Data Acquisition And Research Of Optical Fiber Vibration Signal Based On Φ-OTDR

Nowadays,optical fiber sensing technology has developed speedily in China,breaking through the weakness of traditional devices in sensing range and real-time performance.Among them,distributed optical fiber sensing technology based on phase sensitive time domain reflectometer(Φ-OTDR)is widely used.Because Φ-OTDR has the advantages of large sensing length,full distribution and strong real-time performance,it is mainly used in intrusion detection,energy security detection and many other fields.In order to accelerate the productization of Φ-OTDR optical fiber data acquisition system and ensure the signal quality of Φ-OTDR optical fiber data,a set of optical fiber data acquisition scheme is designed in this paper,which can meet the high speed and low delay required by optical fiber data acquisition.It not only realizes the real-time monitoring of vibration signals,but also reduces the cost and volume of the system.The specific research contents of this paper are as follows:(1)By analyzing the generation principle of Rayleigh scattering in the fiber,the arithmetical model of scattered light is put forward,and the definition of Φ-OTDR fiber vibration sensing is explained.The principle of coherent balance detection is analyzed.According to the key system parameters of Φ-OTDR,the signal acquisition scheme of Φ-OTDR vibration detection is designed.(2)According to the above signal acquisition scheme,the components selection,circuit schematic diagram design,high-speed key signal simulation,PCB design,Lab VIEW host computer software design and FPGA program design.(3)Using Φ-OTDR fiber optic sensing system to conduct simulation vibration experiment in the laboratory.In the simulated vibration experiment,the length of sensing fiber is 1.4km,the sampling rate is 100 MSPs,the SNR of the system is 14.5d B,the spatial resolution is 20 m,and the waveform refresh rate of the host computer is 15 Hz.The dynamic response is good,which indicates the effectiveness of the design scheme.(4)Wavelet transform is used to smooth the curve after the difference accumulation,and the filtering effect of random fading is better under the condition of keeping the SNR.In the case of low frequency vibration,the difference between the difference summation method and the kurtosis method is compared.The results show that the kurtosis method is significantly better than the difference summation method at the low frequency vibration of10 Hz,and the SNR is about 3.6d B higher.At the same time,the machine learning model was built to recognize the vibration signals of different frequencies on the fiber.Finally,the recognition rate of the machine learning model for the 1k Hz specific vibration signals was about 98%,which was better than the general difference accumulation method under the condition of sufficient samples.