基于D-S证据理论高速列车横向失稳故障判别研究  

Study on the lateral instability fault recognition of high-speed train based on D-S evidence theory

作　　者：刘棋[1] 宁静[1] 叶运广 陈春骏[1] 

机构地区：[1]西南交通大学机械工程学院,四川成都610031

出　　处：《中国测试》2017年第7期103-107,共5页China Measurement & Test

基　　金：国家自然科学基金项目(51475387)

摘　　要：为解决高速列车发生横向失稳故障时,转向架的运行情况难以被单一传感器测量得到全面信息以及准确地反应等问题,提出利用多个加速度传感器组成多信息源网络系统,建立基于多信息源的高速列车横向失稳故障决策融合诊断系统。由于高速列车发生横向失稳故障存在复杂的轮轨耦合关系,导致列车横向失稳故障状态诊断难度大,基于此提出D-S证据理论方法融合系统中各个传感器中测量数据信息并应用于高速列车横向失稳故障状态判别。结果表明:基于D-S证据理论方法与任何单一传感器诊断结果相比,识别效果更好,对正常状态、小幅蛇行以及大幅蛇行故障状态的识别率分别达92.3%、82.89%、88.67%,证明该方法有效。For addressing the issue that the operation of bogie is difficult to be reflected by the measuring information of single sensor comprehensively and accurately when the lateral instability of high-speed train occurs, the multi-sources system established by more accelerometers sensors is proposed to build a high-speed train lateral instability fault decision fusion diagnosis system based on the multi -sources. The complex coupling relationship between the wheel and the rail exists when the lateral instability occurs, which will cause that lateral instability fault diagnosis conditions is hard. Therefore, the D -S evidence theory is used to fuse the measured data information of each sensor in the system and applied to identify high-speed train lateral instability fault conditions. The results show that the D -S evidence theory is more accurate than that of diagnosis results of any single sensor, in which the recognition rate of normal state, small hunting and criterion hunting achieves as high as 92.3%, 82.89%, 88.67% respectively. It proves the effectiveness of this method.

关 键 词：高速列车 横向失稳 D-S证据理论 决策融合 故障判别 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]