Estimation and Prediction using Deep Learning Model for the Wave Loads of Mid-Size Fishing Vessel

Abstract

The present study performed hydrodynamic analysis of a mid-size fishing vessel to evaluate the design load of small-and mid-size fishing vessels. The primary motion performance and wave load values were determined using pitch, roll, and vertical wave bending moments. The analysis results were then compared to those of the design loads. The analysis results of pitch and roll showed lower values than the design loads, but the vertical wave bending moment showed similar results to the design load. In addition, the motion performance and wave loads were predicted using a deep learning model. Dataset used in deep learning model was constructed considering various environmental conditions, and the 6 degrees of freedom motion responses and vertical wave bending moments of fishing vessel were predicted using the trained deep learning model. The predicted values by the established deep learning model were in good agreement with the analysis results with high accuracy. The challenges of high analysis skills, time, and expense for the evaluation of fishing vessel design loads may be addressed by the proven deep learning approaches. In addition, this study aim to increase the applicability of established deep learning model by considering various fishing vessels and design conditions through additional research in the future.