Design of PI-neural controller for hybrid active power filter

Abstract

This paper aims to design a control method using an adaptive controller for hybrid active power filter. The controller of designed method includes a traditional discrete proportional Integral controller and a neural regulator. The neural regulator is used to estimate the nonlinear model of Hybrid active power filter and predict an output value in the future to adjust the parameters of the traditional discrete proportional integral controller according to the change of load. Compared to the control method using a conventional Proportional Integral controller, the proposed controller shows the advantages of smaller compensation error and smaller total harmonic distortion and able to online control very well. The simulations have verified the effectiveness of proposed controller.