Fault Identification And Health Status Evaluation Of Urban Rail Transit Signal Equipment

Urban rail transit is the established development strategy for my country’s transportation power building,and it is also an important link in the blueprint for modern transportation development in the new era.With the rapid development of urban rail transit and the continuous increase of train speed,higher requirements are put forward for the safe operation of signal equipment.As the key equipment of the signal system,the switch machine is in a complex environment,which leads to frequent mechanical failures.However,traditional failure handling methods cannot meet the needs of urban rail transit intelligent operation and maintenance.Therefore,the research on fault identification and health assessment methods of switch machines has important theoretical significance and application value for improving the efficiency of signal equipment fault handling and reducing maintenance costs.The main contents of this paper are as follows:(1)The structure and function of the switch machine are comprehensively analyzed,and the failure mechanism of its bearing components is analyzed.The ZD6 point machine is the most commonly used model in the urban rail signal system.First,its structure and functional principles are studied.Second,the core component of the point machine—the bearing failure mechanism and the vibration mechanism during operation are further discussed;Finally,the bearing vibration signal acquisition platform is introduced,and the 9 kinds of faults of the bearing and the vibration signal data set under normal conditions are introduced.(2)Taking switch machine bearings as the research object,the basic ideas and design methods of applying deep learning algorithms to fault identification are studied.By analyzing the basic ideas of the CNN algorithm and LSTM algorithm in deep learning,comprehensively integrating the CNN model’s implicit feature extraction and the characteristics of the LSTM model adapting to the learning time series,a design method for the CNN-LSTM combined model for fault identification is proposed.The fault identification of the track machine’s bearing lays the theoretical foundation.(3)On the basis of realizing the fault identification of the bearing of the switch machine,the health evaluation model of the bearing is established.Due to the mixed sensor data of the switch bearing,the sensor fusion model is introduced on the basis of the health state division,and the K-means algorithm is combined to eliminate redundant sensor data,and the mapping relationship between multi-dimensional sensor information and one-dimensional health value is realized;Refer to expert opinions and use the health value as the standard to divide the full life cycle status of the switch machine bearing into: health—good—attention—failure.(4)Finally,predict the remaining service life of the switch machine bearing under different health conditions.Because the health value degradation laws of different bearings in the same fault state are similar,the similar health value curve is established by establishing a similarity model to extract the life of the bearing corresponding to the bearing,and the weighted summation of the extracted life is weighted by the curve similarity to obtain the remaining life prediction result.(5)An example is used to verify the effectiveness of the fault identification model and health assessment scheme based on the CNN-LSTM algorithm.Taking the bearing vibration signal measured by the experimental platform of Case Western Reserve University in the United States as the research object,the fault identification accuracy of the CNN-LSTM algorithm is 99%,and the remaining service life prediction error of the bearing under good and careful conditions is within 11%.The research results of this paper enrich the methods of point machine fault identification and health assessment,and lay the foundation for the urban rail transit signal system fault identification and rapid response plan design.