Digital Signal Processor based Performance Investigation of Indirect Current Controlled Active Power Filter for Power Quality Improvement

This paper presents a novel Digital Signal Processor (DSP) based control algorithm for 3-phase shunt active power filter (APF) using neural network and indirect current control strategy to eliminate current harmonics and reactive power compensation under ideal and distorted supply voltage. The compensation process is based on sensing of source current instead of filter current, and thus differs from conventional techniques for elimination of switching ripples. Fundamental frequency signal extraction is done using adaptive linear neuron (ADALINE) network from distorted voltage signal without using low-pass filter or phase-locked-loop block. ADALINE gives simple, robust and accurate design for selected frequency extraction. Instead of calculating oscillating and reactive power, the reference generation is used to calculate average active power component. Various simulation results are presented to study the performance of proposed design during steady-state and transient conditions of non-linear loads using MATLAB/Simulink tool under sinusoidal and distorted supply voltage conditions. Simulated response has been validated with a laboratory prototype using a Texas make floating-point TMS320F28335 DSP. Based on the simulated and experimental results it can be concluded that the spectral performance of the input current are in compliance with the harmonic limits imposed by IEEE-519 standard.