Weak Fault Enhancement Of Rolling Element Bearings Based On Deconvolution Cascade Variational Mode Decomposition

As one of the most important machine components,rolling element bearings have been widely used in ships,airplanes and other mechanical equipment.However,bearings are run in the hard environment of high load and variable speed,and failure may occur at any time,thereby inducing the failure of the transmission mechanism or even the entire equipment.Therefore,the condition monitoring of bearings is particularly important for the safe operation of mechanical equipment,especially the monitoring and early warning of weak faults is of great significance to prevent equipment failure.In order to solve the problem that the weak fault signal is difficult to extract the fault feature with low signal-to-noise ratio under strong noise,long transmission path and variable speed.This thesis focuses on three aspects:interference noise suppression,transfer function compensation and variable speed stabilization.In this thesis,acoustic and acceleration sensors are used to monitor the health status of rolling bearings,and the sound and vibration signals reflecting the running status of rolling bearings are processed to determine the fault condition of rolling bearings.The main research contents of the paper are as follows:First,in order to show the influence of load distribution,transfer function and resonance attenuation on the modulation of rolling bearings,the modeling analysis of rolling bearings was carried out.The modeling results revealed the modulation characteristics of rolling bearings,that is,the spectral line characteristics of the envelope spectrum of each bearing component.Through the simulation signal of the rolling bearing,the envelope spectrum and the square envelope spectrum of the rolling bearing fault demodulation detection ability are analyzed.The results show that the square envelope spectrum has a stronger detection capability for the modulation characteristics.In order to quantify the fault enhancement capabilities of different methods,an envelope factor that can reflect the standardized square envelope spectrum of the fault characteristic frequency and its multiple harmonics is proposed as an indicator.Then,aiming at the influence of strong noise during operation,according to the principle of resonance demodulation,cascaded variational mode decomposition is proposed to determine the resonance component.On the basis of variational mode decomposition,multi-layer variational mode decomposition is carried out by cascaded method to determine the center frequency of resonance components,which can effectively overcome the problem that the number of modes in traditional variational mode decomposition is difficult to determine.Then,the balance factor is optimized through the multi-scale method,and the bandwidth of the mode is adjusted so that the interference signal can be removed to the greatest extent while retaining the effective signal.Simulation and experimental results show that the cascaded variational mode decomposition method can effectively enhance the weak faults of rolling bearings,and by comparing with the traditional variational mode decomposition method,the enhancement effect of fault enhancement is obvious.Next,aiming at the problem of non-linear coupling of fault signal and noise with long transmission paths,a deconvolutional cascaded variational mode decomposition method is proposed to compensate transfer function while removing interference noise and improving the signal-to-noise ratio of the signal.First,based on the traditional deconvolution method(minimum entropy deconvolution),the multipoint kurtosis is introduced as the objective function to construct an adaptive filter,which can effectively overcome the problem of single pulse deconvolution with kurtosis as the objective function and slow iteration speed,and improve the periodic pulse of the signal.Then,the cascaded variational mode decomposition is used to extract the resonance components in the deconvolution signal,and finally the standardized square envelope spectrum is used to extract the characteristic frequency of the rolling bearing fault.The experimental results show that the deconvolutional cascaded variational mode decomposition method can effectively enhance the weak fault signal of rolling bearings,and by comparing with some methods,the weak fault enhancement capability of the method proposed in this paper is more obvious.Finally,aiming at the problem that the pulses excited by the fault under variable speed do not meet the periodicity,which makes it difficult to extract the fault features,a time-domain signal is stabilized by order analysis,combined with deconvolutional cascaded variational mode decomposition method.First,the order analysis is used to convert the time-domain signal into angular-domain,and the fault-excited pulse is changed into a periodic pulse.Second,the multi-point optimal minimum entropy deconvolution is used to compensate the transfer function and improve the periodic pulse of the angular domain signal.The periodic pulses are then determined by cascaded variational mode decomposition to determine the resonance component,and finally the bearing fault is located by standardized square envelope spectrum.The simulation and experimental results show that the angular domain deconvolutional cascaded variational mode decomposition can effectively enhance the weak fault signal of the rolling bearing under the condition of variable speed,and by comparing with other methods,the enhancement effect of fault enhancement is obvious.