CN112699927B - Pipeline fault diagnosis method and system - Google Patents
Pipeline fault diagnosis method and system Download PDFInfo
- Publication number
- CN112699927B CN112699927B CN202011558156.3A CN202011558156A CN112699927B CN 112699927 B CN112699927 B CN 112699927B CN 202011558156 A CN202011558156 A CN 202011558156A CN 112699927 B CN112699927 B CN 112699927B
- Authority
- CN
- China
- Prior art keywords
- fault
- pipeline
- node
- fault diagnosis
- bayesian network
- 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
Classifications
-
- 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
-
- 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
-
- 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
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
Abstract
The application discloses a pipeline fault diagnosis method, which comprises the steps of constructing a fault diagnosis training data set, forming a Bayesian network initial structure according to fault types and fault causes, and determining the best matched Bayesian network structure and parameters; learning the corresponding relation between the monitoring data and the fault category, and determining an SVM fault classification model; based on real-time monitoring data of each node in the pipeline, classifying faults according to the SVM fault classification model; and according to the classification result, carrying out reasoning by using a Bayesian network structure and parameters, determining the probability of each node occurrence fault, and taking the node with the maximum probability of occurrence fault and the type of occurrence fault as a fault diagnosis result. The application also provides a pipeline fault diagnosis system. The application solves the problem of insufficient historical fault data quantity of fault diagnosis.
Description
Technical Field
The application relates to the technical field of pipeline fault diagnosis, in particular to a pipeline fault diagnosis system and method based on machine learning and a Bayesian network.
Background
The influence factors of pipeline faults or failures are complex, the types of faults are numerous, and certain causal relation exists between the faults. In most cases, the fault diagnosis and localization cannot be directly performed through obvious fault phenomena. Monitoring pipeline operation indexes by using specific sensors is a basic method of fault diagnosis, the occurrence of faults can lead to the change of index parameters, different faults can lead to different change characteristics of the index parameters, but the specific type of sensors can only sense specific types of faults. However, in practical applications, fault diagnosis is often only possible with several conventional sensors. It is difficult to diagnose or even predict pipeline faults based on such incomplete information.
In recent years, the neural network method is successfully applied to fault diagnosis and fault prediction of a plurality of systems by virtue of strong self-learning and self-adaption capabilities. But the neural network is a black box system, is not interpreted, and requires a lot of data training to be effective. Many times, the amount of historical fault data available for learning by the fault diagnosis algorithm is insufficient, and thus the neural network method is less effective. Therefore, how to realize the visual fault diagnosis under the condition that the fault events are not too many is a technical problem to be solved.
Disclosure of Invention
The application provides a pipeline fault diagnosis method and a system, which solve the problem of how to realize fault diagnosis under the condition of less occurrence of fault events.
The embodiment of the application provides a pipeline fault diagnosis method, which comprises the following steps:
constructing a fault diagnosis training data set, wherein the first data set comprises fault categories and corresponding monitoring data, and the second data set comprises historical records of pipeline fault categories and fault causes;
forming a Bayesian network initial structure according to the fault category and the fault cause; determining the best matched Bayesian network structure and parameters based on the second data set;
based on the first data set, learning the corresponding relation between the monitoring data and the fault category, and determining a fault classification model of a support vector machine (Support Vector Machine, SVM for short);
based on real-time monitoring data of each node in the pipeline, classifying faults according to the SVM fault classification model; and according to the classification result, carrying out reasoning by using a Bayesian network structure and parameters, determining the probability of each node occurrence fault, and taking the node with the maximum probability of occurrence fault and the type of occurrence fault as a fault diagnosis result.
Preferably, the bayesian network structure is obtained by using a K2 algorithm of scoring search on the second data set.
Preferably, the parameter of the bayesian network is a conditional distribution value determined by using a maximum likelihood estimation method for the second data set, and the conditional distribution value represents a dependency relationship of faults between nodes.
Preferably, the SVM fault classification model classifies the fault using one or more bi-classifiers.
Further, the pipeline fault diagnosis method further comprises the following steps:
repairing the node with the highest fault occurrence probability according to the fault diagnosis result, and determining whether the node is true as a fault cause or not;
and adjusting node parameters according to whether the fault cause is true or not, and reasoning again by using the Bayesian network structure.
It is further preferred that the composition comprises,
if the cause of the fault is not true, setting the parameter of the node to be 100% non-occurrence;
if the fault cause is true and the previous level fault exists as the fault cause, setting the parameter of the node to be 100% occurrence;
if the fault cause is true and there is no previous level fault as the fault cause, the fault diagnosis ends.
The embodiment of the application also provides a pipeline fault diagnosis system, and the pipeline fault diagnosis method according to any one of the embodiments of the application comprises a sensor monitoring module, a fault classification module and a fault diagnosis module;
the sensor monitoring module comprises at least one of a temperature sensor, a pressure sensor and a noise sensor and is used for generating the monitoring data;
the fault classification module is used for collecting the monitoring data; the SVM fault classification model is also used for classifying faults based on real-time monitoring data of each node in the pipeline;
the fault diagnosis module is used for carrying out reasoning by using the Bayesian network structure and the parameters according to the classification result to determine the probability of the node fault.
Preferably, the pipeline fault diagnosis system comprises a diagnosis data acquisition module, which is used for constructing a fault diagnosis training data set and comprises the first data set and the second data set.
Further preferably, the fault classification module further comprises a first model self-learning module; the first model self-learning module is used for: and based on the first data set, learning the corresponding relation between the monitoring data and the fault category, and determining an SVM fault classification model.
Further preferably, the fault diagnosis module further comprises a second model self-learning module; the second model self-learning module is used for: based on the second dataset, a best matching bayesian network structure and parameters are determined.
The above at least one technical scheme adopted by the embodiment of the application can achieve the following beneficial effects:
the Bayesian network is an uncertainty knowledge expression and reasoning model based on probability analysis and graph theory, is an information representation framework combining causal knowledge and probability knowledge, and is very effective for reasoning problems based on incomplete information. The fault diagnosis problem of pipeline leakage is expressed as an uncertain decision problem through a Bayesian network, the uncertainty and the imperfection of information are considered, and a practical method is provided for the fault diagnosis of the pipeline.
According to the pipeline fault diagnosis system and method provided by the application, the faults are classified based on the machine learning method, the Bayesian network is utilized to carry out probabilistic reasoning on the fault reasons, and the intelligent diagnosis and the efficient investigation of the pipeline faults are carried out by auxiliary pipeline operation and maintenance personnel. The system has self-learning capability, can perform self-learning along with accumulation of diagnostic data, and realizes continuous optimization of the model.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:
FIG. 1 is a flow chart of a pipeline fault diagnosis method according to the present application;
FIG. 2 is a topological diagram of an initial structure of a Bayesian network for pipeline faults;
FIG. 3 is a schematic diagram of an embodiment of a pipeline fault diagnosis system.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be clearly and completely described below with reference to specific embodiments of the present application and corresponding drawings. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
According to the intelligent fault diagnosis method and system provided by the application, the pipeline fault diagnosis model is constructed based on SVM machine learning and Bayesian network, and the pipeline fault is intelligently diagnosed and efficiently checked by utilizing limited sensor monitoring data. The system has self-learning capability, and can continuously optimize the model along with continuous enrichment of the diagnostic data set.
The following describes in detail the technical solutions provided by the embodiments of the present application with reference to the accompanying drawings.
FIG. 1 is a flow chart of a pipeline fault diagnosis method according to the present application;
the embodiment of the application provides a pipeline fault diagnosis method, which comprises the following steps of 10-20:
step 10, performing self-learning according to the fault diagnosis training data set;
further, step 10 includes steps 101 to 103.
Step 101, constructing a fault diagnosis training data set, wherein a first data set comprises fault categories and corresponding monitoring data, and a second data set comprises historical records of pipeline fault categories and fault causes;
the fault categories include, but are not limited to, at least a portion of: pipeline leakage, perforation, fracture, steel pipe defect, corrosion thinning, stress corrosion, corrosion cracking, third party damage, design defect, misoperation, geological disaster, original defect, construction defect, soil corrosion, corrosion-resistant coating failure, cathodic protection failure, stress concentration, residual stress, severe line pressure pipe, illegal construction, design strength defect, design material defect, operator misoperation, maintainer misoperation, address sedimentation, earthquake and debris flow.
The fault cause refers to the reason that one fault category is another fault category. For example, the failure cause of a pipeline leak includes at least one of a break and a perforation; the failure cause of the perforation comprises at least one of defect and corrosion thinning of the steel pipe; the failure cause of the corrosion cracking includes at least one of corrosion thinning and stress corrosion; the failure causes of the fracture include at least one of steel pipe defects, corrosion cracking, third party damage, design defects, mishandling, geological disasters.
The fault diagnosis training data set can be, for example, an on-site investigation, a pipeline history fault record data acquisition and a fault diagnosis model training data set construction. For example, the monitoring data may be accumulated by a sensor.
The dataset comprises two parts: the first data set is a historical record of operation index monitoring data (temperature, pressure, noise and the like) of a plurality of types of faults of pipelines, and can be used for training a fault classification model by an SVM machine learning method; the second data set is a history of the cause of the pipeline fault and is used for bayesian network structure and parameter learning. When constructing the data set, the sufficiency of the sample and the completeness of the data should be ensured as much as possible.
102, forming a Bayesian network initial structure according to fault types and fault causes; determining the best matched Bayesian network structure and parameters based on the second data set;
102A, analyzing a common fault accident tree and creating a Bayesian network initial structure;
in order to simplify the Bayesian network structure and improve the modeling efficiency, statistical analysis and expert consultation of related faults are firstly carried out on a data set, main fault causes are reduced, direct causes and indirect causes of target events are analyzed, a pipeline common fault accident tree is constructed, the relation of all nodes of the Bayesian network is primarily judged, and the judgment result is used as the experience knowledge of Bayesian structure learning. The topology of the pipeline fault structure obtained by fault tree analysis is shown in fig. 2, and is taken as an initial structure of the bayesian network. The failure of a pipeline leak is mainly classified into perforation and fracture: the reasons for possible perforation are steel pipe defects and corrosion thinning, and the reasons for possible fracture are steel pipe defects, corrosion cracking, third party damage, unreasonable design, misoperation and geological disasters. Wherein, the corrosion of the corrosion reduction Bao Heying force can cause corrosion cracking, and the corrosion reduction can cause perforation and fracture. In fig. 2, the fault in the start block of each segment of the arrow is the one that is the upper level of the fault in the arrow pointing block. For example, from "malfunction" to "line leak", "break" is a cause of failure of "line leak", and "malfunction" is a cause of failure of "break".
Step 102B, bayesian network structure learning
For example, based on the constructed sample second data set, a K2 algorithm based on a score search in the FullBNT bayesian network toolbox is used in MATLAB software to find the bayesian network structure that best matches the sample set. The K2 algorithm needs to provide a node sequence in advance, and the parameter is set according to the initial structure of the bayesian network obtained by the accident tree analysis in step 102A.
The number of samples in the second data set is preferably greater than 50, and theoretically, the larger the number of samples, the more accurate the learned bayesian network structure. The data set is stored in a two-dimensional matrix form, each column represents one sample, each row represents the value of all nodes of the sample, the matrix size is n multiplied by m, n is the number of nodes, and m is the number of samples. In order to make the network structure compact enough, the possible value number of each node is set to be two types of yes and no, and the maximum parent number of each node is set to be 2.
The network architecture learned by the algorithm may still be relatively complex and may even present partially uncorrelated connections. To obtain a more concise and accurate network structure, the node correlation evaluation is carried out on the initial structure by combining expert knowledge, the relation among the nodes is adjusted, the uncorrelated node relation is deleted, and the optimized fault diagnosis Bayesian network structure is obtained.
Step 102C, bayesian network parameter learning
Based on the second data set and the bayesian network structure, for example, the maximum likelihood estimation method in the Full BNT bayesian network tool box can be used in MATLAB software to learn the conditional probability distribution (Conditional Probability Distribution, abbreviated as CPD) parameters in the bayesian network, and the dependency relationship between the node variables can be quantified. And after the network structure is obtained, the BNT tool box inputs a data set according to rules and carries out parameter adjustment, so that CPD parameters can be learned.
Step 103, based on the first data set, learning the corresponding relation between the monitoring data and the fault class, and determining an SVM fault classification model;
and training a fault classification model based on the SVM. For example, building an SVM fault classification model by using a libsvm toolbox in MATLAB software through a first data set, and learning the corresponding relation between sensor monitoring data and specific faults. The SVM has good learning ability and can solve the problems of small samples and nonlinear classification, so that the SVM becomes a first-choice classifier for processing the problems of fault diagnosis of a plurality of systems.
Preferably, the SVM fault classification model classifies the fault using one or more bi-classifiers. The SVM is a classifier at pattern recognition. In the patent, only two faults of perforation and fracture are considered to be classified, and the two faults belong to two classification problems, and an SVM and classifier are built. When faced with multiple classes of fault data, multiple SVM classifiers need to be built.
Step 20, diagnosing pipeline faults based on real-time monitoring data;
further, step 20 includes steps 201 to 202. Optionally, step 203 is also included.
Step 201, classifying faults according to the SVM fault classification model based on real-time monitoring data of each node in the pipeline;
based on monitoring data of sensors such as temperature, pressure and noise, performing fault preliminary classification by using the SVM fault classification model obtained through training in the step 10, judging that the fault belongs to perforation or cracking, and taking the classification result as input evidence of a Bayesian network inference engine.
Step 202, reasoning by using a Bayesian network structure and parameters according to the classification result, determining the probability of each node occurrence fault, and taking the node with the highest probability of occurrence fault and the type of occurrence fault as a fault diagnosis result.
In step 202, a bayesian network model is used for fault diagnosis, the SVM fault classification result is used as the input evidence of a bayesian network inference engine, probability reasoning is performed, namely the posterior probability of the relevant node can be updated, and all possible fault reasons and the occurrence probability of each cause are determined. And sequencing the posterior probability values of the fault reasons, and taking the reason corresponding to the node with the largest posterior probability value as a fault diagnosis result as the basis of fault overhaul.
It should be noted that the Full BNT tool box adopts an engine mechanism, and different inference engines complete model conversion, refinement and solution according to different inference algorithms. In the current application, the Bayesian network nodes are fewer, so that a joint tree reasoning algorithm in an accurate reasoning method is selected, the basic idea is that a Bayesian network model is converted into an undirected tree, then the exponential joint probability distribution is represented by a factor form of the local probability distribution, and finally the reasoning is performed by utilizing message propagation among the joint tree nodes.
And 203, carrying out fault maintenance according to the diagnosis result, and resetting the event fault probability according to the maintenance condition.
In step 203, first, repairing the node with the highest probability of occurrence of the fault according to the fault diagnosis result, and determining whether the fault cause of the node is true; and then, according to whether the fault cause is true, adjusting node parameters, and determining whether to use the Bayesian network structure again for reasoning.
If the cause of the fault is not true, setting the parameters of the node to be 100% non-occurrence, and reasoning again using the Bayesian network structure. At this time, the overhaul result indicates that the fault is not caused by the reason corresponding to the node, which indicates that the reason represented by the node does not occur, and the real reason of the fault needs to be explored. Therefore, the node is set to be 100% non-occurrence, the BN model is input as evidence to continue reasoning, and the factor with the maximum posterior probability value is searched again for maintenance.
If the fault cause is true and there is a previous level fault as the fault cause, setting the parameters of the node to be 100% occurrence, and reasoning again using the Bayesian network structure. At this time, the overhaul result shows that the fault is caused by the reason corresponding to the node, and the reason is the middle layer event of the accident tree, which indicates that the fault is caused by the bottom layer event, and the reason needs to be further explored. And setting the node to be 100% occurrence, inputting the node as evidence into a BN model for continuous reasoning, and finding out the node of the previous level fault with the maximum posterior probability value for overhauling again.
If the fault cause is true and there is no previous level fault as the fault cause, the fault diagnosis ends. At this time, the overhaul result shows that the fault is caused by the reason corresponding to the node, and the reason is located at the bottom event in the accident tree, so that the fault diagnosis is finished, and the reason for causing the fault is determined.
And repeating the related steps according to the overhaul result until the cause of the fault is overhauled.
FIG. 3 is a schematic diagram of an embodiment of a pipeline fault diagnosis system.
The embodiment of the application also provides a pipeline fault diagnosis system, and the pipeline fault diagnosis method according to any one of the embodiments of the application comprises a sensor monitoring module, a fault classification module and a fault diagnosis module;
the sensor monitoring module is used for installing conventional sensing equipment such as a temperature sensor, a pressure sensor, a noise sensor and the like at specified monitoring points of the pipeline. The sensor monitoring module comprises at least one of a temperature sensor, a pressure sensor and a noise sensor and is used for generating the monitoring data;
the fault classification module is used for collecting the monitoring data; the SVM fault classification model is also used for classifying faults based on real-time monitoring data of each node in the pipeline; the fault classification module can be used for carrying out preliminary classification on faults by using sensor monitoring data preferentially.
The fault diagnosis module is used for carrying out reasoning by using the Bayesian network structure and the parameters according to the classification result to determine the probability of the node fault. The fault diagnosis module performs fault diagnosis, overhauling and troubleshooting based on the Bayesian network until the cause of the fault is overhauled.
The fault diagnosis module comprises data of a Bayesian network structure model, data of CPD parameters and a Bayesian network inference engine. The inference engine refers to an operation module of the bayesian network diagnostic program, and the fault diagnosis result in step 202 of the present application is obtained through calculation.
Further, the fault diagnosis module records fault location overhaul data and data related to fault point resolution. After the fault location maintenance is completed, according to the scheme of step 203, the probability of occurrence of the corresponding fault event of the node is reset, and the bayesian network inference engine is driven again to update the fault diagnosis result.
Preferably, the pipeline fault diagnosis system comprises a diagnosis data acquisition module, which is used for constructing a fault diagnosis training data set and comprises the first data set and the second data set. The diagnosis data acquisition module is used for recording key data in the whole fault diagnosis and overhaul process, and providing the key data for the fault classification model self-learning module and the fault diagnosis model self-learning module to optimize a fault classification algorithm and update Bayesian network structure/parameters.
Further preferably, the fault classification module further comprises a first model self-learning module; the first model self-learning module is used for: and based on the first data set, learning the corresponding relation between the monitoring data and the fault category, and determining an SVM fault classification model. That is, the first model self-learning module performs model parameter optimization using the diagnostic overhaul key data collected by the diagnostic data collection module.
Further preferably, the fault diagnosis module further comprises a second model self-learning module; the second model self-learning module is used for: based on the second dataset, a best matching bayesian network structure and parameters are determined. That is, the second model self-learning module uses the diagnosis and overhaul key data collected by the diagnosis data collection module to update the network structure and CPD parameters.
It should also 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 one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are to be included in the scope of the claims of the present application.
Claims (8)
1. A method of diagnosing a pipeline fault, comprising the steps of:
constructing a fault diagnosis training data set, wherein the first data set comprises fault categories and corresponding monitoring data, and the second data set comprises historical records of pipeline fault categories and fault causes;
obtaining a pipeline fault structure topological graph as an initial structure of the Bayesian network according to the fault category and the fault cause; determining the best matched Bayesian network structure and parameters based on the second data set; the Bayesian network structure is obtained by using a K2 algorithm of scoring search on the second data set, and the parameters of the Bayesian network are conditional distribution values determined by using a maximum likelihood estimation method on the second data set, and represent the dependency relationship of faults between nodes;
based on the first data set, learning the corresponding relation between the monitoring data and the fault category, and determining an SVM fault classification model;
based on real-time monitoring data of each node in the pipeline, classifying faults according to the SVM fault classification model; and according to the classification result, carrying out reasoning by using a Bayesian network structure and parameters, determining the probability of each node occurrence fault, and taking the node with the maximum probability of occurrence fault and the type of occurrence fault as a fault diagnosis result.
2. The pipeline fault diagnosis method according to claim 1, further comprising the steps of:
repairing the node with the highest fault occurrence probability according to the fault diagnosis result, and determining whether the node is true as a fault cause or not;
and adjusting node parameters according to whether the fault cause is true or not, and reasoning again by using the Bayesian network structure.
3. The pipeline fault diagnosis method according to claim 1, wherein the SVM fault classification model classifies faults using one or more bi-classifiers.
4. A pipeline fault diagnosis method as claimed in claim 2, wherein,
if the cause of the fault is not true, setting the parameter of the node to be 100% non-occurrence;
if the fault cause is true and the previous level fault exists as the fault cause, setting the parameter of the node to be 100% occurrence;
if the fault cause is true and there is no previous level fault as the fault cause, the fault diagnosis ends.
5. A pipeline fault diagnosis system using the pipeline fault diagnosis method according to any one of claims 1 to 4, characterized by comprising a sensor monitoring module, a fault classification module, a fault diagnosis module;
the sensor monitoring module comprises at least one of a temperature sensor, a pressure sensor and a noise sensor and is used for generating the monitoring data;
the fault classification module is used for collecting the monitoring data; the SVM fault classification model is also used for classifying faults based on real-time monitoring data of each node in the pipeline;
the fault diagnosis module is used for carrying out reasoning by using the Bayesian network structure and the parameters according to the classification result to determine the probability of the node fault.
6. The pipeline fault diagnosis system of claim 5, comprising a diagnostic data acquisition module for constructing a fault diagnosis training data set comprising the first data set and the second data set.
7. The pipeline fault diagnosis system of claim 5, wherein the fault classification module further comprises a first model self-learning module;
the first model self-learning module is used for: and based on the first data set, learning the corresponding relation between the monitoring data and the fault category, and determining an SVM fault classification model.
8. The pipeline fault diagnosis system of claim 5, wherein the fault diagnosis module further comprises a second model self-learning module;
the second model self-learning module is used for: based on the second dataset, a best matching bayesian network structure and parameters are determined.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011558156.3A CN112699927B (en) | 2020-12-25 | 2020-12-25 | Pipeline fault diagnosis method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011558156.3A CN112699927B (en) | 2020-12-25 | 2020-12-25 | Pipeline fault diagnosis method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112699927A CN112699927A (en) | 2021-04-23 |
| CN112699927B true CN112699927B (en) | 2023-10-20 |
Family
ID=75510283
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011558156.3A Active CN112699927B (en) | 2020-12-25 | 2020-12-25 | Pipeline fault diagnosis method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112699927B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102867105A (en) * | 2012-06-29 | 2013-01-09 | 常州大学 | Urban natural gas pipeline failure multi-factor multi-mode probability analysis method and application thereof |
| CN105468917A (en) * | 2015-12-01 | 2016-04-06 | 北京无线电计量测试研究所 | Pipeline fault prediction method and apparatus |
| CN109784724A (en) * | 2019-01-15 | 2019-05-21 | 常州大学 | A kind of subsea production system equipment fault diagnosis method based on Bayesian network |
| CN109858126A (en) * | 2019-01-23 | 2019-06-07 | 北京市燃气集团有限责任公司 | Pipeline network of fuel gas in city safety monitoring method for early warning and system based on settlement monitoring |
| CN110727553A (en) * | 2019-10-15 | 2020-01-24 | 上海交通大学 | Method and device for predicting and diagnosing faults of processor system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8457897B2 (en) * | 2007-12-07 | 2013-06-04 | Exxonmobil Upstream Research Company | Methods and systems to estimate wellbore events |
-
2020
- 2020-12-25 CN CN202011558156.3A patent/CN112699927B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102867105A (en) * | 2012-06-29 | 2013-01-09 | 常州大学 | Urban natural gas pipeline failure multi-factor multi-mode probability analysis method and application thereof |
| CN105468917A (en) * | 2015-12-01 | 2016-04-06 | 北京无线电计量测试研究所 | Pipeline fault prediction method and apparatus |
| CN109784724A (en) * | 2019-01-15 | 2019-05-21 | 常州大学 | A kind of subsea production system equipment fault diagnosis method based on Bayesian network |
| CN109858126A (en) * | 2019-01-23 | 2019-06-07 | 北京市燃气集团有限责任公司 | Pipeline network of fuel gas in city safety monitoring method for early warning and system based on settlement monitoring |
| CN110727553A (en) * | 2019-10-15 | 2020-01-24 | 上海交通大学 | Method and device for predicting and diagnosing faults of processor system |
Non-Patent Citations (1)
| Title |
|---|
| "基于贝叶斯-事故树的石化企业罐区泄漏风险研究";李威臻等;《南开大学学报》;第53卷(第1期);1-5 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112699927A (en) | 2021-04-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Nasiri et al. | Fracture mechanics and mechanical fault detection by artificial intelligence methods: A review | |
| Diez-Olivan et al. | Data fusion and machine learning for industrial prognosis: Trends and perspectives towards Industry 4.0 | |
| He et al. | A fuzzy TOPSIS and rough set based approach for mechanism analysis of product infant failure | |
| Vrignat et al. | Sustainable manufacturing, maintenance policies, prognostics and health management: A literature review | |
| CN102765643B (en) | Elevator fault diagnosis and early-warning method based on data drive | |
| CN101950382B (en) | Method for optimal maintenance decision-making of hydraulic equipment with risk control | |
| CN109800127A (en) | A kind of system fault diagnosis intelligence O&M method and system based on machine learning | |
| Zhao et al. | A fuzzy system for concrete bridge damage diagnosis | |
| CN107273924B (en) | Multi-data fusion power plant fault diagnosis method based on fuzzy clustering analysis | |
| CN111412579B (en) | Air conditioning unit fault type diagnosis method and system based on big data | |
| CN103678881B (en) | Composite fault diagnosis method based on combination of artificial immunity and evidence theory | |
| CN110705812A (en) | Industrial fault analysis system based on fuzzy neural network | |
| CN110555477A (en) | municipal facility fault prediction method and device | |
| CN117014472A (en) | Cloud side end cooperation-based intelligent power plant equipment inspection method and system | |
| CN113339204A (en) | Wind driven generator fault identification method based on hybrid neural network | |
| CN114444582A (en) | Mechanical equipment fault diagnosis method based on convolutional neural network and Bayesian network | |
| Lewis et al. | Metrics for evaluating the performance of complex engineering system health monitoring models | |
| CN113984389B (en) | Rolling bearing fault diagnosis method based on multiple receptive fields and improved capsule graph neural network | |
| CN112699927B (en) | Pipeline fault diagnosis method and system | |
| KR20220089853A (en) | Method for Failure prediction and prognostics and health management of renewable energy generation facilities using machine learning technology | |
| CN116484219A (en) | Water supply network water quality abnormal pollution source identification method based on gate control graph neural network | |
| CN115600695A (en) | Fault diagnosis method of metering equipment | |
| CN115310693A (en) | Comprehensive energy data cleaning method, system, equipment and medium | |
| Smirnov | Fuzzy model of fault finding in systems of responsible appointment | |
| CN114265891A (en) | Intelligent workshop system and method based on multi-source data fusion and storage 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 |