基于城轨列车在途监测数据的安全预测系统开发

发布时间：2018-05-22 16:05

本文选题：数据挖掘 + 城轨列车　；参考：《北京交通大学》2017年硕士论文

【摘要】：近年来随着我国城市轨道交通行业的快速发展,列车的运行安全问题受到了行业的广泛关注。为了保证列车的运行安全,在列车运行中发现并解决故障,本文以城市轨道行业为背景,使用基于在途列车状态的数据分析与挖掘技术解决列车运行中的安全问题。在现有的车辆维保生产管理系统中,使用该系统的列车在途运行数据,运用数据挖掘技术进一步开发子系统-安全预测系统。系统主要包括故障统计、故障关联分析、故障识别和故障预测四个功能模块,本文的主要工作包括:(1)按照系统的设置和用户的要求,使用原始数据表在数据库中建立相应的故障对应关系表,根据不同的条件,在对应表中对故障数据进行统计,生成故障统计图。(2)使用Hadoop对列车运行时产生的故障数据进行数据分析和挖掘,得出故障的属性和故障之间的关联规则,生成关联规则数据表。(3)使用关联规则进行故障识别,在故障表中找到与异常数据有关的故障,在关联规则数据库中查找异常数据的属性与故障的关联规则,对异常数据进行判断,在确认故障隐患后,生成故障隐患单。(4)综合相关数据,运用回归分析知识对故障的发展趋势进行预测,构造故障属性之间的回归方程,建立回归模型,通过回归模型对故障隐患进行预测。安全预测系统将数据挖掘技术引入车辆的维保管理系统中,实现了对列车运行状态的实时监测,经过系统对异常数据的识别和预测,及时通知用户对故障隐患进行处理,保证了列车的运行安全。系统测试结果表明安全预测系统的功能要求已经基本实现,车辆维保生产管理系统已经实现部分功能,相关版本已在广州地铁的具体项目中进行测试验收,软件相关后续开发工作也在进行中。
[Abstract]:In recent years, with the rapid development of urban rail transit industry in China, the safety of train operation has been widely concerned by the industry. In order to ensure the safety of train operation and to find and solve the trouble in train operation, this paper takes the urban rail industry as the background, and uses the data analysis and mining technology based on the state of the train on the way to solve the safety problem in the train operation. In the existing vehicle maintenance and maintenance production management system, the data of train running in the system is used to further develop the subsystem, the safety prediction system, by using the data mining technology. The system mainly includes four function modules: fault statistics, fault correlation analysis, fault identification and fault prediction. Using the original data table to establish the corresponding fault correspondence table in the database, according to the different conditions, the fault data in the corresponding table is counted, Using Hadoop to analyze and mine the fault data generated by train operation, to get the association rules between fault attributes and faults, and to generate association rules data table. 3) to identify faults using association rules. Find the fault related to the abnormal data in the fault table, look up the attribute of the abnormal data and the association rule of the fault in the association rule database, judge the abnormal data, after confirming the hidden trouble of the fault, (4) synthesizing relevant data, using regression analysis knowledge to predict the trend of fault development, constructing regression equation between fault attributes, establishing regression model and predicting hidden trouble through regression model. The safety prediction system introduces data mining technology into the maintenance management system of the vehicle, realizes the real-time monitoring of the train running state. After the system recognizes and predicts the abnormal data, the users are informed in time to deal with the hidden trouble. The running safety of the train is ensured. The system test results show that the functional requirements of the safety prediction system have been basically realized, the vehicle maintenance and production management system has realized some functions, and the related versions have been tested and accepted in the specific projects of Guangzhou Metro. Software related follow-up development work is also under way.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;TP311.52

【参考文献】