Switching angles computations using pso in selective harmonics minimization pwm

Abstract

The preprogrammed pulse width modulation (PWM) method for operating power converters ensures minimum switching loss and at the same time better quality waveform at the output. The Fourier series analysis of the predefined output waveform with the removal of certain low order harmonics yields a set of trigonometric equations in which switching angles are treated as variables. It results in removal of preselected harmonics from the output and control on the fundamental component. The main challenge in implementation of selective harmonics minimization is in solving of these system of highly non linear equations which have different number of solutions for different range of modulation index. The particle swarm optimization technique can be used to obtain switching angles from the system of highly non linear equations by properly formulating an objective function which need to be minimized. The requirement of symmetry constraints on the output waveform can be implemented using penalty function approach. The main advantage of the PSO approach over algebraic and numerical based techniques is that it can be used for computation of large number of switching angles with less dependency on the initial guess of the solution.