Neural network based detection in switched capacitor filter for harmonic mitigation

In this paper, neural network technique used in digital based control for switched capacitor filter (SCF) is proposed. The SCF is an alternative tool to mitigate harmonic distortion in the AC supply system as it has only small capacitor size and low switching frequency. Harmonic detection by using neural network technique is developed to determine harmonic component and hysteresis control is used to generate compensated current. The digital signal processor (DSP) is used to implement digital based control. Simulation and experimental work is carried out to evaluate its control performance. The results show the ability of this technique to achieve desirable harmonic mitigation.