Abstract
The increasing demand for rare earth materials in the electric vehicle and power electronics industries has led to dwindling reserves and rising costs, particularly for Permanent Magnet Synchronous Machines (PMSMs). To address these challenges, rare-earth-free motors such as Permanent Magnet assisted Synchronous Reluctance Machines (PMa-SynRMs) have emerged as promising alternatives. The PMa-SynRM motor design aims to balance cost and performance. However, compared to Interior Permanent Magnet Synchronous Machines (IPMSMs), PMa-SynRMs exhibit more severe flux linkage nonlinearity, along with significant variations in inductance, flux saturation, and cross-saturation phenomena, all of which must be considered for high performance control.This thesis focuses on enhancing both the steady-state performance and dynamic response of PMa-SynRMs. To achieve this, an improved Maximum Torque per Ampere (MTPA) control strategy and an enhanced Extended State Observer (ESO) were proposed.
In Chapter 3, a comprehensive PMa-SynRM model was developed, incorporating inductance variation, flux saturation, and cross-saturation to accurately capture the machine's flux characteristics.
In Chapter 4, three improved MTPA control methods were introduced to optimize steady-state response:
1) A novel Pseudorandom Frequency Signal Injection (PRFSI) method that achieves continuous harmonic distribution, reduced MTPA angle detection errors, and better dynamic response compared to traditional Constant Frequency Signal Injection (CFSI).
2) An online MTPA control strategy based on High Frequency Signal Injection (HSI), which leverages only the permanent magnet flux linkage data. Error analysis and a supplementary control loop were proposed to compensate for MTPA detection discrepancies.
3) An online tracking detection MTPA strategy, robust against resistance variations and parameter uncertainties, featuring an improved convergence function for faster response and a Self-learning Control (SLC) mechanism to adapt to flux characteristics.
In Chapter 5, an enhanced anti-disturbance ESO-based control strategy was proposed to improve dynamic response. A second-order Anti-Disturbance Extended State Observer (A-DESO) was designed to minimize observation errors at low frequencies and enhance noise suppression at high frequencies, particularly for analog position signals and current sensor noise. Additionally, a third-order Improved Extended State Observer (IESO) was introduced to further reduce sensorless control noise, minimize observed position errors, and shorten convergence time.
The proposed methodologies were validated through both simulations and experiments, demonstrating significant improvements in machine modeling, steady-state performance, and dynamic response.
| Date of Award | 13 Jul 2025 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Alan Zhang (Supervisor), Chunyang Gu (Supervisor) & Shuo Wang (Supervisor) |