| With the progress of science and technology and the increasing social demand,mechanical equipment is developing towards high speed,high efficiency and high precision.As the most important transmission part in mechanical equipment,the health of gearbox is closely related to the safe operation of mechanical equipment.Because of the complex structure and terrible working environment,gearbox is one of the most vulnerable parts in mechanical equipment.Therefore,the research of gearbox fault diagnosis method is of great significance to ensure the safe operation of mechanical equipment and avoid the occurrence of major accidents.In this academic paper,signals noises reduction algorithm and the fault feature extraction and diagnosis method which the research object are gearbox are studied.The main research contents are as follows:（1）Study on noises reduction algorithm of gearbox vibration signals.Due to ambient noises and fixed parts loose etc,the vibration signals of gearbox usually contains a lot of noises.An improved VMD-SVD noises reduction algorithm based on variational mode decomposition and singular value decomposition is proposed.In this method,singular value decomposition is embedded into variational mode decomposition,and a large amount of noises is filtered out on the basis of retaining useful information.The effectiveness of the proposed method is verified by comparing and analyzing the results of the simulated signals’ and the measured signals’ noises reduction experiment.（2）Study on fault feature extraction and diagnosis method based on two-dimension convolutional neural network（2-D CNN）.The convolutional neural network avoids the complex feature extraction process in the traditional pattern recognition algorithm and contains feature extraction in the convolution and pooling operation.Aiming at the problem of fault feature extraction and diagnosis of gearbox vibration signals,2-D CNN method based on snake permutation algorithm is used.Compared with other fault feature extraction and diagnosis methods,the method in this thesis has achieved higher recognition accuracy.Through the comparison and analysis of the experimental results,it can be concluded that the fault diagnosis method based on improved VMD-SVD and 2-D CNN is suitable for the fault diagnosis of gearbox under strong background noises. |
