Particle swarm optimization-based gain, delay compensation and filter determination of a repetitive controller for a grid-tie converter

The paper presents evolutionary optimization of a repetitive control system for a grid-tie converter. Based on the internal model principle, the repetitive controller (RC) can reduce the influence of periodic disturbances. The goal of the presented control system is to draw sinusoidal currents under distorted grid voltage conditions. To provide a sufficient stability margin, a phase lead compensation is often introduced in the RC structure. In the described scheme, it is done by a Taylor series expansion based fractional delay filter. Therefore, the compensation time does not have to correspond with the integer number of samples. The tuning procedure is done by a particle swarm-based search for optimal parameters of the repetitive controller (gain, phase lead compensation and filter design) and PI controllers. The optimization results are verified by numerical simulations.