摘要
信号集中监测(CSM)系统通过对无绝缘轨道电路(JTC)指定位置的数据采集,实现了对JTC的故障诊断。为进一步提高CSM系统的诊断水平,提出基于随机森林的JTC故障智能诊断方法。首先,通过传输线理论,对CSM系统所采集的JTC数据进行建模,并利用JTC半实物仿真平台对模型进行验证;然后,采用故障注入技术,利用所建模型,对JTC的典型故障模式下的CSM监测数据进行仿真,以此建立故障特征集;最后,制定包含再训练过程的JTC故障智能诊断策略,基于随机森林模型设计相应的故障诊断算法。实验结果表明,所提方法基于CSM,在不增加采集点的情况下,能够准确定位JTC上31种典型故障,且对未知故障具有智能诊断和学习能力,故障定位准确、识别度高、适应性强。
Currently,the Centralized Signal Monitoring(CSM)system can diagnose the faults of the Jointless Track Circuit(JTC)by collecting the data at the specific locations of the JTC.An intelligent fault diagnosis method for JTC was proposed based on the random forest model to further improve the diagnostic performance of the system.Firstly,the JTC data collected by the CSM system was modeled through the transmission line theory,and the models were verified by the JTC semi-physical simulation platform.Secondly,the fault injection technology on the built models was used to simulate the CSM monitoring data under different typical JTC failure modes so as to establish the fault feature set.Finally,the intelligent fault diagnosis strategy for JTC including the retraining process was developed,and the fault diagnosis algorithm was designed based on the random forest model.Experiments show that the algorithm in this paper based on CSM can locate 31 typical JTC faults precisely without increasing the measurement points in the CSM,and also has intelligent diagnostic and learning abilities for unknown faults,with high accuracy,high recognition and strong adaptability.
作者
刘一博
赵林海
LIU Yibo;ZHAO Linhai(School of Electronic and Information Engineering,Beijing Jiaotong University,Beijing 100044,China)
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2021年第11期78-87,共10页
Journal of the China Railway Society
基金
国家自然科学基金(61490705)。
关键词
无绝缘轨道电路
信号集中监测系统
故障诊断
随机森林
jointless track circuit
centralized signal monitoring system
fault diagnosis
random forest
