Demodulation Analysis of Distributed Vibration Sensor Signals Based on Fast Fourier Transform
Si Zhaopeng,Mao Bangning,Bu Zehua,Gong Huaping,Xu Ben,Kang Juan,Yang Chunjun,Zhao Chunliu
DOI: https://doi.org/10.3788/CJL220902
2023-01-01
Chinese Journal of Lasers
Abstract:Objective Phase- sensitive optical time-domain reflectometry is a type of distributed optical fiber sensing technology that has become one of the most rapidly developing sensing methods in modern sensing technology because of its wide monitoring range, high sensitivity, low monitoring cost, and many measuring parameters. Compared with other distributed optical fiber sensing technologies, this technology can perform distributed multipoint measurement of weak signals over long distances. Distributed acoustic sensing (DAS) technology uses coherent Rayleigh backscattering technology in common single-mode sensing fibers that can realize the continuous detection of remote external vibration, sound, and temperature changes. This technology has strong environmental applicability, electromagnetic interference resistance, chemical corrosion resistance, and good concealment. Therefore, research on the DAS technology of a phase sensitive optical time domain reflectometer (f -OTDR) system has great scientific and practical significance for social engineering applications. A DAS system typically adopts IQ and Hilbert demodulation, but both have certain defects. Although the IQ demodulation method has a simple structure, its demodulation results are easily affected by phase noise caused by the frequency drift of narrow linewidth laser and polarization instability of local light and Rayleigh scattering light, decreasing the demodulation accuracy. Hilbert transform can be used in DVS and DAS systems, but the low noise term cannot be eliminated in low pass filtering; therefore, the anti- noise performance is poor. In order to address the limitations of IQ and Hilbert transform demodulation methods, a signal demodulation method based on fast Fourier transform ( FFT) is adopted in this study to improve the signal-to- noise ratio of vibration signals and to reduce the interference of low frequency noise, pulse optical coherence fading, and harmonics. Methods The digital signal collected by the acquisition card is first passed through a belt pass filter such that the signal frequency near the intermediate frequency signal passes through, and then, the step size is set according to the range resolution of the DAS detection system. To prevent signal vibration at the connection of two steps, the phase of vibration signal extraction is discontinuous. The set step size is transformed by FFT, and the frequency point is set according to the intermediate frequency signal of an acoustooptic modulator. Then, the phase information of the vibration signal is extracted. The signal phase is differentiated to eliminate the phase noise of the signal, and then, the phase signal is unwound in the fiber and time directions. Finally, the extracted phase signal is processed by high- pass filtering to eliminate the influence of system components and low-frequency noise. Results and Discussions Compared with IQ and Hilbert transform demodulation methods, the FFT signal demodulation method has strong phase periodicity, uniform period distribution, better amplitude stability, and less influence from signal coherent fading (Fig. 5). The phase signals demodulated by IQ and Hilbert transform have poor amplitude stability and uneven periodic distribution, and the amplitude of the demodulated signals extracted by both IQ and Hilbert transform demodulation methods is smaller than that extracted by FFT demodulation. Further, the frequency spectrum of the FFT demodulation signal shows that the phase of the vibration signal is not disturbed by harmonic and low frequency noise. Both IQ and Hilbert transform demodulation methods are disturbed by many harmonics and low frequency noise (Fig. 6). Furthermore, the signal- to-noise ratio ( SNR) of vibration signals extracted by three demodulation methods is compared; the FFT demodulation method has the highest SNR (Fig. 7). Conclusions In this study, we adopted a DAS signal demodulation method based on FFT. The phase information of vibration signals is extracted by FFT for the vibration of the wrapped phase difference. Then, the phases secondary wrapped in fiber and time directions are obtained, and finally, a DC signal is achieved with high- pass filtering processing. The 200 Hz sinusoidal periodic signal loaded on PZT is demodulated successfully. Compared with the IQ and Hilbert demodulation methods, the phase signal demodulated by FFT is less affected by the signal coherent fading, and the frequency information display signal is not interfered by harmonics. The signal demodulated by the IQ and Hilbert demodulation methods has good periodicity, and the phase amplitude of FFT demodulation method is larger than that of the IQ and Hilbert demodulation method. The SNR of 200 Hz signal extracted by FFT is 36. 71 dB, that is, 21.04 dB and 20.91 dB higher than that extracted by IQ and Hilbert transform, respectively. Moreover, this method filters the interference of low frequency noise. For sinusoidal vibration signal extraction at 2000 Hz and 5000 Hz, the experimental results show that this method has good applicability.