Research On Fault Diagnosis Method Of Hoist Rolling Bearing Under Changeable And Complex Working Conditions

Mine hoist is an important equipment in mine production,known as the mine "throat".Rolling bearing as a key part of mine hoist,its health status directly affects the reliability of mine hoist operation.Therefore,the realization of fault diagnosis of mine hoist rolling bearing is of great significance to mine safety and efficient production.The components of mine hoist are coupled with each other,and its operation condition is complex and changeable,which leads to the vibration signal of rolling bearing containing a lot of noise.Moreover,the vibration signal data distribution of rolling bearing in the same mode is very different,and it is difficult to extract the sensitive fault features.Aiming at the above problems,this thesis takes digital signal processing method and deep learning algorithm as technical support,and carried out an in-depth research on the fault diagnosis of mine hoist rolling bearing.The main research contents are as follows:In view of the fact that the vibration signals of the rolling bearing of the hoist collected by the sensor contain a lot of noise and can not effectively characterize the health state of the rolling bearing of the hoist,a method of vibration signal reconstruction of the rolling bearing based on parameter optimized Variational Mode Decomposition(VMD)was proposed.Firstly,the average weighted sparse kurtosis was used as the fitness function,and the harris eagle optimization was used to optimize VMD parameters.Then,according to the obtained best parameters,the original vibration signal of the hoist rolling bearing was decomposed by VMD,and the effective weighted sparse kurtosis was used as the index to select the modal component and reconstruct the signal,so as to filter out the interference information and retain the original fault features.The experimental results show that the signal reconstruction method proposed in this dissertation enhances the expression ability of vibration signals and effectively improves the fault diagnosis accuracy of mine hoist rolling bearing.In order to solve the problem that the distribution of vibration signal of rolling bearings in the same mode varies greatly under variable working conditions,and the complex mapping between the vibration signal of rolling bearings and the actual health state of rolling bearings cannot be effectively established.A fault diagnosis model based on Multi-scale Convolution Neural network(MSCNN),Bidirectional Gated Circulation Unit(Bi GRU)and Attention Mechanism(AM)was proposed.In this model,MSCNN was used to extract multi-scale spatial features of rolling bearing vibration signals,and Bi GRU was used to mine the time sequence features of rolling bearing vibration signals.In order to make MSCNN-Bi GRU model have differentiated feature learning ability,this thesis introduces convolution attention module and sequential attention module respectively.The experimental results show that the hybrid neural network model MSCNN-Bi GRU can mine the rich temporal and spatial features of the rolling bearing vibration signals,and the introduction of the AM enables the model to extract the key fault features,which further improves the fault diagnosis accuracy of the model.Combined with the above method research and according to the actual situation and demand of the site,this dissertation designs and implements a mine hoist rolling bearing fault diagnosis system,which provides a strong guarantee for the mine safe and efficient production.Figure [39] Table [7] Reference [70]...