Neural Networks-based Control Strategy for Single-Phase Shunt Active Power Filter

Active power filter(APF) will be the main device in harmonic elimination in future, and the control strategies for APF were generally based on the simplified mathematical models and seldom took the system parametric variation in consideration. Thus the system will destabilize, a novel control strategy based on neural networks is proposed. Discussions on the design of NARMA-L2 model-based controller adopting feedback linearization principle are given. The controller not only discards the system model building, owns a good dynamic response and robustness, but also is convenient for engineering practice. Simulation results are included to demonstrate the effectiveness of the proposed control strategy.