Fiber Optic Acoustic Emission Demodulation System Based On Matched Grating Method

With the advancement of the process of industrial modernization,the normal operation of rotating machinery and equipment plays a decisive role in the production process,gear and bearings work under harsh conditions for a long time and are prone to multiple failures.Acoustic emission(AE)detection technology can be used for real-time monitoring of early damage to equipment,and it can detect more active defects in structural safety.Fiber Bragg Grating(FBG)sensors have the advantages of immunity to electromagnetic interference,small size,light weight,and easy implementation of large-scale distributed measurement.Therefore,combining the advantages of the two,this paper sets up the optical fiber grating acoustic emission detection system,and uses a variety of demodulation methods to make an effective fault diagnosis for the rotating machinery.In this project,AE signals were collected by FBG AE sensors and piezoelectric ceramics(PZT)sensors.The lead-off test on steel plates,faulty bearing and faulty gear detection experiments were carried out.By comparing the signals collected by the two sensors,we can see that the time-domain waveforms are basically the same.The AE signal arrival time is the same,the sensitivity phase is the same.The experiment proves that both the FBG sensor and the PZT sensor can effectively collect the AE signal.In order to realize the demodulation of the acquisition signal of FBG sensor,the spectrometer demodulation method,the tunable F-P filter demodulation method,the matched grating method and the narrow-band light source method are respectively adopted,which are all based on wavelength modulation,and can perform acoustic-optical-electrical conversion.So as to realize the demodulation of AE signal.A large number of contrast experiments show that the combined demodulation method with the matched grating method and the narrow band light source method has the highest demodulation accuracy,and it is the best demodulation scheme in my subject.Different denoising algorithms are selected according to the difference of signal characteristics,including bandpass,sliding mean,wavelet and Calman filter,in which the wavelet threshold and Calman combined denoising filter can reduce the noise amplitude to the maximum,and keep the useful signal waveform unchanged.Finally,the Hilbert-Huang Transform(HHT)is used to process the faulty bearing signals.First,the Empirical Mode Decomposition is performed to decompose multiple eigenmode functions(IMF),then,the Hilbert envelope spectrum is made for each IMF.The IMF with the highest signal amplitude is the bearing fault signal.In the corresponding envelope spectrum,the peak frequency represents the bearing fault characteristic frequency.Comparing with the theoretical calculation formula,it is possible to push back the fault type of the bearing.The degree of fault can also be judged based on the magnitude of the signal amplitude in the IMF.The error between the test and the calculated value is less than 9.4%.