电动汽车用永磁同步电机转子温度分布建模方法和观测技术

The permanent magnet synchronous machine (PMSM) has been widely employed in electric vehicles thanks to its superior performance. However, its rotor permanent magnets (PMs) are quite sensitive to the ambient temperature and even suffer from irreversible demagnetization due to high temperature. Nevertheless, a hardware-based rotor temperature monitoring system will significantly raise the system cost and complexity while existing software-based temperature monitoring methods only can estimate the average temperature of rotor PMs. Therefore, it has become an urgent problem to develop software-based monitoring techniques for the observation of the rotor temperature distribution. Consequently, from the perspective of the multi-physical and multi-frequency analysis of PMSM system, this project is intended to explore and establish a set of modeling approaches to express the relationship between measurable signals and rotor temperature distribution, and build relevant mechanism models afterwards. Furthermore, according to the temporal, spatial, and frequency distribution based multiple signal processing and mathematical manipulation of the developed mechanism models, the observation models for state variables such as the average temperature, the maximum temperature differential, and the local temperature distribution of rotor PMs, will be built respectively. According to those observation models, relevant observers for the rotor temperature distribution will be eventually developed. A novel and generic theoretical framework for the modeling and observation techniques of rotor temperature distribution will be proposed in this project, which has a positive significance on reducing the cost and complexity of rotor temperature monitoring system and will also provide a new theory tool for overheating protection of PMSMs designed for electric vehicles.

永磁同步电机因其优越性能，被广泛应用于电动汽车。但是其转子永磁体对温度比较敏感，在高温下甚至会不可逆退磁。而转子温度硬件监测系统往往会增加系统成本和复杂度，同时现有软件监测方法又只能估算转子永磁体的平均温度。因此，如何研发能观测转子温度分布的软件监测方法，成为当前亟待解决的难题。为此，本项目首先从永磁同步电机系统多物理、多频谱分析的角度，探索建立一套建模方法来表述可量测信号和转子温度分布之间的关系，并建立起相应的机理模型；其次，提出对该机理模型进行基于时间、空间和频谱的多源信号处理和数学变换，建立起转子永磁体的平均温度、最大温度分布差异、局部温度分布等状态变量的观测模型；最后，根据上述观测模型设计相应的温度分布观测器。本项目将为转子温度分布建模和观测技术提出新型、通用化的理论架构，对于降低转子温度监测系统的成本和复杂度具有积极意义，并将为电动汽车用永磁同步电机过热保护技术提供新的理论工具。

基于永磁同步电机的电动汽车是国家当前重点发展的产业，但是电机转子永磁体温度问题制约了其动力系统的可靠性，因此需要高效的温度监测技术来确保其工作于安全温度范围。因此，本项目对永磁同步电机转子永磁体温度估测技术进行了深入研究。主要研究了电压源逆变器非线性因素补偿、永磁同步电机转子温度监测和建模方法、以及基于谐波分析的永磁同步电机匝间短路故障诊断方法。通过项目的研究，主要成果有：.1）提出了多种逆变器非线性因素提取和补偿的方法。所提方法不需要电机参数信息、额外硬件和离线实验，能够同时适应id=0和MTPA控制。通过提取逆变器非线性因素扰动电压项，然后由反馈控制器对其进行在线补偿，能够有效抑制逆变器非线性因素引起的扰动电压，并具有优良的稳态和动态性能。该研究可有效提高参考电压的精度，进而提高转子温度估算的精度。.2）提出了基于一阶槽谐波提取的转子永磁体温度估测方法，所提出的方法和传统的基于信号注入或电机参数的方法相比，不会给系统带来额外的扰动和损耗，也不用考虑逆变器非线性的影响，该方法在id=0和MTPA控制下均表现出优异的温度估测性能。.3）提出了多种永磁同步电机匝间短路故障诊断方法。所提方法能够在电机稳态和瞬态运行工况下检测出故障发生的初期、晚期和故障程度的变化过程，并且对绝缘退化过程具有更高的灵敏度，且不需要额外的传感器、电路、历史数据及电机参数。.4）项目已培养学生21名，其中博士研究生4人，硕士研究生17人；在国内外重要学术刊物上共发表中英文期刊论文21篇，其中IEEE期刊论文9篇，EI收录期刊4篇；另外还发表了3篇国际会议论文和授权发明专利5项。