Model Reference Adaptive System-Based Speed Estimators for Sensorless Control of Interior Permanent Magnet Synchronous Machines

In this work, speed estimators are designed based on a model reference adaptive system (MRAS) for sensorless control of interior permanent magnet synchronous machines (IPMSMs). To provide better filtering performance and reduce the noise contents in the estimated speed, an adaptive line enhancer is proposed to work with a sliding-mode observer to provide an improved reference model for speed estimation. In addition, a heterodyning speed adaption scheme is proposed to replace the conventional speed adaption mechanism. The proposed MRAS-based speed estimator has two different operating modes, which are suitable for generator- and motor-type applications, respectively. The effectiveness of the proposed speed estimators are verified by simulation using real-word vehicle data logged from a test vehicle. Furthermore, experimental results on a test stand for a heavy-duty IPMSM are also provided.