Research On Fault Diagnosis Method Of Offset Rolling Bearing Based On Convolutional Neural Network

With the development of industrial level and the improvement of science and technology,printing equipment is developing towards the direction of complexity,automation and intelligence.Its running state directly affects the safety production and economic benefit of the enterprise.Therefore,it is of great practical significance and economic value to monitor and diagnose the state of the printing equipment.Rolling bearings are important components in printing equipment.The performance of a bearing will directly determine the operational performance of a machine.The extraction and analysis of the fault information contained in the vibration signal of the rolling bearing is a common and effective method in mechanical fault diagnosis.This paper takes the rolling bearing in the ink supply unit of offset printing press as the research object.Based on the non-linear and non-stationary characteristics of the vibration signal in the rolling bearing,the ensemble empirical mode decomposition and Hilbert transformation is used to decompose and characterize the fault vibration signal.The fault classification model of rolling bearing based on convolutional neural network was established,and the fault diagnosis of the roller bearing of ink roller was completed.The main work of the thesis is as follows:(1)The theoretical principle and algorithm process of the ensemble empirical mode decomposition method are studied,and its advantages in analyzing non-linear and non-stationary signals are summarized and applied to the signal processing of the printing press bearings.The experimental comparison proves the inhibition effect of the modal aliasing phenomenon in the traditional empirical mode decomposition method,which greatly improves the accuracy of the original signal decomposition.(2)A fault feature extraction method based on the ensemble empirical mode decomposition and Hilbert transform is proposed.After the bearing vibration signal is decomposed by the ensemble empirical mode decomposition,the effective IMF is selected to carry out Hilbert transformation to obtain the instantaneous frequency of the fault signal,and the fault feature data set of the bearing vibration signal is summarized.The instantaneous frequency of the bearing vibration signal can effectively reflect the time-frequency characteristics of the signal and accurately characterize the motion of the bearing.(3)Taking into account the excellent performance of convolutional neural networks in feature learning and pattern recognition,a fault classification model for rolling bearings based on convolutional neural networks is established.The influence of some important parameter settings in the convolutional neural network on bearing fault diagnosis results was experimentally analyzed,and the optimal parameters of the classification model suitable for the bearing sample conditions in this paper were obtained.It is verified that the fault classification model of this paper achieves a high classification accuracy in the fault diagnosis of rolling bearings.The related research provides important reference for the information acquisition and identification technology of intelligent fault diagnosis of printing presses,and has certain engineering application value.