WaveResist (WIN/MAC): An Artificial Neural Network Tool for Ship Wave Resistance Coefficient Calculation

WaveResist is a program designed to calculate the dimensionless coefficient of added wave resistance (CAW) for ships. This calculation is performed using four sets of artificial neural networks (ANNs). These ANNs utilize the following dimensionless hull shape parameters as inputs:
• L/B: Ratio of the length between perpendiculars to the beam (-),
• B/T: Ratio of the beam to the draught (-),
• L/T: Ratio of the length between perpendiculars to the draught (m),
• CB: Block coefficient (-),
• Fn: Froude numer (-),
• λ/L: Ratio of the wavelength to the length between perpendiculars (-).
To develop these four ANN sets, 919 measurements from model tests of 20 transport ship models were used.
The primary data ranges used for training the ANN sets included:
• L/B in the range from 5.1 to 8.4,
• B/T in the range from 2.5 to 4.5,
• L/T in the range from 15.4 to 30.6,
• CB in the range from 0.503 to 0.87,
• Fn in the range from 0.087 to 0.27.
WaveResist implements four ANN sets with the following input parameters:
• L/B, B/T, L/T, CB, Fn, λ/L,
• L/B, L/T, CB, Fn, λ/L,
• L/B, B/T, CB, Fn, λ/L,
• L/T, CB, Fn, λ/L.
Each of the ANN sets consists of five elementary neural networks.
The added wave resistance coefficient was defined as:

CAW = R / (ζₐ² ρ g (B² / L))

where:
CAW is the added wave resistance coefficient,
R is the wave resistance,
ζₐ is the amplitude of the regular wave,
ρ is the density of seawater,
g is the acceleration due to gravity,
B is the breadth of the ship,
L is the length of the vessel.

Binary files for WaveResist are available for macOS Intel and Silicon (WaveResist.dmg) and Windows (compressed with 7-Zip as WaveResistWin.exe), along with the Pascal source code (WaveResist_src.zip). The WaveResist_src.zip file contains all source files, and the unitconst.pas file includes the weights for all neural networks. The development and compilation of WaveResist were carried out using the Lazarus IDE. The Lazarus project files (WaveResist_src.zip) are also listed below for download.

WaveResist是一款专为计算船舶附加波浪阻力无量纲系数（CAW）而设计的程序。该计算依托四组人工神经网络（Artificial Neural Networks，ANNs）完成，这些ANNs以下列船体形状无量纲参数作为输入：
• L/B：垂线间长与船宽之比（无量纲），
• B/T：船宽与吃水之比（无量纲），
• L/T：垂线间长与吃水之比（单位：米），
• CB：方形系数（无量纲），
• Fn：傅汝德数（无量纲），
• λ/L：波长与垂线间长之比（无量纲）。

为开发这四组ANNs，研究人员采用了20艘运输船舶模型的919次模型试验测量数据。用于训练ANNs的主要参数范围如下：
• L/B：5.1至8.4，
• B/T：2.5至4.5，
• L/T：15.4至30.6，
• CB：0.503至0.87，
• Fn：0.087至0.27。

WaveResist搭载了四组ANNs，各组的输入参数分别为：
• L/B、B/T、L/T、CB、Fn、λ/L，
• L/B、L/T、CB、Fn、λ/L，
• L/B、B/T、CB、Fn、λ/L，
• L/T、CB、Fn、λ/L。

每组ANNs均由5个基础神经网络构成。

附加波浪阻力系数的定义如下：

CAW = R / (ζₐ² ρ g (B² / L))

其中：
CAW为附加波浪阻力系数，
R为波浪阻力，
ζₐ为规则波的波幅，
ρ为海水密度，
g为重力加速度，
B为船舶船宽，
L为船舶总长。

WaveResist的二进制安装包适配macOS Intel与Silicon架构（安装包为WaveResist.dmg）及Windows系统（经7-Zip压缩为WaveResistWin.exe），同时提供Pascal源代码包（WaveResist_src.zip）。WaveResist_src.zip包含全部源代码文件，其中unitconst.pas文件存储了所有神经网络的权重参数。WaveResist的开发与编译工作基于拉撒路集成开发环境（Lazarus IDE）完成，相关Lazarus项目文件已包含在WaveResist_src.zip中，可供下载。