Research On The State Subdivision And Fault Prediction Of Public Traffic Vehicles Based On Big Data

With the rapid development of public transport,the operating intensity of public transport vehicles has been increasing,and there have been a variety of unexpected problems in public transport vehicles during high-intensity and overloaded operation.The bus group company produces a variety of data sources in the process of operation and maintenance.How to effectively use data assets to establish an analysis system of bus big data and guide the safety and maintenance of public transport vehicles is very important.First,this paper analyzes the literature of vehicle condition and fault prediction.In view of the current theory and practice of health prediction based on big data is still at the initial stage,which requires more theoretical exploration and practical application.Based on the vehicle data collected by vehicles of Beijing Public Transport Group,combined with maintenance data and external weather influences,this paper uses data mining and big data analysis methods to select bus data to study the status of vehicle faults,and get vehicle status subdivision and prediction of key faults.Secondly,this article introduces the composition of bus big data and the process of sorting and preprocessing of bus data.From the perspective of vehicle value,we choose the degree of nearness,frequency,and time as the analysis indicators of the state of bus.The k-means clustering method is used to subdivide bus vehicles and assign different vehicle labels respectively.During the subsequent vehicle maintenance and repair,you can refer to this tagging information.According to different vehicle optimization and maintenance strategy,so as to provide some references and suggestions for vehicle maintenance and repair units.Then,according to the type of bus fault information,the short message text mining technology is used to extract the characteristics of bus fault information,and the key fault location of the bus is obtained.The engine fault is the most significant fault feature,followed by the brake status,and it will accompany the occurrence of the leakage phenomenon.In the selection of fault feature parameters,six characteristic parameters have been determined according to the grey correlation analysis of the characteristic parameters of engine fault,and the change of these parameters and the fault location can be focused on in the later period of vehicle maintenance and maintenance,and the maintenance time is reduced.Finally,the paper designs the fault prediction for the key engine parts of the bus,and respectively uses two algorithms,support vector machine and logistic regression.The main characteristic parameters of the engine are used as the input index,and the forecast of the bus engine fault is realized by combining the weather characteristics.The comparison analysis shows that the support vector machine has a higher accuracy compared with the logistic regression,and has a good superiority on the two classification problem.Initially establish a set of practical application system for analysis and prediction of public transport vehicles based on Beijing Bus Group Corporation.Eventually,the purpose of effective prevention and reduction of bus fault and accident rate is to optimize the system of bus maintenance and maintenance,and reduce the adverse effects of vehicle failures on the normal road traffic.