多物理场作用下的轨道交通牵引传动系统健康监测研究

A traction drive system is one of the crucial systems to ensure the safe operation of rail transportation equipments/systems. Health is the eternal pursuit of a reliable system and a real-time health monitoring technology is the core technology to guarantee that the traction drive systems will always be in the reliable working state. Owing to the lack of health monitoring mechanism for system-level real-time state in rail transit traction drive systems, and difficulties of health degradation modeling for the key components/devices and health perception factor mining for the systems, the modeling methodology for the health degradation of the key components/devices will be studied in multi-physics fields of complex conditions. Furthermore, the mining methodology for system-level health state perception factor and the assessment methodology based on the deep confidence network for the health degradation state of the key components/devices will be proposed. In addition, a health monitoring mechanism and the hardware-in-the-loop simulation platform of the traction drive systems will be investigated and constructed for the simulating and experimental validation of proposed real-time health monitoring methodology in multi-physics fields of complex conditions. Finally, a set of methods for real-time health monitoring will be established for the rail transit traction drive systems. These studies will provide a new mode for health monitoring of rail transit equipment/systems, as well as technical support for further realization of remaining life forecasting and health management. This work will be of great scientific significance and significant application value to enhance the service capability of the system, ensure the safe operation and reduce the maintenance cost.

牵引传动系统是确保轨道交通装备/系统安全运行的关键系统之一，健康是永恒的追求，健康监测是保证系统可靠工作的核心。本课题针对轨道交通牵引传动系统缺乏健康监测机制，器件健康退化建模、系统健康感知因子挖掘、实现系统级健康状态监测难等问题，研究多物理场下复杂服役工况的关键器件/部件健康退化机理建模方法；研究不同服役工况系统健康状态感知因子挖掘方法；研究基于自适应的系统健康状态监测方法和基于深度置信网络的关键器件/部件健康退化状态评估方法；构建牵引传动系统健康监测机制与系统仿真平台，用于多物理场下复杂服役工况的系统健康监测方法及其实现技术的验证，从而形成面向主动安全保障的牵引传动系统健康监测理论方法和实现技术，为轨道交通装备/系统的健康监测提供新的解决模式，为进一步实现剩余寿命预测与健康管理提供技术支撑，对增强系统服役能力、确保运行安全、降低维护成本具有十分重要的科学意义和重大的应用价值。

牵引传动系统是确保轨道交通装备/系统安全运行的关键系统之一，健康监测是保证系统可靠工作的核心。本课题围绕轨道交通牵引传动系统缺乏健康监测机制、健康退化建模难、系统健康感知因子难以获取等问题，开展了相关的研究，取得了如下重要成果：①提出了多物理场作用下复杂服役工况的牵引传动系统健康状态建模技术，搭建了牵引传动系统健康退化/失效的机理和信号模型，实现了传感器、变流器、牵引电机性能退化/失效的模拟；②提出了不同服役工况下牵引传动系统健康状态感知因子挖掘技术，构建了时序型、趋势型、分布型故障演化特征库，获取了较为敏感反映牵引传动系统健康退化/失效状态的特征，为健康状态监测奠定基础；③提出了自适应的牵引传动系统健康状态监测方法，实现了牵引传动系统运行状态的实时准确监测；④研制了面向实时健康状态监测的轨道交通牵引传动系统半实物仿真平台，对所提方法进行了测试与验证，为健康监测算法有效应用于牵引传动系统安全运行提供重要理论、技术支撑和平台环境。在项目资助下，发表论文27篇，其中SCI收录12篇、EI收录14篇；申请国家发明专利17项，其中授权10项；获批软件著作权1件。培养副教授1人，博士2名，硕士6名；举办学术会议7次；课题组成员参加10次故障诊断相关国际学术会议（共37人次）。项目执行期间，获批与中车株洲电力机车研究所有限公司合作共建“轨道交通节能控制与安全监测”湖南省重点实验室。