Research On Fault Diagnosis System Of Rolling Bearing Of Main Fan In Coal Mine

Because there are many interference sources and frequent changes in the ventilation network of coal mine,the complex and changeable air flow excitation force is produced on the rotating impeller of the main fan in coal mine,which makes the rolling bearings at both ends of the impeller rotor of the main fan bear the effect of alternating stress,and the centrifugal force of the impeller is also borne on the rolling bearings.Therefore,the rolling bearing is an important factor affecting the operation safety of the main fan in coal mine,and it is also one of the key concerns in the maintenance of the main fan.Research on fault diagnosis of rolling bearing of main fan can not only improve the safety of main fan in coal mine,but also promote the technical progress of maintenance of main fan bearing in coal mine from regular maintenance to real-time diagnosis.Therefore,the research work carried out in this subject not only has important theoretical significance,but also has important social and economic benefits.Firstly,the structure,common failure modes and vibration mechanism of rolling bearings are briefly introduced.The theoretical formulas for calculating the characteristic frequency and natural frequency of rolling bearings are derived from the structural parameters of rolling bearings.Secondly,according to the vibration response characteristics of rolling bearings in fault state,the fault feature extraction and fault type identification methods for rolling bearings of main fan in coal mine are determined: empirical mode decomposition method and fault classifier based on BP neural network.In the feature extraction of rolling bearing fault,this paper uses EMD,EEMD and CEMD methods to extract the feature of unsteady vibration signal,and finds that CEEMD method has the best decomposition result and the smallest reconstruction error.Therefore,in this paper,the CEEMD method is used to extract features of fault signals of rolling bearings of main fan in coal mine,and the intrinsic mode functions(IMF)of fault signals of rolling bearings are obtained.Based on the advantage of sample entropy in processing complex signals,a sample entropy algorithm based on CEMD is proposed to extract fault features of rolling bearings.When extracting bearing fault features,the intrinsic mode function is obtained by decomposing vibration signals with CEEMD.Then the sample entropy parameters are determined by the relationship between sample entropy parameters and calculation time.In this paper,the sample entropy parameters used in extracting rolling bearing fault features are:sample length l=1024,embedding dimension m=2,threshold r=0.1.Finally,the sampleentropy values of IMF are calculated by using sample entropy.Construct eigenvectors.Using the sample entropy algorithm based on CEMD,the fault feature is extracted from the test data of different fault types under the same working condition and different working conditions respectively.The extraction results verify the feasibility of the sample entropy algorithm based on CEMD.A BP neural network classifier for identifying rolling bearing faults is established.The training and testing samples of BP neural network classifier are established by using the eigenvectors constructed by the sample entropy algorithm based on CEMD,and the training and testing are carried out.The results of training and testing show that the BP neural network classifier can accurately identify rolling bearing faults.In this paper,by using CEEMD,sample entropy,BP neural network and other algorithms,the fault of rolling bearing of main fan in coal mine is accurately identified,which provides a new method for the safe operation of main fan in coal mine.