WO2016198683A1 - Method and apparatus for monitoring pipeline using an optical fiber sensor system - Google Patents
Method and apparatus for monitoring pipeline using an optical fiber sensor system Download PDFInfo
- Publication number
- WO2016198683A1 WO2016198683A1 PCT/EP2016/063442 EP2016063442W WO2016198683A1 WO 2016198683 A1 WO2016198683 A1 WO 2016198683A1 EP 2016063442 W EP2016063442 W EP 2016063442W WO 2016198683 A1 WO2016198683 A1 WO 2016198683A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- optical fiber
- values
- disturbance
- scattered
- condition
- Prior art date
Links
- 239000013307 optical fiber Substances 0.000 title claims abstract description 80
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000012544 monitoring process Methods 0.000 title description 6
- 230000005855 radiation Effects 0.000 claims description 29
- 230000003287 optical effect Effects 0.000 claims description 13
- 230000002123 temporal effect Effects 0.000 claims description 11
- 238000011161 development Methods 0.000 claims description 9
- 238000005070 sampling Methods 0.000 claims description 6
- 230000004913 activation Effects 0.000 abstract description 2
- 238000013528 artificial neural network Methods 0.000 description 23
- 239000000835 fiber Substances 0.000 description 21
- 230000000694 effects Effects 0.000 description 16
- 210000002569 neuron Anatomy 0.000 description 16
- 238000004422 calculation algorithm Methods 0.000 description 12
- 230000000875 corresponding effect Effects 0.000 description 12
- 230000008859 change Effects 0.000 description 7
- 230000018109 developmental process Effects 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 210000000225 synapse Anatomy 0.000 description 7
- 238000001514 detection method Methods 0.000 description 6
- 238000010079 rubber tapping Methods 0.000 description 6
- 230000033001 locomotion Effects 0.000 description 5
- 238000001228 spectrum Methods 0.000 description 5
- 238000005452 bending Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000009412 basement excavation Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000005553 drilling Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000002347 injection Methods 0.000 description 2
- 239000007924 injection Substances 0.000 description 2
- 210000002364 input neuron Anatomy 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000010355 oscillation Effects 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 241000735215 Lepidocybium flavobrunneum Species 0.000 description 1
- 241001465754 Metazoa Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000036982 action potential Effects 0.000 description 1
- 238000001994 activation Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000010009 beating Methods 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 210000003169 central nervous system Anatomy 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000011109 contamination Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000002964 excitative effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000009313 farming Methods 0.000 description 1
- 210000003608 fece Anatomy 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000010006 flight Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000007620 mathematical function Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000005693 optoelectronics Effects 0.000 description 1
- 210000004205 output neuron Anatomy 0.000 description 1
- 231100000572 poisoning Toxicity 0.000 description 1
- 230000000607 poisoning effect Effects 0.000 description 1
- 230000000135 prohibitive effect Effects 0.000 description 1
- 230000001681 protective effect Effects 0.000 description 1
- 230000002829 reductive effect Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H9/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by using radiation-sensitive means, e.g. optical means
- G01H9/004—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by using radiation-sensitive means, e.g. optical means using fibre optic sensors
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D5/00—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable
- G01D5/26—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light
- G01D5/32—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light
- G01D5/34—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light the beams of light being detected by photocells
- G01D5/353—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light the beams of light being detected by photocells influencing the transmission properties of an optical fibre
- G01D5/35338—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light the beams of light being detected by photocells influencing the transmission properties of an optical fibre using other arrangements than interferometer arrangements
- G01D5/35354—Sensor working in reflection
- G01D5/35358—Sensor working in reflection using backscattering to detect the measured quantity
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/02—Mechanical actuation
- G08B13/12—Mechanical actuation by the breaking or disturbance of stretched cords or wires
- G08B13/122—Mechanical actuation by the breaking or disturbance of stretched cords or wires for a perimeter fence
- G08B13/124—Mechanical actuation by the breaking or disturbance of stretched cords or wires for a perimeter fence with the breaking or disturbance being optically detected, e.g. optical fibers in the perimeter fence
Definitions
- the invention relates to an optical fiber sensor system and a method for determining a location and type of a disturbance along a pipeline.
- Oil, water, gas and other product pipelines form a critical network in every part of the world and the pipelines form an easy target for intruders.
- the pipelines are also susceptible to earthquakes, tsunamis and to other geohazard incidents.
- Monitoring the pipelines in order to keep the pipelines safe from damage is a major challenge.
- the long distances, often through remote and hostile territory, make the costs of most conventional monitoring systems prohibitive. If an oil or gas pipeline is damaged it can have devastating impacts on human life and health due to explosions, fire and contamination; environment due to poisoning of flora and fauna, as well as the associated financial losses and damage to both image and reputation.
- the pipelines are susceptible to various types of third-party interference, such as deliberate acts (illegal tapping, sabotage) or unintended disruption (construction work, farming).
- This third-party interference can cause huge financial and environmental damage and loss of reputation to the pipeline operators.
- a reliable, real-time pipeline monitoring system is therefore required in order to protect nature, human health and economic interests.
- Illegal tapping is a major problem in emerging countries like India, China and South America.
- Petroleos Mexicanos (PEMEX) counted 1,324 cases of illegal tapping in Mexico. Every day Pemex estimates 40,000 liters of oil and gas, which sums up to an annual damage of more than 1 billion US$ of lose. Such damage could be avoided or substantially reduced, if third party interference was detected before or close to the occurrence of the interference happens.
- the acoustic disturbance can be representative of damage to the pipelines through third party interference or geohazards, as described in the introduction. These systems involve the use of an optical fiber laid alongside the pipeline, which acts as a sensor and detects changes in the pattern of back- scattered radiation in order to sense an acoustic disturbance.
- Pimon GmbH, Kunststoff, Germany sells an apparatus PMS2500-vibrO that utilizes distributed fiber optical sensing technology to detect the acoustic disturbance.
- the PMS2500- vibrO system combines an optical time domain reflectometer (OTDR) with an analysis and pattern recognition software and offers a customized interface with geographic information system (GIS) mapping.
- OTDR optical time domain reflectometer
- GIS geographic information system
- ANN artificial neural network
- the ANNs are computational models and are inspired by animal central nervous systems, in particular the brain, that are capable of machine learning and pattern recognition.
- the ANNs are usually presented as a system of nodes or “neurons” connected by "synapses” that can compute values from inputs, by feeding information from the inputs through the ANN.
- the synapses are the mechanism by which one of the neurons passes a signal to another one of the neurons.
- ANN For the recognition of handwriting.
- a set of input neurons may be activated by pixels in a camera of an input image representing a letter or a digit. The activations of these input neurons are then passed on, weighted and transformed by some function determined by a designer of the ANN to other neurons, etc. until finally an output neuron is activated that determines which character (letter or digit) was imaged.
- ANNs have been used to solve a wide variety of tasks that are hard to solve using ordinary rule-based programming, including computer vision and speech recognition.
- a class of statistical models will be termed "neural” if the class consists of sets of adaptive weights (numerical parameters that are tuned by a learning algorithm) and are capable of approximating nonlinear functions of the inputs of the statistical models.
- the adaptive weights can be thought of as the strength of the connections (synapses) between the neurons.
- the ANNs have to be trained in order to produce understandable results.
- a supervised learning the learning paradigms all have in common that a set of pre-analyzed data, for example a waveform, is analyzed by the ANN and the weights of the connections (synapses) between the neurons in the ANN are adapted such that the output of the ANN is correlated with a known event.
- An improvement in the efficiency of the results of the ANN can be obtained by using a greater number of data items representing the known event in a training set.
- the greater number of data items require, however, an increase in computational power and time for the analysis in order to get the correct results. There is therefore a trade-off that needs to be established between the time taken to train the ANN and the accuracy of the results.
- Deep-learning is a set of algorithms that attempt to use layered models of inputs. Jeffrey Heaton, University of Toronto, has discussed deep learning in a review article entitled 'Learning Multiple Layers of Representation' published in Trends in Cognitive Sciences, vol. 11, No. 10, pages 428 to 434, 2007. This publication describes multi-layer neural networks that contain top-down connections and training of the multilayer neural networks one layer at a time to generate sensory data, rather than merely classifying the data.
- Neuron activity in the prior art ANNs is computed for a series of discrete time steps and not by using a continuous parameter.
- the activity level of the neuron is usually defined by a so-called "activity value”, which is set to be either 0 or 1, and which describes an 'action potential' at a time step t.
- the connections between the neurons, i.e. the synapses, are weighted with a weighting coefficient, which is usually chosen have a value in the interval [-1.0, + 1.0].
- Negative values of the weighting coefficient represent "inhibitory synapses" and positive values of the weighting coefficient indicate “excitatory values”.
- the computation of the activity value in the ANNs uses a simple linear summation model in which weighted ones of some or all of the active inputs received on the synapses at a neuron are compared with a (fixed) threshold value of the neuron. If the summation results in a value that is greater than the threshold value, the following neuron is activated.
- a fiber sensor system for determining a location and type of a disturbance along a pipeline.
- the fiber sensor system comprises an optical fiber, a radiation source for launching radiation into the optical fiber, a detector for detecting back- scattered radiation back- scattered from the optical fiber, and a signal processor connected to the detector for generating a plurality of values from the back- scattered radiation, corresponding to the condition of the optical fiber at a certain time at a certain location of the optical fiber.
- the signal processor comprises a recorder to record the values corresponding to the condition of the optical fiber at a specific repetition rate, an analyzer to analyze the change of those values per repetition over time and create a value over time pattern corresponding to the specific disturbance.
- the radiation source as mentioned above is in one aspect designed to produce a series of optical pulses.
- the fiber sensor system comprises a memory for storing a plurality of patterns representative of known disturbances, and a means to compare the value over time patterns corresponding to the specific disturbance as created with the recorded patterns representative of known disturbances.
- the radiation source produces infrared light in a series of optical pulses.
- the specific repetition rate at which the values corresponding to the condition of the optical fiber are recorded is 250Hz.
- Also disclosed is a method for determining a location and type of a disturbance along a pipeline comprising the steps of launching radiation into an optical fiber buried along a pipeline, detecting back- scattered radiation back- scattered from the optical fiber, generating a plu- rality of values from the back- scattered radiation, corresponding to the condition of the optical fiber at a certain time at a certain location of the optical fiber; recording the values corresponding to the condition of the optical fiber at a specific repeti- tion rate, analyzing the development of those values per repetition over time; and creating a value over time pattern corresponding to the specific disturbance.
- the method for determining a location and type of a disturbance further comprises the step of comparing the value over time pattern with recorded patterns of known disturbance types.
- the values corresponding to the condition of the optical fiber are recorded at a preferred specific repetition rate of 250Hz.
- the preferred spatial sampling rate is 1 meter (this corresponds to 100MHz ADC sampling frequency).
- Other temporal and spatial sampling rates may be used, as the sampling frequencies may be varied within a certain range.
- the optical fiber sensor system taught in this disclosure is able to provide: alert in case of emergency with an accurate GIS location, identify threats in real time in order to reduce the risk from leakages caused by digging, drilling, tapping, sabotage, earthquake etc., monitor around 50km of pipeline with one measuring unit without any additional power supply and with scalability to cover deployments of any length by adding additional units.
- the system reduces false or irrelevant alarms and is capable of differenti- ating multiple events down to 5-15 meter resolution; as well as track potential intruders, moving in different velocities (vehicles or on foot) along the pipeline.
- the system can detect direction and speed of the possible intruders.
- Fig. 1 shows an example of an optical fiber sensor used in the system of the disclosure.
- Fig. 2 to 6 illustrate graphs showing various outputs generated by processing of a back- scattered signal in a processor.
- Fig. 7 is a flow chart, describing the steps of the method of this disclosure. DETAILED DESCRIPTION OF THE INVENTION
- the optical fiber sensor system 10 of this disclosure is based on distributed fiber optical sensing and integrates an optical time domain reflectometer (OTDR) with detection software, as well as a customized interface with geographic information system (GIS) mapping.
- OTDR optical time domain reflectometer
- GIS geographic information system
- the fiber optic cables are often already buried near to or attached to the pipeline for telecommunication purposes.
- the system turns a standard single mode telecommunication fiber into a listening device for events of interest.
- the optical time-domain reflectometer is an optoelectronic instrument used to characterize optical fibers.
- the reflectometer injects series of optical pulses into the optical fiber that is being tested and detects at the same fiber end the returning ("back- scattered") light that has been scattered or reflected back from points along the fiber.
- the OTDR produces a so-called reflectogram from this measurement and performs an efficient, precise and wide analysis of the fiber characteristics.
- a laser diode launches radiation in the form of a plurality of optical pulses into one end of the optical fiber and a photodiode measures the returning light.
- the photo diode is part of the OTDR.
- the back- scattered signal detected by the photodiode provides relevant information about the events of interest along the optical fiber.
- the back- scattered signal is influenced by both the attenuation and reflections, which the injected laser pulse experi- ences on its way through the optical fiber. If a certain area within the optical fiber has a higher degree of attenuation or reflection, for instance caused by a bending or a fiber connector, this higher degree of attenuation or reflection is detected by differences in the back- scattered signals.
- the velocity of the optical pulse is known and the exact location of the event of interest can therefore be determined on the basis of the time difference between the injection of the optical pulse and the return of the back- scattered signal.
- the optical fibers make a good sensor, as the optical fibers are able to measure vibrations over long distances.
- the reflections of the signals in the optical fibers change on vibrations (e.g. caused by vehicles, footsteps, digging, drilling), on temperature alterations (e.g. caused by escaping pressurized gas), or when optical fibers get strained, bent, kinked, or cut off.
- Fig. 1 shows an example of the optical fiber sensor system 10 using the teachings of this disclosure.
- An optical fiber 20 used in this sensor system 10 can be a specially laid optical fiber 20, which has been placed in a region of interest.
- the optical fiber 20 can also be a standard telecommunications optical fiber, which generally carries data.
- the optical fiber 20 can be laid near or directly adjacent to a pipeline 25, as outlined in the introduction to the description.
- the optical fiber 20 is connected to a radiation source 30, which in one non- limiting aspect of the invention is a semiconductor laser producing radiation 35 comprising a plurality of optical pulses at 1.55 ⁇ and launches the radiation 35 into the optical fiber 20 at repetition rates up to 4kHz in the present embodiment.
- a radiation source 30 which in one non- limiting aspect of the invention is a semiconductor laser producing radiation 35 comprising a plurality of optical pulses at 1.55 ⁇ and launches the radiation 35 into the optical fiber 20 at repetition rates up to 4kHz in the present embodiment.
- the returning back- scattered signal 37 from the optical fiber 20 can be observed at the same end of the optical fiber 20 in form of a reflectogram by a detector 40 in which the time delay of signal incidence is shown as distance on the optical fiber 20.
- a highly coher- ent laser as the radiation source 30 is used in order to increase the sensitivity. Due to this high level of coherence, an interference pattern can be observed in the back-scattered signal 37 (fingerprint). Vibrations cause bending in the optical fiber 20 and result in temporary changes in this interference pattern.
- the interference pattern can be used to detect a disturbance 45, for example a third-party interference caused by manual or machine digging. Every kind of incident causes certain multiple vibrations and changes the back- scattered signals 37 leading to differences in the interference pattern.
- the back-scattered signals 37 are received and the signal patterns are digitalized.
- the signal in the detector 40 passes through to a signal processor 50.
- the signal processor 50 is able to classify the dis- turbance 45 and send an alarm to the system operator, if required.
- the signal processor 50 is designed to separate all "regular" sounds from possible disturbances 45 in the back-scattered signal as well as any inoffensive or other irrelevant incidents.
- the processor 50 further includes an input device (not shown) that is used to input information items relating to the reflectogram.
- the information items may include a name or a label generally attached to the reflectogram and/or to one or more features in the reflectogram.
- the input device is connected to a signal processor (not shown) having a memory.
- the signal processor compares the characteristics relating to a particular feature in a stored reflectogram with the inputted information and can associate the particular stored reflectogram (or a portion thereof) with the inputted information. This association is memorized so that if an unknown pattern or feature in a reflectogram is detected by the detector 40, the processor 50 can determine that this unknown pattern is in fact a known pattern and output the associated item of information to the user at an output.
- the pattern recognition system 10 can be trained to recognize a large number of patterns in the reflectogram using an unsupervised leaning process. This unsupervised learning process is carried out using a set of pre-stored patterns representing features of interest from existing reflectograms and running the pre-stored patterns through the processor 50. [0045] The system and method of the current disclosure can be used to determine and classify unknown disturbances 45, as shown in Fig. 2 to 6.
- the radiation 35 is launched into the opti- cal fiber 20.
- the detector 40 receives the back- scattered signal 37 and passes the back- scattered signal 37 to the processor 50.
- a label can be associated with the one or more features in the back- scattered signal.
- the procedure is repeated for a different feature. This different feature creates a different structure within the plurality of layers 180. The learning procedure can then proceed using different ones of the features.
- the measuring data to be evaluated comprises a value matrix spanned by sampled fiber range and the temporal sample steps within a certain observation time. It must be distinguished between detection of location and detection of time of the fiber (20). For each laser pulse injected into the fiber (20), the back-scattered signal is detected. Via the time elapsed between injecting the pulse and detection of a certain part of the back scattered signal, the location is determined by means of the run time of the light (OTDR-Method), from which the scattered signal emanates. An analogue-digital-converter scans the back scattered signal at 100MHz, so that for each meter of fiber a back scattered value can be allocated. The back scattered values for the entire fiber (e.g.
- a fingerprint is a line in the matrix to be evaluated.
- the temporal development of the fiber is obtained by repeated scanning of the fiber, i.e. by repeatedly injecting laser pulses with a repetition rate of e.g. 250Hz.
- the temporal development of a fiber location is derived from evaluating subsequent fingerprints at this location. The values at such a location form the columns of the matrix and are called time slice.
- the evaluation is carried out for the example of seven pickaxe beatings 201 to 207 carried out at a sample installation of the device according to the disclosure. Due to the specific physical layout of the sample installation, the same action is detected in three positions, namely at 175m, 230m and 530m from the measuring distance.
- the classification of disturbances on the optical fiber 20, which is due to activities around the optical fiber 20, is carried out by mainly investigating the temporal development or by recognizing patterns within a single one of the time slices, wherein intermediate results of adjoining time slices are also considered (stripe condition).
- the following steps are carried out:
- Step 1 Use of a Laplace operator.
- This operator generates positive or negative peaks at those locations, at which the input signal shows a substantial change. Small changing values are those values close to zero and steady input values are those values that are precisely zero.
- This Laplace operator improves the signal-to-noise ratio of the measured values.
- the one dimensional Laplace operator is a vector of coefficients which are each to be multiplied with one out of an equal number of subsequent input values. Similar to a scalar product the sum of all products ensues the output value at the position the center coefficient has been applied to. The number of coefficients is odd.
- the coefficients comply three requirements: The middle one has a positive maximum value. The coefficients are symmetric with regard to the central one and behave like a Gaussian or a Lorentz distribution. The sum of all coefficients is zero.
- FIG. 3 the impact of the Laplace operator is shown.
- the graph in Fig. 3c shows the Laplace data, which arise from the signal of a pickaxe, as shown in Fig. 3a.
- the amount is formed by the result of the Laplace operator, so that one works with positive peaks only, see the graph in Fig. 3d.
- Step 2 The Laplace output often contains solitary peaks which are not caused by a digging stroke. Stroke events instead show a whole cluster of peaks that enables one to locate the time of the hit by locating the highest density of peaks in such a cluster. In order to neglect single peaks and to identify the most likely hit time the amount of Laplace out- put within a time slice is averaged as described below. By doing so single peaks are dwarfed and one can assign a hit time to each peak cluster. The latter is a local maxima of the averaged time slice. In order to achieve the above-quoted averaging of the absolute Laplace data, a sliding average is formed for every single point within the time slice, e.g.
- Step 3 Definition of a threshold as a function of the noise background, by multiplying the noise average with an eligible factor, and sorting out the over threshold values, as can be seen in Fig. 3d.
- Step 4 Use of a "step condition": Values which are over the defined threshold val- ues can also be generated by a person walking around over the optical fiber 20 or by a car near the optical fiber. To be considered to be a blow of a pickaxe or a spade, over threshold values must fulfil two additional prerequisites, the step condition and the stripe condition.
- the step condition: Starting point or symmetric point is the temporal position of values which are over the threshold values found in the 3rd step.
- an average is formed in form of the values in the raw signal over the value area (measuring values before step 1) in each case.
- the step condition requires that the amount of the difference of both averages may not surpass a certain value. If it does, the peak is at or close to a clear step in the basic level of the raw signal. Steps can be seen in the signal of a pedestrian. They are, however, extremely rare for a pickaxe signal, which is shown in the graph in Fig. 3a.
- Step 5 Use of a "stripe condition": the remaining values are now compared with other, neighboring values, in the value matrix, meaning that a value will be compared to other values in this time slice, but also with values from the neighboring ones of the time slices.
- This stripe is given by an eligible delta t and an eligible delta x.
- a value fulfils the condition if there are suffi- cient other values next to it in the stripe area as defined by delta t and delta x, so that a defined number of values is reached in a "cluster", e.g. 5.
- Values 201 to 207 shown in Fig. 2 have gone through all those steps. These values can be supposed to have been created in the raw signal due to an excavation or a pickaxe activity.
- Signals of pedestrians can be classified into those signals created by walking across the sensor cable and those signals originating from walking along the sensor cable. If someone walks across the sensor cable, at right angles to the sensor cable, patterns of chronologically separated steps can be recognized in the affected time slice (see Fig. 4 401). Even if the pedestrian walks on the same spot, the measured signal looks like signal 401 in Fig. 4. It will be noted that an offset has been added to the graphs in order to separate the lines clearly from each other in Figs. 4 and 5. This is merely for reasons of clarity.
- a clear deviation from the raw signal which usually shows a length of about 0.5 seconds for single steps is typical.
- This deviation in the raw signal can be formulated as an additional condition in the algorithm described above for the detection of a blow by a pickaxe. If the step condition described above was fulfilled, and the event was rejected as a blow, a pedestrian step event can be ascertained by examining this additional condition of the deviation from the raw signal. This is valid only for single pedestrian steps. If the pedestrian runs along the sensor cable, the steps are not so clearly separated from each other. Then the movement of the pedestrian can be detected by use of the following algorithm.
- a spline algorithm is used to show movements of the pedestrians or the vehicles over the optical fiber 20.
- steps in the raw data or displacements are typical for other movements in the signal.
- the spline algorithm is a mathematical function, which interpolates the development of the measuring data. Narrow peaks in the signal as well as noise are ignored.
- the algorithm comprises the following steps (see Fig. 4 and 5): [0063] Step 1: Filtering out or suppression of narrow peaks. Should a value differ clearly in the raw signal from the average of the values in close proximity, the value is substituted with the average. This is illustrated in the graph 402 and 502 in Fig. 4 for two single steps of a pedestrian and Fig. 5 for a pedestrian walking along the optical fiber, in which a sub- stitution graph is constructed from the raw data in the upper graph 401 and 501.
- Step 2 Smoothing of the data by use of a gliding average as in graph 403 and 503).
- Step 3 Calculation of a spline, which returns only the substantial changes of the signal, graphs 404 and 504.
- Step 4 The amount of the derivative (slope) of the spline (graph 404 in Fig. 4 and 504 in Fig. 5) is always high when there are clear steps in the raw signal, as graph 405 in Fig. 5 and 505 in Fig. 5 indicate.
- Step 5 Display of the values, which are larger than a certain threshold value 410 and 510, as a signal of the pedestrian, graphs 405 and 505.
- Fig. 6 the signal of the pedestrian is shown that walks along the sensor cable in one direction and then turns back. It will be noted that the signal of one pedestrian is shown mirrored in Fig. 6. This is an artefact of the system used. The pedestrian walks first 10-15 meters along the cable and then returns to the starting point. This leads to the slightly crescent- shaped figure.
- Any stationary vehicles with diesel engines tractors, JCBs, trucks
- the Fourier transformation is calculated in a calculation window (segment of a time slice, e.g., 512 values) and the power spectrum derived therefrom.
- the basic oscillation and harmonics of the engines are visible as peaks around the power spectrum.
- the Fourier- transform-algorithm For every monitored fiber position the Fourier- transform-algorithm is applied. This means the previous 512 values of each time slice are taken as the input vector of Fast Fourier Transform. This gives an output vector of 512 complex values. Deriving the square of the absolute output values gives the frequency- power spectra of the harmonic components in the input between zero and the Nyquist- frequency. This is half the frequency that has been used when recording the input values that means for a 250Hz temporal sampling rate the Nyquist-frequency is 125 Hz. A high peak at zero Hz is caused by the constant input signal offset and is to deduct or to neglect. The harmonic vibration of a diesel engine generates clear peaks in the frequency power spectra. The fundamental frequency and its higher-order harmonics are to observe. For setting an alarm state a suitable threshold can be defined that the peak in the power spectra has to surpass in order to switch to an alarm state.
- Moving vehicles can be distinguished from the pedestrians as a function of their weight and their speed by the size of a segment of the optical fiber in which the moving vehicles cause a signal. While the pedestrian crossing the signal generated in the sensor cable is seen in a response area of 3 to 5 meters along the sensor cable, the signal generated in the sensor cable relating to a crossing car is between 10 to 12 meters of the response area. The signal relating to a truck has a response area which is bigger than 20 meters.
- the affected time slices show a typical wave- form, like half an oscillation of a sinus wave, of the measuring signal.
- a gliding average averaging over 50 to 100 values
- the wave form is preserved while the background noise and isolated high peaks disappear.
- Such a wave form is not included in the pedestrian's signal and, hence, points with all probability to a moving car.
- the FWHM of the wave form points to the time needed to cross the cable.
- the system 10 can give an appropriate warning and human intervention can be initiated in order to review the disturbance and classify the unknown feature or to resolve in the other conflicts.
- Fig. 7 is a flow chart, describing the steps of the method of this disclosure.
- the method for determining a location and type of a disturbance along a pipeline comprises the following steps :
- the first step 701 comprises the injection of radiation into an optical fiber buried along a pipeline, followed by the detection of the back-scattered radiation in the next step 702, which was back- scattered from the optical fiber and which is generating a plurality of values, corresponding to the condition of the optical fiber at a certain time at a certain location of the optical fiber;
- the third step 703 comprises the recording of the generated values corresponding to the condition of the optical fiber at a specific repetition rate, analyzing the development of those values per repetition over time; and creating a value over time pattern corresponding to the specific disturbance; the location and the type of a disturbance is determined in the fourth step 704 by comparing the value over time pattern with recorded patterns of known disturbance types.
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
An optical fiber sensor system and a method for determining a location of a disturbance having a signal processor with a plurality of activation cells adapted to react to components of a back-scattered signal and label the disturbance.
Description
METHOD AND APPARATUS FOR MONITORING PIPELINE USING AN
OPTICAL FIBER SENSOR SYSTEM
BACKGROUND OF THE INVENTION Field of the Invention
[0001] The invention relates to an optical fiber sensor system and a method for determining a location and type of a disturbance along a pipeline. Introduction to the Invention
[0002] Oil, water, gas and other product pipelines form a critical network in every part of the world and the pipelines form an easy target for intruders. The pipelines are also susceptible to earthquakes, tsunamis and to other geohazard incidents. Monitoring the pipelines in order to keep the pipelines safe from damage is a major challenge. The long distances, often through remote and hostile territory, make the costs of most conventional monitoring systems prohibitive. If an oil or gas pipeline is damaged it can have devastating impacts on human life and health due to explosions, fire and contamination; environment due to poisoning of flora and fauna, as well as the associated financial losses and damage to both image and reputation.
[0003] The pipelines are susceptible to various types of third-party interference, such as deliberate acts (illegal tapping, sabotage) or unintended disruption (construction work, farming). This third-party interference can cause huge financial and environmental damage and loss of reputation to the pipeline operators. A reliable, real-time pipeline monitoring system is therefore required in order to protect nature, human health and economic interests.
[0004] According to a survey in 2009, about 36% of all worldwide pipeline leaks were caused by third party interference (TPI), such as illegal tapping, sabotage or construction work. Examples of the third party interference includes: 04 October 2001 Fairbanks Alaska (USA), 990 tons of oil due to sabotage; 2000 Tschernigow, Ukraine, loss of 500,000 liters
of diesel due to illegal tapping; 10 June 1999 Bellingham, Whatcorn Creek, Washington (USA), 880 tons of petrol due to construction work, financial damage USD 45 million.
[0005] Illegal tapping is a major problem in emerging countries like India, China and South America. In 2011 for example, Petroleos Mexicanos (PEMEX) counted 1,324 cases of illegal tapping in Mexico. Every day Pemex estimates 40,000 liters of oil and gas, which sums up to an annual damage of more than 1 billion US$ of lose. Such damage could be avoided or substantially reduced, if third party interference was detected before or close to the occurrence of the interference happens.
[0006] Many times the pipeline monitoring is done by walking, driving and flying along the pipeline. The annual costs for walking and driving the line vary from€100 to€350 per kilometer. The additional costs for flying-the-line amount to€4.50 per kilometer. Usually two inspections are performed per month, which adds up to€108 per kilometer (€4.50/km x 2 flights/month x 12 months). The annual costs are between€208 and€458 per kilometer.
Description of the Related Art [0007] A number of systems for the sensing of an acoustic disturbance are known. The acoustic disturbance can be representative of damage to the pipelines through third party interference or geohazards, as described in the introduction. These systems involve the use of an optical fiber laid alongside the pipeline, which acts as a sensor and detects changes in the pattern of back- scattered radiation in order to sense an acoustic disturbance. For exam- pie, Pimon GmbH, Munich, Germany, sells an apparatus PMS2500-vibrO that utilizes distributed fiber optical sensing technology to detect the acoustic disturbance. The PMS2500- vibrO system combines an optical time domain reflectometer (OTDR) with an analysis and pattern recognition software and offers a customized interface with geographic information system (GIS) mapping.
[0008] International Patent Applications No WO 2011/05813, WO 2011/ 015812 and WO 2011/059501 (QinetiQ Ltd) all teach various aspects of using a distributed fiber optic sensing system for establishing events of interest. Similarly UK Patent Application No GB 2
491 658 also teaches a method and system for locating an acoustic disturbance. These patent applications all have in common that the systems analyze the back-scattered radiation from the optical fiber to establish the event of interest. Such systems are useful in determining an event of interest from the acoustic disturbance, but the systems are known to produce "false positives" in which events are identified that are of no interest and fail to identify some events of interest, in particular when such events have not been seen before.
[0009] One solution to the issue of incorrect identification of events would be to use an artificial neural network (ANN) to train the system to recognize the event. The ANNs are computational models and are inspired by animal central nervous systems, in particular the brain, that are capable of machine learning and pattern recognition. The ANNs are usually presented as a system of nodes or "neurons" connected by "synapses" that can compute values from inputs, by feeding information from the inputs through the ANN. The synapses are the mechanism by which one of the neurons passes a signal to another one of the neurons.
[0010] One example of the ANN is for the recognition of handwriting. A set of input neurons may be activated by pixels in a camera of an input image representing a letter or a digit. The activations of these input neurons are then passed on, weighted and transformed by some function determined by a designer of the ANN to other neurons, etc. until finally an output neuron is activated that determines which character (letter or digit) was imaged. ANNs have been used to solve a wide variety of tasks that are hard to solve using ordinary rule-based programming, including computer vision and speech recognition. [0011] There is no single formal definition of an ANN. Commonly a class of statistical models will be termed "neural" if the class consists of sets of adaptive weights (numerical parameters that are tuned by a learning algorithm) and are capable of approximating nonlinear functions of the inputs of the statistical models. The adaptive weights can be thought of as the strength of the connections (synapses) between the neurons.
[0012] The ANNs have to be trained in order to produce understandable results. There are three major learning paradigms: supervised learning, unsupervised learning and reinforcement learning.
[0013] In a supervised learning, the learning paradigms all have in common that a set of pre-analyzed data, for example a waveform, is analyzed by the ANN and the weights of the connections (synapses) between the neurons in the ANN are adapted such that the output of the ANN is correlated with a known event. There is a cost involved in this training. An improvement in the efficiency of the results of the ANN can be obtained by using a greater number of data items representing the known event in a training set. The greater number of data items require, however, an increase in computational power and time for the analysis in order to get the correct results. There is therefore a trade-off that needs to be established between the time taken to train the ANN and the accuracy of the results.
[0014] Recent developments in the ANNs involve so-called 'deep learning'. Deep-learning is a set of algorithms that attempt to use layered models of inputs. Jeffrey Heaton, University of Toronto, has discussed deep learning in a review article entitled 'Learning Multiple Layers of Representation' published in Trends in Cognitive Sciences, vol. 11, No. 10, pages 428 to 434, 2007. This publication describes multi-layer neural networks that contain top-down connections and training of the multilayer neural networks one layer at a time to generate sensory data, rather than merely classifying the data. [0015] Neuron activity in the prior art ANNs is computed for a series of discrete time steps and not by using a continuous parameter. The activity level of the neuron is usually defined by a so-called "activity value", which is set to be either 0 or 1, and which describes an 'action potential' at a time step t. The connections between the neurons, i.e. the synapses, are weighted with a weighting coefficient, which is usually chosen have a value in the interval [-1.0, + 1.0]. Negative values of the weighting coefficient represent "inhibitory synapses" and positive values of the weighting coefficient indicate "excitatory values". The computation of the activity value in the ANNs uses a simple linear summation model in which weighted ones of some or all of the active inputs received on the synapses at a neuron are compared with a (fixed) threshold value of the neuron. If the summation results in a value that is greater than the threshold value, the following neuron is activated.
[0016] One example of a learning system is described in international patent application No. WO 1998 027 511 (Geiger), which teaches a method of detecting image characteris-
tics, irrespective of size or position. The method involves using several signal-generating devices, whose outputs represent image information in the form of characteristics evaluated using non-linear combination functions. [0017] International patent application No. WO 2003 017252 relates to a method for recognizing a phonetic sound sequence or character sequence. The phonetic sound sequence or character sequence is initially fed to the neural network and a sequence of characteristics is formed from the phonetic sequence or the character sequence by taking into consideration stored phonetic and/or lexical information, which is based on a character string se- quence. The device recognizes the phonetic and the character sequences by using a large knowledge store having been previously programmed.
[0018] An article by Hans Geiger and Thomas Waschulzak entitled 'Theorie und Anwen- dung strukturierte konnektionistische Systeme', published in Informatik-Fachberichte Springer- Verlag, 1990, pages 143 - 152 also describes an implementation of a neural network. The neurons in the ANN of this article have activity values between zero and 255. The activity values of each one of the neurons changes with time such that, even if the inputs to the neuron remain unchanged. The output activity value of the neuron would change over time. This article teaches the concept that the activity value of any one of the nodes is dependent at least partly on the results of earlier activities. The article also includes brief details of the ways in which system may be developed.
[0019] Multiple tests have shown that the systems as described in the prior art do not deliver reliable results. A system and a method for determining a location and type of a dis- turbance along a pipeline delivering reliable results is needed therefore, which is comparatively easy to install, run, and maintain.
SUMMARY OF THE INVENTION [0020] The principal of the method and apparatus of determining locations and types of disturbances along a pipeline, as described in this disclosure, is based upon analyzing back- scattered radiation in an optical fiber using a signal processor.
[0021] More specifically, a fiber sensor system for determining a location and type of a disturbance along a pipeline is provided. The fiber sensor system comprises an optical fiber, a radiation source for launching radiation into the optical fiber, a detector for detecting back- scattered radiation back- scattered from the optical fiber, and a signal processor connected to the detector for generating a plurality of values from the back- scattered radiation, corresponding to the condition of the optical fiber at a certain time at a certain location of the optical fiber. The signal processor comprises a recorder to record the values corresponding to the condition of the optical fiber at a specific repetition rate, an analyzer to analyze the change of those values per repetition over time and create a value over time pattern corresponding to the specific disturbance.
[0022] The radiation source as mentioned above is in one aspect designed to produce a series of optical pulses. [0023] In another preferred embodiment, the fiber sensor system comprises a memory for storing a plurality of patterns representative of known disturbances, and a means to compare the value over time patterns corresponding to the specific disturbance as created with the recorded patterns representative of known disturbances. [0024] In another aspect of the disclosure, the radiation source produces infrared light in a series of optical pulses.
[0025] In a further aspect of the disclosure, the specific repetition rate at which the values corresponding to the condition of the optical fiber are recorded, is 250Hz.
[0026] Also disclosed is a method for determining a location and type of a disturbance along a pipeline, the method comprising the steps of launching radiation into an optical fiber buried along a pipeline, detecting back- scattered radiation back- scattered from the optical fiber, generating a plu- rality of values from the back- scattered radiation, corresponding to the condition of the optical fiber at a certain time at a certain location of the optical fiber; recording the values corresponding to the condition of the optical fiber at a specific repeti-
tion rate, analyzing the development of those values per repetition over time; and creating a value over time pattern corresponding to the specific disturbance.
[0027] In another aspect of the disclosure, the method for determining a location and type of a disturbance further comprises the step of comparing the value over time pattern with recorded patterns of known disturbance types.
[0028] In a further aspect of the disclosure, the values corresponding to the condition of the optical fiber are recorded at a preferred specific repetition rate of 250Hz. Currently, the preferred spatial sampling rate is 1 meter (this corresponds to 100MHz ADC sampling frequency). Other temporal and spatial sampling rates may be used, as the sampling frequencies may be varied within a certain range.
[0029] The optical fiber sensor system taught in this disclosure is able to provide: alert in case of emergency with an accurate GIS location, identify threats in real time in order to reduce the risk from leakages caused by digging, drilling, tapping, sabotage, earthquake etc., monitor around 50km of pipeline with one measuring unit without any additional power supply and with scalability to cover deployments of any length by adding additional units. The system reduces false or irrelevant alarms and is capable of differenti- ating multiple events down to 5-15 meter resolution; as well as track potential intruders, moving in different velocities (vehicles or on foot) along the pipeline. The system can detect direction and speed of the possible intruders.
DESCRIPTION OF THE FIGURES
[0030] Fig. 1 shows an example of an optical fiber sensor used in the system of the disclosure.
[0031] Fig. 2 to 6 illustrate graphs showing various outputs generated by processing of a back- scattered signal in a processor.
[0032] Fig. 7 is a flow chart, describing the steps of the method of this disclosure.
DETAILED DESCRIPTION OF THE INVENTION
[0033] The invention is described on the basis of the drawings. It will be understood that the embodiments and aspects of the invention described herein are only examples and do not limit the protective scope of the claims in any way. The invention is defined by the claims and their equivalents. It will be understood that features of one aspect or embodiment of the invention can be combined with a feature of a different aspect or aspects and/or embodiments of the invention.
[0034] The optical fiber sensor system 10 of this disclosure is based on distributed fiber optical sensing and integrates an optical time domain reflectometer (OTDR) with detection software, as well as a customized interface with geographic information system (GIS) mapping. The fiber optic cables are often already buried near to or attached to the pipeline for telecommunication purposes. The system turns a standard single mode telecommunication fiber into a listening device for events of interest.
[0035] The optical time-domain reflectometer (OTDR) is an optoelectronic instrument used to characterize optical fibers. The reflectometer injects series of optical pulses into the optical fiber that is being tested and detects at the same fiber end the returning ("back- scattered") light that has been scattered or reflected back from points along the fiber. The OTDR produces a so-called reflectogram from this measurement and performs an efficient, precise and wide analysis of the fiber characteristics. [0036] A laser diode launches radiation in the form of a plurality of optical pulses into one end of the optical fiber and a photodiode measures the returning light. The photo diode is part of the OTDR. The back- scattered signal detected by the photodiode provides relevant information about the events of interest along the optical fiber. The back- scattered signal is influenced by both the attenuation and reflections, which the injected laser pulse experi- ences on its way through the optical fiber. If a certain area within the optical fiber has a higher degree of attenuation or reflection, for instance caused by a bending or a fiber connector, this higher degree of attenuation or reflection is detected by differences in the back- scattered signals. The velocity of the optical pulse is known and the exact location of the
event of interest can therefore be determined on the basis of the time difference between the injection of the optical pulse and the return of the back- scattered signal.
[0037] Suppose now that a disturbance occurs near the optical fiber. The pressure on the optical fiber will cause at least a small degree of bending of the optical fiber and will result in a change of the returned back- scattered signal. This change can be detected in the OTDR. Suppose that the disturbance is a moving object, then the slight bending of the optical fiber will change location, enabling the speed and direction of movement of the disturbance to be established.
[0038] The optical fibers make a good sensor, as the optical fibers are able to measure vibrations over long distances. The reflections of the signals in the optical fibers change on vibrations (e.g. caused by vehicles, footsteps, digging, drilling), on temperature alterations (e.g. caused by escaping pressurized gas), or when optical fibers get strained, bent, kinked, or cut off.
Example
[0039] Fig. 1 shows an example of the optical fiber sensor system 10 using the teachings of this disclosure. An optical fiber 20 used in this sensor system 10 and can be a specially laid optical fiber 20, which has been placed in a region of interest. The optical fiber 20 can also be a standard telecommunications optical fiber, which generally carries data. The optical fiber 20 can be laid near or directly adjacent to a pipeline 25, as outlined in the introduction to the description.
[0040] The optical fiber 20 is connected to a radiation source 30, which in one non- limiting aspect of the invention is a semiconductor laser producing radiation 35 comprising a plurality of optical pulses at 1.55μιη and launches the radiation 35 into the optical fiber 20 at repetition rates up to 4kHz in the present embodiment.
[0041] The returning back- scattered signal 37 from the optical fiber 20 can be observed at the same end of the optical fiber 20 in form of a reflectogram by a detector 40 in which the time delay of signal incidence is shown as distance on the optical fiber 20. A highly coher-
ent laser as the radiation source 30 is used in order to increase the sensitivity. Due to this high level of coherence, an interference pattern can be observed in the back-scattered signal 37 (fingerprint). Vibrations cause bending in the optical fiber 20 and result in temporary changes in this interference pattern. The interference pattern can be used to detect a disturbance 45, for example a third-party interference caused by manual or machine digging. Every kind of incident causes certain multiple vibrations and changes the back- scattered signals 37 leading to differences in the interference pattern. The back-scattered signals 37 are received and the signal patterns are digitalized. The signal in the detector 40 passes through to a signal processor 50. The signal processor 50 is able to classify the dis- turbance 45 and send an alarm to the system operator, if required.
[0042] The signal processor 50 is designed to separate all "regular" sounds from possible disturbances 45 in the back-scattered signal as well as any inoffensive or other irrelevant incidents.
[0043] The processor 50 further includes an input device (not shown) that is used to input information items relating to the reflectogram. The information items may include a name or a label generally attached to the reflectogram and/or to one or more features in the reflectogram. The input device is connected to a signal processor (not shown) having a memory. The signal processor compares the characteristics relating to a particular feature in a stored reflectogram with the inputted information and can associate the particular stored reflectogram (or a portion thereof) with the inputted information. This association is memorized so that if an unknown pattern or feature in a reflectogram is detected by the detector 40, the processor 50 can determine that this unknown pattern is in fact a known pattern and output the associated item of information to the user at an output.
[0044] The pattern recognition system 10 can be trained to recognize a large number of patterns in the reflectogram using an unsupervised leaning process. This unsupervised learning process is carried out using a set of pre-stored patterns representing features of interest from existing reflectograms and running the pre-stored patterns through the processor 50.
[0045] The system and method of the current disclosure can be used to determine and classify unknown disturbances 45, as shown in Fig. 2 to 6.
[0046] In this example of the system and method the radiation 35 is launched into the opti- cal fiber 20. The detector 40 receives the back- scattered signal 37 and passes the back- scattered signal 37 to the processor 50.
[0047] A label can be associated with the one or more features in the back- scattered signal. [0048] The procedure is repeated for a different feature. This different feature creates a different structure within the plurality of layers 180. The learning procedure can then proceed using different ones of the features.
[0049] Processing of the back-scattered signal 37 in the processor 50 is described with respect to Figs. 2 to 6.
[0050] The measuring data to be evaluated comprises a value matrix spanned by sampled fiber range and the temporal sample steps within a certain observation time. It must be distinguished between detection of location and detection of time of the fiber (20). For each laser pulse injected into the fiber (20), the back-scattered signal is detected. Via the time elapsed between injecting the pulse and detection of a certain part of the back scattered signal, the location is determined by means of the run time of the light (OTDR-Method), from which the scattered signal emanates. An analogue-digital-converter scans the back scattered signal at 100MHz, so that for each meter of fiber a back scattered value can be allocated. The back scattered values for the entire fiber (e.g. 20000 values for 20km) per injected laser pulse is called a fingerprint. A fingerprint is a line in the matrix to be evaluated. The temporal development of the fiber is obtained by repeated scanning of the fiber, i.e. by repeatedly injecting laser pulses with a repetition rate of e.g. 250Hz. The temporal development of a fiber location is derived from evaluating subsequent fingerprints at this location. The values at such a location form the columns of the matrix and are called time slice.
[0051] In Fig. 2 the evaluation is carried out for the example of seven pickaxe beatings 201 to 207 carried out at a sample installation of the device according to the disclosure. Due to the specific physical layout of the sample installation, the same action is detected in three positions, namely at 175m, 230m and 530m from the measuring distance.
[0052] The classification of disturbances on the optical fiber 20, which is due to activities around the optical fiber 20, is carried out by mainly investigating the temporal development or by recognizing patterns within a single one of the time slices, wherein intermediate results of adjoining time slices are also considered (stripe condition). For the detection of manual excavation activities the following steps are carried out:
[0053] Step 1: Use of a Laplace operator. This operator generates positive or negative peaks at those locations, at which the input signal shows a substantial change. Small changing values are those values close to zero and steady input values are those values that are precisely zero. This Laplace operator improves the signal-to-noise ratio of the measured values. The one dimensional Laplace operator is a vector of coefficients which are each to be multiplied with one out of an equal number of subsequent input values. Similar to a scalar product the sum of all products ensues the output value at the position the center coefficient has been applied to. The number of coefficients is odd. The coefficients comply three requirements: The middle one has a positive maximum value. The coefficients are symmetric with regard to the central one and behave like a Gaussian or a Lorentz distribution. The sum of all coefficients is zero.
[0054] In Fig. 3 the impact of the Laplace operator is shown. The graph in Fig. 3c shows the Laplace data, which arise from the signal of a pickaxe, as shown in Fig. 3a. The amount is formed by the result of the Laplace operator, so that one works with positive peaks only, see the graph in Fig. 3d.
[0055] Step 2: The Laplace output often contains solitary peaks which are not caused by a digging stroke. Stroke events instead show a whole cluster of peaks that enables one to locate the time of the hit by locating the highest density of peaks in such a cluster. In order to neglect single peaks and to identify the most likely hit time the amount of Laplace out-
put within a time slice is averaged as described below. By doing so single peaks are dwarfed and one can assign a hit time to each peak cluster. The latter is a local maxima of the averaged time slice. In order to achieve the above-quoted averaging of the absolute Laplace data, a sliding average is formed for every single point within the time slice, e.g. a value is summed using the five direct preceding values and five direct succeeding values. After a division by the number of all those values, here eleven values, the result of the sliding average substitutes the previous value. By doing so, the graphs are smoothed, and artifacts, such as isolated peaks, can be suppressed. [0056] Step 3: Definition of a threshold as a function of the noise background, by multiplying the noise average with an eligible factor, and sorting out the over threshold values, as can be seen in Fig. 3d.
[0057] Step 4: Use of a "step condition": Values which are over the defined threshold val- ues can also be generated by a person walking around over the optical fiber 20 or by a car near the optical fiber. To be considered to be a blow of a pickaxe or a spade, over threshold values must fulfil two additional prerequisites, the step condition and the stripe condition.
[0058] The step condition: Starting point or symmetric point is the temporal position of values which are over the threshold values found in the 3rd step. In a defined time distance from this temporal position two so-called value areas that are equal in size are looked at which are located in time before and after the main peak at the temporal position. For example, suppose that the peak is located at a temporal peak at time t=200, then the two value areas chosen are a right value area stretching from t=220 to t=320 and a left value area stretching from the left from t=80 to t=180. For these value areas an average is formed in form of the values in the raw signal over the value area (measuring values before step 1) in each case. The step condition requires that the amount of the difference of both averages may not surpass a certain value. If it does, the peak is at or close to a clear step in the basic level of the raw signal. Steps can be seen in the signal of a pedestrian. They are, however, extremely rare for a pickaxe signal, which is shown in the graph in Fig. 3a.
[0059] Step 5: Use of a "stripe condition": the remaining values are now compared with other, neighboring values, in the value matrix, meaning that a value will be compared to
other values in this time slice, but also with values from the neighboring ones of the time slices. For pickaxe signals, it is typical that the blow of the pickaxe in the ground generates almost always several values in a relatively narrow stripe (see Fig. 2). This stripe is given by an eligible delta t and an eligible delta x. A value fulfils the condition if there are suffi- cient other values next to it in the stripe area as defined by delta t and delta x, so that a defined number of values is reached in a "cluster", e.g. 5. Values 201 to 207 shown in Fig. 2 have gone through all those steps. These values can be supposed to have been created in the raw signal due to an excavation or a pickaxe activity. [0060] Signals of pedestrians can be classified into those signals created by walking across the sensor cable and those signals originating from walking along the sensor cable. If someone walks across the sensor cable, at right angles to the sensor cable, patterns of chronologically separated steps can be recognized in the affected time slice (see Fig. 4 401). Even if the pedestrian walks on the same spot, the measured signal looks like signal 401 in Fig. 4. It will be noted that an offset has been added to the graphs in order to separate the lines clearly from each other in Figs. 4 and 5. This is merely for reasons of clarity.
[0061] A clear deviation from the raw signal, which usually shows a length of about 0.5 seconds for single steps is typical. This deviation in the raw signal can be formulated as an additional condition in the algorithm described above for the detection of a blow by a pickaxe. If the step condition described above was fulfilled, and the event was rejected as a blow, a pedestrian step event can be ascertained by examining this additional condition of the deviation from the raw signal. This is valid only for single pedestrian steps. If the pedestrian runs along the sensor cable, the steps are not so clearly separated from each other. Then the movement of the pedestrian can be detected by use of the following algorithm.
[0062] To show movements of the pedestrians or the vehicles over the optical fiber 20, a spline algorithm is used. In contrast to the narrow peaks in the raw data relating to the excavation activities, steps in the raw data or displacements are typical for other movements in the signal. The spline algorithm is a mathematical function, which interpolates the development of the measuring data. Narrow peaks in the signal as well as noise are ignored. The algorithm comprises the following steps (see Fig. 4 and 5):
[0063] Step 1: Filtering out or suppression of narrow peaks. Should a value differ clearly in the raw signal from the average of the values in close proximity, the value is substituted with the average. This is illustrated in the graph 402 and 502 in Fig. 4 for two single steps of a pedestrian and Fig. 5 for a pedestrian walking along the optical fiber, in which a sub- stitution graph is constructed from the raw data in the upper graph 401 and 501.
[0064] Step 2: Smoothing of the data by use of a gliding average as in graph 403 and 503).
[0065] Step 3: Calculation of a spline, which returns only the substantial changes of the signal, graphs 404 and 504.
[0066] Step 4: The amount of the derivative (slope) of the spline (graph 404 in Fig. 4 and 504 in Fig. 5) is always high when there are clear steps in the raw signal, as graph 405 in Fig. 5 and 505 in Fig. 5 indicate.
[0067] Step 5: Display of the values, which are larger than a certain threshold value 410 and 510, as a signal of the pedestrian, graphs 405 and 505.
[0068] Furthermore, an alternative algorithm was developed, which is suitable to detect pedestrian movement. This algorithm has a stronger response to digging activities than the Spline- algorithm and may only be used effectively, if by means of the above algorithm (step and stripe condition) a digging activity can be excluded. Similar to the step condition described within each time slice for each point in time the passed raw signal is compared with the following raw signal. These two areas are distributed symmetrically around the position and are separated by a gap. E.G.: Position = 200; Values between 80 and 180 form an area 1; values between 220 and 320 form an area 2. The difference value of the averages of the two areas is a function with particularly large amplitudes for pedestrian activities. To suppress noise values below a threshold are set to zero and only values above the threshold are displayed. A maximum of this function appears for steps in the basic signal, in particu- lar when the step condition above excludes a hit event.
[0069] In Fig. 6 the signal of the pedestrian is shown that walks along the sensor cable in one direction and then turns back. It will be noted that the signal of one pedestrian is
shown mirrored in Fig. 6. This is an artefact of the system used. The pedestrian walks first 10-15 meters along the cable and then returns to the starting point. This leads to the slightly crescent- shaped figure. [0070] Any stationary vehicles with diesel engines (tractors, JCBs, trucks) can be detected very simply by means of a Fourier transformation. To do so, the Fourier transformation is calculated in a calculation window (segment of a time slice, e.g., 512 values) and the power spectrum derived therefrom. The basic oscillation and harmonics of the engines are visible as peaks around the power spectrum. For every monitored fiber position the Fourier- transform-algorithm is applied. This means the previous 512 values of each time slice are taken as the input vector of Fast Fourier Transform. This gives an output vector of 512 complex values. Deriving the square of the absolute output values gives the frequency- power spectra of the harmonic components in the input between zero and the Nyquist- frequency. This is half the frequency that has been used when recording the input values that means for a 250Hz temporal sampling rate the Nyquist-frequency is 125 Hz. A high peak at zero Hz is caused by the constant input signal offset and is to deduct or to neglect. The harmonic vibration of a diesel engine generates clear peaks in the frequency power spectra. The fundamental frequency and its higher-order harmonics are to observe. For setting an alarm state a suitable threshold can be defined that the peak in the power spectra has to surpass in order to switch to an alarm state.
[0071] Moving vehicles can be distinguished from the pedestrians as a function of their weight and their speed by the size of a segment of the optical fiber in which the moving vehicles cause a signal. While the pedestrian crossing the signal generated in the sensor cable is seen in a response area of 3 to 5 meters along the sensor cable, the signal generated in the sensor cable relating to a crossing car is between 10 to 12 meters of the response area. The signal relating to a truck has a response area which is bigger than 20 meters.
[0072] For the moving vehicles, about half of the affected time slices show a typical wave- form, like half an oscillation of a sinus wave, of the measuring signal. After subjecting the time slice to a gliding average (averaging over 50 to 100 values) the wave form is preserved while the background noise and isolated high peaks disappear. Such a wave form is
not included in the pedestrian's signal and, hence, points with all probability to a moving car. The FWHM of the wave form points to the time needed to cross the cable.
[0073] Should the system 10 be unable to identify the feature, then the system 10 can give an appropriate warning and human intervention can be initiated in order to review the disturbance and classify the unknown feature or to resolve in the other conflicts.
[0074] Fig. 7 is a flow chart, describing the steps of the method of this disclosure. The method for determining a location and type of a disturbance along a pipeline, comprises the following steps :
[0075] The first step 701 comprises the injection of radiation into an optical fiber buried along a pipeline, followed by the detection of the back-scattered radiation in the next step 702, which was back- scattered from the optical fiber and which is generating a plurality of values, corresponding to the condition of the optical fiber at a certain time at a certain location of the optical fiber; the third step 703 comprises the recording of the generated values corresponding to the condition of the optical fiber at a specific repetition rate, analyzing the development of those values per repetition over time; and creating a value over time pattern corresponding to the specific disturbance; the location and the type of a disturbance is determined in the fourth step 704 by comparing the value over time pattern with recorded patterns of known disturbance types.
Claims
1. An optical fiber sensor system (10) for determining a location and type of a disturbance (85) along a pipeline comprising:
an optical fiber (20) to be buried along the pipeline;
a radiation source (30) for launching radiation (35) into the optical fiber (20);
a detector (40) for detecting back-scattered radiation (37) back- scattered from the optical fiber (20);
a signal processor (50) connected to the detector (40) for generating a plurality of values (55) from the back-scattered radiation (37), corresponding to the condition of the optical fiber at a certain time at a certain location of the optical fiber;
wherein the signal processor (50) comprises a recorder to record the values corresponding to the condition of the optical fiber (20) at a specific spatial and temporal sampling rate, an analyzer to analyze the development of the values per repetition over time and create a value-over-time pattern corresponding to the specific disturbance, and a comparator to compare this value over time pattern with recorded patterns of known disturbance types.
2. The optical fiber sensor system (10) of claim 1, wherein the radiation source (30) launches a series of optical pulses.
3. The optical fiber sensor system (10) of claim 2, wherein the radiation source (30) launches infrared pulses.
4. The optical fiber sensor system (10) of any of the above claims, further comprising a memory (190) for storing a plurality of patterns (196) representative of disturbances (45).
5. The optical fiber sensor system (10) of any of the above claims, wherein the specific repetition rate at which the values corresponding to the condition of the optical fiber are recorded is 250Hz.
6. A method for determining a location and type of a disturbance (85) along a pipeline, the method comprising:
launching (205) radiation (35) into an optical fiber (20) buried along a pipeline; detecting (210) back- scattered radiation (37) back- scattered from the optical fiber (20);
generating a plurality of values (55) from the back-scattered radiation (37), corresponding to the condition of the optical fiber (20) at a certain time at a certain location of the optical fiber (20);
recording the values (55) corresponding to the condition of the optical fiber (20) at a specific repetition rate;
analyzing the development of those values (55) per repetition over time; and creating a value over time pattern corresponding to the specific disturbance.
7. The method for determining a location and type of a disturbance (85) of claim 6, further comprising the step of comparing the value over time pattern with recorded patterns of known disturbance types.
8. The method for determining a location and type of a disturbance (85) of claims 6 or 7, wherein the values corresponding to the condition of the optical fiber are recorded at a specific repetition rate of 250Hz.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP16729876.9A EP3332272A1 (en) | 2015-06-12 | 2016-06-13 | Method and apparatus for monitoring pipeline using an optical fiber sensor system |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
GB1510267.6 | 2015-06-12 | ||
GB1510267.6A GB2539254A (en) | 2015-06-12 | 2015-06-12 | Method and apparatus for monitoring pipeline |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2016198683A1 true WO2016198683A1 (en) | 2016-12-15 |
Family
ID=53784571
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/EP2016/063442 WO2016198683A1 (en) | 2015-06-12 | 2016-06-13 | Method and apparatus for monitoring pipeline using an optical fiber sensor system |
Country Status (3)
Country | Link |
---|---|
EP (1) | EP3332272A1 (en) |
GB (1) | GB2539254A (en) |
WO (1) | WO2016198683A1 (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107153222A (en) * | 2017-04-28 | 2017-09-12 | 国网上海市电力公司 | A kind of communication platoon pore passage occupies passive on-line monitoring method |
CN107730799A (en) * | 2017-10-13 | 2018-02-23 | 重庆励祺科技有限公司 | Hidden optical cable perturbation sensing system |
CN111024212A (en) * | 2020-01-14 | 2020-04-17 | 辽宁国运通达通信集团有限公司 | Method for converting optical cable distance into landmark position by using distributed optical fiber sensing system |
CN111063174A (en) * | 2018-10-17 | 2020-04-24 | 海隆石油集团(上海)信息技术有限公司 | Pipeline line safety early warning system based on distributed optical fiber sensing |
CN111965693A (en) * | 2020-08-21 | 2020-11-20 | 电子科技大学 | Pipeline trend tracing method and system based on optical cable |
CN112129490A (en) * | 2020-10-16 | 2020-12-25 | 杨继涛 | Multifunctional optical cable detection device and method |
CN115793086A (en) * | 2023-02-07 | 2023-03-14 | 武汉新楚光电科技发展有限公司 | Optical cable laying environment underground cavity judgment method and system based on optical fiber sensing |
US11750282B2 (en) | 2018-10-05 | 2023-09-05 | Centre For Development Of Telematics (C-Dot) | Learning-based method and system for configuring an optical time-domain reflectometer in a gigabit passive optical network |
WO2023220451A1 (en) * | 2022-05-13 | 2023-11-16 | Nec Laboratories America, Inc. | System to measure coil locations and lengths on aerial fiber cables by distributed fiber sensing |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107730800B (en) * | 2017-11-13 | 2019-07-16 | 浙江众盟通信技术有限公司 | Anti-Interference Analysis method based on fiber-optic vibration safety pre-warning system |
EP3776915A4 (en) * | 2018-04-06 | 2021-06-02 | Ava Risk Group Limited | Event statistic generation method and apparatus for intrusion detection |
JPWO2021010407A1 (en) * | 2019-07-16 | 2021-01-21 | ||
EP4001862A4 (en) * | 2019-07-17 | 2022-08-17 | NEC Corporation | Optical fiber sensing system, optical fiber sensing equipment, and abnormality assessment method |
CN111780857B (en) * | 2020-06-05 | 2022-02-15 | 南京曦光信息科技有限公司 | Multi-point disturbance positioning detection method of P-OTDR system based on harmonic accumulation |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2011058322A2 (en) * | 2009-11-13 | 2011-05-19 | Qinetiq Limited | Improvements in distributed fibre optic sensing |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB0605066D0 (en) * | 2006-03-14 | 2006-04-26 | Schlumberger Holdings | Method and apparatus for monitoring structures |
CN201173390Y (en) * | 2008-04-03 | 2008-12-31 | 东南大学 | Limit type signal processing and recognition module for distributed optical fibre disturbance sensing system |
GB0919899D0 (en) * | 2009-11-13 | 2009-12-30 | Qinetiq Ltd | Fibre optic distributed sensing |
WO2014170013A1 (en) * | 2013-04-17 | 2014-10-23 | Eth Zurich | Fibre optic based intrusion sensing system |
GB2525251A (en) * | 2014-04-18 | 2015-10-21 | Mic Ag | Optical fibre sensor system |
-
2015
- 2015-06-12 GB GB1510267.6A patent/GB2539254A/en not_active Withdrawn
-
2016
- 2016-06-13 WO PCT/EP2016/063442 patent/WO2016198683A1/en active Application Filing
- 2016-06-13 EP EP16729876.9A patent/EP3332272A1/en not_active Withdrawn
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2011058322A2 (en) * | 2009-11-13 | 2011-05-19 | Qinetiq Limited | Improvements in distributed fibre optic sensing |
Non-Patent Citations (2)
Title |
---|
LI WANG ET AL: "A Fiber-sensor-based Long-distance Safety Monitoring System for buried Oil Pipeline", NETWORKING, SENSING AND CONTROL, 2008. ICNSC 2008. IEEE INTERNATIONAL CONFERENCE ON, IEEE, PISCATAWAY, NJ, USA, 6 April 2008 (2008-04-06), pages 451 - 456, XP031255209 * |
SUZANA VALE ET AL: "Fiber optic based pipeline monitoring", 18 March 2013 (2013-03-18), XP055295463, Retrieved from the Internet <URL:http://www.youtube.com/watch?v=Y6zJblR7gqw> [retrieved on 20160815] * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107153222A (en) * | 2017-04-28 | 2017-09-12 | 国网上海市电力公司 | A kind of communication platoon pore passage occupies passive on-line monitoring method |
CN107730799A (en) * | 2017-10-13 | 2018-02-23 | 重庆励祺科技有限公司 | Hidden optical cable perturbation sensing system |
US11750282B2 (en) | 2018-10-05 | 2023-09-05 | Centre For Development Of Telematics (C-Dot) | Learning-based method and system for configuring an optical time-domain reflectometer in a gigabit passive optical network |
CN111063174A (en) * | 2018-10-17 | 2020-04-24 | 海隆石油集团(上海)信息技术有限公司 | Pipeline line safety early warning system based on distributed optical fiber sensing |
CN111063174B (en) * | 2018-10-17 | 2022-07-12 | 海隆石油集团(上海)信息技术有限公司 | Pipeline line safety early warning system based on distributed optical fiber sensing |
CN111024212A (en) * | 2020-01-14 | 2020-04-17 | 辽宁国运通达通信集团有限公司 | Method for converting optical cable distance into landmark position by using distributed optical fiber sensing system |
CN111024212B (en) * | 2020-01-14 | 2020-09-25 | 辽宁国运通达通信集团有限公司 | Method for converting optical cable distance into landmark position |
CN111965693A (en) * | 2020-08-21 | 2020-11-20 | 电子科技大学 | Pipeline trend tracing method and system based on optical cable |
CN112129490A (en) * | 2020-10-16 | 2020-12-25 | 杨继涛 | Multifunctional optical cable detection device and method |
WO2023220451A1 (en) * | 2022-05-13 | 2023-11-16 | Nec Laboratories America, Inc. | System to measure coil locations and lengths on aerial fiber cables by distributed fiber sensing |
CN115793086A (en) * | 2023-02-07 | 2023-03-14 | 武汉新楚光电科技发展有限公司 | Optical cable laying environment underground cavity judgment method and system based on optical fiber sensing |
CN115793086B (en) * | 2023-02-07 | 2023-06-06 | 武汉新楚光电科技发展有限公司 | Optical cable laying environment underground cavity judging method and system based on optical fiber sensing |
Also Published As
Publication number | Publication date |
---|---|
GB2539254A (en) | 2016-12-14 |
EP3332272A1 (en) | 2018-06-13 |
GB201510267D0 (en) | 2015-07-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2016198683A1 (en) | Method and apparatus for monitoring pipeline using an optical fiber sensor system | |
US10119845B2 (en) | Optical fibre sensor system | |
US20180357542A1 (en) | 1D-CNN-Based Distributed Optical Fiber Sensing Signal Feature Learning and Classification Method | |
US9739645B2 (en) | Fibre optic distributed sensing | |
Fouda et al. | Pattern recognition of optical fiber vibration signal of the submarine cable for its safety | |
CN104574742B (en) | Optical fiber perimeter security system based on phi-OTDR technology | |
CN104217513B (en) | The method improving phase sensitive optical time domain reflectometer identification intrusion event accuracy rate | |
CN109374116B (en) | Excavation behavior identification method of buried optical fiber sensing vibration detection system | |
Tian et al. | Intrusion signal recognition in OFPS under multi-level wavelet decomposition based on RVFL neural network | |
CN111537056A (en) | Pipeline along-line third-party construction dynamic early warning method based on SVM and time-frequency domain characteristics | |
US12000729B2 (en) | Perpendicular distance prediction of vibrations by distributed fiber optic sensing | |
CN114613116B (en) | External damage prevention early warning method, device, equipment and storage medium | |
CN109120336B (en) | False alarm prevention and false alarm prevention method based on phase sensitive optical time domain reflection sensor | |
Timofeev et al. | Multimodal heterogeneous monitoring of super-extended objects: modern view | |
Mahmoud | Practical aspects of perimeter intrusion detection and nuisance suppression for distributed fiber-optic sensors | |
CN111539393B (en) | Oil-gas pipeline third-party construction early warning method based on EMD decomposition and LSTM | |
CN111951505B (en) | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system | |
Lyu et al. | Two-stage intrusion events recognition for vibration signals from distributed optical fiber sensors | |
US20230024104A1 (en) | Identification of false transformer humming using machine learning | |
Wu et al. | An effective signal separation and extraction method using multi-scale wavelet decomposition for phase-sensitive OTDR system | |
Bublin | Machine learning for distributed acoustic sensors, classic versus image and deep neural networks approach | |
CN101626270B (en) | Event pre-warning and classifying method by external safety pre-warning and positioning system of photoelectric composite cables | |
Tejedor et al. | Towards detection of pipeline integrity threats using a SmarT fiber-OPtic surveillance system: PIT-STOP project blind field test results | |
CN114354744A (en) | Power cable external damage prevention monitoring method, system and device | |
Zhao et al. | Pipeline intrusion detection and location based on phase-sensitive optical time domain reflectometer |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 16729876 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2016729876 Country of ref document: EP |