Simulation and real time implementation of shunt active power filter for power quality enhancement using adaptive neural network topology

Abstract

With the rise in non-linear loads, the need for a veracious power filter to alleviate voltage and current harmonics for meeting power quality has been a challenging requirement of the existing power system. This paper presents a systematic approach to improve power quality with a shunt active power filter (SAPF). The controller consists of a neural network approach to enhance the switching control of SAPF under dynamic and non-linear loading conditions. DC link voltage has been regulated using proportional-Integral (PI) control. To extract reference current, instantaneous-reactive power (IRP) theory has been used. IRP theory employing indirect current control strategy provides reference source current signals as well as regulated PWM switching signals for hysteresis control. An ADALINE network has been devised to extract fundamental signals from the polluted supply voltage. The model has been simulated on the MATLAB®/Simulink platform and validated by SAPF prototype in the laboratory using the TMS320F28335 floating-point DSP controller.