An intelligent ANFIS-based sensorless control scheme for robust rotor position estimation and minimisation of ripples in PMSM drive

Abstract

An intelligent sensorless control strategy for Permanent Magnet Synchronous Motor (PMSM) incorporating super-twisting sliding mode observer (STSMO) with an Adaptive Neuro-Fuzzy Inference System (ANFIS) is proposed in this work to enhance position and speed estimation of PMSM drive. This method leverages the benefits of both traditional STSMO and neuro-fuzzy-enhanced STSMO. Stability is further ensured by the incorporation of speed-dependent adaptive gains. The proposed method successfully addresses issues such as chattering and DC offset, while providing precise and smooth speed and position estimation. Nevertheless, noise in the back-EMF signal and DC offsets might still lead to estimation inaccuracies. The predicted back electromotive force (EMF) signal is refined by integrating ANFIS with STSMO to address these problems. Lyapunov-based analysis establishes observer stability, and computational assessment confirms real-time feasibility. Validation has been conducted through simulations and experimental analysis under various dynamic scenarios, including speed variation, load disturbances, and startup conditions. The results demonstrate that the proposed scheme achieves superior rotor position and minimises torque ripple while ensuring robust and smooth operation. Additionally, it offers faster settling times and less speed overshoot compared to traditional approaches, highlighting its potential for industrial PMSM applications.