Implementation of Synchronous Reference Frame theory for Harmonic Reduction in Computer Loads

This paper presents a non-linear control technique to reduce harmonic current in computer loads using active power filter. Nonlinear loads like personal computer draws non-sinusoidal current and produce distorted current waveform at three phase AC source. The usage of personal computers is increased today, that generates harmonics and creates severe problems in power system. Active filters are used to reduce voltage and current harmonics produced by computer loads on the supply side. The proposed system is implemented with synchronous reference frame (SRF) theory for extracting compensating current reference. The reference signal obtained by SRF theory is utilized to generate switching pulse for active power filter. The performance of SRF controlled time series neural network is implemented with voltage-controlled active power filter to reduce distortions in the supply side. The system is modelled in MATLAB Simulink environment, and the results of neural network reduce the harmonic current to a minimum level when compared to a traditional PI tuned SRF controller.