智能化高速列车健康管理关键技术研究

The high-speed train health management is an important method to ensure the train safety with a high operational speed, it has become a key research in the field of maintenance and repair of the high-speed trains. Some preliminary results have already been applied. However, the existing technology is still facing some problems, such as incomprehensive state detection, disability of detecting tiny fault, low accuracy of fault location, and the low rate, small capacity, limited expansion capability of the detection network, etc. Therefore, the project aims at the key technology of the high-speed train health management, and the research content includes that the intelligent state monitoring and fault location method are studied, and the key technical scheme of the state monitoring data's effective extraction, ault diagnosis prediction and noise restrain are puts forward respectively; the high-speed train communication network(TCN) technology based on Ethernet and the topology structure and the communication protocol stack of the high-speed TCN based on Ethernet are studied as well, the communication scheduling method of the real-time virtual layer in the improved protocol is put forward, and the corresponding simulation calculation and tests are realized. This research can improve maintenance and repair of high-speed train with a new theoretical method and technical support in our country.

高速列车健康管理是保障高速安全运营重要手段，成为高速列车维护维修领域的研究重点，一些成果已初步得到应用。然而，现有技术仍面临状态检测不全面、细微故障无法检测、故障定位准确率低，以及检测网络通信速率底、容量小、扩展能力有限等问题。为此，本项目针对高速列车健康管理关键技术进行研究，内容包括：研究智能化状态监测与故障定位方法，提出状态监测数据有效提取、故障与预测、噪声抑制等关键技术方案；研究基于以太网的高速列车通信网络（TCN）技术，提出基于以太网的高速列车通信网络网络拓扑结构及通信协议栈，改进协议中实时虚拟层的通信调度方式，并进行相应的仿真计算和测试。该项研究可为我国高速列车的维护和维修工作提供新的理论方法和技术支撑。

高速列车健康管理是保障高速安全运营重要手段，成为高速列车维护维修领域的研究重点，一些成果已初步得到应用。然而，现有技术仍面临状态检测不全面、细微故障无法检测、故障定位准确率低，以及检测网络通信速率底、容量小、扩展能力有限等问题。.为此，本项目针对高速列车健康管理关键技术进行研究，内容包括：研究智能化状态监测与故障定位方法，提出状态监测数据有效提取、信息融合、故障诊断、故障预测等关键技术方案。采用HMM算法对车载安全计算机平台的平均故障识别率89%以上，采用DTW算法对列车轴温识别的最大响应时间小于0.4s，误报率低于3%；.研究基于以太网的高速列车通信网络（TCN）技术，提出基于以太网的高速列车通信网络网络拓扑结构及通信协议栈，改进协议中实时虚拟层的通信调度方式，并进行相应的仿真计算和测试。结果表明，基于优先级的调度算法在损失非实时性数据传输时延的条件下，保证了实时性数据传输延迟为251μs，比FCFS算法降低了20μs。该项研究可为我国高速列车的维护和维修工作提供新的理论方法和技术支撑。