Study On Intelligent Fault Diagnosis Meathod Of Key Parts Of Gearbox Under Variable Working Conditions

As a key part of rotating machinery,gearbox is widely used in industry.With the rapid development of science and technology,the increasing demand of industry,the service condition of gearbox is more complex and changeable,and its parts are more prone to damage and failure.The failure of parts may cause a series of chain reactions,leading to the failure of the whole rotating machinery,thus seriously affecting the normal industrial production,causing major economic losses and even casualties.Therefore,the research on the characteristics analysis and diagnosis of key parts of gearbox,such as gear and bearing failures,is carried out to ensure the safe and efficient operation of equipment,avoid economic losses and major safety events Therefore,it plays an important role.In this paper,gear box and bearing are taken as the research object,and the optimal feature extraction of gear local fault,bearing vibration signal analysis and intelligent fault diagnosis method under variable working conditions are studied:(1)In order to improve the accuracy of local fault feature extraction of gearbox under constant working condition,an optimal feature extraction algorithm based on CEEMDAN-SQI-SVD is proposed.In this paper,the selection of effective intrinsic mode function from CEEMDAN decomposition is studied,SQI algorithm is proposed to select the optimal reconstruction signal,and the optimal feature of gear local fault is obtained by combining SVD algorithm,and the experimental analysis of gear local crack fault with different fault levels is carried out,and the effectiveness of the method is verified by using BP neural network as classifier.(2)Aiming at the problem that the conventional method is difficult to deal with the unstable vibration signal under the condition of variable speed,the order spectrum analysis method of bearing local fault vibration signal under the condition of variable speed based on angle domain resampling theory is proposed.In order to verify the effectiveness of this method for the unsteady vibration signal processing,a bearing failure test-bed was built and the bearing failure simulation test under the condition of uniform acceleration was carried out.The results show that the angle domain resampling theory can effectively transform the unstable vibration signal in time domain into the relatively stable vibration signal in angle domain,and the order spectrum analysis combined with CEEMDAN and SQI theory can more effectively diagnose the bearing fault location under variable working conditions.(3)In order to solve the problem that it is difficult to diagnose the bearing fault under variable speed condition,an intelligent fault diagnosis method based on angle domain resampling technology and integration of SDP and DCNN is proposed.Using CEEMDAN to denoise and reconstruct angular vibration signals,the proportion of fault information components in signals is further increased;using SDP to visualize signals,determining SDP parameters by Pearson correlation coefficient method,mapping different fault types of bearing reconstruction signals to polar coordinates to form SDP patterns;using deep convolution neural network to identify and reconstruct the characteristics of SDP patterns classification.The results show that the proposed method can effectively diagnose the bearing fault under variable working conditions.The diagnostic rate of training samples is 97.14%,and the diagnostic rate of test samples is 96.00%,which is superior to several other methods under the same test conditions.