Research On Testing And Life Prediction Of Electromagnetic Contactor For High-speed Train Air Conditioning System

The electromagnetic contactor of the high-speed train air-conditioning system is the key component to ensure the reliable operation of the train air-conditioning system,and its performance will degrade continuously until failure with the increase of train service time.Once it fails,the normal operation of the air-conditioning system will be seriously affected,and the temperature change caused by the abnormality of the air-conditioning system will have a negative impact on the riding comfort of passengers.At present,it is difficult to realize the online monitoring of the state of the high-speed train contactor due to the limitation of technology and funds.The reliable operation of the train is ensured mainly by replacing the contactor in advance regularly,which will cause a great waste of contactor performance.In order to maximize the use of the contactor performance under the premise of the reliable operation of the train,it is necessary to study the residual life prediction of the contactor.In respect of the above issues,this thesis determines and verifies the degradation sensitivity of different performance parameters based on analyzing the failure mechanism of the contactor and establishing its electromagnetic and dynamic simulation models.Furthermore,a performance test system of the contactor is designed and built,and the degradation data of the contactor performance parameters in the whole life cycle are obtained.Then,the performance degradation models based on the pull-in time and the super-path time are established respectively,and the optimal estimation of the parameters for the models is carried out by using the particle filter algorithm,which realizes the residual life prediction of the contactor.In order to further improve the engineering applicability of the prediction model,a prediction model for the remaining life of the contactor based on the RBF neural network is established.Meanwhile,an improved particle filter algorithm is used to update and optimize the parameters of the prediction model,and the performance of the model to predict the remaining life is analyzed and verified by the test data.This thesis mainly includes the following contents:1.The structure,working principle,and failure mechanism of the contactor are studied,and the sensitivity of the two performance parameters of pull-in time and super-path time to the degradation process of the contactor contact is analyzed.The electromagnetic and dynamic simulation models of the contactor are established,and the pull-in motion processes of the contactor under different contact gaps are simulated and analyzed.Further,the functional relationships of different performance parameters with contact gap are obtained,which verifies that the pull-in time and the super-path time are both highly sensitive to the contact degradation process.2.According to the life test standard of the contactor,a performance test system of the contactor is designed and built.The test system takes the test software of the host computer and the IMC data collector as the core,and has completed the degradation test and performance parameter test of the contactor in the whole life cycle.Then,the degradation data of performance parameters are obtained.Wavelet transform and stationary time series analysis are used to eliminate the random noise information in the original degradation data,and the trend term in the degradation data is extracted.3.Combined with the cumulative fatigue damage theory and the contact failure mechanism,a degradation model of the contact gap is established.On this basis,based on the dynamic simulation results of the contactor,the performance degradation models of the contactor based on the pull-in time and the super-path time are established respectively.The parameters of the performance degradation models are optimally estimated by the standard particle filter algorithm,and then the residual life of the contactor is predicted.The results show that the performance degradation models have high accuracy in predicting the residual life.4.In view of the problems that the prediction accuracy of the performance degradation models is affected by the individual differences of samples,the models require a large amount of data for the predicted samples,and the standard particle filter algorithm has deficiencies.Taking the cumulative arcing energy,pull-in time,and super-path time together as input variables,a prediction model for the residual life of the contactor based on the RBF neural network is established.Meanwhile,an improved adaptive genetic algorithm that can adjust the crossover and mutation probability nonlinearly and adaptively with evolution and fitness is proposed.Then the improved adaptive genetic algorithm is used to optimize the resampling process of the particle filter algorithm.Further,the parameters of the prediction model based on the RBF neural network are optimized by the improved particle filter algorithm,and the optimal prediction model of the residual life of the contactor is obtained.The prediction results show that the prediction accuracy of the model is improved,and the model has strong generalization ability and high engineering practicability.