Artificial neural network controller for single-phase active power filter

This paper shows the conception and simulation of a single-phase shunt active power filter for harmonics and power factor compensation of uncontrolled rectifier with battery load. The objectives of this project are to minimize the harmonic distortion and intensity the power factor of the single-phase system with battery charger load. The design of shunt active power filter is verified using the simulations in Matlab/Simulink. Shunt active harmonic filter was controlled by the neural network method where it applied the proportional and integral method in order to adjust the direct current link voltage and the hysteresis current controller is employed to generate signals for switching the voltage source inverter. The error signal has been compensated using the controllers. The reference filter currents signal is obtained by subtracting the line current with the compensated signal from the controller. This reference current is fed to the hysteresis current controller and compare with the sensed filter currents to obtain the switching signal for active power filter. Simulation results are acquired with the conventional PI controller and ANN controller.