Position Sensorless Control of PMLSM by Sum-of-Squares Approach

Abstract

This paper presents an observer-based sensorless drive system for the position control of the permanent magnet linear synchronous motor (PMLSM). To achieve the purpose of position control and maintain the stability of the global system, the drive system is designed using the sum-of-squares (SOS) method. With the help of the SOSTOOLS toolbox, the control algorithm can be established without the need for complicated mathematical derivation. First, to apply the SOS technique, the motor model is treated as a fifth-order polynomial dynamics model. Based on this model, the state estimation of the full order Luenberger observer is studied. The feedback gain of the observer is obtained in polynomial form rather than a constant, thus making the observer nonlinear and time-varying. Compared with existing studies, the SOS observer construction requires no additional assumptions and no adaptive mechanisms to cope with changes in velocity. By combining the SOS method with the Lyapunov analysis, the observer-based control system is systematically developed to ensure the stability of the global system. The experimental results verify the proposed control scheme.