Elman Neural Network with Customized Particle Swarm Optimization for Hydraulic Pitch Control Strategy of Offshore Wind Turbine

Abstract

Offshore wind turbines have garnered significant attention recently due to their substantial wind energy harvesting capabilities. Pitch control plays a crucial role in maintaining the rated generator speed, particularly in offshore environments characterized by highly turbulent winds, which pose a huge challenge. Moreover, hydraulic pitch systems are favored in large-scale offshore wind turbines due to their superior power-to-weight ratio compared to electrical systems. In this study, a proportional valve-controlled hydraulic pitch system is developed along with an intelligent pitch control strategy aimed at developing rated power in offshore wind turbines. The proposed strategy utilizes a cascade configuration of an improved recurrent Elman neural network, with its parameters optimized using a customized particle swarm optimization algorithm. To assess its effectiveness, the proposed strategy is compared with two other intelligent pitch control strategies, the cascade improved Elman neural network and cascade Elman neural network, and tested in a benchmark wind turbine simulator. Results demonstrate effective power generation, with the proposed strategy yielding a 78.14% and 87.10% enhancement in the mean standard deviation of generator power error compared to the cascade improved Elman neural network and cascade Elman neural network, respectively. These findings underscore the efficacy of the proposed approach in generating rated power.