Adaptive Sliding Mode Observers for Speed Sensorless Induction Motor Control and Their Comparative Performance Tests

This paper presents adaptive sliding mode observers (A-SMOs) performing speed estimation for sensorless induction motor drives utilized in both industrial and electrical vehicle (EV) applications due to their computational simplicity. The fact that the constant switching gain (λ0) is used in conventional SMOs (C-SMOs) leads to the chattering problem, especially in low-speed regions. To tackle this issue, this paper proposes two different λ0 adaptation mechanisms based on fuzzy and curve fitting methods. To estimate stator stationary axis components of stator currents and rotor fluxes together with the rotor speed, the proposed A-SMOs only utilize the measured stator currents and voltages of the IM. Here, the difference only between the estimated and measured stator currents is determined as the sliding surface in the proposed A-SMOs. To demonstrate the effectiveness of the proposed fuzzy-based A-SMO (FA-SMO) and curve fitting-based A-SMO (CFA-SMO), they are compared with C-SMO in real-time experiments for different scenarios including wide speed range operations of IM with/without load torque changes. Moreover, the stator and rotor resistances as well as the magnetizing inductance variations are also examined in real-time experiments of the proposed methods and the conventional one. The estimation results demonstrate how positively the λ0 adaptations in FA-SMO and CFA-SMO affect the performance of C-SMO. Finally, two A-SMOs with improved performance are introduced and verified through real-time experiments.