高速列车电力电子牵引变压器复合故障在线诊断方法研究

The traction drive technology based on power electronic transformer (PET) is one of the development trends of the future high-speed train, which are of high efficiency and energy conservation. Due to the addition of a large number of power modules and sensors, the failure rate of the traction drive system is sharply increased. Moreover, it is indeed a harsh operating environment in the traction drive system, for instance, rapid variation of the traction network voltage, time-varying of the control parameters and circuit parameters, switching between the traction mode and regenerative braking mode, and multiple fault modes. Thus, it is a severe challenge to ensure the safe and stable operation of the traction drive system. The project focuses the PET-based traction drive system, including four essential problems: multi-fault modes and effects analysis of the power electronic traction transformer (PETT), multi-fault mechanisms and propagation laws, multi-fault features coupling laws, and fast online diagnosis methods. The PETT multi-fault modes focus on open-circuit fault and short-circuit fault of power module, gain fault and offset fault of sensor, and compound fault of power module and sensor. As a result, the detail research contents including: the accurate dynamic fault model and database of the PETT, the multi-fault characteristics coupling laws, the propagation ways and effect laws of the multi-fault, the feature extraction and feature selection methods, and the diagnosis methods of PETT multi-fault based on time-adaptive data-driven technologies. Therefore, this project can provide an important theoretical basis and technical support for the safe and reliable operation of future high-speed train.

基于电力电子变压器的牵引传动技术是未来高速列车高效、节能牵引传动系统发展方向之一。由于新增大量功率模块和传感器等部件，牵引传动系统故障率增高，保障其安全稳定运行面临严峻挑战。项目以电力电子变压器牵引传动系统为研究对象，针对有潜伏性的故障模式，通过对多输入条件、多参数时变、多工况状态和多故障模式等多边界条件耦合影响下，电力电子牵引变压器多功率模块开路和短路故障、多传感器增益和偏移故障，及功率模块与传感器同时故障等复合故障产生的内在机理、传播和影响评估、特征耦合规律和在线诊断等核心问题的研究，建立电力电子牵引变压器精确动态故障模型和数据库，揭示不同边界条件综合影响下电力电子牵引变压器复合故障特征耦合机理、传播方式和影响规律，分析故障特征解耦、提取和选择方法，提出基于时间自适应数据驱动的电力电子牵引变压器复合故障在线诊断方法，为保障高效节能的高速列车安全可靠运行提供重要的理论基础和技术支持。

本项目以电力牵引传动系统中的功率模块和传感器为研究对象，针对有潜伏性的故障模式，通过对多输入条件、多参数时变、多工况状态和多故障模式等多边界条件耦合影响下，多功率模块开路和短路故障、多传感器增益和偏移故障，及功率模块与传感器同时故障等复合故障产生的内在机理、传播和影响评估、特征耦合规律和在线诊断等核心问题的研究，建立了电力牵引传动系统精确动态故障模型和数据库，揭示了不同边界条件综合影响下复合故障特征耦合机理、传播方式和影响规律，分析故障特征解耦、提取和选择方法，提出了多种基于数据驱动的复合故障在线诊断方法，并建立了多种人工智能技术在电力电子领域应用方案，为保障高效节能的高速列车安全可靠运行提供了的理论/数据基础和技术验证手段。基于该项目研究，共发表或录用SCI期刊论文8篇，EI期刊论文2篇；申请发明专利4项，授权发明专利4项；参加国内外学术交流活动10余人次，培养研究生11名。