高速列车牵引电机转子系统耦合故障的混合诊断方法研究

Rotor system, the core equipment of traction motor of China Railway High-speed(CRH) train, has all along subjected to the high-speed variable conditions and generates the multi-degree-of-freedom movements of traversing, ups and downs, stretching and so on by strong impact. They make the dynamic response of system that is more complex and results in a variety of faults. This will cause system failure or even major fatal accidents. Therefore, the rotor system of traction motor of CRH train under the special service environments is selected as research object in this project. The type, quantity, location and severity of the coupling faults are effective diagnosing goals. The dynamics model of rotor system under coupling faults based on fractional calculus is established. The association relations between fault stimulation and dynamics behavior of rotor system are studied, and the fault dynamic evolution mechanism is clarified. An improved empirical wavelet transform(EWT) based on fractional Fourier and optimal wavelet basis is proposed in order to improve the adaptive decomposition of vibration signal and propose a qualitative diagnosis method of coupling faults of rotor system. Then a quantitative diagnosis method of coupling faults of rotor system based on combining the harmonic balance theory, frequency spectrum energy analysis with the constructed dynamics model is proposed. And the qualitative diagnosis method and quantitative diagnosis method are combined in order to construct a new hybrid diagnosis model and method of coupling faults of rotor system, which can comprehensively diagnose the coupling faults of rotor system. The research results provide an effective method for diagnosing faults of traction motor of CRH train, and the research project has important theoretical significance and application value.

转子系统作为高速列车牵引电机的关键部件，长期处于高速变工况状态，并承受强冲击而产生横移、沉浮、伸缩等多自由度运动，使系统动态响应愈加复杂，导致多种故障发生，造成系统失效甚至发生重大事故。为此，本项目针对特殊服役环境下高速列车牵引电机转子系统，以有效诊断耦合故障的类型、数量、位置和严重程度为研究目标，建立基于分数阶微积分的转子系统动力学模型，探索故障激励与系统动力学行为之间的关联关系，揭示故障动态演化规律；研究基于分数阶Fourier变换和最优小波基的经验小波变换改进方法，以提高振动信号分解的自适应性，提出转子系统耦合故障定性诊断方法；结合动力学模型，研究基于谐波平衡理论和频谱能量分析的转子系统耦合故障定量诊断方法；由此构建一种定性和定量相结合的耦合故障混合诊断模型与方法，实现转子系统耦合故障的有效诊断。本项目为高速列车牵引电机转子系统的故障诊断提供一种新思路，具有重要的理论意义和应用价值。

本研究项目在综合考虑了高速列车牵引电机转子系统存在的旋转、横移、沉浮、伸缩等多自由度运动，以及高速变工况运行过程中振动的非线性特性等影响因素前提下，基于现代非线性动力学理论，引入分数阶微积分，采用数值仿真技术，建立了无故障和含裂纹、不对中等单故障和碰磨-不对中等耦合故障的牵引电机转子系统非线性动力学模型，探讨了单故障和耦合故障引起的激励与转子系统非线性动力学行为之间的关联关系，为高速列车牵引电机转子系统的故障诊断提供了模型基础。针对经验小波变换Fourier频谱自适应分割导致处理强噪声、不稳定信号时存在的不足，研究并提出了基于尺度空间阈值法的改进经验小波变换方法和基于最大最小长度曲线的改进经验小波变换方法，将故障振动信号分解为具有物理意义的一组经验模态，并将改进的经验小波变换方法用于轴承故障振动信号的分析，同时还研究并提出了一种基于改进变分模态分解的振动信号分析方法，实现了牵引电机轴承故障的定性诊断。针对轴承故障不同损伤程度，引入谐波分量及频谱能量思想，研究并提出了基于高阶差分数学形态谱熵的电机轴承故障定量诊断方法；引入深度置信网络、高斯卷积深度置信网络等深度学习方法，实现了牵引电机轴承故障的定性定量诊断。在研究故障诊断技术的基础上，对电机轴承性能评估及寿命预测方法进行了深入研究，提出了一种能够有效反映不同轴承性能退化趋势的新特征指标，进而提出了基于新特征指标的轴承剩余寿命预测新方法，实现了轴承剩余寿命的精确预测。本项目的研究为高速列车牵引电机的健康监测、定性定量诊断与预测提供了理论基础与方法支撑。