摘要
以高铁设备故障诊断为研究背景,提出一种基于样本证据距离和可信度函数的基本概率指派(BPA函数)的生成方法。根据所提出的概率赋值和传感器可信度生成方法,生成初始BPA。为了消除干扰影响,同时考虑证据冲突问题,引入冲突消解的方法,对初始BPA经过冲突阈值判别并进行相应的冲突消解。高速列车转向架故障诊断试验表明,所提出BPA生成的方法简单有效,能够处理冲突较大的数据。
Taking the fault diagnosis of high speed railway equipment as the research background, this paper presents a method to determine the basic probability assignment (BPA function) based on the distance and confidence function of sample evidence. According to the proposed methods to determine the probability assignment and the reliability of the sensor, the initial BPA is generated. In order to eliminate the influence of interference and consider the conflict of evidence at the same time, the method of conflict resolution is introduced. The initial BPA is judged by the conflict threshold and the corresponding conflict resolution is implemented. The fault diagnosis test of high speed train bogies shows that the proposed method for determining the BPA is simple and effective, and it can be used to handle data with major conflicts.
作者
朱菲
金炜东
ZHU Fei;JIN Weidong(School of Electrical Engineering, Southwest Jiaotong University, Chengdu Sichuan 610031, China)
出处
《中国铁路》
2018年第4期82-85,97,共5页
China Railway
基金
国家自然科学基金项目(61134002)
关键词
故障诊断
DS理论
基本概率指派
BPA函数
故障分类
转向架
fault diagnosis
DS theory
basic probability assignment
BPA function
fault classification
bogie