CN110145695A - A kind of hot duct leakage detection method based on the fusion of the depth confidence network information - Google Patents
A kind of hot duct leakage detection method based on the fusion of the depth confidence network information Download PDFInfo
- Publication number
- CN110145695A CN110145695A CN201910478572.3A CN201910478572A CN110145695A CN 110145695 A CN110145695 A CN 110145695A CN 201910478572 A CN201910478572 A CN 201910478572A CN 110145695 A CN110145695 A CN 110145695A
- Authority
- CN
- China
- Prior art keywords
- information
- detection model
- depth confidence
- hot duct
- fusion
- 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.)
- Granted
Links
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
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Examining Or Testing Airtightness (AREA)
Abstract
The present invention provides a kind of hot duct leakage detection method based on the fusion of the depth confidence network information, comprising: S1: selection includes the double-deck information fusion detection model for being limited the depth confidence net that Boltzmann machine and the least square support vector classification based on Gauss-Cos mixed kernel function combine and establishing hot duct leakage;S2: the accuracy of acquisition new data information test information fusion detection model;S3: when accuracy is 95% or more, information fusion detection model meets the requirements, into S6;Otherwise enter S4 to S5;S4: being updated using Gibbs sampling iteration and improves information fusion detection model, until the corresponding eigenmatrix of all initial data was selected;S5: updating depth confidence net weight bias matrix model parameter, completes the update of information fusion detection model, returns to S2 to S3;S6: the eigenmatrix input information of pipeline to be detected is merged into detection model, output channel state classification result.The present invention solves the problems, such as that pipe leakage method low efficiency, precision are low.
Description
Technical field
The present invention relates to heating main pipeline intelligence leak detection fields, specifically, more particularly to a kind of being set based on depth
The hot duct leakage detection method of communication network information fusion.
Background technique
It gradually popularizes and comes in the whole nation along with central heating, the laying of hot duct also gradually increases, for steam heating pipe
For road, the integrity of pipeline provides enough guarantees for heating.Consequent is that heating safety is also gradually taken seriously,
But due to the increase of pipeline total amount, safety accident also gradually increases, and the personal safety and property safety etc. to the people all cause
Huge threat.There are many factors for the generation of safety accident, are not limited solely to natural cause, the aging of pipeline itself, road
Not the paying attention to of road construction, artificial operation error etc. can all cause possible safety problem, once safety accident occurs, not only
It is serious possibly even to can cause casualties along with the loss of economic asset.So being to prevent to the enough attention of pipeline
The key factor of safety accident, if the leakage of hot duct cannot be detected in time, it is possible to will cause serious economic damage
It becomes estranged casualties.Due to the buried underground of hot duct, so be a thing very troublesome to the monitoring of pipeline, it is timely
The leakage for discovering pipeline needs a large amount of detection device and staff, and when the leakage of pipeline occurs, relevant staff needs
The generation of leakage is perceived in time according to the variation of monitoring parameter, and accurately finds, position the accurate location of leakage, is supplied ensureing
Under the premise of warm safety, guarantee heating system even running.
Summary of the invention
The low problem of pipe leakage low efficiency, precision is carried out according to the prior art set forth above, and is provided a kind of based on deep
Spend the hot duct leakage detection method of confidence network information fusion.The present invention mainly utilizes depth confidence net learning algorithm, benefit
Eigenmatrix training, the least square supporting vector based on Gauss-Cos mixed kernel function point are carried out with the double-deck Boltzmann machine
Class machine algorithm realizes leak diagnostics, to achieve the purpose that improve detection accuracy and detection efficiency.
The technological means that the present invention uses is as follows:
A kind of hot duct leakage detection method based on the fusion of the depth confidence network information, comprising the following steps:
S1: for the initial data of acquisition hot duct leakage experiment as initial training sample, the initial data is by just
Hot duct pressure signal and flow signal under normal state and leak condition, extract the feature of pressure signal and flow signal respectively
Matrix, eigenmatrix are made of mean value, root mean square, kurtosis and degree of skewness;
Using the corresponding eigenmatrix of initial data as input information, two depth confidence networks, two moulds are inputted respectively
The pipeline conditions classification results of type output include the normal condition and leak condition indicated respectively with 0 and 1, and two models are carried out
Information fusion: selection includes the double-deck limited Boltzmann machine and the least square based on Gauss-Cos mixed kernel function support to
The depth confidence net that amount classifier combines establishes the information fusion detection model of hot duct leakage, when the pipe of two models output
When road state classification result is consistent, using the pipeline conditions classification results as the pipeline conditions point of information fusion detection model output
Class result;
S2: new data information is obtained, the new data information is the known time of day obtained after training detection model
Hot duct pressure signal eigenmatrix and flow signal eigenmatrix;
New data information input information is merged into detection model, output channel state classification result is for testing information fusion
The accuracy of detection model;
S3: in all data samples that new data information includes, pipeline conditions classification results and the consistent number of time of day
The accuracy of information fusion detection model is indicated according to the ratio that sample accounts for all data samples;
When accuracy is 95% or more, indicates that information fusion detection model meets the requirements, merged as final information
Detection model enters S6;Otherwise enter S4 to S5 to be updated information fusion detection model;
S4: initial calculation position of the corresponding eigenmatrix of an initial data as model is randomly selected each time, is adopted
Repeat K training with the limited Boltzmann machine of bilayer, updated using Gibbs sampling iteration and improves information fusion detection mould
Type obtains the edge distribution of training data using limited Boltzmann machine, as the iterative parameter of Gibbs sampling, until all
The corresponding eigenmatrix of initial data was selected;
S5: according to the edge distribution of training data, depth confidence net weight bias matrix model parameter is updated, completes information
The update of detection model is merged, the accuracy that S2 to S3 retests information fusion detection model is returned;
S6: the spy that the mean value of the pressure signal of pipeline to be detected and flow signal, root mean square, kurtosis, degree of skewness are formed
It levies the final information of Input matrix and merges detection model, export pipeline conditions classification results to be detected.
Further, the energy function of the Boltzmann machine are as follows:
Wherein, v is the neuron node of aobvious layer, and h is the neuron node of hidden layer, and i and j respectively represent every layer of Boltzmann
The number of nodes of machine, b and c are respectively the bias of aobvious layer and hidden node, WijIndicate the weight square between every layer of neuron node
Battle array.
Further, in step S5, depth confidence net weight bias matrix model parameter is calculated using following formula:
W→W+λ(P(h1|v1)v1-P(h2|v2)v2)
b→b+λ(v1-v2)
c→c+λ(h1-h2) (2)
Wherein: λ indicates that learning rate, P indicate that the probability for the aobvious layer activation that neuron is reconstructed in hidden layer, v and h represent aobvious
The node of layer and hidden layer, b and c are respectively the bias of aobvious layer and hidden layer.
Further, in step S1, the least square support vector classification based on Gauss-Cos mixed kernel function is used
The following formula of Gauss-Cos mixed kernel function calculated:
Wherein, σ is the parameter of Gauss-Cos mixed kernel function.
Compared with the prior art, the invention has the following advantages that
Hot duct leakage detection method provided by the invention based on the fusion of the depth confidence network information, using the double-deck glass
The graceful machine of Wurz and least square support vector classification information fusion method based on Gauss-Cos mixed kernel function are to heating pipe
Road detection is modeled, and to realize the prediction to pipeline conditions, depth confidence network method is given, using the double-deck Boltzmann
The method of machine and gibbs sampling, the feature for effectively extracting input information make the meter of leak detection aspect of model matrix update simultaneously
It calculates rapider;In view of pipeline training sample set is limited, least square support vector classification is on Small Sample Database collection
Excellent real-time and classification accuracy are instructed using the least square support vector classification based on Gauss-Cos mixed kernel function
Classification output model is got, BP classification method is faster than in speed and accuracy;Using the double-deck Boltzmann machine and it is based on
The least square support vector classification information fusion method of Gauss-Cos mixed kernel function than traditional depth confidence network,
The methods of least square support vector classification, BP have excellent classification diagnosis performance.The present invention can adapt to pipeline work
The needs of state-detection in the process further effectively enhance model prediction ability, improve the precision of prediction of model.
To sum up, it applies the technical scheme of the present invention and is mainly believed using the double-deck Boltzmann machine progress eigenmatrix training
Number depth characteristic value, least square support vector classification based on Gauss-Cos mixed kernel function realizes leak diagnostics, from
And achieve the purpose that improve detection accuracy and detection efficiency.Therefore, technical solution of the present invention solves the prior art and carries out pipe
The problem that road leak efficiency is low, precision is low.
The present invention can be widely popularized in fields such as hot duct leak detections based on the above reasons.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to do simply to introduce, it should be apparent that, the accompanying drawings in the following description is this hair
Bright some embodiments for those of ordinary skill in the art without any creative labor, can be with
It obtains other drawings based on these drawings.
Fig. 1 is hot duct leakage detection method schematic illustration of the present invention.
Fig. 2 is hot duct leakage detection method flow chart of the present invention.
Fig. 3 is hot duct leakage detection method testing result schematic diagram of the present invention.
Fig. 4 is hot duct leakage detection method testing result schematic diagram of the present invention.
Fig. 5 is hot duct leakage detection method testing result schematic diagram of the present invention.
Fig. 6 is hot duct leakage detection method testing result schematic diagram of the present invention.
Specific embodiment
It should be noted that in the absence of conflict, the feature in embodiment and embodiment in the present invention can phase
Mutually combination.The present invention will be described in detail below with reference to the accompanying drawings and embodiments.
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
It is only a part of the embodiment of the present invention, instead of all the embodiments.It is real to the description of at least one exemplary embodiment below
It is merely illustrative on border, never as to the present invention and its application or any restrictions used.Based on the reality in the present invention
Example is applied, every other embodiment obtained by those of ordinary skill in the art without making creative efforts all belongs to
In the scope of protection of the invention.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to exemplary embodiments of the present invention.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
Unless specifically stated otherwise, positioned opposite, the digital table of the component and step that otherwise illustrate in these embodiments
It is not limited the scope of the invention up to formula and numerical value.Simultaneously, it should be clear that for ease of description, each portion shown in attached drawing
The size divided not is to draw according to actual proportionate relationship.Technology known for person of ordinary skill in the relevant, side
Method and equipment may be not discussed in detail, but in the appropriate case, and the technology, method and apparatus should be considered as authorizing explanation
A part of book.In shown here and discussion all examples, appointing should be construed as merely illustratively to occurrence, and
Not by way of limitation.Therefore, the other examples of exemplary embodiment can have different values.It should also be noted that similar label
Similar terms are indicated in following attached drawing with letter, therefore, once it is defined in a certain Xiang Yi attached drawing, then subsequent attached
It does not need that it is further discussed in figure.
In the description of the present invention, it is to be understood that, the noun of locality such as " front, rear, top, and bottom, left and right ", " it is laterally, vertical,
Vertically, orientation or positional relationship indicated by level " and " top, bottom " etc. is normally based on orientation or position shown in the drawings and closes
System, is merely for convenience of description of the present invention and simplification of the description, in the absence of explanation to the contrary, these nouns of locality do not indicate that
It must have a particular orientation or be constructed and operated in a specific orientation with the device or element for implying signified, therefore cannot manage
Solution is limiting the scope of the invention: the noun of locality " inside and outside " refers to inside and outside the profile relative to each component itself.
For ease of description, spatially relative term can be used herein, as " ... on ", " ... top ",
" ... upper surface ", " above " etc., for describing such as a device shown in the figure or feature and other devices or spy
The spatial relation of sign.It should be understood that spatially relative term is intended to comprising the orientation in addition to device described in figure
Except different direction in use or operation.For example, being described as if the device in attached drawing is squeezed " in other devices
It will be positioned as " under other devices or construction after part or construction top " or the device of " on other devices or construction "
Side " or " under its device or construction ".Thus, exemplary term " ... top " may include " ... top " and
" in ... lower section " two kinds of orientation.The device can also be positioned with other different modes and (is rotated by 90 ° or in other orientation), and
And respective explanations are made to the opposite description in space used herein above.
In addition, it should be noted that, limiting components using the words such as " first ", " second ", it is only for be convenient for
Corresponding components are distinguished, do not have Stated otherwise such as, there is no particular meanings for above-mentioned word, therefore should not be understood as to this
The limitation of invention protection scope.
Embodiment 1
As shown in Figs. 1-2, the present invention provides a kind of hot ducts based on the fusion of the depth confidence network information to leak inspection
Survey method, comprising the following steps:
S1: for the initial data of acquisition hot duct leakage experiment as initial training sample, the initial data is by just
Hot duct pressure signal and flow signal under normal state and leak condition, extract the feature of pressure signal and flow signal respectively
Matrix, eigenmatrix are made of mean value, root mean square, kurtosis and degree of skewness;
Using the corresponding eigenmatrix of initial data as input information, two depth confidence networks, two moulds are inputted respectively
The pipeline conditions classification results of type output include the normal condition and leak condition indicated respectively with 0 and 1, and two models are carried out
Information fusion: selection includes the double-deck limited Boltzmann machine and the least square based on Gauss-Cos mixed kernel function support to
The depth confidence net that amount classifier combines establishes the information fusion detection model of hot duct leakage, when the pipe of two models output
When road state classification result is consistent, using the pipeline conditions classification results as the pipeline conditions point of information fusion detection model output
Class result;
S2: new data information is obtained, the new data information is the known time of day obtained after training detection model
Hot duct pressure signal eigenmatrix and flow signal eigenmatrix;
New data information input information is merged into detection model, output channel state classification result is for testing information fusion
The accuracy of detection model;
S3: in all data samples that new data information includes, pipeline conditions classification results and the consistent number of time of day
The accuracy of information fusion detection model is indicated according to the ratio that sample accounts for all data samples;
When accuracy is 95% or more, indicates that information fusion detection model meets the requirements, merged as final information
Detection model enters S6;Otherwise enter S4 to S5 to be updated information fusion detection model;
S4: initial calculation position of the corresponding eigenmatrix of an initial data as model is randomly selected each time, is adopted
Repeat K training with the limited Boltzmann machine of bilayer, updated using Gibbs sampling iteration and improves information fusion detection mould
Type obtains the edge distribution of training data using limited Boltzmann machine, as the iterative parameter of Gibbs sampling, until all
The corresponding eigenmatrix of initial data was selected;
S5: according to the edge distribution of training data, depth confidence net weight bias matrix model parameter is updated, completes information
The update of detection model is merged, the accuracy that S2 to S3 retests information fusion detection model is returned;
S6: the spy that the mean value of the pressure signal of pipeline to be detected and flow signal, root mean square, kurtosis, degree of skewness are formed
It levies the final information of Input matrix and merges detection model, export pipeline conditions classification results to be detected.
Further, the energy function of the Boltzmann machine are as follows:
Wherein, v is the neuron node of aobvious layer, and h is the neuron node of hidden layer, and i and j respectively represent every layer of Boltzmann
The number of nodes of machine, b and c are respectively the bias of aobvious layer and hidden node, WijIndicate the weight square between every layer of neuron node
Battle array.
Further, in step S5, depth confidence net weight bias matrix model parameter is calculated using following formula:
W→W+λ(P(h1|v1)v1-P(h2|v2)v2)
b→b+λ(v1-v2)
c→c+λ(h1-h2) (2)
Wherein: λ indicates that learning rate, P indicate that the probability for the aobvious layer activation that neuron is reconstructed in hidden layer, v and h represent aobvious
The node of layer and hidden layer, b and c are respectively the bias of aobvious layer and hidden layer.
Further, the mixed kernel function of the least square support vector classification based on Gauss-Cos mixed kernel function
Construction:
Common local functions have RBF kernel function, have stronger learning ability, but Generalization Capability is relatively weak.Often
Global kernel function has Polynomial kernel function, has stronger Generalization Capability, but learning ability is weaker.In order to improve minimum
Two multiply support vector classification performance, and the application uses mixed kernel function mode, and there are two types of mode, the first sides for mixed kernel function
Method is according to Mercer principle and linear weighted function principle, and two kernel functions are added or kernel function.Second method Mercer principle
With the still kernel function that is multiplied after two kernel functions multiplications of principle.First method increases the selection of parameter due to using weight
Calculation amount.Precision of prediction can be improved using multiplication principle in second method, and does not increase parameter and computation complexity.This hair
It is bright to use second method, Gauss-Cos kernel function is constructed, Gauss-Cos is a kind of mixed kernel function, both comprising part letter
Breath, and include global information, precision of prediction can be improved, and do not increase parameter and computation complexity;In step S1, it is based on
The following formula of Gauss-Cos mixed kernel function that the least square support vector classification of Gauss-Cos mixed kernel function uses
It is calculated:
Wherein, σ is the parameter of Gauss-Cos mixed kernel function.
Least square support vector classification algorithm principle based on Gauss-Cos mixed kernel function:
If sample is n-dimensional vector, the L sample in certain region is indicated are as follows: (x1,y1),…,(xL,yL)∈Rn* R, first with one
Nonlinear Mapping Ψ () is sample from former space RnIt is mapped to feature spaceOptimal determine is constructed in this higher dimensional space
Plan function:
X in above formulaiIt is the feature of i-th of vector, ω and b are the parameters that feature space is mapped to higher dimensional space;It is N pairs given
It is trained, calculates hyperplane parameter ω, b, e value satisfaction:
It is solved with Lagrangian method:
Wherein αiIt is Lagrange multiplier, according to optimal conditions:
It can be indicated with matrix form are as follows:
In above formulaY=[yi,…,yN],E=[ei,…,
eN], α=[αi,…,αN];
Matrix solution is:
Using the prediction model of the available Nonlinear Classifier of Mercer condition, in which:
Solve above-mentioned system of linear equations can be obtained classifier prediction model it is as follows:
K is kernel function in above formula;It is to meet any symmetrical kernel function of Mercer condition corresponding to feature space
Dot product.
Fig. 3-6 is used to illustrate the testing result of pipeline leakage detection method of the present invention: as can be seen that using this hair
Precision under the leak detection model of bright technical solution is 3000 in data volume is very high, and prediction numerical value is very accurate, can be true
The operating status under practical pipeline is predicted, the precision under being 6000 in data volume is declined slightly, but by adjusting RBM node layer
Number more can really show the actual motion state of pipeline, and error precision is within 5%;Based on depth confidence net algorithm
Leak detection model accuracy such as table 1, all greater than 95%, the precision for meeting long range hot duct status assessment is asked, tool
There are good ability of self-teaching and stability.
Accuracy of the 1 leak detection model of table under 3000 and 6000 data volumes
Method choice long range hot duct pressure signal and flow signal operation disclosed by the invention and state parameter
The information characteristics such as input of mean value, root mean square, degree of skewness and kurtosis as model, the pipeline conditions to be predicted are as the defeated of model
Out, the history data of leakage experiment is chosen as initial training sample, establishes leak detection using depth confidence network method
Initial model.In addition, being optimized using the double-deck Boltzmann machine to model, and using based on Gauss-Cos mixed nucleus letter
The mode of several least square support vector classifications carrys out the precision of prediction of computation model.
Hot duct leakage detection method proposed by the present invention based on depth confidence net with heating process characteristic change
Change adaptively improved model performance, can be realized the accurate prediction to pipe leakage state, it is real-time to the leakage of collective's heating
Monitoring has important meaning.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify to technical solution documented by previous embodiment, or some or all of the technical features are equal
Replacement;And these are modified or replaceed, the model for technical solution of the embodiment of the present invention that it does not separate the essence of the corresponding technical solution
It encloses.
Claims (4)
1. a kind of hot duct leakage detection method based on the fusion of the depth confidence network information, it is characterised in that: including following
Step:
S1: for the initial data of acquisition hot duct leakage experiment as initial training sample, the initial data is by normal shape
Hot duct pressure signal and flow signal under state and leak condition extract the feature square of pressure signal and flow signal respectively
Battle array, eigenmatrix are made of mean value, root mean square, kurtosis and degree of skewness;
Using the corresponding eigenmatrix of initial data as input information, two depth confidence networks are inputted respectively, and two models are defeated
Pipeline conditions classification results out include the normal condition and leak condition indicated respectively with 0 and 1, and two models are carried out information
Fusion: selection includes the double-deck limited Boltzmann machine and the least square supporting vector point based on Gauss-Cos mixed kernel function
The depth confidence net that class machine combines establishes the information fusion detection model of hot duct leakage, when the pipe-like of two models output
When state classification results are consistent, which is classified as the pipeline conditions of information fusion detection model output and is tied
Fruit;
S2: new data information is obtained, the new data information is the confession of the known time of day obtained after training detection model
The eigenmatrix of the pressure signal in heating coil road and the eigenmatrix of flow signal;
New data information input information is merged into detection model, output channel state classification result is for testing information fusion detection
The accuracy of model;
S3: in all data samples that new data information includes, pipeline conditions classification results and the consistent data sample of time of day
Originally the ratio for accounting for all data samples indicates the accuracy of information fusion detection model;
When accuracy is 95% or more, indicates that information fusion detection model meets the requirements, merge detection as final information
Model enters S6;Otherwise enter S4 to S5 to be updated information fusion detection model;
S4: initial calculation position of the corresponding eigenmatrix of an initial data as model is randomly selected each time, using double
The limited Boltzmann machine of layer repeats K training, and information fusion detection model is updated and improved using Gibbs sampling iteration,
The edge distribution of training data is obtained using limited Boltzmann machine, as the iterative parameter of Gibbs sampling, until all original
The corresponding eigenmatrix of data was selected;
S5: according to the edge distribution of training data, updating depth confidence net weight bias matrix model parameter, completes information fusion
The update of detection model returns to the accuracy that S2 to S3 retests information fusion detection model;
S6: the feature square that the mean value of the pressure signal of pipeline to be detected and flow signal, root mean square, kurtosis, degree of skewness are formed
Battle array inputs final information and merges detection model, exports pipeline conditions classification results to be detected.
2. the hot duct leakage detection method according to claim 1 based on the fusion of the depth confidence network information, special
Sign also resides in: the energy function of the Boltzmann machine are as follows:
Wherein, v is the neuron node of aobvious layer, and h is the neuron node of hidden layer, and i and j respectively represent every layer of Boltzmann machine
Number of nodes, b and c are respectively the bias of aobvious layer and hidden node, WijIndicate the weight matrix between every layer of neuron node.
3. the hot duct leakage detection method according to claim 1 based on the fusion of the depth confidence network information, special
Sign also resides in: in step S5, depth confidence net weight bias matrix model parameter is calculated using following formula:
W→W+λ(P(h1|v1)v1-P(h2|v2)v2)
b→b+λ(v1-v2)
c→c+λ(h1-h2) (2)
Wherein: λ indicates learning rate, and P indicates the probability for the aobvious layer activation that neuron is reconstructed in hidden layer, v and h represent aobvious layer and
The node of hidden layer, b and c are respectively the bias of aobvious layer and hidden layer.
4. the hot duct leakage detection method according to claim 1 based on the fusion of the depth confidence network information, special
Sign also resides in: in step S1, the Gauss- of the least square support vector classification use based on Gauss-Cos mixed kernel function
The following formula of Cos mixed kernel function is calculated:
Wherein, σ is the parameter of Gauss-Cos mixed kernel function.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910478572.3A CN110145695B (en) | 2019-06-03 | 2019-06-03 | Heating pipeline leakage detection method based on deep confidence network information fusion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910478572.3A CN110145695B (en) | 2019-06-03 | 2019-06-03 | Heating pipeline leakage detection method based on deep confidence network information fusion |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110145695A true CN110145695A (en) | 2019-08-20 |
CN110145695B CN110145695B (en) | 2021-01-15 |
Family
ID=67590303
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910478572.3A Expired - Fee Related CN110145695B (en) | 2019-06-03 | 2019-06-03 | Heating pipeline leakage detection method based on deep confidence network information fusion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110145695B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113551156A (en) * | 2021-06-23 | 2021-10-26 | 广州杰赛科技股份有限公司 | Pipeline state monitoring method and device based on deep learning and storage medium |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101196872A (en) * | 2007-11-19 | 2008-06-11 | 清华大学 | Leakage detecting and locating method based on pressure and sound wave information amalgamation |
CN103322415A (en) * | 2013-06-05 | 2013-09-25 | 哈尔滨工程大学 | Two-dimensional reproduction method for petroleum pipeline defects through least squares support vector machines (LS-SVM) |
CN105546352A (en) * | 2015-12-21 | 2016-05-04 | 重庆科技学院 | Natural gas pipeline tiny leakage detection method based on sound signals |
CN105654141A (en) * | 2016-01-06 | 2016-06-08 | 江苏大学 | Isomap and SVM algorithm-based overlooked herded pig individual recognition method |
US20170076563A1 (en) * | 2015-09-15 | 2017-03-16 | General Electric Company | Systems and methods to provide pipeline damage alerts |
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 |
EP3327206A1 (en) * | 2016-11-25 | 2018-05-30 | Tata Consultancy Services Limited | Ranking pipes for maintenance in pipe networks using approximate hydraulic metrics |
CN108730776A (en) * | 2018-06-22 | 2018-11-02 | 大连理工大学 | A kind of Oil Pipeline Leakage Detection based on the fusion of extreme learning machine information |
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 |
US20180347763A1 (en) * | 2017-05-31 | 2018-12-06 | Schneider Electric Software, Llc | Machine learning detection of pipeline ruptures |
-
2019
- 2019-06-03 CN CN201910478572.3A patent/CN110145695B/en not_active Expired - Fee Related
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101196872A (en) * | 2007-11-19 | 2008-06-11 | 清华大学 | Leakage detecting and locating method based on pressure and sound wave information amalgamation |
CN103322415A (en) * | 2013-06-05 | 2013-09-25 | 哈尔滨工程大学 | Two-dimensional reproduction method for petroleum pipeline defects through least squares support vector machines (LS-SVM) |
US20170076563A1 (en) * | 2015-09-15 | 2017-03-16 | General Electric Company | Systems and methods to provide pipeline damage alerts |
CN105546352A (en) * | 2015-12-21 | 2016-05-04 | 重庆科技学院 | Natural gas pipeline tiny leakage detection method based on sound signals |
CN105654141A (en) * | 2016-01-06 | 2016-06-08 | 江苏大学 | Isomap and SVM algorithm-based overlooked herded pig individual recognition method |
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 |
EP3327206A1 (en) * | 2016-11-25 | 2018-05-30 | Tata Consultancy Services Limited | Ranking pipes for maintenance in pipe networks using approximate hydraulic metrics |
US20180347763A1 (en) * | 2017-05-31 | 2018-12-06 | Schneider Electric Software, Llc | Machine learning detection of pipeline ruptures |
CN108730776A (en) * | 2018-06-22 | 2018-11-02 | 大连理工大学 | A kind of Oil Pipeline Leakage Detection based on the fusion of extreme learning machine information |
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 (1)
Title |
---|
陈俊: "基于深度置信网络和支持向量机的计量自动化终端故障诊断方法", 《广西电力》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113551156A (en) * | 2021-06-23 | 2021-10-26 | 广州杰赛科技股份有限公司 | Pipeline state monitoring method and device based on deep learning and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN110145695B (en) | 2021-01-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Cheng et al. | Data-driven modelling of nonlinear spatio-temporal fluid flows using a deep convolutional generative adversarial network | |
Hu et al. | When do extended physics-informed neural networks (XPINNs) improve generalization? | |
Xu et al. | Composite quantile regression neural network with applications | |
CN108664687A (en) | A kind of industrial control system space-time data prediction technique based on deep learning | |
CN113779675B (en) | Physical-data driven intelligent shear wall building structure design method and device | |
CN109492822A (en) | Air pollutant concentration time-space domain interaction prediction method | |
CN110348713A (en) | A kind of platform area line loss calculation method based on association analysis and data mining | |
CN109299262A (en) | A kind of text implication relation recognition methods for merging more granular informations | |
CN110473592B (en) | Multi-view human synthetic lethal gene prediction method | |
CN109991685A (en) | A kind of precipitation prediction technique and device based on more LSTM Model Fusions | |
CN105160400A (en) | L21 norm based method for improving convolutional neural network generalization capability | |
CN106485227A (en) | A kind of Evaluation of Customer Satisfaction Degree method that is expressed one's feelings based on video face | |
Rath et al. | A hybrid empirical-Bayesian artificial neural network model of salinity in the San Francisco Bay-Delta estuary | |
CN109711883A (en) | Internet advertising clicking rate predictor method based on U-Net network | |
CN104634265B (en) | A kind of mineral floating froth bed soft measurement method of thickness based on multiplex images Fusion Features | |
CN108879732A (en) | Transient stability evaluation in power system method and device | |
CN105843829A (en) | Big data credibility measurement method based on layering model | |
CN114245337A (en) | Water supply pipe network leakage positioning sensor arrangement method based on graph convolution network | |
CN110310698A (en) | Classification model construction method and system based on protein length and DCNN | |
Feng et al. | A characteristic-featured shock wave indicator for conservation laws based on training an artificial neuron | |
Fan et al. | An improved approach to generate generalized basic probability assignment based on fuzzy sets in the open world and its application in multi-source information fusion | |
CN110930514B (en) | Three-dimensional grid model self-adaptive simplification method based on local region characteristics | |
Robati et al. | Inflation rate modeling: adaptive neuro-fuzzy inference system approach and particle swarm optimization algorithm (ANFIS-PSO) | |
CN201716727U (en) | Geographical simulation system based on remote sensing and GIS | |
CN114912364A (en) | Natural gas well flow prediction method, device, equipment and computer readable medium |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210115 Termination date: 20210603 |
|
CF01 | Termination of patent right due to non-payment of annual fee |