Research On Identification Method Of Deterioration State Of Key Components Of Urban Rail Vehicle Bogies

The bogie is the core component of urban rail vehicles.During the long-term service of the train,some key components of the bogie will experience performance degradation or damage,which will cause severe vibration and performance degradation at different positions of the train,and even cause dangerous accidents such as derailment and rollover of the train in severe cases.In order to ensure the safe operation of the train,when the performance of a key component has deteriorated to a certain degree,the component must be replaced.How to make full use of the real-time monitoring data of urban rail vehicles for in-depth mining has become an important topic in the field of urban rail vehicle degradation identification and health assessment.In view of the above problems,this article first studied the impact of the performance degradation of key components of the bogie on the dynamic performance of the vehicle under variable conditions.Then,a method based on the multi-channel information fusion model of convolutional neural network to identify the performance degradation of bogie suspension components and a method of identifying the wear status of wheelset treads based on the VGG network are proposed.Finally,integrating the above-mentioned dynamic analysis and key component degradation identification research results,a software system for performance degradation identification and dynamics evaluation of key bogie components was developed.The research results of this thesis mainly include the following four parts:(1)The impact of the performance degradation of key components of the bogie on the dynamic performance of the entire vehicle under variable operating conditions is studied.Use SIMPACK dynamics software to simulate the performance degradation of key components by changing the performance parameters of the key components of the bogie.The speed level is set from 40km/h at 20 to 120km/h in 5 levels;the degradation level is set from 100% to 10% from the original value.There are 6 levels of 50%.The effects of secondary transverse shock absorbers,secondary vertical shock absorbers and air springs on the critical speed,stability and safety of the car body under different speeds and different degrees of degradation are analyzed.It provides a theoretical basis for the identification of the suspension parts in Chapter 3 and the wear identification of the wheelset tread in Chapter4.(2)A method based on the multi-channel information fusion model of convolutional neural network to identify the performance degradation of the key components of the bogie is proposed.This method can accurately realize the "end-to-end" prediction directly from the vibration signal to the degree of performance degradation,and does not require a manual feature extraction process,thereby reducing the dependence on expert knowledge and experience.For a complex system such as a bogie,the vibration signal of a single channel has defects such as lack of information and low signal-to-noise ratio.Therefore,this method improves the prediction accuracy of the model by fusing the vibration information of multiple channels.Finally,it was verified through an example.(3)A method of wheelset tread wear status identification based on VGG network is proposed.Tread wear conditions are more complicated,which makes the recognition accuracy of tread wear low by ordinary convolutional neural networks.Drawing lessons from the idea of the VGG network,by adopting a small convolution kernel and a small pooling kernel,the number of layers of convolution is increased,thereby enhancing the depth of the network and enhancing the feature learning ability.Finally,compared with the ordinary convolutional network,the recognition accuracy of the tread wear state is improved.Finally,it was verified through an example.(4)Developed a software system for identification and dynamics evaluation of the performance degradation of key components of urban rail vehicle bogies.The software system integrates the research results of the first and second parts mentioned above,and can predict the performance degradation degree of key bogie components online and the impact of the degradation degree on the dynamic performance of the whole vehicle through the collected bogie vibration data.An application example shows that the software system has achieved the expected basic design function.