CN109538944B - Pipeline leakage detection method - Google Patents
Pipeline leakage detection method Download PDFInfo
- Publication number
- CN109538944B CN109538944B CN201811464205.XA CN201811464205A CN109538944B CN 109538944 B CN109538944 B CN 109538944B CN 201811464205 A CN201811464205 A CN 201811464205A CN 109538944 B CN109538944 B CN 109538944B
- Authority
- CN
- China
- Prior art keywords
- leakage
- sample
- neural network
- pipeline
- sample characteristic
- 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
Images
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
- F17D5/02—Preventing, monitoring, or locating loss
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Abstract
The invention discloses a pipeline leakage detection method, which solves the problem of low detection accuracy of the existing method. The method comprises the following steps: collecting stress wave signals of each monitoring point in the pipeline with known leakage condition, extracting characteristic values, and constructing a training sample to obtain sample characteristic values; taking the sample characteristic value as an input signal of a support vector machine classifier, and obtaining an optimal classification surface function of the support vector machine for judging the leakage condition according to the actual leakage condition of each monitoring point; and taking the sample characteristic value as an input signal of the neural network regressor, and obtaining a neural network model and a leakage point model position calculated for the leakage point position according to the actual leakage point position. The invention realizes the accurate judgment and calculation of the pipeline leakage condition and the leakage point position.
Description
Technical Field
The invention relates to the technical field of pipeline leakage detection, in particular to a pipeline leakage detection method.
Background
The detection and positioning of the pipeline leakage relate to multiple disciplines such as pipeline hydrodynamics, thermodynamics, sensing technology, weak signal detection, signal processing and the like, the existing pipeline leakage detection method comprises an acoustic theory method, an infrared detection method and a machine learning method, the acoustic theory method depends on expert experience and an existing knowledge base for detection, and the detection precision is low; the infrared detection method needs detectors with different signals for pipelines with different pipe diameters, so that the cost is high and the detection is complex; the machine learning method comprises a neural network method, an artificial intelligence method and a support vector machine method, the monitoring accuracy of the neural network method is low, the artificial intelligence method needs a large amount of sample data to perform statistical analysis, and the support vector machine method can only judge whether leakage occurs and cannot determine the position of the leakage point.
Disclosure of Invention
The invention provides a pipeline leakage detection method, which solves the problem of low detection accuracy of the existing method.
A method of detecting a leak in a pipe, comprising the steps of: collecting stress wave signals of each monitoring point in the pipeline with known leakage condition, extracting characteristic values, and constructing a training sample to obtain sample characteristic values; taking the sample characteristic value as an input signal of a support vector machine classifier, and obtaining an optimal classification surface function of the support vector machine for judging the leakage condition according to the actual leakage condition of each monitoring point; and taking the sample characteristic value as an input signal of the neural network regressor, and obtaining a neural network model and a leakage point model position calculated for the leakage point position according to the actual leakage point position.
Further, before the step of obtaining an optimal classification surface function of the support vector machine for judging the leakage situation according to the known leakage situation of each monitoring point by using the sample feature value as an input signal of the support vector machine classifier, the method further includes: and performing dimensionality reduction processing on the sample characteristic value by a principal component analysis method to obtain a dimensionality reduction sample characteristic value and a dimensionality reduction matrix, and taking the dimensionality reduction sample characteristic value as an input signal of the support vector machine classifier.
Further, the method further comprises: and optimizing the position of the leakage point model by adopting an Adagarad method according to the known actual position of the leakage point to obtain an optimized model for optimizing the position of the leakage point.
Preferably, the kernel function of the support vector machine classifier is a polynomial inner product function, a radial basis function, or an S-type inner product function.
Preferably, the neural network model adopts a three-layer artificial neural network, and the activation function is a Relu function.
Further, the neural network model adopts a standard BP algorithm, a self-adaptive adjusting learning rate BP algorithm, a momentum-improved BP algorithm, a conjugate gradient method, a Gauss Newton method and a Levenberg Marquardt method.
Further, the characteristic values include: the peak value, the average amplitude value, the variance, the root of square mean, the amplitude value of the root of square, the dimensionless waveform index, the peak value factor, the impulse factor, the margin factor, the kurtosis and the kurtosis factor of the stress wave signal.
Preferably, the frequency of acquisition of the stress wave signal for each monitoring point in the pipeline for which the leakage condition is known is 0.1 Hz.
Preferably, the step of performing a dimensionality reduction process on the sample eigenvalue by a principal component analysis method to obtain a dimensionality reduction sample eigenvalue and a dimensionality reduction matrix, and using the dimensionality reduction sample eigenvalue as an input signal of the support vector machine classifier further includes: carrying out standardization pretreatment on the sample characteristic value to obtain a standardized sample characteristic value; screening dimension reduction space data from the standardized sample characteristic values, and sorting the dimension reduction space data in a descending order according to variance contribution rate to obtain the dimension reduction sample characteristic values; and calculating to obtain the dimension reduction matrix according to the linear product of the dimension reduction matrix and the standardized sample eigenvalue of the dimension reduction sample eigenvalue.
Preferably, the dimension ratio after dimensionality reduction to before dimensionality reduction is 0.75.
The beneficial effects of the invention include: the invention innovatively provides a detection method combining a support vector machine and a neural network, the support vector machine is used for judging whether real-time data is leaked, the position of a leakage point of the leaked leakage point data is further calculated through the neural network, and the judgment accuracy is high; meanwhile, the neural network is optimized, the accuracy of judgment of leakage conditions of the leakage points and calculation of leakage positions are further improved, the method is simple, and the detection cost is low.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of an embodiment of a method for detecting pipeline leakage;
FIG. 2 is a flowchart of an embodiment of a method for pipeline leak detection including eigenvalue dimensionality reduction;
FIG. 3 is a flow chart of an embodiment of a method for detecting pipeline leakage including optimization of the location of the leakage point.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention and the accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The technical solutions provided by the embodiments of the present invention are described in detail below with reference to the accompanying drawings.
Fig. 1 is a flow embodiment of a method for detecting a pipeline leakage, which is used to detect a pipeline leakage situation, and as an embodiment of the present application, the method for detecting a pipeline leakage includes the following steps:
In step 101, the leakage of the fluid in the pipeline is mainly characterized in that a multiphase turbulent jet is formed at the leakage position, high-frequency stress waves are generated on the pipe wall after interaction with the pipeline, and the time-frequency curve of the stress waves without leakage and the time-frequency curve of the stress waves with leakage are obviously different, so that whether the pipeline leaks or not can be judged by extracting the characteristic value of the stress wave signal.
As an embodiment of the present invention, the characteristic values include: the peak value, the average amplitude value, the variance, the root of square mean, the amplitude value of the root of square, the dimensionless waveform index, the peak value factor, the impulse factor, the margin factor, the kurtosis and the kurtosis factor of the stress wave signal. The characteristic value may further include other indexes related to leak detection, and is not particularly limited herein.
In step 101, the training samples are priori information, and the training samples are constructed according to the known leakage situation and the leakage position. For real-time pipeline leakage detection, stress wave signals of all monitoring points also need to be collected, and feature extraction is carried out, wherein extracted features correspond to the sample feature values.
In step 101, as an embodiment of the present invention, the collection frequency of collecting stress wave signals for each monitoring point in the pipeline with known leakage is 0.1Hz, and it should be noted that the collection frequency may be other values, which is not particularly limited herein.
And 103, taking the sample characteristic value as an input signal of a support vector machine classifier, and obtaining an optimal classification surface function of the support vector machine for judging the leakage condition according to the actual leakage condition of each monitoring point.
In step 103, the support vector machine transforms the input space to a high-dimensional space by using the nonlinear transformation defined by the inner product function, i.e. the kernel function, and finds the optimal classification surface in this space to obtain the optimal classification surface function of the support vector machine. The optimal classification surface can correctly classify all the sample characteristic values, and the distance from the point, closest to the optimal classification surface, in the sample characteristic values to the optimal classification surface is the largest, namely the confidence range of the popularization boundary is the smallest.
As an embodiment of the present invention, the kernel function of the support vector machine classifier is a polynomial inner product function, a radial basis function, or an S-type inner product function; the variable of the kernel function is the sample characteristic value. The kernel function may be a kernel function of another form, and is not particularly limited herein.
In step 103, an optimal classification surface function of the support vector machine can be obtained according to the sample characteristic value and the actual leakage condition, and for the characteristic value obtained by collecting and extracting the characteristics in real time, the leakage condition of the collected signal can be determined according to the optimal classification surface function of the support vector machine.
And step 104, taking the sample characteristic value as an input signal of the neural network regressor, and obtaining a neural network model and a leakage point model position calculated for the leakage point position according to the actual leakage point position.
The neural network module is a function of a multi-input single-output nonlinear relation, the input component is multiplied by the corresponding weight component and then added with the threshold value to obtain an activation function, and the error between the output value and the target value is minimized by changing the numerical values of the weight component and the threshold value, so that the neural network model is obtained.
In step 104, the input signal of the neural network regressor is the sample characteristic value, the target value is the actual position of the leak, the output value is the model position of the leak, and the model position of the leak is closest to the actual position of the leak by changing the values of the weight component and the threshold value, so as to obtain the neural network model.
As an embodiment of the present application, the neural network model adopts a standard BP algorithm, and it should be noted that the neural network model may also adopt other algorithms, such as an adaptive learning rate adjusting BP algorithm, a momentum-modified BP algorithm, a conjugate gradient method, a gauss-newton method, and a Levenberg Marquardt method, which are not particularly limited herein.
As an embodiment of the present application, the neural network model adopts a three-layer artificial neural network, the Relu function is selected as the activation function, the Relu function has the advantages of saving calculation amount and alleviating the over-fitting problem, and the activation function may also select other functions, which is not particularly limited herein.
In step 104, the neural network model is obtained according to the sample characteristic value of the training sample and the actual position of the leak, and the leak is located. And for the stress wave signals collected in real time, when the leakage condition is judged to be leakage through a support vector machine classifier, the position of a leakage point is further calculated according to the neural network regressor.
The embodiment of the invention provides a pipeline leakage point detection method, which combines a support vector machine and a neural network, the support vector machine effectively judges the leakage condition, and on the basis, the leakage point position of monitoring point data which is judged to be leaked is calculated through the neural network, so that the detection result is high in accuracy.
Fig. 2 is a flow embodiment of a pipeline leakage detection method including eigenvalue dimension reduction processing, where dimension reduction processing is performed on an input signal entering a support vector machine classifier, and as an embodiment of the present application, a pipeline leakage detection method includes the following steps:
And 102, performing dimensionality reduction on the sample characteristic value by a principal component analysis method to obtain a dimensionality reduction sample characteristic value and a dimensionality reduction matrix, and using the dimensionality reduction sample characteristic value as an input signal of the support vector machine classifier.
In step 102, the purpose of performing the dimensionality reduction processing on the sample feature values is to reduce the types of input signals input into the support vector machine classifier and speed up the computation workload.
The principal component analysis method has the basic idea that dimensionality reduction is carried out on a data set which has higher dimensionality and is mutually associated among all dimensionality variables, and the original information is kept in the data set after dimensionality reduction as much as possible. The principal component analysis method converts an original space into a low-dimensional principal component space through linear conversion, the converted new features are called principal components, the irrelevance among the principal components is met, and the principal components are arranged in a descending order according to variance contribution rates in corresponding directions.
In step 102, the sample characteristic values are screened to obtain the dimension-reduced sample characteristic values, as an embodiment of the present invention, the dimension of the sample characteristic values is 20, the dimension of the dimension-reduced sample characteristic values is 15, and the screening ratio is 0.75. It should be noted that the screening ratio is generally 0.7 to 0.95, and in the embodiment of the present invention, 0.75 is selected, and other values may also be selected, which is not particularly limited herein.
In step 102, performing dimensionality reduction on the sample eigenvalue to obtain a dimensionality reduced sample eigenvalue as an input signal of the support vector machine classifier, and correspondingly, performing dimensionality reduction on an eigenvalue extracted from a stress wave signal acquired in real time and multiplying the eigenvalue by a dimensionality reduction matrix.
And step 204, taking the sample characteristic value after the dimension reduction as an input signal of the support vector machine classifier, and obtaining the optimal classification surface function of the support vector machine for judging the leakage condition according to the actual leakage condition of each monitoring point.
In step 204, the input signal of the support vector machine classifier is the feature value of the reduced dimension sample.
And step 104, taking the sample characteristic value as an input signal of the neural network regressor, and obtaining a neural network model and a leakage point model position calculated for the leakage point position according to the actual leakage point position.
In step 104, the input signal of the neural network regressor is the sample characteristic value, and due to the multi-input single-output characteristic of the neural network, dimension reduction processing is not required.
The pipeline leakage detection method provided by the embodiment of the invention has the advantages that the characteristic values are subjected to dimensionality reduction and then input to the support vector machine classifier, and the calculation rate of the support vector machine classifier is improved.
Fig. 3 is a flow embodiment of a method for detecting pipeline leakage including optimizing a position of a leakage point, where the position of the leakage point is optimally calculated, and as an embodiment of the present application, the method for detecting pipeline leakage includes the following steps:
In step 101, stress wave signals are collected for each monitoring point in the pipeline with known leakage conditions, characteristic values are extracted, and accordingly, the same characteristic values also need to be extracted for the stress wave signals returned to each monitoring point in real time.
In step 201, the sample feature values are subjected to normalization preprocessing to eliminate the influence between different dimensions.
In step 201, the sample characteristic value is subjected to a normalization preprocessing, and accordingly, the characteristic value of the stress wave signal obtained in real time is also subjected to the normalization preprocessing.
In step 203, the dimensionality reduction matrix can be obtained by solving according to a lagrangian factor method in linear algebra.
In step 203, the eigenvalue of the stress wave signal obtained in real time is also multiplied by the dimensionality reduction matrix to perform dimensionality reduction processing.
And 204, taking the characteristic value of the dimensionality reduction sample as an input signal of a support vector machine classifier, and obtaining an optimal classification surface function of the support vector machine for judging the leakage condition according to the actual leakage condition of each monitoring point.
In step 204, the dimensionality reduction sample eigenvalue is used as an input signal of the support vector machine classifier, and accordingly, after dimensionality reduction is performed on the eigenvalue of the stress wave signal obtained in real time, the eigenvalue is input into the support vector machine classifier, and the leakage condition of the input signal is determined according to the optimal classification surface function of the support vector machine.
And step 104, taking the sample characteristic value as an input signal of the neural network regressor, and obtaining a neural network model and a leakage point model position calculated for the leakage point position according to the actual leakage point position.
And 105, optimizing the position of the leakage point model by adopting an Adagarad method according to the known actual position of the leakage point to obtain an optimized model for optimizing the position of the leakage point.
In step 105, the positions of the leakage points are optimized and calculated, and the reason for adopting the Adagrad method is that for various leakage situations, the positions of some leakage points can be optimized and updated, and the positions of some leakage points are not optimized, that is, the Adagrad method can keep different output update rates for the input parameters.
In step 105, correspondingly, the leakage classification is performed on the stress wave signals collected in real time, and after the position of the leakage point is calculated, the position of the leakage point needs to be optimized by using an Adagrad method.
The pipeline leakage detection method provided by the embodiment of the invention comprises the step of optimizing and calculating the position of the leakage point, and the position of the leakage point can be provided more accurately.
It is to be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above description is only an example of the present invention, and is not intended to limit the present invention. Various modifications and alterations to this invention will become apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.
Claims (7)
1. A method of detecting a leak in a pipe, comprising the steps of:
collecting stress wave signals of each monitoring point in the pipeline with known leakage condition, extracting characteristic values, and constructing a training sample to obtain sample characteristic values;
performing dimensionality reduction treatment on the sample characteristic value by a principal component analysis method to obtain a dimensionality reduction sample characteristic value and a dimensionality reduction matrix;
taking the characteristic value of the dimensionality reduction sample as an input signal of a support vector machine classifier, and obtaining an optimal classification surface function of the support vector machine for judging the leakage condition according to the actual leakage condition of each monitoring point;
taking the sample characteristic value as an input signal of a neural network regressor, and obtaining a neural network model and a leakage point model position calculated for the leakage point position according to the actual leakage point position;
optimizing the position of the leakage point model by adopting an Adagarad method according to the known actual position of the leakage point to obtain an optimized model for optimizing the position of the leakage point;
the characteristic values include: the peak value, the average amplitude value, the variance, the root of square mean, the amplitude value of the root of square, the dimensionless waveform index, the peak value factor, the impulse factor, the margin factor, the kurtosis and the kurtosis factor of the stress wave signal.
2. The pipe leak detection method of claim 1, wherein the kernel function of the support vector machine classifier is a polynomial inner product function, a radial basis function, or an sigmoid inner product function.
3. The pipe leak detection method of claim 1, wherein the neural network model employs a three-layer artificial neural network, and the activation function is a Relu function.
4. The method of detecting pipe leakage according to claim 1, wherein the neural network model employs a standard BP algorithm, an adaptive learning rate adjusted BP algorithm, a momentum-modified BP algorithm, a conjugate gradient method, a gauss-newton method, a levenberg marquardt method.
5. The method of pipeline leak detection according to claim 1, wherein the frequency of stress wave signal acquisition for each monitoring point in the pipeline for which the leak condition is known is 0.1 Hz.
6. The method of claim 1, wherein the step of performing a dimensionality reduction process on the sample eigenvalue by a principal component analysis method to obtain a dimensionality reduced sample eigenvalue and a dimensionality reduced matrix, and using the dimensionality reduced sample eigenvalue as an input signal of the SVM classifier further comprises:
carrying out standardization pretreatment on the sample characteristic value to obtain a standardized sample characteristic value;
screening dimension reduction space data from the standardized sample characteristic values, and sorting the dimension reduction space data in a descending order according to variance contribution rate to obtain the dimension reduction sample characteristic values;
and calculating to obtain the dimension reduction matrix according to the linear product of the dimension reduction matrix and the standardized sample eigenvalue of the dimension reduction sample eigenvalue.
7. The pipe leak detection method of claim 1, wherein the ratio of the dimensions after dimensionality reduction to the dimensions before dimensionality reduction is 0.75.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811464205.XA CN109538944B (en) | 2018-12-03 | 2018-12-03 | Pipeline leakage detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811464205.XA CN109538944B (en) | 2018-12-03 | 2018-12-03 | Pipeline leakage detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109538944A CN109538944A (en) | 2019-03-29 |
| CN109538944B true CN109538944B (en) | 2020-07-07 |
Family
ID=65852380
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811464205.XA Active CN109538944B (en) | 2018-12-03 | 2018-12-03 | Pipeline leakage detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109538944B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110319982B (en) * | 2019-06-03 | 2021-03-30 | 清华大学合肥公共安全研究院 | Buried gas pipeline leakage judgment method based on machine learning |
| CN110792928B (en) * | 2019-09-24 | 2021-08-10 | 中国石油化工股份有限公司 | Pipeline leakage diagnosis method based on big data combined algorithm |
| CN111639711B (en) * | 2020-05-29 | 2023-04-18 | 中国地质大学（武汉） | Oil pipeline leakage monitoring method based on pressure monitoring time sequence data |
| CN112555689B (en) * | 2020-11-22 | 2022-09-27 | 武汉第二船舶设计研究所(中国船舶重工集团公司第七一九研究所) | Multi-sensing pipeline state intelligent monitoring device |
| CN112613484B (en) * | 2021-01-06 | 2022-05-20 | 西南石油大学 | Gas pipeline leakage identification method based on singular spectrum analysis and support vector machine |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1904581A (en) * | 2006-08-01 | 2007-01-31 | 东北电力大学 | Oil gas water multiphase flow type identification method based on main component analysis and supporting vector machine |
| CN101008992A (en) * | 2006-12-30 | 2007-08-01 | 北京市劳动保护科学研究所 | Method for detecting leakage of pipeline based on artificial neural network |
| CN104061445A (en) * | 2014-07-09 | 2014-09-24 | 中国石油大学(华东) | Pipeline leakage detection method based on neural network |
| CN104373821A (en) * | 2014-11-21 | 2015-02-25 | 天津科技大学 | Natural gas pipeline safety monitoring device based on acoustical science active spurring |
| CN105864643A (en) * | 2016-03-24 | 2016-08-17 | 华北电力大学 | Gas pipeline leakage positioning experimental device and method based on RBF neural network |
| CN206130547U (en) * | 2016-07-07 | 2017-04-26 | 北京信息科技大学 | Gas transmission pipeline leak testing system under multiplex condition |
| CN106706215A (en) * | 2016-11-17 | 2017-05-24 | 深圳市天成智能控制科技有限公司 | Thermodynamic system valve inner leakage monitoring method |
| CN108870090A (en) * | 2018-06-22 | 2018-11-23 | 大连理工大学 | Pipeline leakage detection method based on the fusion of least square method supporting vector machine information |
-
2018
- 2018-12-03 CN CN201811464205.XA patent/CN109538944B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1904581A (en) * | 2006-08-01 | 2007-01-31 | 东北电力大学 | Oil gas water multiphase flow type identification method based on main component analysis and supporting vector machine |
| CN101008992A (en) * | 2006-12-30 | 2007-08-01 | 北京市劳动保护科学研究所 | Method for detecting leakage of pipeline based on artificial neural network |
| CN104061445A (en) * | 2014-07-09 | 2014-09-24 | 中国石油大学(华东) | Pipeline leakage detection method based on neural network |
| CN104373821A (en) * | 2014-11-21 | 2015-02-25 | 天津科技大学 | Natural gas pipeline safety monitoring device based on acoustical science active spurring |
| CN105864643A (en) * | 2016-03-24 | 2016-08-17 | 华北电力大学 | Gas pipeline leakage positioning experimental device and method based on RBF neural network |
| CN206130547U (en) * | 2016-07-07 | 2017-04-26 | 北京信息科技大学 | Gas transmission pipeline leak testing system under multiplex condition |
| CN106706215A (en) * | 2016-11-17 | 2017-05-24 | 深圳市天成智能控制科技有限公司 | Thermodynamic system valve inner leakage monitoring method |
| CN108870090A (en) * | 2018-06-22 | 2018-11-23 | 大连理工大学 | Pipeline leakage detection method based on the fusion of least square method supporting vector machine information |
Non-Patent Citations (2)
| Title |
|---|
| 《埋地燃气管道防腐性评估与破损点识别》;马子超;《工程科技Ⅱ辑》;20180131;第C038-1196页 * |
| 《基于BP神经网络的管道泄漏声信号识别方法研究》;焦敬品等;《仪器仪表学报》;20161130;第2588-2596页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109538944A (en) | 2019-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109538944B (en) | Pipeline leakage detection method | |
| CN111037365B (en) | Cutter state monitoring data set enhancing method based on generative countermeasure network | |
| KR102018330B1 (en) | Apparatus and method for detecting anomaly behavior in plant pipe using multiple meta-learning | |
| Chuang et al. | Leak detection in water distribution pipes based on CNN with mel frequency cepstral coefficients | |
| US11630034B2 (en) | Method for diagnosing and predicting operation conditions of large-scale equipment based on feature fusion and conversion | |
| CN101196872A (en) | Leakage detecting and locating method based on pressure and sound wave information amalgamation | |
| Abdulla et al. | Probabilistic multiple model neural network based leak detection system: Experimental study | |
| CN111896616B (en) | Gas-liquid two-phase flow pattern identification method based on acoustic emission-BP neural network | |
| US20230417622A1 (en) | Methods and systems for leakage analysis of urban pipelines and storage media | |
| CN111695465B (en) | Pipe network fault diagnosis and positioning method and system based on pressure wave mode identification | |
| CN110953488B (en) | Gas-liquid two-phase flow pipeline leakage acoustic emission detection method based on stack self-coding | |
| CN114782745A (en) | Ocean sound velocity profile classification method and device based on machine learning | |
| CN111931983B (en) | Precipitation prediction method and system | |
| CN115169733A (en) | Deep learning-based method for predicting resuspension quantity of deep sea sediments by using internal solitary waves | |
| CN114963030A (en) | Water supply pipeline monitoring method | |
| Zhang et al. | Rolling bearing fault diagnosis based on time-frequency feature extraction and IBA-SVM | |
| CN110701487B (en) | KPCA and Cas-SVDD-based multi-working-condition pipeline leakage detection method | |
| CN116609440B (en) | Intelligent acceptance management method and system for building engineering quality based on cloud edge cooperation | |
| CN116680639A (en) | Deep-learning-based anomaly detection method for sensor data of deep-sea submersible | |
| Zhang et al. | The method for determining optimal analysis length of vibration data based on improved multiscale permutation entropy | |
| CN110619176A (en) | Aviation kerosene flash point prediction method based on DBN-RLSSVM | |
| Atmani et al. | Enhancement in bearing fault classification parameters using Gaussian mixture models and Mel frequency cepstral coefficients features | |
| CN114444544A (en) | Signal classification and identification method based on convolutional neural network and knowledge migration | |
| Chen | A Research on Fault Diagnosis of Wind Turbine CMS Based on Bayesian-GAN-LSTM Neural Network | |
| Pang et al. | Real-time monitoring of fluidized bed agglomerating based on improved adaboost algorithm |
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 |