欢迎访问铁路通信信号工程技术,今天是 English

铁路通信信号工程技术 ›› 2021, Vol. 18 ›› Issue (5): 59-65.DOI: 10.3969/j.issn.1673-4440.2021.05.011

• • 上一篇    下一篇

基于生存分析的AFC终端设备无故障运行时间预测

李 波1,张炳森2,张 宁3   

  1. 1.中铁第四勘察设计院集团有限公司,武汉 430063;
    2.北京城建设计发展集团股份有限公司,北京 100037;
    3.东南大学智能运输系统研究中心轨道交通研究所,南京 210018
  • 收稿日期:2020-08-27 修回日期:2020-08-29 出版日期:2021-05-25 发布日期:2021-08-16
  • 基金资助:
    中国铁建股份有限公司科研计划课题(17-C41)

Prediction for Fault-free Operation Time of AFC Terminal Equipment Based on Survival Analysis

Li Bo1,  Zhang Bingsen2,  Zhang Ning3   

  1. 1. China Railway Siyuan Survey And Design Group Co., Ltd, Wuhan    430063, China;
    2. Beijing Urban Construction Design & Development Group Co., Ltd., Beijing    100037, China;
    3. ITS Rail Transit Research Institute, Southeast University, Nanjing    210018, China
  • Received:2020-08-27 Revised:2020-08-29 Online:2021-05-25 Published:2021-08-16

摘要: 提出一种基于加速失效模型的无故障运行时间预测方法以改善AFC终端设备的维护管理水平。分析人、环境及设备自身状况特征因素对设备故障的影响,在生存分析的理论基础上研究构建基于风险的无故障运行时间预测模型。采用南京地铁油坊桥车站一年的AFC设备故障数据,以机械类故障为例进行估计和验证。根据赤池信息量准则,选择Weibull分布建立无故障运行时间预测模型,以平均绝对百分比误差(MAPE)作为模型性能评价指标,结果显示该方法具有优秀的预测性能,相比基于回归分析的方法表现出明显的优势。本研究结果可为设备维修管理提供有价值的参考。

关键词: 轨道交通, 生存分析, 区间预测, 加速失效模型

Abstract: In order to improve the maintenance management level of AFC terminal equipment, a prediction method of fault free operation time based on accelerated failure model is proposed. This paper analyzes the influence of human, environment and equipment condition characteristics on equipment failure, and constructs a risk-based prediction model of fault free operation time based on the theory of survival analysis. The AFC equipment fault data of Youfangqiao station of Nanjing subway for one year is used to estimate and verify the mechanical fault. According to the Chi Chi information criterion, Weibull distribution is selected to establish the prediction model of fault free operation time, and the mean absolute percentage error (MAPE) is used as the performance evaluation index of the model. The results show that the method has excellent prediction performance, and has obvious advantages over the method based on regression analysis. The research results can provide valuable reference for equipment maintenance management.

Key words: rail transit, survival analysis, interval prediction, accelerated failure model

中图分类号: