Capture Power Prediction of the Frustum of a Cone Shaped Floating Body Based on BP Neural Network

How to improve the power generation of wave energy converters (WEC) has become one main research objectives in field. This paper illustrates a framework on use back propagation (BP) neural network predicting capture frustum cone shaped floating body. Mathematical model single body is derived, and radius, semi-vertical angle, mass, submergence depth, take-off (PTO) damping coefficient, stiffness coefficient are identified as key variables. Commercial software ANSYS-AQWA used for numerical simulations obtain hydrodynamic parameters, then calculated by these parameters. A database containing 100 samples established Latin hypercube sampling (LHS) method, simple feature study conducted. BP with high accuracy designed trained predictions based built database. The results show that forecasting desired outputs great agreement error percentage not greater than 4%, correlation (CC) 0.9, P value close 1, root mean square (RMSE) less 139 W. proposed method provides guideline designers identify basic parameters system coefficient.