Multi-vector model predictive fault-tolerant current control using reconstructed vector distribution for five-phase PM motor

Abstract

<p indent=0mm>To address a single-phase open-circuit fault, this paper proposes multi-vector finite-control-set model predictive fault-tolerant current control. This method includes reconstruction of the post-fault vector distribution and introduces a vector selection method with a reduced computational burden. The reconstructed vector distribution is based on the principle of minimizing the copper loss, and is generated by combining several basic vectors. Standard PWM waveforms and an enlarged modulation area are considered simultaneously. The vector selection method adopts the basic principle of classic finite-control-set model predictive control twice, which avoids evaluation of all the candidate vectors. The computational burden can thus be reduced without a vector selection error. In addition, to further improve the operational performance, duty cycle modulation technology is adopted to optimize the amplitude of the candidate vectors. A thorough experimental evaluation is performed, which demonstrates the effectiveness and superiority of the proposed fault-tolerant method.