Fatigue damage prediction model for floating wind turbine mooring system based on neural network

In this paper, a prediction method of fatigue damage of floating wind turbine mooring system based on neural network is proposed for NREL 5MW semi-submersible wind turbine.The coupling model of wind turbine, platform and mooring system was established by using SolidWorks, FAST and AQWA simulation software. The introduced F2A solved the problem that AQWA could not effectively deal with the aerodynamic-servo-elasticity and hydrodynamic coupling of floating wind turbine, improved the prediction accuracy, and analyzed the tension time history of mooring system under different incidence angles. Based on Miner's linear fatigue cumulative damage theory and rainflow counting method, the fatigue damage under short-term sea conditions was calculated; Build a suitable artificial neural network, train and validate given samples, select the neural network with the highest fitting accuracy, and obtain a set of prediction models that can predict fatigue damage within a certain range given input variables, reducing computational costs. The results indicate that the time cost of using neural network models to predict fatigue damage is significantly reduced, and the prediction effect is good, the prediction efficiency is improved. Therefore, constructing a neural network for predicting mooring fatigue damage is reasonable and effective.