摘要
受电弓-接触网系统作为电气化铁路的重要组成部分,一旦发生故障将严重影响弓网正常受流及行车安全,因此对接触网故障的预警就尤为重要。本文基于接触网结构模型及动力学特性,建立了垂向弓网耦合模型,在不同弓网动力参数下,对受电弓振动响应进行了仿真与分析。为了在接触网可能发生故障时提出预警,提出了一种基于GNG聚类与LS-SVM的接触网故障预警方法。首先运用GNG聚类算法对正常状态数据分类,得到若干聚类点,并计算当前状态与聚类中心的距离得出相似度趋势;同时构建历史矩阵,根据LS-SVM回归模型的回归结果得出预测差值,将相似度趋势和预测差值结合得到故障预警风险系数,对弓网故障状态提出预警。针对接触网正常状态和故障状态,做出了有效预警,表明了该方法的实用性。
The pantograph-catenary system is an important part of the electrified railway.Once a failure occurs,it will seriously affect the normal flow of the pantograph-catenary system and driving safety.Therefore,the early warning of catenary failure is particularly important.Based on the catenary structure model and dynamic characteristics,a vertical pantograph-network coupling model is established.Under different pantograph-network dynamic parameters,the vibration response of the pantograph is simulated and analyzed.In order to provide early warning when the catenary may fail,a warning method for catenary failure based on GNG clustering and LS-SVM is proposed.First,use the GNG clustering algorithm to classify the normal state data,obtain a number of clustering points,and calculate the distance between the current state and the cluster center to obtain the similarity trend;at the same time,construct the historical matrix,and obtain the result according to the regression result of the LS-SVM regression model.The prediction difference is calculated,and the similarity trend and the prediction difference are combined to obtain the failure warning risk coefficient,and an early warning is provided for the failure state of the pantograph and network.According to the normal state and failure state of the catenary,the results give an effective warning,which shows the practicability of the method.
作者
汤路
杨俭
宋瑞刚
袁天辰
TANG Lu;YANG Jian;SONG Ruigang;YUAN Tianchen(School of Urban Railway Transportation,Shanghai University of Engineering Science,Shanghai 201620,China)
出处
《智能计算机与应用》
2020年第12期6-11,共6页
Intelligent Computer and Applications
基金
国家自然科学基金(面上项目)(51575334)。
