Optimized design of damped proportional-resonant controllers for grid-connected inverters through genetic algorithm

Abstract

This paper describes the design principles of damped proportional-resonant (PR) controllers used for the grid-connected inverters. Previous works focused on the frequency domain analysis to select parameters, which can't reflect the time domain performance directly. This paper presents a mathematical model of grid-connected inverter connected with a LC filter and isolation transformer in the stationary reference frame, whose order is reduced by weighted current feedback. Based on the reduced-order model, the principles for designing damped PR controllers in the frequency domain are presented in details. The optimized design of damped PR controller through genetic algorithm using the cumulative amount of current tracking error as indicator of the time-domain performance is proposed. Compared with traditional system, the system with the optimized damped PR controller achieves not only better stability in the frequency domain, but also more excellent performances of current tracking and lower total harmonic distortion in the time domain. Simulation results are presented to verify the feasibility and effectiveness of the proposed approach.