Neural Network Analysis of Passing Ship Forces and Moments in Confined Channels

Abstract

A passing ship induces surge and sway forces and yaw moments on moored ships. Breakouts of moored ships can result in personnel injuries and deaths, ship and shore facility damage, and pollution. In a confined channel, the passing-ship phenomenon involves six interrelated parameters, describing ship sizes and drafts, separation from sidewall or bank, separation between ships, and channel width. It is virtually impossible to develop an interrelated set of equations that can be used in a mooring analysis which adequately accounts for all of these parameters. Hydrodynamic computer programs, which take time and expense, are sometimes be used to calculate surge, sway and yaw for use as input to a mooring analysis. An artificial neural network model has now been developed to calculate surge and sway forces and yaw moments. It will be used in a computer program. This paper presents examples of data produced by that model which show the relationships of ship sizes, depth-to-draft ratios, wall separation, ship separation, and channel width to surge, sway and yaw. This information will be of interest to mooring designers and others concerned about the passing ship problem. This paper describes how the six-parameter neural network model was developed. It will be of interest to those wanting to model other complex multiparameter problems.