Model-Free Predictive Current Control of PWM Rectifier Under Unbalanced and Distorted Network

Abstract

Model predictive current control (MPCC) has been proposed for the control of a pulse width modulation (PWM) rectifier due to its simple principle and quick response. MPCC achieves good performance for ideal grid voltages when an accurate inductance value is known. However, the actual grid voltages are usually unbalanced and distorted, and the inductance may vary with the working condition. In these conditions, the performance of a PWM rectifier deteriorates significantly due to the current harmonics and power ripples. To cope with distorted grid voltages and inductance variations, this paper proposes model-free predictive current control (MFPCC) for a PWM rectifier, which can achieve sinusoidal grid currents and a constant active power even under an unbalanced and distorted network. The robustness against an inductance variation is achieved by using an extended state observer (ESO) based on an ultralocal model of the system. The current reference is calculated based on the principle of active power ripple elimination and then is tracked by predictive control. The presented experimental results prove that the proposed MFPCC scheme achieves good performance under a distorted network even with an inaccurate inductance value.