CN113915536B - Analysis processing method based on pipeline safety early warning system - Google Patents
Analysis processing method based on pipeline safety early warning system Download PDFInfo
- Publication number
- CN113915536B CN113915536B CN202111271313.7A CN202111271313A CN113915536B CN 113915536 B CN113915536 B CN 113915536B CN 202111271313 A CN202111271313 A CN 202111271313A CN 113915536 B CN113915536 B CN 113915536B
- Authority
- CN
- China
- Prior art keywords
- signal
- frequency
- event
- mode
- processing method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 28
- 238000003672 processing method Methods 0.000 title claims abstract description 17
- 230000000694 effects Effects 0.000 claims abstract description 25
- 238000000605 extraction Methods 0.000 claims abstract description 15
- 238000003909 pattern recognition Methods 0.000 claims abstract description 9
- 238000000034 method Methods 0.000 claims description 30
- 239000013307 optical fiber Substances 0.000 claims description 25
- 238000010606 normalization Methods 0.000 claims description 18
- 239000013598 vector Substances 0.000 claims description 16
- 230000035945 sensitivity Effects 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 11
- 238000006243 chemical reaction Methods 0.000 claims description 10
- 238000000354 decomposition reaction Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 6
- 238000005065 mining Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 238000001228 spectrum Methods 0.000 claims description 6
- 238000011161 development Methods 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 5
- 230000003287 optical effect Effects 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 230000001052 transient effect Effects 0.000 claims description 4
- 230000003321 amplification Effects 0.000 claims description 3
- 238000005553 drilling Methods 0.000 claims description 3
- 239000000835 fiber Substances 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 3
- 230000000737 periodic effect Effects 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000005236 sound signal Effects 0.000 claims description 3
- 230000026676 system process Effects 0.000 claims description 3
- 238000013459 approach Methods 0.000 claims description 2
- 238000012546 transfer Methods 0.000 abstract description 2
- 230000010485 coping Effects 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 230000001427 coherent effect Effects 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 238000012876 topography Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000001069 Raman spectroscopy Methods 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013016 damping Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000009290 primary effect Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000011896 sensitive detection Methods 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F17—STORING OR DISTRIBUTING GASES OR LIQUIDS
- F17D—PIPE-LINE SYSTEMS; PIPE-LINES
- F17D5/00—Protection or supervision of installations
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention discloses an analysis processing method based on a pipeline safety early warning system, which relates to the technical field of pipeline safety and aims at solving the problem that a transfer function is unknown, so that the system is difficult to provide quantitative information of a strain value; event classification, more accurate classification of events using a pattern recognition machine, and more detailed extraction of feature parameters. The effects of accurately analyzing the event and rapidly analyzing and coping are achieved.
Description
Technical Field
The invention relates to the technical field of pipeline safety, in particular to an analysis processing method based on a pipeline safety early warning system.
Background
The pipeline transportation is an important way of energy transportation in China, is an important energy means which is essential in daily life of people, has complex and various topography and topography along the pipeline, is different from natural and social environments, occurs when risks of threatening pipeline safety such as geological disasters and third party construction are met, and has the advantages of economic rapid development, continuous promotion of town process, continuous change of the number of high-consequence areas along the pipeline, the length of pipe sections and the like, heavy pipeline protection work task and great responsibility.
On the existing pipeline, the continuous medium paved along the whole pipeline has only optical fibers, the original construction of the optical fibers paved in the same ditch has the primary effects of mainly being used for information communication, and in recent years, along with the progress of technology, the optical fiber sensing technology is rapidly developed, the development of the technology brings a natural sensor to the existing pipeline, the whole-line continuous sensing along the pipeline can be realized on the basis of not carrying out any additional construction, the environment sensing on the periphery of the pipeline is definitely an optimized sensing detection method, the optical fiber linear continuous passive sensing measurement distance is long, the working frequency bandwidth is wide, the dynamic range is large, and the optical fiber linear continuous sensor is a very high-quality linear continuous sensor with low energy consumption and high precision. The optical fiber is a sensor with extremely high sensitivity, can realize micro-magnitude sensitive detection for the measured or field loading, and can realize sensing measurement of physical quantities such as temperature, speed, acceleration, pressure, displacement, rotation, bending, flow, liquid level, sound field and the like. The optical fiber can realize high-speed sensing transmission of a large amount of data under the condition that the optical fiber is thinner than a hair wire, the volume is very light and small, the optical fiber core is externally provided with a carrying sleeve, is flexible and convenient to lay in any form after being protected by a filling medium, the optical fiber is not interfered by electromagnetic interference and various radiations, the optical fiber is particularly suitable for the inflammable, explosive and power supply difficult (high in cost) application occasions with strict space limitation, the distributed optical fiber sound sensing detector DAS based on the coherent optical time domain reflection technology is currently the main stream direction of pipeline safety precaution monitoring, a plurality of processing key points in the DAS system based on the coherent optical time domain reflection technology are provided, firstly, the signal based on the Rayleigh scattering effect is reflected, the signal usually carries a large amount of noise and a background environment signal which is not concerned, the background noise is required to be denoised, and meanwhile, the target signal is amplified, and secondly, what method is adopted to extract and classify the characteristics of threat events is the key point of signal processing analysis, and the method such as heterodyne, two-dimensional edge detection, multidimensional comprehensive denoising, raman diversity detection and small wave diversity detection method is proposed before.
The prior art solutions described above have the following drawbacks: because the pressure wave transmitted by the third party damage event which threatens the pipeline safety through the soil is mostly highly transient, each time the event causes less damping, the propagation speed is high, the dynamic strain change propagates along the optical fiber within a few milliseconds, the traditional methods such as time averaging are not applicable, the transmitted wave needs to be tracked and analyzed on a fast time scale, the position and the influence range of a disturbance point are determined, meanwhile, the distribution of a scatterer in the optical fiber is random due to the diversity and the unknown of the target event, the unknown of a transfer function is caused, and the quantitative information of the strain value is difficult to be provided by a system.
Disclosure of Invention
The invention aims to provide an analysis processing method for better processing, extracting and analyzing signals, which improves the accuracy of distinguishing target events by a system and is based on a pipeline safety early warning system.
In order to achieve the above purpose, the present invention provides the following technical solutions: an analysis processing method based on a pipeline safety early warning system comprises the following steps:
s1: event detection, converting a time domain signal into a frequency domain signal through fast fourier transformation, and then analyzing aiming at frequency, spectrum density, power spectrum and the like:
a1. for non-periodic signals can be expressed as:however, the implementation system processes high-frequency discrete data, so the high-frequency discrete data can be expressed as:
a2: the back scattering signal caused by the target event after the time-frequency conversion needs to be extracted from the measured back scattering signal, wherein the target event is set as y (t), the background signal is set as f (t), the noise signal is set as x (t), and the relation between the back scattering signal and the background signal is set as x (t):
f(t)=y(t)+x(t)
a3: after background noise cancellation, the signal f (t) needs to be digitized f [ c, m ] =y [ c, m ] +x [ c, m ]
Where c is the measurement point (c ε Z) + ) M is a digitized time index, and the interference of background noise generally does not change in a short time, which is different from the behavior of the target signal with high transients, so the number of samples used in the signal power estimation is expressed as:
a4: the noise in the non-target frequency band is filtered by applying the frequency band of the event, and then convolution amplification is carried out by cross-correlation operation of the target signal and the event template signal g (t), namely, sliding multiplication is carried out on two sequences, so that the amplified clear target event signal is obtained: (f×g) (x) = ≡f * (t)g(x+t)dt。
S2: the method comprises the steps of carrying out event reconstruction and feature extraction, adopting a frequency division mode decomposition method to decompose any real value signal into a plurality of discrete modes, wherein Sz, Z=1, … and Z, the modes have the characteristics of reconstructing an input signal, sequencing corresponding values in vectors, compressing each mode around a center frequency, calculating the bandwidth of the mode as a specific norm of Hilbert conversion operation, calculating an analysis signal corresponding to each mode by Hilbert conversion, transferring the analysis signal to a baseband by heterodyne through complex harmonics, adopting an alternate direction multiplication method ADMM to decompose an objective function into a plurality of sub-functions, estimating the bandwidth of the mode by smoothing the demodulated signal, realizing the frequency division mode decomposition method on the basis of the alternate direction multiplication method ADMM,
b1: the problem of optimization is converted into an unconstrained form by introducing an enhanced lagrangian function:
alpha is the balance parameter and lambda is the Lagrangian multiplier. Then adopting a series of ADMM iteration sub-optimization calling methods:
b2: the quadratic problem is then mapped in the corresponding fourier domain
Once the event signal is reconstructed, the characteristic signal is then extracted from the reconstructed signal, using the instantaneous amplitude a (n), instantaneous frequency f (n), instantaneous phase of the instantaneous signalThe true valuable discrete signal s (n), the resolved signal sa (n) can be expressed as +.>
Where I and Q are in-phase and quadrature components, which are derived fromGiven, wherein,Representing the hilbert transform. Thus, the instantaneous signal characteristics (a (n), phi (n), f (n)) can be calculated as
B3: extracting characteristic signals from the reconstructed signals through transient signal characteristics, normalizing all the characteristics, and calculating the skewness, kurtosis and variance of the characteristics;
s3: the method comprises the steps of event classification, more accurate classification of events by using a pattern recognition machine, more detailed extraction of characteristic parameters, classification of the event into mechanical mining, manual mining, drilling, trains, leakage and other events by classifying the target event, analysis of frequency bands of the events, classification of the behavior characteristics of the events, development situation of the events and influence on an optical cable, and thus obtaining an optimized decision combination.
By adopting the technical scheme, the analysis processing method based on Fourier transform and frequency division mode decomposition can better process, extract and analyze the signals, improve the accuracy of distinguishing the target event by the system and ensure the overall good processing and analysis effects.
Further, the pattern recognition in S3 is mainly based on a Gaussian Mixture Model (GMM) to classify in two different patterns.
By adopting the technical scheme, the mode identification mode of classifying under two different modes is convenient for analyzing different events integrally, so that different dangerous events are ensured to be fully identified integrally, and the integral early warning effect is improved.
Further, one of the two modes of the mode identification identifies the machine and the activity of the machine along the pipeline for the machine plus activity identification mode, and the other mode of the mode identification directly identifies whether the activity is an actual threat to the pipeline for the threat detection mode.
Through adopting above-mentioned technical scheme, the activity that machine plus activity identification mode sign machine and machine carried out along the pipeline and threat detection mode direct identification activity are the actual threat of pipeline and carry out dangerous recognition to the inside condition of whole pipeline, prevent that dangerous activity from causing the injury to the pipeline, have increased holistic security.
Further, the pattern recognition system internally comprises a feature extraction unit for reducing the high dimensionality of the signals acquired by the DAS system into more informative and distinguishable feature sets, a feature normalization unit for compensating for changes and sensing positions in the signal acquisition process, and a pattern classification unit for classifying the sound signals into a set of predefined Nq classes.
By adopting the technical scheme, the feature extraction unit, the feature normalization unit and the mode classification unit are used for classifying and identifying the overall internal events, so that the overall events can be fully classified, and the optimal decision combination can be conveniently found out overall.
Further, the feature extraction unit calculates energy for each sound frame using a short-time fast fourier transform (ST FFT), the over-frequency band signal takes a suitable time window first as a base feature vector component, calculates the complete base feature vector and corresponding energy at a given fiber position, and then defines a bandwidth for the initial frequency and the final frequency of the frequency range of the event.
By adopting the technical scheme, the bandwidth is defined for the initial frequency and the final frequency of the frequency range of the event, so that the combination of the solutions is conveniently carried out integrally according to the type of the event, and the detected event is ensured to be fully solved integrally and conveniently.
Further, the characteristic normalization unit performs optical fiber loss compensation normalization and sensitivity-based normalization processing on the acquired signals, and in order to compensate amplitude exponential decay of the signals due to optical fiber loss along with distance, the sensitivity-based normalization is to compensate the influence of sensitivity inconsistency, so that we can obtain signals with equivalent sensitivity along the optical fiber.
Through adopting above-mentioned technical scheme, carry out optical fiber loss compensation normalization and based on sensitivity's normalization processing to the acquisition signal, convenient whole data conversion to the convenience is whole to react according to the data that obtains, and the convenience is whole carries out the early warning to dangerous condition.
Further, the pattern classification unit classifies each feature vector into the most probable class (activity recognition pattern of machines and activity classes in machines) using a GMM-based approach, and employs a posterior maximum probability criterion to assign a given feature vector to the class with the highest probability given by the corresponding GMM.
By adopting the technical scheme, the given feature vector is allocated to the category with the highest probability given by the corresponding GMM by adopting the posterior maximum probability criterion, so that the best solution is matched conveniently and integrally, and the optimal decision combination is obtained, so that the dangerous situation is treated integrally.
In summary, the beneficial technical effects of the invention are as follows:
1. by adopting an analysis processing method based on Fourier transformation and frequency division mode decomposition, signals can be better processed, extracted and analyzed, the accuracy of distinguishing target events by a system is improved, and the effect of accurately analyzing the events is generated;
2. the data acquired by the DAS are analyzed by adopting a mode identification and machine learning method, so that threat events are successfully identified and analyzed, an optimized decision combination is conveniently obtained, and the effect of rapid analysis and response is generated.
Detailed Description
The method of the present invention is described in further detail below.
The analysis processing method based on the pipeline safety early warning system mainly comprises three main steps, namely event detection, event reconstruction and feature extraction, and event classification, wherein the steps are as follows:
s1: event detection, converting a time domain signal into a frequency domain signal through fast fourier transformation, and then analyzing aiming at frequency, spectrum density, power spectrum and the like:
a1. for non-periodic signals can be expressed as:however, the implementation system processes high-frequency discrete data, so the high-frequency discrete data can be expressed as:
a2: the back scattering signal caused by the target event after the time-frequency conversion needs to be extracted from the measured back scattering signal, wherein the target event is set as y (t), the background signal is set as f (t), the noise signal is set as x (t), and the relation between the back scattering signal and the background signal is set as x (t):
f(t)=y(t)+x(t)
a3: after background noise cancellation, the signal f (t) needs to be digitized f [ c, m ] =y [ c, m ] +x [ c, m ]
Where c is the measurement point (c ε Z) + ) M is a digitized time index, and the interference of background noise generally does not change in a short time, which is different from the behavior of the target signal with high transients, so the number of samples used in the signal power estimation is expressed as:
a4: the noise in the non-target frequency band is filtered by applying the frequency band of the event, and then convolution amplification is carried out by cross-correlation operation of the target signal and the event template signal g (t), namely, sliding multiplication is carried out on two sequences, so that the amplified clear target event signal is obtained: (f×g) (x) = ≡f * (t)g(x+t)dt。
S2: the method comprises the steps of carrying out event reconstruction and feature extraction, adopting a frequency division mode decomposition method to decompose any real value signal into a plurality of discrete modes, wherein Sz, Z=1, … and Z, the modes have the characteristics of reconstructing an input signal, sequencing corresponding values in vectors, compressing each mode around a center frequency, calculating the bandwidth of the mode as a specific norm of Hilbert conversion operation, calculating an analysis signal corresponding to each mode by Hilbert conversion, transferring the analysis signal to a baseband by heterodyne through complex harmonics, adopting an alternate direction multiplication method ADMM to decompose an objective function into a plurality of sub-functions, estimating the bandwidth of the mode by smoothing the demodulated signal, realizing the frequency division mode decomposition method on the basis of the alternate direction multiplication method ADMM,
b1: the problem of optimization is converted into an unconstrained form by introducing an enhanced lagrangian function:
alpha is the balance parameter and lambda is the Lagrangian multiplier. Then adopting a series of ADMM iteration sub-optimization calling methods:
b2: the quadratic problem is then mapped in the corresponding fourier domain
Once the event signal is reconstructed, the characteristic signal is then extracted from the reconstructed signal, using the instantaneous amplitude a (n), instantaneous frequency f (n), instantaneous phase of the instantaneous signalThe true valuable discrete signal s (n), the resolved signal sa (n) can be expressed as +.>
Where I and Q are in-phase and quadrature components, which are derived fromGiven, wherein-> Representing the hilbert transform. Thus, the instantaneous signal characteristics (a (n), phi (n), f (n)) can be calculated as
B3: extracting characteristic signals from the reconstructed signals through transient signal characteristics, normalizing all the characteristics, and calculating the skewness, kurtosis and variance of the characteristics;
s3: event classification, more accurate classification of events using a pattern recognition machine, and more detailed extraction of characteristic parameters, classification of the event by classifying the collection of the target events into several types of events such as mechanical mining, manual mining, drilling, trains, leakage, etc., classification of the event by analyzing the frequency segments of the events and the behavior characteristics of the event, the development situation of the event and the influence on the optical cable, pattern recognition mainly based on a Gaussian Mixture Model (GMM), and classification in two different patterns:
1. the machine + activity identification mode identifies the machine and the activities that the machine performs along the pipeline;
2. the threat detection mode directly identifies whether the activity is an actual threat to the pipeline.
The overall system integrates three main phases
Feature extraction, namely reducing the high dimensionality of signals acquired by a DAS system into a feature set with more informativity and distinguishing property;
feature normalization, which compensates for variations in signal acquisition and sensing position;
pattern classification, classifying the sound signals into a set of predefined Nq classes (using a set of signal models, GMMs, which were previously trained from a labeled signal database),
the feature extraction uses a short-time fast fourier transform (ST FFT) to calculate the energy for each sound frame, and the super-frequency band signal takes the appropriate time window first, which is used as the basis feature vector component. The complete basic feature vector and the corresponding energy calculated at a given optical fiber position, then the initial frequency and the final frequency of the frequency range of the event are defined into bandwidths, and then the collected signals are subjected to optical fiber loss compensation normalization and sensitivity-based normalization processing, wherein the optical fiber loss compensation normalization is mainly used for compensating amplitude exponential decay of the signals due to the optical fiber along with the distance loss. The sensitivity-based normalization is to compensate for the effect of sensitivity inconsistencies so that we obtain a signal of equivalent sensitivity along the fiber, the pattern classification module classifies each feature vector into the most likely class using GMM-based methods (machine + active recognition pattern of active class in machine, use a posterior maximum probability criterion to assign a given feature vector to the class with the highest probability given by the corresponding GMM, resulting in an optimized decision combination.
The embodiments of the present invention are all preferred embodiments of the present invention, and are not intended to limit the scope of the present invention in this way, therefore: all equivalent changes in structure, shape and principle of the invention should be covered in the scope of protection of the invention.
Claims (7)
1. An analysis processing method based on a pipeline safety early warning system is characterized by comprising the following steps of: the method comprises the following steps:
s1: event detection, converting a time domain signal into a frequency domain signal through fast fourier transformation, and then analyzing aiming at frequency, spectrum density, power spectrum and the like:
a1. for non-periodic signals can be expressed as:however, the implementation system processes high-frequency discrete data, so the high-frequency discrete data can be expressed as:
a2: the back scattering signal caused by the target event after the time-frequency conversion needs to be extracted from the measured back scattering signal, wherein the target event is set as y (t), the background signal is set as f (t), the noise signal is set as x (t), and the relation between the back scattering signal and the background signal is set as x (t):
f(t)=y(t)+x(t)
a3: after background noise cancellation, the signal f (t) needs to be digitized f [ c, m ] =y [ c, m ] +x [ c, m ]
Where c is the measurement point (c ε Z) + ) M is a digitized time index, and the interference of background noise generally does not change in a short time, which is different from the behavior of the target signal with high transients, so the number of samples used in the signal power estimation is expressed as:
a4: the noise in the non-target frequency band is filtered by applying the frequency band of the event, and then convolution amplification is carried out by cross-correlation operation of the target signal and the event template signal g (t), namely, sliding multiplication is carried out on two sequences, so that the amplified clear target event signal is obtained: (f×g) (x) = ≡f * (t)g(x+t)dt;
S2: the method comprises the steps of carrying out event reconstruction and feature extraction, adopting a frequency division mode decomposition method to decompose any real value signal into a plurality of discrete modes, wherein Sz, Z=1, … and Z, the modes have the characteristics of reconstructing an input signal, sequencing corresponding values in vectors, compressing each mode around a center frequency, calculating the bandwidth of the mode as a specific norm of Hilbert conversion operation, calculating an analysis signal corresponding to each mode by Hilbert conversion, transferring the analysis signal to a baseband by heterodyne through complex harmonics, adopting an alternate direction multiplication method ADMM to decompose an objective function into a plurality of sub-functions, estimating the bandwidth of the mode by smoothing the demodulated signal, realizing the frequency division mode decomposition method on the basis of the alternate direction multiplication method ADMM,
b1: the problem of optimization is converted into an unconstrained form by introducing an enhanced lagrangian function:
s.t.∑ z s z =f.
alpha is a balance parameter, and lambda is a Lagrangian multiplier; then adopting a series of ADMM iteration sub-optimization calling methods:
b2: the quadratic problem is then mapped in the corresponding fourier domain
Once the event signal is reconstructed, the characteristic signal is then extracted from the reconstructed signal, using the instantaneous amplitude a (n), the instantaneous frequency f (n), the instantaneous phase phi (n) of the instantaneous signal; the true value of the discrete signal s (n), the resolved signal sa (n) can be expressed as
Where I and Q are in-phase and quadrature components, which are derived fromGiven, whereinH {. Cndot. } represents the Hilbert transform; thus, the instantaneous signal characteristics (a (n), phi (n), f (n)) can be calculated as
B3: extracting characteristic signals from the reconstructed signals through transient signal characteristics, normalizing all the characteristics, and calculating the skewness, kurtosis and variance of the characteristics;
s3: the method comprises the steps of event classification, more accurate classification of events by using a pattern recognition machine, more detailed extraction of characteristic parameters, classification of the event into mechanical mining, manual mining, drilling, trains, leakage and other events by classifying the target event, analysis of frequency bands of the events, classification of the behavior characteristics of the events, development situation of the events and influence on an optical cable, and thus obtaining an optimized decision combination.
2. The analysis processing method based on the pipeline safety pre-warning system according to claim 1, wherein the analysis processing method comprises the following steps: the pattern recognition in S3 is mainly based on Gaussian Mixture Model (GMM) classification in two different patterns.
3. The analysis processing method based on the pipeline safety pre-warning system according to claim 2, wherein: one of the two modes in the mode identification is a machine plus activity identification mode for identifying the activity of the machine and the machine along the pipeline, and the other mode in the mode identification is a threat detection mode for directly identifying whether the activity is an actual threat of the pipeline.
4. The analysis processing method based on the pipeline safety precaution system according to claim 3, characterized in that: the pattern recognition system internally comprises a feature extraction unit for reducing the high dimensionality of the signals acquired by the DAS system into more informative and distinguishable feature sets, a feature normalization unit for compensating the change and sensing position in the signal acquisition process and a pattern classification unit for classifying the sound signals into a group of predefined Nq types.
5. The analysis processing method based on the pipeline safety precaution system according to claim 4, characterized in that: the feature extraction unit calculates the energy for each sound frame using a short-time fast fourier transform (ST FFT), the over-frequency band signal takes a suitable time window first, which is used as a base feature vector component, the complete base feature vector calculated at a given fiber position and the corresponding energy, and then the bandwidth is defined for the initial frequency and the final frequency of the frequency range of the event.
6. The analysis processing method based on the pipeline safety precaution system according to claim 4, characterized in that: the characteristic normalization unit performs optical fiber loss compensation normalization and sensitivity-based normalization processing on the acquired signals, and in order to compensate amplitude exponential decay of the signals due to optical fiber loss along with distance, the sensitivity-based normalization is to compensate the influence of inconsistent sensitivity so that the signals with equivalent sensitivity can be obtained along the optical fiber.
7. The analysis processing method based on the pipeline safety precaution system according to claim 4, characterized in that: the pattern classification unit classifies each feature vector into the most probable class (activity recognition pattern of machines and activity classes in machines) using a GMM-based approach, and employs a posterior maximum probability criterion to assign a given feature vector to the class with the highest probability given by the corresponding GMM.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111271313.7A CN113915536B (en) | 2021-10-29 | 2021-10-29 | Analysis processing method based on pipeline safety early warning system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111271313.7A CN113915536B (en) | 2021-10-29 | 2021-10-29 | Analysis processing method based on pipeline safety early warning system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113915536A CN113915536A (en) | 2022-01-11 |
CN113915536B true CN113915536B (en) | 2023-12-01 |
Family
ID=79243629
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111271313.7A Active CN113915536B (en) | 2021-10-29 | 2021-10-29 | Analysis processing method based on pipeline safety early warning system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113915536B (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1598516A (en) * | 2004-07-19 | 2005-03-23 | 天津大学 | Interference distributed fibre-optical pipe leakage real-time monitoring method and device |
CN105156901A (en) * | 2015-07-23 | 2015-12-16 | 中国石油天然气集团公司 | Optical fiber early warning system noise reduction method and device based on wavelet analysis |
CN108317402A (en) * | 2017-12-29 | 2018-07-24 | 上海复栋智能科技有限公司 | A kind of distributing optical fiber sensing positioning early warning analysis method |
CN110410684A (en) * | 2019-07-04 | 2019-11-05 | 华中科技大学 | Non-intruding distributed optical fiber pipeline based on acoustic detection monitors system and method |
CN110751073A (en) * | 2019-10-12 | 2020-02-04 | 武汉理工大学 | Pipeline early damage mode identification method based on distributed optical fiber sensing and deep learning |
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 |
CN111911815A (en) * | 2020-07-06 | 2020-11-10 | 天津精仪精测科技有限公司 | Water pipeline safety early warning system based on optical fiber interferometer |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107425906B (en) * | 2017-07-25 | 2019-09-27 | 电子科技大学 | Distributing optical fiber sensing signal processing method towards underground pipe network safety monitoring |
-
2021
- 2021-10-29 CN CN202111271313.7A patent/CN113915536B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1598516A (en) * | 2004-07-19 | 2005-03-23 | 天津大学 | Interference distributed fibre-optical pipe leakage real-time monitoring method and device |
CN105156901A (en) * | 2015-07-23 | 2015-12-16 | 中国石油天然气集团公司 | Optical fiber early warning system noise reduction method and device based on wavelet analysis |
CN108317402A (en) * | 2017-12-29 | 2018-07-24 | 上海复栋智能科技有限公司 | A kind of distributing optical fiber sensing positioning early warning analysis method |
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 |
CN110410684A (en) * | 2019-07-04 | 2019-11-05 | 华中科技大学 | Non-intruding distributed optical fiber pipeline based on acoustic detection monitors system and method |
CN110751073A (en) * | 2019-10-12 | 2020-02-04 | 武汉理工大学 | Pipeline early damage mode identification method based on distributed optical fiber sensing and deep learning |
CN111911815A (en) * | 2020-07-06 | 2020-11-10 | 天津精仪精测科技有限公司 | Water pipeline safety early warning system based on optical fiber interferometer |
Non-Patent Citations (2)
Title |
---|
基于EMD-AWPP和HOSA-SVM算法的分布式光纤振动入侵信号的特征提取与识别;张燕君;刘文哲;付兴虎;毕卫红;;光谱学与光谱分析(02);577-582 * |
基于光纤传感与智能识别的管道安全预警技术;郝尚青;刘建;崔振伟;杨传江;;石油规划设计(02);38-42 * |
Also Published As
Publication number | Publication date |
---|---|
CN113915536A (en) | 2022-01-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Pinnegar et al. | The S-transform with windows of arbitrary and varying shape | |
Kandamali et al. | Machine learning methods for identification and classification of events in ϕ-OTDR systems: a review | |
Lopez et al. | Box-cox-sparse-measures-based blind filtering: Understanding the difference between the maximum kurtosis deconvolution and the minimum entropy deconvolution | |
Zhao et al. | Distributed acoustic sensing vertical seismic profile data denoiser based on convolutional neural network | |
CN107783200A (en) | Joint EMD and TFPF algorithms a kind of all-wave magnetic resonance signal random noise method for reducing | |
CN103617684A (en) | Interference type optical fiber perimeter vibration intrusion recognition algorithm | |
Sheng et al. | Applications in bearing fault diagnosis of an improved Kurtogram algorithm based on flexible frequency slice wavelet transform filter bank | |
CN111442827A (en) | Optical fiber passive online monitoring system and method for transformer winding vibration | |
Sun et al. | Target location method for pipeline pre-warning system based on HHT and time difference of arrival | |
CN111537056A (en) | Pipeline along-line third-party construction dynamic early warning method based on SVM and time-frequency domain characteristics | |
CN103226028B (en) | Method for identifying and detecting disturbance signals of phase-sensitive optical time domain reflectometer | |
CN109272017A (en) | The vibration signal mode identification method and system of distributed fiberoptic sensor | |
Li et al. | Research on noise reduction method for ship radiate noise based on secondary decomposition | |
CN114199362A (en) | Distributed optical fiber vibration sensor mode identification method | |
CN110672875A (en) | Surface water flow velocity detection method based on Chirp-Z transformation | |
CN104316160B (en) | Underwater signal instantaneous frequency demodulation method based on Wavelet Ridge | |
Liu et al. | Investigating DEM error patterns by directional variograms and Fourier analysis | |
Cheng et al. | Enhanced symplectic characteristics mode decomposition method and its application in fault diagnosis of rolling bearing | |
CN114674352B (en) | Distributed disturbance sensing and demodulation method based on Rayleigh scattering spectrum dissimilarity | |
CN113915536B (en) | Analysis processing method based on pipeline safety early warning system | |
Chen et al. | An AI-based monitoring system for external disturbance detection and classification near a buried pipeline | |
Wang et al. | Attribute-guided target data separation network for DAS VSP data | |
Wang et al. | Cyclic correlation density decomposition based on a sparse and low-rank model for weak fault feature extraction of rolling bearings | |
Gong et al. | Nonlinear vibration feature extraction based on power spectrum envelope adaptive empirical Fourier decomposition | |
Biao et al. | Rotating machine fault diagnosis by a novel fast sparsity-enabled feature-energy-ratio method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |