Artificial neural network based three phase shunt active power filter

This work describes artificial neural network (ANN) based control algorithm for three phase three wire shunt active power filter (SAPF) to compensate harmonics and improve power quality. System consists of three phase insulated gate bipolar transistors IGBT based current controlled voltage source inverter (CC-VSI), series coupling inductor and self supported DC bus. Increasing application of non-linear loads causes power quality problem. SAPF is one of the possible configurations to improve power quality. Traditional SAPF have PLL based unit template generator for extraction of fundamental signal. Traditional PLL needs to be tuned to obtain optimal performance for frequency estimation. It requires initial assumptions for fundamental frequency and minimum frequency. With varying frequency, it can't be dynamically tuned for optimal performance. A new ANN based fundamental extraction based on Lavenberg Marquardt back propagation algorithm is proposed. Proposed SAPF is modelled in Simulink environment. Simulated results show the capability of proposed system.