Virtual position predictive control with system delay observer to improve PMLSM position tracking accuracy

Abstract

To improve the position tracking accuracy of permanent magnet linear synchronous motors (PMLSM), this paper introduces a virtual position predictive control (VPPC) with a system delay observer (SDO). Un-like conventional position predictive control (CPPC), which ignores the complexity of prediction models and velocity adjustments, the proposed VPPC combines a simplified position model with an active variable speed control. This design accelerates mover response and increases maximum reference speeds through active adjustments to better estimate the predicted output. Since the prediction period in CPPC often misaligns with the system delay, resulting in additional prediction errors, this paper further explores the relationship between predictive periods and system delay. Based on this analysis and a unified modelling concept, an SDO is included to observe and compensate for system delay, correcting the prediction period and optimising the control model. Experimental results on a PMLSM platform confirm the superior position tracking performance of the VPPC with SDO compared to conventional controllers.