Académique Documents
Professionnel Documents
Culture Documents
TO SHIP DESIGN
A Papanikolaou, National Technical University of Athens, Athens/Greece
S Harries, FRIENDSHIP SYSTEMS, Potsdam/Germany
M Wilken, Germanischer Lloyd, Hamburg/Germany
G Zaraphonitis, National Technical University of Athens, Athens/Greece
SUMMARY
An integrated design and multiobjective optimization approach to ship design is herein presented. It integrates methods
and software tools for the simultaneous evaluation of key measures of merit in the early phase of ship design. The
implemented approach is herein applied to the design of an Aframax tanker for which a variety of parameters related to
payload, steel weight, strength, oil outflow, stability and hydrodynamics were considered within an integrated
multiobjective design and optimization procedure. Required Freight Rates (RFR), Oil Outflow Index (OOI), Energy
Efficiency Design Index (EEDI) and maximum speed for given main engine margins were determined so as to rank
design alternatives. Formal exploration and exploitation strategies were utilized to investigate the design space and,
subsequently, advance competing design proposals into certain directions such as maximum energy efficiency, attainable
speed and environmental protection in case of accidents. The paper focuses on the integration of design methods, of
related software tools and optimization, utilizing the design of a tanker as an elaborated demonstration example to
illustrate the holistic view of the adopted approach.
Inherently coupled with the design process is design In recent time, shipping industry's major ecological
optimization, namely the selection of the best solution concerns are related to energy/ fuel consumption and
out of many feasible ones on the basis of a criterion, or associated green-house gas emissions. This comes on top
rather a set of criteria. A systemic approach to ship of longstanding concerns regarding accidental oil
design may consider the ship as a complex system pollution, particularly by crude oil carriers. The
integrating a variety of subsystems and their components, introduction of the EEDI as put forward by the [11]
e.g. subsystems for cargo storage and handling, raises both awareness and triggers efforts for higher
energy/power generation and ship propulsion, energy efficiency, while high bunker prices continue to
accommodation of crew/passengers and ship navigation. excite economic pressure on the operators. A recent
Independently, considering that ship design should comprehensive study on the risk of large oil tankers
actually address the whole ships life cycle, it may be showed that the potential loss of cargo is dominated by
split into various stages that are traditionally composed grounding and collision accidents, along with fire and
of the concept/preliminary design, the contractual and explosions, [10]. Enlarged double hull width and double
detailed design, the ship construction/fabrication process, bottom height, enhanced compartmentation and varying
ship operation for an economic life and size of tanks can lead to improved environmental
scrapping/recycling. It is evident that the optimal ship protection, without compromising ships efficiency, as
with respect to her whole life cycle is the outcome of a elaborated by Papanikolaou et al. [12].
holistic optimization of the entire, above defined ship
system for its life-cycle (Papanikolaou, [13]). An analysis using Lloyds Register Fairplay WSE
Inherent to ship design optimization are also the Database revealed that one fifth of the existing Aframax
conflicting requirements resulting from the design tanker tonnage would be older than 15 years by 2012.
constraints and optimization criteria (merit or objective Even though current tanker capacity appears to outweigh
functions), reflecting the interests of the various ship anticipated demand of oil transport, the fleet's ageing is
design stake holders: ship owners/operators, ship builders, likely to trigger replacements.
classification society/coast guard, regulators, insurers,
cargo owners/forwarders, port operators etc. Assuming a It is therefore safe to assume that new tanker designs will
specific set of requirements (usually the shipowners be sought in the near future. However, it is not obvious
requirements for merchant ships or mission statement for what will be the main driving forces:
naval ships), a ship needs to be optimized for lowest Safer shipping by containing or mitigating oil
construction cost, for highest operational efficiency or outflow in case of an accident,
lowest Required Freight Rate (RFR), for highest safety Greener operations by reducing emissions per ton-
and comfort of passengers/crew, for satisfactory mile of cargo,
protection of cargo and the ship herself as hardware and Smarter business by increasing returns (higher cargo
last but not least, for minimum environmental impact, capacity and lower fuel consumption).
particularly for oil carriers with respect to marine
pollution in case of accidents. Recently, even aspects of A reasonable combination is likely to be favored over an
ship engine emissions and air pollution need to be extreme, depending on the specific situation and
considered in the optimization of ship design and preference of the stake holders. The more high-quality
operation. Many of these requirements are clearly design data are available the easier it is to understand
conflicting and a decision regarding the optimal ship opposing influences, come to a sound judgment and
design needs to be rationally made (Fig. 2). choose the BEST compromise (Sames et al, 2011).
shift of bulkhead heads
frame spacing
Figure 2: General arrangement along with layout of tanks and selected free variables
3.2 DESIGN APPROACH Cargo tank capacity in full load and design load
conditions,
The process was set up in the FRIENDSHIP-Framework Steel weight of the cargo tank area,
(FFW), combining POSEIDON, NAPA and SHIPFLOW Maximum ship speed at design, ballast and scantling
simulations. The following key measures were drafts,
computed: Probability of oil spill in case of accidents measured
by IMO's oil outflow index (OOI).
A general flow chart is presented in Fig.5. For each Operational impact measured by the energy
variant a hull form is generated within FFW along with efficiency design index (EEDI), combining engine
alternative tank configurations. The structural design in power, deadweight and ship speed according to IMO,
the cargo block area is then determined with POSEIDON Financial attractiveness measured in terms of
in accordance to the prescriptive part of the Common required freight rate (RFR), combining the annual
Structural Rules (CSR) for Double Hull Oil Tankers. The cost of transport via capital, fuel and other operating
hydrodynamic performance of the design alternatives is costs with the number of roundtrips times cargo
determined via a response surface model (RSM) built mass per year.
from a priori flow simulations using the CFD code Free variables of the overall investigations were
SHIPFLOW, in connection with the potential flow code parameters that control the hull form (outer shell), the
(XPAN) and viscous (CHAPMAN) analyses. This is tank layout and geometry as well as the inner structure,
followed by a batch mode execution of NAPA to get the Table I and Fig.2.
stability and trim characteristics plus the probability of Having established the most favourable main particulars,
oil outflow for the generated alternative tank cargo tank arrangement and cargo block scantlings,
configurations and hull form shapes. The process is within a global optimization procedure, the ship's aftbody
complemented by several additional features available was subsequently fine-tuned with regard to wake quality
within the FRIENDSHIP-Framework, which enable the and total resistance. In addition, systematic changes were
gathering, synthesis and analysis of the various results undertaken to study the dependencies of selected
from all conducted external simulations. measures of merit on specific parameters (sensitivity
From the determined values of cargo tank capacity, steel studies), e.g. the change of oil outflow probability by
weight and ship speed two combined performance further increasing the double bottom height of the
measures (indicators) for ecology and economics were foremost tanks.
derived:
For hydrodynamic analyses, see section 4.2, the length, are global parameters of the fully parametric model, the
beam, longitudinal position of the center of buoyancy variations of XCB and displacement were realized by
(XCB) and displacement volume were changed means of a Generalized Lackenby for partially
systematically. While length and beam of the hull form parametric modifications, [1] & [2]. Local parameters
defining the shape of the aftbody's basic curves were takes the hull form, the minimum distance of the inner
changed during the hydrodynamic fine tuning. In this structure to the hull (outer shell) and the longitudinal
phase 12 local parameters were varied, for instance the position of the engine room's bulkhead as inputs. The
fullness of the diagonal starting in the forward clearance collision bulkhead's position is computed according to
point, the forward clearance of the propeller and the IMO rules.
fullness of the aft bulb curve in the midship plane (Tillig,
2010). During the global optimization the side shell width at
deck height, the double bottom height at amidships, the
An existing geometry model from previous studies by angle and width of the hopper plate and the step in the
Papanikolaou et al. [13] was taken as a good starting double bottom towards the foremost tank were changed.
point for the design task and the parametric model was The bulkhead positions were moved discretely according
adjusted to closely resemble the existing hull form. to the frame positions. The total number of frames was
Generating a new variant then simply meant changing controlled by specifying the number of frames per tank.
the selected set of parameters. The first tanks (COT1) and the last tanks (COT6) were
flexible in length by allowing shifts of the bulkhead
4.2 TANK ARRANGEMENT positions by one frame distance forward or aft, Fig. 2.
The tanks associated with a specific design variant were
The cargo tanks were generated within the FFW using represented as an assembly of planar surfaces within the
feature technology, e.g. [4]. The tanks are generated such FFW, Fig. 7, and transferred to NAPA by means of the
that maximum cargo volume is realized while ensuring a edge points for the bulkheads and hopper plates.
minimum distance to the hull form, e.g. 2 m. The feature
Figure 7: Family of parametrically generated hull forms for 6x2 cargo tank arrangements by use of the
FRIENDSHIP-Framework (FFW)
6. DISCUSSION OF RESULTS RFR) falls with rising deadweight (DWT) until a certain
minimum is reached. Cost of transport could be reduced
6.1 EXPLORATION by about 4%. The performance of the heaviest tankers is
slightly less attractive with regard to RFR but the tanker
In the course of the herein implemented optimisation with lowest EEDI is found among them. The best tanker
procedure approximately 2500 variants were generated with lowest OOI turns out to be among the smaller
and assessed. To start with, a Design-of-Experiment designs with a slight penalty in RFR of ~2%.
(DoE) for the exploration of the global design space was
performed, yielding a database with all relevant Normalization was done with the baseline's data to gain a
simulation outputs and the key measures of merit, certain independence from current price levels and their
namely RFR, OOI and EEDI. A conventional Aframax volatility. The RFRs were determined via a roundtrip
tanker served as a reference (baseline) for comparison model for the Caribbean trade on the basis of
and normalization, Table III. For identifying the contemporary cost levels. (Capital costs were based on a
attractive design alternatives with respect to the newbuilding price of 65 Million $, 25 years of lifetime
economic performance for instance, the design variants and an interest rate of 8%. Fuel costs were computed
were ranked according to RFR, Fig. 10. Naturally, any with HFO at 500 $/t and MGO at 800 $/t for the transfer
other preference of the decision makers can be within the ECA. Other operating costs were
considered and the two best designs for oil outflow OOI approximated with 3 Million $/year and presumed to be
and EEDI, respectively, are marked in Fig. 10, too. The independent of the variations.)
diagram shows that the cost of transport (normalized
Normalized RFR
1.050
best OOI
1.025
1.000
best RFR
reference
design
0.975
best EEDI
0.950
110,000 112,000 114,000 116,000 118,000
DWT
Figure 10: Designs established by means of the integrated CAE approach
6.2 REFINEMENTS the FFW without further interactive work, i.e. they are a
direct outcome from the optimizations.
Since a good number of generated designs exhibit nearly
the same RFR, see Fig.10, the variant with the best OOI
among them was selected for further refinements. A local
hydrodynamic optimization, utilizing a deterministic
search strategy, was undertaken for the aftbody, focusing
on the quality of the wake field as an objective. The
aftbody was allowed to change such that the impact on
the cargo tanks previously established in the global Figure 11: Hull form of favored design
optimization was negligible. The fine-tuning of the
hydrodynamics yielded a further increase in the
achievable speed, such that the tanker could be expected
to attain 15.6 kn at design draft and 16.8 kn at ballast Table III: Main particulars of reference & favored design
draft with a level of confidence of 1.3% V. The main
characteristics of this favored design are summarized in Parameter Reference design Favored design
Table III and compared to the reference design. The
associated hull form is presented in Fig.11. The lines Length over all 250 m 250 m
stem from the parametric model and were realized within Beam 44 m 44 m
Depth 21.0 m 21.5 m Speed at ballast draft 15.9 kn 16.8 kn
Design draft 13.7 m 13.7 m 3.541 g CO2 / 3.281 g CO2 /
EEDI (t nm) (t nm)
Block coefficient 0.83 0.85
Inner bottom height 6.3 SENSITIVITIES
COT 2-6 (S+P) 2.50 m 2.10 m
Inner bottom height Finally, in order to understand the robustness of the
COT 1 (S+P) 2.50 m 2.75 m established design with regard to small modifications a
Side shell width 2.50 m 2.65 m separate DoE (sensitivity study) was performed. About
150 additional variants were generated whose free
Angle of hopper plate 50 37 variables changed within 1% of the corresponding
Width of hopper plate 5.25 m 5.20 m parameters of the favored design. Fig.12 presents a
selection of sensitivities, with changes in RFR displayed
Frame spacing 3.780 m 4.400 m
in the upper row and changes in OOI and EEDI in the
Shift of bulkheads 0m 0m middle and lower row, respectively. The favored design
DWT 111 436 t 114 923 t can be regarded as a (local) optimum for RFR while in its
Maximum cargo vicinity only few variants perform slightly better with
volume 124 230 m3 129 644 m3 regard to OOI and EEDI. In general, the sensitivity of
parameters is quite limited. This indicates that the
OOI 0.0138 0.0142 favored design does not represent an extreme breed of
Speed at design draft 15.1 kn 15.6 kn parameters with respect to just one criterion.
Figure 12: Sensitivity of best RFR design (marked by red bullets, band width of abscissas 1%)
6.4 RFR-OOI STUDY the normalized RFR from 0.961 to 0.966, i.e. taking just
3.4% gains instead of 3.9% in comparison to the
The relationship between RFR and OOI was further reference tanker, leads to a further reduction of OOI from
investigated, again utilizing the integrated CAE approach. 0.015 to 0.012. In Fig.13 the design called best RFR is
The tank geometry was systematically varied within the highlighted. It is evident that this design is a good
bounds summarized in Table IV while freezing all other solution for both economic performance and
variables at the values of the best RFR design. Fig.13 environmental safety. Fig.14 offers a synthesized-artistic
opens a view on the compromise between economy impression of the resulting ship.
(ordinate) and safety (abscissa). The smaller the
accidental oil outflow the higher the cost of transport.
This is not unexpected but the diagram quantifies how
much an operator needs to pay for a safety margin
beyond the regulatory limit set by MARPOL. Relaxing
0.968
0.967
MARPOL limit
0.966
0.965
Normalized RFR
0.964
0.963
0.962
0.961
best RFR
0.960
0.959
0.0115 0.0125 0.0135 0.0145 0.0155 0.0165
Oil Outflow Index
Figure 13: Economics vs. safety in Aframax tanker Figure 14: CFD computed wave field plus cut-away
design showing the inner structure of proposed 6x2 design
Table IV: Free variables and their bounds for RFR-OOI study
Free variable Lower bound Upper bound Primary influence
Inner bottom height of cargo oil tanks 2 to 6 (S+P) 2.1 m 3.0 m Tank geometry
Lifting of inner bottom of cargo oil tank 1 (S+P) 0.2 m 2.0 m Tank geometry
Side shell width 2.1 m 3.0 m Tank geometry
Angle of hopper plate 30 60 Tank geometry
Width of hopper plate 4.0 m 6.0 m Tank geometry
7. CONCLUSIONS
An integrated design and multi-objective optimisation but it also reduces the complexity associated with CFD
approach to ship design was developed that covers in a analyses and, hence, allows to already utilize them early
holistic way a multitude of aspects of early ship design: in the process when gains are potentially the highest.
main dimensions, hull form, hydrodynamics and
powering; structures, strength and weight estimates; The presented example showed that once a (quasi-
safety, including intact and damage stability; economics; randomly created) database of variants is available it is
and regulatory requirements. An example application quick and easy to search for the preferred combination of
was presented for an Aframax tanker with the aim of measures of merit. One may then choose a more
realizing better environmental safety (lower OOI), conservative design, being a balanced all-rounder, or
efficiency (lower EEDI) and economics (lower RFR). deliberately decide to favor a more extreme solution,
Formal explorations and exploitations were combined to featuring excellent performance in one measure of merit.
investigate the design space and, subsequently, advance Additional investigations can be done easily once the
competing design proposals into certain directions. CAE environment is established, for instance to gain an
About 2500 variants were realized, each instance having appreciation of the relationship between costs and safety
its individual hull form (outer shell), tank or to check the robustness of the favored design.
compartmentation and an inner steel structural system.
Setting up an integrated approach still requires quite
The integrated system brings together sophisticated some effort at this point with respect to time.
software systems for analysis and simulation. Nevertheless, the necessary software platform is now
Challenging issues, like CFD simulations, can be available and the presented project proved feasibility.
implemented by systematic numerical series and suitable Major prerequisites are parametric models for various
meta-models (RSM). This not only speeds up the time ship types, which allow automation. Significant design
needed for investigations by several orders of magnitude improvement can then be realized even for moderate
deviations from currently established design practice.