An improved binary differential evolution algorithm for optimizing PWM control laws of power inverters

摘要

Stochastic optimization methods inspired by biological evolution system have been widely employed to optimize PWM control laws of power inverters. But the existing approaches impose a serious computational burden and difficult parameter tuning issues. However, the differential evolution (DE) algorithm has the superiority of simple implementation and few parameters to tune. Thus, we propose an improved binary DE (IBDE) algorithm for optimizing PWM control laws of power inverters. The proposed algorithm focuses on the designs of the adaptive crossover and parameterless mutation strategies without imposing an additional computational burden. In numerical experiments, a single-phase full-bridge and two-level three-phase inverters are considered, and the optimal PWM control law is calculated to maximize the closeness of the controlled inductor current to sinusoidal reference current by using the proposed algorithm. Experimental results indicate that IBDE can obtain high quality output waveform that is a very good approximation to the sinusoidal reference waveform. Moreover, the spectrum analysis for the optimal PWM control law obtained by IBDE indicates that the lower odd-order harmonics are eliminated, while the existing peer algorithms cannot do well. We also carry out experiments on sensitivity analysis with respect to several important parameters.