Non-deterministic approach for the movement of the small fishing vessels

Abstract

Estimating dynamic roll motion under fishing operation conditions is one of the unresolved problems with respect to the capsize accident of fishing vessels. In addition to potential non-linear response of the roll motion to the wave, unexpected factors occurred by fishing operations make the roll motion complicated. Considering the estimation for the dynamic phenomena generating on the fishing vessels, the numerical simulation based on non-deterministic approaches is one of the effective methods. Among them, it is well-known that a Neural Network model is a practical methodology. Based on the tank tests, the roll response models in fishing operation conditions were constructed. Appropriate estimates for the actual roll motion with a full-scale showed the effectiveness of the Neural Network model. Furthermore, as the model forecasts properly the critical condition occurring for the shipping water, it is suggested that the model is also effective utilizing as the forecasting model to avoid the fishing disaster.

Modeling Ship Equations of Roll Motion Using Neural Networks

“…Tremendous successes have been seen using more traditional neural network‐based approaches (Hess et al 2006a, b; Kimura and Amagai 2003; Haddara and Hinchey 1995; Khan et al 2005). The fundamental aim of this approach is to yield physics‐based equations with which one might then apply analytical approaches for rapid analysis of stability boundaries and/or couple with a neural network to determine an ideal reduced‐order model to fit the current vessel motion conditions.…”