Gain-adaptive non-singular fast terminal sliding mode control for induction motors under load torque variations

This paper investigates an adaptive sliding mode speed control for induction motors (IMs) to enhance system robustness against uncertainties, especially those caused by load torque variations. Since the load torque is difficult to measure accurately in practical applications, a sliding mode load torque observer (SMLTO) is designed to estimate the unknown load torque within finite time. Based on the torque approximation, a non-singular fast terminal sliding mode control (NFTSMC) scheme is developed to ensure stability of the closed-loop system and achieve the predefined speed convergence rate. Compared with conventional fixed-gain control strategies, the proposed method inherently provides finite-time convergence and adaptive disturbance compensation without relying on manual gain retuning. The proposed control method not only guarantees the desired stability of the closed-loop system with a fast convergence rate but also improves the system robustness against parameter uncertainties and torque variations. Significantly, an adaptive gain adjustment method is advocated to automatically modify the control parameters according to the instantaneous speed deviation, thereby enhancing control accuracy and alleviating unwanted chattering. A real-world experiment is presented, which demonstrates the merits of the proposed method in achieving a quick convergence rate, accurate torque estimation, and guaranteed stability of the controlled system under load torque variations. The experimental results are obtained from repeated trials and show consistent improvements in convergence speed, disturbance rejection, and steady-state accuracy. Experimental data are provided to verify the effectiveness and feasibility of the proposed control method.