Proceedings of the ASME 2016 35th International Conference on Ocean, Offshore and Arctic Engineering

OMAE2016
June 19-24, 2016, Busan, South Korea

OMAE2016-54983

A SYNTHESIS MODEL FOR FLNG DESIGN

Daniel Prata Vieira Rodrigo Sauri Lavieri Thiago Peternella Rocha

Numerical Offshore Tank (TPN) Numerical Offshore Tank (TPN) Numerical Offshore Tank (TPN)
University of São Paulo University of São Paulo University of São Paulo
São Paulo, SP, Brazil São Paulo, SP, Brazil São Paulo, SP, Brazil

Raul Dotta Fabiano Rampazzo Kazuo Nishimoto

Numerical Offshore Tank (TPN) Numerical Offshore Tank (TPN) Numerical Offshore Tank (TPN)
University of São Paulo University of São Paulo University of São Paulo
São Paulo, SP, Brazil São Paulo, SP, Brazil São Paulo, SP, Brazil

ABSTRACT ranking of the generated solutions through an objective

The increasing demand for natural gas is encouraging the function.
development of novel floating units’ designs, capable of Capacities, production rates, equipments, load distribution,
processing large quantities of hydrocarbon. These units called environmental actions, stability, sea keeping and structural
FLNG (Floating Liquefied Natural Gas) are facilities that design estimates are the major areas to consider and will be
produce, process and store liquefied natural gas (LNG) related to one or more mathematical models, constraint and
offshore. Once the topside and tanks of a FLNG are larger and objective functions. The work will present a general overview
more complex than the regular FPSO vessels, a design process of each model separately and how they work together, as well
considering these particularities must be used. as examples of solutions and analyses depending on the input
Once just few FLNG units are under construction and under values.
design and not yet in operation, the information on the design It must be clear that this approach is applicable just in the
first stages is poor. It is difficult to obtain a first hull sizing early stages of design to obtain the first hull sizing. After that it
without taking in account the complexity mentioned above. is necessary to fall back on the traditional iteration process to
Thus, a set-based approach that works with sets of possible rely in a feasible design.
solutions that are analyzed and compared using a merit function
in order to select the best and feasible solutions was used. INTRODUCTION
However, to produce a sufficiently large family of solutions, Although several FPSO (Floating Production and Storage
which includes most of the solution space, either the solution Units) have been designed and built for the last three decades,
descriptions or the models must be simplified. From the only recently the first platforms dedicated to process and store
computational point of view, the analyses of a family of design natural gas (FLNG) are being constructed. Technological
solutions basically relies on an initial parameterization of the issues, particularly concerning the storage of the gas, make the
object and a set of mathematical models that, as a group, will commercial use of this fuel unfeasible. Thus, it was (and still is)
be referred as synthesis model. Additionally, some restrains are reinjected in the well or burned.
also applied to eliminate unfeasible solutions. The output of the The risk associated to a large dimension pressure vessel is
synthesis model is a set of performance quantities that will be unacceptable. At the same time, to store pressured gas in an
used to rank the solutions. array of small cylinders would increase the weight of the
This design approach is particularly useful to deal with structure and would make it inefficient. Therefore, the most
project trade-offs and to optimize multiple characteristics. suitable technology for storage of gas offshore is the
Optimal solutions belongs to a surface (or a hyper surface) liquefaction. So, nowadays both technologies are applied in gas
called Pareto boundary. This paper aims to achieve a platform transportation. Quantities carried by LNG (Liquefied Natural
design capable of producing, storing and offloading liquefied Gas) ships tend to be larger than those carried by CNG
natural gas. It must safely survive under environmental (Compressed Natural Gas) ships.
conditions of Santos Basin in São Paulo, Brazil. In the same The liquefying process of natural gas is achieved by
way, the design should guarantee the shortest downtime as well reducing its temperature to approximately -160ºC. In such
as keep costs, of acquisition and operation, as low as possible. condition, the gas becomes liquid and can be stored at a
Each of these characteristics must be quantified to allow a pressure close to the atmospheric. This liquid density is

1 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 approximately 0,5t/m³, less than crude oil and thus requires different loading conditions as well as for distinct annual
large heavily insulated tanks for storage. production rates.
For that reason, the whole design of the FLNG rests on
supporting a heavy processing plant and carrying large tanks
Production, Fractions of LNG and LPG, Number of
with relatively light load. These two aspects lead to a large operating Days, Number of days between offloads,
dimension hull. Deck Type
Free surface effect inside the tanks is another major issue to
be considered. Finally, the processing plant depends on low tilt
angles and low acceleration levels to operate, (Pettersen, et al.
2013), making a careful seakeeping investigation necessary.
PARAMETRIZATION

SYNTHESIS MODEL OVERVIEW

As mentioned before, a set-based approach was used to
provide more design information in the early stages of design.
This section gives an overview of this approach. TOP SIDE MODEL
The kernel of a parametric design routine is its synthesis
model, considering that all mathematical models, describing
and evaluating each aspect of the design object, are assembled
in a single continuous algorithm. The sequence of routines is PRE-STABiLITY
defined by the amount of information required as input, i.e. in (Draft Estimative)
order to precisely compute the stability characteristics of the
hull, the weight of the structure as well as its 𝐶𝑜𝐺 (Center of
Gravity) must be previously known.
Figure 1 presents the sequence of models from top left to HULL STRUCTURE
bottom right. In every step of the loop, an individual is
generated and described by a set of parameters. Through the
production requirement (i.e. 4 million metric tons per year) and
the time between two offloading operations it is possible to No
DNV RULES
calculate the required tank volume and load weight. Later, a
routine creates an estimate for the topside layout, mass (all
Yes
inertial characteristics) and centroids. Then, the structural
model predicts the mass of the hull considering all the previous
information. Finally, stability and seakeeping are computed as STABILITY
the inertia of the complete structure is known.

No
IMO RULES

Yes

STORED AS A
HULL MOTION FEASIBLE
SOLUTION

Figure 1 – Mathematical models of the synthesis model N=N+1

During this process some constraints are applied and No

unfeasible individuals are eliminated. At the end of each step, a N>Number of
feasible solution is recorded, producing a family of them that desired solutions
can be displayed as a cloud. As shown in Figure 1, each
solution has its geometrical parameters and performance
characteristics which will be further used to evaluate and rank Yes
the solutions. The synthesis model flowchart is presented in
Figure 2. END
The number of solutions generated is limited by
computational time. A whole set of solutions were generated for Figure 2 – Synthesis Model Flowchart

2 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 A maximum downtime of 30 days per year was taken as
constraint. The uptime is evaluated considering the operational After the establishment of the fixed parameters, the model
limits of the system, which can be summarized as: the variables must be defined.
maximum acceleration at the top of the liquefaction tower
should be less than 0.1g and the maximum roll and pitch
combined angle should be less than 1 degree.

PARAMETRIZATION
The choice of the parameters that describe the platform
depends on the mathematical models applied to analyze and
evaluate the solutions. Redundant parameters must be avoided
to guarantee the reliability of the method. Only after the
establishment of all mathematical models in design process, an
optimal set of parameters should be obtained.
Besides the geometric parameters, some fixed values are
used to generate a solution. Basically they are:
 Production per year (MTPA)
 Fractions of LNG and LPG produced Figure 4 – Hull form parameterization
 Number of operating days during the year
The set of parameters that describes the solution are:
 Number of days between offloads
 LOA (Length Overall)
 DeckC (Deck Configuration)
 B (Beam)
The last value mentioned, DeckC, was introduced to  D (Depth)
identify three distinct layouts for modules on the deck. Layout  Hpo (Height of stern chamfer)
1 considered the turret outside the deck at the bow of the hull  Hpr (Height of bow chamfer)
with the flare, all processing modules and the living quarter at  Lpo (Length of stern chamfer)
the stern of the structure. Layout 2 also considered the turret  Lpr (Length of bow chamfer)
outside the deck at the bow of the hull, but the living quarters  Ntanks (Number of storage tanks)
right behind it, allowing the weathervane effect to take away  NtanksT (Number of storage tanks rows)
exhausting gases or any accidental leakage. In Layout 3 the  hDD (Height of double deck)
turret was moved inside the hull structure and the living quarter,  hDB (Height of double bottom)
as well as all generator sets and workshops, were kept ahead  wDS (Wide of the double side shell)
from it. This configuration takes advantage of the
weathervaning effect and avoids the flow of gas underneath the A range for each of the parameters is defined and an
accommodation. However, moving the turret inside the deck individual is randomly created by picking parameters between
area implies the increase of hull length and produces structural these ranges.
issues. Figure 3 shows a sketch of these three different Once the parameterization is known, ranges for the
proposed topside configurations. parameter must be determined. This process, normally based on
similar projects research, aims to reduce the size of the space of
solutions.

TOPSIDE MODEL
The refining process of natural gas depends on three sets of
Layout 1 equipment: Pretreatment, Liquefaction and Utilities . Such
division can be extended to a floating plant (FLNG), except that
this structure has some additional facilities, as presented in
Table 1.
This process has as products the liquid natural gas (LNG),
Layout 2 Liquefied Petroleum Gas (LPG) and condensate. The following
Table 1 shows the main modules of the refining process of
natural gas (Ronceiros, 2008).

Layout 3
Figure 3 – Proposed topside configurations

3 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 1 – Main elements of the gas refinement process The routine to estimate topside uses as Inputs: LOA, B, D,
Pretreatment Liquefaction Utilities DeckC, Production per year, and delivers as Outputs: Total
Turret Refrigeration system Power generators mass of deck modules, Center of gravity of the Deck, Length of
MEG recovery Liquefaction system Air System Deck, Beam of Deck, and Longitudinal Position of Living
Equipment admission BOG compressor Bleeding Quarters. This information is fed into the other blocks of
Removal acid Cooling intake Nitrogen synthesis model, described further.
Dehydration system Subdivision
Mercury removal Flare
Water treatment
Accommodation;
Control Center;
Workshops;
Pipe Rack.

A routine was developed to compute the areas required for

each module in layouts (DeckC previously mentioned), as well
as the center of gravity, total mass, length and beam about the
deck.

Figure 6 – Graphic representation of the topside layout (DeckC =

1). Each volume represents a module and has a particular density
and 𝑪𝒐𝑮

How this work relies on the early stages of the FLNG

design, important issues like the turret are not fully developed.
In this stage is important to have an idea of the weight and the
required space to allocate that. The increasing of the weight in
structure due the required reinforcements is incorporated in the
15% weight margins.
Figure 5 – Length Vs Beam for existing FLNG designs. In further stages, it is necessary to develop the turret sizing
and evaluate the impact on the main dimensions.
After the analysis of existing FLNG designs as those
presented in Figure 5, it was observed that the weight of the STRUCTURES
modules varied according to the equation (1) : In the initial steps of design process, usually, the
information about hull structure is limited. So, in this work, a
𝑊𝑚𝑜𝑑𝑢𝑙𝑒 = 𝐶 ∗ 𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛0,6 (1) method for obtaining initial information was proposed. The
method consists in designing a midship section that is in
where, 𝐶 is a constant derived from each module through the agreement with the primary requirements of Det Norske Veritas
analysis of these similar vessels. Weight margins of 15% were (2012). Once a midship section is defined, it is used to estimate
taken in account inside the constant 𝐶. the hull structural mass and center of gravity. Furthermore, the
Center of Gravity (𝐶𝑜𝐺) of each module was estimated transverse structural elements, as ring girders and bulkheads are
according to similar structures, and adopted as a linear defined to provide more accuracy in the weight estimation.
regression, whereas the 𝐶𝑜𝐺 of the Topside, one of the outputs Obviously this simplified method comprises only a first
of the routine, was calculated considering the weighting center estimative of this parameters, so in other advanced steps it is
of each module. possible to detail the layout obtained in this section and
The length of the topside was calculated by the summation improve the hull structure evaluation according to more
of maximum length between a pair of modules, placed specific rules.
transversely (Figure 6), plus a minimum required longitudinal Two main assumptions were made to evaluate the structure
distance between the modules. arrangement. First, it was assumed that tanks will be separated
An analogous method was applied to calculate the topside in modules which will be installed after the hull construction.
width. The width of the modules, minimum required distance An IMO Type B (SPB) prismatic tank will be used, which have
and the width of the piperack were added. The definition of the the ideal characteristics to store LNG and LPG (IHI Offshore
deck beam was based on the maximum width among all sets of Group 2014).
modules. This assumption is important once the hull structure can be
designed independently of the tank structure. Thus, the DNV

4 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 primary requirements are fulfilled without considering the So, a typical midship section arrangement is composed by
inertia due to the addition of the tank structure. It was the following basic elements:
considered a conservative approach once the tank structure
presence will contribute to the increase of midship section  Side Shell;
properties.  Inner Side Shell;
The second assumption is that the other hull sections are,  Bottom;
initially, considered equal to the midship section. It is also a  Double Bottom;
conservative approach, once the main efforts are located in the
 Deck;
hull center. In future steps a study can be made to evaluate the
impact of this assumption in the hull weight.  Double Deck;
The routine flowchart is presented in Figure 7 and consists And the following secondary elements:
basically in evaluating the midship section arrangement, the
transverse structure arrangement and the hull weight and center  Girders; and
of gravity using an iterative process.  Stiffeners.
The first midship section is generated with basic elements
Inputs: which are kept the same during the iterative process. This
- Main Dimensions
midship section is made with only a few secondary elements.
- Tank Arrangement
The iterative process is responsible for adding elements until
the requirements are achieved. A minimum and a maximum
amount of each type of element are defined. The elements are
added according to a pre-defined order to avoid excessive
First midship section
addition of elements only in a part of the section. Furthermore,
generation
several numbers of configurations that are in accordance with
rules are provided and that with minimum area, consequently
less steel, are chosen.
Figure 8 presents an example of a midship section generated
Midship section by the structures subroutine.
properties evaluation
Iterative
Midship section Process
modification

Minimum
No
requirements
verification

Yes

Transverse Structure
Arrangement

Figure 8 – Example of generated midship section

Outs: Once the midship section is generated the transverse

- Midship section structure is evaluated. First, the transversal bulkheads are
arrangement located between the tanks. So in the tank span, several ring
- Hull weight and CoG girders are distributed according to a pre-defined distance
estimation between them. Later, the isolation is evaluated according to the
Figure 7 – Structures Routine Flowchart tank dimensions. Figure 9 shows an example of generated
structural elements.
The hull main dimensions and the tank configuration are Using the volume of each element and its location it is
provided as input by the synthesis model. Thus, a first midship possible to estimate the hull weight and center of gravity. Also
section is generated using one typical hull structure here, a margin of 15% in the structure weight was used for
arrangement without considering longitudinal bulkheads which safety purposes.
will be provided by the SPB tank.

5 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 the water plane around the hull for any heel angle. Besides the
water plane points, all hull vertices below were identified to
define a convex hull that corresponds to the displaced volume.
This convex hull is determined solving the Delaunay
problem for the set of points. Thus, the boundary of the volume
is divided into several triangular panels and, from them; it is
possible to compute the total volume and centroid of the
underwater portions of the hull.
One may notice that the barge shape, shown in Figure 4,
simplifies the volume computation since all the points can be
analytically found.

Hydrostatic equilibrium
The procedure described above does not guarantee that the
displaced volume maintains the same during the increase of the
heel angle. To do so, one may apply an iterative method to find
the position of the water plane. In the code described here, this
search is done by changing the vertical position of the whole set
of rotated points. The search method is a simple dichotomy
algorithm and the initial interval for the search was selected
considering the amount of rotation. Small changes of angle
Figure 9 – FLNG Structural elements considered in model
produce equally small changes on water plane position while
large changes produce great modifications. It was defined a
STABILITY
Just like new FPSO, the FLNG discussed herein will receive tolerance of 0.1% of the initial displacement to stop the
a new hull, particularly designed for it. Thus, the main focus of iterations. The code described on the previous item is executed
the hull design is the stability and seakeeping characteristic. for each iteration.
Once the current position of the waterline is founded, the
The main objective of this stability model is computing the
center of the volume is computed and saved for future
righting arm (GZ) curve of the parameterized hull, considering
calculations.
free surface effects (FSE) of the tanks. It must be fast and
robust to deal with several non-linearities related to heel angles
and several tank levels, as shown in Figure 10. The main steps Tanks Volume
of the stability routine are presented below: Dividing the number of tanks by the number of tank rows,
the number of divisions in the longitudinal direction is
obtained.
 Hull volume computation
All tanks are supposed to have the same dimensions, so the
 Hull hydrostatic equilibrium
tanks are distributed inside the double hull span, considering
 Tank volume computation the space required for the tank thermal isolation layer.
 Tank equilibrium Furthermore, the tanks are not located below the living
 Righting arm computation quarters, according to the Det Norske Veritas (2012) rules, and
nor in the inclined bottom region.
Hull volume computation For the first approach, it was assumed that the vessel load is
Using a set of parameters, shown in Figure 4, the vertices of a homogeneous liquid equally distributed in the tanks.
the geometry are generated. A rotation around the 𝑥 axis is The volume of the tank was calculated using the same
performed, corresponding to the desired heel angle. The routine routine presented before except for the geometry change (a set
only computes transversal stability, but through the addition of of 6 planes plus the water plane).
other direction (around the y axis) to the rotation matrix it could
also calculate longitudinal stability. Tank Equilibrium
Using the rotated set of points, 8 planes can be defined by In order to guarantee that the volume estimates in the tanks
the normal vectors, calculated using cross product between two are correct even for large heel angles, the equilibria of the tanks
pair of vertices. The water plane was also defined by its normal, must be calculated. The same iterative method applied for each
a positive unitary vertical vector applied at point 𝐭 = [0 0 𝑡]. tank was used to estimate the position of the water plane area.
Two by two, all the defined planes were intersected with the The free surface shape can change drastically while the heel
water plane producing a cloud of points. Only the points inside angle is modified, especially if the tank is near empty or near
the hull volume were selected. This was done by comparing full, as presented in Figure 10.
signals of inner products between planes’ normal vectors and
the point of the cloud. This procedure produces the vertices of

6 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Righting Arm Curve
10
GZ (without FSE)
GZ (with FSE)
8 GMsen()

Figure 10 – Change on the free surface area for 10º heel angle
4

GZ [m]
Once the position of the free surface of the tank is
2
determined, the centroid is calculated and this information is
replicated for all the tanks.
0

Righting Arm Curve (GZ)

-2
Repeating the previously presented process for several heel
angles, an array of 𝐶𝑜𝐵 (center of buoyancy) positions are
-4
calculated, as well as the corresponding 𝐶𝑜𝐺 (center of 0 10 20 30 40 50 60 70 80
Angle [deg]
gravity). The horizontal distance between these two points for
each heel angle produces the righting arm curve (𝐺𝑍). The Figure 12 – Free surface effect on the righting arm curve
displacement of the liquid inside the tanks changes the 𝐶𝑜𝐺
position while 𝐶𝑜𝐵 of the hull is modified as the heel angle is The stability routine uses as Inputs: LOA, B, D, Hpo, Hpr,
increased. Lpo, Lpr, Ntanks, NtanksT, hDD, hDB, wDS, Structural mass
Righting Arm Comparison and CoG, Topside mass and CoG, tanks levels, and delivers as
4 Output: The draft at different loading conditions, righting arm
curve with and without FSE to be compared to IMO criteria,
3 𝐶𝑜𝐺 and moment of inertia of the load.

2
HULL MOTION
Numerical Model
GZ [m]

The evaluation of hull motion analyses was carried out

1
using WAMIT (Wave Analysis MIT), a code that uses potential
theory (Newman, 1977) and BEM (Boundary Element Method)
0 to solve the frequency domain problem of diffraction and
Commercial Software radiation for floating bodies in 6 degrees of freedom. The input
GMsin() to the numerical model is basically:
-1
FLNG routine
Estimated Downflooding  mesh of the wet surface of the vessel,

-2
0 10 20 30 40 50 60 70 information about mass and inertia,
Angle [deg]
 a vector of wave periods (or frequencies), and
Figure 11 – Comparison between the righting arm curve produced
by the FLNG routine and other sources
 a vector of wave incidence angles.
In addition, some configurations parameters depend on the
For small angles, the 𝐺𝑍 curve can be approximated by type of simulation are also provided.
𝐺𝑀𝑠𝑖𝑛(𝜃), where 𝐺𝑀 is the metacentric height of the hull. This The hull wet surfaces are provide by a subroutine
quantity can be analytically calculated for small angles and box responsible by write de Geometrical Data File (GDF). This
shaped bodies. subroutine uses the geometric parameters generated in the
The GZ curve was compared to the linear approximation synthesis model by writing a set of surfaces using NURBS
and presented good agreement, as presented in Figure 11. formulation. Figure 13 shows examples of the hull wet surface
Furthermore, a comparison with commercial software was used to calculate the behavior in waves using WAMIT. In the
performed and good coherence was observed for larger heel cases where DeckC = 3, a moonpool is located in the hull to
angles. provide more accurate motion. The same procedure presented
Figure 12 presents the GZ curves, without FSE and with it, in Malta et al. (2006) was used to evaluate the numerical model
as well as the GMsin(θ) approximation, to show the effect of with moonpool.
the free surface.

7 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 The routine basically uses the data generated by the
synthesis model as main dimensions, weights and centers of
gravity, and performs two tasks: the first is to write the input
data file described above. All information is provided by the
synthesis model. The second task is the generation of
geometrical data file. After, the routine moves all files to a
specific folder and runs the WAMIT code.

Downtime Evaluation
(a) Model with turret outside the hull Another subroutine is responsible to post-processing the
results provided by WAMIT. This subroutine uses a list of
points of interest and a set of local (Santos Basin, Brazil)
environmental conditions to evaluate RAOs, accelerations and
natural periods.
The FLNG will use a turret as mooring system (Leite, et al.
1999), so regardless of the wave direction, an approach using a
30 degrees offset between the vessel and the wave was
considered. It can be not a conservative approach, once in
(b) Model with turret inside the hull extreme environmental conditions this angle can be greater than
Figure 13 – Example wetted surfaces of a FLNG hull.
30 degrees, but it is sufficient to compare the solutions between
themselves. In further design stages, extreme conditions must
For the simulations, the code default settings and higher-
be considered to evaluate downtime, structural resistance and
order method with a panel size equal to 15 meters will be used.
mooring system performance.
More details can be found in WAMIT (2012).
The properties of mass and inertia are obtained in previous
DATA ANALYSIS
calculations and will be inserted in the model through a mass
In this section an example of data analysis is presented. A
and inertia matrix.
To an extensive analysis of the cases an automation of the simulation using the input parameters presented in Table 2 was
carried out and 350 feasible cases were generated.
numerical model was proposed. This automation includes a
computational routine which follows the data stream shown in
Table 2 – Input parameter of example case.
Figure 14.
Production per year (MTPA) 4
Fraction of LNG 75%
Synthesis
Synthesis Model
Model
Fraction of LPG 25%
Number of operating days during the year 300
Number of days between offloads 21
DeckC (Deck Configuration) 1

Geometric data Write input data

file generation files There are different ways to analyze the data provided by the
synthesis model, see for example Tancredi (2008). In this work
we considered as objective function the minimization of
structural mass, and as constraints a maximum of 30 days/year
WAMIT
WAMIT
downtime. So, in the space of generated cases that one which
execution
execution agrees with constraints and has the minimum structural weight
is a good candidate to the best generated solution. Figure 15
presents the mass structure versus the system downtime. The
cases in blue are in agreement with downtime limit and the
Points of Results
Results Post-
Post- Environmental cases in red are not. A possible Pareto boundary is highlighted
Interest Processing
Processing Conditions
to show the family of good solutions.

RAOs, Accelerations
and Natural Periods

Figure 14 – Data flow of the routine evaluation of wave behavior.

8 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Figure 15 – Structure Mass Vs. Downtime for the 350 generated
cases.

Once the downtime is a restriction, the chosen solution was that

with the minimum mass structure (inside the red circle). Table 3
shows the dimensions and characteristics of the chosen
solution.

Table 3 – Dimensions and Characteristics of chosen solution

Figure 16 – Beam, Length and L/B regressions of existing designs
Dimensions
considering de Annual Production
L 473 m
B 77 m
CONCLUSIONS
D 37 m
A synthesis model considering topside sizing, the stability
Hpo 33.3 m
criteria, structure criteria, weight estimation and motion in
Hpr 33.3 m
waves was presented. Through the assembling of distinct
Lpo 22 m
design areas it was possible to obtain the big picture of the
Lpr 22 m
design of such structures, at least by the naval architecture point
hDD 3.2 m
of view. This is an important achievement, since there are
hDB 2.3 m
relatively few data available from existing designs.
wDS 3 m
Characteristics
Among several issues, it was made clear that large deck area
is required; therefore a considerable hull depth is needed, in
Structural Mass 1.49E+05 t
order to make it structurally feasible. Structural mass is directly
Topside Mass 8.80E+04 t
Total Mass 5.95E+05 t
related to these facts and thus was chosen as a ranking criterion.
KG 31.85 m
Three hundred and fifty possible solutions were obtained
Draft 17.1 m
and a seakeeping analyses was carried out to evaluate the
system downtime. The solutions were presented in a cloud of
points considering the structural mass and the downtime. The
To verify the feasibility of chosen solution, the beam, length
and the non-dimensional L/B were plotted in Figure 16 with chosen solution was that with minimum structural mass and in
regressions of the same parameters from existing designs agreement with downtime constraint.
(presented in Figure 5). A good agreement between the The chosen solution was compared with existing designs
dimensions evaluated by synthesis model and existing projects regressions and proved to be consistent.
However, this is only the first step in the FLNG design.
was obtained.
With the initial size provided by the synthesis model it is
necessary to evaluate each area as General Arrangement, Intact
Stabilty, Seakeeping, Risers, Structure, Mooring System,
Weight and Centers, Damaged Stability, Installation design and
Costs. These areas must be evaluated in sequence according to
the traditional spiral method to obtain a refined design.

9 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 FURTHER STEPS IHI Offshore Group. Self-supporting Prismatic shape IMO type
A more extensive simulation considering a larger number of B (SPB) Technology. 12 2014.
cases must be carried out to fill the presented cloud of points http://www.ihi.co.jp/offshore/spbmenu_e.htm.
next to the Pareto boundary.
Leite, André Jacques de Paiva, Alexandre Nicolaos Simos,
Simulations with the different topside layouts must be
Eduardo Aoun Tannuri, and Celso Pupo Pesce.
carried out and comparisons between them must be provided.
"FPSO-Turret System Stability and Wave Heading."
In the structural model, analysis considering transversal
Proceedings of the Ninth (1999) International
efforts should be implemented to accomplish the classification
Offshore and Polar Engineering Conference. Brest,
society rules in the early stages.
France, 1999.
The seakeeping numerical model should take in account the
cargo motions inside hull to provide more realistic data about Malta, E. B., M. F. Cueva, and K. Nishimoto. "Numerical
hull motions and accelerations, which can change the downtime Moonpool Modeling." 25th International Conference
evaluation. on Offshore Mechanics and Artic Engineering. 2006.
Once a solution is obtained further studies as riser and OMAE2006-92456.
mooring analysis, offloading analysis, wind and current efforts
must be carried out to verify the design feasibility. Newman, John Nick. Marine Hydrodynamics. MIT Press, 1977.
Pettersen, J., O. Nielsen, S. Vist, L. E. N. Giljarhus, K.
ACKNOWLEDGMENT Aasekjaer, and B. O. Neeraas. "Technical and
The authors acknowledge Frade Japão Petróleo Ltda for the Operational Innovation for Onshore and Floating
financial support and for the motivation of this work. LNG." LNG17 Proceedings. 2013. 1-12.
Ronceiros, N. G. Simulação do Processo de Liquefação de Gás
REFERENCES Natural APCI C3MR. Rio de Janeiro: PUC-Rio, 2008.
Det Norske Veritas. Rules for Classifcation of Floating Tancredi, Thiago Pontin. Otimização Multidisciplinar
LNG/LPG Production, Storage and Loading Units Distribuída aplicada a Projetos de Engenharia. PHD
(DNV-OSS-103). DNV, 2012. thesis, Universidade de São Paulo, 2008.
WAMIT. "User Manual." MA: www.wamit.com, 2012.

10 Copyright © 2016 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 09/13/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use