高速铁路非正常事件时长分布及场景划分研究

A Study on Duration Distribution and Scenario Classification of Abnormal Events in High-speed Railway

为了提高高速铁路应急水平和实时调度能力,采用最大似然法、最小二乘法、矩估计法3种方法对各类型非正常事件的持续时长进行威布尔分布拟合,对拟合效果进行检验与比较,探究各类事件的持续时长特征;以故障发生时刻、持续时长和影响列车数为聚类指标采用模糊C-均值算法划分非正常事件场景。结果表明,矩估计法的威布尔分布模型拟合度最优,不同类型非正常事件持续时长分布差异大;故障持续时长越长,影响列车数越多,故障强度就越大,为非正常事件下列车运行调整提供前瞻性数据支撑。

In order to improve the emergency preparedness and real-time traffic control ability of high-speed railway,the maximum likelihood method,the least square method and the moment estimation method are used to fit the Weibull distribution of the duration for various types of abnormal events.The fitting effect is tested and compared to explore the duration characteristics of various events.Fuzzy C-mean algorithm is used to divide the abnormal event scene with the time of fault occurrence,the duration and the number of interrupted trains as the clustering index.The results show that the fitting degree of Weibull distribution model is optimal,the duration distribution of different types of abnormal events varies,and the longer the fault duration,the larger the number of trains interrupted,the greater the fault intensity.The results can provide prospective data support for train operation adjustment under abnormal events.