Model-Predictive Sensor Fault Estimation and Accommodation Using Interval Type-2 T–S Fuzzy Models for PMSM Monitoring and Control Platform

Reliable measurement under sensor faults and disturbances/noises is crucial for monitoring and control of electric motor drive systems operating under complex nonlinear dynamics and uncertainties. This work develops a new model-predictive fault estimation (FE) and accommodation scheme for the interval type-2 (IT2) Takagi–Sugeno (T–S) fuzzy models in discrete-time settings, remedying sensor faults and bounded disturbances within ellipsoids to enhance measurement reliability. A flexible and less conservative two-step design scheme is given, which integrates an offline state/fault estimator design and an online optimized model-predictive fault-tolerant controller design subject to multivariable constraints. The former optimizes the estimation error bounds by scaling a scalar to ensure its convergence to a minimal robust positively invariant (RPI) set, and the latter imposes that the overall states converge to a time-varying RPI concentration. The framework guarantees recursive feasibility in controller gain tuning, boundedness of estimation errors, and input-to-state stability of the overall closed-loop dynamics. Validation results on a three-phase surface-mounted permanent magnet synchronous motor (SPMSM) monitoring and control platform demonstrate that the proposed approach achieves accurate sensor fault estimation (FE), preserves measurement reliability, and substantially improves control performance subject to sensor faults and disturbances.