| High-speed trains are an important construction target for the strategic deployment of a powerful country in transportation.High-speed trains have become the main means of transportation for people to travel by virtue of their high efficiency and comfort.As a large-scale ground transportation system that is critical to the safety of passengers’ lives and property,the safety and reliability of high-speed train operation are of paramount importance.On-board equipment is an important equipment of the high-speed train operation control system,and its operation status directly affects the reliability of train operation.Therefore,it is necessary to carry out fault prediction research for on-board equipment to detect and eliminate potential faults in advance to ensure the safe operation of high-speed trains.Based on the problem of fault data analysis and fault prediction of high-speed train control on-board equipment,this paper first proposes a multi-dimensional fault data model for the complex fault data of high-speed train control on-board equipment,and conducts modeling and analysis on operating fault data;The time between failures is selected as the research object,and a method for predicting the time between failures based on the EMD algorithm is proposed to realize the failure prediction of vehicle equipment;The software of fault data statistical analysis and fault prediction is designed and implemented.The main work of this paper includes:(1)Aiming at the operation data of on-board equipment of high-speed railway train control,the multi-dimensional fault data model of on-board equipment is built based on OLAP technology,which solves the problems of standardized storage and management of the complex on-board equipment fault data,realizes the mining and analysis of the implied information of fault data,and provides a reliable data basis for subsequent fault prediction.(2)Aiming at the problem of fault prediction of on-board equipment,taking DMI unit as an example,the time between failures of DMI unit is selected as the research object,and the fault prediction reliability(FPR)evaluation index is proposed for the safety-critical train control system on-board equipment.Based on the nonlinear characteristics of fault interval time series,EMD decomposition algorithm was introduced to carry out time series decomposition.PSO-SVR algorithm,BP neural network and LSTM neural network algorithm were used to predict the decomposed items respectively.The prediction results were integrated,and RMSE and FPR were used to evaluate the prediction accuracy.The validity of LSTM neural network prediction model based on EMD algorithm and the feasibility of the evaluation index are verified.(3)In response to the needs for statistical analysis and failure prediction of on-board equipment fault data,based on the multi-dimensional fault data model and fault prediction research foundation of on-board equipment,this paper designs and implements the time-space characteristics analysis and fault prediction software of train control on-board equipment failure data.DMI,STU-V and VDX field fault data are used to verify the fault prediction method proposed in this paper.Finally,the function verification and analysis of the software are completed,and the high efficiency and intelligence of fault data analysis and fault prediction are realized.The paper proposes and constructs a multi-dimensional model of on-board equipment fault data,uses OLAP technology to analyze the Spatio-temporal characteristics of fault data,and introduces a LSTM neural network prediction model based on EMD algorithm to predict the time between failures of on-board equipment.On this basis,the time-space characteristics analysis and fault prediction software of train control on-board equipment fault data was designed and implemented,and the software functions were verified using on-board equipment field fault data,which proved that the software has high practical value.This thesis contains 7 figures,11 tables and 73 references. |
PDF Full Text Request