ANN-Based Adaptive PI Control for Wind Turbine with Doubly Fed Induction Generator

This paper focuses on developing a novel algorithm which dynamically optimizes the controllers of doubly fed induction generator (DFIG) driven by a wind turbine (WT) to increase DFIG transient performance in all wind speed conditions. Particle swarm optimization (PSO) is proposed to optimize parameters of PI controllers of DFIG’s rotor side/grid side converters (RSC/GSC) at different wind speeds in order to maximize the damping ratios of the system eigenvalues in small signal stability analysis. Based on the optimal values and the wind speed data set, an artificial neural network (ANN) is designed, trained, and it has the ability to quickly forecast the optimal values of parameters. Adaptive PI controllers (including ANN) are designed which dynamically change PI gain values according to different wind speeds. Simulation is done via PSCAD software for a single machine connected to an infinite bus (SMIB) system. The results show that the DFIG of ANN based adaptive PI control could significantly contribute in the transient performance improvement in a wide wind speed range