Fault Diagnosis Of A Rotor-bearing System Based On VMD

With the development of our country’s economy and technology,the process of industrialization is gradually accelerating,the proportion of rotating machinery in the mechanical system is increasing year by year,and the structure and technology in the equipment are becoming more complex,and the research direction is also developing towards intelligence and integration.The rotor-bearing system,as one of the core components of rotating machinery,is affected by the working environment and its own processing errors,and is prone to different types of failures such as rotor cracks and damage to the inner and outer rings of the bearing.As the degree of failure increases,it will affect the operation of the entire equipment.In severe cases,the equipment will be paralyzed,causing economic and property losses,and even threatening the life and safety of related personnel.Therefore,the realization of the fault diagnosis of the rotorbearing system plays a key role in the safe operation and monitoring of the equipment.This paper takes the rotor-bearing system as the research object,establishes a finite element simulation model of the rotor-bearing system based on the finite element method,and builds a rotor-bearing test bench;analyzes the health of the rotor-bearing system,rotor cracks,bearing failures and rotor cracks Dynamic characteristics under four states of coupling fault with bearing;Based on genetic algorithm,the penalty factor and the number of sub-modes of variational modal decomposition(VMD)are optimized,and the optimized VMD can decompose and reconstruct the bearing vibration signal.Bearing fault identification;based on VMD and principal component analysis to reconstruct the rotor vibration signal,using the reconstructed signal as input in the Light GBM algorithm to achieve system health,rotor cracks,bearing failures,and rotor cracks and bearing coupling failures.Classification under status.A finite element simulation model of the rotor-bearing system is established.Based on the strain energy release and multi-body dynamics theory,the stiffness matrix of the crack and the impact load of the faulty bearing are calculated respectively,and the rotorbearing system test bench is built according to the parameters of the simulation model.Through simulation and experimental signals,the dynamic characteristics of system health,rotor cracks,bearing failures and coupling failures of the two are simultaneously analyzed,and the correctness of the simulation model is verified through modal tests.The results show that the 2X and 3X super-harmonic frequency components appearing at 1/2 and 1/3 of the critical speed are important characteristics of rotor cracks;the occurrence of oscillation in the axis trajectory graph and the instantaneous impact frequency in the instantaneous frequency graph are bearing failure Important features.At the same time,the coupling of the two faults will cause the amplitude of the 1X component in the vibration signal to increase,and the amplitude of the 2X and 3X superharmonic components will decrease,which will affect the diagnosis of the fault.In view of the problem that spectrum analysis cannot effectively diagnose bearing faults,after the bearing fails,the rolling elements collide with the raceway to produce periodic impacts,and the weighted index value using the kurtosis value and the autocorrelation coefficient in the vibration signal is proposed.As the evaluation index of the bearing fault characteristic information,the weighted index cumulative value of the decomposed components is used as the fitness function.Based on the genetic algorithm(GA),the number of sub-modes and the penalty factor of the VMD are optimized,and the parameter optimized VMD is used to adjust the bearing The vibration signal is decomposed and reconstructed,and the frequency spectrum of the reconstructed signal is analyzed to obtain the corresponding fault frequency,which realizes the fault diagnosis of the bearing.Aiming at the problem that when the rotor crack and the bearing fault are coupled,the fault characteristic information of the bearing is weak,the amplitude of the superharmonic component of the rotor changes,and the two faults cannot be effectively diagnosed at the same time,a reconstruction signal based on VMD and ICA is proposed.,Use Light GBM algorithm to realize the fault classification and recognition method in different states,and optimize the tree depth and learning rate of Light GBM based on grid search.First,sort the importance of the time-frequency domain characteristics of rotor cracks and bearing faults,and find the best sensor measurement points of the entire system according to the importance ranking;then,use VMD to separate the rotor crack characteristic components in the measurement point signals,and The separated rotor signal and the original measuring point signal are subjected to principal component analysis to obtain the bearing fault information in the signal;finally,the time domain signal after the superposition of the two fault signals is input into the parameter-optimized Light GBM algorithm,and the rotor is successfullyVibration signals under the four types of bearing system health,rotor cracks,bearing outer ring faults and coupling faults have been classified and identified.The research results of this paper can realize the identification of rotor cracks,bearing faults and coupling faults in the rotor-bearing system by using simulation and experimental rotor vibration signals under the premise of using only a single sensor,which has certain engineering practical value.