Research On Fault Diagnosis Method Of High-Speed Rolling Bearing Based On Vibration Signal

As the core component of mechanical equipment,rolling bearings are widely used in aerospace,military,wind power and shipbuilding industries,which are also the vulnerable parts in mechanical equipment.Under high-speed working conditions,rolling bearings are easy to occur abnormal vibration and excessive noise,and the rolling bearing is more prone to failure.Therefore,the fault diagnosis of high-speed rolling bearings is of great practical significance to prevent the occurrence of shaft accidents.In this paper,high-speed rolling bearings were taken as the research object,based on the vibration data of rolling bearings,two kinds of fault diagnosis methods of high-speed rolling bearings were adopted,which provided the new method for the fault diagnosis of high-speed rolling bearings.In this paper,the research on fault diagnosis of high-speed rolling bearings was mainly carried out from the aspects of data preprocessing,fault feature extraction and feature identification.(1)The rolling bearings with slight,moderate and severe cracks and pitting faults were designed and manufactured.Based on the high-performance rolling bearing test rig,the test scheme was designed,and the vibration data under different states were collected for the fault diagnosis research of rolling bearing.(2)Aiming at the problem that it was difficult to extract fault features due to the strong noise of rolling bearings at high-speed condition,a rolling bearing fault diagnosis method based on wavelet thresholding denoising,complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)energy entropy and particle swarm optimization least squares support vector machine(PSO-LSSVM)was adopted.A wavelet threshold denoising was first applied to de-noise vibration signals,and CEEMDAN was used to decompose the vibration signals.Combined with the correlation coefficient and variance contribution rate,the effective components were selected.Energy entropy as the feature vector was input into the PSO-LSSVM classifier for fault diagnosis.The experimental results showed that the average accuracy of this proposed method can reach 95.18%for rolling bearings under different fault states,which proved the feasibility of the proposed method.(3)Due to the parameter optimization of variational mode decomposition(VMD)method and low accuracy of convergence problems in whales optimization algorithm(WOA),a novel fault diagnosis method of rolling bearing combining with wavelet threshold de-noising and genetic algorithm optimization variational mode decomposition(GA-VMD)and the whale optimization algorithm based on von Neumann topology optimization least squares support vector machine(VNWOA-LSSVM)was adopted in this paper,which can be used for accurate identification of failure and improve the diagnosis effect.The parameters of VMD were optimized by GA,and then the optimized VMD was adopted to extract fault feature information.The whale algorithm was improved by using von Neumann topology,and the fault feature vector was trained and classified by constructing VNWOA-LSSVM fault diagnosis model.The results showed that the average accuracy rate of rolling bearing fault diagnosis under different states can reach 97.00%,which improved the accuracy rate of fault diagnosis.(4)In order to assist the fault diagnosis research,based on the GUI module of MATLAB software,the corresponding fault diagnosis systems were developed for the two fault diagnosis methods respectively,which can realize the comprehensive analysis of data reading,preprocessing,feature extraction and fault identification.(5)Based on the high-performance rolling bearing comprehensive test rig,the experimental study under different working conditions was carried out for the preset crack and pitting faults.The identification accuracy and applicable conditions of the two diagnostic methods under different radial loads and rotational speed were compared.The results showed that the wavelet thresholding denoising with GA-VMD and improved WOA-LSSVM method was more suitable for high-speed and heavy-load condition.Based on this method,the reinforced life cycle test of rolling bearings under high-speed and heavy-load was carried out.The results showed that the life cycle faults of bearings can be diagnosed effectively by this method.