Enhanced sensorless control of IPMSM drives: from full-torque startup to dynamic boosting

Abstract

This thesis investigates the degradation of rotor position estimation accuracy and control performance in sensorless control of Interior Permanent Magnet Synchronous Motors (IPMSMs) under heavy-load conditions, which is primarily caused by inductance saturation when employing the high-frequency (HF) square wave injection method. To address this issue, an enhanced phase-locked loop (EPLL) is introduced, which incorporates mutual inductance and current ratio estimation into the rotor position estimation process, thereby enabling more accurate position detection. Additionally, a nonlinear observer is implemented to estimate inductance variation, achieving precise initial rotor position estimation and full-torque startup performance. Furthermore, optimized strategies are proposed to minimize phase shift, with the objective of enhancing both position estimation accuracy and operational efficiency.

This thesis proposes a motor model that incorporates the effects of inductance saturation and cross-coupling. An enhanced PLL is introduced to accurately extract the rotor position signal. The stability and robustness of the proposed EPLL are validated through linear stability analysis, nonlinear stability analysis, and dynamic transition analysis. The control strategy integrates a cross-coupling factor and employs the EPLL to estimate the inductance to current variation ratio in real time. The proposed method enables precise estimation of rotor position and speed, even under conditions of severe magnetic saturation.

To enable full-torque sensorless startup of IPMSM, this thesis proposes a method that integrates magnetic saturation characteristics into the pulsating voltage model. A nonlinear observer is employed to estimate inductance variation, thereby enabling accurate rotor position estimation that is robust to nonlinear load effects. The proposed approach represents a paradigm shift from traditional saturation compensation techniques to the active utilization of saturation-induced inductance variations. This method ensures continuous full-torque production during the startup phase, while maintaining high position estimation accuracy and eliminating torque derating. To mitigate non-ideal effects in the current demodulation algorithm, such as phase delay introduced by filtering, a high-precision current observer is developed to capture specific current signals. By integrating the position observer model with filter dynamics, the proposed current observer effectively suppresses current noise and enhances rotor position estimation accuracy. All control strategies proposed in this research have been thoroughly validated through both simulation and experimental testing.

This thesis presents a sensorless control strategy that incorporates cross-coupling effects to enhance the control performance of IPMSM in the zero-low-speed operating range. The proposed control scheme not only improves system dynamic response but also enables accurate estimation of rotor torque and speed, even under over-saturation conditions. Additionally, the developed initial position estimation method ensures continuous full-torque production during motor startup without requiring torque derating. The introduced high-precision current observer effectively suppresses current noise and enhances steady-state control accuracy. The proposed approach systematically refines and optimizes the control strategy to guarantee stable motor operation during full-torque startup and under heavy-load conditions.

Date of Award	15 Jan 2026
Original language	English
Awarding Institution	University of Nottingham
Supervisor	John Xu (Supervisor), Jing Li (Supervisor) & Dunant Halim (Supervisor)