Naval Surface Warfare Center

Carderock Division
NSWCCD-CISD-2012/004 / July 2012 / Hull Form Design and Optimization Tool Development

West Bethesda, MD 20817- 5700

NSWCCD-CISD2012/004 July 2012
Center for Innovation in Ship Design
Summer Intern Report

Hull Form Design and Optimization

Tool Development
By:
Lacy Thomas
Brendan White

Approved for Public Release: Distribution is unlimited

Form Approved
OMB No. 0704-0188

REPORT DOCUMENTATION PAGE

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data
sources, gathering and maintaining the data needed, and completing and reviewing this collection of information. Send comments regarding this burden estimate or any other
aspect of this collection of information, including suggestions for reducing this burden to Department of Defense, Washington Headquarters Services, Directorate for Information
Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other
provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. PLEASE
DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS.

1. REPORT DATE
September 7 2012

2. REPORT TYPE
Final

3. DATES COVERED (From-To)

May 2012 August 2012

4. TITLE AND SUBTITLE

Hull Form Design and Optimization Tool Development

5a. CONTRACT NUMBER

5b. GRANT NUMBER
5c. PROGRAM ELEMENT NUMBER

6. AUTHOR(S)
Brendan White
Lacy Thomas

5d. PROJECT NUMBER

5e. TASK NUMBER
5f. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) AND ADDRESS(ES)

Naval Surface Warfare Center
Carderock Division
9500 Macarthur Boulevard
West Bethesda, MD 20817-5700

8. PERFORMING ORGANIZATION
REPORT NUMBER
NSWCCD-CISD-2012/004

10. SPONSOR/MONITORS
ACRONYM(S)

9. SPONSORING / MONITORING AGENCY NAME(S) AND ADDRESS(ES)

Chief of Naval Research
One Liberty Center
875 North Randolph Street, Suite
1425
Arlington, VA 22203-1995

11. SPONSOR/MONITORS REPORT

NUMBER(S)

12. DISTRIBUTION / AVAILABILITY STATEMENT

Approved for Public Release: Distribution Unlimited
13. SUPPLEMENTARY NOTES
14. ABSTRACT
George Mason Universitys (GMU) Center for Computational Fluid Dynamics (CCFD) is working with the Naval Surface Warfare Center's
(NSWC) Center for Innovation in Ship Design (CISD) under a five year grant from the Office of Naval Research (ONR) aimed at developing and
validating hull form design and hydrodynamic optimization tools. These tools will be used for early stage ship design and education of junior
naval engineers. The categories of the tools can be further broken down into tools for hull form generation, representation and modification,
design-oriented simple Computational Fluid Dynamic (CFD) tools, and hull form optimization tools. This paper will discuss the further
development of these tools including the implementation and testing of a new optimization algorithm, the improvement of a rapid hull form
generation tool (HulGen), and the creation of a user interface for a SWATH resistance and powering prediction program (SWAD). The work was
performed under the 2012 Naval Research Enterprise Intern Program.
15. SUBJECT TERMS
hydrodynamic, hull form,

generation, optimization, algorithm

16. SECURITY CLASSIFICATION OF:

a. REPORT
UNCLASSIFIED

b. ABSTRACT
UNCLASSIFIED

c. THIS PAGE
UNCLASSIFIED

17. LIMITATION
OF ABSTRACT
SAR

18. NUMBER
OF PAGES
34

19a. NAME OF RESPONSIBLE

PERSON Colen Kennell
19b. TELEPHONE NUMBER
301-227-5468

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Team
Brendan White

Lacy Thomas

Acknowledgements
Chi Yang - GMU
Max Harper - CISD
Bob Lamb - CISD
Colen Kennell - CISD
Jack Offutt CISD
Steven OuimetteCISD
Siu Fung NSWCCD
Matthew Knobloch - SEAP
Center for Computational Fluid Dynamics GMU

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Table of Contents
Team ................................................................................................................................................ i
Acknowledgements .......................................................................................................................... i
Table of Contents ............................................................................................................................ ii
List of Figures ................................................................................................................................ iv
List of Tables .................................................................................................................................. v
Abbreviations ................................................................................................................................. vi
Abstract ........................................................................................................................................... 1
Introduction ..................................................................................................................................... 1
Background and Objectives ........................................................................................................ 1
Deliverables ................................................................................................................................. 1
Hull Optimization Tool Development ..................................................................................... 2
Hullform Generation Tool Development ................................................................................ 2
Resistance and Powering Prediction Tool Development ........................................................ 2
Hull Optimization Tool Development ............................................................................................ 2
Background ................................................................................................................................. 2
Objective ..................................................................................................................................... 3
Algorithm Validation .................................................................................................................. 3
Greiwank Function .................................................................................................................. 4
Algorithm Implementation .......................................................................................................... 6
Single Speed Input Files .......................................................................................................... 7
Multiple Speed Input Files ...................................................................................................... 8
Optimization Tool Validation ..................................................................................................... 8
ii

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Strengths of the Optimization Tool ........................................................................................... 13

Run Time ............................................................................................................................... 13
Alternate Applications ........................................................................................................... 14
Summary ................................................................................................................................... 14
Hullform Generation Tool Development ...................................................................................... 15
Background and Objectives ...................................................................................................... 15
Existing HulGen Tool ............................................................................................................... 16
Input ....................................................................................................................................... 16
Output .................................................................................................................................... 18
Reduction of Mandatory Input Parameters ............................................................................... 18
Summary ................................................................................................................................... 20
Resistance and Powering Prediction Tool Development .............................................................. 21
Background and Objectives ...................................................................................................... 21
SWAD12 Input .......................................................................................................................... 22
SWAD12 Output ....................................................................................................................... 23
Hull Representation ................................................................................................................... 24
Graphic Feedback ...................................................................................................................... 24
Summary ................................................................................................................................... 26
Recommendations for Further Work ............................................................................................ 26

iii

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

List of Figures
Figure 1: TriSWACH Model .......................................................................................................... 3
Figure 2: Rosenbrock's Saddle Function ........................................................................................ 4
Figure 3: Greiwank Function .......................................................................................................... 4
Figure 4: Contour of Greiwank Function........................................................................................ 5
Figure 5: Greiwank Convergence History ...................................................................................... 6
Figure 6: Sidehull Positions for TriSWACH .................................................................................. 7
Figure 7: Contour with 1,200 Points ............................................................................................... 9
Figure 8: Drag Convergence Using Differential Evolution .......................................................... 10
Figure 9: DE Optimal Sidehull Location Plotted on Contour Plot for Froude Number .31 ......... 12
Figure 10: Wave Drag at Optimized Positions ............................................................................. 13
Figure 11: SWATH ship ............................................................................................................... 14
Figure 12: HulGen Main Interface ................................................................................................ 15
Figure 13: Baseline Hull Interface ................................................................................................ 19
Figure 14: Baseline Ship Type Selection Menu ........................................................................... 20
Figure 15: SWAD12 Main Interface............................................................................................. 21
Figure 16: SWAD12 Input Interface............................................................................................. 22
Figure 17: SWAD12 Import and Export Menu ............................................................................ 23
Figure 18: SWAD12 Output Interface .......................................................................................... 24
Figure 19: SWAD12 Coefficients of Resistance Curves .............................................................. 25
Figure 20: SWAD12 Graph of Hull .............................................................................................. 25

iv

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

List of Tables
Table 1: DE Results for Optimal Sidehull Location ..................................................................... 11
Table 2: Comparison of Optimal Sidehull Locations from Contour Plot and DE ........................ 11
Table 3: HulGen Parameters ......................................................................................................... 16
Table 4: HulGen Parameter Categories ........................................................................................ 18
Table 5: HulGen Main Parameters ............................................................................................... 19

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Abbreviations
CISD Center for Innovation in Ship Design
CFD Computational Fluid Dynamics
DE Differential Evolution
DWL Design Waterline
EHP Effective Horsepower
GMU George Mason University
NREIP Naval Research Enterprise Intern Program
NSWCCD Naval Surface Warfare Center Carderock Division
NURBS Non Uniform Rational B-Spline
ONR Office of Naval Research
SSF Steady Ship Flow
SWATH Small Waterplane Area Twin Hull
TriSWACH Trimaran Small Waterplane Area Center Hull
UI User Interface
UNREP Underway Replenishment

vi

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Abstract
George Mason Universitys (GMU) Center for Computational Fluid Dynamics (CCFD) is
working with the Naval Surface Warfare Center's (NSWC) Center for Innovation in Ship Design
(CISD) under a five year grant from the Office of Naval Research (ONR) aimed at developing
and validating hull form design and hydrodynamic optimization tools. These tools will be used
for early stage ship design and education of junior naval engineers. The categories of the tools
can be further broken down into tools for hull form generation, representation and modification,
design-oriented simple Computational Fluid Dynamic (CFD) tools, and hull form optimization
tools. This paper will discuss the further development of these tools including the implementation
and testing of a new optimization algorithm, the improvement of a rapid hull form generation
tool (HulGen), and the creation of a user interface for a SWATH resistance and powering
prediction program (SWAD). The work was performed under the 2012 Naval Research
Enterprise Intern Program.

Introduction
Background and Objectives
Hydrodynamic design of ships involves several stages, from early-stage and preliminary design
to final design. It has become increasingly important to evaluate multiple hydrodynamic aspects
of performance simultaneously and efficiently during the early stage of the design process.
The Center for Computational Fluid Dynamics (CFD) at George Mason University has an
ongoing research project sponsored by the Office of Naval Research (ONR) on the development
of hull form design and hydrodynamic optimization tools. These tools are comprised of three
main components. One component consists of design-oriented CFD tools. The second consists of
hull form generation, representation, and modification tools. The final component consists of
optimization tools.
The 2012 Naval Research Enterprise Intern Program (NREIP) intern team was tasked with the
further development of the tools. As part of this task the team was to implement and validate an
optimization algorithm, improve a rapid hull form generation tool called HulGen, and create a
user interface for a SWATH ship total resistance and powering prediction tool, called SWAD.

Deliverables
The summer work consists of three individual sections. Each of these sections has its individual
set of associated deliverables.
1

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Hull Optimization Tool Development

DE input files for analytical functions

Specified range wave drag calculation tool
Contour plotting tool
DE tool for optimizing a TriSWACH
DE optimization plotting tool
Basic user guide for contour and DE tools

Hullform Generation Tool Development

Reduction of required input parameters for HulGen

Resistance and Powering Prediction Tool Development

New user interface for SWAD

Graphic feedback and accessible output data from SWAD

Hull Optimization Tool Development

Background
One of the objectives for the summer was to create and validate a hull optimization tool using an
existing optimization algorithm. The intent of the optimization tool was to find the optimal
position for the sidehull of a TriSWACH, shown in Figure 1, to minimize the wave drag. Scilab,
open source software, was selected to create this optimization tool. This software is similar to
and has many of the same capabilities as the more widely used Matlab, which is not open source.
The algorithm to be used for the optimization process was the Differential Evolution (DE)
algorithm. This Differential Evolution algorithm is available as open source software. It was
developed by Kenneth Price and Rainer Storn from the University of California at Berkley and is
available for many different programming environments including a version which can be run in
Scilab.
The DE algorithm is used to find the location of the global minimum of a function, not the value
of the global minimum. The algorithm accomplishes this by using a method known as
metaheuristics which allows the algorithm to examine a large area by trying to iteratively
improve an original guess as to where the minimum is located. These iterations are carried out
until an output value less than or equal to the user specified minimum is obtained. After the
initial solution attempt is generated from bounds specified by the user, the algorithm generates
possible solutions by combining previous solution attempts according to some formula within the
algorithm and keeping track of the solution attempt with the lowest value.
2

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Objective
The goal was to integrate the DE algorithm into a hull
form optimization tool used to find the location of the
sidehull of a TriSWACH with minimum wave drag. A
picture of the TriSWACH model can be seen in Figure
1. The algorithm was to be validated before it would be
accepted for use in an optimization tool. This was to be
done with a variety of analytical functions. Upon the
algorithms validation, a new optimization tool was to
Figure 1: TriSWACH Model
be created by integrating the DE algorithm with two
already existing hydrodynamic optimization programs developed at George Mason University.
These programs include a small geometry generation program and the Steady Ship Flow (SSF)
program used for calculating wave drag. Upon completion of this tool, a separate tool was
developed to validate the results.

Algorithm Validation
To validate the DE algorithm, input files were written for nine different analytical functions
including Rosenbrocks Saddle Function, shown in Figure 2 and the Greiwank Function shown
in Figure 3. Two input files are required for each function. One input file contains a computer
coded function representing the mathematical function. The other input file contains the
parameters for running the function code with the algorithm. These parameters include the
number of inputs (dimensions), the minimum value the algorithm is trying to obtain, minimum
bounds for the points, maximum bounds for the point, and a maximum number of iterations. It
should be noted that the minimum and maximum bounds are not constraints, but rather an initial
guess as to where the minimum may be found. This gives the algorithm bounds to generate an
initial guess to find the minimum. The algorithm will still search outside these bounds if
necessary. If the user wishes to put actual limitations on where the algorithm can search, as seen
in Rosenbrocks Saddle Function in Figure 2, constraints can be added to the function file which
can force the algorithm to remain strictly within the given bounds.

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Figure 2: Rosenbrock's Saddle Function

Only functions for which the global minimum was already known were used for validation. The
algorithm searches for the global minimum given by the user and will stop when either a value
less than or equal to the specified minimum is found or when the algorithm has reached the
maximum number of iterations. The algorithm was able to find the global minimum of each
tested function.
Greiwank Function
When testing the DE algorithm, one function was used more than others. This is the Greiwank
Function. It has some special properties that made it a beneficial test function for the algorithm.
Figure 3 shows the general Greiwank Function. For testing purposes, an n value of two was
used.

Figure 3: Greiwank Function

The Greiwank Function has many local minima, which makes the function an interesting test
case for it is able to specifically test the algorithms ability to find the global minimum without
getting stuck in any local minima. Figure 4 shows a contour plot of the function. The local
minima are shown by the blue circles tending towards the top right corner where the global
minimum is located and plotted on the contour.

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Global Minimum

Local Minima

Figure 4: Contour of Greiwank Function

For testing purposes, the program was set to find zero with a precision of six decimal places.
This was to ensure a fairly accurate result with a reasonable number of iterations. The program
generates a random number between the given bounds for the first attempt at finding the
minimum. This causes the DE algorithm to require a different number of iterations for every run
because it generates a different path for finding the minimum each time. The algorithm was able
to find the minimum very quickly, which is a feature beneficial for the future implementation of
the algorithm as a hydrodynamic optimization tool. A sample test result is shown in Figure 5,
which presents a convergence history of the function and the location with reference to the
number of iterations.

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

F(X1,X2)

X1

X2

Figure 5: Greiwank Convergence History

The algorithm was successful in finding the global minimum each time it was run. The results
provided evidence that the algorithm would be valid for the optimization tool.

Algorithm Implementation
After validating the DE algorithm, the next step was to implement the algorithm as a
hydrodynamic optimization tool. Upon completion of this tool, it was to be applied to a
TriSWACH model to investigate the effects of longitudinal and transverse spacing of the
sidehull on the wave drag of the model, and to find the optimal position of the sidehull for low
wave drag. The sidehull spacing definition is shown in Figure 6 which defines the three different
longitudinal positions and three different transverse positions considered in the experiment.
6

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Figure 6: Sidehull Positions for TriSWACH

Single Speed Input Files

To use the DE algorithm for the TriSWACH, the wave drag at each sidehull position had to be
written as a function. Two programs from George Mason University (GMU) were used to help
accomplish this task. The first was a small geometry generation program that created a geometry
file for a TriSWACH given a specific sidehull location. This geometry generation program uses
an input file in which the transverse and longitudinal location of the sidehull can be changed and
generates new TriSWACH geometry with the given sidehull position. The programs from GMU
normalize the ship to a length of one. Since the TriSWACH used for the test was 84 feet long, it
was necessary to divide the considered locations by 84 to obtain the proper format used by the
GMU programs. This conversion issue was considered when writing the input file and a variable
for the length of the ship was included to allow the program to work with different ship lengths.
The geometry generation program produces an output file containing the new geometry which is
used by the second GMU program, the Steady Ship Flow (SSF) program, which calculates the
wave drag for the given geometry. The drag is calculated at a given speed or Froude Number
specified by the user in the SSF programs input data file.
7

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

The first input file for the optimization tool contains a coded representation of a function with the
return value being the value the user is trying to minimize. The tool was being used to minimize
drag, thus code had to be written to implement a function with drag as the return value. This was
done by coding the wave drag as a function of the longitudinal position and the transverse
position of the sidehull. The tool generates longitudinal and transverse position values to be used
as input values for this coded function. This code then writes an input file for the geometry
generation program with the new position coordinates. The geometry generation program is
executed and an output file containing the new ships geometry is generated. Upon generation of
this output file, SSF is called and a wave drag output file is produced. The code then reads this
new output file and adds the current drag values to an array containing all drag values from
previous iterations. The current drag value is the coded functions return value.
The basic information contained in the optimization tools parameter file, the second input file, is
similar to the information used in the parameter files for the analytical functions, with the only
real difference being the minimum value. With the analytical functions, the minimum value is
known, so that value can be input and the algorithm will stop once that value is reached. The
minimum wave drag is unknown, but there is a simple way to work around this problem. The
optimal position of the sidehull is the position with the least amount of drag. Therefore the
minimum value is set to zero. Knowing there was not going to be a location where the drag will
equal zero, the intent is to have the algorithm find the closest value to zero within the maximum
number of iterations.
Multiple Speed Input Files
Once the input files were created and results were obtained using a single speed, the first input
file was changed slightly to accommodate multiple speeds. The optimization tool can find the
optimal position when a ship is traveling at a specific speed, but it is not always the case that
only one ship speed is of interest. This creates the need to incorporate multiple speeds into the
tool. The relative importance of each speed is implemented with different weights. If the user
wants the optimal position for a ship that runs at one speed seventy-five percent of the time and
another speed twenty-five percent of the time, the weights of the numbers can be adjusted. For
the summer work, the weights were assumed to be equal across all speeds. This required only a
simple average calculation. The additional speed values must also be added to the SSFs input
data file before the average calculations can be done.

Optimization Tool Validation

Before the optimization tools results could be accepted, the results had to be validated. This was
done by creating a separate contour tool in Scilab to draw contour plots for wave drag and
comparing the minimum found in the contour plot to the minimum found with the DE algorithm.
8

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

The longitudinal and transverse points are not generated for the contour tool in the same way
they are generated when using the optimization tool. The contour tool has to generate its points
autonomously rather than having the points generated by the DE algorithm. The points are
generated by using a minimum and maximum variable in both the longitudinal and transverse
directions. The user then specifies the number of points desired in each direction and the tool
evenly divides the given dimensions into the correct number of points. The tool iterates through
the points calculating drag at each possible combination of longitudinal and transverse points and
stores these values in an array similar to how the values are stored in the DE implementation.
This array is then stored in a text file. There is a separate plotting tool written for the actual
plotting to allow the user to plot existing values without having to run the entire contour tool.
The plotting tool simply opens the text file where the output is stored and creates a contour plot
with the longitudinal points on the x-axis and the transverse points on the y-axis. The contour
plot has colors ranging from red for the higher values to dark blue and black for the lowest
values. By using the contour plots, the results from the DE algorithm can be validated. The
contour plots provide a visual from which the minimum value can easily be estimated. This value
can be compared to that which was obtained using the DE algorithm.
Several cases were tested and the contour plot results were compared against the results from the
DE algorithm. In Figure 7 the minimum is represented by the circle towards the middle of the
contour plot. This contour plot was created using sixty longitudinal points and twenty transverse
points for a total of 1,200 points. A Froude Number of .31 was used for these calculations.

Global Minimum

Figure 7: Contour with 1,200 Points

9

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Once the contour plot was created, the results were compared with those from the optimization
tool. The results from the tool were consistently close to the minimum after 30 iterations. Trials
were run using up to 60 iterations, but it was found that accurate results were achieved by 30
iterations as demonstrated by Figure 8. To obtain accurate results in a reasonable amount of time,
the optimization tool was configured with a maximum iteration limit of 30.

Figure 8: Drag Convergence Using Differential Evolution

10

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Table 1 presents the optimal sidehull positions obtained using the optimization tool for Fn =
0.31, 0.25, and the average Fn, respectively.
As shown in the second row of Table 1, the optimization tools results for a Froude Number of
.31 agree with the results from the contour plot. This demonstrates that the tool was successful in
finding the optimal position.

Table 1: DE Results for Optimal Sidehull Location

Case
Fn = 0.25
Fn = 0.31
Average Fn

Longitudinal Position
0.580 (48.7 ft)
0.411 (34.5 ft)
0.943 (79.2 ft)

Transverse Position
0.052 (4.4 ft)
0.114 (9.6 ft)
0.052 (4.4 ft)

Table 2 compares the contour plot results to the DE algorithm results for each of the speeds to
further validate the optimization tool.

Table 2: Comparison of Optimal Sidehull Locations from Contour Plot and DE

Longitudinal Position
Fn = 0.25
Fn = 0.31
Average Fn
Transverse Position
Fn = 0.25
Fn = 0.31
Average Fn

Contour Plot Range

0.56 - 0.61
0.40 - 0.45
0.92 - 0.97
Contour Plot Range
0.05 - 0.09
0.09 - 0.13
0.05 - 0.07

11

Differential Evolution
0.580
0.411
0.943
Differential Evolution
0.052
0.114
0.052

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Figure 9 shows the last iteration of the DE algorithm plotted on the contour plot when a Froude
Number of .31 was used. This provides a visual form to show that the results agree. The point
representing the last iteration from the DE algorithm is located in the global minimum circle on
the contour plot.

DE

Figure 9: DE Optimal Sidehull Location Plotted on Contour Plot for Froude Number .31

After finding the optimal positions at different speeds using the optimization tool, those positions
were examined over a range of speeds. Using the previously mentioned SSF program from
GMU, the wave drag for three optimal positions was calculated over a wide range of speeds.
Fifty Froude Numbers between 0.18 and 0.48 were used for these calculations. Figure 10 shows
the calculated wave drag for the geometry created when the sidehull is in the optimized location
determined by the Froude Numbers shown in the legend.

12

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

It can be observed from Figure 10 that the optimization performed for a single design speed will
result in a large drag reduction at the given speed, but might have a drag increase at off-design
speeds. It would be beneficial to consider multi-design-speed optimization.

6.00E-03
Optimized with Fn = .31
5.00E-03

Optimized with Fn = .25

Optimized with Average Fn

Drag

4.00E-03

3.00E-03

2.00E-03

1.00E-03

0.00E+00
0.18

0.21

0.24

0.27

0.3

0.33

0.36

0.39

0.42

0.45

0.48

Froude
Figure 10: Wave Drag at Optimized Positions

Strengths of the Optimization Tool

Run Time
There are multiple benefits to using the algorithm to find the minimum rather than the contour
plots. The optimization tool results provide a specific number as opposed to a range. The run
time for the algorithm is also significantly shorter than the runtime of the contour plots.
Computing the previous contour plot requires approximately twelve hours on a laptop with a
dual-core processor, but the DE optimization tool with thirty iterations requires only three hours
on the same hardware. This represents a significant time difference even with the small amount
of data that was used. As the size of the data increases, the run time benefits of using the DE
optimization tool will increase.
13

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Alternate Applications
After use of the DE algorithm for optimizing the
TriSWACH had been validated, other uses for the
algorithm were considered. The DE algorithm is a
general optimization algorithm and therefore it
can be implemented with a variety of different
programs. Again attempting to use the algorithm
to optimize drag, the algorithm was applied to
SWAD90, which is a power and resistance
Figure 11: SWATH ship
prediction tool for SWATH (Small Waterplane
Area Twin Hull) ships. An example of a SWATH ship is shown in Figure 11. The input files
required to use the algorithm with SWAD90 are similar to those for the TriSWACH
optimization. When attempting optimization for the TriSWACH, the position of the sidehull was
variable in the equation. When optimizing with SWAD90, the variable to optimize was
unknown. The first attempts to optimize with SWAD90 involved trying to optimize the nose and
midbody lengths of the strut to minimize drag, but after several trial runs the algorithm seemed
to have trouble finding a minimum. The results from both the algorithm and the contour plots
were inconclusive. This was interpreted to mean that the drag value predicted by SWAD90 is not
sensitive to the nose and midbody lengths of the strut. Due to time constraints, experimenting
with other variables was left as a future project.

Summary
The Differential Evolution algorithm was implemented into a new hydrodynamic optimization
tool. The algorithm was thoroughly tested with several different analytical functions before the
development of the optimization tool began. Once the DE algorithm was validated, the
optimization tool was implemented. A contour tool was developed to create contour plots to
show the wave drag of the sidehull of a TriSWACH in various locations. The contour tool
calculates wave drag at a specified number of points between given bounds in the transverse and
longitudinal directions. It then creates a contour plot of these points. The contour plots were used
to find the optimal position of the sidehull of a TriSWACH in terms of the wave drag. The plots
created by this tool were used to help validate the results of the new optimization tool by
comparing the minimum found by the optimization tool to the minimum in the contour plots.
Finally, the optimization tool was integrated with SWAD90 to explore the tools versatility and
confirm that the tool could be used for other applications.

14

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Hullform Generation Tool Development

Background and Objectives
During the early stages of the ship design process, it is important to be able to evaluate multiple
hydrodynamic aspects of performance simultaneously and efficiently. The Center for
Computational Fluid Dynamics at George Mason University has an ongoing research project
sponsored by the Office of Naval Research on the development of a methodology for innovative
hydrodynamic ship hull design.
Rapid ship hullform generation tools are useful in early stage ship design and are a precursor to
hydrodynamic performance analysis and design optimization. HulGen is a rapid hull generation
tool that generates a hull from scratch using user input hull form parameters. The latest version
of HulGen was developed at CISD in 2011. The function of HulGen is to facilitate the generation
and modification of monohull forms during early stage ship design.
A new feature has been implemented in HulGen to allow the user to generate a hull form using
fewer parameters than was previously required. Use of the full array of HulGen parameters
makes it suitable for integration with the design and optimization tools being developed at GMU.
The option to use fewer parameters makes HulGen more suitable for use by a human user.

Figure 12: HulGen Main Interface

15

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Existing HulGen Tool

Input
HulGen is implemented in Visual Basic to provide a user friendly interface. The main interface is
shown in Figure 12. The hull lines are polynomials defined by hull form parameters input by the
user. HulGen uses a total of 68 parameters to generate a hull. The list of parameter values used in
HulGen can be seen in Table 3. These parameters are used to create the curves in Table 4 which
are then used to create the body plan of the hull.

Table 3: HulGen Parameters

Parameter #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26

Parameter name
Number of Stations
Ship Length (LBP)
Ship Beam
Ship Draft
Maximum Section Coefficient (Cx)
Number of Segments
Area at FP (SAC)
Slope of Curve at FP (SAC)
Location of Desired Station of Maximum Area (Xmax) (SAC)
Slope of Curve at Xmax (SAC)
Length of Parallel Sections (Lmid) (SAC)
Area at AP (Ctran)
Slope of Curve at AP
Prismatic Coefficient (Cp)
Longitudinal Center of Buoyancy
Offset at FP (LWC)
Slope of Curve at FP (LWC)
Location of Desired Station of Maximum Area (Xmax) (LWC)
Slope of Curve at Xmax (LWC)
Length of Parallel Sections (Lmid) (LWC)
Offset at AP (LWC)
Slope of Curve at AP
Waterplane Coefficient (Cwp)
Longitudinal Center of Floatation (LCF)
Offset at FP (FDEC)
Slope of Curve at FP (FDEC)

16

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64

Location of Flat Middle Section (Xmid)

Length of Flat Middle Sections (Lmid)
Offset at Forward End of Flat Middle Section (Ymid)
Slope of Curve at Forward End of Flat Middle Section (Smid)
Offset at AP
Slope of Curve at AP
Depth at FP
Depth at Midships
Depth at AP
Longitudinal Position of Keel Rise
Section Area Coefficient at AP
Slope Value at FP (Keel)
Slope Rate at FP (Keel)
Location of Flat Middle Section (Xmid) (Keel)
Length of Flat Middle Sections (Lmid) (Keel)
Slope Value at Forward End of Flat Middle Sections (Keel)
Slope Rate at Forward End of Flat Middle Sections (Keel)
Slope Value at AP (Keel)
Slope Rate at AP (Keel)
Slope Value at FP (DWL)
Slope Rate at FP (DWL)
Location of Flat Middle Section (Xmid) (DWL)
Length of Flat Middle Sections (Lmid) (DWL)
Slope Value at Forward End of Flat Middle Sections (DWL)
Slope Rate at Forward End of Flat Middle Sections (DWL)
Slope Value at AP (DWL)
Slope Rate at AP (DWL)
Slope Value at FP (Deck Edge)
Slope Rate at FP (Deck Edge)
Location of Flat Middle Section (Xmid) (Deck Edge)
Length of Flat Middle Sections (Lmid) (Deck Edge)
Slope Value at Forward End of Flat Middle Sections (Deck Edge)
Slope Rate at Forward End of Flat Middle Sections (Deck Edge)
Slope Value at AP (Deck Edge)
Slope Rate at AP (Deck Edge)
Longitudinal Location of the Start of FBC (X1)
Longitudinal Location of the End of Full Section (X2)
Longitudinal Location of the Start of Full Section (X3)

17

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

65
66
67
68

Longitudinal Location of the End of FBC (X4)

Slope Value at the Start of FBC (S1)
Slope Value at the End of FBC (S4)
Half Siding Coefficient

Table 4: HulGen Parameter Categories

Category Name
Basic Parameters
Section Area Curve
Load Waterline Curve
Deck Edge Curve
Sheer Profile
Keel Rise Curve
Slope Curve (Keel)
Slope Curve (DWL)
Slope Curve (Deck Edge)
Flat of Bottom Curve

Parameter #s
1 to 6
7 to 15
16 to 24
25 to 32
33 to 35
36 to 37
38 to 45
46 to 53
54 to 61
62 to 68

Each curve is defined by a polynomial with a specific order. The polynomials are calculated in a
non-dimensional coordinate system and translated afterwards. The coordinates of the
polynomials are normalized using different hull parameters.
Output
HulGen has the ability to create a NURBS surface in Rhinoceros, a commercial 3D modeling
software widely used at the Naval Surface Warfare Center. HulGen is also able to export hull
offsets and diagrams.

Reduction of Mandatory Input Parameters

HulGen has 68 input parameters used to generate a hull form. However, using 68 different
parameters to create a hull form from scratch can be a difficult task for an inexperienced user. A
new feature has been implemented in HulGen which provides the ability to approximate 62
parameters based on the other 6 parameters. The parameters are approximated using modified
versions of formulas provided by a senior Naval Architect at the Naval Surface Warfare Center
with extensive experience in hull design. The formulas used are dependent on the type of ship.
The 6 main parameters which are used to calculate the remaining 62 can be found in Table 5.
18

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Table 5: HulGen Main Parameters

Parameter #
2
3
4
5
14
34

Parameter Name
Ship Length
Ship Beam
Ship Draft
Maximum Section Coefficient
Prismatic Coefficient
Depth at Midships

The 6 main parameters can easily be adjusted within the user interface. The algorithm for
determining each of the 62 remaining parameters is unique to the parameter and dependent on
the type of ship. The user must specify the ship type using a drop down menu. The interface for
this feature can be seen in Figure 13.

Figure 13: Baseline Hull Interface

19

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

The input reduction formulas implemented within HulGen, Figure 14, support the following
types of ships:

Conventional Surface Combatants

Unconventional Surface Combatants
Twin-Screw, Open-Stern, UNREP and Auxiliary
Single-Screw, Closed-Stern, UNREP and Auxiliary
Single-Screw, Open-Stern, UNREP and Auxiliary
Aircraft Carriers

Figure 14: Baseline Ship Type Selection Menu

This new feature allows a user to generate a hull by choosing a hull type and providing 6 main
parameters. HulGen will then fill all of the other parameters in for the user. The 6 provided
parameters and the 62 approximated parameters can then be used as a design baseline. The user
still has the ability to manipulate all 68 parameters before or after using the new feature, thereby
allowing the user to create hull forms that do not lie within the baseline types of ships.

Summary
A new feature has been implemented in HulGen which will allow an inexperienced user to
generate, modify, and export a hull form using only 6 parameters. The user is still able to use the

20

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

other HulGen features to modify hull forms to a greater degree. HulGen is suitable for future
integration with the CFD tools under development at George Mason University.

Resistance and Powering Prediction Tool Development

Background and Objectives
SWAD is a software tool developed and used by the Naval Surface Warfare Center which
predicts total resistance and powering in calm water for Small Waterplane Area Twin Hull
(SWATH) ships. Several versions of SWAD exist. The most commonly used version is
SWAD90, which was written in FORTRAN and is run using text files and command line inputs.
The most recent previous implementation was at CISD as an Excel tool which lacks a detailed
graphic user interface. The goal of this project was to provide SWAD an intuitive graphic user
interface as a stand-alone application. This was done by implementing SWAD in Visual Basic,
creating the newest version SWAD12. The user interface of SWAD12 can be seen in Figure 15.

Figure 15: SWAD12 Main Interface

21

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

SWAD12 Input
An interface similar to that of HulGen has been implemented to provide an intuitive layout. The
input parameters are divided among labeled tabs which the user can easily navigate through.
Example hulls have been included in the program, which may serve as a baseline for the user to
modify. Parameters that have a limited quantity of discrete possible values are represented as
dropdown menus. All other parameters may be easily adjusted through the use of slider bars or
by entering exact parameter values into textboxes. One tab of parameters including a dropdown
menu and several slider bars is shown in Figure 16. Any change to the value of either the slider
bar or the textbox automatically triggers an update to the other. Each slider bar has an associated
function which translates from the slider bars range of integer values to the parameters range of
valid values. Each slider bars function was designed to provide the slider bar with a reasonable
resolution as well as an appropriate minimum and maximum. Most parameters require a linear
scaling function, but some parameters use an exponential scaling function. For example, when
manipulating the slider bar for ship length, the parameter value will increase faster as the slider
bar moves towards the larger values. In previous versions of SWAD, all parameter values were
manually entered into a text file in a specific format and order before running SWAD and could
not be altered as easily. The dropdown menus and slider bars are useful when quickly adjusting
parameter values.

Figure 16: SWAD12 Input Interface

22

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

A key feature in the functionality of SWAD12 is backwards compatibility with previous versions
of SWAD. SWAD12 was implemented to allow the user to import original SWAD input files
compatible with the FORTRAN versions. This import feature is located in the menu bar as
shown in Figure 17.

Figure 17: SWAD12 Import and Export Menu

SWAD12 Output
The output interface for SWAD12 can be seen in Figure 18. SWAD12 is concerned primarily
with the portions of the ship below the Design Waterline (DWL). Portions above the waterline
are included for completeness and because they may be useful when moving the data to another
tool.
SWAD12 estimates the resistance and powering of a user specified SWATH hull form for a
given set of speeds. Coefficients of residuary resistance, coefficients of frictional resistance, and
the effective horsepower are calculated for hulls and struts. SWAD12 also calculates displaced
volume and wetted surface area for hulls, struts, and appendages. All of the output data is
displayed in Imperial units.
The user has the option to export import data compatible with previous versions of SWAD. The
user can export the output data in the same format as FORTRAN versions of SWAD. This
feature may useful in the event that individuals have designed scripts to read in data from
previous SWAD versions. The user can also export the output data to Excel in the same format
as the Excel implementation of SWAD. These options can be seen in Figure 17.
23

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Figure 18: SWAD12 Output Interface

Hull Representation
SWAD12 will model twin hulls or a single demi-hull. It allows one strut per hull or two struts
per hull. SWAD12 can model simple or bulged hulls with either circular or elliptical cross
sections. Hulls are modeled as an elliptical nose cone and a parabolic tail cone with a variable
number of cylindrical or conical sections between. The strut waterplane shape is modeled as an
elliptical or parabolic nose section, a rectangular parallel mid-section, and a parabolic tail
section. Struts are assumed to be wall sided with vertical leading and trailing edges.

Graphic Feedback
Previous versions of SWAD lacked graphic feedback on the users entered information. With the
new interface for SWAD12, valuable feedback is now presented to the user. Coefficients of
resistance curves are calculated and graphed with the output. An example can be seen in Figure
19. The user can specify whether the curves should be graphed as a function of the Froude
Number or the speed in knots. This is significant as in previous versions of SWAD the user must
manually graph the performance curves from the raw output data after exiting SWAD.

24

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Figure 19: SWAD12 Coefficients of Resistance Curves

The hull corresponding to the input parameters is graphed in the main interface to provide
feedback on the dimensions of the hull (Figure 20). The hull is approximated with straight lines
and is not meant to represent the curves of the hull used for calculation. The graphed hull is
useful, allowing the user to directly see if the information entered is reasonable. In addition, pop
up windows have been implemented to warn the user when approximations are made to the hull
used for calculations.

Figure 20: SWAD12 Graph of Hull

25

Naval Surface Warfare Center Carderock Division

Naval Research Enterprise Intern Program
Hull Form Design and Optimization Tool Development

Summary
A new tool, SWAD12, has been implemented in Visual Basic. SWAD12 is a SWATH resistance
and powering prediction tool which uses a graphic user interface to facilitate user input and to
provide graphic feedback. The outputs include plan and profile views of the SWATH, geometry,
resistance curves, and EHP data. SWAD12 is a remake of an old FORTRAN SWAD tool.
Features have been implemented which will allow backwards compatibility with FORTRAN
versions of SWAD.

Recommendations for Further Work

Further development is recommended for all of the tools.
The DE tool currently accommodates up to three speeds, but the tool can be further developed to
increase the number of input speeds. This will be useful when trying to optimize ships with
missions requiring several speeds. The DE tool can also be linked with HulGen and SWAD12
for further optimization studies.
Improvements to HulGen can expand its ship generation capabilities. Improvements include the
ability to model bulbous bows and hard chines. Also, validation of the hull forms generated with
HulGen is recommended. Development and validation of a database containing HulGen
parameters for existing ships would provide the user with a wider range of hulls to build from.
Further improvements to SWAD12 can be made to provide a more polished interface. Validation
of SWAD12 output against SWAD90 output for a variety of SWATH hull forms is
recommended. Development and validation of a database containing SWAD12 parameters for
existing SWATHs would provide the user with a wider range of hull forms to use as a starting
point.
These tools are all part of a larger ongoing tool development effort. Further work needs to be
completed to integrate each of them into the comprehensive hydrodynamic hull form
optimization toolset under development as a collaboration between George Mason University
and CISD.

26