基于混合模型的道岔综合监测系统研究  

Research on Comprehensive Monitoring System for Switch Based on Hybrid Model

作　　者：曹峰[1] 张娟 CAO Feng;ZHANG Juan

机构地区：[1]南京铁道职业技术学院轨道交通工程实践中心,南京210031 [2]南京铁道职业技术学院通信信号学院,南京210031

出　　处：《铁道通信信号》2024年第1期45-51,共7页Railway Signalling & Communication

基　　金：江苏省大学生创新创业训练计划项目(YXKC2022008)。

摘　　要：道岔作为关键的铁路信号设备,也是铁路线路三大薄弱环节之一,其工作质量直接影响列车的运行安全。传统的道岔检测方法过分依赖人工经验,检测效率低下,难以应对现有铁路运行中行车速度快、发车密度高等对道岔维护所带来的严峻挑战,并且现有道岔监测也存在监测项目不全面等问题。为满足工电融合需要,开发了一套基于混合模型的道岔综合监测系统,使用卷积神经网络自动进行特征提取,以获取道岔状态,充分发挥深度学习的自动特征提取优势;采用支持向量机和向量域的混合算法,对正常/故障数据进行分类和异常检测,从而提高故障检测的准确率。测试结果表明:与现有人工巡检方法相比,该系统能够为相关人员提供精准、实时的道岔故障预警,提高维护效率,有效减少人力成本且降低道岔病害的发生概率。As a key railway signal equipment,the switch is also one of the three weak links in railway lines,and its work quality directly affects the safety of train operation.The traditional switch detection methods overly rely on manual experience and have low efficiency,which is difficult to cope with the severe challenges brought by the fast running speed and high departure density on the maintenance of switches in the existing railway operation.Moreover,the existing switch monitoring also has some problems such as incomplete monitoring items.In order to meet the needs of the combination of railway engineering and signalling&communication,a switch comprehensive monitoring system based on the hybrid model is developed.The convolutional neural network is applied for automatic feature extraction to explore the switch states,giving full play to the advantages of automatic feature extraction of deep learning.A hybrid algorithm of supporting vector machine and vector domain is used to classify normal and fault data and detect anomalies,so as to improve the accuracy of fault detection.The test results show that compared with the existing manual inspection methods,the system can provide accurate and real-time fault early warning for the personnel,improve the maintenance efficiency,effectively the labor cost and reduce the probability of switch faults.

关 键 词：混合模型 道岔 综合监测 支持向量机 支持向量域 

分 类 号：U284.7[交通运输工程—交通信息工程及控制]