:
t ~

0 sec

Inlet header break, with an initial discharge of 55 kg/s.

t =

25 sec

Reactor trip on high reactor building pressure signal.

t =

65 sec

Loss of electrical power.

t = 360 sec

Loops isolation on low primary system pressure.

After 600 seconds the system cooling was due to the mechanism of double phase
thermosyphoning in both circuits, with the following flows:
Pass
Intact
loop

Broken
loop

1
2
Group 1
Group 2
Critical Group 3
Group 4
GroupS
Non critical

Flow
[ke/s]
142.0
142.0
38.8
48.3
16.4
24.3
25.2
167.5

The Emergency Core Cooling System action was initiated manually around 630 seconds,
together with the steam generators crash cool down.

In the simulation the emergency coolant injection was directed:

. to all the headers.
. to inlet headers only.
69

Summary of results
The results of the emergency coolant injection at 630 seconds are presented in Table L,
comparing the effects of injecting in the two modes described above.
Figures 4 and 5 show inlet and outlet temperatures for the broken loop.
For thennosyphoning mechanism circulation there must be an important temperature difference
between both headers.
Between 800 and 900 seconds, near to zero temperature difference was observed in the broken
loop, denoting that core cooling would be temporary endangered. The intact loop always
exhibited an appreciable difference.
Figures 18 and 19 show these temperature differences for the case when injection is directed to
inlet headers. A clear difference existed between inlet and outlet headers, anticipating nominal
direction flows for all the groups.
In every case, the emergency coolant injection produces a transitory flow oscillation (fig. 6 to
10 and fig. 20). This is more accentuated when the coolant is directed to all the headers,
predicting periods of very low flow and even momentary flow reversals. In particular, the
channel representing Group 3 reverses definitely. On the other hand, when the emergency
coolant injection is initiated to inlet headers only, all the channels remain with nominal
direction flow.
In spite of these differences, all the channels exhibit an adequate fuel cooling (fig. 11 to 15),
with single phase flow (fig. 16,17 and 21 for example).

Afterwards, the possibility of delaying the Emergency Core Cooling System action was
analyzed.
The injection was carried out around 1000 seconds after the initiation of the accident. It was
also directed to all the headers and to inlet headers only.
The results are shown in Table n again comparing the two possible scenarios.
By 1000 seconds we find a degraded core cooling. The steam generators are not able to
condensate all the vapor produced, and we find void in the broken loop inlet headers (fig. 32).
The high void fraction in the core causes a sheath temperature increase.
At the moment of the injection, the coolant is practically stagnant in Groups 1,2 and 5 and has
reversed in Groups 3 and 4.
Figures 22 and 23 show that in the broken loop inlet and outlet headers temperatures are
practically equal in the critical pass, and in the non-critical pass outlet header is cooler than
inlet header. This would anticipate stagnation or flow reversal.

When emergency coolant is directed to inlet headers the injection keeps a positive temperature
difference between outlet and inlet headers.
The entrance of the emergency coolant to all the headers leads again to a transitory flow
oscillation (fig. 24 and 25); but finally, all the channels remain with nominal direction, except
Group 3, that continues reversed.
When the emergency coolant injection is supplied to inlet headers only a more definite
tendency to reach nominal direction flow was observed in all the channels (fig. 28 and 29).
Finally, the coolant in every group flows in nominal direction by the end of the transient.

When the emergency coolant injection is initiated to all the headers, the simulation predicted
long periods of low flow for Groups 1, 2 and 4, which results in a considerable void increase
(fig. 26 and 27), with the consequent sheath temperature increase.
Injecting to inlet headers only, this was not observed (fig. 30 and 31), except for Group 2 where
coolant stagnation causes vapor generation and temporary increase of sheath temperature.
However, the calculated temperatures did not show values high enough to produce fuel
damage.
70

As mentioned at the begining, during low flow periods the driving forces are the density
gradients in each channel. We considered useful to estimate this driving forces in order to
compare them with pressure differences between headers (inlet and outlet to that channel) to
have an adequate measure of the circulation conditions through the core in the critical pass.
As long as the hydrostatic height between inlet and outlet feeders is greater than pressure
difference between inlet and outlet headers of the critical pass, coolant will flow in nominal
direction. Otherwise, it will reverse.
This analysis confirmed the results described above: supplying emergency coolant injection to
all the headers, all the channels stay with nominal direction flow, except Group 3; injecting to
inlet headers only, all the channels remain with flow in nominal direction.
Broken loop
3

3
e
6| 3

6
3
6

6
1
6
1

6
1

e
4
6
4
6
4
6

6
4

4
6

6
4

4
6

6
1

4
6
4
6
4
6

6
1

2
6

6
2

6
5

5
6

5
6
5
6

e
4

e
4
6
5

e
4

6
5

6
1

6
1
6
1
6
2
6
2
6
2
6
5
6
5

3
6
3
6
1

6
1

e
1
6
2
6

2
6
2

6
2

e
5
6

e
3

s
i
6

Intact loop
6
3
6
1

6
1
6
1

3
6
3
6
1

6
1

TS
7
8
7

6
1

8
7
8
7
8
7

6
2

7
8

8
7
8

2
6
2
6
5

8
7

6
1
6
1
6
e
1
6
2
2
e
6
2
2
6
6
2
6
2
6
2
e
2
61 5
5
6
6 5

6
1

11

10

7
8
7

8
7

e
7

7! 8
8I 7
7i 8
12

13

8
7
8
7
8

7
8
7

7
8
7
8
7
8
7

7
8

7
8
7
8

8
7
8

7
8

7
8
7

8
7

8
7
8
7

7
8

8
7

7
8

7
8

8
7
8
7

7
8

7
8
7

8
7

7
8

8
7
8

7
8

8
7
8

7
8
7
8

7
8

8
7
8

8
7
8

7
8
7
8
7
S

7
8
7
8
7
8

A
B
C
O

8
7
8

7
8
7
8
7

8
7
8

7
8

E
F

7
8
7
8
7
8
7
8
7
8

7
8

7
8
7

L
M

N
O
P
Q
R
S

U
V

7
14

G
H
J
K

15

16

17

18

19

20

21

22

Figure 1 Core channel groups representation

CH3GHSH3D'."" '

Figure 2 Intact loop representation

Figure 3 Broken loop representation

71

Critical pass headers temperature

Flow-Group 4
Pfncyooa<**ft>r

otanrio

Figure 9
Nocvcntical pass headers temperature
(SmpncycooMrttoaf e <<)

Flow-Groups

(ErmycncraxiMnrioMneAMdsii;
MM MM

Figure 5

Figure 10
Flow - Group 1

Sheath temperature - Group 1

(EmoTuney cnXanr

I-

(<>

Figure 6
Flow-Group 2

fMcyooMrtUxneMcderc;

Figure 11
Sheath temperature - Group 2
(E/in>vancr cnolvK > O

M O m
<<)

Figure 7
Flow - Group 3
(SiMiyMCyaxiMloaiAeNMfta;

Figure 12
Sheath temperature - Group 3
f&netyency coolant to Ifte ft

<)

Figure 13

72

Sheath temperature - Group 4

Non-critical pass headers temperature

(Smaymraialtitt *> MM fto^fors)

\-i^^J

^>1

^^

<s= ^N*

==.

Figure 19

Fi gure 14
Sheath temperature - Group 5

Row-Group 3
lEmtrvencraxtrttolntatKXKKtzl

/?
.
x^ ^J
4r

t sgSN^ ?=&-

Figure 15
Figure 20
Core votd - Group 1
xart to at Ihe OMdan;

Core void-Group 3

- 1
TTT
u(>)

Figure 21
Core void - Group 3

Critical pass headers temperature

Critical pass headers temperature

Noo-critical pass headers temperature

r coount ft> o

(Emergency coolant to ! toe headetz)

-htTT

Figure 23

73

Flow-GroupS

Flow-Group 1
(Bmpancy eootanf to 0ie haters)

Figure 24

Figure 29

Flow-GroupS

Core void - Group 1

fi

Figure 30
Figure 25
Core void-Group 1

Core void - Group 3

Figure 26
Core void - Group 3

Broken loop inlet headers void

Flow - Group 1
(Smagaxy axUta to Hal tmaOars)

J___l__

Figure 28
74

-1-

TABLE I: EFFECTS OF EMERGENCY COOLANT INJECTION

Injection time: 632 ace.

In every cue:

Coolant flow was in nominal direction at injection time.

The Injection causes A transitory now oscillation.

Emergency coolant to all the headers

Emergency coolant to inlet headers

End of simulation: 1170 sec.

End of simulation: 1300 sec.

Variable

now

8he*th

Qroup 1

A momentary flow revertiil during the oscillation period WAR observed.

Then it continues in nominal direction.
Group I
After the oscillation period, flow continues in nominal direction.
Group 3
During the oscillation (which is longer in time than in other channels)
stagnation was observed and then a definitive (low reversal
Group 4
After the oscillation period, flow continues in nominal direction.
Group 3
A momentary flow revmiil during the oscillntion period was observed.
Then it continues in nominal direction.
non critical After the oscillation period, flow continues in nominal direction.
pats
Group 1
The prompt effect is a greiit decrease.

tentftnturt
Group!
Group 3

No flow reversal was observed.

It keeps nominal direction.
It keeps nominal direction.
Neither stagnation nor flow reversal were observed.
It keeps nominal direction.

Idem Group 2.
No flow reversal was observed.
It keeps nominal direction.

Idem Group 2.

The final value Is less than 16DC.

Idem Group 1.

Group 4

Because of the stagnation caused by the injection some peaks no greater man In all the channels the prompt effect is a great temperature decrease.
320C were observed. However the final vnlue is less than 200C.
The final value is always less than 140C.
Idem Group 1.

Group 5

Idem Group 1.

non critical Idem Group 1.

past

ON

TABLE II: EFFECTS OF EMERGENCY COOLANT INJECTION

In ovory case:

Tinio injection: 1007 me.

Emergency coolant injection onuses a rnuitltorf flow oscfllfttfon.

b) Prior to the injicrion sheath temperatures begin to increase became of tho low

Coolant flow At injection time:

) It stagnant in Groups 1, ?., 5 find in the non-critical pass.
b)

c)

flows.
The prompt effect of the injection is a great sheath temperature decrease.

Is reverted in Groups 3 and -1.

Emergency coolant to all the headers

Emergency coolant to Inlet headers

End of simulation: 187.0 nee.

End of simulation: 1820 soc,

Variable
Flow

Group 1

A mninontnry flow nrrenn) wa ohiorved during (Its oscillation period. Ttton it wm Neither rcvonnl nor stapiation were observed during the oicfflition period. Kondatl

Group 2

Similnr behaviour as Group 1 win observed.

Group 3

Motnentnry t(gn(lom were observed during the oscilation period, aflor which
nnmlnitl direction flow it reeatahltalied. However, it decreases itnd finally nrvrrsei
definitively.

After the oscillation period nnmlnnl direction flow is reestablished.

Group 4
Group 5

Similnr behaviour us Group 1 was observed.

Similar behaviour as Group 1.

seen a lonjf ttnr)on, but finally How in reestablished in nominal direction.

direction Dow is rec;tablihod.

Neithor reversnl nor stapiation were observed during the otcfflttion period Thtt
there is a tntnsitory flnn'rcvm*!, but finaHy nomraal direction flow is
reestablished.

A very tecentixted onclllntion WAD observed with successive stagnations and flow Simitar behaviour as Group 1.
irveritli, After which nominal direction flow is rcmtablijhed.
Similar behaviour as Group 1.
non critical Emorf^nc^' coolnnt mjection reestAblishes nominal dlrM^ion flow.

pa
Sheath

Group 1

The pronounced peak wa.i not observed. The final value Is loll thin 130C.

Group 2

The low flow cauies an important tmnsitory increaio. The finnl vahio is lens thnn
1WC.
Similnr behaviour ns Group 1 . The fintl vnlue ii ltj thnn 150C.

Group 3

A trnnsitory inr.n:nnc was observed here nlso, but not so important. The final value

Group 4

ii leai than 130C.

Similnr behaviour tit Group 3, although the peak is more pronounced. The final
value is less thnn 130C.

Group 5

The better circulntion condhionB prevent the temperature penk. The finnl vhio Idem Group 3.

t<mpemtnro

A peak wns observed, although not as pronounced as when the emergency cootafl
enten to all the headers. The final value is less than 130*C.
Non important incrsas: wns observed. The final vahie Is lost than 120C.,

Idem Group 3.

it less than 130C.

non critical A sliglit trnnsitory incrense was obncrved. Tho final valiw is less than 150C.
pant

Idem Group 3.

4. CONCLUSIONS

The refill performance of a high -pressure accumulator tank emergency coolant injection system
under natural circulation conditions was evaluated. The scenario consisted of a 55 kg/s reactor
inlet header break coincident with a loss of electrical power supply.
The emergency core cooling system action was initiated manually, supplying injection to all the
headers and to inlet headers only.
This was performed at 630 seconds and at 1000 seconds after the header break.

The results showed a better performance of the mentioned system whenever the injection is
directed to inlet headers only.
The code accounted for some experimental tests carried out at RD-14M loop:
core cooling may be satisfactory even with reversed flows.
emergency coolant injected to inlet headers reestablishes nominal direction flows.
Although the results for the later injection showed a fuel temperature decrease, it is
recommended not to delay the initiation of the emergency coolant injection beyond 600 seconds
because it was seen that void would reach the critical pass inlet header, denoting an insufficient
heat removal by the boilers

REFERENCES

[1] BEDROSSIAN, G. - GERSBERG, S. - Simulacidn de una rotura pequena en colector de

entrada del sistema primario de transporte de color de la Central Nuclear de Embalse
con perdida de summistro electrico normal - Comisi6n Nacional de Energia At6mica,
1025 (1993).
[2] ATOMIC ENERGY OF CANADA LIMITED - Small loss-of-cookmt accident and
emergency core cooling operation - 59 SDM-7 Second Edition (1982).
[3] ATOMIC ENERGY OF CANADA LIMITED - 600 MWe CANDU-PHW, Central
Nuclear en Embalse, C6rdoba, for Comisi6n Nacional de Energia At6mica - Safety Report
(1985)
[4] CALABRESE, R. - GERSBERG, S. - Sistema de calculo de accidentes por perdida de
refrigerante con acoplamiento neutrdnico-termohidrdulico para reactores tipo CANDU.
Informe preliminar - Comisibn Nacional de Energia Atdmica 1046 (1992)
[5] LIN, M R. et al. - FIREBIRD III Program Description - AECL 7542 (1984)

77

A RISK-BASED MARGINS APPROACH FOR PASSIVE

SYSTEM PERFORMANCE RELIABILITY ANALYSIS

N.S. SALTOS, A. EL-BASSIONI

US Nuclear Regulatory Commission,
Washington
C.P. TZANOS
Argonne National Laboratory
USA
Abstract

Passive safety systems (e.g., those used in Westinghouse's AP600 and General Electric's
SBWR designs) rely on natural forces, such as gravity, to perform their functions. Such
forces are relatively small compared to pumped system driving forces, their magnitude varies
from one scenario to another and they are subject to large uncertainties. These uncertainties
affect the passive system thermal-hydraulic (T-H) performance reliability (for short, T-H
reliability). For this reason, the T-H unreliability of the passive systems must be assessed
and accounted for in the Probabilistic Risk Assessment (PRA). The quantification of T-H
unreliability involves a prohibitively large number of computations. This paper presents a
conservative risk-based "margins" approach which eliminates the need to quantify T-H
unreliability for most, if not all, accident sequences. With this approach the issue of T-H
unreliability is addressed by linking it to the passive safety system success criteria and, thus,
to core damage frequency (CDF). This is achieved by assuring that the adopted success
criteria for safety systems and operator actions are conservative enough so that the
contribution to a sequence CDF from T-H uncertainties is significantly smaller than the
contribution from hardware failures and human errors.
INTRODUCTION
Current advanced "passive" reactor designs, such as Westinghouse's AP600 and General
Electric's SBWR designs, have a number of unique features that distinguish them from both
operating and advanced evolutionary light water reactor designs. Although, they use both

active and "passive" systems for accident prevention and mitigation, only the "passive"
systems are classified as safety related by the vendor. The word "passive" is in quotation
marks because these systems are not purely passive. Active components, such as motor
operated valves and check valves, are used for system actuation. The mission of the active
systems is to provide a first line of defense and reduce the challenge rate for the safetyrelated "passive" systems.
The passive safety systems in these designs rely on natural forces, such as gravity and stored
energy, to perform their accident prevention and mitigation functions once actuated and
started. These driving forces are not generated by external power sources (e.g., pumped
systems), as is the case in operating and evolutionary reactor designs. Because the
magnitude of the natural forces which drive the operation of passive systems is relatively
small, counter-forces (e.g., friction) can be of comparable magnitude and cannot be ignored
*
as is usually done with pumped systems. This requires more accuracy in estimating "net"
79

driving forces. Moreover, there are considerable "knowledge-based" uncertainties associated

with factors on which the magnitude of these forces and counter-forces depends (e.g., values
of heat transfer coefficients and pressure losses). In addition, the magnitude of such
"natural" driving forces depends on the specific plant conditions and configurations which
could be existing at the time a system is called upon to perform its safety function. For
these reasons, the reliability of the thermal-hydraulic (T-H) performance of the "passive"
systems must be assessed.
The United States Nuclear Regulatory Commission (U.S. NRC) is currently reviewing the
AP600 design. The assessment of the T-H reliability of passive safety systems is a major
issue in the review of the AP600 design PRA because it can have a large impact on the
success criteria for systems and operator actions as well as on the structure of the event trees
themselves (e.g., by underestimating pressures which would cause safety relief valves to
open and potentially stick open or by not modeling human actions that would change the
course of an accident).
ASSESSMENT OF THERMAL-HYDRAULIC PERFORMANCE UNRELIABILITY
The quantification of T-H unreliability for all accident sequences involves statistical
evaluations which require a prohibitively large number of computations. The conservative
risk-based "margins" approach, presented in this paper, eliminates the need to quantify the
T-H unreliability for most, if not all, accident sequences.

The T-H unreliability of a sequence for which a best estimate T-H analysis predicts no core
damage (success sequence) is defined as the probability that this sequence will actually lead
to core damage because of uncertainties in the predicted T-H performance of passive
systems. With this definition, the problem of assessing the T-H unreliability of passive
systems becomes equivalent to assessing the impact of uncertainties on the predicted value of
the variable used in the core damage criterion, such as peak cladding temperature (PCT) or
water level in the reactor vessel. Through the core damage criterion, the T-H unreliability is
linked to the success criteria of the passive systems and, thus, to the core damage frequency.
In the following discussions, a PCT value equal or greater than 2200 F is assumed to lead
to core damage. However, in the margins approach, presented in this paper, any other core
damage criteria can be used.
The value of the PCT predicted by a T-H computer code can significantly differ from its
actual value because of uncertainties, approximations and errors in the calculation, which are
driven by:

uncertainties in the values of code input variables,

approximations and uncertainties in modeling the physics of the process,

approximations in modeling the system geometry,

the use of numerical methods to solve the system equations.

80

The code input variables include:

initial and boundary conditions, such as initial plant temperatures, pressures, water
inventories (levels), and reactor power

forcing functions, such as decay heat and flow rates of active systems

physical properties, such as densities, conductivities, viscosities, and specific heats

dimensions, such as clad diameter and thickness and fuel-cladding gap

thermal-hydraulic parameters, such as heat transfer coefficients, friction factors and

gap conductance

timing of human actions (e.g., manual actuation of core makeup tanks or manual
depressurization in the AP600 design).

Examples of approximations and uncertainties in modeling the process physics are: (1) the
treatment of a liquid steam mixture as a homogeneous fluid, (2) the assumption of

thermodynamic equilibrium, (3) the use of correlations (e.g., correlations for heat transfer
coefficient, pressure drop coefficient and critical heat flux) and simple models (e.g., models
for slip, flooding, entrainment and critical flow).
To avoid excessive computational times, approximations in modeling the system geometry
are necessary. These involve simplification of complex geometry features and approximation
of a three-dimensional system by a two or one-dimensional system. In numerical
approximations derivatives are replaced by finite differences and the resulting error depends
on grid structure and time step size.

To assess the impact of uncertainties on the predicted value of the PCT, a large number of
calculations is needed. This number can be reduced by: (a) concentrating on sequences that
contribute significantly to core damage frequency, (b) grouping (binning) the accident
sequences and selecting a "bounding" sequence for each group (bin), and (c) performing
sensitivity analyses to determine the "large impact" variables, i.e., those variables whose
uncertainty has a significant impact on the T-H performance of passive systems. The effort

can be significantly reduced if a conservative bounding analysis can be used. A very

conservative bounding analysis can be based on values for all variables that can be justified
as bounding (as defined by their range of uncertainty). For sequences where conservative
bounding is not successful, a statistical approach can be used to assess the probability that the
limiting value of the PCT will be exceeded. Even for these sequences, the effort to assess
the T-H unreliability of passive systems can be reduced if a combination of conservative
bounding and statistical analysis can be used, where the statistical evaluation is performed for
a very small number of "large-impact" parameters, while a bounding approach is used for the
remaining (larger) number of parameters.
The statistical evaluation can be based on the generation of random samples (where
dependent variables are properly treated) of the "large-impact" variables by using a MonteCarlo method. The PCT can be computed for each sample, and the resulting values can be
used to assess the probability that the limiting value of 2200 F will be exceeded due to T-H
uncertainties. Depending on the size of the problem under consideration (number of "large81

impact" variables) and the information available about the problem, direct Monte-Carlo,
Latin Hypercube, or Monte-Carlo Importance sampling can be used. Direct Monte-Carlo
can be inefficient if the number of "large-impact" variables is not small. Similar statistical
approaches have been discussed extensively in the literature. This includes early work
performed by Westinghouse for EPRI (Ref. 1), work performed at national laboratories
(Refs. 2 and 3), and work performed for the CSAU (Code Scaling, Applicability and
Uncertainty) evaluation methodology (Refs. 4 and 5).

Differences between code predictions and actual values due to geometry simplifications (one
or two-dimensional), coarse spatial grids and large time steps can be assessed by comparing
predictions of simplified computational models with those of more detailed models (e.g., two
or three-dimensional, finer spatial grids, smaller time steps). Differences due to
approximations of the process physics should be evaluated by comparing predictions of
simple codes with those of more mechanistic codes and with experimental measurements.
Analyses must show that there are no surprises due to multi-dimensional effects, system
asymmetries, two phase flow instabilities, oscillatory system behavior, and the presence of
noncondensible gasses.
RISK-BASED MARGINS APPROACH

The risk-based margins approach is a graded approach which consists of four basic steps. In
step 1, the accident sequences are grouped into "bins" and a "bounding" sequence for each

bin is selected. In step 2, sources of uncertainty associated with the T-H performance of
passive systems are identified. In step 3, sensitivity analyses are performed to identify those
variables (such as T-H parameters) whose uncertainty has significant impact on the predicted
PCT. In step 4, for each of the "bounding" sequences identified in step 1, the available
"margin" to core damage is explored and this information is used to determine success
criteria for passive systems and operator actions or to decide on an appropriate resolution
path (e.g., need for more detailed analyses, need for regulatory oversight or design changes).

These basic steps of the margins approach are described in more detail below.
1.

Accident sequence grouping and selection of "bounding" sequences

The effort of assessing the T-H unreliability is drastically reduced if the success accident

sequences, modeled in the event trees, are grouped into bins and, for each bin, a "bounding"
sequence is selected. This is done using a combination of probabilistic and deterministic
arguments (e.g., considering sequences initiated by the same event with a frequency above a
certain cutoff and using similarity and bounding arguments to select a "bounding" sequence).
Analyses are then performed only for these "bounding" sequences since the results are good
(i.e., "bounding") for all the other sequences. To reduce the number of T-H computations,
several conservative assumptions can be made. For example, if the initiating event is a
break, the most limiting break in terms of size and location can be determined and used in

subsequent analyses.
When the T-H unreliability of passive systems is taken into account, some success sequences

(i.e., sequences for which a best estimate T-H analysis predicts no core damage) could
actually lead to core damage. Thus, additional cut sets could be generated which would
contribute to an increased core damage frequency (CDF). These new cut sets are the product
of three terms: (1) the initiating event frequency; (2) a non-T-H unavailability term, which
82

is the product of unavailabilities of all failed systems due to hardware failures and human
errors; and (3) a T-H unreliability term.
The magnitude of these three terms provide important clues about the risk importance of the
T-H unreliability term in a certain sequence. For example, if the product of the first two
terms is very small in a certain sequence (e.g., smaller than lE-9/year), then the T-H
unreliability for this sequence is not risk important and can be ignored. This implies that
success sequences with frequency below a certain cutoff can be ignored. Among the
remaining sequences, the lower the product of the first two terms, the higher the T-H
unreliability term must be to have a significant impact on CDF. This implies that success
sequences with many redundant system (and train) failures, such as those close to the bottom
of an event tree, will not become significant risk contributors unless the T-H unreliability is
very high. This is an important observation since such sequences are the most likely
candidate "bounding" sequences. If the margins approach fails to show margin for such
sequences, a combination of this approach with statistical analysis can be used. Since it must
only be shown that the T-H unreliability is not higher than a rather large number, the
computational effort will not be extensive. On the contrary, it is likely that an approach that
uses statistical analysis with only one variable (preferably having the largest impact on PCT)
will be successful. Statistical analyses with only one variable pose minimal computational
effort.
2.

Identify Sources of Uncertainty

For each "bounding" sequence, determined in the previous step, sources of uncertainty
associated with the T-H performance of passive systems must be identified. A systematic
process is necessary to ensure that uncertainties in code input variables as well as in software
correlations and models are considered. This process could start with the review of models
and parameters of a best estimate computer code in conjunction with available information
from existing studies to concentrate on the modeling of phenomena that dominate plant
behavior. For example, available information from the Code Scaling, Applicability, and
Uncertainty (CSAU) program [Ref. 6] could be used.

Since the computer codes used to predict the T-H behavior of a nuclear power plant are
approximate simulators of complex phenomena, the code predicted values of important plant
variables differ from the actual values of these variables. Typically, large T-H system codes
model 100 to 200 parameters associated with mass and energy transfer processes. Although
the value of all of them is known or is predicted with some uncertainty, for any given
accident sequences a small subset of these parameters dominate plant response.

To guide the code assessment (validation) effort and associated experimental programs for
computer codes to be used for audit calculations in the certification of Advanced Reactor
designs, the U.S. NRC is using the Code Scaling, Applicability, and Uncertainty (CSAU)
evaluation methodology, which was developed in the late 80's [Ref. 6]. The CSAU
methodology is based on the Phenomena Identification and Ranking Process (PIRT); i.e., the
identification of phenomena and their ranking in terms of their relative importance in
determining plant response during a given accident sequence. PIRTs provide assurance that
all phenomena having some significance to plant behavior have been identified and the effort
to assess uncertainties in computer code predictions is concentrated on the modeling of
phenomena that dominate plant behavior.

83

Based on preliminary computer code calculations and expert judgement, detailed PIRTs have
been formulated, documented, and peer reviewed for five postulated accident categories.
They are: (1) small loss of coolant accidents (LOCAs), (2) main steam-line break accidents
with depressurization, (3) main steam-line break accidents without depressurization, (4) steam
generator tube rupture accidents with depressurization, and (5) steam generator tube rupture
accidents without depressurization. These results have been directed towards the assessment
of RELAP5 and CONTAIN computer codes.
Future activities planned include validation of the PIRTs, first with sensitivity calculations
and later by the use of the experimental data as they become increasingly available. Refined
PIRTs will be used as the foundation for determination of the uncertainty in computer code
predictions, when these codes will be applied to audit safety analyses involving the transients
used in the development of PIRTs. The results that have been generated, as well as those
that will be generated in the future, by the PIRT process can provide very useful information
in the identification of "large-impact" variables and in assessing computer code biases for the
evaluation of passive system T-H unreliability.

3.

Identify "Large Impact" Variables

Once the sources of uncertainty are determined and the various parameters (variables) whose
uncertainty contributes to the T-H unreliability of the passive systems are known, sensitivity
analyses can be performed to identify "large-impact" variables (i.e., variables whose
uncertainty has a significant impact on the value of the PCT. In these analyses, the change

in the value of the PCT per unit change of the variable under consideration (i.e., 3PCT/3v,
where PCT = peak clad temperature, v = variable under consideration) as well as the
uncertainty range of the variable should be considered. For some parameters the derivative
may be large while the uncertainty range is small, for others the derivative may be smaller
while the range of uncertainty is larger. Since such variables are the most important
contributors to T-H unreliability, their identification is needed for efficient implementation of
the margins approach. Particular attention should be paid in selecting the range of
uncertainty for these variables. If more refined uncertainty analyses are needed to show that
the success criteria are conservative, analyses which would narrow the variation range of
some "large impact" variables would be the most beneficial.
4.

Explore Available "Margin" to Core Damage

Results from the first three steps of the margins approach, described above, contain all the
information that is needed to explore the available "margin" to core damage (for each
"bounding" sequence). A very conservative bounding analysis is performed first to
determine whether the assumed success criteria for passive systems and operator actions are
conservative. If this analyses fails to show that the success criteria are conservative, then
several options can be investigated to ensure conservative success criteria. Such options
include more detailed analyses of the variation range of some "large impact" variables and/or
use of experimental information and/or commitment by the vendor to perform pre-operational
tests and show that certain assumed values of variables are indeed bounding. If such options
are exercised but still cannot be shown that the success criteria are conservative, then the
option of quantifying the T-H unreliability and properly incorporating it into the PRA can be
exercised. This option may not be as demanding in computational effort if a combination of

84

conservative bounding and statistical analysis can be used, where the statistical evaluation is
performed for a very small number of "large-impact" parameters, while a bounding
(margins) approach is used for the remaining (larger) number of parameters. Quantification
of the T-H unreliability can be avoided if changes are made in either the design or the
success criteria which make the bounding analysis successful (i.e., with these changes the
margins approach shows that the success criteria are conservative).
The various steps that are used to explore the "margin" to core damage are listed below.
1.

Each "bounding" sequence should be analyzed first with values of (all) variables that
can be justified as bounding, to determine whether the assumed success criteria are
conservative (i.e., whether the predicted value of the PCT is less than its limiting
value of 2200 F after the computer modeling error has been accounted for).

2.

If the analyses of previous step fail to show margin for one or more "bounding"
sequences, then more refined analyses can be used to show that the safety system
success criteria for such sequences are conservative (e.g., perform more refined
uncertainty analyses on the variation range of some "large impact" variables and/or
use experimental information and/or commit to perform pre-operational tests and
show that certain assumed values are indeed bounding).

3.

If for some sequences the analyses of the previous two steps would fail to show that
the success criteria are conservative, i.e., the peak cladding temperature determined
from these analyses exceeds the limiting value of 2200F), then one or more of the
following options should be considered:

Quantify the probability that these sequences would lead to core damage because
of T-H performance uncertainties, and properly incorporate this probability in the
PRA analysis.

Consider regulatory oversight of active nonsafety-related systems. This would

increase the reliability of such systems and, in turn, would decrease the challenge
rate of the safety-related passive systems.

Consider design and/or success criteria changes which make the analyses of
either one of the first two steps successful (i.e., with these changes the analyses
show that the new success criteria are conservative).

SUMMARY AND CONCLUSIONS

The T-H unreliability of passive safety systems, such as those used in Westinghouse's AP600
design, must be assessed and accounted for in the PRA. However, the quantification of T-H
unreliability involves a prohibitively large number of computations. A conservative riskbased 'margins" approach was developed which eliminates the need to quantify T-H
unreliability for most, if not all, accident sequences. With this approach the issue of T-H
unreliability is addressed by linking it to the passive safety system success criteria and, thus,
to core damage frequency (CDF). This is achieved by assuring that the adopted success
criteria for safety systems and operator actions are conservative enough so that the
contribution to a sequence CDF from T-H uncertainties is significantly smaller than the
contribution from hardware failures and human errors.
85

The margins approach is currently being implemented for the analysis of the passive system
performance reliability of the AP600 design. This analysis, which is performed in support of
the AP600 design certification, has so far indicated that the conservative margins approach
can be a very efficient tool for addressing the issue of passive system T-H reliability in
advanced passive reactor designs. In addition, such an approach can drastically reduce the
effort needed to address this issue. Preliminary indications (no complete results are available
yet) are that there is adequate margin to core damage for most AP600 design accident
sequences. However, the success criteria for a few sequences may have to be changed.
REFERENCES

1.

M. Mazumdar, J.A. Marshall, and S.C. Chay, "Methodology Development for

Statistical Evaluation of Reactor Safety Analyses," EPRI-NP-194 (1976).

2.

J.K. Vaurio and CJ. Mueller, "Probabilistic Analysis of LMFBR Accident

Consequences with Response Surface Techniques," Nuclear Science and Engineering
75, 401-413 (1978).

3.

E.E. Morris, C.J. Mueller, I.E. Cahalan, and H. Komortya, "A Comparison Study of
Reactor Surface Methodology and Differential Sensitivity Theory in Core Disruptive
Accident Analysis," Proc. 1981 International ANS/ENS Topical Meeting on
Probabilistic Risk Assessment. September 1981.

4.

NUREG/CR-5249, "Quantifying Reactor Safety Margins," 1989

5.

M.G. Ortiz and L.S. Ghan, "Uncertainty Analysis of Minimum Vessel Liquid
Inventory During a Small-Break LOCA in a B&W Plant - An Application of the
CSAU Method Using the RELAP5/MOD3 Computer Code,"NUREG/CR-5818, 1992.

6.

C.D. Fletcher, G.E. Wilson, C.B. Davis, and T.J. Boucher, "Interim Phenomena
Identification and Ranking Tables for Westinghouse AP600 Small Break Loss-ofCoolant Accident, Main Steam Line Break, and Steam Generator Tube Rupture
Scenarios," INEL-94/0061 (Draft Report), October 1994.

86

FEASIBILITY AND EFFICIENCY STUDIES OF FUTURE

PWR SAFETY SYSTEMS

P. AUJOLLET
Commissariat a 1'Energie Atomique,
Cadarache, France
Abstract
The French Atomic Energy Commission (CEA) has initiated evaluation

studies of innovative safety systems, in terms of efficiency and reliability in selected

accidental sequences. Three kinds of systems are currently evaluated : direct
depressurization system designed to control PWR primary pressure to prevent in-

pressure core melt down, passive residual heat removal systems and steam injector
systems.
The feasibility and performance of each system is assessed through the
analyses of accidental transients in a reference three loops PWR, making use of the

Probabilistic Risks Analysis (PRA) method and the CATHARE 2 thermal hydraulic
code.
These studies are conducted in four steps : sizing of each system using one
or more reference transients, insertion of each new system ; modification of
accidental procedures and event-trees, calculation of fuel damages probabilistic risk
values during selected accidental transients and ranking of core damages
probabilistic risk values.
In particular, a high pressure single stage steam driven injector which was

tested at SET laboratories was modelled with the thermal hydraulic code
CATHARE.
INTRODUCTION

The present article summarizes some features of the PWR innovative safety systems wich are
currenly studied at CEA Innovative Safety Systems Department mainly with Cathare 2 code:
direct depressurization system sizing and opening setpoint value adjustment
passive residual heat removal: Secondary Condensing System and In-vessel Heat Exchangers

studies
Safety Injection System or Auxiliary Feedwater Systems with Steam injector. ENEL/SIET Steam
injector Mock up tests results (2) studies with Cathare 2 code
Beyond the specificities of these particular systems, current studies objectives are to verify the ability
of the concept to fullfill its function, to determine its physical generic limits, to foresee code developments
and experimental code assessment programs, to eventually improve the systems and determine
complementary safety systems.
DEPRESSURIZATION SYSTEM

This system must permit a direct control of the primary circuit pressure in extreme conditions. Its
objectives are to prevent a in-pressure core melt down by discharging primary circuit to a low pressure
value (fixed at 1 MPa in that case).
System description.
A valve system is located at the top of the pressurizer, the discharging system is connected to a
pool (DR.WST : In-Reactor Water Storage Tank) located inside the containment. These studies are

performed with Cathare Safety code reactor data set to simulate standart 900 PWR.

87

Applications and Objectives.

This system has to achieve a fast (hundred secondes) depressurization of the primary circuit to
reach and to maintain it at a pressure value lower than 1 MPa.
Method.

After a first sizing of the valve, opening setpoint value adjustment is achieved by simulation with
Cathare code of anticipated transients with multiple failures leading to extreme conditions with uncovered
core and primary circuit at a high pressure value.
The first transient chosen was a total loss of Auxiliary Feedwater Systems event with only
accumulators and LPIS available, during this transient the system has to remove heat from nuclear core
primary and secondary circuits.
Then several transients were tested. Among them, a small break Steam Generator tube rupture event
with only accumulators and LPIS available, gave a sizing point at low pressure as the depressurization
system with a decreased flowrate had to remove heat coming from the core, the primary circuit and from
the broken Steam Generator.
The current signal for valve opening is based on reactor vessel water level measurement with an
opening of the valve at hot and cold legs uncovering. CATHARE simulations results show the efficiency of
the depressurization system during these transients.
It was also verified that an inadvertant opening of the valve during reactor normal operating
conditions do not lead to core damages.
The size of the discharging pool was determined by considering the subcooled liquid water mass
needed to condensate the entire primary coolant mass in vapor phase.
Studies current status

Additional calculations are currently performed to verify the efficiency of the system during other
transients such as multiple Steam Generator tubes ruptures, or small break with primary pumps kept on.
All these results are to be compared with a fast depressurization valve opening test experimental
results which is to be performed on BETHSY integral test facility in 1996.
These studies are completed by a definition work of an experimental program on CEA Super Claudia
facility to modelize direct condensation of water/steam jets in a pool. Super Claudia experimental activities
will be focused on separate effect tests in order to measure fixed physical parameters influence on
condensation process.
RESIDUAL HEAT REMOVAL SAFETY SYSTEMS

Two kinds of passive systems are currently assessed :

At primary circuit level (Refroidissement du Reacteur au Primaire system RRP from
CEA/DER/SIS)
At secondary circuit level (Secondary Condensing System SCS)
Reacteur Refroidissement Primaire system [1].

Usually the residual power is removed in an active way in two steps : first by the Steam Generators (SG)
at high primary side pressure, second by Residual Heat Removal System (RHRS) at low primary side
pressure and temperature.
A new passive system supplying similar functions is described and analysed. This studies have been
performed with a standart 900 MWe PWR reactor Cathare 2 data set.
Description and principle. RRP system includes three in-reactor vessel heat exchangers. A layout of
reactor vessel internal structures to allow forced or natural convection and of the external heat sink are
shown in Fig. 1.
A suction system (Fig. 2) allows forced or natural convection around the integrated heat exchangers.
The heat sink introduced into the reactor vessel is placed above the core so natural convection can
occur inside the reactor vessel.
88

The size of the reactor vessel has to be increased with respect to current PWR reactor vessel sizes,
to integrate the heat exchangers.
Functional analysis and efficiency.
RRP system can work in a large range of conditions.
In normal operating conditions, convection in the intermediate circuit is available as primary pumps
are working, the suction system establishes convection around the heat exchangers; the heat removal
capacity of the system is about 80 MW. A cold shutdown can be conducted using only RRP system

If primary pumps are off, a short natural circulation loop between the core and the heat exchanger
is established . Primary water is circulating only inside the reactor vessel Under these conditions, it has
been shown that with only natural circulation heat exchanges in the intermediate circuit 30 MW can still
be extracted. These last results have been evaluated by CATHARE 2 calculation in the cases of a small
break or a total blackout with in both cases Auxiliary Feedwater System (AFS) unavailable These
accidental sequences were conducted in a totally passive way by the RRP system

Figure 1. RRP System

It must be noticed that RRP system still works with a primary low water level (under hot and cold leg
level)
System advantages.
Cold shutdown can be achieved without any extra system.
RRP system can work totally passively with no time limit (atmosphere as heat sink) at hot shutdown.
The primary fluid is always confined in reactor vessel.
RRP is always effective even if vessel water level is at primary hot and cold legs level
For small breaks and blackout kind of accidents, RRP system allows a significant increase of the
intervention delay and only low pressure safety systems needed
System drawbacks.
The inspection and maintenance operations are obviously more difficult to perform
It implies additive penetrations in the reactor vessel
Reactor vessel dimensions must be significally increased.
The system needs additive penetrations in the containment
Intermediate circuit water flows from the core vessel to the outside of the containment during accidental
transients

89

suction system

Figure 2. Reactor Vessel with RRP System

Current studies

.Linked to the drawbacks last point mentionned a RRP system with an intermediate circuit heat sink
created by a in-pool heat exchangers system located in the containment is currently studied
Secondary Condensing System.

Description and principle.

SCS is a passive system using natural convection. It has to fullfill the safety function of the classical

Auxiliary Feedwater System.

After closure of the main steam valves and opening of the SCS valves steam coming from the three
Steam Generators is condensed in heat exchangers immersed in three pools located outside the
containment, condensed water is driven back by gravity to Steam generator inlets. Fig. 3 shows a scheme
of this system.
Heat exchangers sizing is based on ability to remove residual heat during a loss of feedwater transient
with the application of the single failure criterion (that means roughly 45 Mw for one heat exchanger),
cooling rate limits, low steam temperature at low load and degraded steam generator mass inventorie. The
calculations were performed with the two phase flow COPE code to get an optimized couple tubes
number-tubes length.
Efficiency.

In the case of blackout sequence with all AFS unavailable, CATHARE 2 simulations have been
performed. Cathare results shows that, after opening of the SCS steam valves, elapsed time value to reach
the SCS liquid water maximum flowrate value seems to be convenient (less than 20 sec.), stationnary
flowing conditions inside the SCS are established after 40 sec.allowing heat removal from the primary
circuit untill standard low pressure decay heat removal system can be started.
Extra CATHARE calculations with degraded Steam Generator mass inventory or with one or two
SCS systems unavailable were performed, the results show that safe conditions are maintained as far as
natural circulation is maintained in the pools (water level)
Steam Generator Tube Rupture (SGTR) simulations have shown that an automatically started SCS
on Scram signal is particularly efficient even with a one hour delayed operator action.

90

Current studies status

. They are focused on pool characteristics adjustment with 3 D.rwo phase flow code calculations and on
EPICE experimental program definition to improve and validate Cathare in-pool (or low pressure) heat
exchangers modelling.

condensing
pod

Figures. SCS

91

STEAM INJECTOR SYSTEM

A Steam Injector (SI) is a device without moving part, in which steam is used as the energy source
to pump cold water from a pressure lower than the steam to a pressure higher than the steam. In SI all
thermodynamic processes rely on direct contact transport phenomena (mass, momentum and heat transfer)
between fluids, not requiring any moving mechanism. Schemes of both real [2] and modelled SI are shown
in Fig. 4.
SI Description(2).
SI can be divided in four parts :
a steam nozzle, producing a nearly isentropic expansion and partially converting steam enthalpy
into kinetic energy,
a water nozzle, producing a moderate acceleration and distributing the liquid all around the steam,
a mixing section, where steam and water come into contact. Steam transfers to water heat (because
of temperature difference), mass (because of the related condensation) and momentum (because of velocity
difference). The final result is the complete steam condensation, with a liquid outflow at relatively high
pressure.
diffuser, where the liquid kinetic energy at mixing section outlet is partially recovered
A numerical model has been developed using CATHARE 2.

Figure 4. Real and Modelled Steam Injector

Results of the numerical simulation.

Several experimental tests have been reproduced with CATHARE. And some unexperimented situations
have also been simulated. In particular, SI have been shown to still work even with very degraded steam
quality. Later, a similar case has been experimented by CISE, and it has confirmed our observation.
First, a standard case will be shown. Table 1 shows the conditions that have been used.

92

TABLE 1. Test conditions

water tank temperature

(C)
water tank pressure (MPa)
inlet steam pressure (MPa)
inlet steam temperature
(C)
discharge pressure (MPa)

12.7
0.2
5.14
275.3
5.58

Fig. 5 shows the pressure profile that has been calculated

In steam nozzle part, steam flow passes from a subsonic to a supersonic flow
At water source location, pressure is low enough to allow liquid water injection from water tank
In mixing section, a supersonic two-phase flow is observed, until complete condensation of steam
into water
At that full condensation point, a sharp pressure rise occurs, it corresponds to a supersonic-subsonic
transition
In the last diverging part, pressure is recovered from kinetic energy
-IIIr-

.3e06

06

Source

09

X uas(m)

Overflow

Figure 5. Pressure Profile in SI

SI Systems applications.

Several kinds of safety injection systems using SI can be considered.

Primary Side Safety Injection System. For that kind of systems, the most interesting source of steam
is the pressurizer. Indeed, it seems very attractive during an accidental transient to condense steam coming
from the core vessel through the pressuriser with subcooled water coming from a safety tank and re-injecte
the water liquid flow in the primary system. It is obviously a very efficient way to depressurize primary
circuit. It must be noticed that, with an optimal adjustment, it could be possible to continuously replace a
steam mass by the same mass of subcooled water. WHh such a kind of systems, SGTR management could
be particularly eased.that sharp steam quality and inlet pressure variations due to the pressuriser two phase
flow would not stall SI.
The main problems to solve are coming from steam injector unstability (it cannot be restarted after
stalling) and, from overflows control during Steam Injector starting procedure (it is not proved that SI
would start with two phase flow) and, when running, as the overflow has to remain open
It has to be mentionned also that CATHARE simulations results indicated that non consabte gas
presence seems to affect SI functionning.
93

An other point to be explored by experimental means is the behaviour of diluted boron in a

supersonic SI and, in general, the boron concentration control with a steam injector.
Auxiliary Feedwater System. Systems using SI for secondary safety injection seem, a priori,easyer to
built, as there is no diluted boron to inject and no uncondensable gas (at least in unbroken steam
generators in the case of steam generator tube rupture transient)

But the main point is : can SI systems be more reliable than existing systems? . More complete
studies are necessary to answer that question.

Current studies status

They are in two directions:
Steam Injector Cathare 2 modelling improvments to adapt the heat exhanges model (gas phase at a
supersonic velocity) as there are still discrepencies between experimental data results and Cathare results
Auxiliary Feedwater System using several Steam Injectors Cathare 2 PWR data set, coupling of the
steam injector data set with a 900 Mw PWR Cathare 2 data set in the way of accidental transient (Loss of
Feedwater ) numerical simulation.
ABBREVIATIONS

AFS : Auxiliary Feedwater System

IRWST
: In-Reactor Water Storage Tank
LPSIS : Low Pressure Safety Injection System
PWR : Pressurized Water Reactor
PRA : Probabilistic Risks Analysis
RHRS . Residual Heat Removal System
RRP : Refroidissement du Reacteur au Primaire
SCS : Secondary Condensing System
SG
: Steam Generator
SGTR : Steam Generator Tube Rupture
SI
: Steam Injector
REFEBENCES

[1] Gautier, G-M., Dispositif d'Evacuation de la Puissance Residuelle du Coeur d'un Reacteur
Nucleaire a Eau Pressurisee, Patent n 92 05220, April 28, 1992.

[2] Cattadori, G., Galbiati, L., Mazzocchi, L., Vanini, P., A Single-Stage High-Pressure Steam Injector
for Next Generation Reactors : Test Results and Analysis, European Two-Phase Flow Group Meeting,
University of Hannover, Germany, June 7-10, 1993.

94

RISK REDUCTION POTENTIAL OF JET CONDENSERS

A.W. REINSCH, T.G. HOOK

Southern California Edison Co., San Clemente,
California, USA
K.I. SOPLENKOV, V.G. SELIVANOV, V.V. BREDIKHIN
Research Institute of Nuclear Power Plant Operation,
Moscow
I.I. SHMAL, Y. FILIMONOV, Y.N. PAVLENKO
Electrogorsk Research Engineering Center,
Electrogorsk

Russian Federation
Abstract

Ejectors serving as high-pressure condensers produce a self-controlling heat

removal process induced by the dynamic form of natural convection. The simplicity
of passive heat removal systems with ejector condenser (PAHRSEC or jet condenser)
reduces the likelihood of component failures and human errors. Designed for a heat
removal rate exceeding the decay heat, the system has the capability for internal
depressurization of the reactor coolant system without loss of primary or secondary
coolant. In the event of a small loss-of-coolant accident (LOCA), jet condensers
reduce the reactor system pressure to the point where the shutdown cooling system
can operate. Simulations of the response of the 1,100 MWe pressurized-water reactors
at San Onofre with two PAHRSEC loops to transients and small LOCAs determined the
success criteria for the system. Jet condensers designed for internal depressurization
can cope with more than 99% of all initiating events in terms of frequency. Based on
a probabilistic safety assessment, the system would reduce the frequency of core
damage at the San Onofre Plant by a factor of 2. This value applies to internal
initiating events only. If external events are included in the evaluation, the risk
reduction factor would increase significantly. In the event of a steam generator leak
or tube rupture, a release of fission products to the atmosphere would be prevented
by the closed jet condenser loop.

1.

INTRODUCTION

In the early phase of the advanced reactor program, some aspects of nuclear
safety were defined in absolute terms: Passive systems were considered "inherently"
safe and, therefore, essential for advanced reactor safety. The increased use of
probabilistic safety assessment in the design and operation of nuclear power plants
has led to a more balanced approach, replacing qualitative terms with quantitative
measures.
While physical effects utilized in passive systems tend to increase system
reliability because of the absence of moving parts and support systems, the reliability
of core cooling is not necessarily improved significantly by the increased system

95

reliability itself. A more effective way to reduce the core damage frequency (CDF) is
to develop systems with the capability to respond to the majority of initiating events
and, consequently, terminate a broad spectrum of accident sequences.

Another important aspect in the development of advanced reactor systems is

cost. From the beginning of the nuclear era, attention had been focused on safety at
any price. As a result, nuclear power became too expensive. Electric power generated by recent light-water reactors in the United States costs currently twice as much
as power from combined-cycle gas turbine plants. Wall Street analysts predict now
that the competitive pressure developing in the U.S. utility industry since the passage
of the National Energy Policy Act will lead to the permanent shutdown of 20 to 25
nuclear power plants in the next 10 years. In this economic environment, the objective
for advanced safety systems must be a significant risk reduction at the lowest possible
cost.
2.

PRINCIPLE

2.1

PAHRSEC flow scheme

The essential feature of the concept is the utilization of the reactor decay heat
to produce pumping power for coolant circulation [1,2]. Although comprehensive
computer analyses and large-scale testing of the concept is of relatively recent origin,
considerable experience in the design, analysis and performance verification of the
system has been gained in the last years [3].

A high-pressure jet condenser (Figure 1) converts about 5% of the decay heat

into kinetic energy by expanding steam from the steam generator in a convergent
nozzle. For a heat removal rate of 75 MWt, the kinetic energy of the steam jet at sonic
velocity is about 4 MW. The steam mixes in the ejector with water at lower temperature from a heat exchanger and condenses. Kinetic energy of the mixture is converted
at the increased density into a pressure rise of the fluid sufficient to return the
condensate to the steam generator and recirculate coolant through the heat exchanger
serving as heat sink. Most of the kinetic energy in the steam jet is converted into heat
during mixing with water at low velocity. Less than 10% of it produces pumping power
for fluid circulation and heat removal. The overall ejector efficiency in converting the
decay heat into pumping power is less than 1% which reflects the simplicity and
ruggedness of the component.
One important characteristic of the concept is that the whole loop including the
heat exchanger is at the same pressure as the steam, supplying high-pressure water
to the "suction" side of the ejector. This characteristic extends the operating range of
the PAHRSEC ejector to a much higher pressure level than that of conventional singlestage ejectors whose maximum discharge pressure is about 2.5 MPa.
2.2

Dynamic natural convection

Fluid circulation in jet condensers can be considered the dynamic form of

natural convection. Unlike the static form of natural convection which is driven by
hydrostatic pressure differences, dynamic natural convection uses inertial forces to

96

MSIV

Heat Sink

Figure 1: Principle of jet condenser (PAHRSEC) as

heat removal and depressurization system

generate flow [4]. The system's process-inherent control function regulates the steam
pressure and the mass flow rate depending on the decay heat generated in the core.
Two-phase hydrostatic natural convection is susceptible to flow oscillations since
condensation phenomena affect the steam flow to the condensation surfaces. This
feedback between the process and its power supply leads to instabilities such as
chugging. In the jet condenser, critical flow in the ejector steam nozzle separates the
mixing and condensation process downstream of the nozzle from the steam supply
upstream, preventing feedback between condensation and steam flow. Flow perturbations originating downstream of the ejector steam nozzle cannot travel through the
critical flow region at the nozzle exit into the steam pipe upstream of the nozzle. This
characteristic ensures smooth and stable operation under all transient conditions.
3.

SAFETY-RELATED SYSTEM CHARACTERISTICS

3.1

Internal depressurization

The risk reduction potential of jet condensers is determined by the combined

safety functions of heat removal and internal depressurization. Designed for a heat
removal rate exceeding the reactor decay power, the system has the capability to
depressurize the reactor without coolant blowdown by lowering the reactor coolant
system temperature.
The effect of the internal depressurization is similar to steam relief through
atmospheric dump valves or an automatic depressurization system. However, the
coolant is not discharged from the closed PAHRSEC loop into the atmosphere or a
suppression pool at low pressure. Rather, it is returned to the heat source at high
pressure, conserving coolant and preventing the release of fission products to the
environment. Current active designs compensate for the loss of coolant through the
relief valves by coolant injection with feedwater pumps.

97

3.2

Spectrum of initiating events mitigated

Jet condensers can cope with a broad spectrum of initiating events. Passive

heat removal protects the reactor against all transients including loss of offsite power
and station blackout. The additional capability of internal depressurization provides

protection against LOCAs with an equivalent diameter of up to 5 cm. This increases

the spectrum of events covered by jet condensers to over 99% of all initiating events
in terms of frequency. In a typical pressurized-water reactor population, initiating
events that exceed the mitigation capability of PAHRSEC loops would occur only about
every 500 reactor-years.

3.3

Reduced human error potential

A single response - opening of the startup valve - is required from PAHRSEC

for all initiating events to function. Thermal equilibrium between decay power and heat
removal from the reactor coolant system is established rapidly after reactor shutdown.

The water returned to the steam generator by the ejector is heated by mixing with
high-pressure steam, reducing thermal stresses significantly. Because of these
operational characteristics, automatic PAHRSEC actuation is recommended after plant
trip, obviating the need for human interaction. Human errors during maintenance and
testing are unlikely owing to the system's simplicity. Inadvertent actuation of a jet
condenser caused by operator error or a spurious signal results in a reduction of the
steam pressure by less than 5%. This pressure reduction would not trip the plant.

3.4

Process-inherent control

Utilizing the reactor decay heat as power source for the heat removal process
provides process-inherent control of the coolant flow to match the decay heat. Less
decay heat means also reduced pumping power for fluid circulation, keeping the
temperature distribution relatively stable. With decreasing decay heat, the pressure in
the system decreases, lowering the density of the steam in the ejector nozzle and,
consequently, the steam and water flow rates.
The controlling phenomenon of the process is critical flow at the exit of the
ejector steam nozzle. The sonic velocity of steam remains almost constant over a wide
pressure range. Therefore, a constant volume flow of steam is discharged by the
ejector steam nozzle into the ejector mixing section. If the decay heat decreases
requiring less coolant flow, a reduced saturation pressure in the steam generator (and

a lower steam density) satisfy the energy balance by reducing the mass flow rate in
the ejector steam nozzle.

3.5

Prevention of initiating events

The heat removal rate of a PAHRSEC loop is designed to be approximately

equal to the decay heat following reactor trip. Since the flow and heat removal rates
depend on the geometry only and are not influenced by outside parameters such as
valve alignment, control systems, pump speed, electric power and operator actions,
a quasi-equilibrium is reached rapidly after startup of the system. The capability to

98

reduce pressure and temperature transients after reactor trip can prevent certain
initiating events. For example, the lifting of a pressurizer relief valve led to the accident
at Three Mile Island. An instant balance between decay heat generation and heat
removal after reactor trip reduces the potential for initiating events such as pressure
transients, pressurized thermal shock, cooldown transients, pump seal failure and
water hammer.
3.6

Dose reduction

Similar to a condenser driven by hydrostatic forces, the jet condenser forms a

closed loop with the steam generator, condensing high-pressure steam and returning
the condensate to its source. Heat is rejected to the environment by heat conduction
through the heat exchanger walls without mass exchange between the closed jet
condenser loop and the environment. Fission products carried by the steam into the
loop are returned with the condensate to the steam generator. In contrast to hydrostatic natural convection, however, the condenser surfaces are not exposed to twophase flow, preventing flow oscillations, erosion and heat exchanger leaks.

3.7

Removal of non-condensibles

The liquid state of the coolant in the heat exchanger loop makes the separation
of non-condensibles (for example, H2 or N2) feasible. Connected to the pressurizer
steam space of a pressurized water reactor or the steam line of a boiling-water reactor,
PAHRSECs would remove gases from the primary system during a severe accident
exceeding the core design limits. This capability eliminates the failure mode of hydrostatic condensers caused by accumulation of gases blocking heat transfer.
3.8

Passive accident management

Two of the most effective actions in accident management are heat removal and
depressurization. Both functions are immediately available upon PAHRSEC actuation
after reactor trip and would prevent core damage. However, if the startup of the
system would be delayed until the onset of core damage, the system would provide
passive accident management by reducing pressures, temperatures and the hydrogen
concentration in the reactor coolant system without operator action or support
systems. At high steam pressure, the heat removal and depressurization rates would
be high. The rates would decrease with the decay heat and system pressure.

3.9

Reliability of jet condenser

Only one valve has to be actuated to start the system. Its independence from
electric power and support systems increases the reliability, resulting in a low number
of failure modes and high system reliability. However, the reliability of the system
contributes only little to the increased reliability of core cooling. The type of safety
function and the safety-related system characteristics are more important in enhancing
reactor safety than the system reliability. These characteristics of the jet condenser
concept have been discussed in Sections 3.1 through 3.7.

99

During normal operation of the reactor, the startup valves to the PAHRSEC
loops are closed. After reactor trip, the valves are opened to start the heat removal
process. The pressure in the system before startup is essentially the same as the
steam generator pressure. Only a small pressure difference across the closed startup
valve exists which is created by a water column of about 5 meter in the jet condenser.
Owing to the small load on the valve disc, little force is necessary to open the valve for
startup, increasing the reliability of the valve.
4.

SUCCESS CRITERIA AND SYSTEM SIMULATION

The jet condenser success criterion for decay heat removal and
depressurization following a small break LOCA was determined by transient system
analysis using the Computer Code MAAP 3B (Modular Accident Analysis Program).
The code had been used to establish the system success criteria for San Onofre's
probabilistic safety assessment, the Individual Plant Examination, mandated by the U.S.
Nuclear Regulatory Commission. MAAP simulates the effects of a wide range of
accident scenarios and phenomena including the effect of engineered safety features,
operator actions and changes in geometry during a severe accident.
The code was used to analyze the response of a pressurized-water reactor with
3,390 MW thermal power (2 steam generators, 4 reactor coolant pumps, San Onofre
Unit 2/3) to a 5cm LOCA using two 75MW jet condensers as heat sinks. A failure of
all coolant injection pumps and feedwater pumps was postulated.

During a small LOCA, heat removal from the reactor coolant system by the two
PAHRSEC loops competes with the heat loss through the break by coolant leakage.
The energy removed by the systems is not available for forcing coolant out the break.
The net effect is a continuous reduction of the saturation temperature and pressure in
the reactor coolant system, reducing the density of the two-phase flow discharged
through the leak and, therefore, the leakage flow significantly.
Results of the transient MAAP analysis of a 5cm LOCA are shown in Figure 2.
Temperature, pressure, flow and leak rates decrease gradually. The core is never
uncovered during the transient. Without any active safety system available, the water
inventory in the reactor coolant system decreases initially. However, the accumulator

tanks supply coolant at the same rate as the decreasing leakage rate, maintaining a
nearly constant coolant inventory after the first hour of the accident. After the first
hour, the pressure of the reactor coolant system has been sufficiently lowered to allow
the use of the shutdown cooling or low-pressure injection system.
5.

PROBABILISTIC ASSESSMENT OF RISK REDUCTION

An evaluation of the impact of passive heat removal and depressurization

effected by two jet condenser loops on the core damage frequency of a 1,100 MWe
pressurized-water reactor has been performed. The analysis is based on the models
of the Individual Plant Examination (IPE) for San Onofre Units 2 and 3. The IPE was
mandated by the Nuclear Regulatory Commission as a plant-specific probabilistic
safety assessment for all U.S. nuclear power plants.

100

7P = - 9 - 2

Time [s]
Figure 2a:

Primary system pressure and steam generator pressure

as a function of time after reactor trip

Time [s]
Figure 2b:

xlO

10

Peak clad temperature, core water temperature and steam

generator temperature as a function of time after reactor trip

101

The code QUIKRISK was used to compute the reduction in core damage
frequency resulting from heat removal/depressurization by jet condensers. QUIKRISK
was developed to determine the impact of changes in the configuration of a nuclear
power plant on the core damage frequency. QUIKRISK cuts the time for the solution
of probabilistic models significantly, allowing the analyst to investigate various plant
configurations in a limited time. The code uses existing fault tree models of San
Onofre 2 and 3 developed for the Individual Plant Examination. The input for the code
consists of configuration changes caused by maintenance, repair and testing as well
as design changes or component failures.

A base case without any plant changes was quantified with QUIKRISK first. The
core damage frequency of the plant design configuration (without jet condenser) was
computed by QUIKRISK as 2.8E-5 per year. Subsequently, the fault tree models of the
auxiliary feedwater system and LOCAs with less than 5 cm diameter were changed to
incorporate the jet condenser safety function. After adding the jet condenser to the
models, the core damage frequency decreased to 1.4E-5 per year.
These results do not take into account the safety characteristics discussed in
Section 3.5. In addition, the consequences of core damage would be significantly
reduced by the system characteristics discussed in Sections 3.6 through 3.8. External
initiating events are not expected to cause LOCAs with diameters larger than 1 cm.
Medium and large LOCAs are, therefore, not part of the external initiating event
spectrum. Consequently, jet condensers can cope with all external initiating events.
Including external events in a probabilistic safety assessment increases the system's
risk-reduction potential significantly.
6.

EXPERIENCE AT U.S. NUCLEAR POWER PLANTS

Current risk assessment techniques use plant-specific failure probabilities where

available and generic numbers for rare events. In order to gather information on rare
events, the experience with safety-related incidents in the entire population of U.S.
nuclear power plants has to be reviewed.

Cable fires, cooldown transients, steam generator tube ruptures, loss of offsite
power, station blackout and one small LOCA were the events with the highest impact
on public safety at commercial nuclear plants. In areas with frequent seismic activity,
the nuclear risk is increased significantly by potential earthquakes.
A review of these initiating events indicates that jet condensers can prevent core
damage following all of these events and keep transient pressure and temperature
changes in the reactor coolant system within safe limits. For loss of offsite power and
station blackout, core damage is prevented since the safety function of the passive
heat removal system is independent from electric power. Fire affects automatic control
systems and electric power supply which are not required for PAHRSEC operation.
A cooldown transient in a nuclear power plant is highly unlikely if the decay heat
is removed by jet condensers. The fixed geometry of the system makes a sudden
imbalance between decay power and heat removal impossible, maintaining a quasiequilibrium with a steady temperature distribution.

102

Small LOCAs with up to 5 cm diameter have been discussed in Section 5. The

analysis using the MAAP Code proved that core temperatures and leakage from the
reactor coolant system are continually decreasing with 2 PAHRSECs operating. The
core remains covered throughout the transient. A window of several hours is available
for switching from the PAHRSEC loops to shutdown cooling or low-pressure injection.
In the event of a steam generator tube rupture, fission products leaking into the
steam generator would be retained in the closed PAHRSEC loop. Both steam
generators would be effective for heat removal and depressurization. Identification of
the leaking steam generator and/or operator decisions are not required. Similar to
a small LOCA, depressurization of the primary and secondary sides reduces the
leakage flow continually. The loss of reactor coolant inventory is made up by coolant
injection from the accumulator tanks.
Earthquakes contribute significantly to the core damage frequency of nuclear
power plants in seismically active regions. Although the interaction of numerous minor
malfunctions such as relay chatter causes a large increase in risk, no major structural
damage to the nuclear system such as a large LOCA or a vessel failure is expected.
Since PAHRSEC loops are independent from electric power and control systems, they
have the capability to mitigate all consequences of seismic events and effect a safe
shutdown.
7.

SUMMARY

The core damage frequency of light-water reactors can be reduced significantly

by passive heat removal using jet condensers. Utilizing the decay heat to induce
coolant circulation and system depressurization, jet condensers have the capability to
terminate the majority of potential accident sequences in light-water reactors. The
simplicity of the system lowers the likelihood of human errors significantly. Cost
reductions during construction and operation are anticipated.

REFERENCES

[1 ]

SOPLENKOV, K. I., NIGMATULIN, B., Passive heat removal system with

ejector-condenser, American Nuclear Society Winter Meeting, San
Francisco, California, November 10 -14, 1991.

[2]

Reinsch, A. W., Process-inherent operational safety through jet

condensers, Second International Conference on Thermal Hydraulics and
Nuclear Power Plant Operation, Tokyo, May 14-16, 1986.

[3]

Soplenkov, K.I., Selivanov, V.G., BREDIKHIN, V.V. et. al., Design and

testing of passive heat removal system with ejector-condenser, IAEA

Technical Committee Meeting, Piacenza, Italy, May 16-19,1995.
[4]

REINSCH, A.W., Static versus dynamic natural convection, American

Nuclear Society Annual Meeting, Nashville, Tennessee, June 10 - 14,
1990.

103

DIVERSIFIED EMERGENCY CORE COOLING IN CANDU

PJ. ALLEN, NJ. SPINKS

Atomic Energy of Canada Ltd.,
Canada
Abstract

The CANDU low pressure heavy water moderator surrounds the fuel channels and is
available as a heat sink to maintain channel integrity and avoid fuel melting in the unlikely
event of a loss of coolant accident with loss of emergency coolant injection. Existing
CANDU reactors use pumps to circulate the moderator and the cooling water.

A passive moderator heat rejection system is being developed to diversify the emergency heat
rejection paths to the environment. The moderator heat rejection system, as described in
reference 1, incorporates a natural circulation of heavy water driven by flashing to steam as
the heavy water flows to an elevated heat exchanger. The heat exchanger is cooled in turn by
a natural circulation flow of light water to a large reservoir.
Analysis has shown that this system can reject the moderator heat in a stable manner. As

further confirmation, a full elevation, light water, l/60th volume scaled test of the natural
circulation heavy water loop has been carried out to verify the overall concept and the
analysis. In particular, the stability of the flashing driven flow has been confirmed and will
be reported in reference 2.
With careful attention to eliminating common mode failures in the shutdown systems, the
heat transport system, the emergency core cooling system and the moderator system, the core
melt frequency can be reduced to the point that core melt mitigation for events internal to the
plant ceases to be a design concern. These same design provisions will form additional lines

of defence for all but the most extreme external events as well.
Introduction
As in other pressurized water reactors, the primary accidents of concern for CANDU are
those that involve a loss of coolant (LOCA), or failures in systems that could induce a
LOCA (such as a loss of flow, loss of heat sink). For these accidents, the requirements of
shutdown, cool, contain and monitor apply.
Where CANDU differs from standard pressurized light water reactor technology is primarily
the design of the reactor core. The CANDU core design offers some unique possibilities for
design of mitigating measures that are diverse in concept from the traditional technology.
These differences have been exploited to a degree in existing CANDU designs. For future
designs, concepts are being developed to enhance system diversity to the point that core melt
mitigation for internally caused events ceases to be a design concern.
Diversity in Current CANDU Designs

Contemporary CANDU reactor designs employ redundancy and a considerable level of

diversity in the safety systems. Thus redundancy and diversity exist in the two shutdown
105

systems. Both shutdown systems make use of the low-pressure moderator environment (the
calandria) but shutoff rods enter the calandria from above whereas the poison injection
system enters the calandria from the side, as shown in figure 1. Each system has its own
initiating signals. For every postulated accident, there are normally two diverse signals on
each shutdown system to trip the reactor for the complete range of initial operating
conditions. Note that the two shutdown systems are passive in that no operator action and no
external power are needed for shutdown action to occur.
XS

ACT,V,TYMMA*-0

AMD CqMJftOLAKOMC* UNITS

TYftCAl AMAMGCMCNT *ON

ZOC CONTROL AND Vf HTtCAL
H.UX OCTf CTOft UNITS

CALANMIA
VAULT LINCA

cmoeruec
ooo ooa
o oo oo ooo <
oooooooo<
OOOOO OOO<
OOOOOOOO <

aoooouc

OOO

OO O V O <

O O O OO O C
O OOO OO OO <
OO OO O O C > O O O O O O O O <

O O OOOOC

oo ooo o t

OOOOOOOOi

O O OO OO C

OOO OOO OO<

> OOO OOO OO <
OOOOOOOCM

OOO OOOOO
OOO OOO OO

OOO OOOOO
O OO OO O
UOOCftATOM
WLfT
WXCTIOM
SHUTOOVMUNIT
CALANORIASMELL

MODERATOR OUTLCT

FIGURE 1
CANDU CORE SHOWDVG SHUTDOWN SYSTEM ARRANGEMENT

Redundancy also exists in the emergency-core-cooling systems. In the event of a loss-ofcoolant accident, the Emergency Core Cooling (ECC) system uses pressurized gas (or pumps
in some plants) to inject light water into the heat transport system. The water is eventually
recovered from a sump in the reactor building, cooled in a heat exchanger and pumped back
into the heat transport system.
106

The low-pressure low-temperature moderator serves as a redundant and potentially diverse

emergency-core-cooling system. Its availability during normal operation is apparent and it
acts passively to accept heat from the fuel channels in an accident involving loss of coolant
with loss of ECC. However moderator heat rejection is currently done with electrically
powered pumps and pumped cooling water. The electrical supply and cooling water system
also serve the ECC. The common electrical supplies and cooling water limit the combined
reliability of the emergency core cooling systems.

Core melt in CANDU can occur only with a loss of coolant (whether caused by pipe failure
or a support system failure that induces loss of coolant by system overpressure and relief)
with loss of ECC plus loss of the moderator heat sink. Common failures in these systems,
e.g. loss of service water or electrical power to both the moderator and the ECC systems,
are the main contributors limiting the core-melt frequency for internal events to about 4x10*
per year (ref. 3). While this core melt frequency meets targets set for advanced designs, the
ability of the moderator to act as a heat sink provides the opportunity to progress further by
eliminating common links between the moderator and other systems.
A conceptual CANDU design is under study which employs a conventional ECC system with
a passive moderator heat rejection system. Thus passive design techniques are used to
advantage in enhancing the diversity in the two core cooling systems.
Passive Moderator Heat Rejection

Progress on the passive moderator system development was last given in reference 1. Figure
2 illustrates the concept.
LEVEL 1 Steam generators boil stored water
to the atmosphere via main steam
safety valves (MSSV).

LEVEL 2 High-pressure injection

followed by pumped recovery of
emergency coolant. Heat transferred
to pumped emergency water.
LEVEL 3 Moderator thermosyphons through
heat exchanger. Heat transferred
to water jacket.

WATER JACKET Containment wall contains

water-filled annulus. Water
thermosyphons through heat
exchangers for level 3 cooling.
Heat transferred to air flowing
upwards by natural convection.

LEVEL 2

CONTAINMENT Steel shell designed for

pressure and temperature of
large break.

Emergency
Coolant Injection
Emergency Water Supply

Emergency Power Supply

FIGURE 2
CANDU EMERGENCY COOLING

107

The idea is to run the heavy water in the calandria at a temperature near the boiling point but

to allow the water to flash to steam as it rises in a pipe from the calandria to an elevated heat
exchanger. Subcooled heavy water would be returned to the calandria. The difference in
density between the two-phase flow in the riser and the liquid in the downcomer would
provide the buoyancy force to drive the flow.
Reference 1 gives results of simulations using the CATHENA transient thermalhydraulics
code which demonstrates that the normal full power heat load to the moderator can be

transferred in a stable manner with such a design. Note that the heat load to the moderator
during a loss of coolant accident, with the reactor at decay power, is only 30% of normal
full power moderator heat load.
More recently, further CATHENA simulations have been done at reduced powers. They
show a flow oscillation at low power. Also tests have been done in a full elevation loop
having a scale of about 1/60 in power, volume and flow area. They confirm the CATHENA
predictions. The tests will be reported in more detail in reference 2.
As power is increased, flashing is first observed in the transparent glass riser at upper
elevations. The flow is oscillatory with the riser being liquid filled after the high-flow part of
a cycle. No untoward effect of the oscillations is evident. As the power is increased, the
oscillation amplitude decreases and the flow becomes stable. During a rapid increase of
power, only one or two oscillations are seen as the flow overshoots before returning to the

steady-state value. Thus the feasibility of the flashing-driven design is considered to be

established both for the normal operating condition and for accidents.
Loss of. Shutdown

As in other reactors, the CANDU reactor must be shutdown following an accident so that the
mitigating systems can deal with the consequences. The unique CANDU geometry has
allowed the implementation of a second shutdown system which is diverse from the shutdown

system using rods and equally capable in terms of speed and more capable in terms of
reactivity depth. Both shutdown systems are totally independent of the control system (no
rods or controls are shared) and each shutdown system has independent instrumentation and
trip circuitry. Both shutdown systems make use of the low-pressure moderator environment

(the

calandria) but shutoff rods enter the calandria from above whereas the poison injection

system enters the calandria from the side, as shown in figure 1. For every postulated
accident, two diverse signals are normally provided on each shutdown system to trip the
reactor for the complete range of initial operating conditions.
The careful attention given to diversity in the design of CANDU and its two safety shutdown
systems has led to common mode failures being reduced to a very low probability. This fact
means that a simple product of the failure frequency of events leading to overpower (beyond
the capability of the reactor control system) and the unavailabilities of the two shutdown
systems reflects properly the order of magnitude for loss of shutdown events. Based on the
200 unit-year CANDU operating record, improvements made to earlier reactor control
systems and more recent operating experience, the frequency of challenges to the shutdown
systems is now about 10"2 per year. Also the two shutdown systems are each required to

108

have an unavailability of 10~3 and continued operation requires testing to ensure that this
figure is met. Thus the frequency of loss of shutdown is of order
10'2 x 10'3 x 10'3 = 1Q-8 per unit-year.
See also the figure of 2.5 x 10'8 per unit-year calculated in more detail in reference 3.
Core Melt

Core melt can occur in CANDU only with a loss of heat transport system inventory (whether
caused by pipe failure, valve failure or loss of heat sink), a loss of ECC and a loss of the
moderator as a heat sink. Reference 3 gives a detailed account of the accident sequences
leading to core-melt for CANDU 6. The sequences are dominated by common cause events
such as loss of service water and loss of electrical power. These contribute most of the coremelt frequency of 4.4x10"* per unit-year. The triple failure events involving LOCA with loss
of ECC and loss of the moderator heat sink contribute only 0.6x10"* per unit year and this
figure is dominated by the continuing need to keep moderator and ECC pumps running in the
longer term.

With diversity in the heat transport system, the ECC system and the moderator system, the
core melt frequency reduces to a simple product of the failure frequency of the heat transport
system and the unavailabilities of the ECC and the moderator systems. From the 200 unityear operating record for CANDUs, and the single event (Picketing unit 2, 1994) which
required actuation of ECC, the failure frequency is order 10"2 per unit year. Also the ECC
system is required to have an unavailability of 10"3 and in-service testing is done to ensure
that this figure is met. However the ECC pumps have to keep running for an extended period
and an unreliability of 10~2 has been calculated (reference 3) over the mission period.
A passive moderator heat rejection system which does not rely on continued operation of
pumps should be more reliable than a pumped system and is likely more reliable than the
ECC system. Also it is possible to operate the moderator system continuously so that its
availability is assured. It follows that an unavailability better than 10"3 is achievable and
given a low probability of cross link failures, the frequency of core melt becomes on the
order of

10'2 x lO'2 x 10'3 = 10'7 per unit-year.

The key issue is the probability of crosslink failures between the mitigating systems (ECC
and the moderator) and the initiating accidents. Given the diverse nature of ECC and the
moderator, the probability of crosslinks between these two systems is low. The remaining
question is whether the initiating accidents can cause failures in either mitigating system.

As in other water cooled reactors, the ECC system is designed to be independent, to the
degree possible, from events that could lead to a loss of coolant. Standby equipment is

provided to support the ECC function that is also independent of the normally operating
process equipment used to cool the fuel.
The proposed arrangement for cooling the moderator is also independent of the normal
process fuel cooling and the ECC cooling. The main interface with either system is due to
109

the physical arrangement of the calandria. If a fuel channel failure is the initiating event, then
the LOCA discharge is into the moderator fluid. Calculations have been done (reference 1)
for this situation and the results show that the heat load from the discharge can be
accommodated by the moderator cooling system.
The remaining issue is calandria leakage following a LOCA event. Loss of the moderator
fluid would disable the moderator as a emergency heat sink. Two mechanisms exist which
could lead to calandria leakage. They are addressed in existing designs by pumping makeup
water to the calandria. For the passive design, reliance on pumps is to be avoided.
One mechanism is a fuel channel failure which leads to ejection of the channel endfitting.
For a passive design, there needs to be a restraint to limit end fitting movement following a
channel failure. The other mechanism is a inlet feeder pipe break which, with the reactor at
high power, diverts sufficient flow from the channel to cause overheating and channel
failure. Eventually the moderator water could drain through the failed channel and broken
feeder. For the passive design, early reactor trips which can prevent channel failure will be
installed.

Development programs are ongoing to address these issues.

External Events

Given that the passive moderator equipment is inside the containment/shield building, it can
be protected from most external hazards by the surrounding structure. Well tested
engineering design approaches can be used to protect the containment contents from flood,
high wind, tornado missiles, etc. This situation is different than for more conventional core
cooling systems that depend on power and cooling water coming from outside the
containment. Such sources of power and water are more vulnerable to effects from external
events.

In addition, seismic qualification would not seem to be problematic for a passive cooling
system of this type.
Since survival of the passive moderator system can be assured for foreseeable external
events, core melt frequency will also be reduced for external events.

Conclusion

A passive moderator heat rejection system is being developed for CANDU which, combined
with a conventional emergency-coolant injection system, will provide the diversity to reduce
core-melt frequency to order 10~7 per unit-year. This is similar to the approach used in
contemporary CANDU designs in the shutdown systems which results in a frequency of loss
of shutdown on the order of 10* per unit-year.
Testing of a full height 1/60 power-and-volume-scaled passive-moderator-heat-rejection
system has demonstrated its feasibility for removal of heat during normal operation and
during accidents.

110

With the frequency of severe accidents caused by internal events reduced to order 107 per
unit year by these measures, no need should exist for consideration of core damage states
more severe than the moderator acting as a heat sink. Providing a reliable and diverse
alternate emergency heat sink reduces the severe accident challenge to containment integrity
and provides more assurance of the release to the public and the environment being limited.
REFERENCES

1. W.P. Baek and NJ. Spinks, "CANDU Passive Heat Rejection Using the Moderator",
International Conference on New Trends in Nuclear System Thermohydraulics, Pisa, May
1994.

2. H.F. Khartabil and NJ. Spinks, "An Experimental Study of a Flashing-Driven CANDU
Moderator Cooling System", for CNS Conference, Saskatoon, 1995 June.

3. Report AECL-9607,"CANDU 6 Probabilistic Safety Study Summary", 1988 July.

Ill

MITIGATION OF TOTAL LOSS OF FEEDWATER EVENT

BY USING SAFETY DEPRESSURIZATION SYSTEM

YOUNG M. KWON, JIN H. SONG, SANG Y. LEE,

SANG K. LEE
Korea Atomic Energy Research Institute
Taejon, Republic of Korea
Abstract
The Ulchin 3&4, which ore 2825 MWt PWRs, adopted Safety Depressurization System
(SDS) to mitigate the beyond design basis event of Total Loss of Feedwater fTLOFWJ. In
this study the results and methodology of the analyses by CEFLASH-4AS/REM for the
determination of SDS bleed capacity are discussed.
To verify the results of CEFIASH-4AS/REM simulation a comparative analysis has a bo
been performed by more sophisticated computer code. RELAP5/MOD3. The TLOFW
event without operator recovery and TLOFW event with feed and bleed {F&BJ were
analyzed. The predictions by the CEFLASH-4AS/REM of the transient two phase system
behavior are in good qualitative and quantitative agreement with those by
RELAP5/MOD3 simulation. Both of the results of analyses by CEFLASH-4AS/REM and
RELAP5/MOD3 have
demonstrated that decay heat removal and core inventory
make-up can be successfully accomplished by F&B operation during TLOFW event for
the Ulchin 3&4.

1. Introduction
Following the Three Mite Island accident, the potential ability of Power Operated
Relief Valves {PORVsJ to provide an alternate method to remove decay heat from the
primary system was identffted and considered to be beneficial in dealing with severe
accidents. Recent studies [1, 2, 3, 4] have concluded that F&B can be a viable
alternate means of decay heat removal, but successful use of F&B is contingent upon
the implementation of proper procedures, as well as upon the specific plant design.
ABB-CE's latest plant design. System 80+[5], includes manual bleed valves to provide the
F&B capability according to the USNRC's Severe Accident PoBcy.
The Korea Atomic Energy Research hstitute (KAERI) is performing detailed design of
the SDS similar to that of System 80+. In particular, manually-actuated bleed valves are
designed to provide a capability to rapidly depressurize the Reactor Coolant System
(RCS) for TLOFW event. Presented in Reference 6 are the preliminary results of
thermal-hydrauBc analyses of TLOFW for the Ulchin 3&4, which were performed by
CEFLASH-4AS/REM [7]. h this study provided are the final results and methodology of
the thermal-hydrauic analyses to determine the Ulchin 3&4 SDS bleed capacity. Also an
alternate analysts by more sophisticated computer code RELAP5/MOD3[8] is performed
to verify the results and methodology used for Ulchin 3&4 SDS design.

2. Plant description and initial conditions

The Ulchin 3&4 are two loop 2825 MWt PWRs. The Ulchin 3&4 are designed with two
cold tegs per loop and thus contain four reactor coolant pumps. The SDS consists of
two separate lines connected to the top head of the pressurizer and the flow through

each Ene discharges to the containment atmosphere through a rupture disc as shown in
Figure 1. The two bleed paths consist of an isolation valve and control valve in series
per path, and provide redundant paths. The plant nodal diagrams for CEFLASH-4AS/REM
and RELAP5/MOD3 are provided in Figure 2.

113

FRE88URIZER
SAFETY
. VALVES (>

: MOTOR OPERATED

PRESSURIZER

Fig.l A Schematic Diagram of UCN 3&4 SDS

(a) CEFLASH-4AS/REM

(b) RELAP5/MOD3.1

Fig.2 The Nodalization Scheme of UCN 3&4

114

The plant initial conditions are assumed at full power steady state nominal conditions.
Table 1 provides major plant parameters. Abo provided are steady state initial
conditions obtained by two computer codes. The results of initialization indicate that the
two initial conditions are essentially same.
Table 1. Plant kiftial Conditions and Major Plant Parameters
a. Plant Initial Conditions and Steady State Value
Parameter
Design Value
Steady State: CEFLASH/RELAP5
Core power (MWt)
2815
2815/2815
RCS pressure {MPa}
15.5
15.5 / 15.48
RCS flowrate (ton/hr)

55113

55113 / 55113

Cold teg temperature (C)

Hot teg temperature (C)

295.8
327.3

295.8 / 296.7
327.3 / 327.7

SG pressure {MPa)

7.5

RCS Inventory (Kg)

SG hventory at Rx trip(Kg)

N/A
N/A

73. I 73.7

211000/216200
41400 / 41100

b. Major System Parameters

Primary side volume (m3)
329.4
Pressurizer volume, liquid/total (m3)
25.5/51.4
Low SG level reactor trip setpoint (% WR)
38.5
SIAS setpoint (MPa)
12.6
HPSI pump shutoff head (MPa}
12.65
PSV setpoint (MPa)
17.2
PSV capacity (steam at 17.2 MPa), per valve (Kg/hr)
247517
Number of PSVs
3
2
Analytical bleed area (m )
0.0026

3. Analyses methodology
3.1 Design criteria
The use of F&B is a trade-off between allowable time before operator action and the
bleed capacity of the system. The longer the time, the larger the system capacity must
be. A shorter allowable time before operator action increases the possibility of
inadvertent actuation and resultant containment contamination. Therefore, appropriate
design criteria are required. Fallowings are design criteria selected for the Ulchin 3&4: ))

Each SDS flow path, in conjunction with one of two HPSI pumps, is designed to have a
sufficient capacity to prevent core uncovery following a TLOFW if one SDS path is
opened simultaneously with the opening of the PSVs. 2) Both SDS bleed paths are
designed to have sufficient total capacity with both HPSI pumps operating to prevent
core uncovery following a TLOFW event if the feed and bleed initiation is delayed up to
thirty minutes from the time of the PSVs lift.
The analysis procedure for the bleed capacity starts with a base case in which the
bleed paths are not available, i.e., no operator action is assumed. This base case yields
the time of PSVs Bft and core uncovery. The duration between PSVs lift and core
uncovery is the maximum theoretical allowable time for the operator to open the bteed
paths to prevent the core uncovery. All subsequent cases are analyzed with F&B
operation. The analytical bteed path area required to prevent core uncovery were
investigated in conjunction with operator action time for each F&B cases.

115

3.2 Differences in analytical models

This analyse employs two analytical models, CEFLASH-4AS/REM computer code
developed by ABB-CE and REIAP5/MOD3 computer code version 3.1 developed by
NEL CEFLASH-4AS/REM has been improved from the CEFLASH-4AS[9] which is used for

icensing analysis of small break LOCAs. Reference 10 provides the validation of the
CEFLASH-4AS/REM against experimental data to verify the capability of the code for use
in the analysis of a TLOFW event with F&B. The CEFLASH-4AS/REM (simply, CEFLASH)
employs two mass, two energy, and one mixture momentum equations. Since the
CEFLASH solves only mixture momentum equation, various constitutive relations using drift
flux model are employed. RELAP5/MOD3 (simply, RELAP5) employs two-fluid,
nonequifibrium, nonhomogeneous, hydrodynamic model (six equations) for the transient
simulation of the two-phase system behavior.
Lice all other computer codes, RELAP5 and CEFLASH are limited by the phenomena logical models built into the codes, h addition, RELAP5 and CEFLASH have different
nodafization scheme; RELAP5 permits the user to vary the nodafization. On the other
hand, the CEFLASH has a customized nodafization scheme as shown in Fig.2-(a).

4. Simulation results and discussions

4.1 Determination of bleed capacity
CEFLASH is used for the simulation of TLOFW event without operator recovery and
TLOFW event with F&B. The assumptions used in the simulation of these transients are: 1)
The plant initial conditions are at steady state full power condition. 2) A reactor trip
occurs due to tow steam generator level 30 seconds after event initiation. 3) The
Reactor Coolant Pumps (RCPs) are tripped 10 minutes after the reactor trip
per
Emergency Procedure Guidelines[14]. The 10 minutes operator action time is based on

the fact that the operator should trip the all RCPs in the Optimal Recovery
GuidelnefORG) for the Loss of All Feedwater. For the TLOFW event Bke scenarios the
operator can diagnose the event as Loss of All Feedwater easily, since the system
pressure will increase rapidly in couple of minutes after reactor trip due to heat transfer
degradation. Since the operator can trip the RCPs in the main control room, the 10
minutes operator action time can be justified. 4) The operator actions are considered
according to the design criteria discussed in section 3.1.
For the F&B cases the single failure case and no failure case are considered, where
single failure implies operation of only one HPSI pump and opening one SDS bleed path
and no failure means operation of two HPSI pumps and opening of two bleed paths.
The analytical bleed capacity to prevent core uncovery are investigated by varying the
analytical bleed area in TLOFW simulations. The operator actions coincident with PSVs
fft and 10 minutes after the PSVs lift are considered for the single failure case. For the
no failure case 2 minutes, 10 minutes. 30 minutes, and 40 minutes are considered. Figure
3 shows analytical bleed area required to prevent core uncovery for various operator
action times for the single failure case and no failure case determined by CEFLASH.
Since the analytical bleed area to meet the second design criterion (28 cm2) is
smaller than twice the analytical area to satisfy the first design criterion (40 cm2) as can
be seen in Figure 3, the first design criterion is more restrictive with respect to bleed
capacity. Theses results are in the same trend as the preliminary analyses results
presented in Reference 6. However, current analysis does not assume charging pumps
operation, since Chemical and Volume Control System (CVCS) including charging
pumps is not credited as safety system. The design bleed capacity is selected as 26
cm2 by accounting for the various uncertainties, such as, valve stroke time, decay heat
curve, and code uncertainties.
To verify the results and methodology of CEFLASH analyses, comparative analyses
have also been performed by RELAP5. The cases selected for presentations are TLOFW

116

without recovery and single failure case, hi the next sections discussions are focused on
the results of simulations using 26 cm2 bleed capacity.

4.2 TLOFW without recovery

Table 2 provides major chronology of the event predicted by CEFLASH and REIAP5.
Fig.4 shows pressurizer pressures predicted by the CEFLASH and RELAP5. Following
reactor trip the RCS pressure drops due to a sudden decrease in heat generation from
Table 2. Chronology of the TLOFW Event

Bleed Area (m2}

Feed flow

TLOFW w/o Recovery

(REM/RELAP5)
0
No HPSI

Event
Total toss of feedwater
Reactor trip
RCP trip, manual
Steam generator dryout
PSVs open
SDS bleed path(s) opens
HPSI flow on
Hot teg saturation
Core uncovery begins
or
Minimum RV inventory. Kg
occurred at. sec

TLOFW with F&B

(REM/RELAP5)

0.0026
1 HPSI

Time (seconds)
0/0
30/30
630/630
1360/1600
1389/1345
N/A
N/A
2923/2875
5296/ -

0/0
30/30
630/630
1360/1600
1389/1345
1389/1345
1511/1385
1510/1420
N/A
47300/44300
3280/3600

the core. After a short time period, the RCS pressure starts to rise in response to the
power-to-flow mismatch and reaches to a new steady state. After the RCP trip at 630
seconds, the pressurizer pressure increases more rapidly due to RCP coast down. Fig.5
shows steam generator inventory. When both steam generators dry out at about
1360/1600 seconds (The dryout time can be determined from the liquid inventory in the
RELAP simulation. Since only mbcture inventory has physical meaning in the CEFLASH
simulation, the time when the mixture inventory flattens out corresponds to dry out time
in the CEFLASH simulation), the RCS volume expansion and pressurization is accelerated.
Then the pressurizer pressure reaches the PSVs opening setpoint. Since the PSVs have
enough capacity to accommodate the increased volumetric expansion, the pressurizer
pressure is maintained around PSVs setpoint during the whole transient. It is shown that
the pressurizer pressures are in good agreement between two simulations. The primary
temperature rises until it reaches the saturation temperature corresponding to the
pressurizer safety vatve setpoint as shown in Fig. 6. From that time on, the primary
temperature stays constant while void is generated in the RCS.
Pressurizer goes sold at about 2500 seconds as shown in Fig.7, and the discharge
flow becomes single phase liquid. After the RCS reaches saturation condition around
3000 seconds, steam generated "m the core due to decay heat flows from the core to
pressurizer via hot leg and surge fne. Once the surge line begins to draw vapor, the
net inventory in the pressurizer drops rapid ry because low quality mbcture is still flowing
out of safety vah^es. Fig.8 shows the integrated surge flow and PSV discharge flow. The
integral surge flows predicted by CEFLASH and RELAP5 are "m good agreement before
hot teg is highty voided. After that, two predictions deviate slightly, which might be due
to the absence of countercurrent flow model for the surge Pne in the REM.

117

Core uncovery begins at around 5300 seconds into the transient, when the RCS
inventory becomes so tow that the void tractions in the top three nodes of the core
reaches 1.0 in the RELAP5 (refer to Fig.9 for Reactor Vessel(RV) liquid inventory and
Fig.10 for core void fraction). At that time, the collapsed water level "m the core begins
to decrease rapidly and the cladding and core outlet temperatures begin to rise. It is
observed that the vapor fractions in the core region tend to hang up in the 25-40%
range for extended period of time as shown in Fig.10. When core inventory decreases,
it frequently takes place in such a manner that the local vapor fraction jumps from 0.4
to 1.0 over a very brief time interval as shown in Fig.10. This is purely a result of the
flow regime map change in RELAP5. As shown "m Fig.9 the RV inventory behavior is quite
different between two predictions after the RCS starts to void. This difference "m RV
inventories is judged to be from the difference in inventory distribution among various
RCS components including RV upper head, pressurizer. hot legs, cold legs and SG
U-tubes. It is observed that the drainage of RV upper head inventory after hot leg
voiding predicted by RELAP5 is much faster than that predicted by CEFLASH.

4.3 TLOFW with feed and bleed

Presented in this subsection are the results of the case where F&B operation is utilized
to attempt to cool the core and make up the RCS inventory. The assumptions used in
the simulation of this transient are: 1) Operator opens one train of SDS bleed path and
aligns one train of HPSI pump for injection at the time of PSVs' lift. 2) The SDS is
modeled by an orifice located on the top of the pressurizer whose analytical bleed
area corresponds to 26 cm2.
This case is identical to TLOFW without recovery case until PSVs lift. Table 2 provides
chronology of major event scenarios predicted by CEFLASH and RELAP5. Soon after the
bleed path is opened the RCS pressure decreases rapidly as shown in Figure 11, and
hence HPSI injection flow is initiated at 1511/1385 seconds. A significant amount of
energy is removed through the SDS bleed path when the discharge flow is a single
phase steam. However, as the flow leaving the SDS path becomes two-phase, the
energy removal slows down. Hence the RCS pressure decrease also slows down. And
the RCS pressure briefly begins to repressurize when the discharge flow becomes single
phase liquid. At that point the pressurizer becomes solid (refer to Fig.11,12 and 13). The
pressurizer behavior during repressurization is generally in good agreement between
CEFLASH and RELAP5. CEFLASH shows rather smooth pressurizer pressure transient. The
pressurizer inventory keeps increasing after 4200/5400 seconds as can be seen in Fig. 12.
The RCS becomes saturated quickly after the bleed valves open. The steam

generated in the core migrates from the core to the pressurizer. which increases the
amount of steam bubbles in the pressurizer and consequently increases break flow
quafty. As the quafty of bleed flow increases, the energy removal rate increases. This
results in the decrease in the pressurizer pressure. Therefore, the HPSI injection flow s
reinitiated (See Fig. 14). As the RCS depressurizes, HPSI flow with low temperature at 50
C also increases, which contributes to further reduction of RCS pressure as shown in
Fig.11. The combination of opening the SDS bleed path, which results in loss of RCS
inventory, and the HPSI injection of cold fluid, which lowers the RCS average
temperature and therefore leads to contraction of RCS fluid, eventually causes voiding
"m the RCS. Void formation is evident in the core as early as 2000 seconds into the
transient. However the void fraction at the top of the core is maintained below 40%
due to increased HPSI injection fbw (See Fig.15). The reactor vessel inventory reaches
minimum at 3600 seconds in RELAP5 predictions and continuously increases as shown in
Fig. 16. The CEFLASH prediction shows that core mixture level is always maintained
above the top of the core as shown in Fig.17. This resutt indicate that the selected SDS
bleed capacity meets the first design criterion discussed in section 3.1.

118

As shown in Fig.16 the reactor vessel inventory behavior is quite different between
two predictions after RCS starts to void. This difference in RV inventories is judged to be
from the differences in inventory distribution among various RCS components. Therefore,
further study on the nodalization scheme of RV upper head is recommended. However,
this difference does not affect the conclusion of this analysis, since the event scenarios
before the time of minimum RV inventory are almost the same for both cases and both
predictions show that core is covered two-phase mixture.
The peak cladding temperature is calculated to evaluate the impact of core
voiding. As shown in Fig.18 the cladding temperature is well below acceptance criteria,

which assures core to coolant heat transfer is well maintained.

5. Summary and conclusions

The SDS bleed capacity b determined by numerical simulation of TLOFW event
without operator recovery and TLOFW event with F&B by CEFLASH computer code. The
analytical bleed capacity to prevent core uncovery are investigated by varying the
analytical bleed area and number of operating HPSI pumps.
To verify the results of CEFLASH simulation a comparative analysis has also been
performed by more sophisticated computer code. The predictions by the CEFLASH
simulation of the transient two phase system behavior is found to be in good
agreement with those by the RELAP5 simulation, except the RCS water inventory
distribution which shows small difference after the hot leg voiding.
h conclusion, the results of analyses for TLOFW event with F&B by CEFLASH and
RELAP5 have demonstrated that decay heat removal and core inventory make-up can
be successfully accomplished by F&B operation for Ulchin 3&4 Nuclear Power Plants.

CMCttlFAUIRI

x f

HO FAILURE

OCCFLWH

to

I
I
.
S

5
Io

10

20

SO

*>

*O

<0

TOIE AFTER PSV OPEN (MINUTES)

Fig.

3 Analytical Bleed Areas Required to Prevent

Cora Uncovery for Various Operator Times

1000 2000

3000

400O

$000

COM

TIME (SEC)

Fig. 4 Pressurlzer Pressure {TLOFW w/o Recovery)

119

700
X RELAPI
OCEFLASH

<rs

5?
n

tso

400

tOO

1200 HOO

WOO

HOC

1000 2000

TIME (SEC)

Rg.5 SG Inventory(TLOFWw/oRecovery)

3000

4OOO

5000

000

TIME (SEC)

Fig. 6 Hot Leg Temperature (TLOFW w/o Recovery)

1000

2000

3000

4000

SOOO

TIME (SEC)

Fig. 7 Normanzad Presttirixer Level

(TLOFW w/o Recovery)

Fig. 8 Integrated Surge and PSV Flows

(TLOFW w/o Recovery)

REULPS
X h Hod* (lop)
O 5th Hod.
A dHod*

IX Nod* (bottom)

OJO

1_
>

1000 2000

3000

4000

SOOO

(000

TIME (SEC)

Fig. 9 Reactor Vessel Inventories

(TLOFW w/o Recovery)

120

Fig. 10 Core Void Fraction {TLOFW w/o Recovery)

I
t>

I . . , , !

'0

1500 MOO

4MO

000

7500

tOOO

1100 MOO

4500

HOD

TSOO

MOO

TIME (SEC)

TIME (SEC)

Rg. 12 Normalized Pressurtzer Level (F & B)

Fig. 11 Pressurtzer Pressure (F & B)

X RCLAPS

X RHJkPS

OCEFLASH

OCEFIASH

t
15

c!

1500 MOO

4500

WOO

7SOO

*0

1SOO 1000 4500

TIME (SEC)

MOO

7500

MOO

TIME (SEC)

Rg. 13 Bleed Path Discharge Flow Rate(F&B)

Fig. 14 HPSI Injection Flow (F & B)

1.2
KELAft

XRELAPt
OCtFlASH

X Wl Hod* (tap)

1.0

1JO

OtOlNwl*

A*dNo4*
1*H<x (bottom).

IE
0.4

1 "

1SOO

9000

4SOO

COOO

TIKE (SEC)

Fig. 15 Core Void Fraction (F & B)

7500

(000

1900 3000

4JOO

000

7SOO

TIME (SEC)

Fig. 16 Reactor Vessel Inventory (F & B)

121

I i i ' ' I ' ' i

OCEFLASH

RELAPS
X hNod*<top)
OMhNod*
OTltlNod*

(21

A 3rd Nod*
1>t Nod* (bottom)

2"
I

111

cc
UJ

7.5

a.
X

$5

iu
-

7.0

top of dh COT <.73 m)

S

1SOO

3OOO

4SOO

(07500
00

1500

3000

4500

COM

7SOO

TIME (SEC)

TIME (SECONDS)

Rg. 18 Fuel Cladding Temperature (F & B)

Fig. 17 RV Core Mixture Level (F&B)

REFERENCES

1. H. Komoriya and P. B. Abramson, "Decay Heat Removal During a Total Loss ot

Feedwater Event for a C-E System 80 Plant." ANL/LWR/NRC 83-6, Argonne National

Laboratory. 1983.
2. Boyack. B. E. et aL, Los Alamos PWR Decay Heat Removal Studies Summary
Results and Conclusions, NUREG/CR-4471. Los Alamos Laboratory (1985).
3. Loom's. G. G. & Cozzuol J. J., Decay Heat Removal Using Feed-and-Bteed for U.S.
PWR, NUREG/CR-5072. Idaho National Engineering Laboratory, 1988.
4. Young Seok Bang, Kwang Won Seol and Hyo Jung Kirn. "Evaluation of Total Loss of
Feedwater Accident/Recovery Phase and Investigation of the Associated EOF,
Journal of the Korean Nuclear Society, VoL 25. Number 1, March 1993.
5. Asea Brown Boveri - Combustion Engineering. CESSAR-DC Chapter 5. 1989.
6. Kvvon. Y. M., Song, J. H.. & Ro. T. S* Decay Heat Removal Capability of Safety
Depressurization System for Total Loss of Feedwater Event. Proc. 5th hternational
Topical Meeting on Reactor Thermal HydrauEcs. Salt Lake City. VoL 6. 1992.
7. C-E Power Systems. Reafstic Small Break LOCA Evaluation Model. CEN-373-P. 1987.
8. K. E. Carbon et aL, "RELAP5/MOD3 Code Manual Volume k Code Structure, System
Models, and Solution Methods (DRAFT). " June 1990.
9. C-E Power Systems. "CEFLASH-4AS, A Computer Program for the Reactor Slowdown
Analysis of the Small Break Loss of Coolant Accident." CENP-133P. 1974.
10. C-E Nuclear Power Systems, Software Verffication and Validation Report for
CEFLASH-4AS/REM. W-FE-0063, 1992.

11. J. A. Trapp and V. H. Ransom. "A Choked-Ftow Calculation Criterion for

Nonhomogeneous, Nonequi&brium, Two-Phase Flows," International Journal of
Multiphase Flow, 8. 6, 1982.
12. T. M. Anktam and M. D. White, "Experimental Investigation of Two-Phase Mixture
Level Swell and Axial Void Fraction Distribution under High Pressure, Low Heat Flux
Conditions in Rod Bundle Geometry," Proceedings of ANS Small Break SpecieBst
Meeting. 1981.
13. Walfis. G.. One Dimensional Two-Phase Flow. McGraw-Hill Book Company, 1969.
14. C-E Power Systems. Combustion Engineering Emergency Procedure Guidelines,
CEN-152. 1987.

122

DESIGN AND ANALYSIS OF ADVANCED WATER COOLED

REACTOR SAFETY COMPONENTS AND SYSTEMS

(Session III)
Chairman

J. LILLINGTON
United Kingdom

EUROPEAN PRESSURIZED WATER REACTOR

CONFIGURATION, FUNCTIONAL REQUIREMENTS
AND EFFICIENCY OF THE SAFETY INJECTION SYSTEM

F. CURCA-TIVIG
Siemens AG (KWU), Erlangen,
Germany

J.L. GANDRILLE
Framatome, Cedex,
France
Abstract
The paper presents the thermal-hydraulic behavior of the European Pressurized Water Reactor
(EPR) under LOCA conditions and demonstrates the efficiency of EPR's Safety Injection System (SIS). Safety
criteria as well as performance of the SIS - i.e., main functions, functional requirements, design and emergency
core cooling mode - are discussed. The thermal-hydraulic response of the EPR to various LOCAs was analyzed
by performing best-estimate or conservative LOCA-calculations with the advanced computer codes CATHARE,
RELAP5/MOD2 and SPC-RELAP5. A large spectrum of leak sizes, from small leaks ($ 25 mm ) up to double
ended guillotine breaks, was simulated. The calculation results demonstrate the efficiency of the SIS: a) In case
of LOCAs caused by very small breaks ($ ^25 mm) the EPR's SIS is capable - in conjunction with the
automatic partial cooldown of the secondary side - to prevent loop draining, even when only two Medium Head
Safety Injection (MHSI) trains are effective. In case of small and intermediate leaks (<|>150 mm) the SIS
ensures sufficiently high injection rates for preventing core uncovery, again even when only two MHSIs are
effective. For leaks within the range 150 i $ 350 mm. the SIS is able to limit the core uncovery. b) In case of
large break LOCAs the SIS ensures a fast reflood of the core within approximately 120 s; the peak cladding
temperature is 810C. Reflood of the core is achieved by the accumulator injection.

1.

INTRODUCTION

The paper presents the thermal-hydraulic behavior of the European Pressurized Water
Reactor (EPR) under Loss-of-Coolant Accidents (LOCAs) conditions and demonstrates the
efficiency of EPR's Safety Injection System (SIS). Safety criteria as well as performance of
the SIS - i.e., main functions, functional requirements, design and emergency core cooling
mode - are discussed.
The EPR is a PWR which synthesises the experience gained by EdF, German Utilities,
Framatome and Siemens during the design, manufacturing, construction and operation of
numerous nuclear power plants with a total capacity of 100 000 MWe and an operation time
of 900 years. The EPR is an evolutionary reactor; additionally, it has innovative features in
technical areas such as prevention and mitigation of severe accident scenarios. The EPR
Basic Design is performed under the leadership of Nuclear Power International (NPI), which
is a joint company formed by Framatome and Siemens.
2.

SAFETY INJECTION SYSTEM DESIGN

The design of the EPR is based on the defence in depth principle. The choice of the
events to be considered for EPR design and safety assessment is firstly done deterministically. They consist of the normal operational states and are enlarged by systematically
looking for events having the potential of disturbing the reactivity or power control, the heat
removal from the fuel elements and the confinement of radioactivity.

125

A number of representative initiating events is derived from this systematic approach,

which lead to bounding cases for design and assessment of safety-classified systems,

components and structures. The deterministic design basis is in full compliance with the
approach which was already followed in France and Germany in the past. According to their
expected frequencies, the events are divided in four Plant Condition Categories (PCC):
normal operation (PCC 1), anticipated operational occurrences (PCC 2), infrequent
accidents (PCC 3) and limiting accidents (PCC 4).
Probabilistic targets have been set up for the EPR too. They comprise two safety
objectives: firstly, integral Core Melt Frequency (CMF) considering all plant states and all
types of events < 1O5 y1, and secondly, integral large release frequency < 10-6 y1. In order
to meet the safety objectives, the following design targets are defined: a) integral CMF for
internal events, reactor in power state: 1O6 y1, b) shutdown states shall contribute to CMF
less than power states, and c) integral CMF for internal events associated with early loss of
containment: 10-7y1.
In order to meet the probabilistic targets, the deterministic design basis is extended. In
this design extension, a limited number of accidents with multiple failures and coincident
occurrences including the total loss of some safety-grade systems will be selected by
probabilistic assessment This assessment will permit to design a limited number of additional systems or to adapt existing systems for additional backup functions. The accidents
are divided in two Risk Reduction Categories: firstly, prevention of core melt (RRC-A) and
secondly, prevention of large releases (RRC-B).
The present section gives an overview of the SIS design; safety and operational functions, functional requirements and a brief description are included.
2.1. Safety Functions
-

The safety functions of the EPR's SIS are the following:

Rapid reflood of Reactor Pressure Vessel (RPV) and reactor core following a LBLOCA.
Provision for long-term injection of water to the core for small, intermediate and LBLOCA.
Cooling of the injected water in case of LOCA, in order to terminate the release of steam
to the containment atmosphere as early as possible.
Provision for water injection to the Reactor Coolant System (RCS) for small to intermediate break LOCA or Steam Generator Tube Rupture (SGTR), at any pressure less than
the discharge pressure of Steam Generator (SG) safety valves.
Provision for cooling of the In-containment Refueling Water Storage Tank (IRWST), in
case of LOCA.
Mixing of water recirculated during the long term core cooling after LOCA to ensure homogeneous boron concentration and temperature.
Boration of the RCS for PCC 3 and PCC 4 accidents.
In conjunction with discharge from the RCS via the pressurizer safety valves, provision for
injection of water in order to ensure residual heat removal (RHR) from the RCS and
cooldown to a cold condition, in case of loss of RHR via the SGs.
Provision of backup to RHR system, as a different means of ensuring RHR, with the capability for cooling of the RCS when the temperature is < 100 C ano the pressure is low.
Provision for emergency makeup to the RCS in case of loss of water inventory during cold
shutdown or refueling shutdown.

2.2. Functional requirements

The functional requirements of the EPR's SIS are the following:
- For large and intermediate break LOCA the SIS shall be capable of ensuring that
acceptable core damage limits (see Section 2.3) are not exceeded even assuming a
single failure (SF).
- The cooling capability of the SIS shall be sufficient to ensure termination of steam release
to the containment following a LBLOCA, without exceeding a mean temperature of 110 C
in the IRWST, even assuming a single failure.

126

- For small break LOCA of PCC 3, the SIS shall be capable, if necessary in conjunction with
secondary-side heat removal, of ensuring that acceptable core heatup limits (as defined
below) are not exceeded, even assuming a single failure.
- Boration of the RCS in accidents of PCC 3 or PCC 4 shall be ensured at a rate such that
cold shutdown boron concentration can be reached, assuming a single failure and
assuming, in parallel, cooldown by the secondary side at a cooldown-rate not exceeding
the temperature limit that corresponds to an acceptable core shutdown margin.
- SIS capability shall be sufficient to ensure that, following a total loss of secondary side
heat removal, a time delay for operator actions of approximately one hour is adequate to
avoid exceeding acceptable core damage limits. It is assumed that, in conjunction with
SIS operation, discharge of the RCS via pressurizer valves will be initiated (feed and
bleed operation).
- The cooling capability of the SIS shall be sufficient to maintain in case of SBLOCA the
mean temperature in the IRWST < 110 C (for "feed and bleed" transients, this limit may
be exceeded).
- In case of loss of the RHR system, the SIS is used in a RHR mode as a functional backup.
- Provision shall be taken to ensure that, during all plant initial conditions where SIS operation might be required, the operational readiness of SIS equipment for automatic and/or
manual initiation of safety injection is maintained.
2.3. Description

The SIS consists of four separate and independent trains, each providing injection
capability by an accumulator (ACCU), a Medium Head Safety Injection (MHSI) pump, and a
Low Head Safety Injection (LHSI) pump (see Fig. 1). Both the MHSI and the LHSI pumps
take suction from the IRWST and inject into the RCS loops via nozzles located on the upper
part of the loops. The pumps are located in the Safeguard Buildings, close to the
containment. The LHSI pumps inject simultaneously into both the hot and the cold legs of
RCS. The MHSI pumps inject into the cold legs. The accumulators are located inside
containment and inject into the hot legs, using the same injection nozzles as the hot leg
injection of the LHSI. A heat exchanger is located downstream of each LHSI pump. These
heat exchangers are also installed in the Safeguard Buildings and are cooled by the
Component Cooling Water System (CCWS).
TABLE I: CHARACTERISTICS OF EPR'S SAFETY INJECTION SYSTEM

Medium Head Safety Injection

- Location
- No. of pumps
- Shutoff head
- Maximum injection rate

Accumulators
- Location
- No. of accumulators
- Injection pressure
- Water/ Nitrogen volumes
Low Head Safety Injection
- Location
- No. of pumps
- Shutoff head
- Maximum injection rate

cold leg
4
8.0 MPa
64kg/s

hot leg
4
4.5 MPa
32m 3 /15m 3
combined (cold and hot leg)
4
2.0 MPa
125kg/s

127

IXVIV

FIG. 1: EPR SAFETY INJECTION SYSTEM CONFIGURATION

2.4. Emergency core cooling mode

The emergency core cooling (ECC) mode includes the first four safety functions of the
SIS, as described in Section 2.1. The most important new features of EPR's SIS which have
been evaluated in particular with respect to SBLOCA scenarios are the shut-off head of
only 8 MPa for the Medium Head Safety Injection (MHSI), and the secondary side partial
cooldown to about 6 MPa, which is necessary in order to enable MHSI.
With implementation of a partial automatic secondary side cooldown the SIS configuration ensures appropriate mitigation of all types of primary side break (LOCA) in terms of location and size as well as all relevant NON-LOCA scenarios. The preliminary SIS configuration and characteristics are summarized in Table I. The design and safety criteria to be fulfilled by the SIS under LOCA-conditions are listed in Table II. The criteria have to be fulfilled
under consideration of the single failure criterion.
2.5.

Rationale for the selected emergency core cooling mode

2.5.7. MHSI

The reduced head of the MHSI has been selected in order to clearly improve the mitigation of SGTR. There is no need for variable set points of the secondary-side safety
valves (less risk of human error).
The cold leg injection of the MHSI complies best with the requirements resulting from
both SBLOCA (<200 cm2) and NON-LOCA mitigation. For SBLOCA the cold leg and hot leg
injections are practically equivalent from the point of view of ECC-efficiency. However, for
severe NON-LOCA scenarios (e.g. overcooling transients or those resulting in "feed and
bleed") the cold leg injection is advantageous in terms of core cooling and boration.

128

TABLE II: DESIGN AND SAFETY CRITERIA FOR THE SIS

Design criteria

complete core quenching time for 2A LBLOCA <3 min

no core uncovery for break sizes <j> < 150 mm (< 180 cm2)
no loop draining for break sizes <j> < 25 mm (<5 cm2)

Safety criteria

peak cladding temperature < 1200 C

local cladding oxidation < 17% of cladding thickness
core geometry integrity: no prevention of core cooling
Zirconium-water reaction < 1% of total cladding material

2.5.2. Accumulators
The hot leg injection has been selected because it provides major advantages in the
mitigation of large and intermediate break LOCA:
- faster core reflood and lower peak clad temperature (PCT) because of high steam condensation in the upper plenum and the hot legs,
- pressurized thermal shock concerns are excluded.
2.5.3. LHSI

The combined injection, i.e. the simultaneous injection into both the hot leg and the
cold leg of the main coolant line is favorable for mainly three reasons:
- during LBLOCA the hot leg contribution supports the accumulators by minimizing the
quenching time and consequently the steam release to the containment,
- the combined injection ensures an uniform boron concentration distribution, and
- makes any valve switching unnecessary (e.g. from cold leg to hot leg injection) and thus

increases the reliability of the procedure.

3.

SAFETY INJECTION SYSTEM EFFICIENCY UNDER LOCA CONDITIONS

A large number of SBLOCA and LBLOCA analyses have been carried out by
Framatome and Siemens within the aim of SIS sizing. The CATHARE, SPC-RELAP5 and
RELAP5/MOD2 computer codes have been used. Certain scenarios have been defined as
benchmark calculations and have been simulated with both codes in parallel; comparison of
the results allowed the adaptation of model options as well as of various modeling methods
in order to get similar predictions with both codes.
Results of preliminary LOCA analyses are discussed in this section. Conclusions with
respect to the sizing of the SIS-components (i.e. MHSI, LHSI and accumulators) are drawn

3.1 Computer codes used and EPR modeling

The thermal-hydraulic response of the EPR to LOCA-conditions was analyzed with the
advanced computer codes CATHARE [1, 2], RELAP5/MOD2 [3] and SPC-RELAP5. The
well known French code CATHARE was developed by CEA, Framatome and EdF; it is used
by Framatome for both SB and LBLOCA analysis. Siemens used RELAP5/MOD2 for
SBLOCA-and SPC-RELAP5 for LBLOCA-calculations. SPC-RELAP5 was developed by
Siemens Power Corporation (Richland, WA) for performing realistic analysis of accident
thermal-hydraulics in PWRs. It is a RELAP5-based system code incorporating features of
RELAP5/ MOD2 and RELAP5/MOD3 [4], and company improvements. In general, the
improvements and modifications included are those required to provide congruency with the

129

unmodified literature correlations and those to obtain adequate simulation of key LBLOCA
experiments. For LBLOCA-analysis SPC-RELAP5 was coupled with the containment code
COCO. Thus, both the RCS response and the containment response are simultaneously
calculated; this allows more accuracy in the accident simulation. The coupling method is
described in [5].
Depending on the simulated LOCA-scenario, different CATHARE and RELAP5
nodalizations of the EPR have been used. Two representative models are presented below.
3.1.1. EPR Modeling for CATHARE Calculations

The CATHARE nodalization is depicted in Fig. 2 for the RCS; it is composed of an

arrangement of 1-D components (pipes), 0-D components (volumes), tees, junctions
between components, and boundary conditions. The RCS model features three primary
loops: one broken loop, one intact loop, with the pressurizer, and one double loop lumping
the two remaining intact loops. It comprises 12 volumes, 5 in the reactor vessel (upper
region of the downcomer, lower plenum, upper plenum, guide tubes, upper head), 6 for the
SG inlet and outlet plena, 1 for the pressurizer. It also comprises 12 pipes for the core,
downcomer lower region, hot legs (3), SG tubes (3), cold legs (3) and pressurizer surge line,
amounting to more than 400 hydraulic meshes. All hydraulic components are thermally

linked to meshed walls or fuel rods.

FIG. 2: EPR 1400 MW. CATHARE 2 NODALIZATION OF REACTOR COOLANT SYSTEM

The secondary side model is depicted in Fig. 3 and features 3 SG and connected
pipes; it is also described by volumes and pipes, amounting to more than 200 hydraulic
meshes. The secondary side is an axial preheater type SG. The downcomer is split into two
regions; normal feedwater is delivered to the region which supplies the axial preheater zone
of the riser. Heat exchange between preheater and the lower part of the downward cold Utubes is modeled. The fluid of the axial preheater zone mixes above the partition plate with
the mixture of the opposite riser zone. The resulting mixture is then boiled above that plate

130

Normal feedwater

"Cold" downcomer

"Hot"

downcomer

FIG. 3: EPR 1400 MW. CATHARE 2 NODALIZATION OF SG SECONDARY SIDE

by the heat exchange with the upper part of the U-tubes, both upward and downward parts;
the CATHARE model allows to describe these processes.

3.1.2. EPR Modeling for SPC-RELAP5 Calculations

Very similar plant models have been used for RELAP5/MOD2 and SPC-RELAP5
calculations. The SPC-RELAP5 nodalization is depicted in Fig. 4. The EPR model consists
of 620 control volumes, 660 junctions and 640 heat structures (HS) with more than 4000 HS
mesh points. All RCS components - primary and secondary side - are modelled in detail;
this way the model can be used without major modifications for the simulation of various
scenarios (e.g. small, intermediate and large breaks, feed & bleed, etc.). Special attention
was paid to the RPV-model. The active core is represented by three channels: a breakthrough channel which represents 20% of the fuel assemblies (FAs), a hot channel (1 FA),
and a main channel - also called up-flow channel - including the rest of FAs. The breakthrough and main channels are connected by cross flow junctions, whereas the hot channel
is isolated. Every channel is axially sub-divided into 15 axial nodes. The upper plenum (UP)
is also divided into a 20% breakthrough region and a 80% upflow region. This nodalization
is based on UPTF test data and enables the best-estimate simulation of hot leg injection.

131

I I I I** I I I I h-i r-M

I I I* I I I I

FIG. 4: EPR 1400 MW. SPC-RELAP5 NODALIZATION OF REACTOR COOLANT SYSTEM

TABLE III: EPR STEADY-STATE AT 100% REACTOR POWER

STEADY STATE CHARACTERISTIC

Reactor power
Primary pressure (hot leg)
Secondary pressure (steam dome)
Coolant temperature at core outlet

Coolant temperature at core inlet

Temperature difference SG inlet/outlet
Mass flow rates
- Loop
- Core inlet total
- Core reflector (5%)
- Upper head bypass (2%)
Reactor coolant pump velocity
Reactor coolant pump head
Main steam flow rate
MFWS injection rate
MFWS water temperature
Pressurizer water volume
SG water volume secondary side
RCS water volume (PRZ included)

132

VALUE

4250 MW
15.5 MPa
7.25 MPa
328.6 C
291.3 C
35 K

5263 kg/s
21050kg/s
1052 kg/s
421 kg/s

155.5rad/s
1030kPa
601 kg/s
601 kg/s
230 C
42m 3
100m3
412m3

3.2. Initial and boundary conditions for LOCA-analysis

The initial thermal-hydraulic state of the reactor coolant system at 100% reactor power
is summarized in Table III. The main boundary conditions assumed are:
- reactor power 103% of the nominal power,
- 3% of the reactor power are generated directly in the moderator,
- axial reactor power distribution with chopped cosine shape,
- average Linear Heat Generation Rate (LHGR) 159 W/cm, maximum LHGR 450 W/cm,
- peaking factor 1.45 for the average rod and 2.83 for the hot rod,
- primary system pressure increased by 0.2 MPa,
- conservative decay heat, i.e. ANS 71 + 20% or DIN 25463 + 2a,
- containment back-pressure either set constant (SBLOCA) or calculated on-line (LBLOCA).
Concerning the availability of SIS, single failure of the diesel generator of one SIS train
is assumed, i.e. one MHSI and one LHSI are not available. Because of this SF, one train of
the Emergency Feed Water System (EFWS) is also not available, since it is connected to
the same diesel generator. Additionally, the cold leg injection of the SIS which belongs to
the broken loop is supposed to be lost directly into the containment. It is furthermore
assumed that the LHSI is distributed 50% to 50% between the cold and hot legs. The
resulting SIS effectiveness is summen'zed in Table IV.
TABLE !V: ASSUMPTIONS ABOUT SIS AVAILABILITY FOR LOCA ANALYSIS

3.3

MHSI PUMPS

ACCUMULATORS

LHSI PUMPS

HOT LEG

4 out of 4

1.5 out of 2

COLD LEG

2 out of 4

1.0 out of 2

Calculation's results

3.3.1. SBOCA; MHSI-sizing

RELAP5 and CATHARE 2 calculations were performed for different break sizes in
order to assess the required injection rates to be delivered by the MHSI pumps in
compliance with the design criteria specified in Tab. II:
a) "No loop draining for break sizes <j> < 25 mm". Strict accomodation of the Moody
liquid critical flow model results in a requirement of 40 kg/s at 7.5 MPa RCS pressure for the
25 mm break. However, a limited loop draining, consistent with maintaining the two-phase
natural circulation all around the entire RCS loop, with large margin to the RCS residual
mass corresponding to the transition towards reflux condensation, results in a requirement
of 22 kg/s only at 6.5 MPa.
b) "No core uncovery for break sizes <j> <150 mm". According to increasing break sizes,
i.e. for break equivalent diameters of 50, 75, 115 and 150 mm, the requirements to the
MHSI flow rate versus RCS pressure are roughly 20 kg/s at 6 MPa, 45 kg/s at 6 MPa, 80
kg/s at 4.5 MPa and 90 kg/s at 4.5 MPa, respectively. The results rely upon the following
assumptions:
- most conservative core decay heat, according to ANS 71+ 20% margin,
- secondary side partial cooldown from 8.5 MPa to 6 MPa at 100 K/h rate, actuated upon
pressurizer low pressure signal (11 MPa),
- main coolant pumps trip upon fluid saturation signal.
The required flow rates must be delivered by 2 out of 4 MHSI, since one MHSI is lost
because of assumed single failure of the corresponding diesel generator (see Section 3.2)
and a second one is connected to the broken loop and spills directly into containment.

133

The compatibility between these requiremetns and the preliminary design of EPR's
MHSI system was verified: the MHSI effective flow rate injected by 2 out of 4 MHSI pumps
comply with all a.m. requirements.
3.3.2. LBLOCA; accumulator sizing

A double-ended (2A) guillotine break in the cold leg piping system between the ECC
injection nozzle and the reactor vessel was simulated. SPC-RELAP5 calculation results are
presented below. Main events are listed in Table V.
The system behavior during the blowdown phase is independent on the SIS configuration until ACCU injection starts. The pressures in the primary system and the containment
equalize 36 s after break initiation at a pressure of 320 kPa. The ACCU injection starts at 15
seconds, when the primary system pressure has decreased below 4.5 MPa. A few seconds

later highly subcooled ECC-water from the hot legs is delivered to the UP and penetrates
through the tie plate to the core. The water breakthrough occurs only within a defined area
of approximately 20% - 25% of the total core flow area. While a significant portion of the
steam in the hot legs and the UP is condensed by the hot leg injection, ECC-water which
penetrates through the tie plate is still highly subcooled.
Water downflow from the UP initiates core cooling during end-of-blowdown (Fig. 5).
Some of the water downflow is vaporized and steam flows out of the top and bottom of the
core; however, most of the water downflow is heated to near saturation and flows to the LP.
Within the breakthrough region, the core is quenched from the top down by the water downflow from the UP at the end of blowdown. The LP inventory starts to increase about 10s
before the end of blowdown. This level increase is exclusively due to the hot leg injection
which penetrates through the tie plate and core; the cold leg injection (LHSI + MHSI) only
starts at 40s (Fig. 6).

1200
HOT ROD
UPFLOW CHANNEL
HOT ROD
HOT CHANNEL

1000
+

AVERAGE ROD
UPFLOW CHANNEL

800

o
UJ
cc
600
UJ

Q.
UJ

400

200

20

180

FIG. 5: 2A COLD LEG BREAK ANALYSIS. PEAK CLADDING TEMPERATURES

134

200

600
MHSI

20

40

60

80

100

120

LHSI COLD LEG

LHSI HOT LEG

LHSI TOTAL

ACCUMULATOR

140

160

180

200

FIG. 6: 2A COLD LEG BREAK ANALYSIS. SIS INJECTION RATES (ONE TRAIN)

By the completion of blowdown, the LP is filled to the bottom of core barrel. A few
seconds later (i.e. 45 s after break initiation), the vessel fills to the core inlet and refill is
complete. Hence, the end-of-blowdown and refill are overlapping which reduces the time to
core reflood and therefore the heat-up period of the non-downflow regions of the core;
consequently, the cladding temperature is relatively low: < 800 C.
The reflood phase begins at approximately 36 s after break initiation. Initially, the DC
water level increases rapidly as ECC-water injected into the cold legs is delivered to the DC
and ECC-water injected into the hot legs penetrates through the core to the LP and flows
into the DC. When, at 70 s, the DC water level reaches the cold leg elevation, water spills
out the broken cold leg and the water level stabilizes.
Table V: 2A COLD LEG BREAK ANALYSIS: SEQUENCE OF EVENTS
TIME

EVENT

Break opening, reactor trip, turbine trip and

reactor coolant pump trip

[seconds after
break opening]

Beginning of ACCU injection

0
15

Beginning of refill phase

End of blowdown phase
MHSI and LHSI start
End of refill phase
Quenching of average rod in up-flow channel
Quenching of hot rod in up-flow channel
Quenching of hot rod in hot channel
Accumulators empty

26
36
40
45
60
90
115
150

135

During refill/reflood the hot leg injection condenses efficiently most of the steam
produced in the core. The uncondensed steam flows through the loops. However, since
most of the steam is condensed in the UP and hot legs, the loop steam flows are minimal
and the corresponding pressure drop is small; thus the core flooding rate is high.
Hot rods located in the up-flow channel (which represents 80% of the FA) are
quenched 90 s after break initiation; the hot rod included in the hot channel is quenched
within 115 s. The peak cladding temperature (PCT) is reached within the blowdown phase;
the PCT is 810 C. The second peak reached at 35 s is only 800 C (metal-water reaction
and eventual fuel rupture are not taken into account in the present analysis).
The ACCUs empty 150 s after break initiation, i.e. the ACCU injection has a duration of
approximately 135 s. The ACCU injection rate is in maximum 430 kg/s or, averaged over
the time between injection start and total core quenching, approximately 330 kg/s. There is
a delay of approximately 30 s between core quenching and ACCU emptying. This injection
characteristic was achieved by optimizing the water/nitrogen volumes and by throttling the
ACCU injection line in an adequate way.
4.

CONCLUSIONS

The most important new features of EPR's SIS have been evaluated with respect to
SBLOCA scenarios: they are firstly, the shut-off head of only 80 bar for the MHSI and
secondly, the secondary side partial cooldown to about 60 bar. With implementation of a

partial automatic secondary side cooldown the SIS configuration ensures appropriate
mitigation of all types of primary side breaks (LOCA) in terms of location and size as well as
all relevant NON-LOCA scenarios.
Preliminary results of analyses performed with the best-estimate codes CATHARE,
RELAP5/MOD2 and SPC-RELAP5 demonstrate the efficiency of the SIS:
a) In case of LOCAs caused by very small breaks (<{> < 25 mm) the EPR's SIS is capable - in
conjunction with the automatic partial cooldown of the secondary side - to prevent loop
draining, even when only two MHSIs are effective. In case of small and intermediate
leaks (<{>< 150 mm) the SIS ensures sufficiently high injection rates for preventing core
uncovery, again even when only two MHSIs are effective. For leaks within the range 150
j> < 350 mm, the SIS is able to limit the core uncovery.
b) In case of double-ended break LOCAs the SIS ensures the fast reflood of the core within
approxi-mately 120 s; the peak cladding temperature is 810C. The reflood of the core is
achieved by the accumulator injection, which lasts about 135 s and ensures a relatively
large time delay of 30 s between core quenching and accumulator emptying.

REFERENCES

[1]
[2]
[3]
[4]
[5]

136

BARRE, F., BERNARD, M., "The CATHARE Code Strategy and Assessment",
Nuclear Eng. and Design 124 (1990) 257-284
BESTION, D., 'The Physical Closure Laws in the CATHARE Code",
Nuclear Eng. and Design 124 (1990) 229-245
RANSOM, V.H., et al., RELAP5/MOD2 Code Manual, Volumes 1 and 2,
NUREG/CR-4312, Rev. 1, (1987)
CARLSON, K.E., et al., RELAP5/MOD3 Code Manual, Volumes 1 and 2,
NUREG/CR-5535, EGG-2596 (Draft), (1990)
CURCA-TIVIG, F., KOHLER M., "A Code System for Coupled Analysis of Reactor
Coolant System and Containment Thermal-hydraulics", New Trends in Nuclear System
Thermal-hydraulics (Proc. Int. Conf., Pisa, 1994), Vol. 2, 293-301

SWR-1000: THE DIMENSIONING OF EMERGENCY

CONDENSERS AND PASSIVE PRESSURE PULSE TRANSMITTERS

C. PALAVECINO
Power Generation Group (KWU) of Siemens AG,
Offenbach am Mein, Germany
Abstract

The Power Generation Group (KWU) of Siemens AG and the German electrical power
utilities - particularly those operating boiling water reactor plants - are together developing a new reactor type which is characterized in particular by its passive safety
systems.

These passive safety systems are the emergency condensers, the containment cooling
condensers, the passive pressure pulse transmitters, the gravity-driven core flooding
lines, the rupture disks arranged in parallel to the safety-relief valves, and the scram
systems.

This presentation constitutes a report on the emergency condensers and the passive
pressure pulse transmitters.

The most important reasons for introducing passive safety systems are to increase the
safety of future nuclear power plants, to simplify reactor safety systems and to reduce capital costs.
The emergency condensers are heat exchangers consisting of a parallel arrangement
of horizontal U-tubes between two common heads. The top header is connected via
piping to the reactor vessel steam space, while the lower header is connected to the
reactor vessel below the reactor vessel water level. The heat exchangers are located
in a pool filled with cold water. The emergency condensers and the reactor vessel
thus form a system of communicating pipes. At normal reactor water level, the emergency condensers are flooded with cold, non-flowing water. No heat transfer takes
place in this condition. If there is a drop in the reactor water level, the heat exchang-

ing surfaces are gradually uncovered and the incoming steam condenses on the cold
surfaces. The cold condensate is returned to the reactor vessel.
The design of the emergency condensers must meet the requirements dictated by
the given thermal and hydraulic conditions.

The effects of the thermal condition parameters are relatively well known to us from
the evaluation of emergency condenser testing conducted at Gundremmingen Nu-

137

clear Power Station. As we have altered the elevation conditions in the radial direction in comparison to Gundremmingen Unit A, new sizing calculations have been
performed. An emergency condenser test rig was constructed at the Julich nuclear
research center in order to provide experimental verification of our calculations.

Taking into consideration a redundancy degree of N + 2, a specific thermal rating of

63 MW per emergency condenser results for a reactor with an output of 2778 MW.
The total performance of the emergency condenser system is thus 252 MW, or 9.1 %
of reactor output.
Given this emergency condenser rating, accident control of some transients becomes
very interesting:

The heat removal capacity in the lower pressure range corresponds to that of 2
to 3 relief valves.
In the event of a stuck-open relief valve with simultaneous failure of all reactor vessel injection possibilities, the core will not become uncovered until some
24 hours after the onset of accident conditions.
The following can be said of this component:
a)

It is more reliable than components designed for comparable functions.

The probability of failure of the emergency condenser of Siemens' BWR 1000 is

approximately 10-4 per demand, while that of older emergency condenser designs such as at Gundremmingen Unit A is approximately 2 to 3 x 10-3 per demand, and that of the active residual heat removal systems of Siemens advanced boiling water concept about 2 to 3 x 10-2 per demand.

b)

It is considerably less expensive than the residual heat removal systems implemented to date, which comprise a primary circuit, a component cooling system
and a final cooling system, each equipped with pumps, valves and heat exchangers, etc. These latter systems are provided with a diesel generator as a redundant power supply system. The cost of one train (without considering infrastructure elements such as the building, etc.) can be assumed to amount to
some DM 100 million. In contrast to this, the cost for an emergency condenser
system (comprising four emergency condensers) is estimated to cost between
somewhere between DM 10 and 20 million.

138

The reliability of the electrically-operated reactor protection system represents a

limitation to achieving a higher degree of safety. It was therefore necessary to duplicate the functions of the reactor protection system in a different way. These efforts successfully culminated in the development of the passive pressure pulse transmitter.

Passive pressure pulse transmitters function in the same manner as the emergency
condensers. The pressure generated in a heat-exchanger secondary circuit is used to
actuate pilot or main valves. These passive pressure pulse transmitters allow the
scope of reactor protection systems requiring electric power to be reduced considerably, while plant safety and reliability are increased through the combination of
electrically-operated reactor protection systems with passive safety equipment.

No
1

Quantity
Emergency condenser

2 Safely- relief valve

Spring- loaded pilot valve

Diaphragm pilot valve

Passive pressure pulse transmitter

Rupture disk

Flooding line

Containment cooling condenser

Core flooding pool

10

Pressure suppression pool

11

Honzontal discharge vent

42

12

Vertical discharge shaft

12

13

Scram system

Fig. 1. SWR 1000 Passive Safety Systems

139

28,7 m

Fig. 2. SWR 1000 - Containment

Orywell

22.258

17430

Water level at lull cooling

capability of me condenser
9.000

4394

Fig. 3.
SWR 1000 - Isolation condenser (schematic)
(Height in m above 0.0 elevation in RPV)
140

Containmentwall
Emergency or Isolation Condenser

Fig. 4.
KRB - A Emergency condenser diagram
70

60

cold water (30 *C)

50
Design

capability

40

boiling waler {116 C)

30

20
O Measuring point

10

10

20

30

40

50

60

70

Pressure [bar]

Fig. 5.
KRB-A
Cooling capability of the emergency condenser
Measurement on May 10,1975

141

100
1-

90

TO
Q.

80

=5

70

60
50
40
30

20

10

10

20

30

40

60

50

70

Pressure [ bar ]

Fig. 6.

SWR 1000 - Emergency condenser.

Cooling capability as a function of the pressure in

the RPV
Qh
)

^^^

QH = Cooling capability at lower RPV water leval

Qo = Maximum cooling capability

80

^X
70
/

60

/
50
40
30

20

/
10
n

AH[m]

Fig. 7.
SWR 1000 - Emergency condenser.
Cooling capability as a function of loss of water level
in the RPV. (AH in m)

142

400

-300

200

100

20

40

80

60

pressure in RPV [bar]

Fig. 8.
SWR 1000 - Comparison between the cooling capability
of the isolation condenser and those of the safety relief

valves
Steam
Line to
rEmergency

Condenser

Pressure Pulse Line

to Pilot Valve

Condensate
from Emergency
Condenser

Fig. 9.

SWR 1000 - Passive Pressure Pulse Transmitter

143

ou

60
1

1 40
i>
5
j>n

\\
N\
\s

t 20

***

-.

0
0

50

100

150

200

Fig. 10. SWR 1000 - Reaction Time of PPPT

Table 1: Principal Data of Emergency Condenser System

Number of emergency condensers

Design

1 tube bundle comprising a

four-pass U-tube configuration,
with connection to two common headers

Performance of each emergency condenser

63 MW given a primary system

pressure of 70 bar, a core flooding pool temperature of 40 C,
and a reactor vessel water level
drop of 8.20m

Heat transfer area per condenser

138m2, comprising 104 tubes;

tube diameter: 44.5;
wall thickness: 2.9 mm

Design conditions:
Primary side
Secondary side

88 bar, saturated water

0-10 bar

Temperature:
Primary side
Secondary side

300 C
40-180C

Size (diameter) of connected piping

- Supply steam line: 400 mm

- Condensate discharge line:
200mm

- Headers
o steam side:

500mm

o condensate side: 300 mm

144

Table 2: Comparison of Magnitude of Failure

Probabilities per demand of Various Residual Heat

Removal Systems

Failure Probability

Gundremmingen A Siemens-KWU
Series ABWR
Type Emergency
Active RHRS
Condenser

BWR1000
Emergency
Condenser

System

Signal acquisition
and processing

1E-3

1 E-4

Startup failure
(valves, pumps, etc.)

1E-3

1-3E-2

Failure during
accident (7 days)

1E-4

1 E-2

1 E-4

1E-4

1 E-3

1E-5

2E-5
(2E-2)

Failure of piping and

heat exchanger tubing
Failure of power supply

(from this, failure of

emergency power
supply)

Table 3: Principal Data of Passive Pressure Pulse Transmitter

Type

Dimensions
Diameter x Heigth

mm

Volume of Exchanger

Heat-Exchanger Tube

Diameter Thickness
mm
mm

Ratio

water

Area

Volume/Area

Length
mm

Number
~

m"

l/m'

555 x 650

25

1.5

3250

4,05

0,24

17

555 x 650

20

1,2

860

17.60

0.36

50

430 x 560

20

383

30

4., 33

0,69

Savings through technically

advanced plant design

100

-MWel (net)
200

400

600

800

1000 1200 1400

BWR - Degression of Specific Plant Costs as

Fuction of Output and Technical Design
145

DESIGN, FABRICATION AND TESTING OF FULL SCALE

PROTOTYPE FOR PASSIVE COOLING APPLICATIONS

P. GIRIBALDI, F. MAORIS, M. ORSINI, FL. RIZZO

ANSALDO, Geneva, Italy
Abstract
The paper presents the ANSALDO main activities performed during Design,
Fabrication and Site Assembly of the full-scale prototypes of two essential
components for the G.E. Simplified Boiling Water Reactor, i.e. the Passive
Containment Cooler (PCC) and the Isolation Condenser (1C).

The PCC and 1C are two safety related components

The reference Design provides the following advantages:
1)
2)
2.

New safety features i.e. the use of natural circulation

Decreased operator burden: i.e passive safety approach

INTRODUCTION

An essential feature of the SBWR, the advanced BWR developed by General

Electric is the presence of two safety related and passive systems, the 1C and
PCC which enhance the overall plant safety.
They perform their functions (control of reactor and containment pressure) for
72 hours without operator actions.
The key components of these systems are two natural circulation condensing
heat exchangers immersed in a pool water vented to atmosphere, i.e. the
Isolation Condeser (I.C.) and the Passive Containment Cooler (PCC).
The design analysis methodology and the main structural problems faced during
the design development have been already described in detail in doc. [1].
After a brief summary of the main design aspects focused in this document,
emphasis will be given to the manufacturing methodology of the prototypes.
The on-site activities in order to assembly the components coming from the
Manufacturer shop are described.
Finally being recently completed the test campaign on the PCC, some
preliminary evaluations about its structural integrity are given.

147

3.

COMPONENT DESCRIPTION

3.1 Isolation Condenser

Following is a summary of the main functional and design requirements of

the Isolation Condenser (1C).
*

Function:

Remove decay heat from RPV

following isolation events, to
control the RPV pressure and
water level
30 MW per unit
3
ASME Section III Class 2
8.62 MPa, 1250 psi
302 C, 575 F

*
*
*
*
*
*
*
*
*
*
*

Heat removal capacity

No of units
Code Classification
Design Pressure
Design Temperature
Corrosion Resistant Material
Low susceptibility to Intergranular Stress Corrosion (IGSC)
Reliability for 60 years design life
Easy In Service Inspection, maintenance, repair and removal
Fouling and plugging margin included in the design
Continuous operation at RPV pressure, 10 C

*
*

135 severe thermal transients, 10 to 302 C in a few seconds

Periodic venting of small volumes of non-condensable gases

The Isolation Condenser, see fig. 1, is made of two identical modules, each
consisting of a upper header, a vertical tube bundle and a lower header.

Each header is closed at both ends by flanged dished covers; the upper
one is fed by two lines connected to the main steam line from the R.V. and
the lower one is drained by a condenste line to the R.V.

The system is operated by simple automatic opening of a redundant valve

placed on the condensate return line to the R.V., the connection between
the R.V. steam area and the condenser upper headers being always
opened.
The Isolation Condenser is started into operation by draining the
condensate to the Reactor thus causing steam from the Reactor to fill the
tubes which transfer heat to the cooler pool water.

148

FIG. 1 ISOLATION CONDENSER

CONFIGURATION

Two modules per unit (horizontal steam inlet upper header, 120 vertical
condensing tubes 50.8 mm OD * 2.30 mm, horizontal condensate outlet lower
header).
Main steam line enclosed in a guard pipe and one-piece steam distributor with
built-in Venturi flow limiters to provide protection against potential breaks outside
containment.
Two feed lines/one drain line per module
Upper support allowing tubes thermal expansion free downward.
Horizontal dynamic restraints on lower header
Headers removable bolted covers allowing access to the tubes for all required
operations.
149

3.2

Passive Containment Cooler

Following is a summary of the main functional and design requirements of

PCC

Function:

Remove decay heat from RPV

following LOCA/SA, to control

*
*
*
*
*
*
*
*
*
*
*
*

the containment pressure

10MW per unit
3
ASME Section III Class 2
0.76 MPa, 110psi
171 C, 340 F

Heat removal capacity

No of units
Code Classification
Design Pressure
Design Temperature
Corrosion Resistant Material
Reliability for 60 years design life
Easy In Service Inspection, maintenance, repair and removal
Fouling included in the design
Continuous operation at Drywell pressure, 10 C
2 thermal transients, 10 to 150 C
Venting of large volumes of non-condensable gases

The PCC has basically the same geometry of the 1C: upper and lower
header connected by vertical tubes.
The PCC general arrangement is shown in fig. 2

The system is operated in a completely passive way, the connection

between the PC and the Condensers being always opened.
The PCCS loops are initially driven by the pressure differences created
between the containment drywell and the suppression pool during a LOCA
and then by gravity drainage of steam condensed in the tubes, so they
required no sensing, control logic or power - actuated devices to function.
Two modules per unit (horizontal steam inlet upper header, 248 vertical
condensing tubes 50.8 mm OD * 1.65 mm, horizontal condensate outlet lower
header).
One feed line per module
Single bottom nozzle, including concentric pipes for both condensate drainage
and non-condensible venting.
Lower header support, allowing tubes thermal expansion free upward.

150

Horizontal dynamic restraints on upper header

Headers removable bolted covers allowing access to the tubes for all required
operations.
Disc shaped deflector in upper header to prevent flow maldistribution inside
tubes.

FIG. 2

PASSIVE CONTIANMENT COOLING CONDENSER

CONFIGURATION

4.

DESIGN PHILOSOPHY

The sizing loads are those associated with the thermal transients on calling in
service of the system.
In fact both the equipments are subjected to a shock transient when the
stagnant cold water or gas mixture is replaced by hot condensing steam.
As a result the component structures will experience temperature distributions
evolving with time, leading to significant temperature differences between parts
of different thermal capacities, as well to important in wall thermal gradients, due
to high external heat transfer coefficient. Furthermore high pressure loads are
present for the 1C.

These very high stresses are not related to an accident condition, but to a level B
operating transient and to steady state conditions.

Facing the level B requirements expecially on stress ratcheting and fatigue was
the big design challenge.
151

The design approach was:

Use of a modular design which minimizes the required thickness and

hence the thermal gradients.

Keep the primary stresses low in order to avoid the risk of cyclic
progressive incremental deformation caused by very high thermal stresses.

All the peculiar aspects of the design were investiagated (see ref. (1)) from
steam distribution to nozzles, tubes risk of instability due to the imposed
displacements resulting from the header shell thermal behaviour during
transients. Results are always within the Code allowable limits.
5.

FABRICATION OF THE PROTOTYPES

Because of their geometry the two prototypes were shop fabricated as separate
pieces which were assembled directly at site
The following subassemblies related both to the PCC and the 1C have been
manufactured at the ANSALDO GIE - MILANO shop in Milan
a)
b)
c)
d)

Exchanging modules consisting of the two headers and the tube bundle.
Steam line complete with its diffuser
Feed lines
Drain lines

The most interesting aspects are related to:

a)
b)

c)

Extensive use of forgings

Module assembling with particular attention to the Tube-header junction
welding.
Instrumentation installation on the exchange tubes.

Hereafter these technical features are discussed

5.1

Extensive use of torgings

To satisfy the requirement of minimal in service inspection, the design
makes extensive use of forgings, in particular the headers and the covers
are made by single-piece forgins, without any welding.

152

For the headers the nozzles where obtained by extrusion and then
machining to final dimensions.
For the 1C cover, the curved shape was obtained by drop-forging, then the
component was machined to final size.

5.2 Module assembly

While the lines assembly has followed a quite classic proceedings (cutting,
calking, welding and non destructive examinations) the module one may
present a particular interest.
The assembling methodology has followed the following steps

a)

The drilling of the cylindrical wall in order to seat the exchange tubes,
This operation has been performed at the boring machine by a
special tool which created at the same time the hole and the external
circular groove.

b)

Vertical positioning of the two headers within a solid structure for

assembling in order to avoid deformations caused by the contraction
of the welding between tubes/headers

c)

Following a geometrical survey, the exchange tubes cutting

consistent with the measuring, in order to obtain the edge
overlapping required by the tube/header welding.

d)

Positioning of the exchange tube between the headers

e)

Welding execution and non destructive examinations.

The tube/header welding is an automatic orbital TIG type without weld

material obtained by melting the overlapped edges of tube and the welding
bath was produced by a pulsating and rotating current.
A typical weld result is shown in Fig. 3.

The welding machine is automatic and programmable for each of fts

functions.
The hardest effort has been spent to identify all the parameters and the
related tolerances (current intensity, voltage, overlapping of the joint edges,

153

Section B

Circular groove

Section A Welding blow-up

Section B Welding blow-up

Fig.3 Tube-header junction

rotation speed of the plasma torch, number of the qualified remeltings and
the related variations of parameters).
The specification of all these parameters has required long proof
compaignes completely repeated for both the prototypes, since
manufactured with different materials and different tubes thickness.

It must be remarked that the definition of the procedure for the 1C has been
more laborious than for the PCC, due to the material (INCONEL 600)
greater fluidity in the melted state.
It has been also set a repair proceeding by manual TIG with weld mateiral
to be used during manifacturing if some welding accident, had occurred.

154

5.3 Instrumenation installation on the exchange tubes

As a special application, the installation of the thermocouple for reading the
PCC tubes inner wall tempeature must be pointed out.
The procedure consists in creating a very precise groove on the tube
external by a circular milling cutter, leaving only 1/10 mm thickness at the
tube inner surface, then the thermocouple was put in the cavity and the
recess was filled by brazing.
Samples were subjected to pressure tests, up to 4 times greater than the
design one, and to corrosion tests in Na Cl solution, showing a very good
behaviour.
6.

ON-SITE ASSEMBLY

All the subassemblies manufactured at the ANSALDO GIE shops have been
delivered to Piacenza site where they have been assembled in the pools
assigned to the thermohydraulic tests.

The moduls final positioning has been completed and then they have been
connected to the steam and condensate lines
Following the covers were installed with the metallic O rings and the unit
hydrotest was performed.
After resolution of a very small leak from one PCC flange, the prototype was
ready for starting the test compaign.
7.

FINAL EXAMINATIN OF THE COMPONENTS

While the tests on the 1C prototype will start soon the test campaign has been
completed on the PCC.

The unit has undergone

*
*
*
*
*

5
10
1
2
40

Hydrostatic test at 1.5 x Design Pressure, i.e

Pneumatic test with air
Stream flushing (empty pool) at 143 C
Shakedown hot tests
Thermal hydraulic tests

10 + 4

LOCA pressure/temperature transients

11.4 bar, rel

7.6 bar, rel
3.0 bar, rel
4.9-6.9 bar, rei
3.8-6.9 bar, rel
4.0

155

The complete "structural" instrumentation results are not available yet (post
processing the large amount of date related to strain gauges, thermocouples,
accelerometers etc is very lows) and therefore the detailed comparison between

analytical and experimental results has not been made yet.

Nevertheless there is great evidence that the component has survived all the
tests - equivalent to more than 300 years of life - without any damage.

REFERENCE

1)

I. Micheli, M. Orsini
Design by analysis of SBWR key components
SMIRT August 1993.

156

RESULTS OF SAFETY-RELATED COMPONENTS/SYSTEMS TESTS

(Session TV)
Parti
Chairman
S.FRANKS
United States of America

INVESTIGATION ON PASSIVE DECAY HEAT

REMOVAL IN ADVANCED WATER COOLED REACTORS

F.J. ERBACHER
Forschungzentrum Karlsruhe, Institut fur Angewandte
Thermo- and Fluiddynamik
X. CHENG
Technische Universitat Karlsruhe,
Institut fur Stromungslehre und Stromungmaschinen

H.J. Neitzel
Forschungzentrum Karlsruhe, Institut fur Angewande
Thermo- und Fluiddynamik
Karlsruhe, Germany
Abstract

For future advanced water cooled reactor designs passive safety systems should
be investigated in respect of their feasibility and functional performance. The Research Center
Karlsruhe and the Technical University of Karlsruhe have proposed a composite containment
concept which copes with severe accidents. One essential of this concept is the decay heat
removal from the containment by natural air convection and thermal radiation in a passive way.
To determine the coolability of such a passive cooling system and to study the physical
phenomena involved, experimental separate-effects investigations are under way in the PASCO
test facility which is a one-sided heated rectangular channel. In addition numerical calculations
are performed by using the one-dimensional code PASCO and the three-dimensional thermalhydraulic code FLUTAN. Experimental results in respect to the distribution of the fluid
velocities, fluid temperatures and wall temperatures are presented. Based on the experimental
data the essential contribution of thermal radiation to the total heat transfer is illustrated. A
comparison is made between the experimental data and code predictions. The deficiencies and
needs for further research are being discussed. From the results achieved so far it has been
found that the passive containment cooling by natural air convection coupled with thermal
radiation is very promising. A medium-scale integral test facility MOCKA is presently under
design and will serve to study the integral performance of the passive decay heat removal
system proposed and to investigate the coolability limits by natural air convection coupled with
thermal radiation under all conceivable boundary conditions.
1.

INTRODUCTION

For future advanced water cooled reactors the use of passive safety features should be
encouraged and expanded whenever they can provide an adequate level of functional
performance. Passive safety systems rely on naturally available sources of motive power, such
as natural convection. The use of passive safety features in a nuclear power plant is a desirable
method of achieving simplification and increasing the reliability of the performance of essential
safety functions, e.g. reactor control and shutdown, core and containment cooling, and
retention of fission products. In this way passive systems can also promote an improved public
acceptance of nuclear energy [1], [2J.

159

Passive containment cooling by natural air convection has been proposed in the past for several
innovative reactor concepts e.g. MHTGR, PRISM, AP-600 [3], [4]. A review of the
experimental and analytical studies performed show the need for further experimental and
theoretical research work concerning natural air convection and thermal radiation. It became
evident that due to the complexity of the thermalhydraulic processes involved in passive
systems with their inherent small driving forces by temperature and pressure differences,
experiments have to be performed for each specific design. Validated computer codes must be
developed to support the expected operational performance. Also, proper attention should be
given to system reliability and to possible failure mechanisms which do not arise with active
components and systems.
2.

CONTAINMENT CONCEPT

The composite containment proposed by the Research Center Karlsruhe and the Technical
University Karlsruhe is to cope with beyond-design basis accidents. It pursues the goal to
restrict the consequences of severe core meltdown accidents to the reactor plant without any
noticeable release of radioactivity impairing the public [5].
filter
air outlet

air outlet

concrete 13000m3
steel
500m3
water
1 600 m3

inner containment
shell
outer concrete
containment shell

.A section A-A
T

air inlet

air inlet

core catcher

Fig. 1 New containment concept with passive decay heat removal

Figure 1 illustrates schematically the composite containment proposed. It consists of an inner

steel shell of about 60 m diameter and 38 mm wall thickness and an outer reinforced concrete
shell of approx. 2 m wall thickness. The annulus of approx. 80 cm radial gap width is bridged
by longitudinal support ribs fixed in the concrete shell. The ribs are placed on the
circumference with approx. 50 cm spacing and transfer the load of the expanding and
deflecting steel containment to the reinforced concrete wall (composite containment) in a
potential hydrogen detonation. With this concept the two individual containment shells of the
present design remain essentially unchanged and the capability of withstanding higher loads is
achieved by the composite structure which can cope with a maximum static pressure of 1.5
160

MPa. Such a containment concept has been investigated with respect to its feasibility [6]. The
higher costs for this improvement of the containment compared to the present-day design is
estimated at approx. 5 % of the overall plant investment costs.

A core catcher is an integral part of this new design proposal. In a core meltdown accident the
decay heat is converted into steam by direct contact of the melt with the water. The steam
produced condenses on the inner surface of the externally cooled containment shell. By reflux
of the condensate to the core catcher a passive self-circulating steam/water flow is established.
One essential of this new containment concept is its potential to remove the decay heat by
natural air convection in a passive way. The increase in the temperature of the steel shell
results in natural convection of air in the individual chimneys formed by the support ribs in the
annular gap. Moreover, radiative heat transfer takes place between the steel shell, the support
ribs and the concrete shell. Decay heat is thus removed by natural air convection coupled \vith
thermal radiation to the ambient atmosphere in a passive way.
On the basis of containment calculations with the TPCONT code, which was especially
developed for the given purpose, it can be assumed that due to the high heat storage capacity
of the internal structures of the containment (approx. 13000 mj concrete, 500 m5 steel) the
maximum heat flow through the containment wall to the air is reached in a 1300 MWe reactor
after about 10 days and amounts to approx. 8 MW [7]. Figure 2 illustrates the heat flow rates
in the containment.
15.0

decay heat
cooled hull

A uncooled hull
a vapour
structures

-2.5
15

r
20

25

30

35

40

time / d

Fig.2 Heat flow rates in the containment

A preliminary analysis of the heat removal capability of the composite containment proposed
has been performed using the one-dimensional PASCO computer code [8]. All calculations
performed have shown that the contribution of radiant heat transfer to the total heat transfer in
the chimneys is considerable. With high emissivities of 0.9 of the walls the heat transferred by
radiation equals approximately the amount transferred by convection [9]. Figure 3 illustrates
161

the removable heat as a function of the steel containment temperature for different emissivities
of 0.1 and 0.9, respectively. The calculations have been performed for an inlet temperature of
30C and no filtering of the cooling air. The following dimensions have been assumed:
diameter of the steel containment 60 m, height of the steel containment 40 m, depth of the
annulus 80 cm, thickness of the steel containment 38 mm, thickness of the support ribs 10 cm,
circumferential distance of the support ribs 50 cm. It becomes evident from figure 3 that with
an emissivity of 0.9 which corresponds approximately to the technical surfaces involved a
decay heat of 8 MW can be removed at a steel containment temperature of approx. 150 C .
With the additional pressure increase due to the non-condensable gases released in a core
meltdown accident this corresponds to a maximum containment pressure of about 0 8 MPa.
The figure underlines also the significant contribution of the radiant heat transfer which in the
past was often neglected in evaluating passive air cooled systems

0)

a
0)

o
0)

100

150

200

250

steel containment temperature [C]

Fig 3 Heat removal with different emissivities e of the wall

All preliminary evaluations performed show that the composite containment concept with its
passive containment cooling by natural air convection is a very promising alternative for future
advanced water cooled reactors. However, additional experimental and theoretical research
work needs to be done for the given conditions: large channel geometry and strong interaction
between convective and radiative heat transfer. To investigate the basic phenomena involved in
natural convection coupled with thermal radiation the separate-effects test program PASCO is
under way. The MOCKA test facility serves to finally prove the operational performance of the
passive cooling system under all conceivable conditions. This integral test facility is presently in
the process of a design study.
3.

PASCO TEST PROGRAM

The test program PASCO (acronym for passive containment cooling) is presently under way at
the Institute of Applied Thermo- and Fluiddynamics (IATF) of the Research Center Karlsruhe
162

(previously Nuclear Research Center Karlsruhe, KfK) The general aim of this separate-effects
test program is to investigate passive containment cooling by natural air convection and
thermal radiation

crosswise traversing probes

(velocity, temperature)

traversing probe
(velocity, temperature)

heated height: 8 m
channel cross-section
0 5 x 1 Om (variable)

heated plate

nn i

Fig 4 PASCO test facility

The PASCO test facility shown in fig 4 simulates one cooling channel in the annular gap of the
composite containment proposed The test section consists of a vertical rectangular channel of
which one wall is electrically heated The other three walls are thermally insulated from the
ambient surrounding The maximum channel cross-section is 500 mm x 1000 mm, the heated
height is 8000 mm with four individually heatable zones By changing the channel depth and
the heated height the effect of the channel geometry on heat transfer will be studied The wall
emissivity is varied over a wide range, so that the influence of thermal radiation on the total
heat transfer can be investigated To study the methods of enhancing the heat transfer, different
internal structures (e g inclined repeated fins) will be used Table I summarizes the test matrix
163

Table I: Test matrix

Heated wall temperature Th, C
Wall emissivity &
Heated height H, m
Channel depth L, m

100-175
02-0.9
20-8.0
025-1.0

The test faciliu is equipped among others with approx 170 thermocouples to measure the
distribution of wall temperatures Traversing probes for recording the temperature and velocity
of the air are installed at five different elevations Cross-wise traversing probes at the inlet, in
the mid-plane and at the outlet measure the temperature- and velocity distributions over each
individual channel cross-section Moreover, the pressure at the channel inlet, the air humidity
and the heating power are recorded With this comprehensive instrumentation data are
generated for the validation and development of multi-dimensional computer codes.
EXPERIMENTAL RESULTS

4.

Up to now steady-state experiments have been performed under the following boundary
conditions: channel width 500 mm. channel depth 250 - 1000 mm, heated height 8000 mm,
wall emissivity 0 9, heated wall temperature 100 - 175 C
Figure 5 illustrates the wall temperatures measured at the heated plate and the channel walls
for the channel cross-section 500 mm x 1000 mm. The boundary conditions of the test (M019)
are given in Tab.II. It is evident from the figure that with a nominal temperature of the heated
plate of 150 C the uniformity of temperature can be considered as sufficiently good The
temperatures measured at the back wall and the side walls that are heated by radiation
underline the great influence of radiative heat transfer
side nil Uorlfc)
1012

tick Till
511

S7? St.! S.2 0.9 VS 1 <S 19 C S19

KI.5

61 8 S6.I 51 1 52 1 SJ 8 SI 1 S53

583

side fill (sooth)

loii

H2 .S 119.2

SO.'S

SS)

lltllttf flit!

502

119 6 1!-! ISU

?1 2 61 S SO 3 0 51 1 SO 1 13.8 S19

S7t Hi

165 I'J! 115 3

n t a? s? 5 ss i s 2 52.1 s 2 56 6

SS!

1199

K!

P3I 613 58.1 SS6 S3 5 2S S! S71

EOS S7S

1508

HO! IB.!

13 ! 1C)

S61

SO

1B.2

?ii ss ssi Si sos BS stu SE 0

SS 2

SS.;

1193

n 2 ai a> ss o ss.8 s i s s S71

IOC

S7S

151 2 IB 9 151 0

?15 59 Ql S9 SIS 2.S SS

(OS

S71

S/S

52B

1521

19 18 J

1515 ISIS

n.5 H.1 58.J MS SJJ B.S B.O SIS

SS.S S3 9

1531

ru a.i a; a s si 8 ! v s SI

Ml

52 Z

IV 8 ll

V2

IS.!

3.9 S.! Hf

StXJ S 0.1 11 5 IS 9

) 61 US 36.7 315 31 331 39 S 12.!

Ml

137

1515

1112

1530 1113

1171 IB 9 IS3.3

Fig 5 PASCO test results. Wall temperatures (channel cross-section 500mm x 1000mm)
164

Figure 6 shows the profiles of the air temperature and of the air velocity measured at the test
channel outlet for the channel cross-section 500 mm x 1000 mm. The nominal temperature of
the heated plate was 150 C, the emissivity of the channel walls 0.9. It can be clearly seen in
which way the profiles are affected by the side and back wall temperatures induced by thermal
radiation.

The experimental data generated by the present PASCO tests represent an important data base
for the validation and further development of multi-dimensional computer codes, mainly in
respect to a proper modeling of the radiative heat transfer and the mixed convection turbulent
flow. Such 3D codes (e.g. FLUTAN, TRIO-EF), if verified, will finally be used to evaluate and
design the containment cooling system.

channel cross-section: 500 mm x 1000 mm

Fig.6: PASCO test results: Temperature and velocity profiles at the channel outlet
5.

CODE- / TEST EVALUATION

Calculations are being performed by using the one-dimensional code PASCO and the threedimensional computer code FLUTAN. The PASCO code has been developed to predict the
global thermal-hydraulics of the containment cooling and to assess the experimental data of the
PASCO test facility. The more time-consuming code FLUTAN is primarily used for detailed
analyses.

5.7

Test evaluation by the PASCO code

The PASCO code [8, 9] is based on the heat balances at the individual walls and the enthalpyand momentum balances for the air. The heat transfer from the walls to the air takes place by
natural convection. Between the individual walls of the chimney radiative heat transfer occurs.
165

For the convective heat transfer at each wall the following Nusselt-correlation has been used:

Here Nu stands for Nusselt number and Ra for Rayleigh number. The coefficient C of the
Nusselt-correlation needs to be determined for all individual walls of the chimney by the
experiments for the given conditions, i.e. non-symmetrical heated confined channel geometry.
Up to now 16 experiments have been evaluated in order to determine the coefficients C of the
Nusselt-correlation for the individual walls. In the experiments four different channel depths
were employed, i.e. 250 mm, 500 mm, 750 mm and 1000 mm. For each test four various
temperatures of the heated plate were adjusted, i.e. 100 C, 125 C, 150 C and 175 C. The
coefficients evaluated from all these tests resulted in the following C-values: 0.112 for the
heated plate, 0.21 5 for the back wall and 0. 160 for the side walls.
Table II summarizes the measured and calculated values for all temperatures investigated and
for the channel cross-section 1000 mm x 500 mm. In the experiments flow diverters were
installed at the inlet, the emissivity of the walls was 0.9. It is evident that the agreement is
rather good for all experiments even when adopting the same coefficients C in the Nusseltcorrelation.
Table II: PASCO, comparison between experiment and calculation

Experiment (MOU)
Calculation
Experiment (wois)

Calculation
Experiment (Moi9)
Calculation
Experiment (M022)
Calculation

TPI
[C]
100.0

Phi
[%]
58.1

TE
[C]
23.31

TU
[C]
25.37

124.9

42.9

23.56

25.34

149.8

42.8

22.36

24.79

175.3

45.7

23.01

25.04

TR
[C]
41.6
42.8
47.8
49.5
54.0
56.0
62.9
64.8

TS
[C]
40.3
43.5
45.9
50.6
51.1
57.5
59.5
66.6

QPI
[kW]
3.33
3.23
4.79
4.77
6.58
6.60
8.23
8.63

ML
[kg/s]

0.34
0.34
0.41
0.42
0.44
0.46
0.54
0.53

Given conditions of the experiments:

TPI = temperature of the heated plate
Phi = relative humidity of the air
TE = inlet temperature of the air
TU = environment temperature of the air (averaged over the height)
Results:
TR = temperature of the back wall (averaged)
TS = temperature of the side wall (averaged)
QPI= heat removal from the heated plate
ML = mass flow rate of the air

5.2

Test evaluation by the FL UTAN code

The advanced 3D thermal-hydraulic code FLUTAN available at the Research Center Karlsruhe
is a finite-difference code for single-phase steady-state and transient analyses of single- and
multi-component systems in Cartesian or in cylindric coordinates [10, 11].

166

In the FLUTAN code of the present version four turbulence models are available, i.e. constant
turbulent viscosity model, mixing length model, one equation (k-equation) model and twoequations (k-s) model. All the models are based on the turbulent diffusivity concept and take a
constant value for the turbulent Prandtl-number. For models with transport equations (k-, and
k-s-model) logarithmic wall functions for velocity- and temperature-distribution near the wall
are used.

To extend the application of the FLUTAN code to the PASCO test channel, a thermal
radiation model has been developed with the following simplications [12]:
Air is radiatfvefy nonparricipaling.
Wall surfaces are gray and diffuse.
Cartesian coordinate system.

For flow channels in a Cartesian coordinate system where boundary walls are either parallel or
perpendicular to each other, view factors have been derived analytically [12]. In addition the
so-called macro-element method is introduced, to reduce storage need and computins time
[12].
Figure 7 shows the distribution of air temperature and air velocity along the mid-line at the
channel outlet as function of the distance from the heated plate y for the channel cross-section
500 mm x 1000 mm. The temperature of the heated wall is 150 C and the wall emissivity is
0.9. The curves are the results calculated with the FLUTAN code and the symbols are the data
obtained in the PASCO experiments. The temperature distribution is well reproduced by the
FLUTAN code, whereas the FLUTAN code overpredicts the air velocity in the near wall
region and underpredicts it in the central region. This discrepancy emphasizes the need of
improving turbulence-modeling in the FLUTAN code.
o

LO

A U measured

OJ

<0

U calculated
x T measured
-- T calculated

O
CNJ I

Cb

CL
s

\
O
CO

.2

.5

.8

1 .0

Distance Y. m

Fig. 7:

Measured and calculated distribution of air temperature and air velocity

at the test channel outlet

Figure 8 shows the temperature distribution on the side wall and on the back wall at the middle
elevation for the following conditions: channel cross-section 500 mm x 1000 mm, temperature
of the heated wall 150 C and wall emissivity 0.9. The curves represent the calculated results
and the symbols are the experimental data. A good agreement is found between the

167

experimental and the calculated results. On the side wall the maximum temperature appears
close to the heated wall. It decreases rapidly with increasing distance from the heated wall.
Towards the corner where the side wall connects the back wall the temperature increases again
because of higher temperature of air in this region.

Ii

, o

FLUTfiN

Measurement

(HO 3 9 ) \

S
Sj

20

.40

.60

.30

] .00

1 .20

Perimeter

Fig. 8:

(ml

Temperature distribution on the side wall and on the back wall

curves: FLUTAN calculation, symbols: PASCO measurements

1-C

Fig. 9:

1 .40

1-R

2-R/C

3-R/C

5-R

Heat power transferred at different walls and by different heat transfer modes
1 - heated plate, 2 - side wall, 3 - back wall, 5 - inlet, 6 - outlet
C - convection, R - radiation

Figure 9 shows the transferred heat power at different surfaces and by different heat transfer
modes under the following conditions: channel cross-section 500 mm x 1000 mm, temperature
of the heated wall 150 C and wall emissivity 0.9. From the heated wall a heat power of about
3 kW is transferred directly to air by natural convection. About 4 kW heat power is transferred
from the heated wall via thermal radiation. The results show a considerable enhancement of
total heat transfer due to thermal radiation.
Additional numerical calculations with the FLUTAN code show that a strong influence of
thermal radiation on total heat transfer is obtained for the entire range of wall temperature.
Even at low temperature of the heated wall, e.g. 100 C, the heat power transferred by thermal
radiation equals to that transferred by natural convection.
168

6.

FUTURE WORK

Future test series in the PASCO separate-effects program deal mainly with studies using a test
channel with low emissivity surfaces and also with different channel insertions to investigate
means to enhance the heat transfer Also tests with different heated heights and additional
hydraulic resistances are being performed

To finally prove the integral operational performance of the passive decay heat removal s\stem
proposed and to investigate its coolability limits under all conceivable boundary conditions a
medium-scale integral test facility MOCKA is presently under design Figure 10 illustrates
schematically the test facility The main objectives of MOCKA (acronym for Model
Containment Karlsruhe) are
To study local condensation beha\iour in different types and different parts of containment
To examine the influence of non-condensable gases on condensation behaviour
To investigate the consequences of non-symmetric / local release of steam and gas
To examine the effects of containment wall structure and internal structures
To study the effects of water sprays
To study the effect of cooling channel structure (e g geometry, filter, fins)
To investigate coolability limits by natural air convection and thermal radiation
To demonstrate the operational performance under various conditions
To produce a data base for validation and development of advanced multi-dimensional
computer codes for advanced containment designs

A ir (DUtlel *
A A ^^

TITITTTTT

^^ A A

/K A A A A ^^

c3

Containment wall
structure alternatives
(section A- A)

p;it,' ->

)
A

Internal i
$tractu*e$
Air inlet

U U
U1U LUJJ UUJ

-A
Inner
Containment shell

"

^
Outer
/containment shell

Feed pipe for/

non-condensables

Fig 10 MOCKA test facility

The MOCKA test facility will be designed as a flexible equipment to investigate also
containment issues of present-day reactors and the European Pressurized Water Reactor EPR
169

7.

CONCLUSIONS

The present experimental and analytical research work within the PASCO program have
shown that passive containment cooling in natural air convection coupled with thermal
radiation is a promising concept. The tests performed have generated a broad and detailed data
base for the validation and improvement of different computer codes.

The following specific main conclusions can be drawn:

With high emissivities of the walls the decay heat of a 1300 MWe pressurized water reactor
can be removed by natural air convection in a passive way with maximum steel containment
temperatures not exceeding 150 C.
At intermediate and high wall emissivities thermal radiation contributes significantly to the
total heat transfer by natural air convection, even at relatively low temperatures of the
heated wall.
A radiation model has been developed and implemented in the FLUTAN-code.
The numerical results agree well with the experimental data concerning the distribution of
the wall temperatures.
The prediction of the velocity- and temperature profiles of the air within the cooling channel
emphasizes the need of improving the turbulence modeling in natural air convection.
Additional larger scale integral experiments in the MOCKA test facility must prove the
containment coolability under all conceivable conditions.
ACKNOWLEDGMENTS

This work is partially supported by the "Bundesministerium fur Bildung, Wissenschaft,

Forschung und Technologic" under the Project Grant 15NU 0961.
The authors greatly acknowledge the contributions of W. Just, H. Schmidt and E. Arbogast in
preparing and performing the experiments.
REFERENCES

[1]

INTERNATIONAL ATOMIC ENERGY AGENCY, The Safety of Nuclear Power:

Strategy for the Future, Proceedings of a IAEA-Conference, Vienna, September 2 - 6 ,
1991

[2]

INTERNATIONAL ATOMIC ENERGY AGENCY, The Safety of Nuclear Power,

INSAG-5; A report by the International Nuclear Safety Advisory Group, IAEA, Vienna,
1992

[3]

PEDERSEN, D.R., et al., "Experimental and Analytical Studies of Passive Shutdown

Heat Removal from Advanced LMRs", International Topical Meeting on Safety of Next
Generation Power Reactors, Seattle, USA, 1992

[4]

KENNEDY, M.D., et al., "Advanced PWR Passive Containment Cooling System

Testing", International Topical Meeting on Advanced Reactors Safety, Pittsburgh,
Pennsylvania, USA, April 17-21, 1994

[5]

HENNffiS, H.H., KESSLER, G., EIBL, J., "Improved Containment Concept for Future
Pressurized Water Reactors", International Workshop on Safety of Nuclear Installations
of the Next Generation and Beyond, Chicago, EL, USA, August 28-31, 1989

170

[6]

EIBL, I, "Zur bautechnischen Machbarkeit eines alternative!) Containments fur Druckwasserreaktoren -Stufe 3-", KfK 5366, August 1994

[7]

SCHOLTYSSEK, W., ALSMEYER, H., ERBACHER, F.J., "Decay Heat Removal after
a PWR Core Meltdown Accident", International Conference on Design and Safety of
Advanced Nuclear Power Plants (ANP 92), Tokyo, Japan, October 25-29, 1992

[8]

NEITZEL, H.J., "Abschtzung der Warmeabfuhr durch Naturkonvektion bei einem

alternativen Containmentkonzept", KfK 5005, Juni 1992

[9]

ERBACHER, F.J., NEITZEL, H.J., "Passive Containment Cooling by Natural Air

Convection for Next Generation Light Water Reactors", NURETH-5, Salt Lake City,
Utah, USA, September 21-24, 1992

[10] SHAH, V.L., et al., "COMMIX-1B: A Three-Dimensional Transient Single-Phase

Computer Program for Thermal Hydraulic Analysis of Single and Multicomponent
Systems", NUREG/CR-4348 Vol.1 and Vol.2, 1985

[11] BORGWALDT, H.A., "CRESOR, A Robust Vectorized Poisson Solver Implemented in

the COMMLX-2(V)", Proc. of the Int. Conference on Supercomputing in Nuclear
Applications, Mito City, Japan, pp.346-351, 1990
[12] CHENG, X., ERBACHER, F.J., NEITZEL, H.J., "Thermal Radiation in a Passive
Containment Cooling System by Natural Air Convection", International Symposium on
Radiative Heat Transfer, Kusadasi, Turkey, August 14-18, 1995

171

SPES-2, AP600 INTEGRAL SYSTEMS TEST RESULTS

L.E. CONWAY, R. HUNDAL

Westinghouse Electric Corporation,
Pittsburgh, Pennsylvania, USA
Abstract
The SPES-2 test program was performed to obtain data to validate safety analysis computer codes used to analyze
the performance of the Westinghouse AP600 reactor and passive safety system designs. The test matrix included
nine simulated loss-of-coolant-accidents (LOCAs) and four non-LOCA transients. The nine LOCAs were performed

at three different break locations; and at each location, two different break sizes were simulated. The simulated
breaks ranged from a 1-in. diameter break, up to the double-ended break of an 8-in. pipe. Non-LOC A tests included
three steam generator tube ruptures (SGTRs) and a large steam line break (SLB). In all tests, the passive safety
systems performed as expected and mitigated the simulated accidents with no heatup of the reactor heater rods. Test
results, test-to-test comparisons, and passive safety system performance are discussed.
INTRODUCTION

Two core makeup tanks (CMTs) that can provide

SPES-2 is a full-height, full-pressure integral systems

gravity at any pressure.

test of the Westinghouse AP600 reactor design. The

SPES-2 test was performed as pan of the Advanced

Two accumulators that provide borated water to

borated makeup water to the primary system by

Light Water Reactor (ALWR) program sponsored by

the U.S. Department of Energy (DOE) and the
Electric Power Research Institute (EPRI).
Westingbouse, in cooperation with SIET (Societa
Informazioni Esperienze Tennoidraulicbe), ENEL
(Ente Nazkmale per 1'Energia Eleorica), ENEA (Ente

the reactor vessel when/if primary pressure

<700psia.

per le Nuove Technologic, 1'Energia e rambientc),

and Sopren-Ansaldo, performed the SPES-2 tests to

refueling water stor-age tank (IRWST), that can

remove heat from me primary system by natural

obtain data on the integrated behavior and

circulation at any pressure.

performance of the AP600 passive safety systems to

validate the computer codes used to perform the

The automatic depressurization system (ADS),

licensing safety analyses for the AP600.

which is comprised of a set of valves connected

The AP600 is a 600 MWe reactor designed to

to the pressurizer steam space and the two hot

legs. These valves are opened sequentially to

increase plant safety.

A passive residual heat removal (PRHR) heat

exchanger (HX), comprised of a C-shaped tube

bundle submerged inside the in-containment

It has accident mitigation

provide a controlled depressurization of the

features that, once actuated, depend only on natural

primary system, if the CMT water level

forces such as gravity and natural circulation to

perform all required safety functions. These "passive"

significantly decreases.

safety features can also significantly simplify plant

An IRWST that provides a large source of core

cooling water, which drains by gravity after the
ADS has actuated.

systems, equipment, and operation.

The AP600 primary system is a two-loop design.

Each loop has one hot leg, two cold legs, and one
steam generator with two canned reactor coolant
pumps (RCPs) attached directly to the steam generator
outlet channel bead. The passive safety systems,
shown in Figure 1, are comprised of the following:

A passive containment cooling system (PCS) that

utilizes the AP600 steel containment shell to

transfer beat to the environment (ultimate heat

sink) by natural circulation of air and water
evaporation. The PCS was not included in the
SPES-2 experiments.

173

The SPES-2 test facility was designed to model the

AP600 at full-scale elevation and full pressure,
simulating toe full AP600 plant range of power with
a volume scaling factor of 1/395. Portions of the
reactor vessel, all main coolant loop piping and
passive safety systems have been expressly designed
and constructed for SPES-2 to model the AP600
plant.

demonstrated die repeatability of the SPES-i facility

operation and passive safety system performance
(S01703).
Test 4 (S00504). The 2-in. LOCA simulation with
operation of the active, nonsafety systems was
performed to observe the interaction of the passive
safety and nonsafety systems compared with the base
case 2-in. LOCA (Test 3).

SPES-2 Test Matrix

The SPES-2 test matrix examines the

performance/capability of the AP600 passive safety
systems in mitigating the effects of postulated DBEs,
and provide useful data for computer code
development and validation. The initial and boundary
test conditions challenge the passive safety systems,
provide direct comparisons between selected tests,
and/or match the limiting assumptions used in safety
analysis computer codes including the worst single
failure. The resulting test matrix is discussed below.
Loss of Coolant Accidents

The passive safety systems are required to provide

sufficient water for LOCA mitigation over a long
period of time; thus CMT draindown, ADS actuation,
accumulator delivery, and primary system
depressurization to IRWST delivery all occur.
Therefore, eight different LOCA simulations varying
in size and location were tested to observe the
integrated operation of the passive system over a wide
range of conditions. All LOCA tests, widi one
exception, were performed without operation of the
active, nonsafety pumped injection/heat removal. All
tests were initiated from full-power operating
conditions and used the minimum Standard Safety
Analysis Report (SSAR) pressure setpoints for reactor
trip and safety system actuations.

Tests 5 & 7 (S0060S & SOI007). A 2-to. break

simulation in the DVI line and in the cold-leg to the
CMT balance line were performed to observe the

effect of break location on the passive safer, system

mitigation capability as compared to the base case
2-in. LOCA (Test 3).
Test 6 (S00706). A DEG break simulation of one of

two DVI lines was performed to minimize the amount

of safety injection flow delivered to the reactor vessel
(only one of two CMTs, accumulators, and IRWST
injection lines deliver). This was the largest break
that could be reasonably simulated in the SPES-2
facility.

Test 8 (S00908). A DEG break simulation of one

8-in. cold-leg to CMT balance line was performed to
observe the effect on the faulted CMT and provide a
comparison LOCA with the DEG DVI LOCA.
Test 13 (S01613). The 1-in. diameter break
simulation (Test 1) was repeated with the number of

PRHR HX tubes increased from one to three tubes to

maximize the primary system cooling and better
simulate two PRHR HXs in operation in the AP600
plant
Steam Generator Tube Ruptures

The mitigation of an SGTR consists of reducing the

Test 1 (S00401). A 1-in. diameter break simulation

was selected as the smallest LOCA. This was based
on analyses that showed that complete beatup of the
CMT water occurred prior to ADS actuation, so that
any effect of CMT water flashing during
depressurization could be observed.
Test 3 (S00303). A 2-in. diameter break simulation
in the bottom of a cold-leg loop pipe (containing a
CMT balance line) was performed as the base case
LOCA to which other LOCAs would be compared.

This 2-in. diameter, base case break simulation was

repeated at the end of the test program and

174

pressure of the primary system to be equal to or less

than the faulted steam generator pressure to terminate
primary to secondary flow and to prevent overfill of
the faulted generator. At the same time, beat removal
from the primary system must be provided to remove
core decay heat and keep the primary pressure less
than secondary pressure. In current PWR designs,
recovery from an SGTR requires operator actions to
identify and isolate the faulted generator, establish
primary system heat removal using die intact steam
generator, and manually depressurize the primary
system. Three SGTRs were performed at SPES-2.
All these tests modeled a full rupture of a single
steam generator tube. All were initiated from full-

power operating conditions and used low-low

pressurizer level to initiate reactor trip and safety
system actuations.

Test 9 (S01309). An SGTR with operator action and

nonsafety systems operating was performed to observe

the combined effect of manual SGTR recovery actions

(used in current plants) and passive system operation.
Test 10 (S01110). An SGTR without operator
actions and without operation of active, nonsafety,
pumped injection/beat removal was simulated to
observe the capability of the passive systems to
terminate the event without intervention.
Test 11 (S01211). An SGTR without operator
actions or nonsafety system operation but with the
inadvertent actuation of the ADS was performed to
observe the effect of backflow from me faulted steam
generator on ADS depressurization capability and to
obtain data for determining primary system boron
concentration versus time.
Steam Line Break

This test simulated a large single-ended SLB. It

provided a rapid primary system cooldown transient
and demonstrated the ability of the CMT to provide
primary system mass addition without requiring ADS
actuation.

Test 12 (S01S12). This test was performed at AP600

hot, zero power conditions, and no core decay beat
was used. This test was performed with three PRHR
HX rubes to maximize primary system cooling and
better simulate two PRHR HXs operating in the
AP600. The break size was scaled to simulate a
1388
ft.2 AP600 break area (full steam generator
outlet area) and was performed with no operator
actions or active, nonsafety system operation.

Note: Tests 8, 11 and 12 have been designated as

blind tests. The SPES-2 facility responses for the
tests will be compared to analysis predictions without
prior knowledge of actual test results, therefore, these
test results cannot be discussed at the present time.
SPES-2 TEST RESULTS (BLIND TESTS
RESULTS EXCLUDED)

The passive safety systems mitigated consequences of

the design basis events tested in an orderly and
predictable manner. Therefore, LOCA tests followed
the same sequence of primary and passive safety

system responses. For example, as .shown in

Figure 2, in Matrix Test 3 (Test Run S00303) primary
system pressure decreases rapidly when the break
occurs and results in reactor trip and safety system
actuation. Primary pressure is stabilized by flashing
when pressure decreases to the saturation pressure
corresponding to the temperature of the water in the
rod bundle region, upper plenum, and hot legs. The
CMTs and PRHR HX begin operation and after RCP
trip they operate by natural circulation. Primary
pressure slowly decreases as water is lost through the
break with the PRHR HX and CMTs removing heat.
When the cold legs begin to drain, the CMTs
transition from their natural circulation mode of
operation to a draindown mode, greatly increasing
their net injection rate and effective cooling;
increasing the .rate of primary system pressure
decrease. The draining CMTs result in ADS Stage 1,
2, and 3 actuation, which rapidly decreases primary
pressure and results in accumulator delivery. When
the accumulator injection has been completed, the
CMTs continue to drain which results in actuation of
ADS Stage 4.
ADS Stage 4 completes
depressurization of the primary system below the
elevation head of the water in the ERWST, and
injection flow from the IRWST begins. The IRWST
flow subcools and refills the primary system to the
ADS Stage 4 elevation.
Figure 3 illustrates the net change in primary system
(PZR not included) water inventory during the course
of Matrix Test 3. After break initiation, primary
system inventory steadily decreases due to mass loss
through the break. When the primary system
inventory loss reaches -400 Ibs in SPES-2, the cold
legs begin to drain causing the CMTs to transition to
draindown operation, and the rate of inventory
decrease lessens. In fact, the CMTs draindown
injection rate is sufficient to almost match break flow.
When the CMT has partially drained, ADS Stage 1 is
actuated and is shortly followed by Stage 2 and 3
actuation. After ADS actuation, the accumulators
provide injection at a high flow rate, offsetting water
lost into the PZR and out through ADS Stages 1, 2
and 3, and primary system inventory is restored.
When accumulator delivery is completed, the CMTs
continue to drain, but primary inventory steadily
decreases until CMT low level actuates ADS Stage 4
and IRWST injection begins. IRWST injection
steadily increased the primary system water inventory
until the SPES-2 power channel was completely
refilled.

175

Figure 4 illustrates the collapsed liquid level in the

beated portion of the rod bundle and in (be lower

portion of the upper bead just above toe beated rods,
during the course of Matrix Test 3. The rod bundle
region collapsed level measurement shows that two
periods of high steam fraction occur in the rod
bundle. The first period occurs during the early
portion of the transient and is due to flashing of the
water in the rod bundle resulting from decay heat and
the rate at which primary pressure is decreasing. The
steam fraction reached -45 percent at the time the
CMTs begin draindown. The increase in CMT net
injection flow rate causes the stream fraction to
decrease and subsequent accumulator injection rapidly
refills the rod bundle region with subcooled water.
The second period of high steam fraction occurs after
accumulator injection, when the steam fraction of the
water/steam mixture in the rod bundle steadily
increases to 25 percent until IRWST injection begins.

IRWST injection refills the rod bundle with subcooled

water. The collapsed liquid level measurement in the
region above the beated rods provides a good
indication of the steam/water fraction of the fluid
exiting the top of the rod bundle. This parameter
indicates that the two phase fluid cooling the rod

bundle, reaches a maximum steam fraction of

-50 percent just prior to IRWST injection.

Comparison of LOCA Test Results

accumulator delivery, and the net fluid losses during

the subsequent time delay until ADS-4 actuation.
This time delay is relatively fixed by the CMT
draindown rate. Note that the downcomer water level
is below the elevation of the top of the rod bundle for
many events; however, the rod bundle remains fully
covered with a two-phase mixture due to its lower
density.
Comparison of LOCA Locations

The effect of break location can be examined by

comparing Baseline Test S00303 with Tests S00605
and S01007.
These tests demonstrate two important characteristics.
of break location: the elevation of the break and the
affected line. The break elevation affects the amount
of coolant lost through the break, while the line in
which the break occurs can directly affect the CMT
performance.

In all three tests, primary fluid at the break is initially

single-phase water and, therefore, the break flow is
initially at a high mass flow rate. When the primary

system water level drains to the break elevation, the

fluid at the break converts to two-phase fluid or steam
and break mass flow rate decreases. The DV1 line
break is at the lowest elevation in the primary system,
and the break flow continues to be high until the

Figure 5 shows the rod bundle region fluid steam

primary system fluid level decreases into the annular

fraction estimated from collapsed liquid level. This

occurs just prior to CMT draindownADS
accumulator delivery, for all the LOCA tests. The
fluid steam fraction is plotted vs. the rate of primary

downcomer and reaches the DVI nozzle elevation.

As seen in Figure?, more water is lost from the
primary system before the break flow converts to
two-phase fluid for the DVI line break than for the

system pressure decrease.

two other breaks. Similarly, more water is lost in

Test S00303 than in Test S01007, since the break for
S01007 is effectively from the top of the cold leg.

The steam fraction is

directly related to the rate of depressurization,
indicating that flashing is a major contributor to the
two-phase fluid steam fraction in the rod bundle at
this time, in addition to boiling. Since the primary
system inventory is the same for most tests when the
CMTs begin to drain, the flow to the rod bundle, as
dictated by the water elevation on the cold-leg side of
the power channel, are also similar and the rod bundle
fluid steam fraction is break size dependent.

Figure 6 shows the relationship between steam

fractions in and above the rod bundle with the
minimum downcomer level just prior to the IRWST
injection. Figure 6 shows that the high rod bundle
fluid steam fraction, occurring prior to ERWST
injection, corresponds to the downcomer water level
at that time. This water level results from the
primary system water inventory at the end of the

176

Figure 7 shows a decrease in the slope of the SO 1007

breakline catch tank curve first, when water level in
the primary system has decreased to the top of the
cold-leg. This change in slope indicates the
conversion of breakflow from single-phase water
(from a full cold-leg pipe) to two-phase fluid. In Test
S00303 the decrease in break flow occurs later. In
Test S00605 no change in break flow is indicated
until after ADS actuation.
Since the break flows for all three breaks are initially
similar, the water level decreases to the top of the
loop-B cold legs at similar times for all three tests.
This is important for CMT draindown initiation.

In all three tests, CMT transition to the draindown

operating mode began at similar times. ID the DVI
line break test, CMT-A and -B transitioned
simultaneously. In the Cold Leg Break Test, the cold
leg with the break (CL-B2) drained prior to the intact
cold leg: since CMT-B connected to B2 transitioned
slightly earlier than CMT-A. In the balance line
break test, CMT-A transitioned at approximately the
same time as the other tests; however, both CMT-B
recirculation and draindown modes of operation were
affected. Natural circulation was initially suppressed
by the dP caused by the single-phase flow from the
cold leg to the break. When two-phase flow to the
break started, some flow to the top of CMT-B
occurred where the steam was condensed by the cold
CMT water. This resulted in wide variations in CMT
flow until full transition to draindown operation
occurred after ADS actuation.

In summary, break location bad a significant impact

on the simulated 2-in. LOCAs. Break elevation
affected the amount of reactor coolant loss through
the break, which significantly influences the power
channel coolant inventory prior to start of IRWST
injection. Also, a 2-in. break in the cold leg to CMT
balance line will initially prevent CMT recirculation
and delay transition to draindown of the affected
CMT. However, this did not greatly affect overall
coolant inventory or the ADS-4 timing since it is
actuated by the level of the unaffected CMT.

In all three tests, ADS initiation occurred at similar

times, since the transition to draindown for at least
one of the two CMTs occurred at similar times in all
tests.

The effect of break size can be comparing the

baseline test S00303 (2-in. LOCA in the bottom of

Figure 5 shows the rod bundle steam fraction just

prior to accumulator injection is higher for the DVI
line break than for the cold-leg and balance-line
breaks (Tests S00605. S00303. and S01007,
respectively). This indicates that flow through the rod
bundle and the fluid inventory is lowest for S00605
(DVI line), and highest for SO 1007 (balance line).
Also the downcomer level for S00605 rapidly
decreases to the DVI nozzle level before accumulator
injection starts while the annular downcomer is full
for S00303 and SO 1007 prior to accumulator
injection.

Figure 8 shows the coolant inventory for the primary

system (except pressurizer) during Test S00303 and
Test S00401. For the 2-in. break, the coolant
inventory rapidly decreases to the -400 Ibm level

Prior to IRWST injection the rod bundle steam

fractions for Test S00303 (cold leg break) and Test

Comparison of LOCA Sizes

The SPES-2 LOCA tests included four break sizes:
1-in., 2-in., and two DEGs. The 2-in. and DEC

breaks were performed at two or more break

locations.

CLB-2) with test S00401 (1-in. LOCA in the bottom

of the CLB-2).

(corresponding to the elevation of the top of the coldleg) and CMT draindown mode is initiated. Since the
CMT draindown flow rate is less than the break flow
rate, the CMT draindown is uninterrupted and ADS-1
is actuated. The primary system pressure decreased
to just below the accumulator gas pressure when
ADS-1 actuated, so there was very little injection
from die accumulators. When ADS-1 occurred,
essentially all of the accumulator water inventory was
still available for refilling the power channel.

S01007 (balance line break) are very similar, as are

Test S00401, the 1-in. break, is four times smaller

the minimum downcomer levels. In test S00605

(DVI line break) the rod bundle fluid steam fraction,
and minimum level in the tubular downcomer is
lowest (Figure 6). Lower water inventory in Test
S00605 is due in pan to the fact that there is less
downcomer and power channel water inventory after
accumulator injection than in S00303 and S01007.
During the accumulator injection for the cold leg and
the balance-line breaks, the rod bundle is subcooled
and completely refilled, as seen in the increasing
collapsed level for the rod bundle shown in Figure 4.
However for the DVI line break, the collapsed level
in the rod bundle indicates two-phase flow still exists
at the end of accumulator injection.

than the 2-in. break, and therefore, the inventory

decreased at a slower rate. When the cold legs finally
began to void, causing the CMT balance lines to
drain, the CMTs started a period of intermittent short
draindowns followed by refill and natural circulation,
which increased their overall injection rate sufficiently
to keep the water level in the cold leg. In fact, the
CMT draindown injection capability at this time,
exceeds the break flow. The intermittent CMT
draindown and refill extended the time for draindown
and delayed ADS-1 actuation. Primary system
pressure bad decreased below the accumulator gas
pressure and the accumulators injected -25 percent of
their water inventory prior to ADS-1. Therefore, the

177

coolant inventory after accumulator injection just prior

to IRWST injection is less for the 1-in. break than the

2-in. break and resulted from the expenditure of
accumulator inventory prior to ADS-1; so less water
was available for injection between ADS-1 and
ADS-4. The lower inventory is also reflected in the
higher rod bundle steam fraction prior to IRWST
injection for the 1-in. break, as compared to the 2-in.
break.

TestS00706 (DEC of the DVI-B line) is very

different from the 1-in. and 2-in. breaks.

In Test S00706, there is a complete loss of injection

from CMT-B, accumulator-B, and one of two IRWST
injection lines. This, in addition to the high break
flow, results in a very low minimum coolant
inventory. The minimum coolant inventory in the
power channel still provided adequate cooling
although the two-phase fluid flow in the rod bundle
had a much higher steam fraction. Figure 9 shows
the difference between the temperature of Heater
Rod 87 at the highest elevation in the bundle
(TW020P87) and the saturation temperature
corresponding to primary system pressure for S00706
and S00303. The Test S00706 rod temperature
followed the saturation pressure as in the Reference

effect on me test The CMTs' draindown stopped

when the NRHR flow started due to the backpressure
NRHR flow imposes on the CMT discharge line.
ADS-4 never occurred since CMT draindown stopped
before reaching the ADS-4 trip level.

Figure 10 shows the coolant inventory for

Tests S00504 and S00303. The CVCS injection
helped to maintain coolant inventory during the first
part of TestS00504 until ADS-1" occurred. The
inventory losses after accumulator delivery are similar
for the two events until the NRHR injection matched
the primary system inventory loss out the break, then
the primary system started to refill. The minimum
inventory for Test S00504 was higher than in
Test S00303 since refill was not delayed until ADS-4
and IRWST injection.
Comparison of PRHR Performance

The effect of additional PRHR capability on the

LOCA mitigation can be assessed by comparing Test
S00401 (performed with one PRHR tube) with Test
S01613 (performed with 3 PRHR tubes). Both of
these tests are 1-in. breaks in the bottom of cold legB2.

Test S00303, despite the higher void fraction in the

The PRHR performance measured for Test S01613

rod bundle.

was slightly higher than for S004Q1, and showed that

the 200 percent increase in beat transfer area for the

Effects of Nonsafety Systems

The effect of nonsafety systems operation on the

passive safety system response and overall plant
response can be assessed by comparing Baseline Test

PRHR HX yielded some additional heat transfer. The

PRHR How is slightly higher for SO 1613; however,

the biggest difference is that the PRHR HX exit
temperature was lower for S01613 than for S00401.

S00303 with Test S00504. S00504 is a 2-in. break in

Figure 11 shows the coolant inventory for Tests

the bottom of cold leg B-2, with the CVCS and

NRHR pumped injection and SFWS addition to the
steam generators simulated. The CVCS injection
started on the safety systems actuation (S) signal, and
the NRHR started to inject coolant after ADS-3 when
the primary system pressure was reduced to less than
the NRHR pump discharge pressure.

S00401 and S01613. Both tests spend an extended

time period at the -400 Ibm inventory level. In
S01613, the CMTs expended less inventory
maintaining the coolant inventory, due to extra PRHR
HX return flow and additional injection from the
accumulators caused by the lower system pressure.
ADS-1 was therefore delayed for S01613 relative to
S00401. When ADS occurred, less fluid was
discharged through ADS due to the lower initial
system pressure at the start of ADS. Therefore,
S01613 had more coolant inventory than S00401 after
the accumulator delivery. The net coolant losses from
the end of accumulator delivery until the start of
IRWST injection were very similar for the two
events. However, since S01613 started this period
with more coolant inventory, it also ended with more
inventory than S00401, at the point of minimum
coolant inventory in the vessel.

Test S00504 is very similar to S00303 until ADS-1

was actuated. The reactor trip (R) and the S signals

occurred at nearly identical times, and the pressure
decrease vs. time were similar. The rod bundle steam
fraction prior to ADS were very similar for the two
tests. The CVCS injection did not have significant
impact on the initial pan of the test since break flow
greatly exceeds the maximum possible flow from two
CVCS pumps. The NRHR Hows started when
primary pressure was -160 psia and bad a significant

178

The greater PRHR beat removal in S01613 increased

pressurizer completely filled. In SO 1309, tbe top of

the primary system pressure decrease relative to

S00401 prior to ADS-1. The overall effect of the

tbe pressurizer remained superbeated throughout tbe

test, so it never refilled completely. This occurred

lower system pressure mitigated the severity of

because two pressurizer external heaters were kept on.

the test.

In Test SOI 110, the heat losses were sufficient to

reduce pressurizer pressure and temperature. In both
of these tests the ADS was not actuated since tbe
CMTs remained in their natural recirculatioo mode
throughout the test.

Comparison of Steam Generator Tube Rupture

Two steam generator tube rupture events were

performed (SOI 110:

SGTR without nonsafety

systems and SO 1309: SGTR with nonsafety systems

and operator actions). There were significant
differences between the two tests and these are
attributable to the effects of nonsafety systems
operating for SOI309. Specifically the use of the

These tests demonstrate that the primary-to-secondary

flow due to the SGTR can be terminated by the
passive safety systems with no operator actions, or by
the operators using the non-safety systems in a
manner similar to current PWRs.

SFW system and the steam generator-A PORV to

maintain the primary system cooldown rate provided

sufficient heat removal to maintain single-phase flow

Conclusions/Observations

conditions in the primary system. Also, two of the

1. Tbe passive safety systems mitigated the

six pressurizer external beaters were operated in

SO 1309.

consequences of the design basis events tested in

In Test SOI 110, tbe primary to secondary flow

through the SGTR was quickly terminated with no
operator action, as shown in Figure 11.

2. The passive safety systems were able to prevent

dryout or beater rod temperature increase for all
LOCA's up to and including a DEG of an 8-in..
DVI line.

Figure 12 shows primary system pressure for the two

tests. Tbe SO 1309 pressure is lower than SOI 110 due
to the additional cooling provided by CVCS injection,
SFWS addition, and steam generator-A PORV
actuation. The primary system pressure was actually
higher than steam generator-A pressure throughout tbe
test for SO 1309. The steam generator-A provided
heat removal from tbe primary system to maintain tbe
primary system at single phase flow conditions. The

dP measured in tbe steam generator-A U-tubes,

showed that primary system natural circulation flow
was maintained through steam generator-A.
Figure 13 compares tbe rod bundle differential
pressure for the two transients, snowing that twophase flow conditions eventually occurred in the
primary system in Test SOI 110, while S01309
maintained single phase flow through tbe core until

an orderly and predictable manner.

3. Stable long term cooling of the rod bundle was

established by ERWST water injection, through

the rod bundle and out through the ADS Stage 4

flowpaths.
4. Following a single, double-ended SGTR, the
passive systems provided primary system
boration and beat removal, and terminated
primary- to secondary-side leakage with no

operator action, no operation of non-safety

systems, and no actuation of non-safety systems,
and no actuation of ADS.

5. Non-safety system operation improved overall

performance.

REFERENCES

test termination.

Figure 14 shows pressurizer levels differed greatly

between tbe two tests. In SOI 110 tbe pressurizer
refills completely. In SO 1309, tbe pressurizer only
partially refilled in response to venting through tbe
ADS-1 flow path by operator action, and a low level

1. SPES-2 Test Final Data Report, WCAP-14309,

Rev. 0, March 1995.

2. SPES-2 Facility Description, WCAP-14073, Rev.

0, May 1994 (SET Document 00183R192, Rev.
0, April 1994).

was maintained for the rest of tbe test. Measured

temperatures in tbe pressurizer in SOI 110 indicated
that die pressurizer was subcooled, at which time tbe

3. AP600 Standard Safety Analyses Report, AP600

Doc. GWGL021, Rev. 2, March 1995.

179

FIG.

P-027P

1.

AP600 passive safety systems configuration

111

PRZ P r e s s u r e

0 Tap

CO
C/3

cu

a>
!_

=3

co
CO

<JJ

T ime

FIG.
180

2.

( s)

Facility response summary for S00303

i < j I

U .1 s s

( 1 n |i u I

I > I |. u I >

f ,i r

f . - t

X I) I) M 0 U

2 I) 0

11 U I)

Time
F/G. 3.

L_OOOP
I. A I 5 P

483
19

Comparison of break locations

Change in system mass inventory

0
0

0 PC H e a l e d Red R e f i o o
D I'C II p p e r p I r num
II L A

T ime
FIG. 4.

(s)

( s )

SPES - 2 Test: S00303

181

80

70

60

50

40

30

.5

1.5

Depressurization Rate
FIG. 5.

2.5

(psi/s)

Rod bundle steam fraction before ADS actuation

[UPPER PLENUM
|*OD
BUS* [ M E O V T L E T
I R0 n

(T*F)

111' N D I. R

100

DOWNCOMER LEVEL - FEET A B O V E BAF

FIG. 6.
182

Fluid steam fraction in core at minimum coolant inventory

1400
1 ZOO

1000800
600

JJL

4 00 200
0

T ime

(s )

Total break flow

FIG. 7.

Q_______

ZOO

-zoo
- 4 00

-GOO

-too

FIG. 8.

TftlirtT

Comparison of break sizes

Change in system mass inventory

Tlt

to

Uff.r

<<

*}

T t t z i r i T . rstt ! urt" < i i > t i

go

Time

FIG. 9.

(s)

Comparison of break sizes

Rod temperature relative to saturation temperature
183

200

flOO

T ime

FIG.

10.

(s )

Effects of non-safety systems

Change in system mass inventory

200

0 -

.a

-ZOO

-400

600

BOO

T ime

FIG.

11.

Comparison of PRHR performance

Change in system mass inventory

T ime

FIG.
184

12.

(s)

( s)

Primary to secondary SGTRflow rate for SOI 110

3500

2000

1500

1000

SCO

0 -1

Time

FIG. 13.

( s)

Comparison of steam generator tube rupture

Primary system pressure

_ 12

10

Time

FIG. 14.

Comparison of steam generator tube rupture

Rod bundle differential pressure

Time

FIG. 15.

(s )

(s )

Comparison of steam generator tube rupture

Pressurizer level
185

EXPERIMENTAL INVESTIGATION ON AN INHERENTLY

ACTUATED PASSIVE INJECTION AND DEPRESSURIZATION SYSTEM

L. MANSANI, L. BARUCCA
ANSALDO Nuclear Division,
Genova

G.P. GASPARI
SIET, Piacenza
Italy
Abstract

ANSALDO has conceived an inherently

actuated

Passive

Injection

and

Depressurization System (PIDS) to be used in Nuclear Power Plants (NPP).

The adoption of the PIDS would allow to enhance the reliability of the Safety

Systems, with the additional advantage of a simplification of the NPP current

designs.
Due to the innovative concept of this system, an experimental investigation has
been performed at SIET Laboratory aimed at exploring the physical phenomena
governing the behavior of the system, with the basic goal to demonstrate the

concept viability.
After a brief description of the PIDS concept, the paper presents the

experimental tests matrix and some results.

The obtained experimental data confirm the validity of the concept, justifying

and encouraging the continuation of the activities and calling for a further
development and the execution of related experimental activities.

INTRODUCTION
In a Nuclear Power Plant (NPP) one of the means to mitigate the
consequences of a Loss of Coolant Accident (LOCA) is to inject cold water into
the Reactor Coolant System (RCS) at low pressure. This requires RCS
depressurization. The larger effort finalized to work out design solutions for the
next generation of NPPs, capable to obtain a step ahead with respect to the
already satisfactory safety standards as well as a simplification of current
design, has produced the concept of an innovative inherently actuated Passive
Injection and Depressurization System (PIDS). The PIDS is a system, conceived
by ANSALDO, to perform the depressurization function without making use of
any active component or actuation logic. ANSALDO, conscious of the
innovative concept, before proceeding to the actual system design, committed
SIET to perform an experimental investigation aimed at exploring all the physical
phenomena and demonstrating the concept viability.

187

The paper gives a PIDS concept short description and presents the main
experimental results. The final experimental data report is still in progress.

CONCEPT AND WORKING PRINCIPLE DESCRIPTION

ANSALDO conceived a completely Passive Injection and Depressurization
System which does not make use, to perform its function, of any active
component, actuation logic or operator action.
The basic concept is to perform the required depressurization function by
mixing (berated) cold water with the steam present in the RCS. The cold water
injection is inherently actuated on low RCS water inventory, with the
depressurization (down to the containment pressure) completed by a valve
passively actuated on low RCS pressure. The approach of depressurizing by
adding mass, is of interest for a NPP in which the current Automatic
Depressurization Systems intervention causes, in most cases, a large mass
depletion due to the level swell in the RCS while depressurizing.
Figure 1 shows a PIDS basic configuration. A tank of (borated) cold water,
connected to the RCS steam space, is equipped with a syphon shaped
discharge line. The syphon ascending leg connects the tank bottom and
overcomes the top of the tank by some meters, the descending leg connects an
ejector. The ejector has two additional connections: upstream with the RCS
steam space and downstream with a horizontal small condenser. The condenser
outlet line terminates in the RCS. The elevation of the condenser is determined
according to a predefined RCS level set point for system actuation. An injection
spray line connects the line downstream the ejector to the RCS location chosen
for cold water injection.
Syphon

EJECTOR

Self actuated
Depressurization

Valve

Fig 1 - PIDS BASIC CONFIGURATION

188

In normal conditions, the system can be seen as constituted by two

communicating, water filled, vessel subsystems: the first one being the tank and
syphon ascending leg, the second one being the pipe downstream of the
ejector, the condenser and the RCS water space. When the water level in the
RCS decreases below the condenser elevation, the condenser starts to
condense the steam coming from the RCS steam space through the ejector. The
resulting pressure reduction in the ejector throat will clear the vapour space
over the syphon hydraulic seal and will trigger the syphon. From now on, the
cold water injection will be stable until the tank empties.
The cold water may be sprayed into any vapour space of the primary
system available at the syphon triggering time, thus causing the RCS
depressurization up to a pressure value well below the minimum pressure
reached under any operational transient. The remaining depressurization down
to the containment pressure, will be completed by pilot operated Passive
Depressurization Valves (PDV) passively actuated on low RCS pressure.

EXPERIMENTAL TESTS
In principle, the PIDS is applicable both to Boiling and Pressurized Water
Reactor designs. However the PIDS developed so far has been conceived for
application to a PWR design. It is composed of two subsystems: a) the high
pressure subsystem, which consists of the ejector, the condenser, the water

tank and the connection lines to/from the RCS, b)the low pressure subsystem,
which consists of a Passive Depressurization safety Valve (PDV). The two
subsystems operate in sequence: the former is required to depressurize the RCS
below the PDV actuation set point, the latter is required to continue the system
depressurization from the PDV set point down to the containment pressure.
An experimental campaign has been undertaken on a properly scaled test
facility, designed by ANSALDO and constructed by SIET, with reference to a
medium size PWR (like for instance the Westinghouse AP600), to test the PIDS
high pressure subsystem. The specified tests are "conceptual" tests and
therefore the quantitative results (such as depressurization rate and depth) are
not intended nor expected to be representative of the real PWR system ones.
The goals of the experimental tests are:
to confirm that the system is actuated as soon as the water level uncovers
the condenser;
to confirm that, for sufficiently small breaks (less than 3" equivalent
diameter) wherever located, the system is not activated by any spurious effect
until the water level uncovers the condenser;
to confirm that following the syphon triggering the injection water flowrate
is stable until the tank empties;
to asses the PIDS performance in presence of non-condensable gases.

TEST FACILITY DESCRIPTION

The test facility has three tanks simulating in a 1:30 volume and full height
scale, the volume of an AP600 Core Make-up Tank (CMT), Pressurizer (PRZ)
and Reactor Vessel (Fig. 2). Water can be heated up to 300C at a pressure of
8 MPa by means of steam circulating through dedicated U-tubes bundles inside
the RV and PRZ.

189

of,

MS,

COD US

C-SCH.M

Fig. 2 - SKETCH OF THE TEST FACILITY AND INSTRUMENTATION

Table 1 gives the major geometrical data of the main facility components which,
in addition to the above mentioned tanks (CMT, PRZ and RV) include a
Condenser, an Ejector and related piping. The facility includes a device to
simulate the break at various locations. The facility is thermally insulated,
except the upper portion of the pipes normally filled with steam, so that the
heat losses are reduced. The reason for not insulating the mentioned portions of
lines is to prevent an excessive water temperature decrease in the line
downstream the ejector, which might generate a spurious system actuation
during RCS level decrease before the condenser is uncovered. The test facility is
sketched in Fig. 2.

190

Table 1 - GEOMETRICAL DATA OF THE MAIN PIDS COMPONENTS AND PIPES

SCALE

Elevation
Volume

1:1
1:30

28

OVERALL ELEVATION (m)

PRESSURI2ER (PRZ)

Inner diameter (m)

0.363

Height (m)

11.50

Volume (nr>3)

1.2

Inner diameter (m)

0.98

REACTOR VESSEL (RV)

Height (m)

4.2

Volume (m3)

2.1

Inner diameter (m)

0.69

Height (m)

4.20

Volume (m3)

1.5

COLD WATER TANK

HOT LEG / COLD LEG

Inner diameter (m)

0.146 (6" sen 80)

Inner diameter (m)

0.067 (3" sch 160)

Inner diameter (m)

0.038 (1.5" sch 80)

Inner diameter (m)

0.032 (1.25" sch 80)

SURGE LINE

INJECTION LINE

BALANCE, EJECTOR - CONDENSER

LINES

Thermalhydraulic instrumentation is installed for measuring the parameters

of interest. The following types of instrumentation are provided:
- 34 thermocouples to measure the fluid temperature in the pipes and
components;

one flowmeter to measure the injection mass flowrate;

8 pressure transducers to record the absolute pressure within the various

tanks;
18 differential pressure transducers to record pressure drops and liquid
levels in tanks and pipes.
All instruments were calibrated in laboratory before their installation on the
plant. The instruments overall accuracy resulted 1.1 C for thermocouples,
20 kPa for absolute pressure transducers, 0.15-5-0.30 kPa for differential
pressure transducers. The maximum estimated error of the main interesting
quantities (error analysis is underway) is 1.5 C for the fluid temperature,
50 kPa for the fluid pressure, 0.3 -5- 0.5 kPa for the ejector pressure drops
and 0.05 * 0.10 m for the water level.
The test facility has digital data acquisition systems to record all the
instrument signals.

191

TEST MATRIX AND TEST PROCEDURE

Several tests are included in the test matrix given in Tab.2 with different
break sizes and locations. For all the tests, but #12, the initial pressure is 8 MPa
while the RV water temperature is set 250 C, for RV and cold leg breaks, and
275 C for the top PR2 breaks. LOCA analyses /4/ /5/, performed on the
reference NPP, highlighted that:
- the RV water flashes and, as a consequence, the system pressure stabilizes
at a constant value ( 6 MPa);
- for breaks in the RCS water region, the PRZ empties before the RV flashing;
- for breaks in the RCS steam region (i.e. in the top PRZ) a water level still
remains inside the PRZ also after the RV flashing.
Table 2 - PIDS TEST MATRIX
RUN
tt

BREAK

SIZE

INITIAL CONDITIONS

LOCATION

P (MPa)

NOTES

RVTemp (C)

1"

RV bottom

250

Cold Leg
Cold Leg

250

12-

250

3"

Cold Leg

250

1'

Cold Leg

250

0.4 kg of Nitrogen injected under water

1.5"

PRZ top

275

Transient limitations due to CCFL in surge line

1.25'

PRZ top

275

Transient limitations due to CCFL in surge line

1 81"

PRZ top

275

Transient limitations dug to CCFL in Hot Leg nozzle

3"

PRZ top

275

Modified surge line geometry

Transient limitations due to CCFL in Hot Leg nozzle
Modified surge line geometry

10

1"

RV bottom

250

0 2 kg of Helium plus 0 7 kg of Nitrogen

N2 in equilibrium at transient start

both injected under water

He and N2 in equilibrium at transient start
11

1"

RV bottom

250

Test 1 repetition

12

3"

RV bottom

45

250

0 23 kg of Helium plus 0 52 kg of Nitrogen injected into

PRZ steam space

He and N2 not in equilibrium at transient start

Cold water injection was into RV for all runs but #1 which was into Hot Leg

In the experimental facility, due to limitation in the maximum operating

pressure, the above behavior is reproduced by setting 250 C the RV water
temperature for RV and cold leg break tests and 275 C for top PRZ break
tests.
The amount of non-condensable gases used in the tests is a conservative

upper bound which takes into account the RCS hydrogen content in normal
operation as well as the hydrogen and oxygen produced by radiolysis after the
LOCA (a period of three hours has been considered). Helium and nitrogen have
been used in the tests instead of hydrogen and oxygen.
The tests have been conducted using the following procedure. Initial steady
state conditions were established by heating the PRZ and RV with steam
circulating inside dedicated tubes, then the break valve was opened. As soon as
saturation conditions were reached in the RV the system pressure was
automatically kept constant until the tank water injection started by heating the
RV. During the water injection the power to the RV (supplied by steam) was
maintained at the same value as at the system triggering instant.

192

TEST RESULTS
Among the performed tests, the 1" and 3" cold leg break results are here
summarized.
For both tests, when the initial conditions were reached (system pressure 8
MPa,
RV temperature 250C and nominal water level in the PRZ) the break
valve was opened.
1" cold leg break
The system pressure decreased down to 4 MPa in approximately 1000 s, at
which time the RV water started to flash, (Fig. 3). From this time on the
pressure was kept constant by heating the RV until the condenser uncovered
(4615
s) and hence water injection started, from the tank to the RV. The water
level continuously decreased in the PRZ, ejector and injection lines due to the
loss of mass inventory. As soon as the condenser has been uncovered (461 5 s},
Fig. 4, the condensation rate was enhanced, with steam supplied from the PRZ
through the steam ejector. The resulting pressure reduction in the steam ejector
throat, Fig. 5, cleared the vapour space over the syphon hydraulic seal and
triggered the syphon. The water injection started and, from this time on, the
system was depressurized by the cold water injected into the RV, Fig. 3.
3" cold lea break
The same transient events as for 1" cold leg break were observed, of
course, in shorter times. The system pressure decreased down to 4 MPa in
approximately 190 s, Fig. 7, and then kept constant (the pressure peak being
caused by a mismatch in the RV heating control). The condenser uncovered 645
s after the break, Fig. 8, at which time the cold water injection started, Figs. 9
and 10.
The above behavior was observed, though with different time histories, also
in all the remaining tests, including those with non-condensable gases.

800X0

OLBFKIBtesC PRZpresare (PTO2J

I
2X0

xno

COO

TlnW

SXD

6000

Rg. 3 - 1" COLD LEG BREAK - PRZ PRESSURE VERSUS TIME

193

OBRKIBtese 5eoor fine level (EJEU

25 -i
20 -

15

10

s -

I
4X0

3030

JCOO

I
5X0

too

fig. 4 -1" COLD LEG BREAK - EJECTOR UNE LEVEL VERSUS TIME

CLBFK1B test Wet-Treat Rector dtterata pests* (DPTO3)

3SO

oo -

-2SO -

-SQO -

750

KEO

2000

COO

5020

I
3X0

Fig. 5 -1" COLD LEG BREAK - INLET-THROAT EJECTOR DIFFERENTIAL

PRESSURE VERSUS TIME

OERCIBtesC Injection flowrate (F-UO1)

15
3D
iS 20
U UJ 05 -

ICGO

2CQD

3000

<DO

5X0

<OD

Flfl. 6 .1" COLD LEG BREAK - COLD WATER INJECTION FLOWRATE VERSUS TIME

194

CLBRK3 test: PRZ presare (PTQ3

I6CO

Fig. 7-3" COLD LEG BREAK - PRZ PRESSURE VERSUS TIME

CLBRK3 test Rector lire level (EJRJ

CO

8CO

1200

'1600

Fig. 8 - 3" COLD LEG BREAK - EJECTOR LINE LEVEL VERSUS TIME

220
200 -

OBRK3test: Wet-TlToat Rector dffenrtial pressuB (DFTO3

iao -

120
100

(0 -

23
0

,-20
O
60
83

100

120
HO

~r
0

800

1203

IfiD

Fig. 9-3- COLD LEG BREAK - INLET-THROAT EJECTOR DIFFERENTIAL

PRESSURE VERSUS TIME

195

CLBRK3 test :ln|ecaon Uovvrate (FU=01)

4 -

3 -

i
0

tO

800

1200

^ Iffl)

Fig. 10-3" COLD LEG BREAK - COLD WATER INJECTION FLOWRATE VERSUS

CONCLUSIONS
The performed tests have demonstrated the PIDS concept viability. In
particular the tests have confirmed that:
- the system is actuated as soon as and every time the water level uncovers
the condenser. The system actuation has always occurred at about the same
RV level;
- no spurious effects, capable to actuate the system before the water level
uncover the condenser, has been experienced. A wide margin against early
actuation has been evidenced;
- the injected flowrate has resulted stable;
- the presence of non-condensable gases, even at very high concentration, had
no adverse effect on PIDS actuation and did not reduce its performances;
- although the test objective was not to get information of RCS
depressurization rate, a very fast depressurization was observed in all the
tests during the water injection phase.
The successful test results strongly encourage to continue the PIDS
investigation program activity. The next steps will consist in:
- separate effect tests aimed to gain information on the attainable
depressurization rates and on the associated phenomena;
- integral scoping tests execution aimed at assessing the PIDS interaction with
the other NPP systems.
REFERENCES

/1/ Patent No. PCT/EP94/03162 "Depressurization System for Plants Operating

with Pressurized Steam", September 22, 1994.
121 L. Mansani, "Passive depressurization system large scale test specification".
ANSALDO STU 5100 SMEX 0020000 , January, 1994
/3/ G.P. Gaspari, "Passive injection and depressurization system - Test facility
description", SIET 00324 Rl 94, May, 1995
/4/ L. Mansani, R. Lenti, G. Saiu, "Preliminary accident analysis to support a
passive depressurization system design". Proceeding of the Int. Conf. New
trends in nuclear system thermohydraulics, Pisa, May 30 - June 2, 1994
/5/ Westinghouse, 1992 - AP600 Standard Safety Analysis Report (Chapter 15:
Accident Analysis)

196

DESIGN AND TESTING OF PASSIVE HEAT

REMOVAL SYSTEM WITH EJECTOR-CONDENSER

K.I. SOPLENKOV
Research Institute for Nuclear Power Plant Operation,
Moscow, Russian Federation

V.G.
SELIVANOV
Kharkov Aeronautical Institute,
Kharkov, Ukraine
Yu.N.

FILIMONTSEV

Research Institute for Nuclear Power Plant Operation,

Moscow, Russian Federation
B.I. NIGMATULIN
Electrogorsk Research Engineering Center,
Russian Federation
V.V.
BREDIKHIN
Kharkov Aeronautical Institute,
Kharkov, Ukraine

E.I. TRUBKIN
Electrogorsk Research Engineering Center,
Russian Federation
E.Z. EMELJANENKO
Odessa Polytechnic Institute,
Odessa, Ukraine
A.W.
REINSCH
Southern California Edison
California, USA
Abstract

The objective of the analysis, design, and testing of a new type of passive heat removal
system is the development of a concept with the capability to terminate a broad spectrum
of postulated accident sequences. Its principle is based on the dynamic form of natural
convection utilizing inertial forces instead of gravity for fluid circulation. The process
develops in a loop combining an ejector specifically designed for dynamic natural
convection and heat exchanger in a fixed geometry. This simple configuration, independent of electric power and automatic controls, is capable of coping with the majority
of initiating events occurring in light water reactors. Since motive power does not depend
on gravity, heat can be rejected from a high elevation to a lower level, reducing capital
costs by locating the heat sink at the ground level.
A broad analytical and experimental development program for a passive heat removal
system with ejector-condenser (PAHRSEC) has been conducted at in cooperation with
2 3 4
, , . Various ejector types and PAHRSEC flow schemes were tested. The most
197

promising one removes heat from steam generator, and has been tested in a 3.5 MW
PAHRSEC facility at . The system starts up rapidly and maintains stable operation over

a wide range of thermohydraulic parameters and transient conditions. Based on the results
from these comprehensive tests, PAHRSEC loops have been proposed as retrofits for the
currently operating WWER-440 nuclear power plants at Novovoronezh. The analysis of
a station black-out at a WWER-440 plant equipped with 4 PAHRSEC of 10 MW each
was performed with the integral codes DYNAMICA (Russia) and TRAC-PF1/MOD2
(USA). The results prove consistently that heat removal from the reactor core was ensured
for extended time with only two out of four PAHRSECs operating. The last phase of
validation effort, the construction and adjustment of a mil-scale test loop of the PAHRSEC
designed for a WWER-440 plant, is proceeding swiftly. Completion of the facility and
first tests are scheduled for the second quarter of 1995.
The concept of heat removal by steam condensation in an ejector recirculating the
condensate to the nuclear steam supply system is applicable to all nuclear plants with a
steam cycle power conversion system. A modified PAHRSEC can serve as passive heat
sink in the containment under severe accident conditions.

1. Principle of Dynamic Natural Convection

Since me PAHRSEC process functions within a fixed geometry, flow developing

within the system boundaries is a form of natural convection which, in this particular
application, is based on inertia! forces. In contrast, natural convection induced by
hydrostatic or aerostatic pressure differences is caused by gravitational forces. Most of
the natural fluid phenomena such as wind, ocean currents, rivers and fire are controlled
by static natural convection depending on gravity essentially. A common feature of static
and dynamic natural convection is that density differences resulting from energy exchange
in a fixed geometry generate the flow.
For steady compressible flow in a streamline, the equation of motion without viscous
effects in differential form is
p g dz + p w dw + dP = 0

The integral of the first term represents the hydrostatic or aerostatic pressure difference
between two elevations. Static natural convection develops if heat is added at low
elevation to a fluid and removed at high elevation. In an analogous way, dynamic forces
for fluid propulsion are developed if heat is added at low fluid velocity and removed at
high fluid velocity. The second term in the differential equation of motion represents
dynamic pressure gradients. Accelerating and decelerating flow in dynamic natural
convection correspond to ascending and descending flow in static natural convection. In
the PAHRSEC concept, heat is added, for example, in the steam generator at low flow
velocity and removed in the ejector at high flow velocity by mixing accelerated steam
with water.
Since dynamic forces and, therefore, dynamic natural convection are independent of
gravity, the heat sink can be located at a lower elevation than the heat source, reducing
costs substantially.
198

2. Ejector-Condenser

One of devices realizing the idea of dynamical convection is an Ejector-Condenser

(EC). Fig.l shows a typical geometry of EC.
Steam first comes into the Laval Nozzle and speeds up to a high velocity. Next, in
the Mixing Chamber (MC) it mixes with "cold" water that takes off its heat, mass, and
kinetic energy. At the outlet of MC the resulting two-phase mixture gets a supersonic
flow. Then, in the Diffusor the flow kinetic energy turns into potential form. As a result
the EC pressure can exceed essentially initial steam and "cold" water pressures. Fig.l
shows also typical pressure profiles in EC at different values of return pressure.
water

FIG. 1. Ejector-Condenser for PAHRSEC.

Presently there are two trends to use of EC in NPP Safety Systems:

Development of high-pressure EC to supply water hi reactor or in containment while
accident conditions: [1], [2], [3] and others.
Development of closed loops for heat removal from the 1st or the 2nd circuits as
well as from the containment with EC as a pump. In that case EC hi PAHRSEC
should have the special properties:
simple start-up (without bypass in MC)
low EC pressure difference
high "cold" water temperature (about 100 - 110 C)
a quite wide range of parameters for stable operation
199

3. PAHRSEC

In this paper some PAHRSEC study results for operating NPPs and new generation
ones are presented.
Fig.2 shows the heat removal scheme for SG under accident conditions. The scheme
"steam-liquid" has been tested as a most promising one for currently operating and new
generation reactors.
Scheme "Steam-Steam"
(S-S)

Scheme "Steam-Liquid"
(S-L)

1 - Steame Generator (SG); 2 - Ejector-Condenser; 3 - Heat-Exchanger;

4 - Check Valve; 5 - Start-Up Valve; 6-Start-Up Tank

FIG. 2.

Heat removal from the steam generator under accident conditions.

PAHRSEC starts up as the valve 5 opens. Steam from SG (1) enters the EC nozzle
(2) and "cold" water from SG (1) through HE (3) enters MC. After EC the resulting water,
through the check valve (4), returns into SG. The ultimate heat sink is the HE water under
atmosphere pressure. The analysis of PAHRSEC and its testing, presented below, have
shown that PAHRSEC has the following properties:
passive operation regime
simplicity of passive start-up and capability to be restarted up repeatedly (as well as
manually)
< short time interval between an accident and the beginning of heat removal

maintaining stable operation over a wide range of SG parameters

resistance to "strong" external disturbances such as the safety valve being opened
capability to control (automatically or manually if necessary) its power for reliable

cooling.
Again, the System is consistent with other safety systems and uses standard and
operating equipment that is very important for its swift installation at reactors in service.
200

4. Experimental study
Fig.3 shows the started at test facility with maximal heat removal of 3.5 MW. The
facility has all main parts of a real PAHRSEC:
Start-up Tank

Heat Exchanger

Ejector-Condenser
(EC)

Start-up Valve

Supply Tank

(Steam Generator - SG)

FIG. 3. PAHRSEC Test Facility and Ejector-Condenser of Medium Power.

3

Supply tank. Its total height is 5.3 m and its volume is 0.5 m . There is a system of
steam supply and removal of steam and liquid (not shown on Fig.3) which allows
to simulate different emergrncy situations at real units (for example, in SG). Water
level, pressure, and temperature are mesured in the tank. The maximal pressure and
temperature in it can reach 9 MPa and 315 C respectively.
Ejector-Condenser. Its photograf is presented on Fig.3. That EC allows to remove
heat from a heat source with power about 3.5-4 MW.
Heat Exchanger (HE). It is 5 m high and 0.53 m in diameter. There are two pipe coils
in the vessel, each is 80 m long. The total heat echange area is 21 m .
Check valve. It is a typical one, located between EC and Supply tank.
Start-up valve. An air-driven valve, installed hi the start-up line. Its actuation time
is 0.7 - 0.9 s.
Start-up tank. It is 1.5 m high and 0.426 m in diameter.
Data acquisition and processing system. It allows to monitor non-stationary slow
and quick processes (the frequency of channel measurements is 2 Hz and 2 KHz for
quick processes). It uses conventional primary transducers typical for thermal-physical experiments, and made on base of IBM PC AT. (A hardware of "CAMAC"
standard is used as a communication means with the object)
Some experimental results are presented on Fig.4-6.
201

dP,MP

P,MP

6-

2.1

1
4-

2-

^^
***-J ====

FT
1 [

1.4

-^^

-^ -^
0.7

3
0()

0.0

200

400

6C)0

tiec

1 - Steam Generator Pressure (Pig)

2 - Coolant pressure at tiie EC inlet (Pc)
3 - Pressure difference across EC (dP)

P.MPi

200

400

600

tc

1 - Pressure at the active (steam) nozzle outlet (Pn)

2-Pressureattiie EC outlet (PP)
3-Mixing Chamber Pressure (Pmch)

FIG. 4. Start-Up of EC and Determination of The Maximal Attainable Pressure

Difference In PAHRSEC Circuit.
202

dP.MP

P.MPm

6.0

0.1

1 - Steam Generator Pressure (Pv)

2 - Coolant pressure at the EC islet (Pe)
3 - Pressure difference across EC (dP)

P.MPm

500

1000 1500

te

1-Steam Generator Pressure (Pig)

2 - Mi-ring Chamber Pressure (Pmeh)
3 - Pressure at the active (steam) nozzle outlet (Pn)

FIG. 5. PAHRSEC Operation under "Rapid" Pressure Change in SG.

203

dP.MPi

P.MPi

0.1

0.0

1 - Steam Generator Pressure (P)

2 - Coolant pressure at the EC inlet (Pe)
3 - Pressure difference across EC (dP)

P.MP

6.0 -

4 *> -

r^
\/ ,A-

~ .<J

\
3.0 -

\
2

1.5 -

^^

~^~~^

~~T-^ ==5^

0.0 i
0
5(JO

1000 15 00

tiec

(P*)
(p4)

1-St earn Generator Pressure

2-M ixiog Chamber Pressure
3-Pr essure at the active (stet nn) nozzle outlet(Pn)

FIG.
204

6. PAHRSEC Operation under "Slow" Pressure Change in SC.

Typical experimental curves of PAHRSEC start-up and of maximal EC pressure

difference definition are presented on Fig.4. After the start-up valve opens the pressures
Pp, Pn and Pmch drop. One can say that EC comes into operation if the pressures Pn and
Pmch correspond to the saturation pressure at the mixing chamber temperature. One can
see that EC start time is very short (3-5 s.). The start-up of the whole System occurs later
and depends on the start-up tank volume. After the tank 6 fills out, the pressure behind
EC becomes higher then that in the SG and this opens the valve 4. This is the way
PAHRSEC starts up. It should be noticed again thatheat removal from SG begins just
after the valve 5 opening and it is most intensive at this moment.
During one experiment, when time was about 200 s, 400 s, and 600 s (Fig.4) the
hydraulic resistance behind EC was suddenly increased. Despite it, the System worked
in a stable way and MC pressure did not change. The maximal EC pressure difference
was higher than 1.5 MPa.
The PAHRSEC parameters behaviour while step-by-step SG pressure decrease is
shown on Fig.5. In this case the System worked stabUy in the range 1-7 MPa.
During the first 400 s of another experiment (Fig.6) the SG presure was being
increased from 4.6 MPa to 6.8 MPa and then it was being decreased in a monotone way
to 2 MPa. The System again worked stabily.
5. Installation of PAHRSEC at the 3rd and the 4th blocks of

Novovoronezh NPP
VNIIAES (RINPO), EREC, GIDROPRESS and AEP Institutes have developed the
technical requirements for PAHRSEC designe for WWER-440 NPP, where it was shown
that:
the System is compact and can be easily integrated with existing equipment
it uses common technology: HE, piping, valves etc. are standard ones (the "steamliquid" scheme has been chosen just from these reasonings)
its cost is much lower relative to other safety systems of the same purpose
The calculations of "black-out" accidents for WWER-440 have been conducted using
the code TRAC-PF1/MOD2. They showed that one PAHRSEC was sufficient to remove
heat from one SG. The perameters behaviour when two PAHRSECs are put in to operation
is shown on Fig.7. In about 2.8 hours (10 s) the system reaches a quasi-stationary regime
when power of residual heat generation is equal to PAHRSEC heat removal power.
6. Full-scale 10 MW PAHRSEC
The full-scale experimental set (Fig.8) with N = 10 MW is intended for

testing EC for WWER-440 PAHRSEC

study of PAHRSEC behaviour under different accident conditions
The set design takes into account the requirements to meet while introducing at acting

WWER-440, and the experience of running 3.5 MW PAHRSEC.

205

Pressuf zer pressure Woter level

2000

4OOO

ftOOO

BOOO

Moss flow in loops

0
2OOO
*OQ0
ftOOO
BOOO
IOOOO
1 - SC Ihoul PHRS 2 - SC oiUi PHRS
3 - hoi tint

IOOOO

0
1 -

2OOO

Pressure n the SG 3
f

MPe

2000

*OOO

6000

OOOO

1 - SC wilful PHRS 2 - SC *l*

IOOOO

PMRS

. 7. WWER-440, "Black-Out", PAHRSEC-10 MW.

Start-up tank

">ch

Steam generator

SV - Start-up valve
CV - Check valve

Heat-exchanger

PAHRSEC-10 MW

FIG. 8. PAHRSEC Test Facility of 10 MW.

206

4000

SC w thout PMRS

60OO
2 -

6000

IOOOO

SC !|h PHRS

Researches planned:
start up of the System under rated SG parameters

start up under off-design SG parameters

repeat start-up
operation under different disturbances
determining the stable operation range
heat removal under low (less than 1 MPa) SG pressure
experimental and simulated founding serviceability of PAHRSEC in different

accident conditions (black-out, SG's tube rupture, LOCA, etc.)

Fig.9 shows the 10 MW EC and Heat Exchanger. The whole 10 MW PAHRSEC is
nearly ready, and its putting in service is going to be soon.

Fig 9 Ejector-Condenser and Heat-Exchanger for the Test

Facility of 10 MW PAHRSEC
207

7. Future development of PAHRSEC

Fig. 10 shows the scheme with gravitational start-up system. For that kind of EC a
cylindric MC is needed. Fig. 10 shows a general view of the EC. Its tests have been made
and basic serviceability has been demonstrated.
PAHRSEC-2 scheme (Fig.l 1) has been proposed which has
two joints with SG istead of three in "steam-liquid" scheme (Fig.2)
stable operation under low SG pressure

PAHRSEC with Gravitational Start-Up System

(W. Reinsch)

EC with Cylindric Mixing Chamber

FIG. 10.

208

FIG. 11.

PAHRSEC -2

8. PAHRSEC scheme for containment

Operation principle of the scheme on Fig. 12 is the same as that for SG.
That scheme can work in two regimes:
supply of water from the supply tank to the containment and decrease of pressure in
it.
heat removal and water circulation throughout the circuit: containment - air cooled
heat exchanger - supply of "cold" water to EC - injection hi containment
Fig. 12 shows also EC tested under low steam pressures 0.1 < P < 0.5 MPa. It should
be mentioned that a system for water supply to containment by means of EC has been
also studied in [2].

209

kJ1

1./
i/
/
i

c
1

1C

/
/

:
:

1
/

/
J<
1
,
f1
1

Start-Up
valve

- D*3'
Air cooled /
HE
M

LJ

Supply
tank

i
(a) Scheme of containment PAHRSEC for severe accidents

(b) EC designed for operation with low pressure steam

FIG. 12.

1. G.Cattadore, L.Galbiati, L.Mazzocchi, P.Vanini, "A Single-Stage High-Pressure

Steam Injector for Next Generation Reactors: Test Results and Analysis", European

Two-Phase Flow Group Meeting. University of Hannover, Germany, 7 to 10 June,

1993.
2. T.Narabayashi and others, "Thermohydraulics Study of Steam Injector for Next
Generation Reactor", Int.Conference on "New Trends in Nuclear System Thermohydraulics", May 30th - June 2nd, 1994 - Pisa, Italy, v.l, pp 653-661.

3. F.L.Carpentino and others, "The Safe Integral Reactor (SIR)", ABB Preprint.
210

AN EXPERIMENTAL STUDY ON THE BEHAVIOUR

OF A PASSIVE CONTAINMENT COOLING SYSTEM
USING A SMALL SCALE MODEL

D. SAHA, N.K. MAHESHWARI, O.K. CHANDRAKER,

V. VENKAT RAJ, A. KAKODKAR
Bhabha Atomic Research Centre
Bombay, India
Abstract

A Passive Containment Cooling System (PCCS) has been

proposed for the Advanced Heavy Water Reactor (AHWR) being
designed in India. The function of the system is the long term
cooling of the reactor containment following Loss of Coolant
Accident.
The
system
removes energy released into the
containment through Isolation Condensers immersed in a pool of
water. An important aspect of 1C working is the potential
degradation of 1C performance due to
the
presence
of
noncondensables. Tests have been conducted on a small scale
model of PCCS with constant steam flow rate and steam flow rate
simulating decay heat curve using air as noncondensable. The
effect of variation of different parameters like the partial
pressure of noncondensable has also been studied. The studies
carried out within a limited range of values of different
parameters have confirmed the efficacy of the separation of
noncondensables
and removal of energy released into the
containment. This paper deals with the details of the test set
up, the test procedure and results.
1.0

INTRODUCTION

Increasing awareness towards safety and the experience

gained in the past have led to the incorporation of a number of
passive safety systems in the new generation of nuclear
reactors being designed. These systems depend on the natural
laws of gravity, thermal-hydraulics and physics and do not
require intervention of operators or use of externally actuated
electrical or mechanical devices. One such system envisaged for
long term containment cooling of advanced reactors following
Loss of Coolant Accident (LOCA) is the Passive Containment
Cooling System (PCCS). The purpose of the PCCS is to limit the

containment pressure to a value below a predetermined level and

to achieve a walk-away period without operator action. An
analytical evaluation of different PCCS concepts indicates that
the PCCS incorporating Isolation Condenser (1C) may be the best
option [1].
PCCS with 1C has been adopted for the Simplified
Boiling Water Reactor (SBWR) [2], It is proposed to incorporate
a PCCS with 1C in the Advanced Heavy Water Reactor (AHWR) being
designed in India. An important aspect of 1C functioning is the
potential degradation of heat removal capability of 1C due to
the presence of noncondensables. Experimental data available
on the performance of PCCS [3] is very meagre. In view of this
an
experimental
programme to study and understand PCCS
behaviour and performance is being carried out in phased
manner. In the first phase experiments have been conducted on a
211

small scale model to confirm the working principles of PCCS and

to understand the various phenomena involved.
2.0

DESCRIPTION OF PCCS

A simplified diagram of the PCCS is shown in fig.l. The

isolation condenser comprises of a large number of vertical
tubes connected to horizontal cylindrical headers at the top
and bottom. The 1C is immersed in a large pool of water located
at a high elevation as shown in the figure.

Fig.1 PASSIVE CONTAINMENT COOLING SYSTEM WITH

ISOLATION CONDENSER

Steam-noncondensable gas mixture enters the 1C from volume

Vi (drywell) of primary containment following LOCA through a
line which has no valve. Steam is condensed in 1C and
condensate flows by gravity to the Gravity Driven Water Pool

(GDWP) from the bottom header of the 1C. The noncondensables

are led to the vapour suppression pool in volume V2 (wetwell)
through a vent line submerged in water. Due to inflow of
noncondensables into volume 72, the pressure of V2 increases.
When the pressure in V2 exceeds the pressure in VI by a preset
value the vacuum breaker opens and noncondensables return to
VI. When the differential pressure between V2 and VI reduces to
a ..preset value the vacuum breaker closes. The continued
accumulation of noncondensables in 1C causes degradation of the
performance of 1C causing pressure rise in VI. This may again
cause the flow of noncondensables to V2,depending on the
conditions. The PCCS is always available for containment heat
removal. The differential pressure between the volumes VI and

V2 initially provides the driving head for the steam-gas

mixture flow through the 1C. The heat removal capability of
PCCS is affected mainly by the flow path pressure loss,
noncondensables inside the containment and heat
transfer
coefficients in the pool and 1C.
212

3.0

TEST SET UP

Tests
to study system response behaviour have been
conducted on a small scale model of the PCCS (see fig.2), the
volume scaling of the set up being approximately 1:3000.
Elevation and hydraulic resistances could not be simulated in
this small scale model. The configuration of 1C has been
simplified as shown in fig.2. The upflow tube is thermally
insulated. The downflow tube inner diameter is 10 mm and length

300 mm. Instead of natural circulation, a small forced flow of

water is maintained in the secondary side of 1C. Air has been
used as the noncondensable gas during the tests. The steam air
mixture from volume VI flows to 1C steam box or upper header.
The noncondensable gas vent line runs from 1C water box or

SV : SOLENOID VALVE
DPT: INFERENTIAL PRESSURE

TRANSMITTER
LT : LEVEL TRANSMITTER

: THERMOCOUPLE

: SUBMERGENCE DEPTH

COOLING WATER
PT)

CONOENSATE
COLLECTION
PLENUM

VENT TUBE
P

VAPOUR SUPPRESSION
POOL

STEAM

FROM BOtER

Fig. 2 EXPERIMENTAL SET-UP

lower header to suppression pool. The condensate from the 1C

water box flows to a condensate collection plenum which is
vented at top to the gas space of the 1C water box. The level
of water in the plenum is maintained at an almost constant
value using a valve in the drain line from the condensate
plenum. Air is injected into volume VI from a compressor and
steam is introduced from a boiler.
The volume V2 is connected to the 1C through the gas vent
line. The vacuum breaker between the volumes VI & V2 has been

simulated by a solenoid valve and a transmitter provided for

differential pressure measurement between the two volumes.
Whenever the pressure of V2 exceeds the pressure of volume VI
by a specified amount, the solenoid valve opens and air flows
back from volume V2 to VI. When the differential pressure
reduces to a set value, the solenoid valve closes.
213

The pressures of volume VI, V2 & 1C water box are measured

by pressure transmitters. The level in condensate collection
plenum is measured by a level transmitter. The temperature of
volume
VI is measured at three different locations by
thermocouples installed inside the vessel. Air partial pressure
Pa in VI is calculated as the difference between the total
pressure Pt and steam saturation pressure. The flow of steam
into the volume VI is estimated
approximately by
the
condensation rate of steam in the 1C and condensation in VI due
to heat loss. Variation in steam flow rate into volume VI is
estimated by measuring pressure drop across a restriction in
steam inlet line.
4.0

TEST CONDITION

Tests have been conducted under the following conditions,

i) By maintaining a constant steam flow into volume VI
ii) By varying steam flow rate into volume VI as per decay
heat curve. Variation of steam flow with time is
depicted in Fig. 3
f.O;

0.65
20000

5000

25000

THE (SCO

Fig. 3 VARIATION OF FLOW (F) WITH TIME

In

both

the

above

cases,

initially volume V2 was

maintained at atmospheric pressure. Steam flow rate into volume

VI, air content and submergence depth of vent tube
in
suppression pool were varied during the tests.
5.0

TEST PROCEDURE

Volume VI is initially isolated from 1C and volume V2. Air

was purged from volume VI through exhaust line by supplying
steam from boiler for a sufficiently long time. During this
214

process, the VI vessel also got heated up. After air was
removed, the exhaust line was closed. The volume VI was
pressurised to initial air partial pressure by introducing air.
Air line was then closed. Steam was then supplied to the volume
VI to achieve desired initial total pressure. The 1C and volume
V2 were maintained at cold atmospheric condition. Tests were
initiated by opening the valve between volume VI and 1C. The
valve in the steam inlet line was also opened simultaneously.
Variable steam flow rate into VI was achieved by operating the
valve shown in figure 2. Heat loss from volume VI was
ascertained by maintaining constant pressure in this volume
with through flow of steam and measuring the amount of
condensate over a period of time.
6.0

RESULTS AND DISCUSSIONS

With constant steam input to volume VI, a number of

experiments were conducted for different values of parameters
of interest. Results of two such test runs are depicted in
Figs. 4a and 4b. Fig. 4a represents a case with very low steam
flow rate. In this case, because of initial high condensation
rate in 1C, VI pressure first decreases and then increases with
reduction in condensation rate in 1C. Since the 1C pressure
exceeds the V2 pressure by an amount more than the submergence
depth of vent tube, the vent tube water is driven out and air
enters the volume V2. This causes the V2 pressure to rise. At
about 3200 seconds water again enters vent tube and flow of air
into V2 stops. Beyond this time no change in V2 pressure was
observed. VI pressure also remains almost unchanged since the
energy removal rate of 1C matches with the energy input rate.
The tests were continued for about 10,000 seconds. Upto this
time, no further change in the pressures of VI or V2 was
observed. Since VI pressure was always higher than V2 pressure,
during the test, the vacuum breaker simulator did not operate
and there was no return of air from V2 to VI. 1C water box
pressure is also plotted in Fig. 4a.
is q
VI PRESSURE
VI PRESSURE

u-

-1C

PRESSURE

SUBMERGENCE DEPTH 300 MM

P/Pt
STEAM FlOW

, MJ
t17 Kg/Sec

10
V1 PRESSURE

V2 PRESSURE
1C PRESSURE
SUBMERGENCE DEPTH . 5 MM
P/Pt
> M45

100.0-

STEAM ROW

I ' ' ' Ir

1000
3200

I'

woo

8000

MOO

11200

THE tSEO

Fig. 4a PRESSURE TRANSIENTS WITH

CONSTANT STEAM FLOW RATE

5000

WOW

, O.M24 Kj/Stc

ttOOO

20000 25000

30000

TWE CSEQ

Fig. 4b PRESSURE TRANSIENTS WITH

CONSTANT STEAM FLOW RATE

215

Fig. 4b depicts the results of another experiment with

different values of parameters of interest.. Because of higher
steam flow rate the initial sharp decline of VI pressure was
not observed in this case. However, the subsequent trends of
the pressure curves were observed to be the same as in the
earlier case. Though the tests were conducted for a longer
period (25,000 seconds) no decrease in VI pressure was observed
which is necessary for vacuum breaker simulator to operate.
This indicates that there was no significant variation of 1C
heat removal rate during the test period.

Figs. 5a to 5e depict pressure transients in VI and V2 for

variable steam flow rate. It may be observed from fig. 5a that
the pressure in VI dropped below the pressure in V2 and the
vacuum breaker opened (indicated by arrow in figure) at around
17,000 sees, and closed subsequently. The effect of increase in
the initial amount of air in volume VI can be ascertained by
comparing figure 5b with 5a. In the case shown in 5b the vacuum
breaker opened at around 15,000 seconds. However, pressures in
both the volumes were found to be higher for higher initial air
content. Due to limited periods over which the tests were
conducted, subsequent vacuum breaker opening could not be
obtained. Comparison of fig. 5a with fig. 5c reveals the effect
of higher steam flow rate. From fig. 5c if can be observed that
due to higher steam flow, pressure in VI has increased
considerably, but the pressure in V2 remains almost unaltered.
Dpto 20,000 sec vacuum breaker opening was not encountered. The
effect of change in submergence depth is depicted in - figs 5d
and 5e. Increase in submergence depth led to an increase in VI
pressure. However, no significant change of V2 pressure was
observed. For cases depicted in 5d and 5e vacuum breaker
opening did not take place during the limited period for which
the tests were carried out.
2.0

SUBMERGENCE DEPTH = 400.0 MM

Pa/?t
=0.094

VI PRESSURE

V2 PRESSURE

VI PRESSURE

SUBMERGENCE DEPTH * 400-0 MM

P/Pt
0,148
IMTUU. R.OW. F.
3 1.0032 K}/S*c

.15-

------V2 PRESSURE

tu

10vt
vt

IU

8,5-

0.0

WOO

I*
12000
16000
TM (SEQ

OiO

20000

24000

fig. 5a PRESSURE TRANSIENTS WITH

VARIABLE STEAM FLOW RATE

216

o " " iioo' "sobo'''' 12000''" iwoo

TMEISEO

Rg.5b PRESSURE TRANSIENTS WITH

VARIABLE STEAM ROW RATE

20000

24000

2.0 VI PRESSURE

V2 PRESSURE
SUWCKZNCE DEPTH . 6M. MM
P/Pt
> O.H2
MTIAL aOW, f)
. IM2I K/S

is-

U SUBMERGENCE DEPTH 4004 MM

P/Pt
* p.DH
MTUl ROW. F:
Hint

15-

M ii 111 MI ii i a ii [ i in m 11 ii ii ii | m 1 1 1 1 1 1 1 ii i ii

0.0
0

4000

1000

3000
KOOO
THE (SEQ

24000

20000

4MI

MM

I20M

ttO0

2WW

24000

THE ISEO

Fjg.SC PRESSURE TRANSIENTS WITH

VARIABLE STEAM FLOW RATE

Rg.Sd PRESSURE TRANSIENTS WITH

VARIABLE STEAM INPUT

2.0
VI PRESSURE

V2 PRESSURE
SUBTGRGENCE DEPTH . 4M.C HM

Pi^t
MTUL FLOW. F,

15-

> UK
. U021 kj/lK

J.5-

0.0

4M

KM"" izii""uio""2o'ooo""2iio
TK ISEO

Fig.Se PRESSURE TRANSIENTS WITH

VARIABLE STEAM INPUT

7.0

CONCLUSIONS

1.

The tests conducted over the limited range of values of

different parameters have confirmed the efficacy of PCCS in
separating
the noncondensables and removal of energy
released into the containment.

2.

During experiments with constant steam flow rate, pressures

in volumes VI and V2 had attained almost a steady value
after initial transients. Vacuum breaker did not open in
this case since VI pressure was always higher than V2
pressure.

3.

During
experiments with
reducing
steam
flow rate
(simulating decay heat curve) VI pressure dropped below the
pressure of volume V2 causing vacuum breaker to open. The
reduction of VI pressure can be attributed to the decrease
in steam flow rate and improvement of 1C performance.
217

ACKNOWLEDGEMENT
,Shri

The authors gratefully acknowledge the help rendered by

M.L. Dhawan of RED, BARC. Thanks are also due to

Instrumentation Section of RED.

NOMENCLATURE

: Flow

Fi
h

: Initial Flow
: Submergence depth

Pa
Ft

: Air partial pressure

: Total pressure

REFERENCES

1.

Hirohide Oikawa
et al., "Heat Removal Performance
Evaluation of Several Passive Containment Cooling Systems
During Loss of Coolant Accident" J. of Nucl. Science &
Technology, V 28, N10, Oct., 1991.

2.

M. Brandani et al. "SBWR - Isolation Condenser and Passive

Containment Cooling: An Approach to Passive
Safety."
Proceedings of IAEA TCM, IAEA-TECDOC-677, Rome, Sept. 1991.

3.

Seliechi, Yokobori
et.al.
Isolation Condenser Applied
Cooling

System."

The

"System Response Test, of

as a Passive
Containment

1st JSME/ASME

Joint International

Conference on Nuclear Engineering, Tokyo, November, 1991.

218

RESULTS OF SAFETY-RELATED COMPONENTS/SYSTEMS TESTS

(Session IV)
Part 2
Chairman
E.F. fflCKEN
Germany

STEAM INJECTOR DEVELOPMENT FOR

ALWR's APPLICATION

G. CATTADORI
SIET- Piacenza

L. GALBIATI, L. MAZZOCCHI
CISE Technologic Innovative - Segrate, Milan
P. VANINI, V. CAVICCHIA
R&D Department, Nuclear division,
ENEL, Rome
Italy
Abstract
Steam Injectors (Si's) can be used in Advanced Light Water Reactors (ALWR's) for high
pressure makeup water supply; this solution seems to be very attractive because of the "passive"
features of Si's, that would take advantage of the available energy from steam without introduction
of any rotating machinery. In particular. Si's could be used for high pressure safety injection in
BWR's or for emergency feedwater in the secondary side of evolutionary PWR's.
An instrumented Steam Injector (SI) prototype, operating at high pressures, has been built and

tested. The experimental results confirm the capability of tested SI to operate at constant inlet water
pressure (about 0.3 MPa) and inlet water temperature up to 50C, with steam pressure ranging from
2.5 to 9 MPa (4.5-9 MPa at maximum inlet water temperature). The discharge pressure target (10%

higher than steam pressure) was fulfilled in the operating range. It should be noted that the minimum
operating limit can be lowered to 1.5 MPa with some modifications. To achieve these results an
original double-overflow flowrate-control/startup system, patented by ENEL/CISE in 1993, has been
used.
1.

INTRODUCTION

A Steam Injector (SI) is a device without moving parts, in which steam is used as the energy
source to pump cold water from a pressure lower than the steam to a pressure higher than the steam.
In fact, a large amount of heat, available from steam condensation, can be partly convened into
mechanical work useful for pumping the liquid. In this respect, SI can be regarded as equivalent to
different devices, like turbine-driven pumps, where steam thermal energy is used to pressurize a
liquid; in comparison to this kind of equipment, the main difference is that in SI all thermodynamic
processes rely on direct contact transport phenomena (mass, momentum and heat transfer) between
fluids, not requiring any moving mechanism.
The SI can be divided into the following regions:
a)
steam nozzle, producing a nearly isentropic expansion and partially converting steam enthalpy
into kinetic energy; it can be a sonic nozzle or, if a stronger expansion is required, a
supersonic nozzle with the typical converging-diverging shape;
b)
water nozzle, producing a moderate acceleration and distributing the liquid all around the
steam nozzle outlet;
c)
mixing section, where steam and water come into contact. Steam transfers to water heat
(because of temperature difference), mass (because of the related condensation) and
momentum (because of velocity difference). The final result is the complete steam
condensation, with an outflowing subcooled liquid at relatively high pressure. The shape of
the mixing section is usually converging, for reasons reported in [1];

221

d)

diffuser, where the liquid kinetic energy at mixing section outlet is partially recovered
producing a further pressure rise.
It should be noted that the arrangement of steam and water nozzles could be inverted, creating
a circular water nozzle and an annular steam nozzle as reported in [2]. In another type of
arrangement. [3]. the annular water jet is accelerated both by central and outer steam jet. However,
is authors' opinion that the central steam jet/outer liquid configuration should be more convenient
because it minimizes the perimeter of the steam nozzle and avoids the direct contact between steam
and mixing section walls: in this way viscous dissipations should be reduced.
Si's were used as feedwater supply devices in locomotives and in the Merchant Marine since
World War II and are manufactured by a few companies for applications in food and paper industry.
In nuclear field, several systems based on high-pressure steam-injectors were proposed in [1], [4].
[5] and [6].
It should be noted that, with respect to nuclear systems, the attractiveness of Si's is quite
evident, because a high pressure water supply can be essential to very important emergency functions
iiike emergency core cooling, feedwater supply for decay heat removal and so on) and usually a steam
supply is easily available in power plants. Moreover, SI can be regarded to a great extent as a passive
system, as it does not require any external energy supply or moving mechanical part.
However, commercially available Si's operate at "low" pressure (< 2 MPa). Moreover, no
single-stage high-pressure steam injector is available from previous research programs undertaken in
[4] and [5] (for application to a BWR). In [6], a high pressure SI has been developed for application
to CANDU reactors but the maximum discharge pressure was limited at about 4 MPa. In [6] a singlestage high-pressure steam injector has been developed for application to the secondary side of a
VVER-440 but the inlet liquid pressure was, basically, at the same pressure of the steam generator
so that the pressure gain requested to the component was very limited.
In 1991, ENEL, the Italian Electricity Generating Board, which is engaged in a wide range

of activities concerning ALWR's. decided to evaluate the applicability of a single-stage SI for a

ALWR's application: so, CISE was charged to perform an SI development project in cooperation with
ENEL itself. The results of this project can be found in [7]: tested SI operated with steam pressure
ranging from 2.5 to 8.5 MPa. always fulfilling the discharge pressure target (10% higher than steam
pressure), with inlet water temperature up to 37C. However, in order to improve the SI performance
(reducing, in particular, the steam consumption, see ch. 3.3) an additional optimization study on the
injector design, was undertaken by CISE in cooperation with ENEL. The corresponding experimental
activity was performed at SIET laboratories in 1994. Results are reported below.
2.

EXPERIMENTAL SETUP

The test section used in the experimental activity was made of stainless steel (figure 1). As
can be seen, on the basis of the results obtained in [7], two overflow openings (primary and secondary
overflows) where realized on the mixing chamber in order to allow SI start-up and discharge flowrate
control. The SI was tested at SIET laboratories where a steam-water test facility was designed and
built. Figure 2 shows a schematic diagram of the experimental plant which supplies metered flows
of steam and water to the test section and which includes equipment for controlling and measuring
the thermalhydraulic parameters. An open loop including pumps, heat exchangers and a pressurizer
(volume = 1 m3) provided demineralized water to the SI at the required conditions; water flowrates
up to 25 kg/s were available, with fluid pressure and subcooling controlled at the test section inlet.
A line coming from an external boiler provided superheated steam up to 5 kg/s flowrate and 9 MPa
pressure; liquid injection was used to control the steam temperature at the test section inlet. Opensystem operation was straightforward, with the total discharge directed to condenser. The test section
discharge pressure was varied by means a control valve. The following measurements were taken
during the tests:
D inlet liquid and steam flowrates;
G inlet and outlet fluid pressures and temperatures:
LJ axial pressure profiles in the different SI components: steam nozzle, water nozzle, mixing
section, diffuser (totally 60 measurement stations).
G primary and secondary overflow temperatures and flowrates.
222

LIQUID IMET

1 STEAM SOZZU
2 LKJL1D RESERVOIR
5 MIXING SECTION

UQUID OUTLET

Figure 1 - Test section schematic

- WCH ntSSUR fUMK

ELECTUCAL BOILEX

LOW rtESSUKE VALVIS

Figure 2 - Schematic of
the test facility
223

Flowrates were measured by means calibrated orifice plates (accuracy = 2% of reading).

Temperatures and pressures were measured by K-type thermocouples (accuracy = 1C) and by thin
film-type pressure transmitters (accuracy = 0.1% FSl. respectively.
3.

TEST RESULTS AND DISCUSSION

3.1 Independent and dependent variables

There were four independent variables to be controlled or selected for each test run. These
four can be described as:
3 steam supply pressure:

3 steam superheat;
D water supply pressure:
3 water supply temperature.
To these four indipendent variables, concerning fluid conditions, it should be added:
3 backpressure valve setting;
3 OF2 valve setting (OF1 valve could be always closed after startup).
On the other hand, several dependent variables (for any given set of indipendent variables) were
selected and measured.
They were:
D water supply flow rate:
D water/steam flow rate ratio (w):
CD steam nozzle pressures;
D mixing section pressures;
D primary and secondary overflow pressures/flowrates;
3 diffuser pressures.
3 discharge liquid pressure:
3 discharge liquid temperature.
The range of the independent variables explored in the test program is shown in table I.
As can be noticed, the tests were conducted in a wide range of steam pressure (2.5 MPa - 9 MPa).
Steam superheat was varied (between 1-16C). The inlet water pressure to the test section was kept
constant at about 300 kPa. The inlet water temperature was varied from 15C to 50C. The reason
for this was to check the injector expected performance degradation at increased water temperature.
Main dependent parameter variations are summarized in table II.

table I - RANGE OF INDIPENDENT VARIABLES

224

Steam supply pressure

2.5 - 8.9 MPa

Steam superheat

1 - 16 C
(max. steam quality =
1.06)

Water supply pressure

300 kPa

Supply water temperature

15 - 50 C

table II - RANGE OF DEPENDENT VARIABLES

Inlet liquid flowrate

16-19kg/s

Liquid-to-steam flowrate ratio (w)

3-12

Discharge liquid pressure

2.8 - 9.8 MPa

Discharge liquid temperature

78 - 197 C

Primary overflow pressure

77 - 476 kPa

Secondary overflow pressure

247 - 2994 kPa

Discharge flowrate

3.5 - 14.5 kg/s

3.2 Discharge pressure

The injector discharge pressure for the entire test series is summarized in figure 3. This figure
shows the pressure at the diffuser outlet as a function of the injector inlet steam pressure. In figure
4, the axial pressure profile at maximum inlet steam pressure is shown (the profile is similar at lower
supply steam pressures). It should be remembered that discharge pressure was fixed at an
approximately constant value (10% higher than inlet steam pressure), in order to meet the
specifications. Tested one-stage SI developed sufficient discharge pressure except when steam supply
pressure was less than about 2.5 MPa. It should be noted that maximum inlet steam pressure could
not be increased over 9 MPa because of limitations of the experimental facility. Actually, the
component can develop sufficient discharge pressure even at higher inlet steam pressures (provided
that steam condensation is complete).
1OOOO-

I
/

i
!

""""

Figure 3 - SI discharge pressure:

<g

experimental results

g:

5OOO-

4000-

IU

en
o

6000

200O
1000
n

1
1
i
i
i
!/ \
/i

J1

1t/
/

C
^^X*

t^
1
S\
I

1000 2OOO 3000 4000 5000 6000 7000 8000 9OOO 10000
INLET STEAM PRESSURE (kPa)

STEAM

Figure 4 - Experimental pressure

profile along the mixing section

225

3.3 Discharge Flo>vrate

Discharge flowrate is shown in Fig 5 at different inlet liquid temperatures It can be noted
that discharge flowrate decreases when inlet liquid temperature increases

16-

Figure 5 - Discharge Flowrate at

different inlet liquid temperatures,

14
& 12

OF2 optimal setting

D T1=27'C. D T1=42'C TI=50*C

10
_l

u_

HI

g
<n

2000

3000

4000

5000

6000

7000

8000

9000

INLET STEAM PRESSURE (kPa)

3.4 Steam Consumption

Steam consumption is an important parameter in order to evaluate the performance of an

injector In particular specific steam consumption (Steam Rate, SR) can be defined as

where Fc and FL are. respectively. the steam and liquid flowrates entering the SI and rOF2 is the liquid
flowrate leaving the SI from secondary overflow
In figure 6 SR values at different inlet water temperatures and steam pressure are shown in
case of optimized OF2 setting (Appendix 1, test series I, optimized data) It can be seen that SR
increases with inlet steam pressure and inlet water temperature (test series II)

Figure 6 - Steam Rate (SR) at different

temperatures, OF2 optimal setting

Tl. 27-C
Tl .42-C

Tl . SO'C

INLET

226

STEAM

PRESSURE (UPa)

3.5 SI Characteristic

Figure 7 shows a typical SI characteristic (obtained using the procedure reported in App. I.
test series III). The inlet steam pressure is 5 MPa and the inlet water pressure is 0.3 MPa. The inlet
water temperature is 15C. The characteristic is similar to that for a centrifugal pump: the discharge
pressure drops as the discharge flow rate becomes larger and larger.
It should be noted that, in case of "low"-pressure injector, without any overflow, the injector
performs as a positive displacement pump so that a constant flowrate is discharged against a range
of backpressures, as noted in [8]. So, an SI can be regarded as a positive displacement pump at "low"
pressures (where high mixing section area contraction ratios are not required and overflow is not
necessary) and a centrifugal pump at "high" pressures (where secondary overflow spillage is needed).

Figure 7 - SI characteristic
(PLO=300 kPa, TL=15C, PGO=5MPa)
UI

DISCHARGE FLOWRATE (kg/s)

3.6 Liquid-to-Steam Flowrate Ratio

The liquid-to-steam flowrate ratio (w) was ranging between 3 and 12. w is a significant
variable because the principle of operation of the SI relies on the steam being condensed by the water
in the mixing section and producing vacuum. If the water/steam flow rate ratio is too low, the steam
will not be completely condensed in the mixing section. On the other hand, according to available
theories, the maximum pressure which can be developed by the injector decreases as the water/steam
ratio increases. Thus for the best pressure performance, the injector should be operated at low
water/steam ratio but not so low that the steam will not condense completely. The water/steam flow
ratio for a test could be varied by changing the steam supply pressure to the injector: the lower limit,
for which condensation can be complete, depending also on backpressure setting, was not found. In
fact, as reported above, because of design limitation of the experimental facility, the inlet steam
pressure could not be increased over 9 MPa.
3.7 Flowrate Spillage

The tested SI spills some of the water being pumped. In figure 8 spillage at constant inlet
liquid pressure (300 kPa) and different temperatures is shown: inlet steam pressure being costant,
spillage increases with temperature.
K

16-

Figure 8 - OF2 spillage: temperature effects

(PLO=300 kPa)
D TI=27*C: C TI=42*C; T1=50'C

SPILLED FLOWRATE (kg/s)

"

14-

"

12-

ID

S'

_,

64-

2n.

2000

3000

4000

5000

6000

7000

8000

9000

INLET STEAM PRESSURE (kPa)

227

3.8 Low Quality Tests

Several tests (Test series IV) were performed at low-quality values of the steam entering the
test section (down to x = 0.28-29). The SI operated in a very good way. The steam nozzle expansion
ratio decreased at low inlet steam quality. For this reason, the liquid suction in the mixing chamber
decreased. Nevertheless, being the two-phase flow mass entering the steam nozzle higher than the
mass of pure steam, the total mass entering the SI was found to be approximately constant at different
inlet steam quality values. On the other hand, it has been found that the secondary overflow spillage
tends to decrease at low steam quality, while discharge flowrate, as a consequence, has an opposue
trend.
As far as SR is concerned, this parameter (if it is referred to the pure steam entering the SI)
decreases at low quality tests in comparison to tests at x=l or x > l . However, it is possible to
introduce a further parameter. SRa, defined as the ratio between the steam nozzle outlet steam
flowrate (TCl) and the net flowrate pumped by the steam injector (TL - rOF2):

SRa -

- r 0h ~>.

The steam nozzle steam outlet flowrate can be derived assuming an isentropic expansion in the steam
nozzle and calculating the steam quality x, at the nozzle exit as:
x., = (s., - s")/(s'-s')

where s,, is the specific entropy at steam nozzle inlet section and s', s" represent the specific liquid
and steam entropies at steam nozzle outlet section, calculated at the pressure of the steam after the
expansion in the nozzle. When SRa is calculated (Fig. 9), this parameter seems to be only slightly
affected by any change of inlet steam quality (even if tends to decrease at low steam qualities). In
fact, it has been found that the mass of steam at the mixing section inlet shows just a slight decrease
at low inlet steam qualities. In conclusion, it can be said that, at low inlet steam quality, the injector
performance is only slightly affected by the higher quantity of liquid entering the steam nozzle: SI
performance depends on the actual steam flowrate after expansion and on the actual steam nozzle
expansion ratio (which keeps approximately constant the total mass flowrate entering the injector)
U.3-

0.45-

'

0.4-

Figure 9 - "Low" quality tests: modified

steam rate (SRa)

1
!

0.3n oc

0.2-

0.15-

i
i

1 -

.
"

;
i

0.10.05
n

0.2

0.4

0.6

0.8

INLET STEAM QUALITY

3.9 Modified Test Section

The main objective of the described tests was to minimize the steam consumption: indeed,
at inlet liquid temperature of 37C and at maximum inlet steam pressure (9 MPa), steam rate was
reduced about 5 times in respect to [7].
However, if a higher steam consumption is accepted, SI operating range can be extended. In
fact, using a modified test section (derived from the previous configuration by a reduction of the
liquid nozzle area), several tests were performed at constant liquid pressure (about 300 kPa) and
different inlet liquid temperatures (up to 42C). SI operating range was wider (1.5-9 MPa) but, at the
same time, steam rate was higher (about 2.2 at 9 MPa and TL=27C), figure 10.
228

2.5-

Figure 10 - Modified test section:

steam rate (SR) at different
temperatures

"5>
.X

O>

Ul

5 1-5D T1=27'C; T142*C

if =

1
i

cr

LU
1-

05

0.5n-

1000

. L
2000

j
3000

4000

5000

6000

7000

8000

9000

INLET STEAM PRESSURE (kPa)

4.

APPLICATION TO ALWR's

4.1 Application to BWR's

Si's can be used in BWR's in a high pressure coolant injection system. Using only reactor
dome steam and steam pressure to actuate OF2 valve (see after), the SI system would be capable of
drawing cold water from selected sources (e.g. the condensate storage tank) injecting this water into
the reactor.
An automatic double overflow flowrate control/startup system can been envisaged using two

valves, connected to OF1 and OF2 discharge lines. The OF1 line is provided for startup when the
injector is full of water. Steam entering the injector would initially purge standing water in the mixing
section and a steam/water mixture would flow through OF1 discharge line to the suppression pool for
several seconds until a sufficient suction flowrate is established to enable full condensation of the
steam in the mixing section and sufficient pressure is obtained to direct flow to the reactor vessel. As

steady state flow has been established, the relief valve in the OF1 line will close due to the drop in
the static pressure inside the mixing section, resulting from the acceleration of flow in the mixing

section itself.
As regards the secondary overflow, a valve controlled by steam pressure could be inserted
on OF2 line: in fact, optimal discharge flow through OF2 line depends on inlet steam pressure. This
valve would tend to open at low inlet steam pressure, allowing a suitable liquid flowrate discharge:
on the contrary, OF2 valve would be almost close at high inlet steam pressure.
An additional possibility would be to modulate the inlet liquid flowrate by means of a valve
controlled by inlet steam pressure: in this case, at constant discharged liquid flowrate, water
consumption would be reduced and operating range would be extended down to very low steam
pressures.
Comparing the experimental results with the performance requirements specified in [1] for

a 600 MWe BWR application, it can be derived that the present SI fully satisfy the above
requirements in terms of pressure, with a scaling factor about 1:4 in terms of discharge flowrates.
4.2 Application to PWR's

Si's can be applied to the secondary side of PWR's. In this case, steam from the steam
generators operating at a pressure ranging in the field of safety/relief valve opening set-point can be

used to run a SI supplying feedwater back to the steam generators. Cold water could be taken from
the demineralized water storage tank. As far as system automation is concerned, the same startup/flowrate control system described in the previous paragraph could be used. A simulation with
RELAP5/MOD2 has been performed, [9], to verify the expected performance of the SI taking into
account the various interactions between this component and the secondary side of a PWR. The
analysis has demonstrated the adequacy of SI in removing long term decay heat and preventing
excessive heatup of the reactor coolant system. In particular, simulation results show that, assuming
a 600 MWe PWR reactor (2 loops), the SI can operate till 12 days after scram (when its lower
operating limit, 2.5 MPa, is reached). It should be noted that the present SI is characterized by a
scaling factor about 1:1.5 (in terms of discharge flowrates) in respect to the full size component
(designed for one loop).
229

5.

CONCLUSIONS

The test results showed that CISE one-stage high-pressure SI can operate, at constant liquid
pressure (about 0.3 MPa), with steam pressure ranging from 2.5 to 9.0 MPa and supply water
temperature ranging from 15 to 50C, providing water to a discharge fixed pressure about 10 percent
higher than steam pressure.
It should be noted that:
D steam consumption is quite low and tends to increase with inlet liquid pressure and
temperature and with steam pressure;
D the tested high-pressure SI spills some of the water being pumped (at least at constant
supply water pressure). Spillage increases as the supply-water temperature increases;
D the high-pressure SI (due to secondary overflow spillage) can be regarded as a centrifugal
pump.
REFERENCES
[1]

[2]
[3]

CATTADORI. G.. GALBIATI, L.. MAZZOCCHI, L. & VANINI. P., Steam Injector Analysis and
Testing. European Two-Phase Flow Group Meeting, The Royal Institute of Technology, Stockholm.
Sweden (1992).
FITZSIMMONS, G. W. Simplified Boiling Water Reactor Program. Steam Injector System. Fir.al
Report, GE-NE, GE FR 00876 (1990).
NARABAYASHI. T., ISHIYAMA. T.. MIYANO, H., NEI, H. & SHIOIRI, A. Feasibility and

Application on Steam Injector for Next-Generation Reactor. 1st JSME/ASME Joint International
Conference on Nuclear Engineering (1991) p. 23-28.
[4]

[5]
(61
[7]

[8]
|9)

NARABAYASHI. T., IWAKL C,

NEI. I.. MIZUMACHI, W.. SHIOIRI. A..
Thermalhydraulics Study on Steam Injector for Next Generation Reactor. New Trends in Nuclear

Systems Thermohydraulics (1994) p. 653-661.

SUURMAN. S. Steam-Driven Injectors Act as Emergency Reactor Feedwater Supply, "Power". 3
(1986) 95.
SOPLENKOV. K. Passive Heat Removal System with Injector-Condenser (PHRS-Id.
Electrogorsk Research & Engineering Centre of Nuclear Plant Safety (1994).
CATTADORI G.. GALBIATI, L.. MAZZOCCHI, L. & VANINI, P., A Single Stage High-Pressure
Steam Injector for Next Generation Reactors: Test Results and Analysis, International Journal of
Multiphase Flow, in press (1995)
GROLMES, M.A. Steam-Water Condensing-Injector Performance Analysis with Supersonic Inlet
Vapor and Conx'ergem Condensing Section, ANL-7443 (1968)
GALBIATI, L.. MARTINI. R., Application of a high-pressure steam injector to the secondary side
of a PWR.
Report C1SE-SPT-94-43 (1994)
NOMENCLATURE

A
OF1
OF2
T

cross sectional area

primary overflow
secondary overflow
temperature

Greeks
T
mass flowrate
u>
liquid-to-steam fiowrate ratio
Subscripts

230

steam

liquid

OF2

secondary overflow

a
e
o

steam nozzle outlet, mixing section inlei

mixing section outlet, diffuser inlet
rest conditions

APPENDIX 1
TEST PROCEDURES AND TEST PROGRAM
The startup sequence in the tests was the following

a) at the startup both overflow valves and backpressure valve were opened,
b) water supply valve was opened first and most of the initial water flow was discharged through the
overflow lines,
c) when the steam supply valve was opened, condensation began in the mixing section,
d) a strong vacuum developed in the mixing section and the primary overflow valve could be closed
At this point a performance optimization sequence was adopted

e) the OF2 valve was panially closed (avoiding any interferences with the flow developed in the
mixing chamber),

0 the discharged backpressure could be increased by closing the backpressure valve till discharge
pressure was 10% higher than inlet steam pressure At this point data from data acquisition system
were recorded,
g) the OF2 valve was closed step by step (setting every time the backpressure valve to give discharge
pressure always 10% more than inlet steam pressure) till stalling was reached
After every backpressure valve setting, data were recorded
Four test series were performed in resting program

Test series I Using the above startup-optimization sequence tests were performed at constant inlet IIQU d
pressure (300 kPa) an temperature (21C) for different inlet steam pressures

Test series II At constant inlet liquid pressure (300 kPa). water temperature was increased up to 50C. using
again the described startup-optimizaiion sequence
Tesi series III Tests were performed and an SI "characteristic" (similar to that of a centrifugal punp
characteristic) was denved using the following procedure inlet liquid pressure and temperature were fixed
together with the inlet steam pressure Then, at different backpressure valve position, OF2 valve was gradual'v
closed till stalling
Test series IV In order to investigate the SI behaviour m case of a two-phase mixture entering the steam nozz e

some tests were performed \\ith an -nlet steam quality less than 1 (down to x=0 28) The above optimization
sequence was also used in these tests

231

TESTS ON FULL-SCALE PROTOTYPICAL PASSIVE

CONDENSERS FOR SBWR APPLICATION

P. MASONI
ENEA ERG FISS,
Bologna, Italy
A. ACHILLI
SIET,
Piacenza, Italy

P.P. BILLIG
GE Nuclear Energy,
San Jose, California, USA
S. BOTTI, G. CATTADORI, R. SILVERII
SIET,
Piacenza, Italy
Abstract

The Simplified Boiling Water Reactor (SBWR) is an evolutionary design in boiling water
reactors. A key feature of the SBWR is the use of simple passive systems to respond to any
type of design basis event (DBE). Two of these systems are the Passive Containment Cooling
System (PCCS) and the Isolation Condenser System (ICS).
As a part of the SBWR design and U.S. certification program, GE and the Italian companies
ANSALDO, ENEA, and ENEL are sponsoring a test program of full-scale, prototypical
condensers for both ICS and PCCS, at full pressure, temperature and flow conditions. The
PCC and 1C full-scale prototypes were designed, and were manufactured by ANSALDO. The
tests are performed by SIET at the Performance ANalysis and Testing of HEat Removal
Systems (PANTHERS) facility. ENEA has the responsibility to co-ordinate the tests and to
perform pre- and post-test analysis. Both the thermal and structural performance of the heat
exchangers are measured in these tests at various conditions the units might experience in the
SBWR. The test facility consists of two separate loops, one for the PCC testing, the other for
the 1C testing. At the time of writing this paper (April 1995), the PCC testing was completed
and the facility preparation for the 1C testing is hi progress. The 1C tests are scheduled to begin
in June and to be completed by the end of 1995.
The paper gives a brief description of the PANTHERS experimental program aimed to
demonstrate the thermal-hydraulic and structural performances of a full scale prototype of the
PCC . Preliminary results of the experimental tests are given. These results show the thermalhydraulic performance of the heat exchanger as a function of inlet pressure and of the air mass
fraction for: some steady-state performance tests; for a test in which the water level in the PCC
pool is allowed to drop and the PCC tubes to uncover and for a test with non-condensable gas
build-up.
The experimental results are very satisfactory and show a very good repeatability. The
structural design of the heat exchanger is very robust: the unit has survived ten LOCA cycles
and more than 100 thermal-hydraulic performance tests.
In addition to the PCC results, a short description of the test objectives and the test matrix for
the PANTHERS-IC testing is given.
1.

INTRODUCTION

The Simplified Boiling Water Reactor (SBWR) is an evolutionary design in boiling water
reactors (BWRs). The SBWR has been developed by an international design team from North
233

America, Europe, Mexico and Asia and led by the General Electric Company (GE). The
design extensively uses the technology of operating BWRs, as well as new developments
found in the Advanced BWR (ABWR). A key feature of the SBWR is the use of simple
passive systems to respond to any type of design basis event (DBE) [1]. These systems utilize
passive forces, such as gravity head, natural circulation, or naturally induced pressure
differences, to operate.
One of these systems is the Passive Containment Cooling System (PCCS). The SBWR
containment is similar to existing GE BWRs which have the reactor in a drywell region. The
drywell is connected to a wetwell through submerged pipes in the suppression pool, a part of
the wetwell. The PCCS consists of three Passive Containment Condensers (PCC) connected to
the upper drywell gas space. During a postulated loss-of-coolant accident (LOG A) steam in the
drywell is driven into the PCCS by the pressure difference between the wetwell and drywell in
combination with the vacuum produced by condensation. The condensate flows down into the
Gravity-Driven Cooling System (GDCS) pools in the drywell. The GDCS is another SBWR
passive system which provides makeup to the reactor. Non-condensable gases, such as
containment nitrogen, are separated in the PCC and vented to the wetwell through submerged
pipes in the suppression pool. All piping on the PCCS contain no valves, which results in a
complete passive operation.
Another important passive system is the Isolation Condenser System (ICS). Its main purpose
is to limit the over pressure in the reactor system at a value below the set-point of the Safety
Relief Valves, as a consequence of a main steam line isolation. Three condensers are
submerged in a compartmentlized pool of water located in the reactor building and above the
reactor containment. The primary side of the three Isolation Condensers (1C) are connected by
piping to the reactor pressure vessel. Closed valves in each condensate return line prevent
condensation during normal power operation of the plant. When operation of the 1C system is
required, the valves are opened, the condensate is returned to the reactor vessel by gravity, and
the steam flows directly from the reactor to the condensers. The rate of flow is determined by
natural circulation. Vent lines are provided on the 1C to remove non-condensable gases
(radiolytic hydrogen and oxygen) which may reduce heat transfer rates during extended
periods of operation.
As a part of the SBWR design and U.S. certification program [2], GE and the Italian
companies ANSALDO, ENEA, and ENEL are sponsoring a test program of full-scale,
prototypical condensers for both ICS and PCCS, at full pressure, temperature and flow
conditions. The two PCC and 1C full-scale prototypes were designed, and manufactured by
ANSALDO. ENEA has the responsibility to co-ordinate the tests conducted by SEET at the
Performance ANalysis and Testing of HEat Removal Systems (PANTHERS) facility [3]. Both
the thermal and structural performance of the heat exchangers are measured in these tests at
various conditions the units might experience hi the SBWR. The test facility consists of two
separate loops, one for the PCC testing, the other for the 1C testing.
At the time of writing this paper (April 1995), the PCC testing was completed and the
facility preparation for the 1C testing is in progress. The 1C tests are scheduled to begin in June
and to be completed by the end of 1995.
Both PCC and 1C tests are conducted in accordance with the requirements of NQA-l/la1983 Quality Assurance Programme.
2.

PCC TESTING

2.1. Test Objectives

The test objectives of the PANTHERS PCC Test Program are:
(a) Demonstrate that the prototype PCC is capable of meeting its design requirements for
heat rejection. (Component Performance)
(b) Provide a sufficient database to confirm the adequacy of TRACG to predict the quasisteady-state heat rejection performance of a prototype PCC , over a range of noncondensable flow rates, steam flow rates, operating pressures, and superheat conditions,
that span and bound the SBWR range. (Steady State Separate Effects)

234

(c)

Determine and quantify any differences in the effects of non-condensable build-up in the
PCC tubes between lighter-than-steam and heavier-than-steam gases. (Concept
Demonstration)
(d) Confirm that the mechanical design of the PCC is adequate to assure its structural
integrity over a lifetime equal to that required for application of this equipment to the
SBWR. (Component Demonstration)

2.2. Test Unit

The PCC is described in reference [3]. The full-scale PCC prototype consists of two identical
modules, and the complete two-module assembly is designed for 10 MW nominal capacity, at
the following conditions:
Pure saturated steam in the tubes at 308 kPa absolute and 134 C.
Pool water at atmospheric pressure and 101 C temperature.
Fouling on secondary side of 9x10"^ C/W-m2.
The PCdf design pressure is 759 kPa (gauge) and the design temperature is 171 C.
23. Test Facility Description

The PANTHERS-PCC test loop (Figure 1) includes a pool tank in which the full-scale
prototype of the two-module unit PCC is submerged in water. Figure 2 is a photograph of the
PCC in the water pool. In the picture, the front wall and the roof of the pool are removed. The
steel frame structure, visible around the PCC, holds in fixed position the pool instrumentation
and the cables of the PCC instrumentation. The PANTHERS-PCC test loop is described in ref.
[3-4].
1C POOL

LCV

VENT TANK

FIRST

DESUPERHEATING
STEAM
SUPPLY

Thermal Power
Primary Pressure
Primary Temp.

Steam Flowrate
Air Flowrate

\Pool Capacity

14 MW^
1 MPa
200 C
6.5 kg/*
0.9 kg/s
173 m3/

Legend:
PCV m Pressure Control Valve

FCV Flow Control Valve

TCV = Temperature Control Valve
LCV = Level Control Valve

Figure 1 - Schematic of PANTHERS-PCC test loop

235

Figure 2 - PCC installed in the PANTHERS pool

The test loop elevations are full-scale relative to SBWR. The required power to test the full
unit PCC is obtained using superheated steam from a power station adjacent to the
PANTHERS facility.
Test facility process instrumentation is used to measure global mass and energy balances and
to characterize the state of the test facility outside the condenser unit. Thermal-hydraulic
performance instrumentation is provided to monitor the primary and secondary side heat
transfer capability of the prototype and other pressure and temperature details of interest.
Structural instrumentation is provided on the condensers to confirm design stress levels and
monitor vibration frequencies. A summary of the instrumentation used to monitor the thermal hydraulic performance and mechanical behavior of the PCC is given in ref. [3-4].
Wall thermocouples are brazed at 9 different axial levels on 4 PCC tubes. These instruments
measure the differential temperature across the tube wall, to derive the local heat flux. This
wall thermocouple instrumentation is useful to identify the length of the tube over which
condensation occurs, when the PCC heat removal exceeds the thermal power of the supplied
air/steam mixture.

2.4. Test Matrix

The majority of the PANTHERS PCC testing is steady-state performance testing. For these
tests, the facility is placed in a condition where steam or air-steam mixtures are supplied to the
PCC at a steady rate, and the condensed vapor and vented gases are collected. All inlet and
outlet flows are measured. The vented gases are released to the atmosphere. Once steady-state
conditions are established, data are collected for a period of 15 minutes. The time-averaged
data are reported and analyzed. The table I shows the PANTHERS PCC Test Matrix in which
the tests have been divided into groups described below.
Test Group 1 is used to determine the baseline heat exchanger performance over a range of
saturated steam flow rates without the presence of non-condensable gases. Test Group 1
data are compared with design requirements to meet Test Objective 1.
Test Group 2 addresses the effect of superheat conditions in the inlet steam.
Test Groups 3 through 6 are PCC steady-state performance tests with air/steam mixtures.
As noted previously, the independent variables are steam mass flow rate, air mass flow
rate, steam superheat conditions, and absolute operating pressure.
236

Table I - PANTHERS-PCC TEST MATRIX

Test
Test
Group Conditions

1
2
3
4
5
6
7
8

9
a

6
4a
7a
2*
18
6
3
10

Description

PCC steady-state performance; saturated steam

PCC steady-state performance; superheated steam

PCC steady-state performance; air/steam mixtures

PCC steady-state performance; air/steam mixtures

PCC steady-state performance; air/steam mixtures

PCC steady-state performance; air/steam mixtures
PCC performance; non-condensable build-up
PCC performance; pool water level effects
PCC component demonstration; LOCA cycles

Each test is performed at five inlet pressures with fixed steam and air inlet flow rates

Test Group 3 is used to compare heat rejection rates over a range of air flow rates to the
saturated, steam-only condition determined from the pure steam series. Holding steam
flow constant at near rated conditions, these tests yield the effect of air on the condensation
process.

Test Group 4 supplements Test Group 3, in that it defines condensation performance at the
extremes of the SBWR air/steam mixture ranges, and at several intermediate points.
Test Group 5 further supplements Test Group 4 by extending the effect of noncondensable gases over the superheated steam range.

Tests included in Group 6 are lower priority tests. They supplement the previously
identified tests by increasing the data density within the already established air/steam flow
map.
The PANTHERS PCC test matrix includes transient tests which are used to establish noncondensable build-up effects and PCC pool water level effects. They are not intended to be

systems transient tests.

Test Group 7: six test conditions are specified in this Group. In these test conditions, steam
is supplied at a constant rate, and steady-state conditions established. Air, helium, or

air/helium mixtures are then injected into the steam supply, with the vent line closed, and
the transient degradation in heat transfer performance is measured, as a function of the
total non-condensable mass injected. Air is similar to nitrogen in molecular weight, and is
heavier than steam. Helium is lighter than steam, and is expected to behave in a manner
similar to hydrogen. Hydrogen can be present in the SBWR containment as result of
radiolysis and reaction of the Zircaloy fuel cladding with reactor water. Test Group 7 data
are evaluated to meet the requirements of Test Objective 3.

Test Group 8. In these test conditions, steam and air/steam mixtures are supplied to the
PCC , and steady-state conditions established, similar to the steady-state performance tests.

In these tests, however, the water level in the PCC pool is allowed to drop and the PCC
tubes to uncover as a result of boil off. Both the PCC pool level and the PCC heat rejection
rate are monitored as a function of time.
Test Groups 1 through 5, 7 and 8 provide a database for TRACG qualification and meet Test
Objective 2.
Component testing (structural tests) of the prototype PCC is performed using the same
hardware and test facility. The component demonstration tests are conduct in a similar manner

to the thermal-hydraulic testing. Structural data are collected during the thermal-hydraulic tests
as well as the structural performance tests. The approach taken to address Test Objective 4 is

to subject the equipment to a total number of pressure and temperature cycles in excess of that
expected over the anticipated SBWR lifetime.
Test Group 9: Simulated LOCA cycles are performed by pressurizing the PCC with steam,
so that both the temperature and pressure effects of a LOCA are simulated. The PCC pool
is at ambient temperature at the beginning of a test, but is allowed to heat up to saturation

as each cycle proceeds. Each LOCA cycle lasts approximately 30 minutes. Ten cycles are

performed.

237

3.

TEST RESULTS

The testing was completed at the end of December 1994. At the time this paper is written
(April 1995), the analysis of results is still in progress; we can only outline some preliminary
conclusions.
- The thermal-hydraulic performance satisfies the design requirements for the PCC unit.
- The PANTHERS/PCC facility operation is "smooth". Steady state conditions are rapidly
reached and kept.
- No unexpected phenomena have been detected.
- In the steady-state performance tests with pure steam (Group 1), the inlet pressure of the
PCC is a dependent variable and increases quasi linearly with the steam mass flow rate.
Figure 3 shows the experimental points and the regression line.
- In the steady-state performance tests with air/steam mixtures, the condenser efficiency
decreases as the air mass fraction (ratio between inlet air mass flow rate and total inlet
mass flow rate) increases. Figure 4 shows the experimental results for the tests with
constant inlet steam mass flow rate and constant inlet pressure. While four of the tests are
with saturated steam, one is with 20C superheated steam: the effect of superheating is
negligible. The condenser efficiency increases as the inlet pressure increases. Figure 5
shows the condenser efficiency as a function of the inlet pressure for tests with constant
steam mass flow rate, at four different values of the air mass fraction. The figure shows
also the results for tests with 20C and 30C of superheating. The effect of superheating is

PCC Inlet Pressure (kPa)

350 -i:-

^*^
300 -|
i
250 -1-
200 -i_^^

^^
Pun Steam

Test Results

Linear Regression '

150 -Ij^ ^^'

100

^~

-I
0
02.
0.4

0.6

0.8

Fraction of Maximum Steam Row Rate

Figure 3 - Pure steam tests: PCC inlet pressure vs. steam

flow normalized to the maximum steam flow rate
i

0.9

>.o.s ^
^_

f\

n -7
0> 0.7

O rt (*
?~

U4

>05
<n
W - .
c 0.4

Cor sfsnf Steam Flaw anri

Cot. slant Inlet Pressure

05

T3 n a

0 0.2
0.1
0
0

Saturated Steam

Superheated Steam (20*C)

0.05

0.1

Air Mass Fraction

Figure 4 - Condenser Efficiency vs. air mass fraction at

constant steam mass flow rate and constant inlet pressure

238

0.15

1 -

0.9
g.08-

'

"3

n-7
_ 0.7

S 0.6

6"

Constant Steam Flow Rate

13

j Air Mass Fract = 0.015

t *

j Air Mass Fract = 0.032

!

- oX

6
<D 0.5W - ,
c03 0.4

j * Hf Mass Fract = 0.074

i
Air Mass Fract = 0.147

"o n <a

Air Mass Fract = 0.147

20"C Superheat

0 0.2

* Air Mass Fract. = 0.148

0.1
n

30C Superheat

100

300

500

700

900

Inlet Pressure (kPa)

Figure 5 - Condenser efficiency vs. inlet pressure at ditterem

air mass fractions and constant steam flow rate.

again very small. The lower the air mass fraction, the lower the inlet pressure at which the
efficiency approaches unity.
Figure 6 shows the inlet pressure as a function of the 1C pool water level in a test of Group
8 (slow transient test) with steam only. While decreasing the water level in the PCC pool,
the inlet pressure decreases from the initial value to approximately 80 % the initial value,
until the tube bundle starts to be uncovered. From this moment on, the pressure increases.
Refilling the pool, the pressure follows the same trend, but with a small hysteresis. When
the tube bundle was almost completely covered with water, the inlet pressure again
reached its minimum and then followed the same values as when the level decreased The
PCC completely recovered its condensing efficiency.
Figure 7 shows the effect of the air build-up in the condenser tubes (Test Group 7), hi
terms of tube wall AT (normalized to the initial values) as a function of time. After
reaching steady state with steam only, air is slowly injected with a constant flow rate
(Time -1000 s) Because the venting is prevented, the air, heavier than steam, is
accumulated in the bottom of the PCC. Tube wall AT is proportional to the local heat flux.
It drops to zero when the portion of the tube where the thermocouples are located is filled
with air, preventing any further condensation. The lower is the location of the wall
thermocouple in the tube, the sooner the heat flux drops to zero.
2.5

Constant Steam Flow

c

Top] of Condenser Tube i

_ffi

1.5

si

3 <D

Sol

i
Water Level Decrease

0.5

Water Level increase

0.2

0.4

0.6

0.8

Pool Water Level / Normal Water Level

Figure 6 - Inlet pressure (normalized to the initial inlet

pressure) vs. pool water level (normalized to normal water
level)
239

2000

4000

6000

8000

10000

Figure 7 - Tube wall AT (normalized to the initial values) vs.

Time, at different axial tube location, in an air build-up test
4.

1C TESTS

4.1. Test Program Objectives

The test objectives of the PANTHERS-IC Test Program are:
1.
Demonstrate that the prototype 1C is capable of meeting its design requirements for
heat rejection. (Component Performance)
2.
Provide a sufficient data base to confirm the adequacy of TRACG to predict the
quasi-steady heat rejection performance of a prototype 1C, over a range of operating
pressures that span and bound the SBWR range. (Steady-State Separate Effects)
3.
Demonstrate the start-up of the 1C unit under accident conditions. (Concept
Demonstration)
4.
Demonstrate the non-condensable venting capability of the SBWR 1C design, and
condensation restart capability following venting. (Concept Demonstration)
5.
Confirm that the mechanical design of the 1C is adequate to assure its structural
integrity over a lifetime equal to that required for application of this equipment to
the SBWR. (ComponentDemonstration)
The thermal hydraulic specific objectives are:
a)
measure the steady-state heat removal capability over the expected range of the
following SBWR conditions:

steam pressure

concentration of noncondensible gases

pool-side bulk average water temperature

pool-side water level

b)
confirm that the vent lines and the venting strategy for purging non-condensable
gases perform as required during 1C operation.
c)
confirm that tube-side heat transfer and flow rates are stable and without large
fluctuations.
d)
confirm that there is no condensation water hammer during the expected start-up,
shutdown and operating modes of the 1C.
e)
confirm that the condensate return line performs its function as required during
steady state and transient operation and that water level oscillations and
condensation induced flow oscillations do not impair heat removal capacity
f)
Measure the heat loss from the 1C when it is in the standby mode, with the
condensate drain valves closed
g)
Measure the drain time for the 1C upper plenum during the 1C start-up transient.
The structural specific objectives are:
a)
measure the temperature and the stress levels at the critical locations of the 1C in all
the test conditions
b)
measure the vibration at critical locations on the 1C resulting from flow and/or
condensation.
240

c)
d)

verify through pre- and post-test non-destructive examination (NDE) of selected

header/tube weld joints that a specified fraction of thermal cycles results in no
excessive deformation, crack initiation or excessive crack growth rate
measure the stress levels at critical locations on the 1C, resulting from flow and/or
condensation induced vibration during expected periods of 1C operation

4.2. Test Unit

The 1C is described in ref. [3]. The full-scale 1C prototype is one module of the two-module
SBWR 1C design. The single module is designed for 15 MW nominal capacity, at the
following conditions:

Pure saturated steam at 289 C

Pool water at atmospheric pressure and 100 C
Tubes plugged 5%
Fouling factor on secondary side, 9x10'5 C/W-m2
The design pressure is 8.62 MPa and the design temperature is 302C. The 1C material is
INCONEL 600.
Figure 8 shows the schematic of PANTHERS-1C test loop.
TO STACK

1C POOL

VENT LINES

GAS VOLUME
MEASUREMENT

NON
CONDENSABLE
GASES

DE-SUPERHEATING
LINE

STEAM FROM POWER STATION

Fig. 8 - Schematic of PANTHERS-IC

43. Test Matrix

Seven types of tests are planned, three of which are structural tests. The majority of the
thermal hydraulic 1C tests are steady-state performance tests. The transient tests will be used to
demonstrate the start-up of the 1C under full-scale thermodynamic conditions. Table n shows
the 1C test matrix,
a)
Test type 1: steady-state performance tests
These data will establish the 1C heat rejection rate as a function of the inlet pressure.
In general, the procedure for the steady state tests will be as follows:
The steam vessel and 1C will be purged of initial air. The 1C pressure may be either
241

b)

c)

d)

e)

f)

g)

242

design pressure, or a lower value, depending on whether or not the test is also being
used as a structural demonstration cycle. The 1C is then placed in operation by
opening the 1C drain valve. Steam supply to the steam vessel is then regulated such
that the vessel pressure stabilizes at the desired inlet value. Data will then be
acquired for a period of approximately 15 minutes. At this point, the steam supply
may be increased or decreased to gather data at a different operating pressure, or the
test may be terminated. In all cases, flow into the 1C will be natural circulation
driven, as is the case for the SBWR.
Test type 2: start-up tests
These tests will be performed in much the same manner as the steady-state
performance tests, but transient data will be recorded over the course of the
experiment. The objective is demonstrate the start-up and operation of the 1C in a
situation comparable to a reactor isolation and trip. These tests will be performed
with an initial pressure of 9.48 MPa, and, after the opening of the drain valve, the
inlet pressure is stabilized at 8.618 MPa
Test type 3: non-condensable gas
Non condensable gas effects tests begin in a similar manner as he steady-state
performance tests, until the pressure has been stabilized at the desired value. In this
case, a mixture of air and helium will be injected into the 1C supply line at a very
low flow rate. The ratio of air to helium in the injected flow will be 3.6:1, simulating
the composition of radiolytic gases. Gas injection will continue until the 1C inlet
pressure increases to 7.653 MPag. The lower 1C vent is then opened, and the 1C
vented until pressure returns to the initial operating pressure, or stabilizes at an
intermediate value. If the pressure has returned to the initial value, the test is
terminated. If the inlet pressure has stabilized, the 1C top vent will be opened, and
the performance monitored until venting is complete, and the inlet pressure returns
to the initial value. The test is then terminated.
Test type 4: pool water level
Water level tests also begin with the 1C in stable operation at the desired initial inlet
pressure. The 1C pool water level is then reduced and the 1C performance monitored,
water level will be reduced until the pool level is at mid height of the condenser
tubes, or the 1C inlet pressure reaches 8.618 MPag (1250 psig) whichever comes
first. The pool water level will then be increased to normal and the 1C performance
allowed to return to normal. The test is then terminated.
Test type 5: normal 1C operation structural cycle.
This type of test is essentially representative of the cyclic duty expected of the 1C as
used in operation in the SBWR. The test is conducted hi a very similar manner as a
steady state performance test (test type 1) but the operating pressure is equal to the
design pressure, 8.618 MPa and the initial pool water temperature is less than 32 C.
Thermal hydraulic tests can be qualify as a structural type 5 cycle if these limits are
respected and the inlet pressure is held for 2 hours.
Test type 6: reactor heatup/cooldown without 1C operation.
This type of test is essentially representative of the cyclic duty expected of the 1C as
used in stand-by mode in the SBWR. The test consists in a pressurization of the 1C
up to 8.618 MPag and in a cooldown and de-pressurization. The initial pool water
temperature is less than 32 C. Thermal hydraulic tests can be qualify as a structural
type 6 cycle if these limits are respected.
Test type 7: Anticipated Transient Without Scram (ATWS) event simulation. The
test is very similar to test type 5, the main difference being a rapid pressurization of
the 1C up to 9.480 MPa, before opening the drain valve.

Table H PANTHERS-IC TEST MATRIX

Test Cond. No. of tests Test Type
No.
1
2
3
4
5

6
7
8
9
10
11
12
13
14
15
16
17
18

5.

3
1
1
1
1
1
1
1
1
1
1
1
1
1
1
20
5
1

2
1
1
1
1
1
1
1
1
1
1
3
3
4
4
5
6
7

Inlet

Pressure
(MPa)
8.618
7.920
7.240
6.21
5.52
4.83
4.14
2.76
1.38
0.69
0.21
0.48
2.08
0.48
2.08
8.618
8.618
8.618

CONCLUSIONS

The PANTHERS experimental program has successfully tested a full-scale prototype of the
PCC. Data from these tests are now being used in the certification work for the SBWR plant.
The broad selection of data allow analysts to study many applications of the PCC
performance, data collected by PANTHERS-PCC are used to compare the heat removal from a
prototype condenser with the SBWR design requirements. By using the steady-state tests with
steam only, a correlation can be developed to derive the condenser performance at design

conditions.

The database from the steady-state performance tests will be able to confirm the adequacy of
computer codes, such as GE's TRACG, to predict the quasi-steady-heat rejection performance
of a prototype PCC, over a range of non-condensable gas flow rates, steam flow rates,
operating pressures, and superheat conditions, that span and bound the SBWR range. The
transient tests can also provide data for computer code validation. For example, from one set
of transient tests, analysts can determine and quantify any differences in the effects of noncondensable gas build-up in the PCC tubes between lighter than steam (helium) and heavierthan steam (air) gases. From tests where the pool water level was allowed to drop and partially
uncover the tubes, the results can be used to study the performance of the heat exchanger with
varying heat transfer surface area.
Results from structural tests demonstrate that the PCC is a very robust design. After ten
LOCA cycles and over one hundred thermal performance tests, the unit emerged unscathed.
The final evaluations of the performance data from the PANTHERS-PCC are currently
taking place; however, at present it appears that all the test objectives will be met
In addition to the PCC test results, a short description of the test objectives and of the test
matrix for the PANTHERS-IC testing has been given. Based on the PANTHERS-PCC
experience, the authors are confident that the 1C testing program will be successfully

accomplished in the planned time.

243

REFERENCES

[1] UPTON, H. A., COOKE, F. E., SAWABE, J. K., 1993, "Simplified Boiling Water Reactor
Passive Safety Features", Proc. of the 2nd ASME-JSME Nuclear Engineering Joint
Conference ICONE 2, Vol. 1, (PETERSON, P.P., Ed.), San Francisco(CA) March 21-24 1993.
[2] Rao, A. S., 1993 "Simplifying the BWR" ATOM No 430, September/October 1993.
[3] Masoni, P., Botti, S., Fitzsimmons, G. W., 1993, "Confirmatory tests on full-scale
condensers for the SBWR", Proc. of the 2nd ASME-JSME Nuclear Engineering Joint
Conference ICONE 2, Vol. 1, (PETERSON, P.F., Ed.), San Francisco(CA) March 21-24 1993.

[4] Masoni, P., Bianchini, G., Billig, P.P., Botti, S., Cattadori, G., Fitch, J.R., and Silverii, R.,
1995. "Tests on Full-scale Prototypical Passive Containment Condenser for SBWR's
Application", Proc. of the 3rd ASME-JSME Nuclear Engineering Joint Conference ICONE 3,
Kyoto (J) April 23-27, 1995

244

TESTING STATUS OF THE WESTINGHOUSE AP600

E.J. PIPLICA, J.C. BUTLER
Westinghouse Electric,
USA
Presented by L. Conway

Abstract

Westinghouse has developed AP600~a 600 MWe two-loop, advanced, simplified

passive plant-in response to the Advanced Light Water Reactor Program sponsored by the
Electric Power Research Institute and the U.S. Department of Energy.

The principal safety functions of primary coolant inventory, reactivity control, reactor
residual heat removal, and fission product containment are accomplished with safety systems
that are based on logical extensions of proven technology and rely on natural forces such as
gravity, convection, and evaporation.

AP600 technology is supported by rigorous testing. The test programme is the single
most visible portion of the AP600; it is a global effort, with the cooperation of the U.S.
government (including the U.S. NRC), industry, and the academic world. With the
completion of this programme, the AP600 has become the most thoroughly and rigorously
tested reactor system in the world. The results have demonstrated the performance of the
components and systems unique to the passive safety concept upon which the AP600 is
based.
Westinghouse has developed AP600-a 600MWe two-loop, advanced, simplified passive
plant-in response to the Advanced Light
Water Reactor Program sponsored by the
Electric Power Research Institute and the
U.S. Department of Energy see figure 1).
The AP600 is currently scheduled to receive
Final Design Approval from the U.S.
Nuclear Regulatory Commission (NRC) in
September 1996. In support of the design
and the design certification review, the
AP600 has undergone the most extensive test
program ever conducted on a nuclear power
plant design.
The overall AP600 plant design follows in
the decades-long tradition of Westinghouse
two-loop PWRs (see figure 2), which have

consistently operated with average lifetime

availabilities of 83 percent-significanfly
better than the U.S. national average.

AP600 passive safety systems simplify

safety functions that have traditionally been
provided by active safety systems (see figure
3). The principal safety functions of primary
coolant inventory, reactivity control, reactor
residual heat removal, and fission product
containment are accomplished with safety
systems that are based on logical extensions
of proven technology and rely on natural
forces such as gravity, convection, and
evaporation.
Passive Residual Heat Removal
A natural circulation heat exchange loop
connected to the reactor and located inside
the containment transfers reactor heat to a
water pool inside the containment, serving
the same function as a conventional
emergency feedwater system (see figure 4).

245

Passive Core Cooling System

A few simple valves are used to align and
automatically actuate the passive safety
systems (see figure 5). To provide high
reliability, these valves are designed to
actuate to their safeguards positions upon
loss of power or upon receipt of a safeguards
actuation signal. However, they are also
supported by multiple, reliable power
sources to avoid the possibility of
unnecessary actuations.
The passive core cooling system (PXS) uses
three passive sources of water to maintain
core cooling through safety injection. These
injection sources include the core makeup
tank (CMT), the accumulators, and the incontainment refueling water storage tank
(IRWST). These injection sources are

directly connected via two nozzles for direct

vessel injection (DVI) so that no injection
flow wil] be spilled for the larger break
cases. The CMTs inject at any RCS
pressure, using gravity to provide injection
flow. For larger leaks, additional water is
provided by the accumulators, which inject
water pressurized by compressed nitrogen at
700 psig. Long-term injection water is
provided by gravity from the ERWST, which
is designed for injection at atmospheric
pressure. The automatic depressurization
system (ADS) is composed of four stages to
permit a relatively slow, controlled RCS
pressure reduction. The ADS depressurizes
the RCS to atmospheric pressure.
Passive Containment Cooling System
The AP600 steel containment vessel is a
robust containment system that prevents
radioactive releases to the environment and
serves as the vehicle for heat transfer for
reactor residual heat removal in the event of
postulated accidents (see figure 6). Outside
air is ducted by air baffles between the steel
containment shell and the conaete building,
cooling the vessel's outer surface by natural
convection. In accident scenarios, this aircooling is enhanced by draining water stored
in a 400,000-gallon annular tank in the roof
of the shield building onto the steel
containment shell. This tank has sufficient
water to providefor three days of cooling,

246

after which time additional cooling water

could be provided by operator action to
maintain low containment pressure and
temperature. But even if no additional water
was provided at this time, air-cooling alone
would be sufficient for continued public
safety.
AP600 Test Program
AP600 technology is supported by rigorous
testing. The test program is the single most
visible portion of the AP600; it is a global
effort, with the cooperation of the U.S.
government (including the U.S. NRC),
industry, and the academic world. With the
completion of this program, the AP600 has
become the most thoroughly and rigorously
tested reactor system in the world. The
results have demonstrated the performance of
the components and systems unique to the
passive safety concept upon which the
AP600 is based.
The AP600 test program was initiated in
1985 with the overall objectives to provide
design information to verify component

designs, to simulate the AP600 thermalhydraulic phenomena and behavior of the

passive safety systems, to provide high
quality, qualified data to validate the
computer codes used in the Westinghouse
safety analyses, and to support the U.S.
NRC design certification review of the
AP600. The AP600 test program was
completed in 1994 and successfully achieved
all of these objectives. The test results
confirmed the exceptional behavior of the
passive systems and were instrumental in
facilitating code validations.

The AP600 tests can be grouped into three

general categories:

circulated around the bearing and a motor

was used to spin the shaft at pump-operation
speeds (see figure 8).

Component Design Verification Tests:

These tests confirmed the performance of
AP600 design features.

The testing successfully demonstrated the

design and construction of a full-scale

Passive Containment Cooling System (PCS)

Tests: These tests confirmed the operation of
the PCS and helped validate the computer
code modeling.

encapsulated, depleted-uranium journal. The

bearing journal, radial bearings, thrust
bearings, and friction-dynamometer test rig
operated smoothly with no significant
vibration over the entire speed and load

Passive Core Cooling System (PXS) Tests:

These tests confirmed the operation of the
PXS and were invaluable in the development
and validation of computer code models.
Component Design Verification Tests
Component design verification tests were
used to confirm performance of AP600
features associated with the use of canned
motor reactor coolant pumps (RCPs) and
top-mounted nuclear instrumentation.

Reactor Coolant Pump (RCP) Air and

Water Flow Tests
Because the AP600 RCPs are attached to
and take suction from the steam generator
channelhead, an RCP air flow test was
performed to observe the flow patterns in the
steam generator channelhead (see figure 7).
The test demonstrated that no adverse flow
patterns that could affect pump performance
will occur and established the pump
hydraulic parameters. Water flow tests were
also performed to verify the hydraulic
performance.
High-Inertia RCP Rotor Test
To ensure that a large margin to departure
from nucleate boiling (DNB) is maintained in
the AP600 design following a complete loss
of motor electrical power to the RCPs, a fullscale, high-inertia .shaft using a stainlesssteel-encased, depleted-uranium journal was
constructed with bearing journals, radial
bearings, and thrust bearings. This assembly
was mounted on a shaft and placed in a
dynamometer test stand.
Water was

range. The test results demonstrated that a

large margin to DNB will be maintained
following a loss of power to the RCPs.

Incore Instrumentation System Test

This test investigated interference from the
control rod drive mechanisms (CRDMs) on
the fixed detector signal (see figure 9). A
full-scale CRDM was operated next to a
simulated fixed detector and the output signal
was compared to the input signal.

The test results demonstrated that the AP600

incore instrumentation system will not be
affected by electromagnetic interference from
the CRDMs.

Reactor Internals Air Flow Test

A scaled model of the reactor vessel
downcomer and bottom head was constructed

of clear plastic to allow visualization of flow

patterns in the vessel lower head (see figure
10). Flow tracers were injected to illustrate
these flow patterns. The test was conducted
at various flow rates to ensure that no flow
anomalies/vortexes can occur in the AP600
vessel due to removal of bottom-mounted
instrumentation.

This test demonstrated that there will be no

abnormal flow distribution.
Passive Containment Cooling System
(PCS) Tests
The operation of the AP600 ultimate heat
sink using natural, passive safety systems
has been confirmed by a comprehensive test
program. This program included large-scale

247

integral testing of the passive containment

cooling system (PCS). The PCS tests
included the heated plate test, the integral
PCS test, the air flow path resistance test, the
large-scale heat transfer PCS test, the water
distribution tests, and the wind tunnel tests.

Heated Plate Test

The initial data for and demonstration of the
PCS heat and mass transfer processes that
occur on the outer surface of the containment
were obtained by experiments on a thick steel
plate 2 ft. (0.61m) wide by 6 ft. (1.8m) tall
that was heated on one side and had an
evaporating water film and ducted air flow
on the other side (see figure 11). The plate
was heated to simulate the temperature of the
containment wall that would occur in an
actual plant following a postulated accident.
The total plate-heating capacity installed
was approximately 12,000 watts.
The resulting test data confirmed the
calculalional models for heat removal from
the containment surface for both a wetted
and dry surface. The test also snowed that a
stable water film can be easily formed and
that even high air velocities would not strip
water from the surface.

Integral PCS Tests

A 3-ft. (1-m) diameter by 24-ft. (7.3-m) high
steel pressure vessel was built to simulate the
entire PCS heat transfer processes occurring
both on the inside and outside containment
surfaces (see figure 12).
The steel pressure vessel was initially filled
with one atmosphere of air and heated on the
inside with dry steam. Three different steam
inlet arrangements were used: one with a
steam inlet near the top of the vessel to
minimize air steam mixing; one near the

bottom of the vessel to promote air steam

mixing; and one where steam was uniformly
introduced along the full height of the vessel
during the test The vessel was surrounded
by a full-size cooling air flow path. The
coolina air could be heated and humidified to

248

simulate a full-range of expected conditions.

The outer surface of the vessel was wetted by
applying water to the top of the external
surface.

This test demonstrated that even with the

least effective steam inlet arrangement, the
overall heat removal capability of the PCS
met or exceeded analytical predictions. The
dry heat transfer results of up to 40 psig
confirmed the capability of the PCS to
provide adequate cooling to maintain
containment pressure below its design
pressure when the initial water supply was
used (within three days), and if no resupply
occurred. There was sufficient heat removal
even without external water, preventing
containment failure following several
accident scenarios.
Air Flow Path Resistance Test
A 1/6-scale replica of a 14 degree section of
the entire PCS air flow path was constructed
to quantity the air flow path resistance (see
figure 13). This test resulted in the addition
of two aerodynamic improvements to the air
flow path-a rounded entrance into the air
cooling annulus and rounded air baffle
supports from the containment The final air
flow path resistance demonstrated was used
in subsequent analyses.
Large-Scale Heat Transfer PCS Test
A 1/8-scale AP600 containment vessel was
constructed to examine the performance of
the PCS on a large computer code
verification and to demonstrate that the code
results were scalable (see figure 14). This
vessel simulated the inside containment
structures and volumes in order to achieve
prototypical internal circulation patterns and
air/steam mixture ratios, as well as a scaled
external
cooling
air
flow
path.
Instrumentation was installed to measure the
vessel steel shell inner and outer wall
temperatures at multiple locations, inside
steam/air temperature, steam condensation
rate, cooling air temperature rise, and cooling
air velocity.

Testing was conducted for natural convection

air-only cooling and forced convection with
various degrees of water coverage on the
outer vessel. Controlled steam flows were
used to establish the test conditions providing
internal pressures up to 40 psig. The results
of the water distribution tests were used to
establish water coverage on the vessel dome
and walls, which ranged from full coverage
to partial water coverage using a striping
pattern. Cooling air flow rates were also

varied to simulate the range of expected air

flow velocities using both forced and natural
convection.
A condensation measuring
system was used to measure internal steam
condensation in these tests, focusing on both
the external heat transfer mechanisms and the
internal heat transfer mechanisms, effects of
noncondensable, and transient conditions
similar to those that may be encountered in a
severe accident scenario.
Testing
demonstrated PCS heat transfer capability
over a wide range of pressure, steam flow
rate, and noncondensable gas conditions.
Water Distribution Tests
These tests provided a large-scale
demonstration of the capability to distribute
water on the steel containment dome outer
surface and top of the sidewall for the entire
range of expected water flow rates. The test
apparatus is a full-sized, 1/8 sector of the
AP600 containment dome (see figure 15).

Testing has been completed, and the results

confirm the detailed design of the water
delivery and distribution system.
Wind Tunnel Test
The AP600 wind tunnel tests were performed
at the University of Western Ontario's
boundary layer wind tunnel. Detailed scale
models of the AP600 structures (see figure
16) were used in the test to simulate the
structural details of the shield building air
inlet and exhaust and surrounding buildings.

The tests demonstrated that wind will haveno adverse impact on the natural convectioninduced draft flow in the containment
annulus. The tests also provided data that
were used to assess baffle loadings under
severe wind conditions.

Passive Core Cooling System (PXS)Tests

The test programs for the AP600 passive
core cooling features included the passive
residual heat removal heat exchanger (PRHR
HX) test, the automatic depressurization
system (ADS) test, the check valve test, the
core makeup tank (CMT) test, the fullpressure, full-height integral test, and the
long-term cooling integral systems test
Passive Residual Heat Removal Heat
Exchanger (PRHR HX) Test
This test characterized the thermal
performance of the PRHR HX and me
mixing behavior of the in-containment
refueling water storage tank (IRWST), using
prototypical PRHR HX tubing material, tube
diameter, pitch, and length with the HX tubes
located inside a scaled IRWST (see figure
17). The test conditions covered a full range
of expected flow rates, including forced
PRHR cooling and natural circulation flow
rates by varying the pumped flow through
the tubes.
This test was the largest of its kind and
provided the data necessary to develop heat
transfer correlations for vertical tube HXs.
Automatic Depressurization System (ADS)
Test
The automatic depressurization system
(ADS) tests were a full-sized simulation of
the operation of the ADS function of the
PXS and was conducted at ENEA's Vapore

facility in Casaccia, Italy (see figure 18).

The purpose of the tests was to confirm the
design of the spargers, determine the
dynamic effects on the IRWST structure, and
confirm the operability of the ADS valves.

249

The tests demonstrated the successful

performance of the sparger. These tests were
performed over a range of water
temperatures and water levels using steam
flow rates greater than those anticipated in

the actual transient. System performance

was demonstrated over a wide range of plant
conditions using a full-scale simulation of an
ADS flow path.
Check Valve Test
Tests were conducted to demonstrate the
capability of the passive safety injection
check valves to open under low pressure
differential conditions that exist during
gravity drain injection (see figure 19). The
tests used a hydraulic loop to determine the
opening and operating differential pressure
for each of the check valves.
Test results demonstrated the suitability of
standard check valves for the passive safety
injection functions.
Core Makeup Tank (CMT) Test
This test verified the gravity drain behavior
of the CMT, the steam pressure balance
piping over a full range of flow rates and
pressures, and the operation of the tank level
instrumentation, which acts as a control for
the ADS. A 1/6-diameter and 1/3-height
scale CMT was constructed and
instrumented (see figure 20) to obtain the
condensation rates within the tank to verify
the analytical computer model. The test
facility was designed to simulate the CMT
operating modes over a wide range of
conditions.
The CMT test did, in fact, evaluate the
operation of the CMT over a full range of
conditions and provided invaluable data for
CMT computer code model development
High-Pressure Integral Systems Test
A large-scale, full-height, full-pressure
integral systems test was performed using the
SPES test loop at the SET facilities in
Piacenza, Italy (see figure 21). The purpose

250

of this test was to simulate the AP600

thermal-hydraulic phenomena and behavior
of the passive systems following specified

small-break loss-of-coolant accidents, steam

generator tube ruptures, and steam line
breaks. The test loop, known as SPES-2,
was modified to represent the AP600 loop
configuration and PXSs and included: two
cold legs, one short leg per loop, CMTs,
accumulators, PRHR HX, ADS, and DVI
lines. A series of tests were performed to
simulate high-pressure system interactions as
the result of a loss-of-coolant, steam
generator tube rupture, or steam line break
accident
Testing was completed in 1994. The tests
covered a broad spectrum of break sizes,
break locations, and system interactions.
The test results have been used extensively to
validate the computer codes and modes used
in AP600 safety analyses.
Long-Term Cooling Integral Systems Test
The long-term cooling integral systems test
simulated the operation of the PXS from
-300 psig, the transition to the natural
convection post-accident, long-term cooling
mode for the AP600, and demonstrated the
operation of the long-term gravity makeup
path from the IRWST and long-term core
cooling via the natural circulation flow path
from the flooded containment The test used
a scaled model made of stainless steel to
simulate the reactor vessel, IRWST, CMT,
RCS (including the pressurizer), and lower
containment structure (see figure 22). Water
was the working fluid, permitting direct
modeling of the test with existing
\Vestinghouse analysis codes. The core was
simulated with electric heater rods scaled to
match core decay heat
The tests were performed at Oregon State
University and completed in 1994. Since
completion of the testing, the test results have
been used to validate AP600 computer code
modeling. The U.S. NRC, in recognition of

the value of the test facility, has performed

additional confirmatory testing.

Design Certification
On June 26, 1992, Westinghouse submitted
the Standard Safety Analysis Report (SSAR)
and Probabilistic Safety Study (PSS) to the
U.S. NRC for review in support of the
application for design certification. The
SSAR includes a complete plant design

minus the as-site and as-procured

information. The PSS evaluates the AP600
design's risk to the public. It has also been
used by the AP600 design team as a design
tool, in that conclusions drawn from the
preliminary PSS have been factored into the
detailed design.

In November 1994 the NRC issued a draft

safety evaluation report on the AP600
design. The NRC safety evaluation report

FIG. 1.

(NRC SER) reported on the review of the

design certification submittals by the NRC
staff and by the Advisory Committee on
Reactor Safeguards, a committee of
independent experts from outside the NRC.
NRC technical approval of the AP600
design, know as Final Design Approval, is
expected in September 1996.
The successful completion of the AP600 test
program has facilitated the design
certification process and has culminated in a
thoroughly tested design. The AP600 test
program provides fundamental engineering
data not only to support the licensing process
leading to NRC design certification, but also
to provide utilities with confidence in the
innovanve aspects of the AP600. The
AP600 test program supports the licensing
schedule leading to PDA in September 1996.

AP600 - The new Westinghouse standard 600 MWe plant

251

AP600
Pressurizer

Nuclear
Steam Supply
System

Steam Generator

Hot Leg Pipe

Cold Leg Pipe

Reactor
Vessel

FIG. 2.

High Inertia
Canned Motor
Pumps

AP600 nuclear steam supply system

Standard 2-Loop
Ultimate Heat Sink

AP600

Ultimate Heat Sink

^*--

FIG. 3.

252

AP600 simplified safety systems rely on natural forces

IRWST-

Screen

FIG. 4.

Passive safety systems

Core Makeup
Tank (1oI2)

Pumps

FIG. 5.

AP600 - Passive core cooling systems

253

PCS Water
Storage Tank

Air Inlet

Air Inlet

Concrete
Shield
Building
Air Flow
Baffle

Steel
Containment
Vessel

AP600
6/16/94

AP600-C-GF.11

FIG. 6.

FIG. 7.
254

AP600 - Passive containment cooling systems

AP600 reactor coolant pump sir and water flow tests

FIG.

FIG. 9.

8.

AP600 high-inertia RCP rotor test

AP600 incore instrumentation system test

255

Exhaust

Support

Cold Leg

Radial Keys

Lower
Plenum

Secondary
Core Support

Vortex Suppression
Device
AP600C-LF.38

FIG. 10.

AP600 reactor internals airflow test

AP600C-LF.44

FIG. 11.
256

AP600 heated plate test

FIG. 12.

AP600 integral PCS tests

-24

PT#17Chimney, A = 0.75 ft.2

PT#16

Screen, A (Total) = 1.38 ft 2>

Inlet
A = 2.18 ft.2

-20

PT#15-

-16
PT#1

Distance
Above
-12
Operating
Deck (ft.)

PT#2-

PT#12
PT#11

-8

PT#10

PT#3-

Containment Annulus,
'A = 0.50 ft 2

Shield Building Annulus,

A =1.30 ft.2

PTM-

-4

'PT#8
PT#7

PT#5

PT#6
I

I I

12

I I I

Lo
I

I I I I I

Radius of Section (ft.)

FIG. 13.

AP600 PCS air flow path resistance test

257

FIG.

14.

FIG. 15.

258

AP600 large-scale heat transfer PCS test

AP600 water distribution tests

FIG. 16.

FIG. 17.

AP600 wind tunnel tests

AP600 passive residual heat removal exchanger test

259

FIG. 18.

AP600 automatic depressurization system test

niabT SUPPORT FIXTURE

' -*'"

FIG. 19.

260

AP600 check valve test

FIG. 20.

FIG. 21.

AP600 core makeup tank test

AP600 full-pressure, full-height integral systems test

261

FIG. 22.

262

AP600 long-term cooling integral systems test

TESTING FOR THE AP600 AUTOMATIC

DEPRESSURIZATION SYSTEM

T. BUETER, L. CONWAY
Westinghouse Electric Corporation,
Pittsburgh, USA
P. INCALCATERRA, C. KROPP
ENEA C. R. Casaccia, Rome,
Italy

Abstract
In order to test the design of the ADS and develop data for the
The Automatic Depressurization System (ADS) of the
Westinghouse AP600 reactor will be used to provide controlled
depressurization of the reactor coolant system (RCS). This will.
ID turn, allow the initiation and long term operation of gravity
driven cooling flow in the RCS. ADS tests were conducted at the
VAPORE test facility in Casaccia. Italy through a Technical
Cooperation Agreement between Westinghouse, ENEA,
SOPREN/ANSALDO. and ENEL to produce data for the
development and verification of computer codes to simulate the
system. The test program also provided insights about the
operation of valves supplied from various vendors that could be
used in the AP600 ADS. The data gathered from the tests
showed the ability of the ADS design to fulfill its function over
the range of conditions expected in the AP600. The tests also
demonstrated the abilities of gate and globe valves from several
vendors to initiate and terminate an ADS blowdown as could be
required in the AP600.

INTRODUCTION
The Automatic Depressurization System (ADS) in the
Westingbouse AP600 Reactor will be used to provide staged

depressurization of the reactor coolant system (RCS) to allow the

initiation and continued long term gravity driven cooling flow to
the RCS. The AP600 ADS design consists of 2 independent sets
of ADS flow paths. Each of the flow paths consists of 4 separate
stages. Each stage will contain 2 valves in series, an isolation

development of computer codes mat model the AP600 ADS. a full

scale model of one of the ADS flow paths (stages 1, 2, and 3)
was constructed at ENEA's VAPORE (VAlve and Pressurizer
Operation Related Experiments) facility in Casaccia. Italy. The
ADS tests were performed at VAPORE in three phases. The first
phase, phase A. consisted of steam blowdowns performed to
measure the steam condensation pressure pulses generated by the
sparger. That testing was completed in 1992 and produced
valuable data for the design of the IRWST structure as well as for
the computer codes. Other papers have been published on this
aspect of the ADS testing, and this paper will focus on the
remaining phases of the testing.
The phase B testing was done through a technical cooperation
agreement of Westinghouse. ENEA. SOPREN/ANSALDO, and
ENEL to gather data on the ADS performance. The Phase 8
testing also was done to indicate the loadings that can be expected
in die ADS piping, to gather data on the types of valves that
could be used in the AP600 ADS, to gather data on the effects of
the ADS blowdowns on the valves, and to benchmark the safety
analysis computer codes. The phase B testing was further divided
into two parts, Bl and B2. Phase Bl produced data with which
to benchmark the computer codes used for the design certification
of the AP600. Phase B2 showed the effects of the blowdown on
the valves and gathered information about the various loads
produced by the blowdown on the valves and piping of the ADS.

valve and a control valve. Stages 1 through 3 are positioned at

TEST FACILITY

the top of the pressurizer. Stage 4 is connected to one of the two

hot legs in the RCS loops. Each of the stages 1. 2, and 3 of the
ADS will discharge through a sparger into the In-containment

valves for power plants. It has a 1400 ft* (39.6 m5) PWR
pressurizer and a discharge line at the top of the pressurizer for

Refueling Water Storage Tank (IRWST). which will condense the

steam discharge.

The VAPORE test facility was originally designed to test safety

steam discharges through the valves and into an in-ground,

reinforced concrete discharge tank. The pressurizer has a volume

263

of about 40 m3. and a design pressure and temperature of 19.7

MPa (about 2900 psig) and 365C (62SF). respectively. There
are t .6 MW of electric heaters available to beat the water for a

locations in the VAPORE ADS piping and the sparger discharge

test at VAPORE. This facility met the needs of the tests of the
ADS design, although some significant modifications were needed

in various positions on the discharge pool walk and in the

immediate vicinity of the sparger. Typically, there were 43
pressure transducers, 10 pressure transmitters, IS strain gages, 10
load cells. 9 accelerometen, 2 level transmitters. 1 A? cell, 3
valve position indicators, and 46 thermocouples in the set of

for the AP600 testing program.

Modifications were made to the discharge tank, including the

addition of a prototypical sparger, and the ability to beat the
condensing water in the tank to up to 100C with superheated
steam from the C. R. Casaccia district heating system. For the
phase B testing the sparger was repositioned at a higher elevation
in (he pool, because the analyses from the Phase A part of the

testing indicated the new position would allow the steam

discharge to condense more quickly and reduce the loads on the

tank walls.
A water line with two isolation gate valves was added at the

bottom of the pressurizer, and a full size prototypical ADS loop

was constructed. It was built to allow it to be connected to the
steam discharge line or the water discharge line through the use
of a common elbow/spool piece that bad to be moved from the
steam line to the water line (or visa versa) for a test.
The ADS loop at VAPORE consists of three stages of piping

and valves. Each stage consists of a pair of valves, a gate valve

upstream of a globe valve. The three stages are arranged in a
loop as shown in Figure 1. The AP600 design calls for two
redundant loops to be supported on top of (he pressurizer. The

pool. These data points included the temperatures and pressures

before and after each of the valves, the temperature* and pressures

instruments used on the ADS loop and in the sparger pool during
a test
For the Phase B2 tests, the ADS valves were also instrumented

with a data acquisition system developed by ITI MOVATS. This

system included instruments to measure the valve torque and
thrust, the valve stem position and the movement of the valve
spring pack at any point during the valve stroke.
Examples of die Phase B1 data for a blowdown from one of the
tests is shown in Figures 2. 3, 4, and S. Figure 2 shows the
pressure and temperature of (he fluid as it eaters the ADS loop.
As flow was initiated by opening the 12-inch gate valve used for
isolation of the pressurizer, the rapid increase in temperature and
pressure can be seen at the entrance to the ADS loop. The rather

quick initial drop in the temperature shown in this figure at the

beginning of the blowdown is a result of the cold water from the
bottom of the pressurizer traveling through the piping (where the
water discharge line was located). After the few seconds required
for (he plug of cold water to pass, the temperature increase ia the

fourth stage of the AP600 ADS which connects to the hot legs
of the RCS, was not modeled at VAPORE.

blowdown fluid as it enters (he ADS loop is evident

The pressure in the piping before the ADS loop decreases as the
blowdowo progresses until the isolation valve is closed, when it

PHASE B TESTING
The phase Bl testing was performed in the latter half of 1994.

levels off. The pressure and quality at the entrance of the ADS
loop was controlled through (be use of one of the 12-inch

The tests were designed to produce flow regimes similar to those

that the ADS on the AP600 would experience. That is, the phase

insolation gate valves. VLI-2, as a kind of orifice. The valve was

left in a carefully measured, partially closed position in order to

Bl tests were designed to produce data (hat would bound (he mass

flow and quality of expected transients in the AP600. The tests

showed that the ADS design was capable of performing its

create a flow orifice to limit the mass flow. This also created a
source of a pressure drop and flashing to steam in the saturated
fluid. This use of the valve allowed for the input fluid quality to

depressurization function as well as provided data for the

be varied to as much as 20%, while the mass flow was kept in the

validation of computer codes that could simulate the AP600 safety

systems. The Phase B1 data also provided the responses of the
sparger discharge pool to the revised elevation of the sparger.

required range of the AP600 design.

Figure 3 shows the temperature and pressure in the middle of
(be ADS stage through which the blowdown traveled, in this case
(be 4-inch piping, stage 1. The rapid increase of the pressure and
temperature correlates to the time of the opening of the 12-inch
isolation valve plus a short time for the fluid to transit to the ADS

The Bl tests were performed with only two 8-incb gate valves,
stages 2 and 3. and one 4-inch globe valve, stage 1. The 8-incb
globe valves in stages 2 and 3, and 4-incb gate valve in stage 1
were simulated with orifices in some of the tests to bound the

data on the blowdown with all six of the valves installed. Some
of the Bl tests were performed with no orifices installed to bound
the maximum mass flow of (he ADS loop.
The Phase Bl tests included several series. One test series

consisted of blowdowns through only one stage, a gate/globe

(orifice/valve for the Bl tests) pair in the ADS loop. Other tests
called for the blowdown to go through 2 stages together, and
other tests required 3 stages together.

The tests included blowdowns with 2 phase flow, i.e., flow from
the bottom of the pressurizer through the liquid line. There were
also tests with saturated strain through the steam discharge line in

loop. For this test the flow area through the 12-inch gate valve.
VLI-2. used to control the mass flow and quality, was slightly less
than dje flow area of any of (be piping or valves in stage 1.
Figure 4 shows (be pressure and temperature of the fluid in (be

16-inch discharge line for the ADS loop. The data trace mimics
those ID the previous figure, albeit at a lower pressure due to the
pressure drop through the ADS stage piping and valve and the
increase in the size of the piping as the fluid exits the ADS loop.
The time delay is slightly longer as the fluid had to travel
completely through the loop before it arrived at toe discharge line.

Figure 5 shows the discharge pool response to the blowdown

from an instrument near the lower part of one of the sparger arms.

the top of the pressurize!. These two types of tests showed the
influences of relatively large mass/small volumetric flows versus

The initiation of the discharge can be clearly seen in the

temperature data. The pressure changes cannot be discerned,

relatively small mass/large volumetric flows in the ADS and the

discharge pool.

because they are so small. This is because the pressure increases

only slightly as the condensing water remains subcooled with

For all of these tests data was taken from approximately 130

atmospheric pressure at the top of the pool. The water in the

264

discharge pool very rapidly cooled the saturated, two phase fluid
from tbe blowdowo. causing the rapid local temperature increase
in tbe fluid where tbe sparger arms were located.

ool follow tbe torque as closely as expected. The test valve

operators bad sufficient torque to open and close tbe valves, and
it was expected (bat tbe thrust traces would all look the same as
tbe torque traces. But, there were tests where this was not true.

Some of tbe Phase B1 tests required that the discharge pool be

cold, i.e., less than 60C. while others required that the discharge
pool be at saturation temperature, 100C. Tbe hot pool tests

This was thought to be due to tbe design of tbe threads on tbe

valve stem, ihe type of grease used, or tbe forces imposed by tbe

were performed to gather data on tbe effects of tbe blowdowns on

fluid flow through tbe valve.

tbe discharge tank for those analyses where it is postulated that

The valve was opened under tbe pressure of tbe fluid from tbe

tbe pool may have been heated to saturation prior to tbe ADS

pressurizcr. remained open for a few seconds, then closed. Tbe

initiation. The cold pool tests were performed to collect data on

valve was fully open at the point in the traces where an almost

tbe expected usual operating conditions of tbe ADS.

As discussed earlier the cold pool tests showed that the ADS

horizontal line can be seen in tbe data traces. Another increase

in the torque and thrust occurred when tbe valve started to close

loop discharge was rapidly cooled, condensing the steam in tbe

two phase flow, or completely condensing the saturated steam
flow for the steam tests. The average temperature of the

again. The running values can be seen until the valve contacts tbe
seat, at which point tbe gate is forced into the seat and the torque
rises to its initial value. The torque is reduced slightly on tbe

discharge pool after tbe test was increased nominally with tbe

close stroke as the force of tbe fluid helps to close the valve.

blowdowa, usually 5C to 10C. In Figure 5, a higher overall

temperature can be seen at the end of the test than was at tbe
beginning of tbe test This is a local, transient, effect seen

This can be seen at tbe peak at tbe end of the trace where tbe
torque drops rapidly a few percent just after it reaches tbe peak.
Figures 6 and 7 can be compared to the fluid pressure upstream
and downstream of the ADS loop in Figures 8 and 9, respectively,

immediately after a test was completed. After a relatively short

time, tbe temperature of the pool reached equilibrium at a level

5C to 10C higher than it was before the test

Tbe Phase B2 tests were designed to examine tbe valves
installed in the ADS loop at VAPORE. Gate valves and globe
valves were designed, built by several different vendors, and
installed in tbe test facility. Two of tbe gate valves and one globe

to see the relative fluid temperature and pressure at the various

stages of tbe valves' operation. As with tbe Phase Bl tests tbe

pressure in the ADS loop exhibits a rapid increase as the valve

opened, peaks, then drops off as the valve was closed. Tbe
pressure upstream of the ADS loop steadily decreases as the

valve used in the Phase 62 tests were used for tbe phase Bl tests
as well. In the Phase Bl tests the valves were always used either

pressurizcr is drained. Again, there is a sharp, short lived

decrease in tbe fluid temperature in tbe entrance to the ADS loop
as tbe cold water from tbe bottom of the pressurizer passes past

in the fully open or fully closed positions, i.e.. they were not

tbe entrance to the ADS loop.

opened or closed with a differential pressure across them.

ID contrast, tbe Phase B2 tests were designed to test the ability

Since one of the concerns with gate valves is the thermal

binding of tbe gates in the seats, cold stroke tests were performed

of each valve to initiate and terminate an ADS blowdown at the

various pressures expected to be seen in tbe AP600. One series

valve pair. Another series of tests called for the initiation and
termination of the blowdown flow with tbe gate valve in tbe
stage. These tests were done in a range of pressures from

after each hot blowdown test, to show any significant changes in

the torque or thrust required to operate a valve when it was
allowed to cool down in tbe closed position.
The results of those tests showed that tbe torque and thrust
required to unseat tbe gate of tbe valve after it had cooled were
usually lessened after tbe valve had cooled in the closed position.
In fact, after the valve had been allowed to cool from operating

approximately 400 psig (30 bar) to 2235 psig (154 bar) to cover
tbe range of pressures for the AP600 ADS.

temperatures, then stroked one time, tbe opening torque and thrust
were usually slightly higher on subsequent cold stroke tests. Thai

Figures 6, 7, 8 and 9 show some of tbe results from a typical

Phase B2 blowdown test. In Figures 6 and 7, tbe data traces for
tbe torque and thrust, respectively, applied to the 4-inch gate
valve during a blowdowo can be seen. This blowdown was
initialed and terminated with the 4 inch gate valve. Tbe "steps"
in the torque signal as tbe valve closes are due to tbe effects of
tbe fluid flow on the valve gate. That is. a static valve test would
not show these characteristic plateaus in the torque during the

is. for the first cold stroke tbe torque and thrust required to open
tbe valve were usually less that what was required when the valve
was hot. The torque and thrust values returned to their usual
values on tbe second and third cold stroke tests. In all cases tbe
torque and thrust required to open the valve were essentially tbe
same when the valve was not as when the valve was cooled in tbe
closed position before it was opened.

of tests called for tbe initiation and termination of tbe blowdown

flow with tbe globe valve in the ADS stage, i.e., a gate/globe

valve stroke. In this test, as in all of the B2 tests, tbe ADS valve
successfully initiated and terminated tbe flow through tbe ADS
loop.
Tbe maximum torque and thrust for tbe valve can be seen a few

seconds after the beginning of the valve stroke, just before tbe
torque and thrust drop from their maximum to their minimum..
Tbe maximum torque is usually when tbe valve is in its seat, as
is the case in Figure 6. The torque will reach a 0 point just after

it starts to open as seen in the figure where the trace drops in a

vertical line to tbe minimum value. Tbe thrust trace in Figure 7

is similar in shape to tbe torque data, as the thrust would be

expected to be. In some of tbe tests the valve thrust values did

SUMMARY
The ADS testing program performed at VAPORE produced data
for the development and refinement of computer simulations of
tbe AP600 ADS. The tests demonstrated the ADS design can
successfully depressurize tbe AP600 RCS in a controlled manner
for tbe range of pressures and fluid qualities that may be seen in
the AP600. Tbe phase B1 tests covered a range of combinations

of flow paths that could be used if the ADS was called upon to
depressurized the RCS. This phase of tbe testing also
demonstrated the suitability of the submerged depth of the

discharge tank sparger for the AP600 design. The phase B2

tests demonstrated tbe ability of gate and globe valves from a

variety of vendors to successfully initiate and terminate a

blowdown of tbe AP600 RCS. Phase B2 also covered a wide
range of supply pressures to test the valves in conditions
representative of the AP600.

265

N>

FIG. 1.

ADS test loop

1.20

1.00

-.

0.80

m
_

0.60

I- I
0.40

0.20

0.00

-0.20

Time

Entrance to ADS Loop

(B1 sage 1)

FIG.

Pressure and temperature at entrance to ADS loop

2.

1.20

1.00

0.80

-~
S

060

~~

0.40

f
*

0.20

000

J.

020

Time

FIG.

3.

Pressure and temperature at middle of ADS loop

267

1.20

1.00

0.80

I- I
0.40

020

0.00

-020

Time

Discharge of ADS Loop

(B1 - stage 1)

FIG. 4.

I^U

Pressure and temperature at discharge of ADS loop

1
:

j
!

i
ii

1.00

';/ >. ;

;
: :, i = l ; Vfe

' i . .= \\ \

"x"

ID

'

0.80

tl

"

:r

1
g-

;:

* Pressure
r
j

0.60

0.40

oi2

I
i

020

Time

Discharge Pool
(B1- stage 1)

FIG. 5.

268

Pressure and temperature in discharge pool

- Temperature

1.20

1.00

0.80
X
19

0.60

Torque |
o
0.40

0.20

0.00

-0.20

Time

FIG. 6.

Torque applied to gate valve during blowdown

1.20

1.00

0.80

e 0.60
~5
0.40

0.00

-0.20

Time

FIG. 7.

Thrust applied to gate valve during blowdown

269

098

'

I ' !
085

Tkm
c* to
AOS Loop|
Loop!
Entfanc*
(82?-Ugl)

FIG. 8.

Pressure and temperature at entrance to ADS loop for phase B2 blowdown

1 10

100

i
1

0>0

oeo
T

'

\
jf!

'

f
i

'

070
1

i ,
I i

000

010

1
i

i
[i

1 i

020
0 10

i
jr1T"
i/pI ' j
i i
i

tt

040

030

i j;
:

ItI* '. .
t |

1
| | 050

'

. !

i i i

Oscharp* ol ADS Loop]

<B2 lapoll

FIG. 9.

270

Pressure and temperature at discharge of ADS loop for phase B2 blowdown

THE STUDY OF THE EFFECTIVENESS OF THE

EMERGENCY CONDENSER OF THE BWR

600/1000 IN THE NOKO TEST FACILITY

E.F. HICKEN, H. JAEGERS, A. SCHAFFRATH
Forschungszentrum Juelich, ISR,
Juelich, Germany

Abstract
The BWR600/1000 is a new innovative passive boiling water reactor concept which is being
developed by Siemens. The concept is characterized in particular by passive safety systems
(i.e. four emergency condensers, four building condensers, eight passive pressure pulse
transmitters, six gravity-driven core flooding lines, eight rupture disks arranged in parallel to
the relief valves and two scram systems).

For experimental investigations of the effectiveness of the emergency condenser the NOKO
test facility has been constructed at the Forschungszentrum Julich in cooperation with
Siemens. This project is sponsored also by the BMBF and some German Utilities. The test
phase started in early 1995.
The NOKO facility has a maximum operating pressure of 9 MPa and maximum power of 4
MW for steam production. The emergency condenser consists of 8 tubes and is fabricated
with original geometries and materials.

The objectives of the experiments are the study of the effectiveness of the BWR600/1000
emergency condenser under all expected conditions and (including non-condensibles), the use
of the experimental data for computer code validation and model improvements.
Some results from the start-up phase will be given.

After finishing the emergency condenser test series several other components (e.g. building
condensers and passive pressure pulse transmitters of the BWR600/1000) shall be tested in
the NOKO facility. Because of the multi-purpose design of the NOKO test facility only few
reconstructions are necessary for other designs.
1.

Introduction

The BWR 600/1000 is a new innovative passive boiling water reactor concept which is being
developed by Siemens. The concept is characterized in particular by passive safety systems
(i.e. four emergency condensers, four building condensers, eight passive pressure pulse

transmitters, six gravity-driven core flooding lines, eight rupture dishes arranged in parallel to
the safety valves and two scram systems. For completeness fig. 1 and 2 show the design of the
BWR 600/1000.
For experimental investigations of the effectiveness of the emergency condenser (and later for
other components) the NOKO test facility has been constructed at the Forschungszentrum
Julich (KFA) in cooperation with Siemens, which was responsible for the detailed planning
and for most components. This project is funded by the BMBF, some German Utilities,
Siemens and KFA.
2.

The NOKO Test Facility

The general design is shown in fig. 3. An electrical heater, which has been used before for in
the HDR facility, produces a two-phase flow mixture, which is separated in the separator. The
pressure vessel with an inner diameter of 0.5 m and a height of 12.7 m is mainly used to
control the water level. Connected to the pressure vessel are the feed line and the down-pipe
271

of the emergency condenser. Both pipes lead the headers; eight pipes are connected to the
headers. For the simulation of the water pool the emergency condenser is placed within a
vessel. This vessel, named condenser, contains 20 irr of water and can be operated up to 1
MPa. The heat removal from this condenser is evaporation of water. The steam is condensed
in the condensation tank, which is used also as a water storage tank and as a depressurisation
system for relief and safety valves. The heat removal from the condensation tank happens
through a cooling loop which transfers the energy to a river or a cooling tower.
The power level of the electrical heater can be controled in 8 steps; the maximum pressure
within the primary loop is 9 MPa.
The detailed planning of the test facility started in October 1993; in December 1994 the
facility was running for a short time with 1.35 MW.

3.

Instrumentation

In addition to the instrumentation used for operational purposes the test instrumentation is
shown in fig. 4. As a result from start-up tests some more instruments, mainly mass flow
instrumentation will be installed.
The accuracy of instrumentation will repeatedly be checked with calibrated instruments; the
power of the electrical heater can be measured with an accuracy of 0.2 %.
4.

Test Programme

Although not yet fixed we expect about 100 test runs within 1 year. The variables to be studied
are power, pressure, pressure vessel level on the primary side, pressure, temperature and
water level on the secondary side, length of down pipe and flow resistances in steam line and
down-pipe. In addition, the mass flow through the emergency condenser in case of a flooded
pressure vessel will be studied.
5.

Results from Start-Up Tests

After the usual difficulties during the start-up period we can state that the electrical heater
and the control systems operate appropriately. The recirculation pump is sensitive to
subcooling due to requirements from the component protection system.

The instrumentation shows no difficulties. The heat up period is about 2-4 hours and the time
for one test between 30 and 60 minutes.

In fig 5 and 6 preliminary date are shown. The solid curves show the expected values as
calculated by the vendor. Both figures contain experimental data.

6.

Future Tests

It is planned to test the passive pressure pulse transmitters, see fig. 7, in parallel to the
emergency condenser.
Following the tests with the emergency condenser, the building condenser, fig. 8, will be
tested.
Following these tests steam injectors, see fig. 9 are candidats for being installed in the NOKO
facility.

272

Conuinmenl cootng
condense* (1 oM)

F/G. 7.

5WR 600/1000 containment

273

condenser (total of 4)

8 Containment cooling
condenser (total of 4)
9 Core flooding pool

2 Safety-relief valve (total of 8)

3 Spring-loaded pilot valve

4 Diaphragm pilot valve

10 Pressure suppression pool

5 Passive pressure pulse transmitter

11 Horizontal discharge vent

6 Rupture disk (total of 8)

12 Vertical discharge shaft

7 Flooding line (total of 4)

13 Scram system

FIG. 2.
MT> Simulator

BWR 600 - passive systems

Reactor Pressure
Vessel

Emergency Condenser

p = 10 bar.
V = 20 m .

Ysteam
' Generatoi

Pressure
Vessel
( h = 12 7 m
-05m]

Header 0 2 m. I = 6 m

' V /; i

Ram
Water
Channel

River
Water

Condensation Tank

FIG. 3.
274

Cooling Circuit

General design of the NOKO test facility

Condenser

1
T

M | H

68

1 i! 1

Pressure
Vessel

17

O
Water

-XI
FIG. 4.

Instrumentation scheme

>H ( % 1
)0

1AA.

QH = Cooling capabll iy at lower RPV waist lev<>l

3U

Qo = Maximum cooling capability

80
70

O
60

50

30

20

/\

O test during start up phase preliminary evaluation

5.

\/

40

10

/ ^

^\

i
i
\

1
1

SWR 1000- emergency condenser. Cooling capability as a function of loss of

water level in the RPV (AH in m)
275

100

21
?

90

"

80

O
O)

70

60

50

40

30

10

10

20

30

40

50

Pressure [ bar ]

FIG. 6.

SWR 1000 - emergency condenser.

pressure in the RPV

FIG. 7.

216

60

70

^-

Cooling capability as a function of

Passive pressure pulse transmitter

steam

outlet

inlet

77

77

FIG. 8.

Building condenser

RPV

RPV

FIG. 9.

Steam injectors

211

EXPERIMENTAL STUDY OF ISOLATION CONDENSER

PERFORMANCES BY PIPE-ONE APPARATUS

R. BOVALINI, F. D'AURIA, G.M. GALASSI, M. MAZZINI

Dipartimento di Costruzioni Meccaniche e Nucleari,
Universita degli Studi, Pisa, Italy
Abstract

The paper discusses an experiment performed in PIPER-ONE facility which simulates a

General Electric BWR-6 with volume and height scaling ratios of 1/2200 and 1/1,
respectively. The apparatus was properly modified to test the thermalhydraulic characteristics
of an isolation condenser-type system. This system consists in an once-through heat
exchanger, immersed in a pool with water at ambient temperature and installed at about 10 m
above the core.
The analysis of a first test on isolation condenser behaviour, named PO-SD-8, performed in
1992 showed that the RELAP5 code predicts well the overall thermohydraulic behaviour, but
discrepancies were identified in predicting local phenomena occurring in the pool and in the
isolation condenser. Therefore a second test, (named PO-IC-2), has been performed with
improved instrumentation.
The paper describes the results obtained in test PO-IC-2 and discusses the capabilities of
the Rclap5/Mod3.l code.
1. INTRODUCTION

Innovative reactors (essentially AP-600 and SBWR) are characterized by simplification in

the design and by the presence of passive systems for emergency core cooling. Experimental
and theoretical researches are needed to qualify the new components introduced in the design
and to characterize the thermalhydraulic scenarios expected during accidents. Available system
codes arc not retained suitable to evaluate the thermalhydraulic performances of the new
systems, especially in case of long lasting transients evolving at low pressure /!/.
In the frame of the activities carried out at University of Pisa related to the analysis of
thermalhydraulic situations of interest to the mentioned reactors (e.g. refs. 121 and /3/), three
series of experiments have been carried out utilizing the PIPER-ONE facility. They were
aimed at the experimental investigation of the behaviour of systems simulating the main
features of the Gravity Driven Cooling System (GDCS, first scries of experiments, rcfs. /4/)
and of the reactor pressure vessel Isolation Condenser (1C), (second and third" series of
experiments, rcfs. /5/ to 111 and rcf. /8/, respectively). At the same time Rclap5/mod2 and
mod/3 codes /9/, /10/ have been extensively applied as best estimate tools to predict the
transient scenarios of both SBWR and AP-600 reactors (sec also rcfs. /I I/ and /12/).
PIPER-ONE is a General Electric BWR experimental simulator specifically designed in the
early '80 to reproduce small break LOCA transient scenarios (e.g. rcfs. /13/ to /15/).
The above mentioned experiments were essentially devoted to a qualitative investigation of
the thermalhydraulic conditions typical of the new components foreseen in the abQyc reactors
(essentially SBWR) and setting up a data base suitable for code assessment. The distortions
that characterize PIPER-ONE hardware, in comparison with a scaled loop simulating SBWR,
completely prevent applications of the measured data to reactor conditions: the core and
downcomcr heights, as well the distance between the Bottom of Active Fuel (BAF) and the 1C
top arc important parameters not considered in the model.
279

On the other hand, Rclap5/mod2 and mod3 arc worldwide known codes developed at

Idaho National Engineering Laboratory (USA). The aforementioned applications to innovative

reactors scenarios emphasized the codes inadequacies in producing reliable results when

system pressure attains values below 0.5 MPa. Numerical deficiencies, limited ranges of
validity of the utilized correlations and lack of user experience (i.e. difficulty to develop a
suitable code use strategy) are retained mostly responsible of this situation.
The purposes of the activity described in the present document, that is the direct follow up
of the post-test evaluation of the previous 1C experiment PO-SD-8, (ref. /?/), are essentially:
a) to give an outline of test results obtained during the high pressure part of the test PO-IC-2,
also related to the study of the isolation condenser performance;

b) to evaluate the capabilities of the latest version of Relap5/mod3 code (i.e. version 3.1) in
reproducing the experimental scenario, giving emphasis to the attempt of fixing the code
limitations and the criteria for the optimal use of the code.
c) to demonstrate possible improvements in the latest code version with respect to the
previous one especially in relation to condensation heat transfer.
In order to achieve these objectives the original Relap5/mod2 nodalization of the PIPERONE apparatus /16/, also modified to perform pre-test and post-test calculations of the test
PO-SD-8 (e.g. refs. Ill and /17/), was completely renewed. In particular, the number of nodes

was roughly doubled also considering the available computer code capabilities. This
nodalization was qualified by the calculation of the test PO-SD-8 and was used extensively for
stability analysis in PIPER-ONE loop, (ref. /18/).
Some tens of sensitivity calculations were performed in the present context to identify some
uncertain boundary conditions in the experiment (e.g. spatial distribution of heat losses to the
environment) and to optimize the user choices, like the countercurrent flow limiting option in
the annular region of steam separator to predict liquid deentrainment from the mixture flowing
out from the core.
2. EXPERIMENTAL FACILITY

The PIPER-ONE apparatus is an integral test facility designed for reproducing the
behaviour of BWRs in thermalhydraulic transients, dominated by gravity forces.
The ENEL BWR plant installed at Caorso (I) was formerly taken as the reference prototype in
the design of the apparatus. The reactor is a General Electric BWR-4 plant, equipped with a

Mark II containment, but it has some features of the latest GE design (e.g., 8x8 fuel rod
assembly). The BWR-6 plant, equipped with 624 fuel bundles was assumed as reference for
the first test carricd-out on the PIPER-ONE facility, chosen by OECD-CSNI as ISP 21 /19/;
then, the latter was used as reference plant for all the LOCA tests, which had already been
performed.

The apparatus is constituted by the main loop, the ECCS simulators (LPCI/CS, HPCI/CS)
and the systems simulating ADS, SRV and steam line, as well as the blow-down line.
Nine zones can be identified in the main loop: lower plenum, core, core bypass (outside the

core), guide tube region, upper plenum, region of separators and dryers, steam dome, upper
downcomcr, lower downcomcr and jet pump region.
The volume scaling factor is about 1/2200, while the core cell geometry and the
piczomctric heads acting on the lower core support plate arc the same in the model and in the
reference plant.

The heated bundle consists of 16 (4x4) indirectly heated electrical rods, whose height, pitch
and diameter arc the same as in the reference plant. The maximum available power is about

320 kW, corresponding to 25% of scaled full power of the reference BWR.

280

The one-dimensionality as well as the overall simplicity of the apparatus- .have to be

highlighted; this is the direct outcome of the main objective of the research. In fact, the
primary circuit was designed in such a way to have as far as possible:
- one dimensional cylindrical volumes (nodes in code calculations);
- connections between adjacent nodes clearly defined (geometric discontinuities, Venturi
nozzles, orifices);
- lack of items (such as pumps and control systems) which can originate confusing situations
in code calculations.
The instrumentation system has features consistent with the fundamental philosophy of the
facility design. The data acquisition system can record 128 signals, with a frequency of up to
10 Hz for each signal.
As already mentioned, the facility hardware was modified by inserting the 1C loop, which
can operate at the same pressure of the main circuit (Fig. 1).
The main component of the isolation condenser loop is a heat exchanger consisting of a
couple of flanges, that support 12 pipes, 22 mm outer diameter and 0.4 m long; it is immersed
in a tank of 1 m^ volume, containing stagnant water, located at 4th floor of the PIPER-ONE
service structure. The heat exchanger is connected at the top and at the bottom respectively
with the steam dome and the lower plenum of the main loop. In order to enhance natural
convection inside the pool, a sort of shroud has been installed that divides the pool into two
parts.
The isolation condenser loop is instrumented with a turbine flow-meter and a differential
passive transducer on the hot side; a series of almost 30 thermocouples in various position of
the 1C and of the pool as shown in Fig. 2, complete the 1C instrumentation.
Hardware restrictions preclude the possibility to have a system correctly scaled with respect
to those provided for the new generation nuclear reactors, particularly the GE SBWR. This is
evident from Table I, where relevant hardware data characterizing the isolation condenser
loop installed in PIPER-ONE facility are compared with SBWR related data. In particular the
item "ratio" of Table I demonstrates that the heat transfer area of isolation condenser in
PIPER-ONE is roughly five times larger than the ideal value. The distance between bottom of
the active fuel and the isolation condenser and the height of the core itself are two of the most
important parameters differentiating PIPER-ONE from SBWR. These essentially prevents any
possibility of extrapolating PIPER-ONE experimental data to SBWR.

LP BOTTOM

ZERO LEVEL

Fig. 1 - Connection between isolation condenser loop and PIPER-ONE loop.

281

tube A
cut A-A

OJ-A

tubes B.C

Fig. 2 - Sketch of isolation condenser pool with temperature measurement locations

Tab. I - Comparison between isolation condenser related data in PIPER-ONE and in SBWR.
QUANTITY

UNIT

PIPER-ONE

SBWRC)

RATIO <*>

[IPER-ONE/SBWR

IDEAL (*> VALUE

OF THE RATIO
PIltER-ONE/SBWR

Primary system
volume
Core nciEhl
Maximum
nominal
core

m-1

0.199

595.

1/2990

1/2990

3.710
0.250

2743
2000.

135/1
1/8000

1/1
1/2990

0.041 H

60.

1/1463

1/2990

0.206

0.1

2.06/1

1/1

mz

0.301

184

1/610

1/2990

m-l

1.512

OJ1

49/1

1/1

m->

0.0015

2334

1/1556

1/2990

00075

0.0038

1.97/1

1/1

m^/rCw

7.341

3066

239/1

1/1

8.64

24.75

1/2.86

I/I

0.020

00508

1/2.54

1/1

0.0023

0.0023

l/l

1/1

Mm

power
4

Value of 3'/. core

Mw
power
Ratio (3V. core Mw/nv1

power) /primary

system value
Isolation

condenser heat
transfer area t"*"4")
7

Isolation

condenser heal
transfer area over
8
9

primary system
volume
IscLiLMi concenter
volume
Isolation

condenser volume
10

over
primary
system volume
Isolation
condenser
heat
transfer area over

11

3V. core power

Height of
isolation
condenser top

related to bottom
of active fuel
12 Diameter of a
single tube
13 Thickness of a angjc

tic

282

(*-) non-dimensional
(*+) only tube bundlci
(*) 1C data have been taken from r
(-} related to BWR6

3. TEST DESCRIPTION

The experiment comprises two phases characterized by different levels of heating power
(40 and 75 kW, respectively, later indicated as phases A and B of the test). They correspond
to the scaled value of the core decay power and to the capability of heat rcmovat-by the 1C
device, found in test PO-SD-8.
The test specifications foresaw constant heating power for 5-MO minutes for both
experimental phases, in such a way that quasi-steady state conditions could be reached by the
main thermal hydraulic quantities (pressure, flowratcs, collapsed levels, etc.).
In the test, the primary circuit was pressurized at the specified value by single-phase natural
circulation; the heat source was provided by the core simulator and the structure heating
system. Then, liquid was drained from the primary circuit for establishing the test specified
liquid levels, roughly at steam separate top elevation. After some hundreds seconds of steady
conditions, with heat losses compensated by the heating cables, the test started .by supplying
power to the core simulator and opening the valves of the 1C loop.
As usual for the experiments in PIPER-ONE facility, the test was designed on the basis of
pre-test calculations performed by Relap5/Mod3 code.
The initial and boundary conditions measured during the two phases of test PO-IC-2 are
given in Tabs. II and III.

Tab. II - PIPER-ONE test PO-IC-2: initial conditions

PARAMETER
LP pressure
LP fluid temperature
Core level
Downcomcr level
1C line fluid temperature
1C pool fluid temperature

SIGN

UNIT

VALUE

PA-LP-1
TF-LP-1
LP-CC-1
LP-LD-1
TF-IC-1
TF-SC-1

MPa
C
m
m
C
C

5.1
262.5
11.9
11.9
17.5
17.5

Tab. Ill - PIPER-ONE test PO-IC-2: boundary conditions

PARAMETER OR EVENT

TIME

(s)

Test initiation
Power versus time
1C top valve opens
1C bottom valve opens
1C top valve closes
1C lop and bottom valve open
1C top and bottom valve close
End of lest

0.0
see Figs. 3 to 6
4.0
32.0
508.0
602.0
1106.0
1184.0

The most significant results measured during the phases A and B are reported in. Figs. 3 to 6.
Figure 3 shows the different phases of the transient; when the core power is 40 kW, the
primary pressure shows a slight decrease (0.3 MPa/min) demonstrating that the power
exchanged through the isolation condenser is greater than supplied power. Roughly a steady
condition is reached at 75 kW core power (phase B), with a decrease of primary pressure of
only 0.02 MPa/min. According to the results of test PO-SD-8, the overall power removed by
283

1W.

5.750

PlPCn-ONETttPO-IC.2A

??-

s * ~* *

0.000

5.000
4.750

3,
4.500

5
a

\-V-

<.2SO

60.

40.

'
.

3.750

?
~

^\

4.000

A^^

'

0.250

1"

0.150

0.100

a.

20.

400

1000

(00

(00

L-

I*-'

.tn

200

0.

o.ooo

200

0.050

-_s^^_.

^ /

. . ** . J,X-~

I .

. . . . . . . JJI*

PIPEfl-ONE Tl PO-IC-2A

"

200

300.

,'ivi^^w/A_v^>_i__<^^

u_

250.

200

400

(00

800

| _^j-**~v.*&.J

^^^

PIPER-OME Ttt PCMC-2A

XXX EiMTF.sc.i
III

/j,^^*VVfV"vr

^^ftyv-A^/^V^^^

"

,~ *.

80.0

f.

ISO.

"

'

II

too.

50.

0.

200

>l

r*'r

/c^i^iir^zzr

If

^^__ $ ______________ J__________

70.0

e
X.

(o.o

50.0

E i

I!

GT

*s

^
i i i i i i i i

s^r''

jT ' 2
hr'^

tu

40.0

*r

20.0
10.0

200
Tim* ()

Pip. 5 Measured and calculated trends of 1C tubes fluid tmperatur

along the nxls

1*

30.0

^y*

s"^'
*f

^>T

I ^w*^-*^v*v^A**s^TtA.
S^M

ICWTEUPflilOIOOOO

00.0

"' * 'j'

f+'X* '<_ -^.Viwjw^OM - . .

|^_,~* > ** rr-n ^**^

^.^

1200

YYY ETJATF.SC-i

1000

^ _____

"" '

yl

1000

110.0

EIZATF.IC-J
EBA1F.IC.4
ICMTEWF8I5OJOCCO
ICntEMPFSI JOWXX)
IC1EMPFM8XIOO

1
200.

....... >X.

Fig. 6- Measured and calculated trends of 1C pool fluid temperature

along th axis

XXX EUAtF.IC.i
YYY
HZ
III
AAA
M

Vwya
...

Tim (i)

Fig. 4 Measured and calculated trends of 1C outlet mass (low rate

\-Tr*T"~IT"T"

. .

.0 O^rt

1200

T1m (s)

2
g

tMUflOWJJSOOOOOOO.

.
2 0.200

3.500

350.

XXX tUAWWjc-i
YYY

0.050
80.

5.250

PIPER-ONE Tt>t PO-IC-2A

ZZZ ICWP1XNJ10000
XXX CtJAWHPOWtR

5.500

U.400

WWA-IP.1

200

400

(oo

800

1000

1200

m.

Fig. 3 ' Measured and calculated trends of pritnary pressure and power

the isolation condenser is of the order of 80 kW, corresponding to an average heat flux of
about 200 kW/m^ in the considered conditions. The condcnsatc mass flowratc, registered at
the drain line (Fig. 4), appears quite constant along the two phases of the test (the turbine was
initially blocked, but started to operate with about 3 minutes of delay from the opening of 1C
valve).
In Fig. 5, the fluid temperatures measured in three positions along the Isolation Condenser
heat exchanger arc compared with the steam temperature at 1C inlet (equal to the saturation
temperature corresponding to the primary system pressure); in particular, the bottom curve
gives an idea of the subcooling conditions attained by the liquid exiting from the heat
exchanger.
Finally, the curves in Fig. 6 essentially show the strong fluid temperature stratification in
the pool: in the upper zone temperature increases up to about boiling conditions at test end,
while the bottom part of the pool remains at ambient temperature during all the test (fluid
temperature less than 20 C).
4. RESULTS OF CODE APPLICATION
4.1 Adopted nodalization

The standard version of the Relap5/mod3.1 has been adopted as reference code in the
present study; necessary adjustments of initial and boundary conditions, within the
experimental uncertainty bands, have been done with this code version ("reference" code
version) aiming at getting the "base calculation". The adopted standard version of
Relap5/mod3.1 runs on an IBM RISC 6000.
Once the "base calculation" results have been obtained, previous code versions
Rclap5/mod2.5, Relap5/mod3 7J and Relap5/mod3 v80, running on Cray mainframe, IBM PC
486 and IBM RISC 6000 respectively, have been utilized, too.
The overall strategy of the performed analysis aimed at assessing the capabilities of the
"reference" code version and at identifying possible improvements with respect to previous code
versions; sensitivity calculations considering variations in the nodalization have been
performed in this framework.
Additional objective of the analysis was to confirm the explanation given for the reasons of
constant 1C flow when core power is varied.
Considering the above objectives, the acquired experience in the use of the latest RelapS
code versions, the code user guidelines, (rcf. /19/), and the criteria proposed in rcf. /20/ for
nodalization development and qualification, a new PIPER-ONE input deck has been
developed.
The sketch of the new nodalization is shown in Fig. 7. Actually this has been developed in
the frame of the BIP (Boiling Instability Program) proposal, (e.g. ref. /18/), and already used
in that frame for planning of experiments.
The main difference with respect to the original nodalization (e.g. rcf. /16/) lies in the
number of nodes, that is now roughly three times larger. Moreover, the hardware
modifications in the primary circuit of the facility have also been considered. In particular, the
fccdwatcr line has been added and the connection zone between downcomcr and lower
plenum and the region around the top of the separator have been slightly modified to better
reproduce natural circulation in the reference reactor.

285

Fig. 7 - Nodalization of PIPER-ONE loop

286

4.2 Analysis of post-test results

The prc-tcst results IT1L that have been the basis for the design of the PO-lC-2 experiment,
arc not comparable with the actual data because of the differences between specified and
actual values of the initial and boundary conditions. In particular, as already mentioned, the
electrical power supply to the fuel rod simulator was limited to less than 100 kw.

For these reasons, three phases of the post-test analysis can be distinguished: _
a. achievement of good comparison between calculated and predicted trends of available
primary circuit data;

b. definition of a reference calculation on the basis of the above activity;

c. execution of a series of sensitivity calculations aiming at reaching the stated objectives.
In the frame of the analysis at item a., boundary and initial condition values related to the
primary loop have been changed within the presumed experimental error bands, up to getting a
satisfactory comparison between predicted and measured time trends in the primary circuit itself.
Once this process ended (this required a somewhat large effort because of the relative low
power of the experiment), the reference calculation results were also compared with the
experimental values in the 1C system (phase b.).
The phase c. of the analysis consisted of five different steps (Tab. IV) aiming at the
evaluation of:
3 a) influence of code version;
3b) influence of selected initial/boundary condition values;
3c) user effects;

3d) nodalization effects;

3e) sensitivity of code results when a geometric/hydraulic relevant parameter is changed in the
input deck.
4.2.1 Steady state calculations (Phase a)

As already mentioned, the main objective of the calculations was to match the primary
circuit pressure trend.
GROUP

VARIED PARAMETER

3a)

USED CODE VERSION

INFLUENCE OF
CODE VERSION

RELAP5/Mod3 1

RELAP5/Mod3V7J

RELAP5/Mod3V80
REU.fi/Mixi2S

Tab. IV - Documented calculations

3b)
INFLUENCE OF
INITIAL AND BOUNDARY INITIAL AND
CONDITIONS
HOUNDARY CONDITIONS

EFFECT ON PREDICTED
1C PERFORMANCE
(TS4C-7) - (HTTEMPI5050SIO)
ICOj
iOO.s
1000.3
-169
- 1 7 1 -12
-146 - 0 1
-148
-04
-147
-148 - 1 7 1

-169
-129

171
-!2
- 1 3 3 -46

-169
-167
-169

171
-12
-167 -118
-169 - 1 1 4

-169

-17.2

-169
-171

-17 I
-12
-175
-124

-169
-106

-171 -12
-146 -84

-169
-17J

-17.2 -12
-17.5 -12J

-169

- 1 7 1 12

-43 7
-30

-44
34

-166

1 7 J 2 319

3c)

USER EFFECT

IN/OUT FLOW TO

1C CONTROL VOLUMES
3d)

MESH NUMBER OF 1C

NOOAHZATION EFFECTS PIPE WALL

10
20

NODE NUMBER OF 1C

PIPE AND CORRESPONDING

INTERNAL POOL ZONE

5
10

1C
PIPE
MATERIAL
THERMAL CONDUCTIVITY

3=)

STANDARD STEEL
STANDARD STEEL '2
HEATED DIAMETER IN THE

CHANGES OF RELEVANT EXTERNAL SIDE OF 1C PIPE

PARAMETERS
WALL
0005 m

0 I m
0001 m
LP-VALVE CLOSURE AT
300 s IN THE TRANSIENT

-37 3
0.3

287

The initial condition at the assumed "time zero" in the experiment was achieved at zero
core power and zero flows inside feedwater and steam lines and at core inlet; structures
heating systems were active to compensate heat losses to environment.
The calculations was carried out changing:
- the fluid temperature around the loop
- the heat losses to environment
- the downcomcr and core region levels
- the opening/closure time of 1C valves
4.2.2 Reference calculations results (Phase b)

Three minor discrepancies arc shown in the lower plenum pressure trend (Fig. 3):
at the transient beginning, owing to the initial fluid temperature stratification not correctly
considered in the code;
- during the interruption of electrical power given to the rods, mostly due to inadequate
consideration of heat input to the fluid from the structures heating system;
- during the second part of the test as a consequence of the above mentioned cause.
It should be noted nevertheless a good agreement between the experimental and the
calculate trends during the part of the test of main interest.
The mass flowratc across the 1C is quite well predicted by the code (Fig. 4); the zero flow
resulting from the turbine signal in the first 180 s is originated by a malfunction as already
mentioned.
The fluid temperature at the inlet and long the 1C axis arc represented in Fig. 5. The
agreement between measured and calculated trends is quite good.
The comparison of fluid temperatures at the 1C outlet, (Fig. 6), confirms that the overall
heat transfer to the pool is satisfactorily predicted by the calculation: this appears to be true
also considering separately the tubes region (where both condensation and heat transfer to the
pool from subcoolcd liquid take place) and the single tube region (where only heat transfer to
the pool from subcoolcd liquid takes place).
The data in Fig. 8 show an ovcrcstimation of surface temperature by the code: this is true
-

for all the surface temperatures. Considering that the overall power exchanged is quite well
predicted, one can deduce a code error in predicting the heat transfer coefficient either in the

inner surface, in the outer surface, either in both the surfaces. This conclusion is also
supported by the observation that measured and calculated surface temperatures are very close
when flowrate and exchanged power are nearly zero.
The good agreement between measured and calculated trends of pool temperature at
different axial elevations (Fig. 9) constitutes an independent proof of the agreement in the
overall exchanged power. Furthermore, the temperature increase in the pool results from
Fig. 10: the smooth increase of fluid temperature in the pool leads to a smooth decrease of the
temperature difference between tubes inner wall surface temperature and pool temperature
itself. Removed power remain constant: this means an average increase of the overall heat
transfer coefficient during the experiment.
4.23 Sensitivity calculations (Phase c)

Five main groups of calculations have been distinguished in Tab. IV, where the varied
parameter ranges arc reported, if applicable. In all cases six quantities have been selected to
characterize the influence of the varied parameter.
- lower plenum pressure;

- 1C tubes surface temperature (external side level 3);

- 1C tubes internal fluid temperature (level 3);
288

1C outlet fluid temperature;

tC HTC inside tubes (level 3);
1C HTC outside tubes (level 3).

300.

PIPER-ONE Test PO-IC-2A

Post-test calculation
RELAP5/Mod3.1

250.

XXX IC31HTTEMPS1S000101
YYY EI2ATS-IC-2

200.

150.

o>
Q.

E
0)

100.

50.

-200

200

400

600

800

1000

1200

Time (s)

Fig. 8 - Trends of 1C tubes internal surface temperature (top level)

120.

100.

PIPER-ONE Test PO-JC-2A

Post-test calculation
RELAP5/Mod3.1

XXX IC31TEMPF561010000
YYY IC31TEMPF562010000

222

E12ATF-SC-1

80.
o>
L*

E
o

60.

Q.

40.

20.

0.
-200

200

400

600

800

1000

1200

Time (s)

Fig. 9 - Trends of 1C internal pool fluid temperature (top level)

289

250.
PIPER-ONE Test PO-IC-2A
Post-test calculation

XXX IC31CALC
YYY EI2AEXP

RELAP5/Mod3.1

200.

150.

100.

<D
Q.

E
0)

50.

-50.

-200

200

400

600

800

1000 1200

Time (s)

Fig. 10 -Trends of temperature difference between internal side of 1C tubes and 1C pool fluid
(top level)

In the last part of Tab. IV, the differences between measured and calculated values of
surface temperature in the lowest node of the 1C tubes, are reported too. This is done at 100,
SOOjind 1000 s into the transient It can be noted that in almost all the cases calculated outer
surface temperature values are larger than measured values (the consideration about possible
thermocouples "fin" effect should be made here, too). Assuming that heat transfer power and
fluid temperature are the same in the experiment and in the calculation, it can be concluded
that calculated HTC values are lower than in the experiment.
Primary pressure is not very much affected by code version, while the unphysical behaviour
for fluid temperature, already noted in the analysis of the PO-SD-8 test (7J code version),
appears in the application of V80 version of the code. Internal heat transfer coefficient is
oscillating only in Relap5/Mod3.1.
A large number of boundary conditions has been varied: the influence of the considered
changes cannot be retained relevant as far as the 1C behaviour is concerned.
The studied nodalization effects include the increase in the number of conduction heat
transfer meshes and the increase in the number of hydraulic nodes. The effects of these
variations at a global level (primary circuit pressure) are negligible. However, local effects can
be very important as shown by the inner wall surface temperature and by the heat transfer
coefficient. Some calculations leads to the conclusion that heat transfer coefficient should be in
someway connected with node dimensions. The change of the hydraulic diameter in the 1C
secondary side has a strong effect on the primary loop pressure and on the 1C related
quantities.

290

5. CONCLUSIONS

The present document reports the analysis performed with RclapS in relation to the PIPERONE experiment PO-1C-2. The test is the follow-up of the similar experiment PO-SD-8 and
aims at characterizing phenomena connected with the operation of isolation condenser in a
geometric configuration typical for SBWR. The test studied was conducted at high pressure
(around 5 MPa) to study code performance in this situation.
Conclusion can be drawn in relation to:
A) overall system performance;
B) thcrmalhydraulic phenomena inside the 1C:
C) code capabilities;
D) effect of various changes in the calculation conditions.
It has been confirmed that a constant removed power characterizes the 1C performance
(item A)), whatever arc the primary loop thermalhydraulic conditions. From the analysis of
experimental data supported by specific code calculations, the reason why 1C exchanged
power remains constant when core power conditions are varied, has been made clear. This is a
consequence of 1C flowrate that is determined by the pressure difference created in the
downcomer of the main loop that remains almost constant during the performed experiment.
Constant 1C flowrate means constant condensation length in the 1C tubes and constant
transferred power to the pool.
The shroud put in the pool was not effective in provoking natural circulation because of the
high conductivity across its wall that led to fluid temperature stratification with the same
characteristics in the inner and outer zones of the pool.
Heat transfer coefficients arc several (3-7) times larger in the condensing zones of the 1C
tubes than in the single phase liquid region (item B)); the fluid pool temperature increase (up
to 70 K.) has a negligible role in this situation (high pressure steam). Liquid level in the 1C

remains almost constant when varying core power.

A new detailed nodalization of PIPER-ONE apparatus was adopted for this study. The
standard version of Relap5/mod3.1 is able to catch the overall phenomenology during the four
main periods of the experiment (item C)).

Discrepancies have been identified mainly concerning the 1C tubes surface temperatures
that are overestimated on the outer surface: this means underestimation by the code of the
outer heat transfer coefficient provided that the overall exchanged power is well predicted.
However the possible "fin" effect originated by thermocouples has not been considered in this
conclusion. An unphysically high heat transfer coefficient is produced in the output by the
code although apparently not used. During the period of power shut-off the code was not able
to simulate the condensation shock occurring the in 1C line.
An extensive series of sensitivity calculations has been carried out (item D)). These
demonstrated:
* some improvements in the code results when passing from the earliest to the latest
Relap5/mod3 code versions; however oscillations in condensation heat transfer value are
much larger in the reference code version. Furthermore the transition logics from heat
transfer mode 2 to 3 (and viccversa) could be improved;
* changes in the equivalent diameter of the pool side of the 1C has an important effect on the
calculation of the local quantities like heat transfer coefficient and temperatures;
* changes in nodalization can also have a noticeable effect as far as the calculation of the
above mentioned quantities is concerned. The last item stressed the need to define some

291

relationship between the (condensation) heat transfer coefficient and the average node
dimensions: this seems necessary to get reproducible results when condensation heat
transfer is involved.
Further activity in this area includes the analysis of a companion experiment during which
nitrogen gas was injected into the primary circuit and actually caused large degradation of the
1C removed power.

REFERENCES
/!/

121
/3/

/4/

/5/
161

111

D'Auria F., Modro M., Oriolo F., Tasaka K.: "Relevant Thcrmalhydraulic Aspects of
New Generation LWR's", CSN1 Spec. Meet. On Transient Two-Phase Flow - System
Thcrmalhydraulics, Aix-En-Provcncc (F), April 6-8, 1992.
D'Auria F., Galassi G.M., Oriolo F.: "Thcrmalhydraulic phenomena and code
requirements for future reactors safety analysis", Int. Conf. on Design and Safety of
Nuclear Power Plants (ANP) - Tokyo (J), October 25-29, 1992.
Andreuccctti P., Barbucci P., Donatini F., D'Auria F., Galassi G.M., Oriolo F.:
"Capabilities of the RELAP5 in simulating SBWR and AP-600 Thermalhydraulic
behaviour", IAEA Technical Committee Meet. (TCM) on Progress in Development and
Design Aspects of Advanced Water Cooled Reactors, Rome (I), September 9-12, 1991.
Bovalini R, D'Auria F., Mazzini M.: "Experiments of Core Coolability by a Gravity
Driven System performed in Piper-One Apparatus", ANS Winter Meeting, San
Francisco (CA), November 10-15, 1991.

Bovalini R., D'Auria F., Mazzini M., Vigni P.: "Isolation Condenser performances in
PIPER-ONE Apparatus", 1992 European Two-Phase Flow Group Meet. , Stockholm
(S), June 1-3, 1992.
D'Auria F., Vigni P., Marsili P.: "Application of Relap5/Mod3 to the evaluation of
Isolation Condenser performance". Int. Conf. on Nuclear Engineering (ICONE-2) - San
Francisco (US), March 21-24, 1993
Bovalini R., D'Auria F., Galassi G.M., Mazzini M.: "Piper-One research: the experiment

PO-SD-8 related to the evaluation of isolation condenser performance. Post-Test

analysis carried out by Relap5/Mod3-7J code", University of Pisa Report, DCMN - NT
200 (92), Pisa (I), November 1992.
/8/ D'Auria F., Galassi G.M:, Mazzini M., Pintore S.: "Ricerca Piper-One: specifiche
dettagliate della prova PO-IC-2", University of Pisa Report, DCMN - NT 234(94), Pisa
(I),Giugnol994.
/9/ Ransom V.H., Wagner R.J., Trapp J.A., Johsen G.W., Miller C.S., Kiser D.M., Riemke
R.A.: "Rclap5/mod2 Code Manual - Vol. 1: code structure, systems and solution
methods", NUREG/CR-4312 EGG 2396 - EGG Idaho Inc., March 1987.
/10/ Carlson K.E., Riemke R.A., Rouhani S.Z., Shumway R.W., Weaver W.L.:
"Rclap5/Mod3 code manual - volume II. User guide and input requirements",
NUREG/CR-5535, June 1990.
/I I/ D'Auria F., Oriolo F., Bella L., Cavicchia V.:"AP-600 Thcrmalhydraulic
Phenomenology: A Rclap5/mod2 Model Simulation", Int. Conf. on New Trends in
Nuclear System Thcrmohydraulics - Pisa (I), May 30 - June 2 1994, Int. Top. Meeting
on Advanced Reactor Safety - Pittsburgh (PA) April 17-21, 1994

292

I\2I Bartmcci P., Bella L., D'Auria F., Oriolo F.: "SBWR thermalhydraulic performance: a
RELAP5/MOD2 modd simulation", 6th Int. Top. Meet, on Nuclear Reactor
Thcrmalhydraulics, Grenoble (F), Oct. 5-8 1993
/13/ Bovaltni R., D'Auria F., Di Marco P., Galassi G.M., Giannccchini S., Mazzini M.,
Mariotti F., Piccinini L., Vigni P.: "PIPER-ONE: a facility for the simulation of
SBLOCA in BWRs", Spec. Meet, on Small Break LOCA Analyses in LWRs, Pisa (I),
June 23-27 1985
/14/ Bovalini R., D'Auria F., Mazzini M., Pintorc S., Vigni P.: "PIPER-ONE Research:
overview of the experiments carried out", 9th Conf. of Italian Society of Heat
Transport, Pisa (I), June 13-14 1991

/15/ Bovalini R., D'Auria F., Mazzini M., Pintorc S., Vigni P.: "PIPER-ONE Research:
Lesson learned", 9th Conf. of Italian Society of Heat Transport, Pisa (I), June 13-14
1991
/16/ D'Auria F., Fruttuoso G.:" OECD CSN1 ISP 21, PIPER-ONE test PO-SB-7: post-test
analysis performed at Pisa University by RELAP5/MOD2 code", University of Pisa
Report, DCMN - RL 386 (89), Pisa (I), March 1989, OECD CSNI 2nd Workshop on
ISP 21, Calci (I), Apr. 13-14 1989
/17/ D'Auria F., Galassi G.M., Mazzini M.: " PIPER-ONE research: specifications of POSD-8 experiment" (in Italian), University of Pisa Report, DCMN - NT 190 (92), Pisa
(I), Jan. 1992
/18/ Ambrosini W., D'Auria F., Galassi G.M., Mazzini M., Virdis M: "Preliminary planning
of BIP tests to be performed in Piper-One apparatus", University of Pisa Report, DCMN
- NT 207 (93), Pisa (I), May 1993.

/19/ D'Auria F., Mazzini M., Oriolo F., Paci S.: "Comparison Report of the OECD/CSNI
International Standard Problem 21 (PIPER-ONE Experiment PO-SB-7)", CSNI Report
Nr. 162, Paris (F), Nov. 1989
/20/ Bajs T., Bonuccelli M., D'Auria F., Debrecin N., Galassi G.M.: "On Transient
qualification of LOBI/MOD2, SPES, LSTF, BETHSY and KRSKO Plant nodalizations
for RELAP5/MOD2 code", University of Pisa Report, DCMN - NT 185(91), Pisa (I),
Dec. 1991
/21/ Brandani M., Rizzo F.L., Gesi E., James A.J.: "SBWR-IC & PCC System: an approach
to passive safety", IAEA Technical Committee Meet (TCM) on Progress in
Development and Design Aspects of advanced Water Cooled Reactors, Rome (I),
September 9-12,1992
1221 Bovalini R., D'Auria F., Galassi G.M., Mazzini M.: "Post-Test analysis of PIPER-ONE
PO-IC-2 experiment by RELAP5/MOD3 codes", University of Pisa, DCMN Report, NT
254 (95), March 1995.

293

COMPUTER MODEL DEVELOPMENT AND VALIDATION

(Session V)
Chairman
A. LERTOON
France

ANALYSIS OF PACTEL PASSIVE SAFETY

INJECTION TESTS WITH RELAP5 CODE

R. MUNTHER, J. VIHAVAINEN
Lappeenranta University of Technology,
Department of Energy Technology
J. KOUHIA
Technical Research Centre of Finland,
Nuclear Engineering Laboratory

Lappeenranta, Finland
Abstract

The gravity driven emergency core cooling (ECC) systems are utilized as important components
of passive safety coolant systems of advanced reactors. Most of the published investigations
have been primarily concerned with the presentation of new concepts, a few of their computational analysis and even fewer studies have been addressed to the experimental investigation of
these systems.

PACTEL (Parallel Channel Test Loop) is an experimental out-of-pile facility designed to

simulate the major components and system behavior of a commercial Pressurized Water Reactor
(PWR) during different postulated LOCAs and transients /!/. The reference reactor to the
PACTEL facility is Loviisa type WER-440. The recently made modifications enable experiments to be conducted also on the passive core cooling. In these experiments the passive core
cooling system consisted of one core makeup tank (CMT) and pressure balancing lines from the
pressurizer and from a cold leg connected to the top of the CMT in order to maintain the tank
in pressure equilibrium with the primary system during ECC injection. The line from the
pressurizer to the core makeup tank was normally open. The ECC flow was provided from the
CMT located at a higher elevation than the main part of the primary system. A total number of
nine experiments have been performed by now.

A preliminary series of experiments with gravity driven core cooling was conducted with
PACTEL facility in November 1992 /2/. The simulated transient was a small break loss-ofcoolant accident (SBLOCA) with a break in a hot leg. In order to investigate this behavior more
precisely, a second series of experiments with an improved instrumentation of the facility was
performed in November 1993 with a small break in a cold leg. In these tests a rapid condensation of vapor interrupted the emergency core cooling flow several times. The tests indicated also
that steam condensation in the CMT can prevent continuous ECC and even lead to partial core
uncovery. The initiation of condensation and the thermal stratification in the CMT were found
very difficult to model in the RELAP5 analysis of the experiments.
However, it should be underlined that these tests presented here are not directly applicable to the
safety analyses of any suggested design, because of the major differences in the geometry
between these concepts and PACTEL. Our objective has been only to simulate the gravity
driven ECC and thus to enhance the understanding of the physical phenomena important in
passive safety systems working with low differential pressures.

297

1. INTRODUCTION
Along with the normal evolution in LWR reactor designs several new interesting concepts have
been presented. These ALWR designs aim at plant simplifications and safety and operability

improvements. The principal tool being used to achieve a safer and simpler reactor is the use of
passive system designs. Unfortunately, it is not easy to confirm that passive safety systems
operate as intended under all the relevant conditions. More work is needed to evaluate the extent
of improvements in safety which can be realized.

This paper provides the presentation of gravity driven emergency core cooling experiments with
PACTEL, their analysis and discussion of the phenomena related to the experiments. The
recently made modifications enable experiments to be conducted also on the passive core
cooling. Firstly this paper describes the experimental facility, the modifications needed for
Gravity driven ECC tests and test conditions. Secondly the RELAP5 represantation of PACTEL
is described. Then the experimental results are compared to the calculational results. Finally
conclusions are drawn both from the experiments and the their modelling with RELAP5.
2. PACTEL FACILITY

The PACTEL facility simulates the major PWR components and systems during small- and
medium-size break LOCAs. The facility consists of a primary system, the secondary side of the
steam generators, and emergency core cooling systems (ECCS). The reactor vessel is simulated

by a U-tube construction including downcomer, lower plenum, core and upper plenum.
The facility is a volumetrically scaled model of the 6-loop WER-440 PWR (The Finnish
Loviisa plant being the reference) with three separate loops and 144 full-length, electricallyheated fuel rod simulators arranged in three parallel channels. The fuel rod simulators are heated
indirectly.
The reference reactor has certain unique features differing from other PWR designs. The WER-

440 has six primary loops with horizontal steam generators. Due to the construction of the steam
generators the driving head for the natural circulation in small break LOCAs is relatively small.
The primary loops have loop seals in both the hot and cold legs. The loop seal is a U-shaped
bend in the leg piping connecting the steam generator to the pressure vessel. It is interesting to
note in the current context that the basic design of the VVER-440 reactor exhibits certain
inherent safety features that are again found by the new designs. The reactor core has low power
density and the primary circuit water inventory is large relative to the power. These characteristics ensure smooth behaviour in transient conditions.

Volumetric scaling (1:305) preserving the elevations has been applied in the PACTEL design.
Maintaining system component heights and elevations is important for realistic simulation of
small break and natural circulation transients. The main characteristics of the facility are
presented in Table I.

298

Table L Main characteristics of PACTEL facility.

Reference power plant
Volumetric scaling factor
Scaling factor for elevations
Number of primary loops
Maximum heating power
Number of fuel rod simulators

WER-440
1:305
1:1
3
1 MW
144

Outer diameter of fuel rods

Heated length of fuel rods

9.1 mm
2420 mm

Axial power distribution

Axial peaking factor
Maximum temperature of fuel rods
Maximum primary pressure
Maximum operating temperature
Maximum secondary pressure

chopped cosine
1.4
800 C
8.0 MPa
295 C
4.65 MPa

3. PACTEL MODIFICATIONS AND TEST MATRICES FOR PASSIVE CORE COOLING

EXPERIMENTS

The passive core cooling system used in the experiments consists of one core makeup tank and
pressure balancing lines from the pressurizer and from a cold leg connected to the top of the
core makeup tank in order to maintain the tank in pressure equilibrium with the primary system
during injection. The line from the pressurizer to the core makeup tank is normally open. The
core makeup tank is located above the reactor coolant loops and steam generators, so the motive
force for injection is the gravity head, Figure 1. The makeup tank used limits the primary
pressure to 5 MPa in the experiments. Since the PACTEL facility is not a model of any of the
proposed passive ALWR designs, the modifications in the facility are intended only to simulate
the gravity driven flow to the primary system. Neither automatic depressurization system (ADS)
nor special valves are simulated. The primary system is depressurized from the pressurizer relief
valve.
3.1 First series of experiments
The first stage of each experiment involved heating the facility to the proper temperature. Before
the tests the core power was set to 80 kW corresponding to 1.8% of the 1375 MW thermal
power of the Loviisa reactor. The fluid temperature and pressure reached a quasi steady state
near 220 C and 40 bars and at this point the pressurizer heater power was reduced to 2 - 4 kW.
These conditions were maintained for about half an hour permitting the fluid to attain a more
uniform temperature and allowing the heat losses through flanges and support structures to
approach an equilibrium. The SG feedwater injection was adjusted manually to keep the water
level in the SGs constant Because of the large water inventory on the secondary side no fast
automatic control was needed. Before each experiment, the CMT was filled to the top with
water at a temperature and pressure of about 40 C and 38 bar, respectively.

The experiments were started by opening the break simulation valve in hot leg number 1 at time
t = Os. Three different break sizes (0 2,4 and 6mm) were used. Simultaneously with the break
valve opening, the ECC line valve and the cold leg PEL valve were opened. The power of the
299

pressurizer heaters was turned off. The first two tests, GDE01 and GDE02, were terminated
when a rapid condensation of vapor in the CMT vapor space depressurized the CMT. Check
valves prevented the collapsed vapor space in the CMT to be filled with liquid drawn from the
ECC line. In order to investigate the flow restrictions in the ECC line the armature of the line
was varied during the three first tests. Neither the primary system nor the secondary system
were depressurized by the operator in the GDE01 and GDE02 tests.

In tests GDE03, GDE04, and GDE05 the secondary side valve was also opened. The primary
system was depressurized in stages through the pressurizer relief valve before the anticipated
CMT flow interruption in the GDE03 and GDE04 tests. For the large 6 mm (in dia.) break in
the GDE05 test no extra depressurization was needed. These three tests were terminated when
a thermal hydraulic status quo and a low pressure level was reachedThe first series consisted
of five experiments, Table n.
PEL

PEL

Figure 1. The passive ECC system of PACTEL.

300

TABLE II
TEST MATRIX/PHASE 1

DESCRIPTION

FORESEEN PHENOMENA/AIMS

NOTES

GDE01/

Investigation of the btowdown

process at low pressure. Gravity
driven ECC effectiveness. Flow
regimes. Phase separation and
stratification. Condensation. Break
flow. Coolant distribution.

for all of PCC-tests pressure - 40

bars. Steady-state initial conditions.

GDE02/
2.0% (4 mm) hot leg
break with gravity
driven ECC.

As above.

Initial conditions as above. Minimized ftow restrictions in the ECC

line.

GDE03/

Effect of primary side depressurization.

low power
(< 80 kW) 2.0% (4
mm) hot leg break
without ECC.

2.0% (4 mm) hot leg

break Gravity driven
ECC with sudden
depressurizatJon of
primary system.
GDE04/
2.0 % (2 mm) hot leg
break Gravity driven
ECC with sudden
depressurizatJon of
primary system.

Effect of break area.

GDE05/
4.4% (6 mm) hot leg
break. Gravity driven
ECC

As above.

As above. Primary and secondary

side depressurization. Optimized
flow restrictions in the ECC line.

As above.

As above. No depressurizations.

3.2 Second series of experiments

The gravity-driven ECC behaviour was investigated mote in the second phase of the tests with
particular emphasis on break location, pressure reductions, reproducibility of the condensation
maneuvered experiments and system operation for the case of a small break LOCA. The major
parameters and phenomena of concern during experiments are the break mass flow rate and the
associated total primary coolant mass inventory, coolant distribution, different types and
alternate paths of natural circulation in the loops, condensation and related heat transfer
characteristics. For the second set of experiments the instrumentation of the facility was
improved. In order to investigate the temperature stratification in the CMT ten thermocouples
were installed to the upper part of the CMT, Rg 2. The water level in the CMT was measured
with a pressure difference transducer. One loop of the three loop facility was isolated. When
301

compared to the first series of experiments, the main differences are that the second series was
carried out with two active loops, insulated PBLs and an improved instrumentation in the CMT.

Fig. 2. Temperature measurement in the CMT tank.

302

The experiments were started by opening the break simulation valve in cold leg number 1 at
time t = Os. Two different break sizes (0 4 and 2mm) were used. Simultaneously with the break
valve opening, the ECC line valve and the cold leg PEL valve were opened. The power of the

pressurizer heaters was turned off. The first two tests, GDE11 and GDE12, were terminated by
operator at t= 3000s. Neither the primary system nor the secondary system were depressurized
by the operator in the GDE11 and GDE12 tests.
In test GDE13 the secondary side valve was also opened and the primary system was depressurized in stages through the pressurizer relief valve before the anticipated CMT flow interruption.
This test was terminated at t=2000s by the operator.

For the small 2 mm (in dia.) break in the GDE14 test no depressurization was used. A high
water level in the pressurizer was used in the initiation of the test in order to achieve circulation
through the CMT in the early stage of the transient The test was interrupted immediately after
the condensation initiation at t= 1170s. The second phase of the experiments was performed in
November, 1993. The test matrix is presented in Table ILL
Table HI Test matrix for the second series of passive core cooBng experiments with the PACTEL facility.

TEST MATRIX/PHASE II

TEST/ DESCRIPTION

FORESEEN PHENOMENA/AIMS

NOTES

GDE11/

Gravity driven ECC effectiveness.

How regimes. Phase separation and
stratification. Condensation. Break
flow. Coolant distribution.
Investigation of the btowdown
process at low pressure. Break
location.

For all PCC-tests pressure = 40

bars. Steady-state initial conditions.
No depressurization of the primary
or the secondary system.

As above. ReprodudbiHty.

As above. Initial conditions as well

tow power
(<80kW), 2.0%
(4 mm) cold leg
break with
gravity driven
ECCS.
GDE12/
2.0% (4 mm)

as possible same as above.

cold leg break

with gravity
driven ECCs.

GDE13/
2.0% (4 mm)
cold leg break
with gravity

Effect of depressurization.

As above. Depressurization of the

primary and the secondary systems.

CMT natural circulation.

As above without depressurizations.

High pressurizer level at test initiation.

driven ECCs.
Sudden depressuriza-tion of
the primary
system.
GDE14 /
0.5% (2 mm)

cold leg break

with gravity
driven ECCs.

Effect of break size.

303

4. THE RELAP5 REPRESENTATION OF PACTEL

A base RELAP5/Mod3 input deck for PACTEL was modified to include the gravity driven
emergency core cooling system. The additions included the CMT and associated pressurizer and
cold leg pressure balancing connections. The model was composed of 257 hydrodynamic volumes, 284 junctions, and 394 heat structures. Although this input served as a starting point for

the calculations, many modifications were made to it during the course of the analysis.
Revisions were made to the original model as new information became available and as input
deficiences were discovered. Those modifications that were expected to have the most effect on
these calculations, and the corresponding input changes are discussed next

In the CMT, modelled as a cylinder, the effect of nodalization was investigated by changing the
number of CMT nodes. These calculations showed that there was no significant difference
between 2, 5 and 10 node CMT models for the overall CMT behavior. However, the amount of
rapid deprcssurizations of the CMT varied between 2, 5 and 10 node models and none of the
models corresponded to the amount or timing of the depressurizations in the tests. The results
with a CMT modelled as a branch did not give any prediction for rapid pressure drops in the
CMT. In the analysis of the first test series the presented calculations were made with 5 node
model. For the calculations of the second series a 30 node model of the CMT was used. No
condensation was then observed in the simulations.
The junctions between the cold legs and the downcomer, and between the upper plenum and the
hot legs were at first modelled as crossflow junctions, but later modified as normal junctions in

order to achieve realistic flow paths and water levels in the upper plenum and the upper part of
the downcomer. The modelling of these junctions also had an effect on the heat loss distribution
in the primary system and this way to the primary pressure when the coolant flow was near

stagnation.
The subcooled discharge coefficient at the break was also varied for a better presentation of the
leak mass flow of the experiments.
The nodalization schemes for both experimental series are presented in the Appendix I.
5. RELAP5 ANALYSIS OF THE FIRST TEST SERIES

The test results from the transients performed in the PACTEL loop were compared to computer
simulations by the RELAP5/Mod3 program /3/. The actual starting steady state conditions in
individual tests were used as input to the computer simulations. All the calculated transients
began with the opening of the break valve. Also the ECC line valve and the cold leg PEL valve
were opened simultaneously. Condensation of steam in the CMT was observed in all experiments.
In the calculation of the GDE01, there were five rapid pressure peaks against the measured one
at 1860 s, Fig 3. The experiment was terminated after this. It was found that changing the
maximum time step had an effect on the peak appearance. On the other hand, RELAP5 changed
the flow chart from vertically stratified flow to bubbly flow in the CMT at the initiation of

condensation. Also the pressure of the pressurizer, the ECC flow and the vapor content of the
upmost CMT node increased at the condensation initiation.

304

40

35
30

J.

EXPERIMENT

25

LU

CC

=D
CO
CO

LU
DC
Q.

20

15
10
5
0
500

1000 1500 2000

TIME [S]

Fig. 3. The CMT pressure in GDE01

The best approximation for the condensation induced pressure peaks was achieved in the
modelling of GDE03 experiment, where also the oscillatory period after the condensation was
modelled, Fig 4. However, there were extra pressure peaks also here.

40

35

30

EXPERIMENT

25
LU

CC

ID
CO
CO
LLJ
CC
Q_

20
15
10
5
0
500

1000 1500 2000

2500

3000

TIME [s]

Fig. 4. The CMT pressure in GDE03

305

It was also found that the modelling of pressure losses in the PBLs had a significant effect on
CMT depressurization behavior. Unfortunately no measured data was available for pressure
losses in the PBLs. A sensitivity study on pressure losses in the cold leg PEL, pressurizer PEL,
and the ECC line was performed and it was found that depressurization modelling was very
sensitive especially for the value of pressure loss in the cold leg PEL.
6. RELAP5 ANALYSIS OF THE SECOND TEST SERIES
In the second series of experiments condensation behavior differed a lot from that observed in
the preliminary tests. As the ECC flow in the first tests stopped totally several times because of
rapid and very short condensations there was now only one condensation phase which lasted
much longer. Good reproducibility was achieved in GDE11 and GDE12 test. The CMT
pressures in GDE11 and GDE12 tests are shown in Fig. 5. In both experiments there was a
condensation phase starring at about 1700s and lasting for 300s.

During the long condensation period in the GDE11 and GDE12 experiments the water level
decreased to the top of core and even slightly below. The uncovery lasted only a short time and
no significant heat-up in the core was found. In the GDE13 and GDE14 experiments no core
uncovery was found.
GDE11/12
45

PCMT EXP 11

40

35
3O

I3
uj 2

o_
15

1O
5
O

5OO

10OO

15OO

2OOO

25OO

3OOO

TIME [s]

Fig. 5. The CMT pressures in GDE11 and GDE12 tests

In the RELAP simulation of the GDE11 experiment no condensation was observed. However,
the overall modelling of the transient is rather good. Fig. 6 shows the measured and calculated
CMT pressures in the GDE11 test A thick 30 node model of the CMT was used in all
simulations of the second series. With the thicker nodalization of the CMT a better presentation
of temperature stratification in the CMT was achieved.
306

GDE11_PACTEL
45
EXPERIMENT
RELAP

4O
35

20
15

10

5OO

1000

15OO

2000

25OO

3OOO

TIME [s]

Fig. 6. Measured and calculated CMT pressures in GDE11 test

The operator activated primary system depressurization in stages affected to the total collapse
of the vapour space, and in the GDE13 test there were three short condensations observed in the
CMT, Fig. 7.
GDE13 PACTEL

45
EXPERIMENT
RELAP

4O
35
_3O

20
OL
"***^***

15
1O
5
O
50O

1OOO

1500

20OO

TIME [s]

Fig. 7. Measured and calculated CMT pressure in GDE13

The first condensation was already at t = 1100s straight after the depressurization initiation.
Similar period of short condensations were observed in the experiments of first series in both
experiments with or without depressurizations.
307

A very steep vertical temperature gradient was formed inside the CMT in all tests. Rg. 6. shows
that the temperature difference just before the condensation in the GDE11 experiment was 180
K in a water layer 0.15 m thick (the thermocouple numbering corresponds to that shown in Hg.
2.). The measured results are shown in Hg. 8. After the condensation at t= 1700s hot water is
sucked to the CMT from the cold leg PBL mixing and filling the CMT. The CMT is repressurized after a new stratification of temperature is formed inside the CMT. With RELAP only the
temperature stratification before condensation can be modelled, Fig. 9.
GDE11

25O

EXPERIMENT
TFCMT 1
TFCMT 2
TFCMT 3
TFCMT~4
TFCMT~5

500

1000

15OO
TIME [s]

2000

25OO

3000

Fig. 8. Measured temperature distribution in the CMT at GDE11 test

GDE11

RELAP

24O
CMT Tf1
CMT~Tf2

CMT Tf3

5OO

10OO

15OO
TIME Is]

2OOO

2500

Fig. 9. Calculated temperature distribution in the CMT at GDE11 test

308

3000

An effort for preventing the rapid condensation was done by carrying a thick, insulating level
of hot water to the CMT with a natural circulation loop formed between the CMT and the
primary system via the cold leg PEL and ECC line. For this reason the water level in pressurizer was set high and a small break size was chosen at the GDE14 test initiation. This natural
circulation phase of the CMT was also in the ROSA-V/LSTF experiment /4/. With these
preconditions a short natural circulation phase was then observed in the GDE14 experiment.
However, this natural circulation phase was not effective enough to form a sufficient layer of
hot water in the CMT. In PACTEL the total water volume above the CMT is small since there
are horizontal steam generators.
7. CONCLUSIONS

No core uncovery was found in any of the tests of the first series. However, the emergency core
cooling flow from the core makeup tank was stopped when rapid condensation collapsed the
core makeup tank pressure. The tank repressurized rather quickly and the emergency core
cooling flow was provided until the next condensation phase.
In the second series of experiments only two of the three loops of the facility were used as in
the first series of experiments all the loops were active. The break was now located to the cold

leg and two different break sizes were used. In one of the tests both the primary system and the
secondary system were depressurized. In all the four experiments performed steam was flowing
into the CMT and then later condensed to the cold water of the CMT. There were striking
changes in the vertical temperature gradient of the CMT. It was experienced that condensation
was then initiated easily by steam or water flow from the PBLs as the steep stratification in the
CMT was broken. Especially the changes in water level in the pressurizer seemed to be
responsible for most of the condensation periods.
We also simulated the both experimental series gravity driven core cooling experiments with
RELAP5/mod3.1. The comparison of calculations and experiments show a good agreement both
in magnitude and time of occurrence for most of the different physical events. The main
observed discrepancy was due to limitations in the RELAP5 code to accurately predict rapid
condensation in the CMT. The most critical aspect in the calculational results was that the
appearance of condensation was dependent also on computational features, such as the time step
and the nodalization.

From the results presented, we conclude that: 1) condensation modelling of the PACTEL
experiments can not be satisfactorily achieved with the modelling capabilities of the current
version of RELAP5. The forthcoming versions of the code should be equipped models where
stratification and the possibility for the breaking of stratified layers can be evaluated. 2) the
thermal hydraulic balance in the system is very sensitive and therefore experimental data should
be used in the calculations whenever possible.
Condensation of steam in the CMT could be avoided with some technical arrangements in the
test facility. However, even though improvements were made to gravity driven ECC systems,
we cannot guarantee that computational models will provide accurate answers. Therefore, to
build this confidence more experimental data has to be obtained and new computational models
developed.

309

REFERENCES

/!/ T. Kervinen, V. Riikonen, J. Kouhia, "PACTEL, Facility for Small and Medium Break
LOCA Experiments," Proceedings of ENC'90 Conference. European Nuclear Society, Lyon,
France, September 23-28, 1990
/2/ Munther, R., Kalli, H., Kouhia, J., Kervinen, T. Passive core cooling experiments with
PACTEL facility. ENS TOPNUX'93, Haag, Netherlands, April 25-28, 1993.
/3/ Munther, R., Vihavainen, J., Kalli, H., Kouhia, J., Riikonen, V., RELAP5 analysis of gravity
driven core cooling experiments with PACTEL. ARS'94, INTL topical meeting on advanced
reactor safety, Pittsburgh, USA, April 17-21, 1994. ISBN 0-89448-193-2.

/4/ T. Yonomoto, Y. Kukita, Y. Anoda, "Passive Safety Injection Experiment at the ROSA-V
Large Scale Test Facility," Proceedings of the ANS National Heat Transfer Conference, p. 393,
American Nuclear Society, Atlanta, Georgia, August 8-11, 1993.
APPENDIX: NODALIZATION SCHEMES FOR PACTEL GRAVITY DRIVEN CORE
COOLING EXPERIMENTS
First series:

RELAP5/MOD3.1

NODALIZATION OF
PACTEL FACILITY
FOR GDE-TESTS
stucure

[X] dosing

IMU

[XI control

r^^,^

Second series:

RELAP5/MOD3.1
NODALIZATION OF
PACTEL FACILITY
FOR GDE-TESTS

310

HEAT TRANSFER TO AN IN-CONTAINMENT HEAT

EXCHANGER IN NATURAL CONVECTION FLOW:
VALIDATION OF THE AEA TECHNOLOGY
COMPUTATIONAL FLUID DYNAMICS CODE CFDS-FLOW3D

R. O'MAHONEY, J.N. LILLINGTON

AEA Technology, Winfrith Technology Centre,
Dorchester, United Kingdom
Abstract

Validation is presented of an appropriate computer code for modelling heat transfer from the
containment atmosphere to an in-containment heat exchanger using new data from ENEL.
This work has been carried out in collaboration with ENEL, CISE and ANS ALDO. The study
helps to identify conditions under which natural circulation induced by the heat exchanger
does initiate. The Computational Fluid Dynamics (CFD) code CFDS-FLOW3D, developed by
AEA Technology, has been used, initially in a 2-dimensional mode, to simulate the natural
convection flow generated within a test vessel by an internal heat exchanger operating in a
steam-air gas mixture. The model incorporates a calculation of the heat exchanger
condensation rate based on local conditions. Calculational parameters have been identified
which allow the transient timesteps to converge sufficiently but without using excessive CPU
time. Results of pre-test calculations performed for 2 different geometrical configurations are

presented. These calculations suggest that the heat exchanger will operate as intended and, at
the design values of pressure and temperature, would exceed the planned test power by up to
28%. Post-test simulation results are presented for the first test performed. Good general
agreement with major measured parameters is found and a possible explanation for the high
upward velocity measured outside the heat exchanger exit is offered. The simulation
underestimated the total condenser power by about 16%; this is believed to be due to
underpredicting the steady state steam fraction in the vessel. CFDS-FLOW3D is found to be
a suitable tool for simulating the details of complex buoyancy driven flows, including noncondensibles, in passive containment cooling applications. The code is sufficiently flexible to
be able to represent correctly heat exchanger condensation effects and to be able to simulate
the resultant natural convection flows in either 2-D or, if required, in 3-D.

1.

INTRODUCTION

A characteristic feature of innovative advanced reactor designs is their reliance on passive

safety systems and, in particular, natural convection to transport decay heat away under
accident conditions. This approach results in slow heat-up rates thus avoiding the need for
operators' intervention for an extended period.
In many current designs the route for decay heat removal from the core to the 'ultimate heat
sink' is as follows:
1.

Decay heat is removed from the core by natural circulation via the reactor coolant
system (RCS) to a large condensing pool or tank, thence imparted to the containment
atmosphere through evaporation,

311

2.

The heat must then be transferred from the atmosphere in the interior of the
containment to the exterior of the containment barrier and finally to the outside
atmosphere (ie the ultimate heat sink).

There are two types of containments being considered in current designs. Steel containments,
e.g. as in AP600, are attractive from the point of view of good heat transfer from the interior
to the exterior but concerns have been expressed that these would not be sufficiently strong
to meet European licensing requirements. Concrete containments are strong but they are poor
heat conductors and heat exchangers will need to be introduced to transfer heat from the
interior to the exterior.
The main purpose of this paper is to present the validation of an appropriate computer code
for modelling heat transfer from the containment atmosphere to such a heat exchanger using
new data from ENEL. This work has been carried out in collaboration with ENEL, CISE and
ANSALDO. The study helps to identify conditions under which natural circulation induced
by the heat exchanger does initiate.
2.

HEAT EXCHANGER EXPERIMENTS

ENEL are evaluating the feasibility of a Passive Containment Cooling System (PCCS)
applicable to a large dry double-barrier concrete containment. The internal heat exchangers
are connected to external heat exchangers sited in large tanks of water above the containment
and heat is transferred by evaporating water in natural circulation. The ENEL 1994 work
programme involved a series of tests performed with a mock-up heat exchanger at 1:40 scale.
The technical design criteria include the following:

the system is completely passive, both internal and external to the primary containment

boundary,

the system is able to maintain its performance for an indefinite period,

the peak containment pressure should not exceed the containment design pressure for
LBLOCA,

the system performance should be such as to lower containment pressure to less than
half the design pressure within 24 hours of the peak,

Passive operation of the PCCS is obtained by utilising natural convection of the relevant
fluids, ie steam and non-condensables (air, hydrogen (for severe accidents)) in the containment
atmosphere and water in pools as part of closed thermo-syphon systems.
ENEL are undertaking a programme of experimental tests to verify the behaviour of internal
containment heat exchangers (HXs), both with respect to their heat transfer characteristics and
also to the natural circulation of the steam-air mixture through the HX. The gas circulation
and heat transfer will be strongly coupled. The gas velocities expected to be very low, less
than 1 m/s. An important aspect of the tests will be the identification of conditions under which
the HX gas natural circulation does, or does not, initiate.

312

These tests are being performed at the CISE facility in Milan (figure 1); the first phase was
carried out during late 1994 and early 1995. The tests are aimed to validate the HX design
criteria, in particular.

to verify the design correlation for HX behaviour as a function of the containment

atmosphere composition and concentrations,
to verify the HX behaviour as a function of the geometry of the chimney (height and
outlet area),

to identify the presence of any recirculating flow inside the chimney, as a function of
the chimney bottom shape.

Figure 1.

ENEL Heat Exchanger (HX) Test Section

313

Tests performed later in the experimental schedule will have objectives to:
verify the effect of light non-condensable gases (eg hydrogen) on the HX behaviour,

verify the HX behaviour as a function of geometrical design, eg different HX tubes

and numbers of rows,

2.1.
Heat Exchanger Facility
The heat exchanger experimental facility (figure 1) consists of an insulated vessel with a
design pressure of 5 bar and design temperature of 150*C. The vessel is approximately 8m
high and 1m internal diameter with a heated pool of water at the bottom. A 1:40 scaled heat
exchanger, with a variable height chimney below it, is located near the top of the vessel. A
ckcular screen is located just below the chimney bottom to allow the outlet flow area to be
varied. During the tests the pool is heated producing evaporation; this pressurises the vessel
and causes the air-steam mixture to circulate. The heat exchanger (HX) secondary side is

connected in a closed loop to an external heat exchanger (ie a condenser) to control its
temperature. The steam-depleted gas mixture exiting the HX will be heavier than the bulk gas
mixture and will therefore tend to flow down the chimney and generate a natural circulation
gas flow.
Data have been obtained by ENEL for a total of 6 tests; 4 with a 3m chimney configuration
and 2 with a 5m chimney configuration.

3.

PRE-TEST ANALYSIS

AEA pre-test analysis has concentrated on two aspects. Firstly, using correlations developed
by HTFS [1,2,3], to predict the heat transfer performance of the HX under the design
conditions of pressure and temperature and, secondly, using the computational fluid dynamics
code CFDS-FLOW3D, to predict the gas mixture flow patterns at start-up and close to the
predicted operating point A prediction of the HX heat transfer performance can be used to
confirm the applicability of the selected test conditions. The CFDS-FLOW3D analysis can be

used to give confidence that the HX circulation flow will initiate and continue as intended.
Heat exchanger correlations developed by HTFS [1,2], adapted for condensation using the
Reynolds-Colburn mass-transfer/heat transfer analogy, have been used to determine the HX
condensation rate, and thereby the buoyancy driving force, as a function of the HX inlet
velocity. Similarly, using the HTFS HX pressure drop correlation [3] together with an estimate
of the additional losses, the total system frictional pressure drop has been determined as a
function of inlet velocity. The buoyancy force and total pressure drop have been evaluated for
design conditions of a saturated air-steam mixture at 2.6bar, HX inlet temperature of 109*C
and HX secondary coolant temperature of 100'C with a 5m high chimney. The condensation
buoyancy driving term decreases with increasing velocity while the total system pressure drop
increases with increasing inlet velocity. The crossover of these curves indicates the predicted
operating point of the system, approximately 0.275 m/s. This velocity corresponds to a total
power transferred due to condensation of approximately 170% of the planned test power.
Therefore, for these conditions, the HX should easily remove the design power and in practice
the system would be expected either to reach a steady state at a lower gas temperature or,
operate with a slightly higher secondary coolant temperature.

314

3.1.
Predictions of Flow Patterns
A 2-dimensional CFDS-FLOW3D model has been created representing one half of the facility
(using a plane of symmetry) for the first design configuration, ie with a 5m high chimney
which has its outlet area reduced to Vb of the inlet area in order to maximise the outlet
velocity. The CFDS-FLOW3D model is shown in figure 2. The initia] conditions are the test
design conditions together with zero flows. Steam is added to the bottom layer of cells at a
rate equivalent to the net heat input to the water pool. Steam is removed from the gas mixture
at the HX level inside the chimney at a rate determined from the HTFS correlations using the
local conditions. These correlations have been added as user coding to CFDS-FLOW3D. The
HX pressure drop calculation has been implemented in a similar manner.
TEST
SECTION
VESSEL

..*..

cc

I HEADER

HEAT EXCHANGER
TUBES

CO

LL
O
in

<r>

D_

CO

LU

CHIMNEY

HI

CO
CO

CC
hUJ
0
_J
LU
CO
CO
HI

'^
SCREEN

A EXPECTED GAS FLOW

DIRECTION
POOL SURFACE

0.55 metres - 22 cells

Figure 2.

CFDS-FLOW3D Model of ENEL Heat Exchanger Test Section

with 5m Chimney

A CFDS-FLOW3D calculation has been run, as a transient, for 45 seconds allowing the flows
to develop from zero. Although the gas mixture was predicted to flow up the chimney towards
the HX for a short initial period (4V$ seconds), the flow quickly turned round as the local
mixture density increased and the desired circulation pattern was established. The predicted
velocity contours (figure 3) show the overall flow pattern; of particular interest are the high315

5 m C h i m n e y : 4 5 . 5 sees R u n

Speed C o n t o u r s

6.4750E-01
5.3958E-O1
4.3167E-O1
3.2375E-01
2.1 5 83 E -01
1.0792E-01
O.OOOOE + 00

Figure 3.

CEDS-FLOW3D Pre-test Predictions for 5m Chimney:

Speed contours at 45 seconds

speed flow exiting the chimney bottom and the small 'dead' or low-speec region just inside
the edge of the HX inlet. There is no sign of re-entrant or recirculating flow at the chimney
outlet; the facility is therefore predicted to operate as intended, for these design conditions.
The predicted gas temperature contours (figure 4) clearly show the colder gas inside the
chimney exiting at the bottom and mixing with the wanner gas in the outer annulus. The
predicted HX inlet velocity at 45 seconds was 0.26m/s and the predicted total power was
128% of the design power. Although lower than the power estimated by the previous analysis
this still easily exceeds the design power.

Pre-test analysis using CFDS-FLOW3D was also performed for the second design
configuration, ie with a 3m chimney which has its outlet area equal to the inlet area in order
to maximise the potential of re-entrant flows. This analysis indicated that the test would
proceed as intended, without re-entrant or recirculating flows at the chimney outlet, and that
the HX power would exceed the design power. Post-test analysis of this configuration is
described in the next section.

316

5 m C h i m n e y : 4 5 . 5 sees R u n

Temperature Contours

3 8194E+02
3.8141E + 02

3.8088E-4-02
3 8O35E+02
3.7982E + 02
3 7929E+02
3.7877E + 02

I
Figure 4.

4.

CFDS-FLOW3D Pre-test Predictions for 5m Chimney:

Gas Mixture Tempo nil lire Contours at 45 .seconds

POST-TEST ANALYSIS

Post-test analysis has been performed for the first test for which data were available, test 8a.
The first stage of the analysis repeated the HTFS correlation prediction method of section 3
but using the actual test conditions. This enabled the HX inlet velocity to be estimated and
the sensitivity to the assumed fabe interface temperature to be identified. For the pre-test
analysis the tube mterf8ce temperature was assumed to be 103C, based on a secondary
coolant temperature of 100C plus a 3 temperature difference from the coolant to the gas-side
tube interface. The actual test 8a data indicate that the rube interface temperature must lie
between 103.25*C (secondary coolant temperature) and 106.93C (HX gas mixture outlet
temperature). If a tube interface temperature of 104.6*C (Tcoolant+1.25*) is assumed, then the
correlation method of section 3 predicts that the system will operate at an inlet velocity of
0.21m/s and an HX condenser power of 101% of the design power with an HX outlet
temperature of 107.08*C. This predicted power and outlet temperature agree very well with
the measured values for this test; this tube interface temperature was therefore used in the
second stage of analysis - the CFDS-FLOW3D simulation described below.

317

4.1. Simulations of Flow Patterns

A 2-dimensional CFDS-FLOW3D simulation of test 8a has been run, as a transient, for 133
seconds starting from zero flows. At 133 seconds the simulation was reasonably close to
steady conditions as can be seen from the HX inlet velocity (figure 5) and HX condensation
power (figure 6). The predicted condenser power at 133 seconds was 79.5% of the measured
power and the inlet velocity was 0.168m/s. The 'sensible' heat transfer, ie the convective heat
transfer in addition to the condensation heat transfer, predicted for these conditions by the
RTFS correlation [1], is approximately 4.3% of the measured power. Thus the total power
removed by the HX is predicted to be 84% of the measured power removed. The difference
is likely to be due, at least in part, to a reduction in the predicted steam mass fraction of the
gas mixture during the 133 seconds of the transient This could probably be addressed in
future simulations by choosing slightly different initial conditions for the transient.

Time (s)
ENH.TBT8. - 3m CHIMNEY

Figure 5.

CFDS-FLOW3D Post-test Simulation for 3m Chimney, Test 8a:

HX Inlet Velocity vs Time

Time (s)
ENELTESTBa 3m CHIMNEY
Half Fic.l.t)'Simulated

Figure 6.
318

CFDS-FLOW3D Post-test Simulation for 3m Chimney, Test 8a:

HX Condensation Power vs Time

The u-velociry contours (vertical velocity) and velocity vectors from the CFDS-FLOW3D
simulation of test 8a (figures 7 & 9 respectively) show the rising gas mixture (+ve velocity)
in the outer annulus and the falling gas mixture within the HX chimney. Of particular interest
is the small region of high upward velocity (shown red) close to the outer wall above the edge
of the screen. This region is close to the annulus gas velocity measurement point and its
existence may, at least partly, explain the high annulus gas velocities measured for some tests.
(Annulus gas velocities would, based on the respective flow areas, be expected to be slightly
lower than those at the HX inlet).
The gas temperature contours (figure 8) show the cooling effect of the HX. The CFDSFLOW3D simulation predicts an HX outlet gas temperature of 106.75'C and an HX gas inlet
temperature of 108.55*C. These compare very well with the measured values.
CFDS-FLOW3D v3.3
3m C h i m n e y Test 8a
Run mO7 - 8663

U V e l o c i t y C o n t o u r s m/s
3.4I33E-01
2.4389E-01
1.4645E-O1
4 900SE-02
-4.8437E-O2
-1.4S88E-01
-2.4332E-OI

I
Figure 7.

CFDS-FLOW3D Post-test Simulation for 3m Chimney, Test 8a:

U-velocity (upward) Contours at 133 seconds

CFDS-FLOW3D v3.3

3m C h i m n e y Test 8a
Ron m07 - 8663

Temperature Contours K
,> 3.84OOE+O2
3.8348E-I-O2
3.8277E+O2
3.82OSE+O2
3.8134E+O2
3 8O63E-hO2
3.80UE+02

I
Figure 8.

CFDS-FLOW3D Post-test Simulation for 3m Chimney. Test 8a:

Gas Mixture Temperature Contours al 133 seconds
319

; 3rrr Chirrmery-T^st

Chiomnew
m 7 -

Ve-loci-ty- Ve-ct-ots
3.-4133E-01

-5.-6&88E-02
OlOO'DOE+DO

Figure 9.

320

CFDS-FLOW3D Post-test Simulation for 3m Chimney, Test 8a:

Gas Velocity Vectors at 133 seconds: HX Inlet & Chimney Outlet

5.

SUMMARY AND CONCLUSIONS

1)

Pre-test analyses were performed, including CFDS-FLOW3D predictions, for 2 of the

tests; 1 test for each of the 2 chimney configurations that were to be used. These
indicated that the tests would operate as intended and that the HX power was sufficient
for the design conditions. The actual test data have confirmed these predictions.

2)

Post-test analysis, including a CFDS-FLOW3D simulation, has been performed for the
first test conducted, test 8a. The simulation is in good general agreement with major
measured parameters and also offers an explanation for the high upward gas velocities
measured in some tests. The simulation underestimated the total measured condenser
power by about 16%; this is believed to be due to underpredicting the steady state
steam fraction in the vessel.

The following findings relate to the capability of CFDS-FLOW3D for this analysis:

3)

CFDS-FLOW3D is sufficiently flexible to allow correlations to be added to represent

the HX condensation and the HX pressure drop. This is achieved by providing
additional "user" subroutines written in fortran.

4)

CFDS-FLOW3D is sufficiently flexible to represent the CISE facility geometry in

either 2-D or, if required, in 3-D.

5)

For the 2-D model used so far the code is relatively slow running, taking timesteps of
up to 30ms. The post-test simulation of test 8a took 32 hours CPU time to calculate
133 seconds of transient on a SUN SPARCstation 10. It is thus just practical to
compute a complete transient, approaching the steady state, without using a CRAY.

6)

The CFDS-FLOW3D simulation has provided a detailed insight into the flow patterns
likely within the test vessel and has provided a plausible explanation of otherwise
unexpected results.

5.1. Further AEA Analysis of ENEL Data

The following future activities are planned:

Perform post-test analyses of the remaining 5 tests for which data are provided.

In achieving these there will also be an aim to:-

Further improve the running speed of the calculations - perhaps by spreading the
condensation region.

Move to a 3-Dimensional representation of the facility.

6.

ACKNOWLEDGEMENTS

The authors acknowledge the support of the UK Department of Trade and Industry who
funded this work. They also acknowledge fruitful technical discussions with P Vacchiani of
CISE and V Cavicchia and P Vanini of ENEL who supplied the data.

321

REFERENCES

[1]

Heat Transfer and Fluid Flow Service, Airside Heat-transfer Coefficients for Staggered
and In-line Arrays of Finned Tubes, HTFS Handbook Paper AMI, Harwell, 1985.

[2]

Heat Transfer and Fluid Flow Service, Fin Efficiency and Surface Effectiveness, HTFS
Handbook Paper AM7, Harwell, April 1986.

[3]

Heat Transfer and Fluid Flow Service, Airside Heat-transfer Pressure Drop for
Staggered and In-line Arrays of Finned Tubes, HTFS Handbook Paper AM3, 1980.

322

ATHLET MODEL IMPROVEMENT FOR THE DETERMINATION

OF HEAT TRANSFER COEFFICIENTS DURING CONDENSATION
OF VAPOR IN HORIZONTAL TUBES

E.F. HICKEN, H. JAEGERS, A. SCHAFFRATH

Forschungszentrum Jiilich ISR, Jiilich, Germany
Abstract

Pre- and posttest calculations of the NOKO experiments shall be performed with the
ATHLET code. ATHLET which is being developed by the Gesellschaft fur Anlagen- und
Reaktorsicherheit (GRS) mbH is intended to cover in a single code the entire spectrum of
loss-of-coolant and transient accidents in Light Water Reactors (LWR).
The present work is sponsored by BMBF. The objectives are in detail:
the modeling of the emergency condenser and the NOKO test facility,
the improvement of the ATHLET condensation model to describe the condensation
heat transfer of pure vapors and vapor/non-condensible mixtures in horizontal and
inclined tubes in ATHLET
pre- and posttest calculations of selected NOKO experiments.
The progress to date is that ATHLET and its pre- and postprocessors are installed on the
IBM RISC workstation cluster of the Zentralinstitut fur Angewandte Mathematik (ZAM) and
DEC workstation of the institute. Furtheron an input data set for NOKO has been developed
and first calculations have been performed.
As a first step for an ATHLET model improvement the present condensation model was
analysed and a literature study was performed. The present ATHLET condensation model is
only able to determine heat transfer coefficients for annular flow with turbulent or laminar
films in vertical tubes. Therefore, the correlations of Nusselt and Carpenter & Colburn are
used. During an ATHLET run, the heat transfer coefficients are determined by using both
approaches and the maximum value is used for further calculation.
In the literature, condensation heat transfer in horizontal and slightly inclined tubes is
discussed in the same manner. The flow conditions can be determined by using the flow chart
of Tandon with the improvement of Palen (transition regime between annular and stratified
flow). For every flow condition several empirical or semi-empirical correlations are available.
In the first section of the tubes, the steam velocity is relatively high and there is a nonnegligible influence of vapor shear on the condensate film and the heat transfer. During
condensation along the tubes, the steam velocity and its influence on the heat transfer
decreases.
For the determination of heat transfer coefficients during condensation in horizontal and
inclinced tubes the stand-alone modul KONWAR (Kondensation in waagerecht und leicht
geneigten Rohren) has been developed. After the input of actuar~geometry and flow
parameters, the water steam properties are calculated by using ATHLET functions, which are
linked to KONWAR. In a next step, the flow conditions are determined following by the
calculation of the heat transfer coefficients.
1.

Introduction

The BWR 600/1000 is a new innovative boiling water reactor concept which is being
developed by Siemens. The concept is characterized in particular by passive safety systems;
they are decribed in another paper. One of the important safety systems is the so-called
emergency condenser. This safety system has to transfer decay heat from the vessel/core
region to a heat sink in the containment, for the BWR 600/1000 concept into a large water
pool.
Parameters influencing the effectiveness of the emergency condenser are the primary pressure
in the pressure vessel, the pressure and temperature of the containment, non-condensibles

323

and the geometrical design. The experimental data will be gained with the NOKO facility,
which is described in another paper.
Pre- and posttest calculations of these tests shall be performed with the ATHLET code which
has been developed and validated by the Gesellschaft fur Anlagen- und Reaktorsicherheit
(GRS)
Germany. The Institute for Safety Research and Reactor Technology (ISR) of the
Research Center Jiilich (KFA) received the task from the BMBF to use the ATHLET code
for the emergency condenser and - if necessary - improve or modify models. In specific, the
tasks listed below are given.
The modeling of the emergency condenser and the NOKO test facility,
the improvement of the ATHLET condensation model to describe the condensation
heat transfer of pure vapors and vapor/non-condensible mixtures in horizontal and
inclined tubes in ATHLET
pre- and posttest calculations of selected NOKO experiments.

2.

The Physical Phenomena

Fig. 1 show schematically one pipe (out of 4 x 104 pipes in a BWR 600/1000) with an
arbitarity water level inside. The length is about 10m, the tube inside diameter 38.7 mm and
the wall thickness 2.9 mm. The inlet steam velocity ranges from about 6 m/s at 7 MPa to
about 20 m/s at 1 MPa; the outlet water velocities are up to 0.25 - 0.3 m/s. It is evident that
depending on velocities and pressures differences flow characteristics of the condensate exist.
Fig. 2 shows some possibilities. It should, however, be mentioned that - due to relatively thick
tube wall - the main thermal resistance is with the metal wall.
The outside heat transfer mechanism is mainly boiling. Due to the relatively narrow bundle
boiling from the lower branch of the tube may influence the boiling of the upper branch. In
addition, larger bubbles may be formed and produce some unsteady behavior.

The temperature decrease inside the tube below the water level is also of interest because the
density differences between the down-pipe and the pressure vessel determines the water level
difference, see fig. 3.

The effectiveness of the emergency condenser in case of the existance of non-condensibles

cannot be calculated with sufficient accuracy. For very low concentrations of non-condensibles
these will be solved in the condensate and thus be removed from the emergency condenser.
For higher concentrations non-condensibles will be accumulated within the tubes; the bend
between down-pipe and reactor vessel may block the removal of the non-condensibles. It will
be of interest which concentrations of non-condensibles will exist along the tubes.
It will also be studied if flashing inside the tubes due to pressure transients will influence the
effectiveness.
8 tubes

average tube length

inner diameter
wall thickness

9.8 m
37.8 mm
2.9 mm

WD < 6 m/s
( WD < 20 m/s at p = 1 MPa

Wc < 0.3 m/s

FIG.

324

1. NOKO tube characteristical operating conditions for 7 MPa.

Flow pattern

low inlet
velocity

high inlet

symmetric unsymmetric
annular flow annular flow

^-^__-_

v wiwviiy

>

^^^
annular
flow

wavy
flow

^Qzoo
~

-^>. .- s. y-~^xD

slug, plug
flow

stratified
flow

. _

^
single
phase fluid
flow

FIG. 2. Flow patterns occuring during condensation inside horisontal tubes.

PPV

FIG. 3. Water level difference between emergency condenser and pressure vessel.

325

3.

The Heat Transfer Package in ATHLET

The general aspects of the ATHLET code have been described elsewhere.
The heat transfer package in ATHLET is controlled by levels; level 1 is characterized by heat
transfer from fluid to structures. Within this level 3 different modes, see fig. 4, are possible.
Within each mode heat transfer coefficients will be determined by using different options. For
the further calculation the maximum value will be used. It should be noted that experimental
data for condensation inside tubes are mainly available for vertical tubes.

Fig. 5 shows a more detailed flow regime, as proposed by Tandon/Palen. In fig. 6 the different
correlations are listed. For the emergency condenser these correlations will be tested. If they
do not fit, model improvements would be necessary.
Fig. 7 shows flow regimes with the pressure and the power as parameter the which are
expected in the emergency condenser.
It has not been decided which method will be used in case of non-condensibles inside the
tubes. As a start the reduction of the inside heat transfer will be related to the decrease in the
(steam) saturation temperature.
Mode

Fluid

Correlation

Mechanism

single
phase
liquid

Dittus-Boetter
Me Adams
mm

forced convection
natural convection
20 W/m1 K

two
phase
fluid

Chen
Nusselt
Carpenter-Colbum
mm

forced convection
laminar film condensation
turbulent film condensation
20 W/m' K

single
phase
steam

Hausen
Dittus-Boelter
Me Eltgot
mm

forced convection
forced convection
forced convection
lOW/m'K

FIG. 4. ATHLET heat transfer correlations of level I'.

2 10'

10'-

10'

slug
flow
stratified flow

(wavy)

001

10'-

plug
flow
10'

x M

5 -10' 10

1-e

10'

[ Wallis ]

= liquid/steam cross sectional area ratio

"o

FIG. 5. Flow regime map for condensation inside horizontal tubes according to TANDON.
326

Flow pattern

Fluid

Correlation

Spray flow

two

Solimann

phase

Nusselt, Kosky & Staub,

Akers,
Carpenter & Colburn,
Chen, Kruzihlin,
Baehr

annular and
semiannular flow

stratified flow

Jaster & Kosky,

Rufer & Kezios

fluid
bubbly, slug, plug
flow

Brebber

steam

Hausen,
Dittus-Boelter

water

Sieder-Tate,
Dittus-Boelter

FIG.

6. Correlations which will be tested for NOKO calculations.

10'

0.01

1
2
3
4
5
6
7

spray flow
annular flow
transition regime
stratified flow
bubbly flow
slug flow
plug flow

0.001
0.0001

0.001

A p = 75bar,4MW
B p = 55bar.4MW
c p = 28bar,4MW
E p=10bar,4MW
D p=75bar,1MW

0.01

0.1

10

1-e

FIG.

7. Flow regimes inside the NOKO tubes.

327

4.

ATHLET Modeling of the NOKO-Facility

Fig. 8 and 9 show the nodalisation of the NOKO facility. It includes 3 thermofluiddynamic
systems, 60 network-objects, constisting of 280 nodes and 240 junctions, 60 heatslabs and
special components (e.g. pump, heat exchanger, electrical heater). All control systems can be
simulated.
A special nodalisation has to be used for the "secondary" side of the condenser, see fig. 10,
due to the recirculation flows.

steam line

steam line

emergency condenser
secondary side

' separator

FIG.

FIG.

328

8. NOKO nodalisation scheme.

9. Nodalisation scheme of NOKO secondary side.

6.00r

emergency condenser power

4.50L

FIG.

0.00
iiiiiiir
500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000
Time in s

10. Emergency condenser power, water level in pressure vessel and pressure
(4 tubes, MD = 0.3 kg/s, p = 28 bar).

5.

Some Results from Calculations

In the following, an ATHLET run at 2.8 MPa pressure is discussed, see fig. 11. During the
calculation, 4 emergency condenser tubes are open. The pressure in the system is fixed by the
control system, the water level decrease is specified by the user- in the input data set. During
the decrease of the water level, heat is removed to the condenser secondary side. At nearly
800 s, the removed heat is larger than the energy supplied by the steam and the pressure
decreases. The increase of the emergency condenser power at 800 s results from the
uncoverage of the control volumes behind the tube bends. In the horizontal part of a U-tube,
the increase of heat transfer surface caused by water level decrease is much stronger than in
the tube bend. After heatup of the emergency condenser secondary side (see T6 - T8 in Fig.
11) and the increase of the fluid temperature at the end of the tubes with time, the emergency
condenser power slowly decreases and, at 3500 s, the amount of condensate reaches the
amount of steam again. In Fig. 11, the temperature distribution in the emergency condenser is
shown. The temperature measurement positions Tl - T4 are located inside the bundle (Tl =
inlet, T2 = before tube bend, T3 = after tube bend, T4 = outlet). The saturation temperature
in the different control volumes identify uncoverage. After total condensation of the steam,
the fluid subcools. The degree of subcooling is influenced by the temperature outside the
bundle and the degree of its uncoverage. On the secondary side of the emergency condenser,
there is a temperature increase with a gradient of 0.01 K/s.

329

i
<*
I

Temperature in C

RELAP5/MOD3 PRE-TEST PREDICTIONS FOR

THE SPES-2 1" C. L. BREAK TEST S01613

A. ALEMBERTI, C. FREPOLI, G. GRAZIOSI

ANSALDO Nuclear Division,
Genoa, Italy
Abstract

SPES-2 is a full height, full pressure experimental facility scaled 1/395 respect to the
Westinghouse AP600 plant. The SPES-2 facility designed and operated by SIET in Piacenza
is the evolution of the previous existing SPES-1 facility. The SPES-2 test matrix provides a
complete set of experiments from Cold Leg break accidents to Steam Generator Tubes
Ruptures and Main Steam Line break. The SPES-2 test program is performed under the
technical cooperation agreement among ENEL, ENEA, ANSALDO and WESTINGHOUSE.
In the frame of the SPES-2 activities ANSALDO carried out pre-test calculations for the
facility as well as comparisons with full plant behaviour to support the facility scaling criteria.
SPES-2 calculations were made using a noding developed by ANSALDO for the
Relap5/mod3/V80 code.
Experiment SOI 613 was the last test of the SPES-2 test matrix, a 1" break on the Cold Leg
bottom performed with a total of three PRHR tubes instead of the normal SPES-2
configuration where only one tube is used to simulate one of the two scaled AP600 PRHR
heat exchangers.
The main purpose of this test was to investigate the facility response when an overcooling
capability is provided to the primary system.
The paper presents first the Relap5/mod3 noding for the SPES-2 facility with highlights on
particular aspects of the nodalization used for the simulation.
Then a comparison between experimental data and pre-test calculations is performed to show
the agreement between pre-test predictions and experimental data.
The code results shows that Relap5/mod3, when the noding is carefully tuned on the facility
using the informations provided by hot and cold shakedown as well as some of the previous
transient tests, is able to predict the facility response for the main parameters of the transient.
A brief comparison between the pre-test predictions made for test SOQ401 (1" Cold Leg
Break test with one PRHR tube) and test S01613 is also presented to analyze the different
response of the facility when an overcooling capability is provided to the primary system.
INTRODUCTION

The paper presents a comparison of the pre-test prediction made by ANSALDO using
RELAP5/Mod3/V80 with SPES-2 test data for transient S01613, a 1" C.L. break made with
3 PRHR tubes.
The SPES-2 noding developed by ANSALDO was tested first against cold and hot
shakedown tests and tuned to obtain a current reproduction of the SPES-2 primary pressure
losses, primary and secondary heat losses etc. Once the first SPES-2 experimental results
were available modification to the noding were made to better simulate the facility behaviour
and also results from full 3-D simulations were used to improve the RELAP5 response [4].
The above work leaded to the SPES-2 noding presented in Figure 1.
The agreement between pre-test prediction and experimental data for test SO 1613 was very
good and it will he discussed in the following.
Then a comparison between the calculated data for this test and the 1" C.L. break with 1
PRHR tube is performed. Final conclusions are then provided.
331

SPES2

STEAM GENERATOR B

STEAM GENERATOR A
970

CMTs

IRWST
680681

Figure 1 - SPES-2 Relap5/mod3 noding

S01613 PRE-TEST CALCULATION AND EXPERIMENTAL DATA COMPARISON

Figure 2 presents the primary side pressure behaviour. The agreement between the
experimental data and the pre-test prediction is very good. After a phase of fast primary
pressure decrease oscillations in RCS pressure take place due to an unstable natural
circulation phenomena while a constant and slow primary pressure decrease takes place due
to the PRHR heat exchange.
When CMT injection switch from single to two-phase flow in the balance line (shown by
CMT injection increase in Figure 3 and 4) the primary pressure decrease is affected due to
the lower entalphy of the flow entering the core.
First stage ADS tuning is well calculated by the code as it is shown in Figure 2 by the sudden
decrease of the primary pressure.
Figure 3 and 4 shows the decrease of the CMT injection flow in the first phase of the
transient due to the CMT heat-up and a consequent decrease of the head available for
injection until the CMT Cold Leg Balance Line flow is in single phase conditions. As the
Balance lines flow become two-phase the CMT injection increases. However the CMT
injection is not very stable as in other transients [1], [2] due to the presence of slugs in the
balance line. The code was able to compute in the correct way the above phenomenology.
Figure 5 presents the break flow comparison showing a very good agreement between

predictions and experimental data.

Figure 6 is a comparison of the pressurizer level. The pressurizer level decrease in the first
part of the transient is well predicted, being conditioned by the primary mass loss. The
pressurizer level increase at the end of the transient due to ADS valves opening is well
calculated by the code while the subsequent pressurizer draining is qualitatively predicted
with a short time delay.
332

Figure 7 and 8 presents the PRHR mass flow and PRHR inlet-outlet temperatures
comparison.
PRHR mass flow oscillations frequency is well predicted while the amplitude is higher in the
prediction respect to experimental data. The average value is any-way very close to
experiment.
PRHR inlet temperature is obviously well predicted while the strong PRHR outlet subcooling
is also calculated of the correct amount.
RelapS pretest predictions and SO1613 exrp. data
--P-3S5O1OOOO
...- pO27-O

1"

SO1613

FTgi_jre 2. pressurTzer pressures

RelapS pretest predictions and SO1613 eacp. data

mflowj-94OO1OOOO
1"
F.B4.QE-O_____________SO1613

Figure 3 CMT B Injection f lows

RelapS pretest predictions and SO1613 exp. data

mflowj-99OO1OOOO
V
F.A4-OE-O____________SO1613

Figure A- CMT A injection flows

333

RelapS pre-test predictions arid SO1613 ercp. data

mflowj998OOOOOO

brkflO

_______SO1613

Fio,Ljf-e 5 GreaU r-oass flows

RelapS pretest predictions and SO1613 e3cp. data

1"

cntrlvor2O
prszLO___

SO1613

F~i<gi_ire & Rressurizer levels

Relap5 pretest predictions and SO1613 escp. data

mf I ow j66O010OOO

1''

prhrFLO_____________SO1613

F'iguire ~7 PRHR mass flows

334

RelapS pr-etest pr-edi.ct.ions and

- ---

tempf 66O0 1 0OOO

tempf 64OO5OOOO
Toutlet 3

SO1613 escp. data

i
1
SO1613

Tinlet 2

SO1613
~

(v

""--,__

*" V _

"it ___

. !,"-' " ~

pgflwawy.-i^w:, - -

^VX^v,^

-i

F"!gi_jre S RRMR inlet and outlet temperatures

RelapS pretest predictions and SO1613 ex.p>. data

mf I ow J-8OO010OOO
F.A2OE-O

V
S01613

Figure 9 Accumulators injection flows

RelapS pretest predictions and SO1613 escp. data

---

mflowj69OOOOOOO

1"

-\

'

f'^

':
-

Figure

1O IRWST injection flows

335

Figure 5 compares the accumulator behaviour. The injection flow is small until ADS opening
takes place, the qualitative behaviour and effects of second stage opening is correctly
calculated by the code.
Figure 10 presents the IRWST injection flow comparison. Timing is very well calculated while
some difference is present in the injection behaviour.

COMPARISON OF RELAP5 PREDICTIONS FOR SPES-2 TEST S01613 AND TEST

S00401

A comparison of the RELAP5 results for the 1" C.L. break tests with one (S00401) and
three (SOI613) PRHR tube is here briefly presented.
Figure 11 presents the primary pressure behaviour. As can be seen the primary pressure value

is reduced when three PRHR tubes are used in the PRHR due to the overcooUng capability of
the PRHR. However main phenomenologies of the two transients are very similar note that
the first stage ADS timing is not strongly affected.
This is confirmed by the break flow comparison presented in Figure 12. Since the two break
flows are very similar the RCS residual mass is very close in the two runs so that also CMT
level (Fig. 13) as well as CMT injection flow (Figure 14) are not influenced by the PRHR
overcooUng capability.
Accumulators flows (Figure 15) shows some difference in the peak value but the injection
behaviour is the same as well as starting and end time.
IRWST injection timing (Figure 16) is close between the two transients.
PRHR mass flows (Figure 17) shows as higher tendency of the run with 3 PRHR tubes to
oscillations while PRHR inlet outlet temperatures are affected obviously by the higher cooling
capability due to the greater heat exchange area.
PRHR inlet temperature is affected by the saturation pressure difference while outlet
temperature is lower in the case with 3 PRHR tubes.
The difference is anyway limited due to the already low value reached also in the case of only
1 PRHR tube.
The main phenomenologies predicted by the code during the two transients were the same and
also quantitative values does not change greatly due to the overcooUng capability of the
PRHR heat exchanger.

Comparison of Reia.p5 prediction, for- test. SOO4-O1 and SO1613

---. p-3esoioooo

-365O1OOOO

11 pressurtzer pressures

336

Comparison of RelapS prediction for test SOOtOl and SOiei3

- - - mflowj-998OOOOOO
1"-1tube
mflowi-998OOOOOO_____V-3tube

Figure 12 BreoK mass flows

Comparison, of RelapS prediction for test SOO4-O1 and SO1613

1"-1tube
1"-3tube

- - - cntr Ivor3O

cntrlvor3O

Figure 13 CMT B levels

Coi-nparison of ReiapS prediction for test SOO-4rOl an.d SO1613

--- mf low!94OO1OOOO

1"-1tube

1"-3tube

mf low r-94OO 1OOOq

Figure

1-4- CMT B injection mass f lows

337

Comparison, of RelapS prediction for- test SOO4-O1 and SO1613

- - - mflowj8OOO1OOOO

V-1tube

mflowi8OOO1OOOO_____1"-3tufae

15 AoeumuIators injection mass

Comparison, of RelapS prediction, for- test SOOtOl a.n.d SO1613

-- mflowj69OOOOOOO
mflowi69OOOOOOO

1"-1tube (kg/sec)
V '-3\\Joe (kg/sec)

tf

F"igLjre 16 IRWST injection moss flows

CJonnpar-ison. of RelapS prediction, for- test SOO-tOl a.nd

----

mflowj66OO1OOOO

V-1tube

_mflowj66OO1OOOO

V-3tube

Figure

338

1 "7 PRHR moss

SO1613

Comparison of RelapS prediction, for test SOO-tOl and. SO1613

- temp*66OO1OOCX)

----- --

tempf-64OO5OOOO
terrpf-66OO1 DOCK)
teiTyf-64OO50OOO

1"1tube

1"-1tube
1 "-3tube
1"-3tube

F~i<gure IS RRHR Inlet orxd ootlet temperatures

CONCLUSIONS
A comparison between RELAP5/mod3/V80 pre-test predictions made by ANSALDO and 1"
SPES-2 C.L. break with three PRHR tubes has been presented.
The results indicate that all phenomenologies of the transients were captured by the code.
Qualitative and quantitative values of pressure, injection flows, temperatures etc. were very
well predicted by the code.
A comparison has also been presented between code results for 1" C.L. break with one and
three PRHR tubes.
The comparison shows that the response of the facility is slightly influenced by the PRHR
overcooling capability. Main phenomenologies remain the same while timing of the events is
very close in the two cases.

REFERENCES

[1]

Alemberti A, Frepoli C., Graziosi G. 1994

SPES-2 cold leg break experiments: Scaling approach for decay power, heat losses
compensation and metal heat release.
International Conference - New Trends in Nuclear System Thermohydraulics - Pisa,
May 3 Oth-June 2nd

[2]

Alemberti A , Frepoli C., Graziosi G. 1994

Comparison of the SPES-2 pre-test predictions and AP600 plant calculations using
RELAP5/mod3
International Conference - New Trends in Nuclear System Thermohydraulics - Pisa,
May 3 Oth-June 2nd

[3]

M. Bacchiani, C.Medich, M.Rigamonti (SIET) - L.E. Conway (Westinghouse)

SPES-2, AP600 Integral System Test - Inadvertent ADS opening and Cold Leg Break
Transients
International Conference - Pittsburgh Pa April 17-21 ARS-94

[4]

Alemberti A., Frepoli C., Graziosi G. 1994

SPES-2 RELAP5/mod3 noding and 1" cold leg break test S00401
22nd Water Reactor Safety information Meeting- Washington D.C. October 24-26 '94

339

DEVELOPMENT AND INITIAL VALIDATION OF

FAST-RUNNING SIMULATOR OF PWRs: TRAP-2

E. BREGA
ENEL-ATN
C. LOMBARDI, M. RICOTTI, R. SORDI
Department of Nuclear Engineering,
Polytechnic of Milan

Milan, Italy
Abstract
Fast running computer programs with versatile, user-friendly, interactive features are
very useful for the design and operation of nuclear reactors. TRAP-2 tailored to AP600
reactor belongs to such a category.
The main simplifying hypotheses are the incompressible but thermally expandable
fluid and the homogeneous/thermal equilibrium approach for two-phase flow. Distributed
parameter components and lumped parameter components are adopted to describe the
whole circuit. The first ones are solved by the Method of Characteristics. The system is
modelled by several circuits in parallel including primary loops and safety circuits: the
analysis of the first ones proceeds by implicit method, while that of the latter ones by
explicit method.
Either a zero-dimensional or a mono-dimensional approach is available for neutron
kinetics: the solution is obtained at the end of the thermohydraulic time step. Several
nested iterative procedures are adopted to find the relevant parameters in a fixed time
and space grid. Heat exchangers are solved iteratively between primary and secondary
side, while the pressurizeris described by a two-fluid two-phase model. In quasi stagnant
conditions the primary circuit can be collapsed in a two separate phase volume, while
maintaining the current models for safety circuits: this increases substantially the
computation speed of the program, at present ranging from 15 times to about 1 the real
time on PC.
Validation of the program is in progress. Satisfactory comparisons have been
obtained with RELAP results of safety transients of a passive reactor. Other comparisons
with experimental data are being performed regarding SPES transient tests for the
AP600 system.
1.

INTRODUCTION

The analysis of any reactor concept concerns two aspects: safety and operability.
Remaining within the framework of a pressurized light water concept, here we consider
PWRs, both of conventional and advanced design. Safety aspects in relation to large loss
of coolant accidents (LOCA) would appear thoroughly studied and documented: complex
computer programs are available for a complete description of these accidents. On the
other hand in any reactor system, "slow" accidents are more important than previously
supposed, both for the delayed triggering of the protection system, especially when of
"passive" nature, and the higher effect on the transient of the dynamic behaviour of many
components in view also of their control system.

341

As far as operability is concerned, the analysis of different operating conditions,

particularly start-up and stability behaviour, may yield important information about reactor
flexibility.
These considerations had led us to the conclusion that a simple but complete
computation model for plant transients might be a useful tool for both parametric studies
during design activity and on-line simulation, when the reactor is in operation.

The slowness of the transient to be analyzed and the limited production of steam
inside the primary circuit allowed us to adopt the hypotheses suitable for an incompressible
fluid, and in particular to disregard the sound speed effects. Moreover, two-phase flow is
taken into account in a simplified manner, by applying the same equations adopted for the
liquid phase, corrected for the increased specific volume.
When the plant is in quasi stagnant configuration in the primary circuit, such as in
very low flow rate/power condition, a single volume-two fluid model is adopted to simulate
the primary circuit including the vessel, maintaining the current models for the safety
systems. This assumption substantially simplifies and accelerates the computation.
This model has been implemented in several computer programs tailored for
different reactor systems, such as AP600, conventional PWRs, PIUS [1], MARS [2] and
ISIS [3].
This paper is devoted to a synthetic description of the model [4] and to its first
validation activity concerning comparisons with the experimental safety transients obtained
in SPES [5] facility for AP600 and with RELAP transients obtained for the safety analysis of
the MARS reactor.
2.

BASIC THERMOHYDRAULIC MODEL

The thermohydraulic model is based on the general conservation equations of mass,

momentum and energy, and on the constitutive laws to close the system.
The circuit is assumed monodimensional and described by a series of constant
cross section ducts, in which the fluid flows along the axis direction. Then convection and
diffusion in the other two directions are taken into account by means of an integration
process, which implies the adoption of two constitutive equations for the diffusive terms in
energy and momentum balance, i.e. for the heat flux and the friction factor.
The fluid is incompressible but thermally expandable. Its density is only a function of
enthalpy, assuming the pressure value at saturation condition (this hypothesis can be .easily
removed in the future).
Two phase flow, when present, is treated according to the homogeneous model with
equal steam and water velocities at thermal equilibrium. The presence of separated steam
and water zones, such as in the pressurizer and the steam generator dome is described by
mass and energy balance equations applied to each zone, disregarding the momentum
balance equation but using a Bernoulli type equation for pressure loss evaluation.
The overall reactor system (primary, secondary and safety circuits) are characterized
by a number of parallel circuits, which are coupled hydraulically and/or thermally and may
function in natural convection. To describe correctly their functioning and their activation
(when applicable) a detailed and precise spatial distribution of their relevant parameters,
such as density, pressure, boron concentration, is needed. This yields to consider
distributed parameter components in the case of pipes, reactor channels and heat
exchangers. By assuming constant the cross section of each component, or subcomponent
in which the actual component has been divided, one obtains:

342

Mass

dt
~

02

=0

Boron

l2

~^" + ~ir" = 0

Energy

Momentum

-- + -V2 [ [ = - - p m gcosff -/
""*

" 3r

(2)

Q dz
2r

"^*\
2

(4)

In two-phase flow the above quantities are substituted by those obtained by the
homogeneous model with equal steam and water velocities.
The circuit plena do not require such a detailed description, because the intense
turbulence here existing justifies the hypothesis of a complete fluid mixing, thus
disregarding any information about spatial distribution of the relevant parameters. Then a
lumped parameter treatment is applicable, integrating the balance equations within the

volume. For instance this implies an instantaneous and homogeneous mixing of all boron
contained inside the volume, which is opposite to the distributed parameter case, where its
value shifts rigidly without any mixing between the inlet and outlet sections (piston
behaviour).
In lumped parameter components the outlet-inlet pressure drop is obtained by the
"mechanic" energy equation (Bernoulli type), valid for incompressible fluids, which leads to
an ordinary derivative system. This subdivision in several distributed parameter components
with constant cross section requires the presence of connecting areas, which allow sudden
variation of the cross section. For these special components mass and "mechanic" energy
equations are solved, neglecting the inertia and the head terms, while considering the
friction term through a number of kinetic heights, given by the usual hydraulic resistance
handbooks.
The constitutive relationships concerning the friction factor and the heat transfer
coefficient or the heat flux are function of the local conditions of the fluid. The correlations
are derived by the literature and the main ones are detailed in Table I (SI units).
In the heat exchangers, including the steam generator, the heat flux is calculated by
assuming constant, within each time step, the thermoydraulic data of one fluid, when
calculating the evolution of the other and viceversa. In natural convection heat exchangers,
the procedure may be iterated, within the same time step.
A special component is the accumulator, which is pressurized with inert gas and
injects water in the primary circuit, when its pressure falls below the accumulator one. A
rather simple model describes the transient of the inert gas pressure versus the injected
flow rate. In this case an iterative procedure is adopted within each time step to take into
account the counterpressure determined in the primary circuit by the injected fluid.
3.

REACTOR CORE MODELS

The reactor core model presents three aspects: channel thermohydraulic, fuel rod
behaviour and neutron power kinetics.

343

From a thermohydraulic viewpoint the core may be approximately described as a

bundle of separate channels, each containing one fuel element. With their different power
inputs, these channels have also slightly different flow rates in order to satisfy the boundary
condition in obtaining the same core pressure drop. During transient conditions a steamwater mixture may be obtained in the hottest channel and this would imply a slightly higher
asymmetry in channel flow rate distribution and above all a "void" generation, which involves
a negative reactivity feedback. A thorough description of this thermohydraulic configuration
is really too complicated for our program aims and so a simplification has been adopted.
TABLE I - MAIN CORRELATIONS INCLUDED IN CURRENT TRAP-2 VERSION.

CORRELATION

AUTHORS
(analytic)

APPLICATION FIELD
internal laminar flow

Re

SELANDER

internal turbulent flow

y _

Nu - 4.36 (for uniform th.flux)

Nv - 3.66 (for uniform temper.)

(analytic)

M/ = 0.023 Re%Prn
(heated n=0.4; cooled n=0.3)

DlTTUS-

BOELTER

CHURCHILLCHU

free convection,
single-phase external flow
(liquid)

0.387 Ray<

= 10.60 + -

-(o.5S9/Pr)'/u
ROHSENOW

fg

g (Pi-Pv)
Pr,

JENSLOTTES

(25)4

forced convection, laminar

flow
forced convection,
single-phase turbulent flow
(liquid or vapour)

free convection,
external flow,
nucleate boiling

two-phase and subcooled

flow

The reactor core is described through a single power channel representing the core
average channel. This is viewed as a "pipe" with constant cross section, and a power input
derived from the fuel rod subroutine. However at the end of each time step and maintaining
the already calculated core boundary condition constant, the parallel channel behaviour is
determined out of line. Three channels are examined, each representative of those of the
central, intermediate and peripheral zone respectively. In this way, a more precise
temperature and void distribution can be calculated to determine reactivity feedback data.
This out-of-line calculation follows an iterative procedure to correct the flow rate of each
channel at the same overall flow rate and in order to converge to the same pressure drop.
The fuel rod temperature distribution is described in the radial direction. The cross
section is divided into five annular zones: three in the fuel zone, one in the gap and one in
the cladding, and their temperatures are the unknowns. The general Fourier equation was
solved by referring to these five temperatures. The boundary conditions are: the power
generated in each zone, which is proportional to its area (deriving from the neutron kinetic
subroutine), the coolant temperature and the corresponding heat transfer coefficient. In

344

conclusion, five linear algebraic equations are obtained, the solution of which yields the
desired temperature distribution and thus the thermal power transferred to the coolant.
The thermal conductivities and specific heats of zircaloy and uranium oxide are
taken as temperature dependent, while their density is assumed constant. Gap conductivity

is assumed to be bum-up and linear power dependent and so an empirical correlation

based on a rather complex fuel performance model was developed.
The neutron power is calculated out-of-line at the end of each time step; the
resulting value is the power input for the calculation of the fuel rod temperature distribution
in the following time step. The reactivity is calculated as the sum of five terms due to: fuel
temperature (Doppler), moderator temperature, moderator density, moderator voids, boron
coolant concentration. Each one is calculated by multiplying the corresponding parameter
variation with respect to a reference condition by a given input value (reactivity coefficient),
which can be a function of the parameter itself.
The neutron kinetic equations (Nordheim-Fuchs) are written for six delayed neutron
groups and solved by the simplified Runge-Kutta method, according to the procedure
indicated in ref.[6]. The solution can be optionally obtained either by this zero dimension
approach or by a monodimensiona! one. Obviously in the latter case the above procedure is
different in the sense that the thermohydraulic parameter effect is calculated directly on
neutron cross sections instead of on the overall reactivity. However, the experience gained

during the tests showed that the monodimensional solution coincides practically with the
zero dimensional one in the "slow" transients. The solution is obtained by a shorter time
step then the one used in the thermohydraulic part of the program. Therefore the neutron
kinetic integration is carried out with constant input data throughout the whole
thermohydraulic time step.
4.

NUMERICAL SOLUTION

The differential equations are discretized by subdividing the space and the time in
given finite intervals. The time grid is in general fixed, but it can be varied along the
transient, while the space grid is fixed. An option foresees a variable space grid, based on
the length run by the fluid in each time step: this procedure is apt to describe fast transients
with low axial numerical diffusion, but it is time consuming when the flow rate falls to very
low values in any single circuit.
The solution is obtained through the "Method of Characteristics", which seems
efficient in reducing numerical diffusion along the fluid pattern.
The multiple circuits solutions are coupled by implicit ("on-line") or explicit f out-ofline") method in time advancement. Those solved with the first method, which in the present
program can be two, are the primary loops, while all the safety systems and the secondary
side are coupled explicitly. For on-line circuits, an iterative procedure is set up: starting from
given flow rate and pressure values in a reference cross section of one selected circuit, the
thermohydraulic parameters of the fluid are calculated along all its possible patterns and
finally the pressure in the starting cross section is obtained. This value must be equal to the
initial one within a given convergence error, otherwise the flow rate is modified accordingly.
The starting cross section is the connecting point between the pressurizer surge line and
the primary circuit This pressure is determined through an iterative procedure between the
pressurizer and the circuit pressure. In conclusion, a number of nested iterative procedures
are to be solved, i.e.: first and second circuit coupling parameters, pressurizer and first
circuit pressure, primary and secondary circuits inside heat exchanger for heat flux, first
circuit flow rate.

345

The out-of-line circuits, typically the emergency cooling systems, are solved at the
end of each time step, keeping constant their boundary conditions i.e. the thermohydraulic
conditions of the connecting points with the primary circuit.
When the transient reaches a quasi stagnant condition in the primary circuit, low flow
rate and power, an important simplification is adopted. All the primary circuits, including the
reactor vessel, are collapsed in a single volume, where all the existing liquid and steam are
collected in two separate zones. The enthalpy is uniform in the zone and is obtained by
weighing the previous values in each component by their corresponding masses. On the
contrary, the safety systems maintain their geometry and are solved without any change in
the procedure. The transient behaviour of the two zone volume is calculated by a model
similar to that already adopted for the pressurizer.
The control system is modelled by detailing the function of the various controllers,
which can be proportional, integrative or derivative. In the present version the main
controlled quantities are indicated in Fig. 1 and refer to: pressure and liquid level into
pressurizer and steam generators, core fluid temperature, grid load.

Fig. 1: TRAP-2 set of controls for a standard PWR.

5.

TRAP-2 PROGRAM CAPABILITIES

The transient analyses performable by TRAP-2 program are listed in Table II. The
modular structure of the implemented model allowed us the generation of different
programs for each reactor configuration (AP600, PIUS, MARS, ISIS), with particular
attention to the characteristic components of each one.
TABLE II - AVAILABLE ANALYSES IN CURRENT TRAP-2 VERSION.

Accident transients
Station Blackout
Depressurization
Boron Dilution
Steam Line Break
Loss of Feedwater
Steam Generator Tube
Rupture
Control Failure (1 control or
more)
Small Break LOCA
Loss of Pump supply (1
pump or more)

346

Operational transients
- Hot and Cold Shutdown
- Start-up
- Load Following

Stability analysis
- Superposition of
disturbances on critical plant
parameters (different
amplitude and frequency)

Both the input data file and the analysis set up file (ASCII format) are easily
modifiable by the user so that parametric studies are supported. The output data are in
column ASCII format and any graphic program can import and plot the results. A synoptic
diagram of the circuits shows the main parameter values during the transient.
TRAP-2 is written in FORTRAN and developed on personal computers under MSDOS/Windows operative system. The dimension of the executable files is less then 1
Mbyte. The program runs on PC and the transients are usually executed with time steps of
about 100 ms; the CPU time required, depending on the type of analysis and the
nodalization besides the time step, ranges from almost the real time to one order of
magnitude.
6.

TRAP-2 VALIDATION

ENEL has started to validate TRAP-2 computer program by means of comparisons

with experimental data and a further validation activity is in progress by comparing it with a
reference program.
6.1.

SPES-2 2" LOCA accident.

An experimental program has been commissioned by Westinghouse in cooperation

with ANSALDO, ENEA and ENEL to support the design certification for AP600. SPES-2 is
an experimental facility made in SIET with the overall objectives to simulate the AP600
thenmalhydraulic behaviour, in particular referring to the passive systems.
The SPES-2 test facility nodalization as simulated by TRAP-2 program is reported in
Fig. 2. The circuits are subdivided in 93 components, with a total number of 628 sections

(belonging to distributed parameter comps.) plus 14 volumes (belonging to lumped

parameter comps.). The simplified nodalization corresponding to the stagnation model is
shown in Fig. 3 (460 sections plus 4 volumes).

Fig. 2: TRAP-2 nominal nodalization.

Fig. 3: stagnation model nodalization.

347

The TRAP-2 program is currently hosted on a PC586/66 MHz and the dimension of
the executable file is 960 Kbytes. The CPU time needed in this case ranges from 15 time
the real time to about 5 and is obviously dependent on the time step. The time interval used
during the accident simulations here reported is 25 ms, shorter than the usual one.
The test of SPES-2 considered for the initial TRAP-2 validation simulates a 2" cold
leg break of AP-600; the break is located between the cold leg to CMT balance line and the
pressure vessel.
The system, initially at nominal conditions, undergoes a slow depressurization owing
to loss of coolant, which is opposed by the pressurizer heaters action as long as the liquid
level into the pressurizer is greater then the heaters cut-off threshold. When the system
pressure reaches the limit of activation of the reactor trip on low pressure, the signal is sent
to the Steam Generators. After 2 s the SG isolation valves are closed. After about 4 s the
power channel heaters are reduced at 20% of nominal power and start to follow the SCRAM
curve. Since this instant the pressure slows down more rapidly because the SG exchanged
power is greater than the residual power given to the fluid from the channel heaters. During
a short period this unbalance is reversed and causes a small rise of the pressure, together
with the circuit flow rate reduction due to the RCP trip.
The activation of the CMTs and of the PRHR system cools down the primary fluid, so
opposing to the pressure rise and maintaining the system pressure practically constant for
few minutes.
We resolve to proceed with the TRAP-2 stagnation fluid model after the primary flow
rate has reached its natural convection value, sufficiently low to justify a dramatic
simplification of the simulation. This passage allows a noticeable computation time
reduction without lack of accuracy for the passive safety systems.
The CMTs cooling flow rate increases, due to the entrance of some steam in the
rising ducts toward the tanks, and then a slow pressure descent starts. The injection of the
accumulator fluid at room temperature begins when the system pressure is under the

intervention threshold: this event does not modify the pressure descent because it reduces
substantially at the same time the natural convection into the CMT loops. The CMT is
described with a distributed parameter duct, and this prevents to calculate the actual water
level. Therefore, the first train of ADS valves opens at a fixed time, thus leading to a
sensible augmentation of the depressurization trend; the subsequent increase of the
accumulators flow rate stops finally the circulation in the CMT loops.
The result of this first level of validation seems satisfactory as far as the system
pressure is concerned. However, a more rapid pressure descent in the first part of the
simulation in comparison with SPES-2 data leads to an anticipation of the SCRAM signal
and of the subsequent activation of CMTs and PRHR loops, while a higher accumulators
flow rate causes a steeper decreasing slope after ADS activation. The first type of
discrepancy is probably due to an incorrect evaluation of the circuit heat losses for their
insufficient knowledge, the second one to an underestimation of the stagnation model in
predicting the exact value of the pressure at the connection of the DVI line with the pressure
vessel. The LOCA flow rate follows the pressure behaviour and fits quite well the mean
value of the experimental data.
Other differences between TRAP-2 prediction and experimental data are referred to
the CMT flow rates when the steam-water mixture reaches the inlet of the riser duct of the
two loops: TRAP-2 depicts an analogous behaviour for both, i.e. a lower flow rate and a
sudden reduction when the accumulators start to inject their own fluid, while SPES-2 data
show a delayed flow rate peak of one CMT with respect to the other. Even this is caused by
the approximation used by the stagnant model: a lack of simulation of the primary circuit
components (cold and hot legs, steam generators, pumps, pressurizer, downcomer), of the
consequent circuit asymmetry and of particular flow conditions (stratified and countercurrent
348

flow regimes) give incorrect predictions on pressure and flow distribution.

Finally, the comparison of the PRHR behaviour shows a certain difference in the
initial and mean flow rate and in the outlet temperature, probably caused by the absence of
heat losses in the connecting pipes (unknown) and by a lower value of the calculated heat
transfer coefficient with respect to the real one.
6.2.

MARS Steam Line Break accident.

TRAP-2 program has been utilized by ENEA-RIN for a further analysis of the MARS
project safety features [7]. The main safety evaluation has been carried out with
RELAP5/mod2.5 program and the availability of these results allows us a validation of
TRAP-2 with a best-estimate code.
The 600 MWth MARS (Multipurpose Advanced Reactor Inherently Safe) one single
loop reactor relies on a totally inherent and passive safety concept. The key issue of
residual heat evacuation in case of accident is solved through a completely passive
Emergency Core Cooling System (ECCS) which consists of two independent circuits based
on natural circulation triggered by passive check valves, which are activated by the primary
pump trip. Each train consists of:
- an ECCS Primary Loop directly connected to the pressure vessel at the same pressure
as the primary system;
- an ECCS Secondary Loop at the same temperature and pressure as the ECCS Primary
Loop;
- a tertiary Pool and Condenser Loop consisting of a water reservoir (Pool) at the ambient
temperature and pressure and an air-condenser connected to the pool.
The MARS nodalization as used by TRAP-2 (version /2M especially devoted to
MARS system) is sketched in Fig. 4 and is the same utilized by RELAP (91 nodes). The set
of transients performed includes Station Blackout, Steam Generator Tube Rupture and
Steam Line Break [8].
The TRAP/RELAP validation is still in progress: in this paper we report two graphs
as an example, referring to the initial period (400 s on a computed total of 7000 s) of Steam
Line Break accident. The first (Fig. 5) relates to the core flow rate, the second (Fig. 6) to the
thermal power exchanged between the ECCS Loops: the comparison is really good apart a

Fig. 4: TRAP-2 nodalization of MARS reactor.

349

-IIIL_

Fig. 5: TRAP/RELAP comparison on MARS SLB accident.

Fig. 6: TRAP/RELAP comparison on MARS SLB accident.

slight discrepancy in the power due to a different evaluation of the heat transfer coefficient
in such a low flow condition.
The transients were performed on a PC586/66 MHz with a time step of 100 ms,
requiring a CPU time of 1.1 second per simulated second.
7.

CONCLUDING REMARKS

A fast running program easily tailored to several reactor systems belonging to PWR
concept has been presented.
It showed to be versatile, user-friendly and useful for parametric studies concerning
safety and operability. The first validation effort shows its numerical consistence, and a
satisfactory agreement with a best-estimate program and experimental tests.
This result allows us to proceed in the validation together with a further refinement of
the model, always retaining the initial hypothesis to disregard sound speed effects.

350

NOMENCLATURE
SYMBOLS

cp
Cb
De
/
g
h
Nu
p
Pr
q"
Ra
Re
t
T
2

GREEKS

specific heat at constant pressure

boron concentration
^
equivalent diameter
P
friction factor
a
gravitational acceleration
r
specific enthalpy
9
Nusselt number
h
pressure
Prandtl number
INDEXES
superficial heat flux
fg
Rayleigh number
I
Reynolds number
Is
time
m
temperature
w
spatial abscissa
V

n
n

roughness
dynamic viscosity
density
superficial tension
mass flow rate
inclination angle

heated perimeter
cross section area
latent heat of evaporation

liquid
liquid at saturation
mixture
wall
vapour

REFERENCES

[1] BABALA, D., HANNERZ, K., Pressurized water reactor inherent core protection by
primary system themnohydraulics, Nud.Sci. & Eng. 90 (1985) 400-410.
[2] CAIRA, M., CUMO, M., NAVIGLIO, A., "Energy security: alternatives to Oil", MARS

nuclear plant: an Italian proposal for an 'inherent safety1 nuclear reactor, US DOE and
Italian MICA meeting, Argonne, Illinois, USA (1988).
[3] CINOTTI, L, DAFANO, D., Contributi italiani al nucleare di 2 generazione: il sistema

ISIS, Energia Nucleare 1 (1990) 48-58.

[4] RICOTTI, M., TRAP: modello peranalisi di sicurezza e operabilita di PWR convenzionali
e di nuova concezione, .applicazioni e approccio neurale alternative, PhD Thesis,
Polytechnic of Milan, Milan (1993).
[5] RIGAMONTI, M., SPES-2: the full height full pressure Italian test facility simulating

AP600 plant, main results from the experimental campaign, these Proceedings.
[6] CHAO, Y., ATTARD, A., A resolution of the stiffness problem of reactor kinetics,
Nud.Sci. & Eng. 90 (1985) 40-46.
[7] ENEA-Nuclear Energy Division, MARS project: Safety Evaluation Report of MARS
Reactor (RVS), MT.GLM.15 (1994).
[8] VETTRAINO, F., RICOTTI, M., SORDI, R., "ICONE-3 International conference on
nuclear engineering", Dynamic response of MARS reactor under design basis accident
conditions, Kyoto, Japan (1995).

351

HEAT AND MASS TRANSFER PHENOMENA IN

INNOVATIVE LIGHT WATER REACTORS

W. AMBROSINI, F. ORIOLO
Universita degli Studi di Pisa,
Dipartamento di Costruzioni Meccaniche
e Nucleari, Pisa

G. FRUTTUOSO, A. MANFREDINI
THEMAS s.r.l., Pisa
F. PAROZZI, M. VALISI
ENEL-ATN, Milan

Italy
Abstract
The paper describes the work performed in the development and the
application of models for predicting heat and mass transfer phenomena
relevant in innovative LWRs. Film condensation and falling film
evaporation are addressed comparing model predictions with separate
effect experimental data. Development of models for non-equilibrium code

application is also discussed.

1. INTRODUCTION

Heat and mass transfer phenomena play an important

role in innovative light water reactors, being at the basis
of some passive safety features designed to mitigate
postulated accident consequences.
Passive containment cooling, in particular, is one of
the key issues in innovative reactors that has
contributed to the new interest for these phenomena in
the nuclear field In fact, some passive reactor
containments are conceived to allow for decay heat
removal by means of a combination of condensation
onto a inner heat transfer surface and convectionradiation or convection-evaporation on the outer
surface. AP600 [1] and SBWR [2] plant concepts are
relevant examples of this strategy of containment
cooling, asking for a better understanding of heat and
mass transfer phenomena.
In consideration of the requirement to maintain
these plants in safe conditions for a very long period
(three days or even more) after a postulated accident,
with no operator action or external intervention, the
need is felt for an improved knowledge of the involved
phenomenology, in order to achieve models capable to
predict the behaviour of reactor plants during such long
lasting transients with a reasonable accuracy.
As a consequence, the enlargement of the available
data base on evaporation and condensation in the
presence of noncondensable gases has been the target of
experimental investigations carried out by different
organizations (see e.g., refs. [3], [4], [5], [6] and [7]).
These studies involve both separate effect and integral
test facilities and are aimed at achieving a better
knowledge of the capabilities of passive cooling
systems.

On the basis of these new data, models available in

the literature can be reconsidered and assessed to
ascertain their capabilities in providing quantitative
estimates of the related heat and mass transfer rates.
Both heat transfer and fluiddynamic aspects are
involved in this respect, the greatest interest being on
falling film evaporation and filmwise condensation.
Moreover, the way to account for the presence of
noncondensable gases in the prediction of heat and
mass transfer rates in other conditions, as pool
evaporation or condensation, requires an improvement
with respect to previously adopted models.
It must be also emphasized that besides the interest
that heat and mass transfer phenomena have for passive
reactor concepts, they are relevant also for present
commercial nuclear power plants in many operating
situations. In particular, during some beyond-DBA
conditions, a degraded performance of the active
safeguards can make passive heat removal mechanisms
to turn out as the ultimate possibility for mitigating
accident consequences.
In the present work problems related to heat and
mass transfer in innovative reactors and in code
applications are reviewed collecting material partly
produced in the frame of a cooperation between the
University of Pisa, ENEL and THEMAS s.r.l., partly
developed independently by the University. The
common feature connecting the presented applications
is the adoption of the heat and mass transfer analogy for
predicting evaporation or condensation in conditions of
interest for light water reactors.

353

In the following, after short notes about the

treatment of heat and mass transfer phenomena in the
mentioned analogy, evaporation of falling water film,
filmwise condensation and general thermal nonequilibrium modelling problems will be discussed.

whenever vapour diffusion occurs from or toward

the interface, an advection fluid velocity is
generated, given by the relationship
w=

2. PHYSICAL BASIS OF THE MODELS

2.1 Modelling Heat and Mass Transfer Phenomena

Whenever at an interface between two phases of a
multicomponent mixture evaporation or condensation
occur, the mass transfer is accompanied by a
simultaneous heat transfer. This characteristic makes
heat and mass transfer processes very peculiar in the
class of phenomena governed by diffusion of chemical
species.
A simple energy balance (or jump condition) at the
immaterial surface separating the phases (interface) is
used to put a relation between the heat flux and the
specific evaporation-condensation rate (cfr. e.g., [8])
(1)
Z-i (qk.int+ k.mt hk.int)=
k=l
This equation, together with the mass jump condition
2
k=l
puts a relation between the heat exchanged by
convection between each phase and the interface and
the related mass transfer, stating that no mass transfer
can occur without a corresponding heat transfer.
In the presence of noncondensables in the gaseous
phase the evaporation-condensation rate depends on the
resistance within the boundary layer located at the
interface, in which the major changes in the
concentration of the components occur. In this narrow
region close to the interface, the multi-D continuity

equations can be combined with the Pick's law of

diffusion, to give the overall mass conservation
equations. For a two-component mixture made of
vapour and a noncondensable gas these equations are

(5)

Pn
which is the result of the condition that the
noncondensable gas has zero net flow across the
interface;
because of the superposition of diffusion and
transport, given the partial pressures of the vapour
in the bulk fluid and at the interface, the specific
mass transfer rate has a logarithmic form

(6)

v.mt

assuming that the partial pressure of the vapour at

the interface, where phase change occurs, is equal
to the saturation pressure at the local temperature,
the mass transfer rate is finally linked to liquid
phase conditions.
In the above equation (6), the thickness of the
boundary layer is needed to calculate the mass transfer
rate. To this purpose it is noted that in two dimensions
equation (3) for an incompressible fluid can be written
as
5p

dp

dp

By comparing this equation with the energy equation

for a fluid with negligible frictional heating
(8)

the well known analogy between heat and mass transfer

is established, allowing to relate the boundary layers for
heat and mass transfer.
For our purposes, this analogy puts a
correspondence between the dimensionless groups
adopted in the semi-empirical correlations to be used
for heat and mass transfer (cfr. [11]):
Nu

Sh=-

g = ' k0

(3)

Scg =

ag
(4)

Restricting to one-dimensional steady processes in

the proximity of a plane interface permeable only to the
vapour and orthogonal to the y Cartesian coordinate and
assuming perfect gas behaviour for the two components
and constant pressure, some well known results can be
easily obtained from the above equations (cfr. e.g., [9],
[10]):

354

(9)
'vn
(10)

As a result, assuming a dependence of heat transfer

coefficient on Pr-33, the following formulation is
obtained for the mass transfer constant [12]

(11)

Finally, since the interfacial heat fluxes in equation

(1) can be expressed as a function of the interfacial
temperature, the energy jump condition resulting after
substitution of (6) can be taken as an implicit definition

of this variable [13], [14]

2

between the constant film thickness and flow rate

obtained in the absence of mass transfer are used to
represent the relation between the local values of the

Km X pv In
p

* 0

k=l
(12)

where the phasic interfacial heat fluxes are evaluated

using appropriate heat transfer coefficients.
Due to the nonlinearity of the diffusive expression of
mass transfer rate, iterative techniques are needed to
solve the above equation, unless lower level
approximations with, eventually, correction factors [15]

are retained sufficient for particular purposes.

2.2 Modelling Falling Film Phenomena

When a liquid film falls down an inclined surface, a

velocity field is established within its thickness which
depends on fluid properties, flow rate per unit perimeter
and boundary conditions. The liquid film Reynolds
number, defined as
4T
(13)
Reif=

" u
is generally used for discriminating between the
possible flow regimes. In particular [see e.g., 11]:
for Rey < 30, a laminar film with a smooth
interface is generally reported;
for 30 < Rejf < 1800, superficial waves are
observed although the film is still laminar;
for Rejf > 1800 the falling film is completely
turbulent.
The Nusselt theory [16] can be adopted for
evaluating the thickness of laminar-smooth falling
liquid films. In the presence of interfacial shear, the
relation between the film thickness and the flow rate per
unit perimeter is the following:
g cos6 pi (pi -pBe) m3 PIm 2 Ti
(14)
-
Depending on the value of the interfacial shear stress,
TJ, various flow regimes may be identified, ranging from

same variables when mass transfer occurs.

Under these assumptions, the change along the
surface in the film flow rate due to mass transfer is

given by:
- G
(18)
dz ~" v.int
In the classical treatment of the condensation of a pure
vapour by Nusselt, the above equation, together with
(14) (in which the shear stress is neglected) is used to
obtain a relationship for the derivative of the film
thickness
^dm_ u l k l( T sat- T w)
m dz
Pl(Pl-P v )g*adopted to calculate the film thickness at the bottom
end of the surface. On the basis of equation (18) a more
complicated expression is reached in the present work
accounting for the presence of noncondensable gases
(see Sect. 5).
The value of the film thickness is needed to
determine the thermal resistance between the interface
and the wall. In the classical Nusselt theory, the
equivalent conductance of the liquid film is due only to
heat conduction

When the effective heat transfer coefficient between

the bulk fluid and the wall owing to the combined effect
of convection and mass transfer must be derived, the
averaged values of film thickness and interfacial and
wall temperatures are used to give:
Mnf *
(21)
H
eflF-%T g -T,
It is reported that the application of the Nusselt
theory leads to underestimate the effective heat transfer

(flooding)

coefficients. The reason is that both waviness and

turbulence actually reduce the equivalent thickness to be

conditions.
A more compact way to express the same result is

considered in equation (20) and contribute to reduce the

diffusion boundary layer.

complete downflow,

to

reverse flow

the following

The stability of the falling film has been an

,
^ o.s.
mf =2 Relf

(15)

interesting subject of research since the pioneering work

by Kapitsa [18]. In particular, theoretical and
experimental evidence has been reported supporting the

(16)

where
mw
v
l

following conclusions:

a liquid film falling down a vertical plate is

unstable at all the Reynolds numbers, i.e.

superficial waviness is always present; surface

and
(17)

having the form of relations often adopted for annular

flow in pipes (see e.g., [17]).
The above formulations hold for a steady and
smooth laminar film in the absence of mass transfer. In

the case of evaporation or condensation, it is assumed

that the effect of mass transfer on the velocity field
inside the film is negligible. Moreover, the relationship

inclination has a stabilizing effect [19];

condensation tends to stabilize the liquid film,

while evaporation has the opposite effect (see e.e.
[20]);
the structure of waviness is rather chaotic and at
sufficiently high Reynolds number consists of large
roll waves riding over a thin smooth substrate (see

355

The effect of turbulence and waviness on heat

transfer is not specifically addressed in the models
described hereafter and will be matter of further studies.
Anyway, in the presence of noncondensable gases these
effects are often second order ones.
3. FALLING FILM EVAPORATION
3.1 Experimental set up

A facility has been built at the University of Pisa to

investigate the evaporation of falling water films in
conditions typical of the AP600 Passive Containment
Cooling System.
As known, the containment system of this
innovative reactor consists of a steel safety envelope
located inside a concrete building (Figure 1) in which
natural circulation flow paths exist for the external air
in order to cool down the steel surface by convection
and radiation after a postulated accident Water tanks
are also provided for spraying water on the outer steel
surface, thus enhancing removal of decay heat by falling
film evaporation.

Air Outlet

Steel
Envelope

J*J \L

Spray
Tanks

Concrete
Building

Figure 1 - Sketch of AP600 Containment System

The experimental facility (Figure 2), designed
considering the conditions occurring during loss of
coolant accident scenarios [23]. consists of a AISI 304
L flat plate (2 x 0.6 x 0.022 m.). supported by a metallic
frame. The surface underwent a preparation by
sandblasting and by a spray painting of about 0.3 mm of
metallic zinc, in order to make uniform the thermal
emissivity of the wall and to reproduce the mean surface
characteristics of the containment shell of AP-600.
The supporting frame allows for a rotation of the
plate up to 90 degrees with respect to the vertical axis,
so that the behaviour of liquid films can be analyzed
using different inclinations of the surface; situations
occurring in the elliptical containment dome can be so

reproduced. To simulate the heating of the containment

wall that would occur in an actual plant following a
postulated accident, the test plate is heated from the
back using 100 modular electric heaters (195 x 95
mm.), subdivided into three different electrically
356

Figure 2 - Facility for film evaporation tests

equilibrated groups. The total heating capacity installed
is approximately 36 kW and is supplied by three FED
regulated electric generators with a maximum voltage
difference of about 50 V ic . The electric power
supplied to the heating resistances is measured with an
on-line electronic wattmeter.
A Plexiglas baffle plate, located at a distance of 10
cm from the plate, faces the surface along which the
liquid film flows. Air from an axial fan flows up
through the duct: the fan has a variable speed motor
drive that allows varying air speed from 0.5 to 10 m/s.
Cooling water is preheated with electric resistances
up to a selected temperature and is supplied at a
prescribed rate to some interchangeable sprays nozzles.
The spray system is provided with a centrifugal pump;
the cooling water flow rate can be regulated with a bypass flow path with related valves. The measurements of
the following thermofluid-dynamic variables can be
performed in the apparatus:
inlet and outlet liquid film temperatures:
temperatures within the plate thickness;
film thickness;
water and air flows;
electrical power supplied to the system;
average temperature and humidity of air flow.
All the measurements of test parameters are
recorded and processed on a personal computer.
3.2 Correlation of first results

During the last year, the facility was subjected to a

complete revision, comprising the displacement of the
structure in a more convenient location and acquisition
of new instrumentation. Therefore, the results here
presented are those obtained in a previous shakedown
campaign [24].
Three series of tests have been run during the
shakedown of the facility:

1. tests for characterizing the plate from the point of

view of temperature distribution and heat losses;
2. tests for evaluating the heat transfer capabilities in
dry conditions (no water film);
3. tests for evaluating the heat transfer capabilities in
wet conditions (with water film).
During the first series, the needed information on
the plate behaviour during heating was achieved; the
heat losses were also estimated for further use in data
processing.
The data collected in the second series of tests
should represent containment cooling by natural
circulation of air between the baffle and the steel
envelope. A good agreement could be obtained between
data and a semi-empirical correlation for heat transfer
in a short duct (2 < LfD^ < 20) [9]:

As the above described results demonstrated the

reliability of the heat and mass transfer analogy for
describing evaporation phenomena, this approach was
taken as a basis for setting up a more complete model
obtained reviewing and improving relationships
previously developed for the FUMO code [26]. In this
model, the laminar-smooth film relationships (see Sect.

2) have been adopted for calculating the film thickness

and an estimate of the film temperature evolution along
the plate is obtained with the following energy balance
equation

&T
n
n
= q
"G
dz w *
where the evaporation rate is evaluated according to Eq.
(6). The following formulation is adopted for
temperature profile in the film thickness.

Nu / Pr = 0.023 [ 1+ (Dfc/L) -7] Re -8 (22)

For low Reynolds numbers, anyway, the convective heat

transfer for the dry plate surface appeared to be
conservatively underpredicted by the correlation.
The senes of data with falling water film was
correlated making use of the heat and mass transfer
analogy. Adopting Eq. (22) for heat transfer in dry
conditions, the corresponding relation for mass transfer
is:

Sh / Sc 33 = 0.023 [ 1 + (Dh/L)0-7] Re -8

(23)

The experimental value of the Sherwood number is

calculated on the basis of the observed evaporation mass
velocity, Gv.
In turn, Gv should be determined on the basis of the
experimental values of injected and collected water
during each test. Since the measurement chains for
measuring the injected and collected mass flow rates
were not yet qualified, an energy balance was adopted to
infer the evaporation rate on the basis of the power
supplied to the plate and of the measured film internal
energy variation from inlet to outlet.
Thus, two different estimates of the evaporation rate
were obtained:
a maximum value, calculated assuming that the
difference between the power fed to the plate and the
variation of the liquid film sensible heat represents
the e\aporation power;
a Best Estimate (BE) value, obtained taking into
account theoretical predictions of the heat losses and
of the power exchanged by convection from the film
surface to the external air.
The values of both maximum and best estimate
values of the dunensionless group Sh/Sc033 have been
reported in Figure 3 . It can be noted that:
maximum and best estimate data appear very close
to each other, confirming the adequacy of the
adopted approach to infer evaporation data;
BE data are well represented by the relationship
reported in Eq. (23) for short tubes, although in
agreement with a similar research by Westinghouse
[25] an overall underprediction seems to be obtained
by the theoretical estimates.

T(y,z) = C0(z)

y2

+ q(z) y + C2(z)

(25)

where the functions C0, Cj and Cj can be found on the

basis of the boundary conditions and of an integral form
of the energy balance equation (24).
These relationships are then solved in axial meshes
in which the plate has been subdivided, adopting an
iterative technique for estimating the interfacial
temperature. Despite of the simplicity of the models
adopted for film fluiddynamics, a good performance in
predicting the evaporation rate in the first performed
tests has been obtained (Figure 4).
1000

K"
.J***
100 t^^-

Experiment BE

Experiment max

Theory short
duct

10

10000

100000
Air Reynolds number

1000000

Figure 3 - Correlation of first data for wet tests

0.003

S
0.002
21
o Ol
6.Jt

0.001

0.001 0.002

0.003

Experimental (BE) evaporation rate [kg/s]

Figure 4 - Calculated vs. expenmental evaporation rates

4. FILM CONDENSATION
4.1 Considered experimental data
The SBWR reactor is equipped with a passive
containment cooling system capable to remove decay
heat after postulated accidents [2].
In particular, two different systems are introduced to
accomplish with this function:

357

 the isolation condenser system (1C);

the passive containment cooling system (PCC).
The first one consists of a loop connected with the
reactor pressure vessel containing a condenser
submerged in an external water pool; it is intended to
allow for reactor depressurization by directly
condensing the vapour produced in the core. On the
other hand, the PCC circuit is connected to the dry-well
atmosphere and contains a separate condenser, also
submerged in an external pool.
Although the function of the two systems is nearly
the same, their ranges of application and the operating
conditions are different. In particular, the 1C performs
decay heat removal by condensing pure vapour at high
pressure conditions, while the PCC system has to cope
with the problem of noncondensable gases present in
the containment atmosphere and operates at far lower
pressures.
Both systems have been the subject of interesting
experimental researches aimed at providing data on the
SBWR plant accident behaviour during their operation
(see e.g. [4], [5], [6] and [7]). In particular, the data
provided by the PANTHERS experimental program [6]
are used in the present paper for validating models for
the prediction of heat and mass transfer in filmwise
condensation conditions.
The program is part of a more general experimental
effort including also downscaled tests for the study of
blowdown (GIST) and PCCS performance (GIRAFFE
and PANDA) [27], [28]. In this frame, PANTHERS
tests supply information on full-scale behaviour,
making use of prototypical condensers for both 1C and
PCCS.
Notwithstanding the wide knowledge existing in the
nuclear field about condensers, the full-scale
experimental investigation is justified by the
peculiarities of the design of 1C and PCCS. In
particular, in the case of PCC the most challenging
modelling aspect is the operation of the condenser with
considerable fractions of noncondensable gases. The
main objective of the tests is to confirm the adequacy of
the design of both apparatuses in meeting the
requirements for use in SBWR.
Figure 5 shows a sketch of the PCCS configuration
of the PANTHERS facility installed at SIET in
Piacenza, It consists of the following main parts:
the PCCS two-module unit submerged in a pool
tank;
piping for injecting vapour or vapour-air mixtures at
prescribed conditions of flow rate, pressure and
temperature;
a condensate tank in which liquid from the
condenser is drained; during tests pressure in this
tank is maintained equal to the pressure of the inlet
mixture;
a vent tank, maintained at a pressure lower than the
inlet mixture, to which noncondensable gases from
the condenser outlet are purged;
a waier make-up system to restore the required level
in the condenser pool.
358

Steam

supply

||

LCV/I

Figure 5 - Sketch of the PANTHERS Facility in the

PCCS configuration [from Ref. 6]
The loop elevations are the same as in SBWR; in

particular, the position of the pool levels and of the loop

seal in the drain line have been kept as in the reactor.
The required power to feed vapour in the test rig is
supplied by a nearby power station. A maximum of 6.0
kg/s of steam at 17 MPa and 540 C can be obtained.
The tests considered in the present work are related
to PCCS performance. They have been carried out by
injecting known flow rates of vapour and air into the
prototype condenser and measuring the corresponding
condensate flow rate in steady-state conditions.
To quantify the effect of the noncondensable gas on
the capability to suppress steam, a condensation
efficiency is defined by the relationship [6]
__ .
Condensate Flow Rate
Efficiency = Inlet Steam Flow Rate
<26>
which is one of the most interesting data to be predicted
by models.
4.2 Model characteristics and performance
In consideration of the ongoing experimental program,
a cooperation started between ENEL, THEMAS s.r.l.
and University of Pisa to set up and qualify a
component model for 1C and PCCS. The objective of
the activity was to acquire experience in the simulation

of condensation phenomena in the presence of

noncondensable gases in order to set up models for long
term analysis of the accident behaviour of SBWR.

The ICONA (Isolation CONdenser Analysis)

program was then conceived as a numerical component
model to be validated as a stand-alone module. The
main characteristics of the code are the following:
mixture balance equations for mass and energy are
solved in control volumes;
allowance for an arbitrary number of

noncondensable gases is made;

thermal equilibrium between liquid, vapour and
gases is considered;
control volumes are connected by junctions through
which homogeneous flow is assumed;
allowance is made for countercurrent flow
conditions at junctions;
a full range heat transfer package is adopted and a
cubic coarse-mesh algorithm is used for treating
heat conduction in plane, cylindrical and spherical
structures;
phase separation in volumes is allowed to simulate
water pools;
a semi-implicit numerical method is adopted to
solve balance equations.
Filmwise condensation is dealt with in the code by
the heat and mass transfer analogy. A steady-state
energy balance equation is written for the film assuming
the form

been also represented.

The performance of the code has been evaluated
comparing the calculated condensation efficiency with
the experimental values. Figure 6 reports the results
obtained in this comparison for the available data
points. The agreement between the calculated and the
experimental efficiency is remarkable, also considering
the limitations of the adopted numerical frame and the
simplicity of the nodalization.
^s~

g.1.00

io.80
+ l*-y
b) 0.60

^ ^>>

0.40
3

^^
nnn

v-v^

^^
J^^

-20%

-""

^^

0.00

0.20 0.40 0.60 0.80 1.00

Experimental Efficiency

Figure 6 - Overall model performance for PCCS tests

Figure 7 and Figure 8 report the same information

in the form of the ratio between the calculated and the
experimental efficiencies as a function of test pressure
and air molar fraction. Some clear trends can be noted:
the efficiency tends to be slightly overestimated at
W
I*-Tw)
(27> low pressure and underestimated at high pressure;
the spread around unity is increasing with
P' Psat(Tint)
increasing air fractions.
P-Pv
The first observation reflects the tendency of the
model to slightly underestimate the slope of the
which expresses that the heat flux transferred by the
efficiency curve as a function of pressure. The second is
liquid film to the wall is equal to the summation of the
partly
due to the fact that with a higher noncondensable
two contributions due to convection and condensation
fraction
a lower efficiency is generally obtained, thus
[13]. This is equivalent to assuming no variation in the
enhancing
the relative error of the calculation.
sensible heat of the liquid film along the structure.
The convective heat transfer coefficients are
evaluated according to classical single phase convection
1.4
relationships. The difference between the gas and the
film velocity is taken into account for calculating the
gas Reynolds number entering the heat transfer
; .
correlations.
The film thickness is evaluated in each hydraulic
0.8
mesh in consideration of the calculated liquid film flow
rate by the Nusselt theory, with allowance for interfacial
0.6
0.2
OJ
0.4
0.5
0.6
0.7
shear (equation 14).
Test Pressure [MPaj
Equation (27) is then solved iteratively for the
interfacial temperature. Then, an overall condensation
Figure 7 - Model performance versus test pressure
heat transfer coefficient is evaluated by the relationship
H

cond =

T g -T w

(28)

This heat transfer coefficient is finally used to define

the convective boundary conditions of the condenser
structures.
In the simulation of the considered PCCS tests, the
ICONA code has been used with six axial nodes to
represent the condenser pipes and two nodes for the
steam and water boxes. The external pool, the
condensate drainage line and the external tanks have

iu.
A '

'

9.8

AA

0.02

0.04 0.06

0.08 O.I

Air Molar Fraction

Figure 8 - Model performance versus air molar fraction

359

0.12

The above results must also be considered taking

into account that the experimental data are preliminary.
More decisive conclusions will be obtained on the basis
of the final data release.

allowance for countercurrent flow at junctions is

made by proper correlations; models for critical flow
are also included;
a full range heat transfer package is adopted and a
cubic coarse-mesh algorithm is used for treating

5. NON-EQUILIBRIUM CODE APPLICATIONS

It is well known that the treatment of heat and mass

transfer is of particular concern in codes for the
thermal-hydraulic analysis of nuclear power plants
when thermal non-equilibrium is adopted. In fact, with
respect to the older equilibrium models, the separate
mass and energy balance equations require the explicit
definition of terms for mass and heat transfer between
the phases.
In the presence of noncondensable gases, the
evaluation of evaporation and condensation can be
performed by means of diffusion approaches as those
outlined in the previous sections. In the thermalhydraulic module of the aerosol transport code ECART
[29] this approach has been used to treat pool interfacial
heat and mass transfer and wall condensation.
ECART (ENEL Code for Analysis of
Radionuclide Transport) [30] is based on a mechanistic
approach to vapour and aerosol phenomenology and is
aimed at unifying reactor coolant and containment
system analysis, representing the current state-of-the-art
of LWR severe accident aerosol codes. The code has
been developed by ENEL with the support of Synthesis,

heat conduction in plane, cylindrical and spherical

structures;
phase separation in volumes is allowed to simulate
water pools with the related scrubbing effects;
a semi-implicit numerical method is adopted to
solve balance equations.
The interfacial heat and mass transfer terms
appearing in balance equations are calculated
combining approaches adopted in the FUMO
containment code [31] with an updated treatment of
interfaces. At present, interfaces are considered at the
two locations of the pool surface and of the condensate
film on heat structures facing the atmosphere.
In the former case, due to the presence of
noncondensable gases, the interfacial temperature is
determined making use of a linearized diffusive
approach included in the energy jump condition across
the pool surface:

(29)

where

Themas, University of Pisa, and Politecnico of Milano.

Most of model development and validation actions

related to this computer code were partially sponsored
by the European Commission in the frame of safety
research programs since 1989. Through an agreement

between ENEL and EdF and joint research actions on

LWR severe accident studies, EdF started a significant
financial and technical contribution to the development
and validation of ECART, that became of ENEL and
EdF common property.
ECART
mainly
consists
of
three
phenomenological sections: one for vapour and aerosol
phenomena, one for the chemistry, and one section
dedicated to thermal-hydraulics. The latter has been
recently developed in order to be directly coupled with
the other two sections and to supply them with the
boundary conditions required for a realistic evaluation
of the physical phenomena, with reference to a steamwater-noncondensable gas mixture, in conditions that
are expected during radionuclide release and transport.
The main characteristics of the thermal-hydraulic
module are summarized as follows:

the model makes use of separate balance equations

for mass and energy in the liquid pool and in the gas
atmosphere;
thermal equilibrium is assumed within the pool and
the atmosphere;
up to nine noncondensable gases can be modelled,
chosen among the most important ones for severe
accidents and source term experiments;
control volumes are connected by junctions through
which homogeneous flow is assumed;

360

[
JBj, lrsat
m |_Pv + j^ PnJ ~dT

(30)

and H mt and HI mt are convective heat transfer

coefficients calculated assuming natural convection on a
flat horizontal plate.
The last term at the R.H.S. of Eq. (29) represents the
heat flux corresponding to mass transfer
WPv>)

(31)

defined as positive for evaporation and negative for

condensation. An intuitive criterion to establish the
conditions for evaporation and condensation as a
function of the vapour partial pressure in the bulk
atmosphere is readily obtained from the above equation.
A similar approach is adopted in the case of steady
condensation on vertical walls, using the Nusselt theory
for film dynamics and considering the effect of
noncondensable gases. The balance equation required to
calculate the evaporation rate involves now the wall
temperature

rw .)
int

(32)

where
is evaluated from a formulation similar
to Eq. (30). As the condensation rate depends on the
film thickness in a more complicate way than in the
case of condensation of pure vapour, the use of Eq. (18)
leads to a more complicate relation for film thickness
derivative than expressed by Eq. (19)
m ,dm =a b + c m

dT d

(33)

where m is the film thickness and a, b, c, d and e are

constants depending on fluid properties and heat
exchange conditions. A double analytical integration of
Eq. (33) over the whole length of the cooling surface
allows the evaluation of the total condensation rate and
of the average values of film thickness and heat transfer
coefficient.
The above relationships need a thorough
qualification on the basis of separate effect experimental
data to be assessed. A preliminary verification of their
adequacy was obtained in the first shakedown tests of
the code in which a pressurization transient in the
Caroline Virginia Tube Reactor (CVTR) containment
simulator was analyzed [32].
Figures 9,10 and 11 report the results obtained by
the stand alone version of the code. Although they must
be considered preliminary, it can be noted an overall
agreement which depends critically on the formulations
adopted for heat and mass transfer.
260000

Although the discussion was mainly focused on

innovative reactor features, it is clearly understood that
the validity of the conclusions reached is general and
hold for a wider spectrum of applications.
The above reported results point out the adequacy of
the heat and mass transfer analogy for the evaluation of
heat and mass transfer in nuclear reactors. It is
remarkable to note that very simple models, mostly
proposed in the first half of the century, have the
capability to catch the main features of delicate
phenomena as evaporation and condensation in the
presence of noncondensable gases in a large range of
parameters.
Subtle aspects of the related physics require to be
furtherly investigated to provide more reliable
predictions of heat and mass transfer rates. It is the case
of film and vapour-gas mixture fluiddynaniics,
particularly in relation to the effect of superficial waves
and film disruption, which have been neglected in the
present treatment.
The agreement obtained between experiments and
predictions is nevertheless encouraging to progress
along the same line of thought.

REFERENCES
100000

500

1000 1500 2000

2500

3000

Figure 9 - CVTR test No. 3: containment pressure

400

3000

Figure 10- CVTR test No. 3: upper compartment

temperature

500

1000 1500 2000

2500

3000

Figure 11 - CVTR blowdown test: lower compartment

temperature
6. CONCLUSIONS

In the present paper works performed in relation to heat

and mass transfer phenomena have been collected in the
attempt to attain an organic presentation of data and
modelling techniques.

[1] HJ. Bruschi and T.S. Andersen "The

Westinghouse AP600: the leading technology for
proven safety and simplicity", IAEA TCM Meeting
on Progress in Development and Design of
Advanced Water Cooled Reactors, Rome (I),
September 9-12,1991
[2] R.J. Me Candless, A.S. Rao, C.D. Sawyer "SBWR Simplifications in plant design for the 90's" IAEA
TCM on Progress in Development and Design
Aspects of Advanced Water Cooled Reactors,
Rome (I), September 9-12,1991
[3] F.E. Peters, A.T. Pieczynski, M.D. Carelli
"Advanced Passive Containment Cooling
Experimental Program", International Conference
on New
Trends
in Nuclear
System
Thermohydraulics, May 30-June 2, 1994, Pisa,
Italy
[4] H. Nagasaka, K. Yamada, M. Katoh and S.
Yokobori "Heat removal tests of isolation
condenser applied as a passive containment
cooling system", 1st JSME/ASME Joint
International Conference on Nuclear Engineering.
November 4-7, 1991, Keio Plaza Hotel, Tokyo.
Japan
[5] S. Yokobori, H. Nagasaka, T. Tobimatsu "System
response test of isolation condenser applied as a
passive containment cooling system", 1st
JSME/ASME Joint International Conference on
Nuclear Engineering, November 4-7, 1991, Kek>
Plaza Hotel, Tokyo, Japan
[6] P. Masoni, S. Botti, G.W. Fitzsimmons
"Confirmatory Tests of Full-Scale Condensers for
the SBWR", ASME, March 1993
[7] F. D'Auria, P. Vigni, P. Marsili "Application of

361

RELAP5/MOD3 to the evaluation of Isolation

Condenser performance", Int. Conf. on Nuclear
Engineering (ICONE-2), San Francisco (US),
March 21-24
[8] N.E. Todreas and M.S. Kazimi "Nuclear Systems"
Vol. I-II, Hemisphere Publishing Corporation,
1990
[9] F. Kreith "Heat Transfer Principles" (in Italian),
Liguori Editore, Napoli 1974

[10]R.B. Bird, W.E. Stewart, E. N. Lightfoot

"Transport Phenomena", John Wiley & Sons, 1960
[11] A. Bejan "Heat Transfer", John Wiley and Sons,
1993
[12]J.G.

Collier

"Convective

Boiling

and

Condensation." McGraw-Hill Book Company,

[25] T. Van De Venne, E. Piplica, M. Kennedy And J.

Woodcock "The Westinghouse AP600 Passive
Containment Cooling Test and Analysis Program."
ANP '92 Congress, Tokyo, October, 25-29, 1992
[26] A. Manfredini, M. Mazzini, F. Oriolo, S. Paci
"Validazione deH'efficacia dello spruzzamento

esterno del contenimento in impianti nucleari a

maggiore sicurezza passiva." 8th National
Congress on Heat Transfer of the UIT, Ancona,
June 28-29, 1990

[27]A.S. Rao, J.R. Fitch, and P.F. Billig "Safety

Research for the SBWR", The 20th Water Reactor
Safety Meeting, Bethesda, MD, 1992
[28] F. Magris, A. Villani, C. Medich and M. Bolognini

"Design and Experimental Verification of Isolation

Condenser and Passive Containment Cooler for
SBWR", International Conference on Design and

1972

[13] A.P. Colburn and O.A. Hougen "Design of Cooler

Condensers for mixture of vapours with noncondensing gas", Ind. Engng. Chem., 26,1178-82,
1934
[14] D. Butterworth, G.F. Hewitt "Two-Phase Flow and
Heat Transfer", Oxford University Press, 1979
[15] HJ.H. Brouwersand A.K. Chesters "Film models
for transport phenomena with fog formation: the
classical film model", Int. J. Heat Mass Transfer,
Vol. 35, No. 1, pp. 1-11,1992
[16] W. Nusselt "Surface condensation of water
vapour". Z. Ver. dt. Ing. 60 (27), 541-546; 60 (26)
569-575

[17] W. Ambrosini, P. Andreussi and B.J. Azzopardi

"A physically based correlation for drop size in
annular flow". Int. J. Multiphase Flow, Vol. 17,
No. 4,1991, pp. 497-507

[18] P.L. Kapitsa - Zh. Eksper. Teoret. Fiz. 18,3, 1948

[19] T.B. Benjamin "Wave formation in Laminar Flow
Down an Inclined Plate", J. Fluid Mech., 2, 554,

Safety of Nuclear Power Plants, 1992

[29] F. Oriolo, W. Ambrosini, G. Fruttuoso, F. Parozzi,

R_ Fontana "Thermal-Hydraulic Modelling in
Support to Severe Accident Radionuclide
Transport", International Conference on 'New
Trends in Nuclear System Thennohydraulics', Pisa,
May 30-June 2, 1994
[30] R. Fontana, E. Salina and F. Parozzi "ECART
(ENEL Code for Analysis of Radionuclide
Transport). Definition of the code architecture and
construction of the aerosol and vapour transport
module", ENEL Report. No. 1019/2, 1991
(unpublished work)
[31] A. Manfredini, M. Mazzini, F. Oriolo, S. Paci "The
FUMO code: Description and Manual" (in Italian)
- Universita di Pisa, Dipartimento di Costruzioni
Meccaniche e Nucleari, RL 416 (89), 1989
[32]R_C. Schmitt, G.E. Bingham, J.A. Norberg

"Simulated design basis accident tests of the

Carolinas Virginia Tube Reactor containment"
Idaho Nuclear Corporation, Final Report IN 1403,
Prepared for U.S.A.E.C, Contract No. AT(lO-l)1230,1970

1957

[20] G. Kocamustafaogullari "Two-Fluid Modelling in

Analyzing the Interfacial Stability of Liquid Film
Flows", Int. J. Multiphase Flow Vol. 11, No.
1,1985, pp. 63-89

[21] T.D. Karapantsios And AJ. Karabelas "Surface

Characteristics of Roll Waves on Free Falling
Films", Int. J. Multiphase Flow. Vol. 16, No. 5, pp.
835-852, 1990

[22] C.E. Lacy. M. Sheintuch and A.E. Dukler

"Methods of Deterministic Chaos Applied to the
Flow of Thin Wavy Films", AIChE Journal, Vol.
37. No. 4, pp. 481-489, 1991
[23] A. Manfredini, F. Mariotti, F. Oriolo and P. Vigni

"A facility for the evaluation of heat flux from a

plate cooled by a water film, with counter-current
air flow" 11th National Congress on Heat Transfer
of the UIT. Milano (I). June 24-26 1993
[24] W. Ambrosini, A. Manfredini, F. Mariotti, F.

Oriolo, P. Vigni "Heat Transfer from a Plate

Cooled by a Water Film with Counter-Current Air
Flow", International Conference on 'New Trends in
Nuclear System Thennohydraulics1, Pisa, May 30June 2, 1994

362

NOMENCLATURE

Roman Letters
Cp
specific heat at constant pressure [J/(kgK)J
C0
temperature profile coefficient
[K]
[K/m]
Cj
temperature profile coefficient
[K/m2]
C2
temperature profile coefficient
Dn

hydraulic diameter

[m]

.fiyj!

vapour diffusion coefficient

[m2/s]

G
h

mass velocity
fluid specific enthalpy

[kg/(m2s)]
[J/kg]

H
k

heat transfer coefficient

thermal conductivity

Km
L

mass transfer coefficient

length

m
tn

film thickness
molecular weight

Nu

Nusselt number

[W/(m2 K)]
[W/(mK)J
[m/s]
[m]
[m]

p
Pr
q"
Q

pressure
Prandtl number
heat flux
power

Re
Sc

Reynolds number
Schmidt number

[Pa]
[W/m2]
[W]

Sh
Sherwood number
T
temperature
[K]
[m/s]
w
fluid velocity
[m/s]
w*
shear velocity
[m]
x
space coordinate
y
space coordinate within the film [m]
z
space coordinate along the plate [m]
Greek Letters
[m2/s]
a
thermal diffusivity
F
mass flow rate per unit perimeter [kg/(ms)]
8m
thickness of mass transfer boundary
layer
9
inclination angle with respect to
downward vertical
X
latent heat of evaporation
[J/kg]]
H
dynamic viscosity
[kg/(ms>]
[m2/s]
v
kinematic viscosity
[kg/m3]
p
density
' ftl/11
characteristic shear stress

Pcg/m3]

tj
tw

interfacial shear stress

wall shear stress

[Pa]
[Pa]

Subscripts
cond condensation

conv
eff

convective
effective

gas-vapour mixture

int
k
1
If
n

interfacial value
k-th phase index
liquid
liquid film
noncondensable gas

sat
v

saturation
vapour

wall

Superscripts

ps
w
+

pool surface
wall
dimensionless value

Abbreviations
BE
Best Estimate
EdF
Electricite de France
ENEL Italian Electricity Board
1C
Isolation Condenser
LOG A Loss Of Coolant Accident

PCCS Passive Containment Cooling System

PUD
Proportional Integral Derivative
R.H.S. Right Hand Side

[Pa]

363

APPLICATION OF THE UMAE UNCERTAINTY METHOD

IN ASSESSING THE DESIGN AND THE SAFETY OF NEW
GENERATION REACTORS

F. D'AURIA, G. FRUTTUOSO, G.M. GALASSI,

S. GALEAZZI, F. ORIOLO

University of Pisa, Pisa

L. BELLA
ENEL CRT, Pisa
V. CAVICCHIA, E. FIORINO
ENEL ATN, Rom

Italy
Abstract

The present paper deals with the problem of uncertainty evaluations in the predictions of
transient behaviour of nuclear power plants by complex thermalhydraulic system codes. This is
relevant to the design and the safety assessment of nuclear reactors.
The UMAE (Uncertainty Methodology based on Accuracy Extrapolation) methodology,
recently proposed by the University of Pisa is briefly outlined and the main results are
summarized.
Emphasis is given in the paper to the application of the method in the domain of
advanced reactors; in this frame a few results obtained from the analysis of Spes-2 (AP-600
simulator) data are discussed.
1. INTRODUCTION

Evaluation of power plant performance during transient conditions has been the main
issue of safety researches in the thermalhydraulic area carried out all over the world since the
beginning of the exploitation of nuclear energy for producing electricity, e.g. [1] and [2].
A huge amount of experimental data have been obtained from the operation of test
facilities simulating the behaviour of nuclear plants. Several complex system codes have been
developed and qualified that allow the characterization of plant transient performance, ref. [3].
In the last few years uncertainty methodologies have been proposed to evaluate the
errors in the prediction of transient scenarios outcoming from the use of the system codes, refs.
[4] and [5]. The error quantification is a necessary step to prove the safety of the existing
plants and to optimize the design and the operation.
The UMAE (Uncertainty Methodology based on Accuracy Extrapolation) has been
recently developed at the University of Pisa and fully applied to the calculation of error bands
in the time trends of relevant thermalhydraulic quantities in PWR following a typical small
break Loss of Coolant Accident, ref. [6]. The use of the UMAE requires the availability of a
qualified computer code and of relevant experimental data obtained in properly scaled
simulators, refs. [7] and [8].

The objective of the present paper is to propose a framework for the application of the
UMAE to the AP-600, with main regard to the deterministic evaluation of the safety margins
during anticipated incident conditions. This requires, among the other things, the availability of
specific experimental data like those obtained or being obtained by the Spes-2, the Rosa-V
and the OSU experimental programs. Such data can be combined, as far as similar phenomena
are concerned, with those available from facilities simulating current generation PWR.

365

To this aim, an outline of the mentioned uncertainty methodology is part of the paper and
the results of the application of a few steps are discussed. These essentially include the
independent qualification of the adopted computer code with the demonstration that the error
in predicting the so-called relevant thermalhydraulic aspects is within acceptable limits. The
analyses performed in relation to experimental data available from the AP-600 simulator Spes2, ref. [9], arc utilized hereafter, ref. [10].
2. BASIS AND DEVELOPMENT OF THE UMAE

The fundaments of the methodology have been discussed in previous papers, e.g. refs.
[11] and [ 12] and can be drawn from the considerations 1 to 6 below, related to the test
facilities and the codes. It is assumed that both of these are representative of the actual state of
the art: i.e. facilities are "simulators" of a reference LWR and codes are based on the "six
balance equations model" and must be intended as generically "qualified", ref. [3].
1. The direct extrapolation of experimental data is not feasible; nevertheless, the time trends of
significant variables measured during counterpart tests in differently scaled facilities are
quite similar: this fact must be exploited.
2. Phenomena and transient scenarios occurring in larger facilities, being nearly constant the
other conditions (e.g. design criteria, quality of instrumentation, etc.), are more close to
plant conditions than those recorded from smaller facilities.
3. "Qualified" codes are indispensable tools to predict plant behaviour during nominal and offnominal conditions.
4. The confidence in predicting a given phenomenon by the code, must increase when
increasing the number of experiments analyzed dealing with that phenomenon.
5. The uncertainty in the prediction of plant behaviour cannot be smaller than the accuracy
resulting from the comparison between measured and calculated trends for experimental
facilities; furthermore, accuracy (uncertainty) must be connected with the complexity of the
facility (plant) and of the considered transient.
6. The effects of user and of nodalization must be included in the methodology.
The basic idea is -to get the uncertainty from considering the accuracy. The main
problems to achieve this arc connected with the availability of experimental data that must be
'representative' of plants, with the quantification of accuracy and with the justification of any
relationship between accuracy obtained in small dimensions loops with accuracy of the plant
calculation, i.e. uncertainty.
The use of data base from counterpart and similar tests in Integral Test Facilities was of
large help in this context, with main reference to the last mentioned problem. In particular,
similar tests arc those experiments performed in differently scaled facilities that are
characterized by the occurrence of the same thermalhydraulic phenomena; counterpart tests are
similar tests where boundary and initial conditions are imposed following a scaling analysis.
Two main aspects have been considered at a preliminary level to judge the realism of the
accuracy extrapolation ref. [6]; from the experimental side, the design of the facilities, the
boundary and initial conditions of the experiments, the suitability of the instrumentation, the
quality of the recorded data have been evaluated together with the similarity of the phenomena;
from the code side, the general qualification process, the capability to simulate the relevant
phenomena identified, the qualification of the nodalization and of the code user have been
independently assessed.
Several parameters of geometric or thermohydraulic nature can be used, in principle, to
derive uncertainty from accuracy values; furthermore, the use of a unique parameter is
preferable to minimize the possibility of counterfeiting information from the available data
base. The selected parameter should constitute a link between the experimental data and the
366

data foreseeable in the plant in the case of occurrence of the conditions of interest, it must be
representative of the involved phenomena at a global scale (i.e. the equivalent diameter in the
core simulators is representative of thermalhydraulic phenomena in the core but does not affect
substantially pressure behaviour during a depressurization transient), it must take into account
the present status of the technology (i.e. the height of the facilities or the initial pressure might
not be suitable parameters because they are nearly the same in the different facilities and in the
reference plant).
Several parameters fulfil the above requirements, nevertheless the volume of the facility
or, in dimensionless form, the ratio between primary system fluid volumes in the facilities and
in the reference plant was selected as 'extrapolation' parameter.
The situations depicted in Figure 1 are expected from using such an approach ref. [11].
The volume scaling ratio is reported on the horizontal axis; this quantity actually varies by four
orders of magnitude from the smallest facility to the plant ref. [12]. The shaded area represents
the range of facilities, the smallest one and the largest one are characterized by a volume
roughly 2000 times and 50 times, respectively, lower than that of a 1000 MWe plant. On the
vertical axis the generic quantity Y and the ratio YE/YC (experimental over calculated value)
are represented where Y is the value of any quantity that is relevant in a given transient.

Y, VE

Range of
facilities

Yr

Extrapolation

* Facility
| Plant

10 -3

10 -4

10 -2

10 -1

Kv

Figure 1 - Possible trends resulting from scaling analyses

2.1 UMAE flow diagram

The use of calculated and measured data related to counterpart and similar tests,
especially with the help of the code, directly led to the achievement of quantities connected
with the uncertainty. "Dispersion bands" and "extrapolated" plant behaviour were previously
defined and evaluated ref. [6].
The UMAE procedure aims at calculating the uncertainty and, although making use of
the same above-mentioned data base, involves the rationalization of the various steps including
the use of statistics in order to avoid or minimize the expert judgement at the various levels.
367

Generic experiments in integral facilities and related calculations other than counterpart and
similar tests, can be processed by the UMAE, provided the availability of data base related to a
reasonable set of individual phenomena that envelope the key phenomena foreseeable in the
selected plant scenario.
A simplified flow-diagram of the UMAE is reported in Figure 2. The way pursued to
evaluate the data base and the conditions to extrapolate the accuracy arc synthesized hereafter.

Nodalization

and user
qualification1'

methodology developed
(Dashed blocks represent blocks that are common to UMAE and CSALJ)

Figure 2- Simplified flowsheet of UMAE.

2.1.1 Evaluation of the Specific Data Base

The specific data base is constituted by the signals recorded during the considered
experiments and by the results of the code calculations. Each test scenario (measured or
calculated) should be divided into 'Phcnomenological Windows' (Ph.W). In each Ph.W. 'Key
Phenomena' (K.Ph) and 'Relevant Thcrmalhydraulic Aspects' (RTA) must be identified. K.Ph
characterize the different classes (e.g. small break LOCA, large break LOCA, etc.) of
transients and RTA are specific of the assigned one; K.Ph are always applicable; the definition
of RTA for small break LOCA in PWR has been done in ref. [8]. K.Ph and RTA qualitatively
identify the assigned transient; in order to get quantitative information, each RTA must be
characterized by 'Single Valued Parameters' (SVP, e.g. minimum level in the core), "NonDimensional Parameters1 (NDP, e.g Froude number in hot leg at the beginning of reflux
condensation), Time Sequence of Events' (TSE, e.g . time when dryout occurs) and 'Integral
Parameters' (IPA, e.g. integral or average value of break flowrate during subcooled
blowdown).
368

2.1.2 Accuracy Extrapolation (block "1")

If the following conditions are fulfilled, accuracy in predicting SVP, NDP, IPA and
TSE can be extrapolated, ref. [13]:
- the design scaling factors of the involved facilities are suitable;
- the test design scaling factors of the involved experiments are suitable;
- the experimental data base is qualified;
- the nodalizations and the related users are qualified;
- RTA are the same in the considered experiments if counterpart or similar tests are involved;
otherwise, the same RTA can be identified in different experiments;
- RTA are well predicted by the code at a qualitative and a quantitative level;
- RTA are the same in the plant calculation 'facility Kv scaled' and in the experiments;
parameters ranges (SVP, NDP, TSE and IPA), properly scaled, are also the same. This must
be interpreted in different ways depending upon the availability of counterpart tests;
- in the plant calculation 'realistic conditions' Ph.W and K.Ph are the same as in the considered
experiments; SVP, NDP, TSE and IPA may be different: reasons for this are understood.
The extrapolation of accuracy is achieved with reference to the above mentioned
parameters through the use of the statistics ref. [14]. The ratios of measured and calculated
values of SVP, NDP, TSE and IPA are reported in diagrams like that shown in Figure 1
assuming that they are randomly distributed around the unity value. This is also justified by the
huge number of variables affecting the considered ratios. In this way 'mean accuracy1 and '95th
percentile accuracy' are derived that are applicable to the plant calculation. The measurement
errors, the unavoidable scaling distortions and the dimensions of the facility are directly
considered.
It should be noted that only one calculation, performed with a qualified Analytical
Simulation Model (ASM in Figure 2), is necessary to get the reference plant scenario. The
ASM is qualified in the frame of the block "k" of Figure 2; the final part of the qualification
process, ref. [15], essentially derives from comparing RTA of the plant calculation with those
obtained in the experimental facilities during the same accident situations. Qualitative and
quantitative accuracy evaluation steps are used in this frame, ref. [16]. The "extrapolated
accuracy" values are superimposed to the reference plant trends to get the final value of
uncertainty.

3. APPLICATION OF THE UMAE TO THE LICENSING PROCESS

The procedures aiming at the evaluation of the uncertainty in codes predictions of

Nuclear Plants related scenarios (herein called uncertainty methodologies) have as main output
error values. These can be at a given time, either error bands to be superimposed to a "base"
value, cither ranges of variation of the assigned quantity, cither limit values (boundary value
approach).
The Peak Cladding Temperature following dryout of the surface of the fuel rods is the
most important quantity to be predicted in safety or licensing calculations. The result obtained
from the application of the UMAE procedure is outlined hereafter ref. [6].
The UMAE has been applied so far to the uncertainty evaluation in a small LOCA
assumed to occur in the Krsko Westinghouse plant installed in Slovenia. The plant is a two
loop 630 Mwe PWR. The considered transient is originated by a rupture in the cold leg
between the pump and the vessel; the break area is about 5% of the cross section area of the
main pipe. Scram, isolation of secondary side and pump trip are foreseen during the transient;
failure of high pressure injection systems is assumed, while accumulators are supposed in
operation. The calculated transient is stopped before allowing the intervention of low pressure
injection systems.

369

The result concerning the rod surface temperature trend in the core region when the
maximum temperature is predicted, is given in Figure 3. The calculated error bands are
superimposed. The PCT during both the dryout periods is calculated with an error of the order
of 100 K.
1000

ASM result
upper envelope
lower envelope

? 800

3
CO

O)

CL

600

400

800

time [s]

1600

Figure 3 - Uncertainty resulting from ASM calculation of a Small Break LOCA in the KRSKO

Westinghouse plant.
Another application of the UMAE was dealt with a data base including experiments in
separate effect facilities, ref. [17]; this led to an error for the PCT of about 270 K in a
thermalhydraulic scenario different from what previously considered.
Two safety/licensing indications can be outlined:
1) all Design Basis Accident simulations, either from experimental facilities either from the
use of best-estimate codes, led so far to PCT (where applicable) in the range 500-1000
K; so the addition of uncertainties in the range 100-300 K still keeps the rod surface
temperature values well below the safety limit (1473 K); this leaves some margin for
possible changes of plant operating conditions;
2) a no-dryout situation is shown as the most acceptable code calculation result in Figure 3.
This comes out from time uncertainty values for the dryout start and for the rewet
occurrence and constitutes a typical bifurcation.
The final remark here is that the application of uncertainty methodologies within the
DBA area, should lead to sufficiently reduced error bands for PCT; e.g. an error 700 K
obtained from uncertainty study could be useless testifying deficiencies either in plant design
either in the methodology/code itself. The former possibility should be excluded, with present
knowledge, within the DBA boundaries.
4. USE OF DATA RELEVANT TO AP-600

In order to apply the UMAE to reactor calculations, the availability of a qualified code
and of "relevant" experimental data is needed; "relevant" experimental data must be obtained
from suitable facilities and directly related to transient scenarios of interest in the plant.
Furthermore, codes can be fully qualified when the comparison of predicted results with
370

"relevant" data is successfully made (see also before).

Most of the thermalhydraulic phenomena important for new reactors are also important
for present generation reactors, although the ranges of parameters can be different. A
comprehensive evaluation of the differences between present and new generation reactors can
be found in rcf. [18] as far as thermalhydraulic phenomena arc concerned. As a result of the
above investigation, most of the experience gained in qualifying system codes on the basis of
situations of interest in present generation reactors can be used for next generation reactors.
However specific testing is also requested: the behaviour of the Core Make-up Tanks
(CMT), of the Passive Residual Heat Removal (PRHR) heat exchanger, etc., must be
experimentally qualified
in integral and separate effect test facilities (ITF and
SETF); the system code capability in predicting the resulting scenarios should also be
demonstrated.
In this framework, specific programs started in different organizations, e.g. Spes-2,

Rosa-V, OSU in relation to AP-600 and Panthers, Panda, Toshiba, Piper-one (see ref. [19] for
this last case) in relation to SBWR.
A few steps of the overall code assessment and uncertainty evaluation processes
carried out with reference to the application of the RelapS code to Spes-2 experiments, are
discussed.
4.1 Spes-2 2" cold leg break test scenario

The Spes-2 facility reproduces all the main zones of the AP-600 primary circuit. A sketch
of this can be seen in Figure 4, see also ref. [20].
The analyzed event is a small break LOCA, originated by to a 2" break in the AP600, in
one of the cold legs in the loop not connected to the pressurizer. This test, No. 3 of the SPES2 Tests Matrix, assumes the actuation of all passive Engineered Safety Features, with the
exception that the area of one of the 4th stage SPES-2 ADS valves has been reduced, to
simulate the failure to open of one AP600 4th stage ADS valve. Conversely, the active systems
(i.e. the normal RHR and the CVCS) are assumed not to operate.

CONDENSER

Figure 4 - Simplified flow sheet of SPES-2 facility

371

From a qualitative point of view the following phenomena occur:

- After the break device is opened, the primary pressure drops to the reactor trip set point and
the main steam line isolation valves are closed, causing a slight increase of SG pressure. The
heater rod power is controlled to match the scaled AP600 decay power, with an extra
contribution to compensate for the heat losses. The compensation power is switched off
when the first stage ADS opens.
- When the pressure reaches the "S" signal setpoint, the main feedwater is isolated, the RCPs
are switched off, inducing a small "bump" in the primary pressure and the PRHR and CMTs
isolation valves are opened.
- The RCS cools down to the same temperature of the secondary side and the coolant
becomes saturated; in this period the primary and the secondary pressures are of course the
same.
- As long as the primary circuit inventory allows, water is circulated through the cold
leg, the balance line and the CMT discharge line. In this period the CMT remains
filled with water, although coolant mass is added to the primary circuit, since less dense
water from the RCS is substituted with denser water from the CMT. Once
the siphon breaks, the CMTs start to empty, causing a significant jump in their injection
flow.
- When the CMT level falls to the proper setpoint, the ADS valves open, depressurizing the
RCS and allowing the accumulators to inject. Since the accumulators and the CMTs share
the same injection line, CMTs injection is reduced in this period.
When the accumulators empty, the CMTs continue to provide coolant to the RCS, at an
increased rate.
- At the proper CMT level, the 4th ADS stage is actuated, reducing the RCS pressure to a
very low value, close to the atmospheric pressure, and allowing stable coolant injection
from the 1RWST.
4.2 Overview of code results
A detailed nodalization has been developed for the Spes-2 facility including all the
geometrical details (Figure 5), sec also rcf. [21].

Figure 5 - Nodalization of the SPES-2 facility for the RELAP5/MOD2.5

372

The analysis of the 2" Cold Leg Break Test has been performed assuming the initial
conditions measured on the facility. The calculated and the experimental event timings are in
good agreement.
The agreement between calculated and experimental break flow is good as long as

subcoolcd blowdown occurs through the break. After, while in the calculation a sharp decrease
in the mass discharge happens at about 550. s, testifying a mixture quality transition of the
flowrate at the break, experimental results do not evidence this phenomenon, and the break
flow reductions occurs later, when the 1st stage ADS valve opens, resulting both in system
depressurization and in mass drawing towards the pressurizer, with a related coolant diversion
from the break.
Phenomenological differences between experiment and calculation again appear during
the accumulator injection phase: bypass of injected liquid occurs in the calculation, whereas
accumulator injected mass is essentially bleed through the ADS valves during the test.
The trends of the primary and secondary pressure versus time are shown in Figure 6: the
agreement with the measurements is good. The slight underprediction of the secondary
pressure is due to inaccuracies in the evaluation of the heat losses.
Mainly due to the very good prediction of the primary pressure, both the hot and cold
SPES-2: 2" CLB

XXX RELAPS/MOD25
YYY RELAPS/MOD25
ZZZ EXP PRIMARY
CXP SrCONDARY

- PhW 2

3 PhW 1

HiW

ACC I n j .

, 9

J
0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

Time (s)

Figure 6 - SPES-2 test # 3: Primary and SGs secondary side pressure

o
o
"!)

i
I

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

Time (s)

Figure 7 - SPES-2 test # 3: Fluid temperature jump across the PRHR

373

I
o

200

400

600

800

1000 1200 1400 1600 1800 2000

2200

Time (s)

Figure 8 - SPES-2 test # 3: Discharged flow rate from CMT B

XXX
YYY
ZZZ
Iff

RELAPS-MOMJ
RELAPS'MOOZ.S
RELAPS/MO02S
EXP-CMTA-TOP

AAA

EXP-CMTA.MIO

SU

EXP-CMTA-BOT

49

200

400

600

800

1000 1200 1400 1600 1800 2000

2200

Figure 9 - SPES-2 test #3: Fluid temperature at various elevations in CMTA

legs temperatures are very well predicted. The main events during the test are also shown.
Figure 7 plots the difference between the outlet and the inlet PRHR temperatures. A
difference as large as 40 C can be observed: in fact, while the PRHR inlet temperature
(essentially the hot leg temperature) is very well calculated, the outlet temperature is
overpredicted, resulting in an underprediction of the heat exchanger performance.
Figure 8 presents the discharge flow from the CMT connected to the same cold leg
where the break is located; the discharge flow from the other CMT is qualitatively similar.
The CMT flow during the recirculation period is correctly predicted, although a delayed
inception of the CMT emptying phase (by less than 100 s) is calculated and the related flow is
somehow overpredicted.
A substantial thermal stratification is observed in the CMTs. which is slightly
undcrpredicted by the SPES-2 model (Figure 9).
374

The "spikes" in the calculated discharge flow, which are not as much evident in the
experiment, are due to sudden condensation phenomena in the CMTs.
The accumulator injection, whose onset is predicted within 100 s and whose flow is very
well predicted, reduces the draining from the CMTs, although not as much as predicted.
The overall prediction of the CMT mass balance is good enough to allow to predict the
1st and the 4th ADS stage openings with an error of 20 and 220 s, respectively.
The ADS flow rate (stage 1-2-3) is plotted in Figure 10, while Figure 11 presents the
differential pressure across the prcssurizer. In the two figures, a large water suction from the
RCS to the PRZ is evident, which allows a large ADS discharge flow.
The parameters more directly related to the AP600 plant safety are the heated rod
temperatures and the core differential pressure (which allows a more direct
calculation/measurement comparison than the core level).
The heated rod temperatures are predicted with very good accuracy. In agreement with
the experiment, no dry-out has been calculated to occur.
The pressure drop across the core is shown in Figure 12: the calculation
matches very well the measured value, allowing to deduce a good prediction of the vessel
water inventory.
XXX RELAP&MO02.S

YYY REWPSIM002.S
ZZZ EXP-ADS12->3
III

200

400

600

800

CXP-ACCUM-A

1000 1200 1400 1600 1800 2000

2200

Time (s)

Figure 10 - SPES-2 test # 3: Discharged flow rate from ADS valves 1,2,3
and Accumulator

200

400

600

800

1000 1200 1400 1600 1800 2000

2200

Time (s)

Figure 11 - SPES-2 test # 3: Pressurizer differential pressure

375

SPES-2: 2" CLB

XXX RELAP5/M002 5
YYY EXPOP-CORE

PhW 1
V-

PliW 2

I'hW

I'hW '. '

r*

I
5

8
O

'

OL
0

Y'- *" *-yv

I
^v

L
1

**- +
-

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

Time (s)

Figure 12 - SPES-2 test # 3: Core differential pressure

4.3 Qualitative and quantitative accuracy evaluation

In order to evaluate qualitatively the accuracy following the procedure mentioned in sect.
2, the transient has been subdivided into 4 phcnomenological windows; more than
30 RTA have been identified which arc characterized by around 70 NDP, IPA, SVP and TSE.
The Ph.W and the agreement between measured and calculated trends of the RTA can be
seen in Table 1.
Table 1 - Agreement between calculated and experimental RTA in SPES-2 test # 3

Wkxbw
(PhW)
SOBCOOLED

TbcnnlbydnuUc
Aipc( (RTA)

Panmeler
(S\'P)

Event
(TSE)

PnMuraM-MnfKwis

Parameter
(NDP)

(IPA)

i1

SLOWDOWN

PnWl

_
j
Turnt- 9

SCRAM (MM (>

Pticnp Cfcwt

PkW,db(.)

FWcfcwle(0

bLclwjr*lmM<<}
TinM of f praMure

E
E

230 (MP.)

Scoaduy Syttern Pivflcuv

bri>v.our

I
Pott. Prmun to otiul

,
PtMMTCKMOTtftHfrK

' R

tt
I
CMTB*.^

Drv,lvr (DV)

Tim-

ft

'

M~

nvxfe n CMTA (>

j CMTA (few

Fud of recnuUttoo

j CMTe flow
1 ttne of lve opconc
I')

j PRHR now

j PRHR power
Vt

aodlOOi
'
1
to wnl fluid fltrUirtcdMci

Legend:

376

( Co (K
"

SdSco)asOCboani (DC E

E
U

Excellent
Unqualified

Reasonable

NE Not [:Mluatcd

Minimal

Table 1 - Agreement between calculated and experimental RTA in SPES-2 test # 3 (Cont'd)
dul* V>ln^4

Wkufew

(PhW
?MT DtsCBAXCt
KODC
TnWl

Ereot

fbennattiydrftuUc
Aipcct (RTA)

rrst)

PRZpt*MM

rm*b*aP$p<<u

PtnaMtcr

ranmeCer
<SVf)
R

(IP A)

(NDf)

Aw*f*PS-&S pnM

ttfMMK* Amf tePbWl

Y-h SS pmm* (}

SG &S betuviotK

W2

Awnc* SGH Pw R
FhW2
R VMM! fluid itraoficMMXi

Si*ic*ahnK tf DC hoaam (DC

<Mb4)tf 750*

dfcoBolwK tf IX'tttOM* (IK.'

<Mk4) tw wJ of PhW 2
Riv^Flow

F nd i>m W?

JBfMka
SMII^PhWI

TwwtBafirCMT
uAi<te4M(vic<0

CMTtxtuvwir

CMTA bvri 200 before

EwJoffkW:

|cMTA ow
S.fl<rf diKbMf 4fMOd

DTbMwMoCMTAa)

CMTnlev12<MWM*

E^rfF.W7

ADS I KUMOM

CMTB(-~"'f-*"

SHf ich*p*fod

Avnm|v *4>c*ahat tt tic-

PRHRMMVMKt

CMafM'**}

EftdafFi*:

Mbcbo*~aMOidS50.

jPRJtRpow*
A m(v Mho*ott it d
M
uOM bctwMo *1 ind 850 .

jPRKRftrw

fli

sn<nw:

SMM^rkwi

E*dtrtiw;
jCopoW

SMx^rkw;
ACClnbtview

TnfACCAt(>)

E*dtfdu
JACCA flow
SHH<dta/V

TefACCfeAMt<)

Cn4O'4<ct<
J ACCB o

S(io<4kKp

Toe f manure OMB

(>

Pnnury SyMem bbtvwuc

MwwranHi(k(}

Tout CMT dtfchu^a

ADSISTAKTTO
iRWST I.NTEJIVEMT10N

nUM/PS ntM ! tbe od of

PbV,-2

MOMMXB nun'iabil mm

Rjt t><-o K.404 unul PS

N6

R
A-*wm^iJ*/A(K>PStbo
pariodbetweaaftH PbW3 iuri
odl230.i(MP/.)
j.Xt
AM^ dP/Al io PS ad
rbW3 (MP^)
MioMwa mMi b PS a tte R
pnoodWwMo PbWJ jun
od 1250.) <kg>

Pnmuy Synmud &KodMy

Sy*B blaviour

PnW3

BMkAow

Snefl\V) . I,

JBcMkflow

lMofrbW)
EodofttWJ

K^

ADS-1 atonvatMotne R

ADStebtvmr

Ritio brt*o fee m f OE

ADS fl<~ ourcnl. ud tbt

(0

S.-:rtW]
EixJrffY*!

NE

bfNL fo ntcgnl

AOS-2 AKVMLMOUOH R
(')

NE

J 8mL fl<7

NE

JAOS-lflow

FUtia bMao 4* ra of dx
AOS (W atopxlt tad tte
MMJPSmui

KC

SknfPiW)
4<rrtV)

NE

JADS 709w

i sorfii.-i
R

NE

Etrft>-i

(0

JADS 3flw

sorn')
ADS-4 otofvwfMourM M
(J

C*d <T Pll

NE
JADS 4ow

s*nf n*'3
Eod of ACC- A 1 KfmA

ACCbetuviour

di*ctunn> (>

Efldof ACCBlKpwd

FUuo t*f-<yi 4**u&ofACC

S*

llo* acrpb
UMtuic^lPSmu

ditch*** (i)
CMTMuvxxr

' fUtitf Nrt^^to (he wn of

CMT tfe* arfnJt >d MMM!
PSlMJM

NE

Eaf fWJ

NE
H

JCMTA*C*<T(VW
SHE tf Oit 4taphB

Legend:

Excellent

Reasonable

Unqualified

NE

Not Evaluated

Minimal

Some discrepancies can be noted in the comparison between measured and calculated
trends, but a good positive overall qualitative judgement allowed us to perform the subsequent
step, e.g. the quantitative accuracy evaluation.
377

The application of the FFT based method, ref. [16] gave the results shown in Table 2. It
should be noted that the overall accuracy value is within the limit 0.4 that has been fixed for the
acceptability of a calculation, ref. [6].

Table 2 - Average accuracy for parameters selected for FFT application and global accuracy of the code
calculation

PARAMETER

AA

IAVF

DP-CORE

0.32

25.5

DP-PRZ

0.63

24.2

DP-SQADC

046

158

1 AOIP

035

21 4

l: A40H

1 SI

260

IO\80U

1 29

186

1 B20I-

070

20.2

F B401:

148

238

IF030P

020

274

II-040P

251

139

INTK_005

0 10

605

P-027P

009

295

P-A04S

026

264

T-A03PL

007

27.6

T-A40IE

020

161

T-A420E

053

244

T-A82E

0.59

22.2

T-A83E

040

271

T-B40IE

0 19

980

T-B420E

020

22.9

TW020P87

0 10

33.3

GLOBAL ACCURACY

0.40

29.1

5. CONCLUSIONS
The UMAE methodology has been presented that allows the calculation of uncertainty in
the predictions of off-normal conditions in nuclear reactors by thermalhydraulic system codes.
Essentially, the uncertainty is derived from the extrapolation of accuracy obtained by
comparing measured and calculated variables trends that are relevant to the assigned accident
scenario. The use of data from counterpart tests, was of large help in developing the concepts
at the basis of the methodology.
One complete example of application of the UMAE to a small break LOCA in PWR has
been given. In that case dryout is foreseen and peak cladding temperature, including the error
of the order of 100 K, remains well below the licensing limits.
The main peculiarity of the UMAE is the direct exploitation of the huge experimental
data base now available from integral test facilities: this is strictly necessary' and is retained
suitable to characterize any foreseeable accident scenario in Light Water Reactors. Moreover,
all the aspects contributing to the uncertainty are taken into consideration: the user effect, the
378

nodalization qualification, the errors in introducing boundary and initial conditions, the intrinsic
capabilities of the code in predicting the phenomena of interest, are fully included in the final
evaluation of the error bands.

Few additional biases to be included in the final value of uncertainty that are not fully
considered by the UMAE, have been identified in the frame of the present research although
not discussed in this paper. Among these, the potential occurrence of important
multidimensional phenomena represents a critical issue if only codes based on one-dimensional
models are available: experimental data from specific facilities should be used to accertain the
codes capabilities in this connection.
The application of the UMAE to situations relevant to the new generation reactors,
and in particular AP-600, seems feasible once data from differently scaled facilities

have been analyzed. It should be noted that most of the experience gained in the analysis of
transients in present reactor configuration can be directly utilized: specifically, in the
considered small break LOCA case a large number of RTA are the same in PWR and AP-600
scenarios.
REFERENCES
[1]

CSNI Group of Experts. 1989. "Thcrmohydraulic of Emergency Core Cooling

in Light Water Reactors", Report No. 161, Committee on the Safety of Nuclear
Installations.
[2] US-NRC 1989. "Compendium of ECCS Research for Realistic LOCA Analysis",
NUREG-1230, U.S. Nuclear Regulatory Commission.
[3] D'Auria F. 1987a. "Experimental Facilities and System Codes in Nuclear Reactor
Safety", Int. Seminar State of the Art on Safety Analyses and Licensing of Nuclear
Power Plants, Varna (BG).
[4] D'Auria F., Lconardi M., Glacscr H., Pochard R. "Current status of methodlogies
evaluating the uncertainty in the prediction of thermal-hydraulic phenomena in nuclear
reactors", Int. Symposium on "Two-Phase Flow Modelling and Experimentation", Roma
(Italy), October 9-11, 1995.
[5] Boyack B.E., Carton I., Duffcy R.B., Griffith P., Katsma K.R., Lellouche G.S., Levy S.,
Rohatgi U.S., Wilson G.E., Wulff W., Zuber N. 1990a - "An overview of the code
scaling, applicability, and uncertainty evaluation methodology", J. Nuclear Engineering
and Design Vol. 119.
[6] D'Auria F., Debrecin N., Galassi G.M. 1994. "Outline of the Uncertainty Methodology
based on Accuracy Extrapolation (UMAE)", CSNI Workshop on Uncertainty
Evaluation, London (UK).
[7] D'Auria F., Debrecin N., Galassi G.M. 1993a. "Application of RELAP5/MOD2
system code to small break LOCA counterpart experiments in PWR simulators", IX
Brazilian Meet, on Reactor Physics and Thermalhydraulics, Caxambu, Minas Gerais
(BRA).
[8] D'Auria F., Ferri R., Vigni P. 1993. "Evaluation of the data base from the Small Break
LOCA Counterpart tests performed in PWR Experimental Simulators", llth Conf. of
Italian Society of Heat Transport, Milan (I).
[9] Alemberti A., Frepoli C., Graziosi G. "SPES-2 col leg break experiments:
Scaling approach for decay power, heat losses compensation and metal heat
release", Proc. New Trends in Nuclear System Thermohydraulic 1,679-687 (1994), Pisa,
(Italy)
[10] D'Auria F., Fruttuoso G., Oriolo F., Bella L., Cavicchia V., Fiorino E. "SPES-2: AP600
Integral Test Facility Results: Confirmatory Plant Behaviour and Relap5/Mod2.5 Code
Assessment", ICONE - 3, The Third Int. Conf. on Nuclear Engineering, Kyoto (J), April
23-27, 1995

379

[11] Bovalini R., D'Auria F. 1993. "Scaling of the accuracy of Relap5/mod2 Code", J.
Nuclear Engineering and Design, vol 139 Nr. 1.
[12] D'Auria F., Karwat H. 1989. "OECD CSNI State-Of-the-Art-Report on
thermalhydraulics of Emergency Core Cooling Systems - Review of the Operation of
Experimental Facilities", University of Pisa Report, DCMN - NT 138(89), Pisa (I).
[12] Bovalini R., D'Auria F., Galassi G.M. 1993. "Scaling of complex phenomena in System
Thermalhydraulics", J. Nuclear Science and Engineering.
[13] Bovalini R., D'Auria F., De Varti A., Maugeri P., Mazzini M. 1992. "Analysis of
Counterpart Tests performed in BWR Experimental Simulators", J. Nuclear Technology,
Vol. 97 Nr. 1.
[14] Belsito S., D'Auria F., Galassi G. M. "Application of a Statistical model to the evaluation
of Counterpart Test data base", J. Kerntechnik. Vol 59 Nr 3, 1994
[15] Bonuccelli M., D'Auria F., Debrecin K, Galassi G.M. 1993. "A methodology for the
qualification of thermalhydraulic codes Nodalizations", 6th Int. Top. Meet, on Nuclear
Reactor Thermalhydraulics, Grenoble (F).
[16] D'Auria F., Leonardi M., Pochard R. 1994a. "Methodology for the evaluation of
thermalhydraulic codes accuracy", Int. Conf. on New Trends in Nuclear System
Thermalhydraulics, Pisa (I).
[17] D'Auria F., Faluomi V., Aksan S.N. "A proposed Methodology for the Analysis of a
Phenomenon in Separate Effects and Integral Test Facilities", Int. Conf. on New Trends
in Nuclear System Thermohydraulics - Pisa (1), May 30 - June 2 1994
18]

D'Auria F., Modro M., Oriolo F., Tasaka K. "Relevant Thermalhydraulic Aspects of
New Generation LWR's", CSNI Spec. Meet. On Transient Two-Phase Flow - System
Thermalhydraulics, Aix-En-Provence (F), Apr. 6-8, 1992, J. Nuclear Engineering &
Design, Vol 145, Nrs. 1&2, 1993
[19] D'Auria F., Vigni P., Marsili P. "Application of RELAP5/MOD3 to the evaluation of
Isolation Condenser performance". Int. Conf. on Nuclear Engineering (ICONE-2) - San
Francisco (US), March 21-24, 1993
[20] Bacchiani M., Mcdich C., Rigamonti M. "SPES-2 AP600 Integral System Test:
Inadvertent ADS opening and Cold Leg Break Transients" ARS 1994 Int. Topical

Meeting on Advanced Reactors Safety, APril 17-21, Pittsburgh, PA

[21] D'Auria F., Fruttuoso G., Galassi G.M., Oriolo F. "Post-test calculation of SPES-2
experiment No. 3 by Relap5/Mod2.5 code", Universita degli Studi di Pisa, DCMN
Report, NT 241 (94), Ottobre 1994.

380

LIST OF PARTICIPANTS

Allen, PJ.

AECL
Sheridan Park Research Community
2251 Speakman Dr.
Mississauga, Ontario L5K 1B2, Canada

Alloggio, G.

ENEL-ATN
Via Monfalcone, 15
20132 Milano, Italy

Artaud, C.J.

Gamma-Metrics
San Diego, C A, USA

Aujollet, P.

Centre d'etudes de Cadarache

CEA/DRN/DER/SIS
Batiment 211
C.E.N. Cadarache
B.P. 1, 13108 - Saint-Paul-Lez-Durance, France

Bandurski, T.

Paul Scherrer Institut

CH-5232 Villigen PSI, Switzerland

Banergee, S.

Consultant to NRC
Washington D.C. 20555, USA

Bedrossian, G.C.

Comision Nacional de Energia Atomica

Avda. del Libertador 8250
1429-Buenos Aires, Argentina

Billig, P.F.

G.E. Nuclear Energy

Advanced Boiling Water Reactor Program
175 Curtner Ave., M/C 781
San Jose, CA 95125, USA

Bredikhin, V.

Research Institute of Nuclear Power

Plant Operations, (VNIIAEA, RINPO)
Moscow, Russian Federation

Bregaenel, E.

Area Attivit' Nucleari Gruppo Nucleare di Milano

Via Monfalcone 15, 20132 Milano, Italy

Budylin, B.V.

Minatom RF
109180, Staromonethy, 26
Moscow, Russian Federation

381

Cattadori, G.

SIET
Via Nino Bixio, 27
29100 Piacenza, Italy

Cavicchia, V.

ENEL-ATN
Viale Regina Margherita 137
00198 Roma, Italy

Choi, Y.S.

Korea Electric Power Corporation

103-16, Moon Jee Dong
Yu Song Ku, Dae Jeon, Rep. of Korea

Cleveland, J.

Division of Nuclear Power, IAEA

Wagramerstrasse 5
P.O. Box 100
A-1400 Vienna, Austria

Conway, L.E.

Westinghouse Electric Corp.

P.O. Box 355
Pittsburgh PA 15230, USA

Curca-Tivig, F.

Siemens AG KWU NA-T

Koldestrasse 16
91050 Erlangen, Germany

D'Auria, F.

Universit' Degli Studi di Pisa

Departimento di Costruzioni Meccaniche e Nucleari
Via Diotisalvi, 2
1-56126 Pisa, Italy

Delnero, G.

ANPA
Via Vitaliano Brancati
48-00144 Rome, Italy

El-Bassioni, A.A.

U.S. Nuclear Regulatory Commission

Washington D.C. 20555, USA

Eltawila, F.

Nuclear Regulatory Commission

Washington D.C. 20555, USA

Erbacher, F.J.

Forschungszentrum Karlsruhe
Institut fuer Angewandte Thermo- und
Fluiddynamik
Postfach 3640
D-76021 Karlsruhe, Germany

Fernandez, A.

EDF- SEPTEN
12-14, Avenue Dutrivoz
69628 Villeurbanne - Cedex, France

382

Ferreli, A.

ANPA
Via Vitaliano Brancati
48-00144 Rome, Italy

Fiorini, G.L.

Cantre d'etudes de Cadarache (CEA)

DRN/DER/SIS
13108 St. Paul Les Durance Cedex, France

Fiorino, E.

ENEL, Rome
Viale Regina Margherita 137
1-00198 Rome, Italy

Fornaciari, P.

ENEL, Nuclear Energy Division

Viale Regina Margherita, 137
00198 Rome, Italy

Franks, S.M.

NE-45
Advanced Light Water Reactors Division
Office of Nuclear Energy
U.S. Department of Energy
Washington D.C. 20585, USA

Galbiati, L.

CISE, Technologie Innovative

V. Reggio Emilia, 39
20090 Segrate (Milano), Italy

Gaspari, G.

SIET
Via Nino Bixio, 27
29100 Piacenza, Italy

Graziosi, G.

ANSALDO Nuclear Division

Corso Perrone 25
16161 - Genova, Italy

Hicken, E.F.

Forschungszentrum Juelich
ISR, D-52425 Juelich, Germany

Hittner, D.

FRAMATOME
Tour Fiat Cedex 16
92084 Paris la Defense, France

Hoeld, A.

Gesellschaft fuer Anlangen- und

Reaktorsicherheit (GRS) mbH
Forschungsgelnde
Walter-Meissner-Strasse
D-85748 Garching, Germany

Hsii, Y.

U.S. Nuclear Regulatory Commission

Washington D.C. 20555, USA

383

Hutchings, G.

Scottish Nuclear Ltd.

3 Redwood Crescent, Peci Park

East Kilbridge, G74 5PR
Scotland, United Kingdom
Kim,

B.S.

Korea Electric Power Corporation

103-16, Moon Jee Dong
Yu Song Ku, Dae Jeon, Rep. of Korea

Kropp, C.

ENEA
Via Anguillarese 301

00060 - S.M. di Galeria (Rome), Italy

Kwon, Y.M.

Korea Atomic Energy Research Institute (KAERI)

System Safety Analysis Dept.
P.O. Box 105, Yousung
Taejeon, 305-600, Rep. of Korea

Kymlinen, O.J.

IVO International Oy
Rajatorpantie 8, Vantaa
01019 IVO, Finland

Leridon, A.M.

Chef du SCC
Centre d'etudes de Cadarache (CEA)
DER/SCC/DIR - BAI 727
13108 St. Paul Les Durance Cedex, France

Lillington, J.

Safety and Performance Services Dept.

240/A32, AEA Technology, Winfrith
Dorchester
Dorset DT2 8DH, United Kingdom

Lombardi, C.

Politechnico di Milano
Via Ponzio, 34/3
20133 Milano, Italy

Lyu,

WANO Tokyo Centre

T.-H.

2-11-1 Iwatokita, Komae

Tokyo 201, Japan
Malave, M.

Nuclear Regulatory Council

CI Justo Dorado, 11
Madrid 28040, Spain

Manfredini, A.

Universit' Degli Studi di Pisa

Via Diotisalvi, 2
1-56126 Pisa, Italy

384

Mansani, L.

ANSALDO
Corso Ferrane 25
16161 - Genova, Italy

Marcila, G.

ANPA
Via Vitaliano Brancati
48-00144 Rome, Italy

Martinelli, R.

ENEA
Via Anguillarese 301
00060 - S.M. di Galena (Rome), Italy

Mazzocchi, L.

CISE, Technologie Innovative

V. Reggio Emilia, 39
20090 Segrate (Milano), Italy

Mcpherson, G.D.

U.S. Nuclear Regulatory Commission

Washington D.C. 20555, USA

Medich, C.

SIET
Via Nino Bixio, 27
29100 Piacenza, Italy

Monti, R.

ANSALDO
Corso Perrone 25
16161 - Genova, Italy

Munther, R.K.

Lappeenranta University of Technology

P.O. Box 20
Fin-53851 Lappeenranta, Finland

Noviello, L.

ENEL-ATN
Viale Regina Margherita 137
1-00198 Rome, Italy

Orazi, A.

ANPA
Via Vitaliano Brancati
48-00144 Rome, Italy

Oriolo, F.

Universit' Degli Studi di Pisa

Via Diotisalvi, 2
1-56126 Pisa, Italy

Orsini, M.

ANSALDO
Corso Perrone 25
16161 - Genova, Italy

385

Palavecino, C.

Siemens AG
Berliner Strasse 295-303
63067 Offenbach 111, Germany

Palma, A.

Politechnico di Milano
Via Ponzio, 34/3
20133 Milano, Italy

Parozzi

ENEL-ATN
Via Monfalcone, 15
20132 Milano, Italy

Pedersen, T.

Division of Nuclear Power, IAEA

Wagramerstrasse 5
P.O. Box 100
A-1400 Vienna, Austria

Proto, G.

ANSALDO
Corso Perrone 25
16161 - Genova, Italy

Py, J.-P.

FRAMATOME
Tour Fiat Cedex 16
92084 Paris la Defense, France

Reinsch, A.W.

Southern California Edison

5000 Pacific Coast Highway
San Clmente, A 92672, USA

Ricotti, M.

Politechnico di Milano
Via Ponzio, 34/3
20133 Milano, Italy

Rigamonti, M.

SIET
Via Nio Bixio, 27
29100 Piacenza, Italy

Sah, D.

Bhabha Atomic Research Centre

Reactor Safety Division

Hall No. 7, BARC, Trombay

Bombay 400 085, India
Saltos, N.T.

U.S. Nuclear Regulatory Commission

Washington D.C. 20555, USA

Sandrelli, G.

ENEL-ATN

Via Monfalcone, 15
20132 Milano, Italy

386

Scandola,

SIET
Via Nino Bixio, 27
29100 Piacenza, Italy

Shugars, H.

Gamma-Metrics
San Diego, A, USA

Silveri, R.

SIET
Via Nino Bixio, 27
29100 Piacenza, Italy

Soplenkov, K.I.

Research Institute of Nuclear Power

Plant Operation, (VNIIAEA, RINPO)
25 Ferganskaya St.
Moscow, Russian Federation

Sordi, R.

Politechnico di Milano
Via Ponzio, 34/3
20133 Milano, Italy

Tarantini, M.

ENEA ERG FISS

Via Martiri di Monte Sol, 4
40129 Bologna, Italy

Tavoni, R.

ENEA - Energy Department

Via Martiri di Monte Sol, 4
1-40129 Bologna, Italy

Trubkin, Y.

Electrogorsk Research Engineering Center (EREC)

6, Bezymiannaya St.
Electrogorsk, Moscow Region
142530, Russian Federation

Valisi, M.

ENEL-ATN
Via Monfaleone, 15
20132 Milano, Italy

Vidard, M.L.

EDF-SEPTEN
12-14, avenue Dutrivoz
69628 Villeurbanne Cedex, France

Zro, S.

FRAMATOME
Tour Fiat Cedex 16
92084 Paris la Defense, France

387