Study On Fault Identification Of Orbit Structure In Service State Based On Intelligent Algorithm

With the growth of China’s economic level and the construction of railway lines,the mileage of railway operations has increased year by year.Monitoring railway bridges and railway tunnels is an important means to ensure the safety of railway operations.Intelligent fault identification research has been used more and more widely in the safety monitoring of railway bridges and railway tunnels.At the same time,intelligent public transportation is one of the development directions in the era of artificial intelligence and big data,which is in line with the outline of scientific and technological development planning with intelligent processing of transportation information management systems as a key priority research theme.With the increase of railway operating mileage,the improvement of train speed and the dynamic interaction between wheels and rails,the safety of train operation cannot be guaranteed,and the comfort of passengers is also affected accordingly.At the same time,the increasing wheel and rail forces tend to lead to the destruction of engineering structures such as track,so it is urgent to monitor the service status of track structures in time.In order to solve this problem,this paper puts forward a research method of fault identification of orbit structure in service state based on intelligent algorithm.By obtaining the time-history curve of the vibration response characteristics of the track structure in service caused by train operation,combined with intelligent algorithms,the service structure change law of the track structure is analyzed.The intelligent fault diagnosis of track structure is realized,which can replace or partially replace the shortcomings of current manual maintenance of skylight and inspection of track inspection vehicle,laying a foundation for fault diagnosis and maintenance of track structure.Firstly,based on the theory of vehicle line dynamics,the vibration acceleration of track structures(such as sleepers and track beds)is obtained and analyzed under different train operating speeds and different track structure conditions(such as: sleeper empty cranes,track bed compaction,track bed turning).Service status of track structure under four operating conditions.In this paper,the vibration response of vehicle track coupling is studied under the train running speed of 200km/h,150km/h and 100km/h under the condition of sleeper empty crane,slab bed and road bed turning over slurry,which serves as the data source for fault classification and identification of track structure.Secondly,this paper studies compressed sensing and sparse reconstruction methods.Combining the principle of compressed sensing and sparse reconstruction with the vehicle line dynamic model,the fault identification method of track structure is studied.By changing the stiffness(kb)and damping(cb)of the track structure model parameters,the vibration response under four operating conditions(normal,sleeper empty,track bed slab knot and track bed turning)is obtained,so that the time domain sample data of the track structure is sparse Refactoring.Due to the relatively complex structure of the vehicle line dynamics model,related to the solution of nonlinear equations,and considering the redundancy of data,the paper presents the original data sparsely,replacing the original data as the data source for fault pattern recognition of the track structure,so as to reduce the computation time of the simulation model.The simulation results show that the sparse reconstruction error of the sparse reconstructed sleeper vibration acceleration data is close to zero.Finally,the paper studies the fault classification and identification of time-domain data of track structure vibration response.Based on the vehicle-track coupling dynamics model,the time-history curves of vibration response of rails,sleepers,and track beds under different train operating speeds and normal conditions,sleeper empty crane conditions,track bed turning conditions,and track bed slab binding conditions are adopted.Comparison and analysis of three algorithms,such as class algorithm,dynamic time warping algorithm and support vector machine,realize the fault identification and classification of the service status of the track infrastructure.In the process of studying the density clustering algorithm,in order to improve the accuracy of the classification of orbital infrastructure,the k-distance simulation diagram is adopted for the parameter selection of the nearest neighbor and the minimum number of core points of the density clustering algorithm,and the relatively gentle curve segment is selected as the density clustering parameter.The simulation analysis shows that the determination of the parameters has a greater impact on the results of the clustering algorithm,and this paper distinguishes between normal operating conditions,sleeper empty crane operating conditions,track bed turning conditions,and track bed forming conditions,which are track structure failures.On-line monitoring of service status and intelligent identification of failure modes have laid the foundation.