Neural Network-Based Control Parameter Identification Method for DFIG

In response to the challenge of obtaining high-precision control parameters for doubly-fed wind turbine electromagnetic models under transient conditions, a method for high-precision identification of control parameters for doubly-fed wind turbines based on Long Short-Term Memory (LSTM) neural networks is proposed. Firstly, the hardware-in-the-loop (HIL) data of the doubly-fed wind turbine controller is measured and obtained using the RT -LAB semi-physical simulation platform, and an identification model of the doubly-fed wind turbine model is constructed in Matlab/Simulink. Then, the Pearson correlation coefficient method is used to extract highly correlated features and train the neural network. Finally, the proposed LSTM neural network is used to identify the control parameters of the doubly-fed wind turbine and compared with the measured data, validating the feasibility, effectiveness, and practicality of the proposed method. The results indicate that, compared to traditional identification methods, the proposed LSTM neural network parameter identification method can effectively improve the identification accuracy of the control parameters of the doubly-fed wind turbine electromagnetic model under transient conditions.