Sensorless Control of High-Speed Linear Synchronous Motor with Online Parameter Identification Based on Model Reference Adaptive System

Abstract

Electromagnetic sleds require high-performance sensorless control for linear synchronous motor (LSM) propulsion systems, yet model-based methods suffer from position estimation errors caused by parameter mismatches due to temperature rise and stator segment transitions. This paper proposes an integrated sensorless control scheme with online parameter identification based on the model reference adaptive system (MRAS). First, the position estimation errors induced by both uniform parameter deviation and inter-phase asymmetry are theoretically quantified, revealing that resistance deviation causes errors inversely proportional to speed, while inductance deviation introduces speed-independent errors. Subsequently, adaptive laws for online stator resistance and inductance identification are designed using Popov’s hyperstability theory, ensuring global asymptotic stability. The identified parameters are fed back to the back-EMF observer in real time. Simulation and experimental results demonstrate that the MRAS algorithm maintains uniform ultimate boundedness under initial parameter mismatches and ensures accurate tracking during 3 g acceleration. Under full closed-loop sensorless control with parameter adaptation, the position error is confined within 9.3% of the pole pitch. The proposed scheme satisfies the operational requirements of the electromagnetic sled and provides an effective solution for sensorless control.