Second-Order Sliding-Mode Observer With Online Parameter Identification for Sensorless Induction Motor Drives

Abstract

Parameter identification plays an important role in speed estimation schemes. This paper presents a speed estimation scheme based on second-order sliding-mode supertwisting algorithm (STA) and model reference adaptive system (MRAS) estimation theory, in which both variations of stator resistance and rotor resistance are deliberately treated. A stator current observer is designed based on the STA, which is utilized to take the place of the reference voltage model of the standard MRAS algorithm. The observer is insensitive to the variation of rotor resistance and perturbation when the states arrive at the sliding mode. Derivatives of rotor flux are obtained and designed as the state of MRAS, thus eliminating the integration. Furthermore, in order to improve the near-zero speed operation, a parallel adaptive identification of stator resistance is designed relying on derivatives of rotor flux and stator current. Compared with the first-order sliding-mode speed estimator, the proposed scheme makes full use of the auxiliary sliding-mode surface, thus alleviating the chattering behavior without increasing the complexity. The robustness and effectiveness of the proposed scheme have been validated experimentally.