LAUSR

For speed-sensorless induction motor (IM) drives, the high-order sliding-mode observer (SMO) can improve the robustness of speed estimation. However, it is difficult for the existing high-order SMOs to ensure finite-time convergence, resulting in performance degradation or even instability in the wide-speed range. To address this problem, a high-order SMO with adaptive gain is proposed based on the motor operation speed. First, a high-order sliding-mode surface is designed to achieve finite-time convergence of the rotor flux estimation in the reference model. Then, an adaptive gain is proposed to guarantee the stability of the high-order SMO in the wide-speed range. Second, an improved discretization method is introduced to ensure the stability and reduce the discretization error in the adaptive model. On this basis, the rotor speed is further obtained by the model reference adaptive system. Compared with the existing SMOs, the investigated method can achieve finite-time convergence and guarantee the estimation stability in the wide-speed range, including field-weakening range. Finally, the experimental results from a 3.7 kW IM test bench reveal the effectiveness of the investigated strategy.