Fault Tolerant CCS-MPC for Asymmetrical DTP-PMSM Drives Assisted by Fault Detection Algorithm

Fault tolerance is a highly desirable feature for safety-critical electric drive applications. This paper proposes an open-circuit fault tolerant model predictive controller (FT-MPC) for asymmetrical dual three-phase (DTP) permanent magnet synchronous machine (PMSM) drives, which is formulated using the continuous control set (CCS) approach. As an alternative to conventional control methods, this predictive controller solves an unconstrained quadratic programming (QP) problem to obtain the voltage references. The FT-MPC is able to operate in a reduced form under healthy operation. To enable the transition to the post-fault control mode, a fault detection and diagnosis algorithm, based on the natural coupling concept, is proposed. The control solution is implemented in an OPAL-RT OP4510 real-time digital system and is experimentally validated. Thanks to the optimisation efforts carried out during implementation, the computational burden of the proposed algorithm is kept low: the developed software requires only about 7.6 % of the controller’s sampling time to compute during healthy operation (reduced form), and about 10.1 % during the post-fault mode, demonstrating its potential applicability in automotive grade microcontrollers. On average, open-phase fault detection times of the order of 15 ms are achieved, allowing for a fast and smooth transition between healthy and post-fault operating modes. Finally, dynamic performance is addressed under standardised driving cycle profiles. The ability of the FT-MPC to keep the torque-producing sub-space on a circular trajectory during the fault detection stage is confirmed, thus avoiding additional undesired torque ripple.