Dead-Beat Predictive Control of DFIG-DC Based on Data-Driven Neural Network Predictor

This paper is concerned with the dead-beat predictive control of a doubly fed induction generator in the presence of fully unknown motor parameters. Only the current and voltage values of the rotor are measured in the control system. Firstly, data-driven neural network predictors are proposed to approximate the model uncertainties as well as the unknown control gains. Based on the predictors, a dead-beat predictive controller is designed without any prior knowledge of model parameters. The input-to-state stability of the closed-loop system is proved by using the cascade theory. Simulation results show the effectiveness of the proposed dead-beat predictive control of a doubly fed induction generator based on data-driven neural network predictors.