ADAPTIVE SUPER TWISTING DISTURBANCE OBSERVER BASED ROBUST SPEED CONTROL FOR PI-BASED FIELD ORIENTED INDUCTION MOTOR DRIVES

Abstract

This paper proposes an adaptive super-twisting disturbance-observer-based robust speed control strategy for proportional–integral (PI)-based field-oriented control (FOC) induction motor drives. Although PI regulators are widely used in industrial FOC systems because of their simple structure, low computational burden, and practical reliability, the outer speed loop remains sensitive to load disturbances, parameter variations, and unmodeled dynamics, resulting in degraded transient performance and limited robustness. To address these limitations while preserving the conventional industrial FOC architecture, an adaptive super-twisting disturbance observer is incorporated into the speed loop, whereas standard PI controllers are retained for current and flux regulation. In that, an ASTDO observer is designed to reconstruct the lumped disturbances affecting the mechanical dynamics and generate a compensation term for the speed control loop. Through an adaptive super-twisting mechanism, the observer gains are adjusted according to the disturbance estimation error, enabling the control system to respond effectively to load torque variations and parameter uncertainties without requiring prior knowledge of disturbance bounds. The proposed PI–ASTDO structure therefore provides an adaptive and robust control mechanism while retaining the conventional FOC architecture and PI-based current and flux regulation. Lyapunov analysis is employed to establish the stability and finite-time convergence of the proposed observer. Furthermore, a hybrid genetic algorithm–particle swarm optimization (GA–PSO) method is applied to determine suitable controller parameters and further improve the dynamic characteristics of the drive. The effectiveness of the proposed structure is evaluated under rated-load operation, load torque disturbances, and rotor resistance variations. The results demonstrate improved speed-tracking performance, reduced overshoot and undershoot, smoother torque response, and enhanced disturbance rejection compared with conventional PI-based FOC and PI–STADOb control. The proposed approach also maintains low computational complexity and the practical implementation advantages of conventional industrial induction motor drives.