KCS & KVLCC2 hulls: from fitting to defor­ma­tion and design exploration

Fitting Existing Hull Forms

The Com­po­nent-Based Ship Workflow in CAESES provides a struc­tured method for creating para­met­ric ship hull geome­tries for modern ship design and naval archi­tec­ture. This approach also enables the para­met­ric fitting of bench­mark ship hulls devel­oped by the Korea Research Insti­tute of Ships and Ocean Engi­neer­ing (KRISO), includ­ing the KCS (KRISO Con­tainer Ship) and KVLCC2 (KRISO Very Large Crude Carrier). These hulls provide high-quality datasets widely used for CFD val­i­da­tion, hydro­dy­namic analysis, and ship flow studies. Para­met­ric versions of these models are avail­able in the CAESES sample library for sim­u­la­tion-driven ship design and optimization.

Com­par­i­son Between Original and Para­met­ric Geometry

A com­par­i­son of the station curves between the original hull geometry (red) and the para­met­ric recon­struc­tion created with the Ship Mod­el­ling Workflow (green) demon­strates the accuracy of the fitting process while enabling para­met­ric control for further hull form explo­ration and optimization.

A com­par­i­son of the hydro­sta­t­ics between the original geometry and the para­met­ric ship modeling workflow (SMW) model is also performed.

Table 1: Hydro­sta­t­ics Com­par­i­son for KCS

Table 2: Hydro­sta­t­ics Com­par­i­son for KVLCC2

From KCS to KVLCC2

Both para­met­ric hull models use the same Ship Modeling Workflow (SMW) para­me­ter­i­za­tion in CAESES, including:

• Section-Based Aftship
• Water­line-Based Foreship
• Single Skeg
• Basic Bulb

Because both hulls are gen­er­ated using the same para­met­ric modeling frame­work, they belong to the same para­met­ric hull design family, even though they rep­re­sent dif­fer­ent ship types. The KCS is a con­tainer ship, while the KVLCC2 is a tanker, result­ing in sig­nif­i­cant dif­fer­ences in hull fullness, pro­por­tions and cargo capacity requirements.

Due to the shared para­me­ter­i­za­tion struc­ture, however, the SMW model enables a smooth geo­met­ric tran­si­tion between the two hull forms. By adjust­ing a single design variable, the geometry can con­tin­u­ously morph from the KCS hull form to the KVLCC2 hull form. This demon­strates the flex­i­bil­ity of para­met­ric ship hull modeling and shows how a unified para­me­ter­i­za­tion can rep­re­sent a wide range of ship hull designs within one para­met­ric framework.

In the ani­ma­tions, ship appendages includ­ing the rudder and pro­peller are also incor­po­rated to provide a more real­is­tic rep­re­sen­ta­tion of the propul­sion con­fig­u­ra­tion. The pro­peller geome­tries are based on designs created using the Advanced Pro­peller Workflow in CAESES and imported into the project as STEP files for sim­u­la­tion. The KCS uses the 5‑bladed KP505 pro­peller, while the KVLCC2 is equipped with the 4‑bladed KP458 pro­peller, both designed within the same para­met­ric workflow before inte­gra­tion into the model. The radial dis­tri­b­u­tions of the pro­peller design para­me­ters for both pro­pellers are shown below, high­light­ing the key geo­met­ric char­ac­ter­is­tics of the two pro­peller configurations.

Defor­ma­tion

Based on these ref­er­ence models, the iden­ti­fied para­me­ters of the Ship Modeling Workflow para­met­ric hull model were defined as the baseline geometry. This baseline served as the starting point for addi­tional targeted shape mod­i­fi­ca­tions, intro­duc­ing further geo­met­ric flex­i­bil­ity to analyze the hydro­dy­namic impact of local hull vari­a­tions. Two shape defor­ma­tion tech­niques were applied.

First, free-form defor­ma­tion was used on the KCS bulbous bow, enabling smooth and con­trolled bulb shape vari­a­tions while pre­serv­ing surface con­ti­nu­ity. This allows sys­tem­atic explo­ration of alter­na­tive bulbous bow designs without com­pro­mis­ing hull geometry quality. Second, RBF-based B‑Rep morphing was applied to the aftbody of the KVLCC2 hull, enabling precise and local­ized stern geometry mod­i­fi­ca­tions directly on the CAD model while main­tain­ing high geo­met­ric fidelity.

Together, these methods demon­strate how par­tially para­met­ric defor­ma­tion tech­niques can be applied to ship hull models to enable targeted design explo­ration and hydro­dy­namic opti­miza­tion within a sim­u­la­tion-driven design frame­work. The ani­ma­tions below illus­trate the influ­ence of the four design vari­ables used in each par­tially-para­met­ric model.

Sim­u­la­tions with SHIPFLOW

SHIPFLOW is a Com­pu­ta­tional Fluid Fynamics (CFD) software spe­cial­ized in the hydro­dy­namic analysis of ships and marine pro­pellers. It is widely used in naval archi­tec­ture for pre­dict­ing resis­tance, propul­sion per­for­mance, and flow char­ac­ter­is­tics around ship hulls. CAESES is well con­nected with SHIPFLOW, enabling seamless inte­gra­tion between para­met­ric geometry modeling and hydro­dy­namic sim­u­la­tion. Through this con­nec­tion, hull forms created and modified in CAESES can be directly trans­ferred to SHIPFLOW, allowing effi­cient design explo­ration, auto­mated sim­u­la­tions, and opti­miza­tion of ship performance.

KCS Setup – Resis­tance Simulation

The original KCS design point cor­re­sponds to a design draft of 10.8 m and a service speed of 24 knots, con­di­tions for which the bulbous bow geometry was opti­mized. This con­fig­u­ra­tion serves as the baseline hull design for eval­u­at­ing hydro­dy­namic per­for­mance. To assess per­for­mance across dif­fer­ent oper­at­ing con­di­tions, addi­tional sce­nar­ios were con­sid­ered: a reduced draft of 9.5 m rep­re­sent­ing lighter loading, slow steaming at 18 knots for energy-effi­cient oper­a­tion, and a high-speed con­di­tion of 26 knots.

Hydro­dy­namic sim­u­la­tions were per­formed using XPAN from SHIPFLOW, a poten­tial-flow solver widely used for ship resis­tance pre­dic­tion and hull-form eval­u­a­tion. A fine com­pu­ta­tional mesh was applied to capture the flow char­ac­ter­is­tics around the hull and bulbous bow region. The analysis focused on key resis­tance com­po­nents, includ­ing fric­tional resis­tance (Rf) and wave-making resis­tance (Rw), pro­vid­ing insight into the vessel’s hydro­dy­namic per­for­mance across oper­at­ing conditions.

KVLCC2 Setup – Self-Propul­sion Simulation

The KVLCC2 baseline con­fig­u­ra­tion cor­re­sponds to a design draft of 20.8 m and a service speed of 15.5 knots. The study focused on self-propul­sion sim­u­la­tions to evaluate the inter­ac­tion between the ship hull and pro­peller. Sim­u­la­tions were con­ducted using the XCHAP from SHIPFLOW, a RANS solver that combines pro­peller – hull inter­ac­tion modeling with propul­sion analysis. Cal­cu­la­tions were per­formed at model scale using a coarse mesh, enabling the eval­u­a­tion of multiple design variants with rea­son­able com­pu­ta­tional cost.

Propul­sion was modeled using a Wagenin­gen B‑series pro­peller, a widely used ref­er­ence pro­peller series in naval archi­tec­ture and hydro­dy­namic studies. The eval­u­a­tion focused on key propul­sion metrics, includ­ing total resis­tance (RS), power deliv­ered (PD), and total propul­sive effi­ciency (ηDS). These para­me­ters provide a com­pre­hen­sive assess­ment of propul­sion per­for­mance and energy efficiency.

Design of Exper­i­ments (DoE)

For each par­tially-para­met­ric model, four design vari­ables con­trolled the geo­met­ric mod­i­fi­ca­tions. The Sobol sequence sampling method was used for design space explo­ration, pro­vid­ing well dis­trib­uted coverage of the para­me­ter space. Fol­low­ing common DoE guide­lines, typ­i­cally 5 to 10 samples per design variable, a total of 30 design variants were gen­er­ated for each study. This sampling strategy provides a suf­fi­ciently rich and well-dis­trib­uted dataset for the devel­op­ment of sur­ro­gate models, enabling the training of reliable machine learning models to approx­i­mate the under­ly­ing sim­u­la­tion responses.

KCS Resis­tance (DoE)

Resis­tance sim­u­la­tions were per­formed at 18 knots and 26 knots to evaluate hull per­for­mance under dif­fer­ent oper­at­ing con­di­tions. For each of the 30 sampled designs, flow-field visu­al­iza­tions were analyzed to under­stand the hydro­dy­namic effects of geo­met­ric vari­a­tions. The ani­ma­tions illus­trate pressure dis­tri­b­u­tion, stream­lines, and wave ele­va­tion, pro­vid­ing insight into how bulb geometry mod­i­fi­ca­tions influ­ence flow behavior and ship resis­tance across the two oper­at­ing scenarios.

During post-pro­cess­ing, cor­re­la­tion and regres­sion analyses in CAESES were used to evaluate how the design vari­ables influ­ence hydro­dy­namic per­for­mance. These analyses help identify para­me­ter sen­si­tiv­i­ties and trends, reveal­ing how each variable affects ship resis­tance within the explored design space.

A key obser­va­tion emerges when com­par­ing results at the two oper­at­ing speeds. Changing speed sig­nif­i­cantly alters the influ­ence of the design vari­ables on wave-making resis­tance, showing that para­me­ter impor­tance varies across oper­at­ing con­di­tions. This effect is also visible in the ani­ma­tions. At 18 knots, the results favor a shorter and more slender bulbous bow, which helps reduce wave resis­tance during slow steaming. At 26 knots, the design explo­ration leads to a longer and fuller bulb, better suited for higher-speed oper­a­tion and dif­fer­ent wave patterns.

KVLCC2 Self-Propul­sion (DoE)

Sobol sampling was used to generate 30 design variants based on four design vari­ables, ensuring a well-dis­trib­uted design space explo­ration. The ani­ma­tions illus­trate pressure dis­tri­b­u­tion, stream­lines, and nominal wake for the dif­fer­ent designs. These visu­al­iza­tions help analyze how stern geometry vari­a­tions influ­ence hull flow, pro­peller inflow and overall propul­sion per­for­mance, pro­vid­ing insight into the vessel’s hydro­dy­namic and propul­sive efficiency.

Cor­re­la­tion and regres­sion analyses were carried out during the post-pro­cess­ing stage using CAESES, allowing the influ­ence of the design vari­ables on the hydro­dy­namic per­for­mance to be sys­tem­at­i­cally eval­u­ated. The analysis indi­cates that a wider skeg in the lower region tends to increase the overall resis­tance of the vessel. However, at the same time, an increase in total propul­sive effi­ciency (ηD) is observed. In several cases, the improve­ment in propul­sive effi­ciency is larger than the cor­re­spond­ing increase in resis­tance, result­ing in an overall reduc­tion in the required propul­sion power.

This behavior is partly related to the char­ac­ter­is­tics of the ITTC extrap­o­la­tion method, which is commonly used for resis­tance and propul­sion analysis at model scale. While the method provides con­sis­tent com­par­a­tive trends for design eval­u­a­tion, the exact balance between resis­tance and propul­sion effi­ciency improve­ments may not fully rep­re­sent the behavior at full-scale oper­at­ing conditions.

Con­clu­sion

This study demon­strates the capa­bil­i­ties of fully para­met­ric modeling using the Ship Modeling Workflow in CAESES, suc­cess­fully repro­duc­ing existing hull forms such as KCS and KVLCC2. Par­tially-para­met­ric modeling was also applied, high­light­ing the robust­ness and effi­ciency of para­met­ric approaches.

A seamless inte­gra­tion between CAESES and SHIPFLOW enabled an auto­mated sim­u­la­tion-driven design frame­work. Resis­tance sim­u­la­tions were per­formed using the XPAN solver, while self-propul­sion analyses were con­ducted with XCHAP.

The results high­light the power of sim­u­la­tion-driven design, where numer­i­cal sim­u­la­tions combined with para­met­ric modeling and struc­tured design explo­ration reveal complex rela­tion­ships between hull geometry and hydro­dy­namic per­for­mance. This approach enables design­ers to identify optimal design trends and make informed deci­sions early in the design process.

The study also empha­sizes the increas­ing impor­tance of data-driven engi­neer­ing. Managing and ana­lyz­ing large volumes of sim­u­la­tion data allows mean­ing­ful insights to be extracted, sup­port­ing smarter design processes and more effi­cient devel­op­ment of high-per­for­mance ships. The inte­gra­tion of Arti­fi­cial Intel­li­gence (AI) and machine learning tech­niques further enhances this process by enabling the devel­op­ment of pre­dic­tive models that can approx­i­mate sim­u­la­tion results, identify complex patterns within the design space, and support faster design explo­ration and optimization.

Overall, the com­bi­na­tion of para­met­ric modeling, sim­u­la­tion-driven design, and data-driven engi­neer­ing provides a powerful and struc­tured workflow for modern ship design, enabling effi­cient pro­gres­sion from concept devel­op­ment to opti­mized hull performance.