Design optimization of asymmetric wave energy converter using artificial neural network model

Abstract

The present study aims to improve the mean extracted power of a Wave Energy Converter (WEC) by mapping the parameters of its ballast weight and position, wave frequency, viscosity, and Power Take-Off (PTO) damping using an Artificial Neural Network (ANN) model. A total of 25 types of WEC rotors are designed with varying ballast weights and positions. The hydrodynamic coefficient and response of each rotor are determined using linear potential theory and viscous damping is estimated using computational fluid dynamics. The optimal design parameters are obtained by applying the trained model to a large randomly generated input dataset and the prediction output is evaluated to determine the best design parameters. According to the findings of the study, a well-trained model can predict and adopt to the nonlinear behavior of the given dataset as well as provide the optimal design parameters for the selected pitch-type WEC rotor.