Multi-Scale Signal Processing Technology In ?-OTDR System

Phase-sensitive time-domain reflectometer(?-OTDR)has the advantages of fast response and high sensitivity,so it can realize the non-contact detection of vibration events using the long-distance transmission of vibration wave and shows great application potential in health monitoring of building structures,safety monitoring of oil/gas pipelines and intruder detection of national borders.?-OTDR can be divided into amplitude-discrimination ?-OTDR and phase-discrimination ?-OTDR according to demodulation object.Amplitude-discrimination ?-OTDR locates vibration events by demodulating the amplitude of Rayleigh backscattering signal(RBS),and phase-discrimination ?-OTDR realizes the quantitative measurement by demodulating the phase of RBS.Studies show that the spatial resolution of the two ?-OTDRs should be matched with vibration events to obtain the best signal-to-noise ratio(SNR).Since different vibration events usually have different lengths,it is necessary for the ?-OTDR system to provide multi-scale observation.Researchers have proposed a multi-spatial resolution ?-OTDR system based on frequency division multiplexing,but it will increase hardware complexity.As one of the signal processing methods,wavelet denoising enables the multi-scale observation in time domain but fail to match vibration events with different lengths in spatial domain.Moreover,due to the discrete wavelet transform(DWT),threshold selection and inverse DWT process performed during processing,it has the disadvantage of large computation load.In order to solve the above problems,this paper proposes a multi-scale matched filtering method.The optimization of system performance could be realized when the scale of matched filtering matches the length of vibration events.For different lengths of vibration events,the filter scale is changed and the SNR of each scale is evaluated to find optimum filter scale.Experimental results have shown that the proposed multi-scale matching filtering method enables amplitude-discrimination D-OTDR to reach the lowest locating uncertainty and the highest locating SNR with low time consumption even under strong noise interference.In addition,a multi-scale phase superposition method is proposed.After phase demodulation,several adjacent phase curves are superimposed along the distance axis.When the number of superposition curves equals the total number of curves contained in the vibration region,the phase SNR reaches its maximum.For different lengths of vibration events,the optimum output of the system can be obtained by changing the number of superposition curves and determining the optimum scale according to SNR.Multi-scale phase superposition greatly improves the phase SNR and the locating effect in phase-discrimination ?-OTDR.In summary,both multi-scale matched filtering and multi-scale phase superposition can effectively suppress noise,provide optimum measurements for unknown events and reduce computation load without increasing hardware complexity,which is of great significance to promote the better applications of ?-OTDR sensing system in engineering.