A novel technique for voltage flicker mitigation using dynamic voltage restorer

Abstract

This paper deals with mitigation of voltage flicker using an intelligent dynamic voltage restorer (DVR). Voltage flicker is produced in the distribution system due to an arc furnace which is a highly nonlinear load in nature. The control scheme of the proposed DVR is based on a neural network (NN) controller whose weights are trained using hybrid of particle swarm optimization (PSO) and gradient descent (GD). A comparative analysis is done for three different controllers: proportional integral (PI), NN with GD, and NN with hybrid of PSO and GD. Simulated results based on peak overshoot and maximum percentage total harmonic distortion (THD) of load voltage shows the superiority of the proposed NN controller.