Resistance Coefficient Estimation for a Submarine’s Bare Hull Moving in Forward and Transverse Directions

Abstract

Resistance of the bare hull of the tourist submarine with spherical heads, moving in forward and transverse directions is analyzed in OpenFOAM using Computational Fluid Dynamics. The resistance coefficients of the submarine are estimated for different length-to-diameter ratios and Reynolds numbers. The Artificial Neural Network with the optimum number of neurons is then trained to predict the resistance coefficients. Two simplified Artificial Neural Network models and Nonlinear Least Squares Marquardt-Levenberg algorithm are employed to fit the results in the form of equations that may be used in the initial design of this type of submarines. The comparative analysis of different prediction models is performed and guidelines for their practical application are given.

Keywords

• submarine
• resistance coefficient
• Artificial Neural Network
• Nonlinear Least Squares Marquardt-Levenberg algorithm
• Computational Fluid Dynamics