Research On Fault Diagnosis Of Gearbox Based On Fast Spectral Correlation

The gearbox is one of the most important mechanical components,and its health directly affects the operation of the entire plant.If the early weak fault of the gearbox can be detected in time,the loss caused by the gearbox failure can be effectively avoided.Therefore,it is of great engineering significance to study the new method of gearbox fault diagnosis.Rolling bearings and gears are the core components of the gearbox and have a high failure rate.Therefore,this paper takes the two as the main research objects,and focuses on a new kind of cyclostationary analysis method-fast spectrum correlation,from the weak fault diagnosis of gearbox,composite fault diagnosis and fault intelligent identification,the main research contents as follows:First of all,it explains the background of the project and the necessity of the work done in this paper.The domestic and international status and development of gearbox fault diagnosis research are summarized and summarized.The research status of the cyclostationary theory is introduced,which paves a good foundation for the development of the thesis.Then the basic principles of fast spectrum correlation are briefly described,and the fault diagnosis examples of rolling bearings and gears are used to show that the fast spectrum correlation can effectively extract fault impact characteristics and lay a theoretical foundation for the paper.Due to the harsh working environment of the gearbox,the fault characteristics of the vibration signals of the gears and rolling bearings are often submerged in strong background noise.In order to effectively extract the rolling bearing fault information under strong noise background,a fault feature extraction method based on total variation denoising and fast spectral correlation is proposed.The total variation denoising preserves the fault characteristic information while reducing the original signal.The fast spectrum correlation effectively extracts the fault impact characteristics of the noise reduction signal,which is conducive to accurate gearbox fault diagnosis.In addition to the fault characteristic signal,the composite fault signal also contains an interference signal,which is difficult to accurately separate the fault signal by conventional methods.A method based on singular value decomposition and fast spectral correlation is proposed for gearbox composite fault diagnosis.The singular value decomposition is used to separate the trend component containing the fault feature information from the interference component,and then the integrated fast spectral correlation method is used to resonate the trend component to realize the composite fault separation.Aiming at the problem of intelligent identification of gearbox faults,a fault diagnosis method based on fast spectral correlation and particle swarm optimization support vector machine is proposed.Firstly,fast spectral correlation analysis is performed on the fault vibration signal to obtain fast spectral correlation spectrum.Then,four cyclic frequencies are selected in the fast spectral correlation and the energy average is obtained to obtain the characteristic energy matrix of the signal.Finally,it is used as the eigenvector.Input to the particle swarm optimization support vector machine for training and testing to achieve fault mode intelligent diagnosis.The effectiveness of the algorithm is verified by the fault diagnosis of rolling bearings.