- 44 -

NEW AUSTRIAN GUIDELINE FOR THE TRANSPORT OF

DANGEROUS GOODS THROUGH ROAD TUNNELS

Diernhofer F.1, Kohl B.1,

Hörhan R.2
1
ILF Consulting Engineers, Linz
2
Austrian Ministry of Transport, Innovation and Technology, Vienna

ABSTRACT
In the Austrian guideline RVS 09.03.11 the Austrian Tunnel Risk Analysis Model TuRisMo
defines how to assess the risk for tunnel users. The same guideline stipulates that the specific
risk involved in the transport of Dangerous Goods (DG) through road tunnels should be
assessed in a separate procedure. Consequently, based upon European Directive 2004/54/EC
and the Austrian Road Tunnel Safety Law, a uniform risk assessment procedure for the
transport of DG through road tunnels has been developed. For a methodical risk analysis
approach, the OECD/PIARC-Model DG-QRAM was chosen. The results shall be published
in new Austrian guideline RVS 09.03.12 in 2010.
The main objectives of the research project are the verification of existing DG transport data,
the development of a complete risk assessment procedure in line with the new ADR tunnel
regulations and the definition of decision criteria for each level of risk evaluation.
Keywords: dangerous goods, road tunnel safety, quantitative risk assessment,
expected value, FN-curve

1. LEGAL ASPECTS OF TUNNEL SAFETY

1.1. Directive 2004/54/EC

This Directive defines minimum safety requirements on an international level to be met by
road tunnels forming part of the Trans European Road Network (TERN). The remarkable
aspect about this Directive is the fact that it combines both a “regulation-based” and a “risk-
based” approach. According to Article 13, a risk analysis shall be performed for all tunnels
featuring a special characteristic, taking into account all design factors and traffic conditions.

1.2. Austrian Road Tunnel Safety Law

The Road Tunnel Safety Law translates the requirements contained in the EC-Directive into
Austrian law. It defines the following measures for the transport of dangerous goods (DG):
• Prior to the definition or modification of regulations and requirements regarding the
transport of DG through a tunnel, a risk analysis is to be performed.
• To enforce the regulations, appropriate signs indicating alternative routes, are to be
posted ahead of the last possible exit before the tunnel and at tunnel entrances.
• In individual cases, specific operating measures designed to reduce the risks related to
some or all of the vehicles transporting DG in tunnels are to be checked (e.g. escorted
passage in convoys).

5thInternational Conference ‘Tunnel Safety and Ventilation’ 2010, Graz

 - 45 -

1.3. ADR 2007 / 2009

When the European Agreement concerning the International Carriage of Dangerous Goods by
Road, commonly known as ADR was revised in 2007, so called tunnel restriction codes were
assigned to all dangerous substances according to their potential of damage, amount (mass)
and carriage type. These codes serve as a basis for a uniform European regulation governing
the transport of dangerous goods though road tunnels.
To enforce restrictions of DG transporting vehicles through a tunnel the relevant authorities
shall assign the tunnel to a category defined in the ADR (Table 1). The new tunnel
regulations of the ADR have been valid since the 1st of January 2010.

Table 1: ADR tunnel categories and signature

Tunnel categories Restrictions Sign Traffic Sign

A No restrictions for the transport of dangerous No sign -

goods

B Restriction for dangerous goods which may lead Sign with additional panel
to a very large explosion bearing the letter B

C Restriction for dangerous goods which may lead Sign with additional panel
to a very large explosion, a large explosion or bearing the letter C
a large toxic release

D Restriction for dangerous goods which may lead Sign with additional panel
to a very large explosion, to a large explosion, bearing the letter D
to a large toxic release or to a large fire

E Restriction for all dangerous goods other than Sign with additional panel
UN Nos. 2919, 3291, 3331, 3359 and 3373 bearing the letter E

2. DATA BASE
At the beginning of the research project great emphasis was put on the data base. From 2006
to 2007 investigations of DG transports were carried out at 12 different cross sections on
Austria’s main traffic routes. In March 2009, the results of these earlier investigations were
evaluated and expanded by a detailed review of DG transports in cooperation with the police.
For one month (March 2009) the police stopped DG vehicles and controlled their transport
documents. The investigation was accomplished in all Austrian federal states and assigned
detailed information about the UN number, amount, carriage type and destination of the
dangerous substances.
The applied risk model includes for the most part very “severe” scenarios with huge amounts
of dangerous substances involved. For example scenario 4 stands for a pool fire of a 28 ton
tank of diesel/petrol. Using the model scenarios without an adjustment would lead to an
overestimation of risks. For that matter the data collected in the detailed investigation were
used to define more realistic allocation rules (see example in Figure 1).

5thInternational Conference ‘Tunnel Safety and Ventilation’ 2010, Graz

 - 46 -

ADR-Class 2

TANK PACKEGED / MIXED CARGO

FULL EMPTY

D a nge rc . 2 3 : D a nge rc o de 2 3 : D a nge rc o de 2 3 : D a nge rc o de 2 0 / 2 2 :

> 0 kg D a nge rc . 2 0 / 2 2 : > 0 kg D a nge rc . 2 5 / 2 2 5 : 0 kg > 0 kg D a nge rc o de 2 5 / 2 2 5 : > 0 kg
D a nge rc . 2 5 / 2 2 5 : D a nge rc . 2 0 / 2 2 :

Sc. 7, 8, 9 Sc. 13 no Sc. Sc. 3 Sc. 13

Figure 1: Allocation key for flammable gases (ADR class 2)

In addition, the results of the investigations revealed that in Austria the composition of the
DG carriage varies only slightly on different traffic routes and that these variations have only
little influence on the risk faced by tunnel users. Eventually, a standardized, averaged scenario
allocation for Austria based upon a detailed DG investigation was found (Table 2).

Table 2: Standardized scenario allocation for Austria (DG-QRAM)

Accident scenarios - DG-QRAM Proportion

Scenario 3: BLEVE of flammable gases in 50kg-cylinder (e.g. Propane) 0,0226

Scenario 4: pool fire of flammable liquids in bulk (e.g. Petrol, Diesel) 0,4270
Szenario 5: VCE of flammable liquids in bulk which may lead to explosive air
fuel mixture (e.g. Petrol), assigned as fraction of scenario 4 0,3610

Scenario 6: release of very toxic gases in bulk (e.g. Chloride) 0,0010

Scenario 7 / 8 / 9: BLEVE / VCE of flammable gases in bulk (e.g. Propane) 0,0113

Scenario 10: release of toxic gases in bulk (e.g. Ammonia) 0,0010

Scenario 11: release of toxic fluids in bulk (e.g. Acrolein) 0,0133

Scenario 12: release of toxic fluids in 100kg-cylinder (e.g. Acrolein) 0,0113

Scenario 13: burst of a tank of non flammable gases (e.g. liquefied refrigerated CO2) 0,0452

3. OECD/PIARC MODEL (DG-QRAM)

To calculate the risks involved in transporting DG an international accepted risk analysis
model called DG-QRAM is applied. The risk model was developed on behalf of
OECD/PIARC and is widely used on an international basis, but is obviously not the only
method available for assessing the risk resulting from the transport of DG.
The results of the risk analysis model are depicted as Expected Values (EV) or F-N curves,
illustrating the relation between accident frequency (F) and accident consequences (N number
of fatalities). Whereas the EV represents the average expected number of fatalities as a result
of all DG accidents, the F-N curve gives more comprehensive information on the extent of
damage in relation to the probability of individual accidents (Figure 3).

5thInternational Conference ‘Tunnel Safety and Ventilation’ 2010, Graz

 - 47 -

4. DEVELOPMENT OF A COMPLETE RISK ASSESSMENT PROCEDURE

In the year 2009, the Austrian Federal Ministry for Transport, Innovation and Technology
(BMVIT) launched a research project with the objective of establishing a complete
investigation and assessment procedure concerning risk analyses for DG (using DG-QRAM).
The research project was exclusively focused on assessing the risk of DG accidents, whereas
mechanical accidents and conventional fires are addressed by the risk model TuRisMo. The
final results have been presented in a new Austrian guideline (draft). The finalized version
should be published in 2010.
The research project served the purpose of defining a clearly structured risk assessment
procedure in line with the new ADR tunnel regulations, based upon reliable DG data. In the
course of this project, risk reference criteria for every step of the assessment procedure had to
be laid down. The project was supported by a work group comprising technical and legal
experts of the BMVIT and the Austrian Ministry of Internal Affairs, the Austrian federal
provinces, the Austrian Chamber of Commerce, the ASFINAG, the fire brigade, the transport
industry as well as ILF Consulting Engineers.
In principle, the risk involved in the transport of DG is determined in a multi-stage assessment
procedure.

4.1. Stage 1 – Classification Matrix

Stage 1 involves using a simple classification matrix (Figure 2) to define DG risks of road
tunnels. The application of the matrix shall permit a simple identification of tunnels with a
low DG transport risk.
The classification matrix takes into account the following main risk factors:
• the tunnel length
• the type of tunnel (bi-directional or uni-directional traffic)
• the ventilation system (natural, longitudinal or transverse)
• the traffic volume
• the percentage of heavy goods vehicles (HGV)
The respective parameters of the matrix were defined in a former study performed in 2008.
Then, a systematically risk calculation were performed for a set of selected reference tunnels
using DG-QRAM. As decision criteria an expected risk value of EV = 1 x 10-3 fatalities/year
was applied for the elaboration of the matrix.
If a tunnel is assigned to a dark field of the matrix, a risk analysis has to be performed.
A first examination of the Austrian road tunnels revealed, that approximately half of the
tunnels require no further risk investigation. These tunnels could, in line with the ADR,
directly be allocated to tunnel category A.
The application of the matrix is only admissible, if a set of requirements is fulfilled (e.g.
proportion of heavy goods vehicles ≤ 25%, longitudinal gradient ≤ 3%, no extraordinary
proportion of DG, etc.).

5thInternational Conference ‘Tunnel Safety and Ventilation’ 2010, Graz

 - 48 -

ventialtion tunnel length daily average traffic volume [veh/day]

type of system [m] ≤ 10000 ≤ 15000 ≤ 20000 ≤ 30000 ≤40000
traffic
HGV [%] ≤ 10 ≤ 15 ≤ 20 ≤ 25 ≤5 ≤ 10 ≤ 15 ≤ 20 ≤ 25 ≤5 ≤ 10 ≤ 15 ≤ 20 ≤ 25 ≤5 ≤ 10 ≤ 15 ≤ 20 ≤ 25 ≤5 ≤ 10 > 10
natural ≤ 250 5,62E-04 7,52E-04 9,17E-04 1,10E-03

natural ≤ 500 1,00E-05 6,96E-04 9,17E-04 1,31E-03 1,70E-03 5,51E-03 9,34E-04 1,81E-03 1,81E-03

longitudinal ≤ 700 4,10E-05 1,97E-04 9,29E-05 2,50E-04 2,73E-04 3,00E-04 4,44E-04 2,18E-04 5,00E-04 6,38E-04 8,21E-04 1,22E-03 6,31E-04 1,22E-03 1,22E-03 1,22E-03 3,00E-03 1,02E-03 1,02E-03 1,02E-03

uni longitudinal ≤ 1000 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 6,00E-04 1,06E-03 2,66E-03 8,00E-04 1,50E-03 1,50E-03 1,50E-03 1,50E-03 1,00E-02 1,01E+00 1,01E+00

directional longitudinal ≤ 1500 6,76E-05 3,26E-04 1,54E-04 5,00E-04 4,52E-04 6,00E-04 7,38E-04 3,62E-04 7,17E-04 8,92E-04 1,38E-03 1,43E-03 1,35E-03 5,15E-03 1,04E-02 1,43E-03 1,04E-02 2,19E-03 2,19E-03 2,19E-03

longitudinal ≤ 2000 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 8,71E-04 1,31E-03 1,38E-03 2,72E-03 1,33E-03 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,00E-02

longitudinal ≤ 2500 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,09E-03 1,00E+00 1,38E-03 2,75E-03 1,65E-03 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,00E-02

longitudinal ≤ 3000 1,06E-04 5,12E-04 2,42E-04 6,00E-04 5,00E-04 4,00E-04 5,12E-04 5,69E-04 1,13E-03 1,69E-03 1,00E-01 2,72E-03 1,73E-03 8,14E-03 9,30E-03 2,72E-03 9,30E-03 3,54E-03 1,54E-03 2,60E-03

transverse ≤ 4000 1,00E-04 1,00E-04 1,00E-04 1,00E-04 1,00E-04 8,12E-04 8,98E-04 1,00E-04 1,76E-03 1,80E-03 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,00E-02 1,01E+00 1,00E-02 1,01E+00 1,01E+00

transverse ≤ 7000 1,95E-04 8,43E-04 3,46E-04 7,00E-04 9,28E-04 1,19E-03 1,29E-03 2,05E-04 1,27E-03 1,80E-03 2,27E-03 2,75E-03 1,51E-03 6,14E-03 1,04E-02 2,75E-03 1,04E-02 2,54E-03 1,08E-03 1,41E-03

natural ≤ 250 ##### ##### ##### ##### ##### 7,00E-04 9,74E-04 ##### 6,76E-04 5,23E-04 ##### 6,16E-04 7,70E-04 9,17E-04 5,99E-04 0,00E+00 1,00E-02

natural ≤ 500 1,95E-04 8,43E-04 3,46E-04 7,00E-04 9,28E-04 7,00E-04 9,74E-04 2,05E-04 6,76E-04 1,04E-03 1,27E-03 7,42E-04 1,13E-03 2,75E-03 2,75E-03 6,14E-03 8,90E-04 1,00E-02 1,00E-02

natural ≤ 700 1,00E-05 5,00E-05 5,00E-04 5,00E-04 5,62E-04 7,00E-04 9,74E-04 3,55E-04 6,76E-04 1,17E-03 1,30E-03 1,65E-03 7,42E-04 5,00E-03 1,65E-03 1,65E-03 5,00E-03 1,00E-02 1,00E-02 1,00E-02

natural ≤ 1000 3,52E-04 1,14E-03 1,39E-03 6,00E-04 1,19E-03 1,00E-02 1,00E-02 1,20E-03 1,30E-03 1,00E-02 1,00E-02 1,00E-02 1,01E+00 1,00E-02 1,00E-02 1,00E-02

bi longitudinal ≤ 1500 8,22E-05 3,87E-04 4,37E-04 8,55E-04 1,12E-03 1,75E-03 2,10E-03 9,71E-04 1,72E-03 2,57E-03 1,00E-01 4,17E-03 1,94E-03 8,65E-03 4,17E-03 4,17E-03 8,65E-03 1,00E-02 1,00E-02 1,00E-02

directional
longitudinal ≤ 2000 1,24E-03 1,00E-02 1,00E-02 1,00E-02 1,22E-03 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,10E+00 1,00E-02 1,00E-02 1,00E-02

longitudinal ≤ 2500 7,46E-04 1,45E-03 3,00E-03 1,50E-03 5,00E-03 1,50E-03 1,50E-03 1,50E-03 1,00E-01 7,00E-03 5,00E-03 9,50E-03 7,00E-03 7,00E-03 9,50E-03 1,00E-02 1,00E-02 1,00E-02

longitudinal ≤ 3000 9,82E-04 1,01E-03 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-02 1,00E-02 1,00E-02

transverse ≤ 3000 ##### 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-02 1,00E-02 1,00E-02

transverse ≤ 4000 8,31E-04 1,01E-03 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-01 1,00E-02 1,00E-02 1,00E-02

transverse ≤ 7000 7,62E-04 1,13E-03 1,49E-03 1,85E-03 2,07E-03 2,07E-03 6,15E-03 1,24E-02 9,80E-03 4,16E-03 7,50E-03 1,24E-02 1,00E-01 1,97E-02 8,73E-03 4,03E-02 1,97E-02 1,97E-02 4,03E-02 1,00E-02 1,00E-02 1,00E-02

Figure 2: Classification matrix (stage 1)

4.2. Stage 2 – Detailed Approach

All those tunnels which indicate a relevant DG risk in stage 1 are subsequently reviewed in a
detailed, tunnel-specific risk analysis using DG-QRAM (Stage 2a).
Whereas the decision criterion in stage 1 is based upon the expected value, in stage 2 the
assessment is assigned to the log scaled F-N diagram. The results (scenario-based FN-curves)
are evaluated by comparing them to a defined reference line in the F-N diagram (see example
in Figure 3). If the reference line is exceeded, additional risk reducing measures are to be
investigated (Stage 2b).

Uni-directional tunnel; length = 1,000m; 55,000 veh/day; HGV 20%; DG 3%

1,E+00
All Scenarios - EV = 2,870E-2

1,E-01
Scenario 4: Motor spirit pool fire
Scenario 5 : VCE of motor spirit
F cummulated frequency [1/ tunnel, year]

Scenario 7: BLEVE of LPG in bulk

1,E-02
Scenario 8: VCE of LPG in bulk
Scenario 9: Torch fire of LPG in bulk
1,E-03
reference line for risk assessment

1,E-04

1,E-05

1,E-06

1,E-07

1,E-08
1 10 100 1000 10000
N number of fatalities

Figure 3: Reference line as assessment criterion in the F-N diagram (stage 2a)

5thInternational Conference ‘Tunnel Safety and Ventilation’ 2010, Graz

 - 49 -

4.3. Stage 3 – Alternative Route

Generally, the transport of DG on the road is not restricted as long as the requirements of the
ADR are met. If the assessment procedure determines that the investigated tunnel possesses
an intolerably high risk, restrictions to the DG transport for a tunnel are to be examined. In
this case the existence of an adequate alternative route is investigated. A road only qualifies as
an alternative route if the entire road segment is suitable and approved for heavy goods traffic.
This requires such aspects as: the number of lanes, the longitudinal gradient, the road width,
curve radii, etc. (be reviewed on a case-to-case basis).
The examination of the alternative route follows the principle that a generally allowed,
existing transport volume of DG should be carried on that route which shows the slightest
transport risk. Therefore, it must be shown that the risk of transporting DG on the alternative
route is significantly lower for the resident population than the risk of unrestricted transport
through the tunnel for the tunnel users.
The risk calculations for the alternative routes (open road sections) are also performed by
application of the risk analysis model DG QRAM.

Alternative 2

City

Investigation area
Alternative 2
Investigation area
Alternative 1
Tunnel
City

Highway Village 2
Alternative 1
2 km Tunnelroute Village 1

Figure 4: Alternative route (stage 3)

5thInternational Conference ‘Tunnel Safety and Ventilation’ 2010, Graz

 - 50 -

5. MEASURES
If in stage 2a the reference criterion in the F-N diagram is exceeded, the DG risk is rated as
unacceptable and special additional risk reduction measures needs to be investigated (Stage
2b).
Consideration made to improve traffic safety predominantly focuses on organisational and
operational measures. In this context, it should be noted that not every measure is equally
suited or efficient. The envisioned measures will thus have to be checked individually for the
respective tunnel system in an in-depth risk analysis.
The current regulation for highways BGBl. 395/2001 issued by the Federal Ministry for
Transport, Innovation and Technology (BMVIT) already stipulates operational measures
depending on tunnel length:
• Tunnels with a length between 1,000 m and < 5,000 m require a flashing warning light
on the vehicle.
• Tunnels with a length of > 5,000 m require a flashing warning light, an escort vehicle
following the transport unit and the distribution of transport documents to the
personnel of the escort vehicle.
Additional operational measures aimed at reducing the transport risk of DG may include:
• Introduction of an overtaking ban
• Introduction of a speed limit
• Installation of an information system
• Installation of a speed control system

6. DEFINITION OF RISK CRITERIA

As mentioned earlier, the societal risk is usually expressed in a graph (F-N diagram). To
determine whether the safety level is acceptable or not, an assessment of the societal risk can
be made. This is based upon a risk reference criterion which is often determined specifically
for the project in question.
For example, an officially established risk limit line concerning DG is used in Switzerland (in
the Swiss Accidents Ordinance).

6.1. Reference line for risk assessment

In the detailed analysis (stage 2), the assessment is based upon a defined assessment criterion
in the F-N diagram. This reference line was calculated by the following formula:
10 −1
For N ≥ 10 fatalities: F= ... for 1 km of tunnel (1)
N2
The slope of the equation reflects the risk aversion level. If the number of fatalities increases
by a factor of 10, the acceptable occurrence frequency decreases by a factor of 100.
The reference line was defined taking into account several aspects:
• risk level in comparable reference systems (e.g. aviation)
• reference criteria used in other countries (e.g. Switzerland, Netherlands)
• special model characteristics of DG-QRAM (e.g. adjustment for length)

5thInternational Conference ‘Tunnel Safety and Ventilation’ 2010, Graz

 - 51 -

6.2. Adjustment of the reference line for tunnel length

A risk assessment based upon the defined reference line requires a standardized tunnel length,
as the reference criterion (1) is usually based on 1 km. For the assessment criterion to be used
in the F-N diagram, an adjustment for the length of the tunnel had to be made.
Concerning the tunnel safety topic, the length of the tunnel is an especially critical factor.
Therefore, the adjustment should not be linear to the length of the tunnel. Hence, the
relationship between tunnel length and risk was modelled as an exponential function:
10 −1
For N ≥ 10 fatalities: F= × L0.5 ... L = tunnel length [km] (2)
N2
To evaluate the reference criterion (2) transport risks of several common road tunnels in
Austria were calculated and assessed in the F-N diagram.

7. INTERNATIONAL OUTLOOK
The ADR’s tunnel regulations have been valid since the 1st of January 2010. Although the
European countries have already assigned many tunnels with restrictions to ADR’s tunnel
categories, this implementation process isn’t completely accomplished. The network-wide
risk assessment and the assignment to risk categories will take some time to be finalized;
especially for states with a great number of road tunnels.
Thus current development trends in the transport industry ought to be observed and taken into
account. The transport industry is likely to face certain additional costs for the transport of
DG as alternative routes tend to result in longer transport times and longer carriage distances.

8. REFERENCE LIST
Directive 2004/54/EC of the European Parliament and of the Council of 29 April on minimum
safety requirements for tunnels in the Trans-European Road Network
Bundesgesetz über die Sicherheit von Straßentunneln (Straßentunnelsicherheitsgesetz STSG),
BGBL 54/2006, Österreichischer Nationalrat, 2006
ADR 2007 European Agreement concerning the International Carriage of Dangerous Goods
by Road, United Nations 2006
Beschränkungen für Beförderungseinheiten mit gefährlichen Gütern beim Befahren von
Autobahntunneln, BGBl. II Nummer 395/2001, Verordnung des BMVIT für Autobahnen
Schweizer Störfallverordnung StFV (Swiss Accidents Ordinance), EDMZ, 1991
Transport of Dangerous Goods through road tunnels – Quantitative Risk Assessment Model,
OECD/PIARC/EU, 2003
RVS 09.03.11 Tunnel Safety - Tunnel Risk Model, FSV (Österreichische Forschungsgesell-
schaft Straße – Schiene - Verkehr), 2008
RVS 09.03.12 Tunnel Safety – Risk Assessment of Dangerous Goods Tansports in Road
Tunnels (draft), FSV (Österreichische Forschungsgesellschaft Straße – Schiene - Verkehr),
2008

5thInternational Conference ‘Tunnel Safety and Ventilation’ 2010, Graz