Research On Rolling Bearing Fault Detection Method Based On Variational Mode Decomposition

As the core component of the rotating machinery,the running condition of the rolling bearing directly affects the operation of the mechanical equipment and even the stability of the whole system.At the same time,the rolling bearing is one of the important fault sources of the mechanical equipment,and most of the faults of the equipment are closely related to the bearing.Because of the influence of the processing technology,the working environment and the artificial operation of the bearing,the bearing will be damaged and the normal operation of the system will be affected.Therefore,the fault diagnosis of the bearing has a very important practical significance.In this thesis,the fault diagnosis of ball bearing is mainly studied.From the processing of bearing vibration signal,the main research contents are as follows:Briefly expounds the meaning and purpose of the rolling bearing fault diagnosis system;introduces the rolling bearing type,basic structure,fault frequency and fault diagnosis development of rolling bearing and other related knowledge;introduces briefly the research status of the rolling bearing fault signal feature extraction and fault pattern recognition.Introduce the empirical mode decomposition,ensemble empirical mode decomposition and local mean decomposition,according to the disadvantages of modal aliasing of the modal decomposition method,introduce the variational modal decomposition based on optimized leapfrog algorithm,effectively solve the modal aliasing phenomenon,but also avoid the artificial decomposition parameters of objectivity.The feasibility and effectiveness of the variational mode decomposition method based on the leap-frog algorithm is verified by the decomposition of the simulation signals and the measured signals.The initial fault signal through the parameter optimization of variational modal decomposition to obtain a series of modal functions,using the envelope correlation spectrum coefficient to select the optimum component matrix,singular value decomposition of the matrix to obtain the singular value of the matrix,the singular value as the characteristic vector of fault pattern recognition of fault type.The least squares support vector machine is used to identify the fault types and the Bayes algorithm is used to optimize the parameters of the least squares support vector machine.Finally,the accuracy and reliability of the least squares support vector machine based on Bayesian Optimization in fault type identification is verified by the actual fault type identification.At the same time,the effects of different decomposition methods and different pattern recognition methods and the number of training samples and test samples on the pattern recognition results are compared.