摘要
针对轨道电路传统维修模式存在的效率低、维修不及时等问题,结合轨道电路的工作原理,分析了轨道电路的故障模式,提出了红光带故障诊断方法。采用随机梯度下降逻辑回归模型,建立了基于数据驱动的轨道电路故障智能预测及预警方法,以实现对具有递增或递减趋势的监测数据的预测及预警。以某站某轨道电路为案例,应用该方法进行趋势预测。试验结果表明:该方法对不同的轨道区段和不同的监测量均有较强的适用性,可同时对多个监测量的数值变化情况进行预测,实现对轨道电路的故障预警,提高轨道电路维修的及时性和效率。
In view of the problems of low efficiency and untimely maintenance of the conventional maintenance mode of the track circuit,combined with the working principle of the track circuit,the failure mode of track circuit is analyzed,and the diagnosis method of the′red band′fault is proposed.Using the stochastic gradient descent logistic regression model,a data-driven track circuit fault intelligent prediction and early-warning method is established to realize the prediction and early-warning of monitored data with increasing or decreasing trends.Taking certain track circuit in certain station as an example,this method is used to predict the trend.Experiments show that the method has strong applicability to different track sections and different monitoring variables and can simultaneously predict the future numerical changes of multiple monitor-ing variables,realizing the fault early-warning of track circuit,thereby improving the timeliness and efficiency of track circuit maintenance.
作者
纪玉清
欧冬秀
李永燕
JI Yuqing;OU Dongxiu;LI Yongyan(Shanghai Key Laboratory of Rail Infrastructure Durability and System Safety,201804,Shanghai,China;不详)
出处
《城市轨道交通研究》
北大核心
2022年第7期30-33,共4页
Urban Mass Transit
基金
国家重点研发计划基金项目(2018YFB1201403)。
关键词
轨道电路
数据驱动
随机梯度下降逻辑回归模型
趋势预测
故障预警
track circuit
data-driven
the stochastic gradient descent logistic regression model
trend forecasting
fault early-warning
