高速列车电力牵引系统的故障诊断研究

The electric traction system of electric locomotive is the power source of trains, and it is also the high risk area of locomotive fault. Its security and stability is important for the operation of railway. This project is going to research the fault diagnosis theory of electric traction system of electric locomotives, and try to achieve a practical breakthrough on fault diagnosis theory and application of locomotive electric traction system. Then a proper area divide method for distribute fault diagnosis of locomotive electric traction system will be proposed to solve many problems of the large-scale systems such as massive calculation problem, complex coupling relationship of parameters, bad real-time performance and so on. Several typical key devices will be chosen, the causes and state characteristic of their faults are to be analyzed so as to find out the suitable fault diagnosis method which fulfill their structure and characteristic. By doing this, fault model is no need to be precise, signal matches status easily, the resource of fault is less harder to be positioned, so the accuracy and reliability of fault diagnosis system for locomotive electric traction system can be advanced. Developing the train communication network to solve low efficiency of information sharing which is caused by different communication modes. By putting all above researches together, this project try to make substantive breakthrough in basic theory and application technology of security forecast and control for the electric traction system; further perfect and improve the fault diagnosis technology for locomotive electric traction system, and support our nation's railway development.

电力牵引系统既是高速列车的动力来源，也是列车上故障的高发区，电力牵引系统的正常工作，是高速铁路安全正点运行的重要保障。本项目通过对列车电力牵引系统故障诊断方法的研究，提出适合铁路高速列车系统分布式故障诊断方法的系统分区方法，解决列车故障诊断计算量大、变量耦合关系复杂、实时性差的问题；选取具有代表性的关键设备，深入分析其故障成因及状态特征信号，以期创建符合其结构和特征的故障诊断方法，从而较好解决故障模型难以确定、信号与状态的关系难以对应、故障源难以定位等问题，提高高速列车系统状态监测及故障预警系统的准确性和可靠性。开展列车通信网络的研究，解决目前车载设备通信制式不统一所带来的信息共享困难问题。结合以上研究内容成果，力争在高速列车电力牵引系统的安全性预测与控制的基础理论和应用技术上取得实质性突破，完善和提高我国高速列车电力牵引系统故障诊断技术，为铁路的发展保驾护航。

本项目通过对列车电力牵引系统故障诊断方法的研究，提出了适合铁路高速列车系统的基于因果模型的分布式故障诊断方法的系统分区方法，解决列车故障诊断计算量大、变量耦合关系复杂、实时性差的问题；针对牵引变压器、牵引变流器、车载整流电源、牵引电机等关键电气设备，深入分析其故障成因及状态特征信号，各提出了符合其结构和特征的故障诊断方法：对于变压器提出了一套完整的电力机车牵引变压器DGA诊断流程和一种基于双绕组变压器输出电流(i2)和输入输出电压差(Δv)之间的关系轨迹“(i2−Δv)”的交流机车/动车组牵引变压器绕组状态在线监测方法；对于牵引变流器，提出了一种非侵入式的开路故障诊断和定位方法，具有很强的适应性，可以在一个载波周期内定位故障；针对车载整流电源，提出了一种不增加变流器传感器和复杂程度的，可实现任一开关管开路的快速诊断/定位的诊断方法，并提出一种容错控制策略，可保持故障整流器的全负载输出；针对牵引电机轴承故障提出一种新的基于多隐层马尔可夫模型与蚁群聚类算法(ACC) 和神经网络相结合的方法；针对牵引电机齿轮箱提出了一种新的机车齿轮箱振动信号去噪的模型及算法；从而较好解决了故障模型难以确定、信号与状态的关系难以对应、故障源难以定位等问题，提高了高速列车系统状态监测及故障预警系统的准确性和可靠性。此外，本项目还开展了列车通信网络方面的研究，所提出的方法对于解决目前车载设备通信制式不统一所带来的信息共享困难的问题，具有一定的参考价值。综上所述，结合以上研究内容成果，本项目在高速列车电力牵引系统的安全性预测与控制的基础理论和应用技术上取得了实质性突破，成果对于完善和提高我国高速列车电力牵引系统故障诊断技术，为铁路的发展保驾护航具有意义。