Research On Optimal Method Of Rolling Bearing Fault Diagnosis Considering Multiple Fault Features

Mechanical equipment accidents and failures can cause serious financial losses and even affect people’s lives.As one of the most commonly used parts in mechanical equipment,rolling bearings play a vital role in mechanical equipment.Its working environment is usually complicated and difficult,so it also has a high failure rate.In the case of a bearing failure,how to accurately identify the location and type of the failure and formulate corresponding maintenance strategies to ensure the safe and reliable operation of the equipment is very important for improving economic benefits and ensuring production safety.In order to solve the problem of high misdiagnosis rate caused by redundant or insufficient fault signal characteristic information in the fault diagnosis of rolling bearings,this article takes rolling bearings as the research object,fully considers the connection between the various processes of fault diagnosis,and extracts and selects multi-domain features of the vibration signals of different types of faults such as inner ring,ball,outer ring and so on.And through a suitable classifier model to identify the state information contained in the vibration signal.First of all,considering the complex operating environment of the rolling bearing,the vibration signal contains a large amount of irrelevant external noise during the operation,and the noise of the original collected signal is reduced.Through analyzing the typical components and failure modes of rolling bearings,the failure mechanism of bearings,etc.,it is proposed to introduce wavelet packet decomposition to process the vibration signals of rolling bearings.The irrelevant noise that covers the vibration signal of the rolling bearing is removed,so that the weak vibration signal of the rolling bearing can be revealed,and the basis for extracting the accurate characteristics of the rolling bearing vibration signal as much as possible is provided.Secondly,multi-domain feature extraction is performed on the rolling bearing fault signal to fully extract the fault feature information of the rolling bearing.In terms of time domain and frequency domain feature extraction,typical information features such as statistical sample parameters of fault signals are extracted respectively.In the time-frequency domain,wavelet packet decomposition is used to extract the signal sample entropy after decomposition.The extracted time-domain feature parameters,frequency-domain feature parameters,and time-frequency domain feature parameters are combined to form a multi-domain feature set,which provides key diagnostic conditions for the fault diagnosis of rolling bearings.Then,in terms of feature selection,it is proposed to diagnose and analyze bearing faults through the best feature data set,and use an improved particle swarm optimization(GDPSO)algorithm to screen the extracted multi-domain feature data set.The iterative rules and weight coefficients of the traditional particle swarm algorithm are optimized,and the fitness function is set to achieve the purpose of extracting the most suitable feature vector for fault diagnosis of the classification model,forming a rolling bearing fault diagnosis method based on the combination of improved particle swarm and decision tree.Finally,the method proposed in the article is applied to the measured fault vibration signal of rolling bearing to verify the method.The influence of a single or a limited number of faults on the limitation of the fault diagnosis result is eliminated,and the diagnostic effect of the GDPSO-DT rolling bearing fault diagnosis method on the rolling bearing fault vibration signal under different fault conditions is analyzed,and the fault diagnosis result is obtained.The experimental results show that the fault diagnosis method proposed in this article has a higher diagnostic accuracy and a better actual diagnosis effect.