Research On High-speed Train Service Life Of Rubber Seals Evaluation Techniques

Rubber ring is one of the high-speed train high speed train braking system criticalcomponents for sealing, its performance, especially the length of life, directly affectsthe braking performance of the train and the train ’s security. Rubber seals with goodsealing performance, high-speed train group snorkel air brake system is sealed toprovide sufficient power to ensure the normal operation of the train braking system,widely used in high-speed rail passive vehicle. But with the use of time, the aging ofthe rubber material damage will occur, performance and service life will be greatlyreduced, so be accurately predicted, comprehensive evaluation of its performance andlife, develop a reasonable maintenance and replacement procedures, it is veryimportant.Although EMU high-speed rail rapid development in our country, but thedevelopment of a relatively short time, many technology is still immature,specifically for the study of EMU high-speed rail operating conditions, theenvironment, little rubber seals, especially for its performance and life evaluationtechnology research remains blank. To meet the EMU and the high-speed rail along-running process safety requirements, its use of the rubber seals of performanceand service life of an accurate evaluation to determine proper maintenance andreplacement program for all EMU present, China is the operator and high-speed railfacing the urgent need to address the problem. Through BP artificial neural networkmodel and kinetics model of aging life of rubber seals were studied, and the twomethods were compared, the main contents are as follows:(1) An article by referring to a large amount of literature on the research status ofrubber aging life are discussed, while the current domestic and international hot airaccelerated aging test several common data processing methods were studied, andthen aging mechanism of rubber and BP neural network techniques were evaluated indetail(2) This paper experiment used the same high-speed train ring rubber sealrefining own recipe, then the accelerated aging test compound with hot air acceleratedaging experiments at different temperatures measured at different times of physical andmechanical properties of low temperature, the subsequent data processing provides experimental support.(3) Using BP artificial neural aging model to predict the life of rubber seals, selectthe appropriate network architecture and the training sample, and compare severalalgorithms, eventually using Levenberg-Marqurdt aging rubber seals algorithm topredict the life expectancy, predicted rubber seals aging life.(4) Currently more mature mathematical statistics conducted on the rubber seallife prediction, first of all parameters of the basic theory of aging life predictionequation estimated by linear regression method and mathematical methods to try. Bythen conducted a statistical analysis of the related formula validation, and ultimatelycome to a predictive model of aging life of rubber seals, rubber seals predicted aginglife.(5) Curve of BP neural network forecasting model and dynamics model of thepredicted results are compared, and through the high-speed train3,4,5trim off theseal it can validate the two methods, finally, two methods have their own advantagesand disadvantages, the BP neural network prediction model to predict process is simple,flexible, fast, strong fault tolerance, don’t need to make clear the specific relationship,but the prediction accuracy is lower than the dynamics curve model, dynamiccomputation complex curve model, but the prediction result is superior to the BPneural network evaluation technology, the prediction results of two methods are allwithin the scope of accepting. Finally by using SPSS software to BP neural networkprediction model is revised, get more precise life prediction model.