Analysis Of Fatigue Characteristics And Reliability Life Prediction Of High Speed Rail Bearings

With the continuous growth of China’s high-speed rail mileage,important parts related to high speed rail have also attracted the attention of people.The fatigue characteristics and life prediction problem of high speed traction motor bearings have become the focus of attention.The research object of this paper is high speed traction motor bearing.It is necessary to provide effective prediction method for fatigue life of high speed rail bearing.1)according to the basic parameters of bearing rotor and its different vibration acceleration,the relationship between bearing’s different working conditions and bearing clearance and life is obtained.The bearing life of the bearing under different working conditions is obtained according to the calculation formula of the revised reference rating life of ISO16281,and the effect of coordination and temperature on the clearance of bearing is considered,and the initial radial clearance of the bearing is given.2)according to the fatigue exfoliation phenomenon of the high speed iron traction motor bearing during the working process,the fatigue life prediction of the bearing is carried out by the combination of the finite element theory and the fatigue analysis theory.Taking the bearing statics simulation result as the fatigue analysis model,the load dynamics simulation results are processed to obtain the load spectrum as the condition.The fatigue life of bearings is predicted by using nCode DesignLife fatigue analysis software,and the fatigue life of bearings is simulated.The fatigue life test of the bench is carried out.The average error between the test result and the simulation result is 7.1%,which verifies the accuracy of the simulation results.3)it expounds the concept of deep learning and its application direction,introduces the supervision,unsupervised and reinforcement learning algorithms of deep learning,and explains the main steps of deep learning to deal with general problems.According to the bench test results,the BP neural network algorithm,support vector machine algorithm and logistic regression algorithm are used to predict the bearing failure of the bearing.