Model Predictive Virtual Flux Control to Improve Performance Under Distorted Input Voltage Conditions

Abstract

This paper proposes a model predictive virtual flux control (MPVFC) method that can mitigate the degradation of the total harmonic distortion of rectifier input currents experienced by active frontend (AFE) rectifiers when input voltages are distorted by harmonics. The proposed method utilizes a cost function based on virtual fluxes, which are robust to input voltage distortions, to control the rectifiers. For the calculation of the proposed cost function which consists of a predictive value of the rectifier virtual flux and a reference of the rectifier virtual flux, the dynamics of AFE rectifiers and the definition of the integration in the discrete time domain are used. Also, the phase of the reference current which is needed when calculating the reference of the rectifier virtual flux is obtained by the input virtual flux. The proposed MPVFC method shows better input current waveforms with lower total harmonic distortion than those obtained with the conventional model predictive control method under input voltage distortion conditions, which was proved via simulations and experiments.