Research on Early Warning of Temperature Fault of Wind Turbine Converter Based on GA-PSO-BP Neural Network

In order to solve the effect of fault deterioration on the performance of the converter during the operation of the wind turbine, and to improve the accuracy of its fault prediction, this paper proposes an early warning method for temperature faults of wind turbine converters based on GA-PSO-BP neural network. Firstly, select input variables that are closely related to the temperature of the converter, combine the SCADA system data to build a BP neural network model, and use genetic algorithm and particle swarm algorithm to optimize; Secondly, establish a sliding time window model based on the temperature residual output from the neural network model, to monitor the operation status of the wind turbine converter; Finally, the accuracy of the model was verified through actual operating data of a wind farm in Xinjiang. The results show that the wind turbine converter temperature fault early warning method based on GA-PSO-BP neural network has high accuracy in monitoring the temperature status of the converter, it can effectively identify the abnormal state of the converter, and can realize early fault warning 48 hours in advance. Which is of great significance to the stable operation and safety maintenance of the wind turbine converter.