Research On Bearing Fault Monitoring Based On Fiber Bragg Grating Sensor

Rolling bearings are one of the important parts of mechanical equipment.The health of the bearings affects the working conditions of rotating machinery.Most of the failures of rotating machinery are caused by bearing failures.A failed bearing will indirectly cause accidents and then bring catastrophic consequences.Therefore,it is necessary to carry out effective detection or condition monitoring of early failure of the bearing.In this thesis,rolling bearing is used as the experimental research object.The QPZZ-II rotating machinery fault simulation experimental platform is used to collect the acoustic emission signals of rolling fault bearing with fiber grating sensor and piezoelectric ceramic sensor respectively.The two signals are compared,and the waveform trends are found to be the same,which proves The fiber grating sensor can be used for bearing fault detection,and the bearing fault diagnosis method based on fiber grating sensor is studied.Based on the acoustic emission detection technology and fiber grating sensing technology,as well as the cause of the acoustic emission signal of the rolling fault bearing,in the signal acquisition experiment of the faulty rolling bearing,the fiber grating sensor and the acoustic emission detection technology are combined for bearing fault monitoring.In the experiment of faulty rolling bearing,the performance of fiber grating sensor is very important,However,different packaging forms and packaging materials have an impact on the strain sensitivity coefficient of fiber grating sensors,and the wavelength shift depends on the strain sensitivity coefficient,so the detection of wavelength shift is particularly important.Crucially,in order to determine the strain sensitivity of strained fiber grating sensors and stainless steel packaged sensors,a lead-breaking experiment based on fiber grating sensing was built,and the peak finding algorithm was used to analyze the experimental data,find the wavelength drift,and determine the strain sensitivity,in order to improve peak finding Accuracy,the adaptive peak finding algorithm is improved,the peak finding is fitted with the Gaussian function coefficient optimized by the LM algorithm,and the waveform is corrected with the correction factor α to obtain the peak position.Finally,a stainless steel packaged sensor is selected to collect the faulty bearing signal in the experiment.When the rolling bearing fails,the acoustic emission signal generated is actually a signal superimposed by various modal elastic waves and complex background noise.In order to deal with this nonlinear and non-stationary signal,this thesis uses EEMD and Hilbert envelope spectrum Combine.EEMD is used to decompose multiple inherent modal components,the correlation coefficient between the IMF component and the denoised signal is calculated,the signal with larger correlation coefficient is selected to reconstruct the signal,and then the envelope spectrum of the reconstructed signal is analyzed.For the drifted center wavelength,the improved adaptive peak-finding algorithm is used to improve the peak-finding accuracy.Due to the randomness of the rolling bearing fault signal,the use of ensemble empirical mode decomposition can better reduce the problem of modal aliasing.The peak frequency of the spectrum represents the characteristic frequency of the bearing failure,and the error between the theoretical calculation value is very small.