IEEE Transactions on Transportation Electrification | 2021
Multi-Vectors Model Predictive Control with Voltage Error Tracking for Five-Phase PMSM Short-Circuit Fault-Tolerant Operation
Abstract
Model predictive control (MPC) has been successfully extended into fault-tolerant operation for the five-phase permanent-magnet synchronous motor (PMSM) under an open-circuit fault. However, compared with the open-circuit fault, the influence of short-circuit fault is more serious. Therefore, this paper proposes a new multi-vectors model predictive control (MV-MPC) with voltage error tracking for a five-phase PMSM under short-circuit faults, including single-phase short-circuit fault and interphase short-circuit fault. The key of this method is to directly add short-circuit compensatory currents to reference currents that benefited from the advantages of model predictive control. Due to the loss of the faulted phase, the shapes of ordinary sectors under faults become asymmetrical. Then, a multi-vectors selection method is presented by constructing the cost function with voltage error terms and determining optimal vectors based on the shortest distance principle and sine theorem. Compared with the existing methods, the total harmonic distortion of the current and torque ripple is reduced greatly under the proposed method. Finally, the effectiveness and superiority of the proposed method are validated by experiments.