基于灰关联的道岔故障诊断方法  

Method of Turnout Fault Diagnosis Based on Grey Correlation Analysis

作　　者：赵林海[1] 陆桥[1] 

机构地区：[1]北京交通大学电子信息工程学院,北京100044

出　　处：《铁道学报》2014年第2期69-74,共6页Journal of the China Railway Society

基　　金：铁道部科技研究发展计划(2011X021-C)

摘　　要：本文基于道岔转辙机典型故障下的动作功率曲线变化规律,提出一种基于灰关联的道岔故障诊断方法。通过采用Fisher准则对相应故障模式所对应的特征向量进行选择,通过优化分辨系数,并计算待测样本与各故障模式特征间的灰关联度,将满足门限要求的最大灰关联度所对应的故障模式作为待测样本的故障诊断结果。实验表明,该方法无需训练即可对道岔转辙机故障进行诊断,且有较好的适应性和较高的准确性。On the basis of the changing rule of the power curve corresponding to the typical faults of the switch machine , the turnout fault diagnosis method was proposed by grey correlation analysis . Using the Fisher crite-rion to select feature vectort of the corresponding failure mode , optimizing the distinguishing coefficients and calculating the gray relational degrees of the features between the test sample and each failure mode , the diag-nosis result of the test sample was obtained according to the failure mode with the maximum gray relational de-gree meeting the threshold requirements . Experiment results show that the method can achieve fault diagnosis without the need of sample training and also has good adaptability and high accuracy .

关 键 词：道岔 转辙机 故障诊断 FISHER准则 灰关联 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] U298[自动化与计算机技术—计算机科学与技术]