Simulation Research On Weak Fault Detection And Performance Degradation Prediction Of Urban Rail Vehicle Suspension System

In recent years,the urban rail traffic,which has the characteristics of large capacity,high efficiency,comfort,and energy saving,has become an important mode of transportation for many cities in China.Suspension system is a key component of urban rail vehicles,of which the early weak fault detection and overall performance attenuation prediction can not only improve the operation safety but also establish the maintenance plan based the equipment status and reduce the cost.In this paper,the weak fault detection,the safety grade evaluation and the performance attenuation prediction of the urban rail vehicle suspension system are researched systematically and comprehensively.The research methods are verified by the simulation platform of urban rail vehicle suspension system based on the professional multi-body dynamics software SIMPSCK.The main contents are as follows:1.A new method of weak fault detection of urban rail vehicle suspension system is proposed.For the problem of weak fault detection of the suspension system of the urban rail vehicle,the fault data is constructed in tensor form,which enriches the fault information.A multilinear dynamic principle component analysis(MDPCA)is proposed to improve the quality of fault feature extraction.The Eros similarity measure algorithm is introduced to detect the change of the data feature when the fault occurs,and the falut location separation is achieved by distributed method.An experiment is made using the fault simulation platform for urban rail vehicle suspension system,compared with traditional DPCA,MPCA,the method proposed by this paper shows better ability to identify weak attenuation like below 10%of suspension components.The problem of weak fault detection of urban rail vehicle suspension system is solved,which provides a powerful technical support for the safety monitoring and early warning of urban rail transit.2.The safety evaluation of the suspension system of the urban rail vehicle is researched based on empirical correlation analysis.The ensemble empirical mode decomposition(EEMD)is used to achieve the extraction of IMF,which overcomes the problem about the mode mixing,then using the correlation coefficient between empirical modes to characterize the safety level of the system.An experiment is made using the fault simulation platform for urban rail vehicle suspension system,the results show that the method used in this paper avoids the defects of the traditional FAHP evaluation algorithm,such as the strong dependence of human factors and the low effectiveness of evaluation results.The reliability of the evaluation results of safety grade of suspension system is improved.3.A method of performance attenuation prediction about whole urban rail vehicle suspension system is proposed.The performance attenuation evaluation index of the whole urban rail vehicle suspension system is established,and the performance attenuation prediction model is built by least squares support vector regression(LSSVR),of which the parameters are optimized based on the improved particle swarm algorithm(SPSO).An experiment is made using the fault simulation platform for urban rail vehicle suspension system,the results show that compared with traditional prediction model,the SPSO-LSSVR model has higher prediction efficiency and prediction accuracy.In addition,according to the actual application,the prediction effect of vehicle body and bogie acceleration as model input is analyzed and a comparative study was carried out in the horizontal and vertical directions,which verifies the feasibility of the method used in the field engineering.The research of this paper provides a valuable reference for the passive maintenance to the active maintenance based on the state of the suspension system.