A Model Predictive Control-Based Open-Circuit Fault Diagnosis and Tolerant Scheme of Three-Phase AC-DC Rectifiers

Three-phase ac-dc rectifiers are extensively employed in critical power conversion applications. This paper proposes an open-circuit fault diagnosis and tolerant scheme based on model predictive control (MPC) to enhance their reliability. The proposed fault diagnosis method is directly implemented by checking the mismatch signal between the applied switching state and the estimated one. The switching state error is independent of the voltage or current signals. Therefore, it is applicable to different voltage rating and loading scenarios. As long as the faulty switch conducts current, the open-circuit fault can be isolated very fast within several sampling periods. By introducing the healthy leg variables, a generalized model including both normal and faulty operations is proposed and the fault-tolerant scheme can be integrated into the MPC scheme for normal operations. With the proposed generalized model and fault-tolerant scheme, the rectifier can autonomously recover from open-circuit faults and the continuous operation of three-phase ac-dc rectifiers is maintained. The proposed fault diagnosis and tolerant scheme are applicable under balanced and unbalanced grid conditions. Experimental results demonstrate the effectiveness of the proposed integrated fault diagnosis and fault-tolerant scheme.