Distributed fiberoptic monitoring systems

  • Infrared energy stream

    Infrared energy stream

    Our approach based on the use of the high vibrosensitivity of the infrared energy stream injected into ordinary optical fiber (buried in the ground near the monitoring object) by means of semiconductor laser of low power.

    This optical fiber will be called a fiber optic sensor (FOS). Typically FOS length is 40-50 km. In the systems of this class, all relevant information is transferred to Processing Center (PC) by the optical fiber, which is not only a sensor (FOS) but at the same time an effective and reliable channel for ordinary data transmission.

  • RELSEN

    RELSEN

    We will call the systems of this class as optical fiber classifiers of seismic pulses (RELSEN), which by the principle of operation belong to the multitude of so-called C-OTDR systems.

    So, the vibrosensitive infrared stream injects into a FOS by means of a coherent semiconductor laser at the wavelength of 1550 nm. The simplified scheme of RELSEN represented on Fig. 1. Thus, the laser probes the FOS with usage of infrared stream. This probing is carried out in the pulsed mode. Pulses have a length of ~ 20-200 ns, with an interval of ~ 50-300 μs. The optical fiber is put into the ground, at the depth of 30-50 cm, at the distance of 5-10 m from the monitoring object and, as a matter of fact, it is an optical fiber sensor.

  • Speckle images

    Speckle images

    When a pulse is moving along the optical fiber, the Rayleigh elastic backscattering is realized on its natural irregularities (impurities), which due to high coherence of the used laser of 3B class leads to formation of the so-called stable interference structures of chaotic type, otherwise called speckles or speckle images.

    A sequence of speckles is received in the point of emanation using an ordinary welded coupler or a circulator. The central moment of the concept is the phenomenon that any seismic vibration arising on the surface of the optical fiber due to propagation of seismoacoustic waves from the sources of elastic oscillations, changes its local refractive index. Changes of the local refractive index are reflected in the time-and-frequency structure (TFS) of the respective speckle.

  • Analysis of the sequence

    Analysis of the sequence

    Knowing the pulse duration and the velocity of wave propagation in the optical fiber, it is easy to determine the section where the TFS speckle deviation took place.

    Analysis of the sequence of speckle structures using wavelet conversion apparatuses (the phase of singling out of primary signs of target signals) and Lipschitz classifiers (the phase of classification of target signals) makes it possible not only to reliably detect the target source of seismoacoustic radiation, but also to determine its type and area of occurrence.

    More info

RELSEN main principles

Core RELSEN functionality is based on the modern methods of reflectometric interferential spectroscopy and high vibrosensitivity of a coherent flow of infrared energy that is injected into a dedicated fiber optic cable through a conventional infrared laser with a wavelength of 1550.116 nm. The RELSEN system is designed to operate in a fully autonomous mode and allows detecting and classifying suspicious (target) seismic acoustic activities in the area along protected objects. The detection accuracy of the mechanical activity is ~ 5-10 m. Using ROPA technology (remote optical pumped amplifier, implying the use of four dedicated optical fibers) RELSEN range is up to 75 km. Without ROPA RELSEN range is ~ 40 km. Thus, there is a real opportunity to remotely monitor backbone communications across the country or to monitor of railroad ballast prism over thousands kilometers. 

Coherent reflectometry is a process of analyzing characteristics of Rayleigh backscatter radiation, initially injected into an optical sensor by impulsive coherent infrared laser. A monomode fiber-optic sensor (FOS) is connected to coherent OTDR (optical time-domain reflectometer). The resulting system is designed for simultaneous detection and classification of the seismic activities in the monitoring object vicinity. The analysis process interprets characteristics of the reflected Rayleigh scattered signal as a function of time.

Seismoacoustic sensing technology provides information about detected events that occur at near-surface layers of soil. The targeted signals are created by the structural waves generated by mechanical effects on the soil, or as a result of a seismic activity.

Walking or running man, traffic and excavation including hand digging are typical sources of acoustic emission (structural sound waves). The range of the sound wavelength in this case is from 0 Hz to 1 kHz. Most of the energy signal needed to identify the type of impact is concentrated in the frequency range from of 0 Hz to 400 Hz. Thus, detected micro-seismic effects (target seismic acoustic vibrations) can be induced by various events including pedestrian activity, sounds of a passing automobile and railway transport, digging or strikes on the ground. Targeted seismoacoustic elastic vibrations (SEV or target SEV - TSEV) are identified throughout the entire length covered by RELSEN in the close vicinity of the fiberoptic sensor deployment. SEV are propagated in soil and inducing secondary seismic acoustic vibrations of the same frequency in the optical fiber and causing longitudinal microstrain, which in turn entails local changes in the amplitude and frequency of the Rayleigh backscattered signal distributed by the fiber. Sound pressure exerted on the ground by various events is estimated in decibels.

Value examples of sound pressure exerted on the ground as a result of vibro-acoustic pulse effects are:

  • walking or running man: 10-30 dB;
  • petrol car engine: 10-40 dB;
  • diesel engine (truck): 40-70 dB;
  • passenger traffic vehicle at a speed of 30 km / h: 60-80 dB;
  • truck at a speed of 30 km/h: 70-90 dB;
  • hand digging: 30-50 dB;
  • tracked vehicles: 100-120 dB;
  • mechanical digging : 80 - 120 dB.

In the propagation of an infrared pulse stream along FOS Rayleigh reflection occurs over its entire length, which continuously extends in a direction opposite to the pulsed laser flow direction. Thus, the energy of the reflected signal is about -40 dB with respect to the initial energy injection, but frequency and speed of signal propagation do not change along the FOS. The reflected signal is created by static impurities in the FOS and defects in the fiber microstructure. The reflection is identified by the mechanism of elastic scattering of Rayleigh type, which is one of the main reasons for the infrared energy loss during transmission over fiber optic cables, and is not less than about 0.2 dB / km. Thus, the reflected signal has considerably weaker energy and significantly longer compared with the energy and duration of the probe pulse initially sent by the infrared laser. The same type of a backscatter signal is reflected in the absence of external seismic impacts at the FOS site with length of approximately 5 -10m.

More precisely, the analysis is made on the result of so-called chaotic interference of backscattered radiation scattered by different parts of the fiber. The result of this chaotic interference is called speckle structure. Intensity of temporary speckle-structure at any particular time depends on the phase difference between different components of the scattered light due to Rayleigh backscattering arriving at the photodetector. These components stood out as result of backscattering of light pulse when it was in the appropriate spatial location of the fiber. For simplicity, speckle-structure realization at the specific timeframe is called a background signal. Naturally, the particular type of the background signal depends on the FOS type, level of its isolation, type of soil and number of other factors. It is important to note that all these factors are stable, or so-called long scale factors (LSF). When seismic acoustic emission occurs, it turns out to be a destabilizing factor for the system of reflected signals or so-called short scale factor (SSF). Detection of SSF on the LSF background is the core idea behind RELSEN unit functionality.

image1

Let’s consider in detail a physical nature of the small scale impact on the light flux inside the optical sensor. The fact that the instantaneous intensity of the reflected signal (temporary speckle-structure) depends on the conditions affecting the phase of the light that is available in a very specific position, within a range of the spatial extent of the light pulse. In other words, we are talking about conditions that affect local refractive index of the fiber. Typically, changes in these conditions induce the emergence of the FOS local longitudinal microstrain (LLM). In turn, the LLM is the cause of changes of the FOS local refractive index. For example, the emergence of a local source of elastic waves near FOS location immediately induces seismic-acoustic wave, which reaching the FOS and cause LLM appearance on a relatively small part of FOS. This is where the local refractive index of the fiber will temporarily change and lead to a change in the phase of the light flow. In turn, the phase change implies a difference of luminous flux values ​​of the instantaneous intensity of the speckle patterns corresponding to the same FOS portion but recorded at different times. These points in time, of course, correspond to different light pulses emitted by the FOS laser probing. In a coherent system spectrum of the reflected signal carries information (phase, the spatial distribution of the reflected energy) sufficient to classify physical impact, which has caused distortion of the reflected signal.

Since the initial pulse has well-defined frequency and duration, knowing the speed of its energy propagation in a fiber optic medium can accurately determine not only the FOS area, which has been reflected from the received signal, but also obtain a spectral-temporal pattern that made a distortion to the timing spectrum of the specific FOS. The level and type of the backscatter modulation carries useful information about the type of a seismic impact (SSF) that induced this modulation. Therefore, seismic vibrations implemented in close proximity to FOS will not only be detected, but properly classified as well.

RELSEN fundamental design features:

  • Characteristics of the reflected scattered Rayleigh signal as a function of time (not the time-averaged value) are matched at the registration data array;
  • Reflected signals contain information sufficient for their classification and therefor enable to identify specific SEV type;
  • Pulses have a length of ~ 50-200 ns, with an interval of ~ 50-300 ms;
  • Pulse is filtered by fiber Bragg gratings, which remove amplified spontaneous (spurious) radiation;
  • The energy of each pulse is 0.2 micro joule;
  • Scattered radiation is amplified and filtered to remove any radiation other than one with a central wavelength of 1550.116 nm;
  • Photodiode Gallium Arsenide (GaAs) is connected to an optical fiber and used as a detector to record the amplified radiation;
  • Two pulses are compared at different times in order to determine the difference between reflected signals. This difference indicates a change of physical configurations of fibers (degree of deformation) that occurred on a particular way at the FOS location, due to exposure to acoustic longitudinal microstrains;
  • Difference between pulses is measured using robust correlation methods, including those based on the wavelet transformation; analysis of similarity degree of pulses is used to estimate amount of change in signal parameters caused by the SEV influence.

Physically RELSEN system will include rack with two coherent laser reflectometers, each of which is designed to service FOS with length of 75 km. Thus, one RELSEN unit is able to control about 150 km. FOS is usually deployed in directions opposite to the RELSEN unit location. A special control procedure is used to detect the external seismic impact in a specific area and automatic classification of the TSV, which includes 6-10 classification types. Configuration of target classes is set during RELSEN installation.

image2

An important feature of this technology is its applicability to monitor and protect not only communications infrastructure but also other strategically important large area objects, including oil and gas pipelines, railways tracks and more. Conceptually RELSEN system is implemented into the existing cable facilities with minimal modifications of the latter. The monitoring system includes networked RELSEN units, each of them connected to a FOS.  FOS is using single-mode fiber-optic cables (ITU-T G.652,654 , 655 standards). RELSEN units are placed at communication facilities or simply within cabling throughout the protected object area at intervals of not more than 75 km (specific locations are determined in accordance with characteristics of the existing cabling system topology). One or group of existing fiber cables is used as a FOS.

Mathematical foundation for the seismic activity classification is found in the interference pattern analysis at the C-OTDR output, which is made using Lipchitz Classifiers theory of Confidence. This theory is currently the most powerful method to classify noise-like signals of a quasi-stationary type. During classification of seismic and acoustic signals the special role is played by the feature space analysis: while it is important to make a right choice of this space, so-called «feature selection» phase and optimization phase of the computing feature values, the so-called «feature extraction» phase. The C-OTDR interference pattern (speckle-structure) is analyzed in this phase. Existing approaches to reflectometric data analysis are usually based on different variations of the FFT (Fast Fourier Transform). In contrast, in the framework of this project is used the wavelet analysis.  Wavelet analysis together with Lipchitz classifiers ensures the best quality and reliability of the TSV classification.  In addition, special attention was paid to the so-called macro-features for classification. These marco-features are based on the analysis of the SEV boundaries dynamics. Next diagram shows the sequence of influences and

Stages of Data Processing of RELSEN system:

image2

image2