A Deadbeat Predictive Sliding‐Mode Observer and Finite Position Set‐PLL for Sensorless Control of PMSM Drives

ABSTRACT To improve the estimation accuracy and transient performance for sensorless control of permanent magnet synchronous motors (PMSMs), a deadbeat predictive sliding‐mode observer (DPSO)‐based back electromotive force (back‐EMF) estimation and a finite position set‐phase‐locked loop (FPS‐PLL) based on golden section search (GSS) algorithm is proposed. Firstly, by incorporating a super‐twisting algorithm in discrete‐time predictive control, a novel observer based on deadbeat prediction principle is proposed. The fast transient characteristic of deadbeat prediction is preserved, whereas the robustness and estimation accuracy is also enhanced. Then, the concept of finite control set‐model predictive control (FCS‐MPC) is introduced into the PLL, and a finite position set‐PLL is designed (FPS‐PLL). The rotor position and speed estimation is reformulated as a unimodal optimization, and it is efficiently solved by using the GSS algorithm. In addition, a tracking differentiator‐based speed observer (TDSO) is employed to achieve smooth speed extraction. Finally, comparative experimental results demonstrate that the proposed method achieves superior position and speed estimation performance in terms of accuracy, dynamic response, and robustness.