Research On Fault Diagnosis Of Rolling Bearing Based On QPSO-RVM

Railway transportation has made outstanding contributions to China's economic construction;it is a specific expression of the strength of the country's economic development.At present,the tamping truck is an important and commonly used mechanical equipment for railway operation and railway maintenance,and its role cannot be underestimated.Because of its long-term harsh working environment,the damage of corresponding components such as bearings is also relatively large.Once a failure occurs and is not resolved in time,the operation of other trains will also be affected accordingly.Therefore,how to accurately and quickly find the early failure of the rolling bearing of the tamping car has become the focus of future research.1)In this paper,Firstly,we perform repeated row preprocessing,approximation preprocessing,and balanced multi wavelet processing on GHM multi wavelet,CL4 multi wavelet,and SA4 multi wavelet.At the same time,the corresponding multi wavelet after preprocessing is applied to the rolling bearing fault signal denoising stage.Meanwhile,carry out a comparative analysis of the corresponding effects.The comparison of the experimental results shows that the CL4 multi wavelet with balanced multi wavelet preprocessing has the best effect on the rolling element fault signal,the inner ring fault signal,and the outer ring fault signal;2)Then the overall average empirical mode decomposition(EEMD)method is used to process the vibration signal of rolling bearing.After decomposing,many intrinsic modal functions(IMF)can be obtained.Next,IMF is input into the Relevance Vector Machine model in the form of feature vectors to compl ete the classification process.3)In terms of the specific selection of classification models,this paper mainly uses the quantum particle swarm optimization(QPSO)to optimize the parameters in the Relevance Vector Machine(RVM)classification model,and uses the RVM classifier optimized based on the quantum particle swarm optimization to classify the fault s,and then accurately diagnose the faults of rolling bearingsThis article introduces the multi wavelet and the corresponding pretreatment method in detail,and discusses the actual effect difference of different pretreatment methods.It also analyzes the impact of the multi wavelet application in the field of rolling bearing fa ult analysis after the different methods are processed in the form of simulation.In addition,on this basis,a method of extracting related fault features of EEMD tamping rolling bearing based on overall empirical mode decomposition is studied.This method can effectively anal yze some types of non-linear and irregular signals.Additionally,this method can also effectively analyze the type of rotation: The vibration signal source is decomposed accordingly,and the corresponding forms of different IMFs are obtained,and also the specific frequency is decomposed into each IMF.Therefore,the IMF energy is the feature vector input to the diagnostic device.At the same time,this paper also conducts corresponding research on the model of bearing fault classific ation,and builds an optimized Relevance Vector Machine model based on particle swarm optimization algorithm,which can effectively improve the accuracy of the entire diagnosis process.In this paper,a diagnostic model of QPSO-RVM is finally constructed.After the sample is input,it can be self-trained and the Gaussian radial basis kernel is selected,and the optimal function can also be adaptively obtained by QPSO.The model can achieve intelligent output for its monitoring samples.Additionally,the model can effectively identify the type of faults existing in the bearing,and can further improve the accuracy of the analysis.It has a good effect in the actual application process.