基于金属磁记忆效应的高速列车轮对微小故障实时动态预测研究

Wheelset is the critical component of walking step of highspeed train and its running state is directly related to security of operation which catastrophic failure has been caused by incipient fault. At present, flaw detection pattern on wheelset is mainly in repair base which is called ground detection. The pattern can only be used for the flaw which has been macro existed afterwards and not for the incipient fault during operation and the trend of tracking of it. This is a big risk during operation. Aiming at the deficiency of the afterwards nondestructive testing, so Metal Magnetic Memory Method which has early diagnosis function for the incipient fault is used to predict the incipient fault for the realtime dynamic testing on the wheelset in service. The research is as following to solve the operation risk of the afterwards testing and improve the safety of highspeed train : Firstly, highspeed synchronous detection and tracking device should be designed for obtaining the realtime dynamic Magnetic Memory Information of wheelset and large capacity distributing network communication should be developed also. Secondly,the distribution characteristics described with more time-frequency characteristics should be extracted for the metal magnetic memory signal based on the sampling multi-wavelet transform precision for weak magnetic memory signal de-noising. Finally,the danger classes of incipient fault should be quantitatively assessed with small sample characteristics and relevance vector machine for inproving the real-time assessment round the accuracy and robustness of the incipient fault.

轮对作为高速列车走行部的关键部件，其运行状态直接关系到列车的安全运行，而由应力集中产生的微小故障是造成轮对突然失效的主要原因。目前我国无损探伤检测主要在维修基地或维修库内进行落地检测，只能被动事后检测已经存在的轮对裂纹，不能提前预测早期微小故障，并且无法实时跟踪运行期间萌生的微小故障，存在巨大运行风险性。因此，针对高速列车轮对事后无损检测的不足，将具有早期诊断功能的金属磁记忆方法应用到在役轮对的实时动态检测，实现微小故障的预测诊断。拟做如下研究：设计高速同步跟踪检测装置、大容量分布式网络通信获取实时动态的轮对磁记忆信号；基于非抽样多小波变换对微弱的磁记忆信号精确消噪，提取全面的时—频联合多特征量描述磁记忆信号的分布特征；采用小样本的相关向量机对多特征量和微小故障的危险等级进行定量评估，提高实时评估轮对微小故障的精确性和鲁棒性，解决事后检测带来的运行风险，提高高速列车运行的安全性。

轮对作为高速列车走行部的关键部件，其运行状态直接关系到列车的安全运行，而由应力集中产生的微小故障是造成轮对突然失效的主要原因。目前我国无损探伤检测主要在维修基地或维修库内进行落地检测，只能被动事后检测已经存在的轮对裂纹，不能提前预测早期微小故障，并且无法实时跟踪运行期间萌生的微小故障，存在巨大运行风险性。因此，针对高速列车轮对事后无损检测的不足，将具有早期诊断功能的金属磁记忆方法应用到在役轮对的实时动态检测，实现微小故障的预测诊断。拟做如下研究：设计高速同步跟踪检测装置、大容量分布式网络通信获取实时动态的轮对磁记忆信号；基于非抽样多小波变换对微弱的磁记忆信号精确消噪，提取全面的时—频联合多特征量描述磁记忆信号的分布特征；采用小样本的相关向量机对多特.征量和微小故障的危险等级进行定量评估，提高实时评估轮对微小故障的精确性和鲁棒性，解决事后检测带来的运行风险，提高高速列车运行的安全性。通过研究设计并研制了实时动态测试系统2套，高速通信控制器两套，并针对复杂环境中轮对的力磁关系进行了研究，对轮对材料所涉及到的静态力磁关系进行了量化研究、对轮对材料的疲劳力磁关系进行了量化研究和热磁关系的量化研究，这些数据将对更深入的对复杂环境的轮对寿命的预测研究提供数据支撑。