Phase sensitive optical time domain reflectometer (φ-OTDR) can retrieve vibration waveforms based on linear relationship between phase change and external events. Yet, it is difficult to identify different events due to the complexity of working environment. How to accurately determine the type of vibration events and thus reduce false alarm rate is important in many practical engineering applications. The existing deep learning (DL) algorithm can directly extract the original data feature, without manual extraction. Hence, DL is usually used to classify and recognize multiple events in φ-OTDR. In this work, a dual input deep convolutional neural network (Di-DCNN) is applied to distinguish six kinds of actual vibration events (including walking, tapping, blowing and raining, vehicle passing, digging and background noise). The features of these two inputs are extracted, respectively, and fused together to identify six vibration events. For comparison, network models with other five inputs are employed for event recognition, including single input of 1D time-demain or 2D image of phase (amplitude) data, and dual input of 1D time-demain and 2D image of amplitude data. Here, the 2D image denotes the transformation of 1D data by Markov Transition Field (MTF). Experimental results show that the Di-DCNN with 1D time-domain phase waveforms and its 2D MTF image being the two inputs considerably improve the recognition accuracy. The average recognition rate of six kinds of vibration events is higher than 94%.
Reducing false alarming of vibration and shortening data processing time are one of the key problems of phase sensitive optical time-domain reflectometer (φ-OTDR). Generally speaking, there are two demodulation methods to locate vibrations: phase demodulation and amplitude demodulation. At present, an often-used method is phase-based crosscorrelation, which shows a comparatively reliable detection performance. Compared with phase cross-correlation, energy/power cross-correlation between different positions is simpler and has certain advantages in practical applications. In this paper, we use φ-OTDR to collects periodic vibration signals (power signals) and transient vibration signals (energy signals). Amplitude differentiation is firstly calculated along slow time axis for the Rayleigh backscattering trajectories. For periodic vibration, power spectrum is then obtained at each position, and cross-correlation coefficients between any two spectrums are computed. If the vibration is transient, average energy is calculated along fast time axis and average energy cross-correlation is performed between any two locations. With the cross-correlation values, we are able to determine whether there is vibration on the optical fiber. In the experiment, periodic vibration is simulated by a sine-driven PZT and transient vibration is mimicked by pencil-break. These results demonstrate that power (energy) cross-correlation coefficients work well to locate periodic (transient) vibrations.
Phase-sensitive optical time domain reflectometer (φ-OTDR) has been extensively investigated in fields of intrusion detection and structural health monitoring. It should be noted that phase noises would keep accumulating during pulse transmission. By subtracting an initial phase at the input point from demodulated phases at other positions, the noises related to the laser itself except random noises can be considerably reduced. In order to further decrease the impact of random noises on waveform retrieval of external vibrations, it is necessary to eliminate the accumulated noises before vibration position as much as possible. In this work, a sliding root mean square method (SRMS) is firstly applied to locate vibration events. By the SRMS, the demodulated phase at ~10 m before vibration point is regarded as the modified reference. Then, the vibration waveform can be retrieved after phase subtraction. For comparison purpose, both the input and modified references are employed to retrieve temporal vibration signals. Experimental results show that the SRMS shows good noise performance for vibration location. In terms of signal retrieval, the vibration waveform can be recovered with better noise suppression by the modified reference compared to the input one.
Phase-sensitive optical time-domain reflectometry (φ-OTDR) is highly sensitive to strain changes of sensing fiber caused by external vibration, by which we are able to locate the vibration. In practice, interference fading will inevitably occur in backscattered Rayleigh traces of φ-OTDR due to the use of highly coherent light source, which increase the possibility of failure detection. In order to reduce the influence of interference fading on vibration detection, both frequency-division multiplexing (FDM) and rotated-vector-sum (RVS) over both time-and frequency-domain are employed in our method. Based on the method, we perform φ-OTDR experiment to locate vibrations. By extracting 3 frequency components of the beating signals (~200 MHz) and carrying out dual rotation, interference fading can be suppressed to a large extent, the vibration-induced phase changes are precisely recovered. One point should be noted is that we found that there is a certain correlation between each frequency component extracted from the beating signal, resulting in interference fading points cannot be completely removed.
Phase-sensitive optical time-domain reflectometer (φ-OTDR) is widely used for safety monitoring of large-scale civil objects, by which external vibrations along the sensing fiber can be detected. It has to be noticed that the category of vibration signals should be accurately distinguished for many real applications. At present, an extensively approach of signal recognition is deep convolutional neural network (DCNN). In the work, the one-dimensional DCNN (1D-DCNN) is applied to recognize different sound-induced vibrations based on their time-domain intensity signals detected by an amplitude-demodulated φ-OTDR system. It is turned out that the DCNN successfully shows the capability of recognizing walking, rock drill, explosion, hand hammer, car siren, and background noises with a high accuracy. Additionally, the 1D time-domain intensity vectors are rearranged into 2D matrices and the 2D-DCNN is accordingly employed to identify these vibration signals. The confusion matrices demonstrate that the 1D-DCNN has a higher average recognition accuracy to identify the concerned sounds with respect to the 2D-DCNN.
Four-wave mixing (FWM) in few-mode fibers (FMFs) has been extensively investigated to develop mode-related alloptical signal processing, such as wavelength conversion, parametric amplification and mode conversion. Compared to the FWM processes in single-mode fibers, intermodal FWM in FMFs shows more flexible phase-matching condition by tailoring the modal dispersion of each optical mode. Generally, there are two mainly different types of FWM processes, namely, Bragg scattering (BS) and phase conjugation (PC). In this paper, we focus our interest on the PC-FWM in both graded-index (GI) and step-index (SI) FMFs to probe mode conversion. In the PC-FWM, the energy transfers from pump modes to both signal and idler waves. From the point of phase matching, the modal dispersions of the two FMFs is firstly optimized by genetic algorithm (GA) to design optimal core radius and core-cladding refractive difference. We then investigate the effect of the small deviations of these two parameters from their optimal values upon the phase mismatch. Numerical results show that both SI and GI fibers are able to convert the LP01 mode to the LP02 mode with the phase matching condition of the SI fiber being more sensitive to the changes of fiber parameters. In addition, we analyze the dependence of mode conversion performance (bandwidth and efficiency) on fiber length and pump level. It is shown that the 3dB-bandwidth increases with the pump power in the PC-FWM, which can be attributed to the nonlinear phase shift induced by the high pump power compensate for the linear phase mismatch.
In practical application, it is found that single-path phase-sensitive optical time-domain reflectometers (φ-OTDR) is susceptible to noise and random interference, which increases the probability of missing detection over external perturbations by conventional amplitude demodulation. In the work, a dual-channel system based on two fibers extracted from an armored four-core cable is investigated to enhance the robustness of the φ-OTDR. In signal demodulation, by combining the conventional differential accumulation algorithm (DAA) and standard deviation algorithm (SVA) a multipath information fusion algorithm (MIFA) is accordingly proposed to conclude whether the vibration signal is present. The MIFA-based dual-channel φ-OTDR is experimentally demonstrated on a highway of 9 km to position a running vehicle, indicating a considerable performance improvement of vibration identification compared to the DAA and SVA.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.