CN115392797B - Operating tunnel structure disease rapid intelligent diagnosis method based on Bayesian network - Google Patents
Operating tunnel structure disease rapid intelligent diagnosis method based on Bayesian network Download PDFInfo
- Publication number
- CN115392797B CN115392797B CN202211326292.9A CN202211326292A CN115392797B CN 115392797 B CN115392797 B CN 115392797B CN 202211326292 A CN202211326292 A CN 202211326292A CN 115392797 B CN115392797 B CN 115392797B
- Authority
- CN
- China
- Prior art keywords
- disease
- node
- bayesian network
- cause
- tunnel structure
- 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
- 201000010099 disease Diseases 0.000 title claims abstract description 307
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 title claims abstract description 307
- 238000003745 diagnosis Methods 0.000 title claims abstract description 69
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000011161 development Methods 0.000 claims abstract description 27
- 238000002405 diagnostic procedure Methods 0.000 claims description 12
- 230000006872 improvement Effects 0.000 claims description 10
- 230000001364 causal effect Effects 0.000 claims description 8
- 238000012795 verification Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 6
- 238000012544 monitoring process Methods 0.000 claims description 6
- 230000007547 defect Effects 0.000 claims description 4
- 238000011158 quantitative evaluation Methods 0.000 claims description 3
- 230000009471 action Effects 0.000 claims description 2
- 206010063385 Intellectualisation Diseases 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 7
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 7
- 230000006866 deterioration Effects 0.000 description 3
- 239000007788 liquid Substances 0.000 description 3
- 230000007847 structural defect Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 208000035474 group of disease Diseases 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000005297 material degradation process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000007789 sealing Methods 0.000 description 1
- 208000037921 secondary disease Diseases 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Educational Administration (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention relates to a Bayesian network-based rapid intelligent diagnosis method for operating tunnel structural diseases, which comprises the following steps: s10: establishing a tunnel structure disease diagnosis Bayesian network; s20: carry out tunnel structure disease diagnosis based on tunnel structure disease diagnosis Bayesian network, include: s201: inputting basic characteristic data of tunnel structure diseases; s202: inferring disease cause according to a bayesian network; s203: verifying whether the inferred disease cause occurs; s204: predicting the development trend of tunnel structure diseases under the condition of not taking measures; s205: and recommending treatment measures according to the Bayesian network. The Bayesian network reasoning method is introduced into the disease diagnosis of the urban rail transit tunnel structure, the cause of the tunnel structure disease can be deduced through simple condition input, and a recommended treatment scheme is given, so that the automation and the intellectualization of the disease diagnosis are realized, the speed and the efficiency of the disease diagnosis are improved, and the diagnosis result is more practical.
Description
Technical Field
The invention relates to the technical field of underground engineering, in particular to disease diagnosis of underground engineering, and specifically relates to a method for rapidly and intelligently diagnosing operating tunnel structure diseases based on a Bayesian network.
Background
With the development of urban rail transit in China, more and more rail transit lines are put into operation. The service life of the rail transit design is usually one hundred years, the line operation frequency is high, and the requirement on the line operation reliability is high. But is restricted by various factors, and the tunnel structure may have different degrees and different types of diseases in the tunnel operation process, including uneven settlement, transverse convergence deformation, cracks, water leakage, shield segment damage, material degradation and the like. For example, since the beginning of operation, there is a large range of uneven settlement in a plurality of sections in Shanghai subway No. one and No. two.
The tunnel structure diseases directly threaten the operation safety of the rail transit line. In order to ensure the operation safety of the rail transit line, the current measures are to monitor and detect regularly, find the diseases in time and manage the diseases. The premise of treatment is to make accurate diagnosis of the disease and clearly judge the cause and development trend of the disease.
For the diagnosis of the tunnel structure diseases, the current situation is mainly judged in a manual mode. The method has the main defects of low diagnosis speed, often requiring several days or even one week from the discovery of problems, the selection of experts, the site survey, the conference discussion to the formation of opinions, long diagnosis period, high cost, limitation of personal experience, difficulty in guaranteeing the comprehensiveness and consistency of diagnosis by the expert diagnosis and limited accumulation of diagnosis experience of subsequent engineering.
In addition, the Bayesian Network (BN) can use the field data to establish the BN network structure and learning parameters through corresponding algorithms. The inventor researches and discovers that although the Bayesian Network (BN) is introduced in the prior art, the bayesian network has a single consideration on node types, does not consider causal association among different diseases, is not practical, does not consider a targeted development trend and a control measure, only can evaluate the current health situation of the tunnel structure, cannot predict a development trend, and cannot directly provide a disease control measure, and the like.
In a word, the high-density operation of the rail transit line provides high requirements for rapid diagnosis and treatment of tunnel structure diseases. The time window is very short, typically only a few hours. The existing rail transit diagnosis and treatment technology is slow in time, and accurate diagnosis can be made only after several days. At present and in future, frequent operation of urban rail transit lines urgently needs a technology capable of quickly and accurately solving the problem of diagnosis and treatment of tunnel diseases.
Disclosure of Invention
In order to solve the problems, the invention provides a Bayesian network-based rapid intelligent diagnosis method for operating tunnel structure diseases, which introduces a Bayesian network reasoning method into urban rail transit tunnel structure disease diagnosis, can deduce the cause of tunnel structure diseases through simple condition input and give a recommended treatment scheme, realizes automation and intellectualization of disease diagnosis, improves the speed and efficiency of disease diagnosis, and ensures that the diagnosis result is more practical.
The invention is realized by the following steps:
the invention first provides a Bayesian network-based rapid intelligent diagnosis method for operating tunnel structure diseases, which comprises the following steps: s10: establishing a tunnel structure disease diagnosis Bayesian network, comprising: s101: determining a tunnel structure disease diagnosis Bayesian network structure; s102: setting the type and value range of the state variable of the Bayesian network node; s103: determining the probability distribution of state variables of the Bayesian network nodes; s20: the method for diagnosing the tunnel structure diseases based on the Bayesian network comprises the following steps: s201: inputting basic characteristic data of tunnel structure diseases; s202: inferring disease cause according to a bayesian network; s203: verifying whether the inferred disease cause occurs; if the verification result is 'no', modifying the value of the disease cause node state variable, returning to S202 to re-infer the disease cause until the inferred disease cause happens right after verification; s204: if the verification result is yes, predicting the development trend of the tunnel structure diseases under the condition of not taking measures; s205: and recommending treatment measures based on the tunnel structure disease development trend and by the Bayesian network. The Bayesian network structure comprehensively considers 'disease-cause-measure-trend' nodes and causal association among different diseases, can quickly deduce the cause of the diseases of the operating tunnel structure, can directly predict the development trend of the diseases, gives a recommended treatment scheme, and improves the speed and efficiency of disease diagnosis.
In some embodiments, the S101 determining the tunnel structure disease diagnosis bayesian network structure includes: s1011: determining a disease type set considered in a tunnel structure disease diagnosis Bayesian network according to actual requirements of tunnel structure disease diagnosis; s1012: aiming at a single disease type, determining a disease cause set and a measure set according to expert knowledge; s1013: aiming at a single disease, dividing nodes related to the disease into four types of causes, diseases, trends and measures, representing a Bayesian network structure by adopting a directed acyclic graph formed by disease-cause-measure-trend, and generating corresponding nodes according to a determined disease cause set and a measure set; s1014: establishing association between the generated nodes; s1015: repeating S1012-S1014, traversing all the disease types in the disease type set, and establishing an independent disease Bayesian network for each disease type; s1016: and according to expert knowledge, considering the internal association among different types of diseases, and determining the causal association among the Bayesian networks of the diseases in the disease type set.
In some embodiments, said S1014 establishing an association between the generated nodes comprises: (1) The cause type node and the measure type node are used as root nodes, the trend type node is a leaf node, the cause type node takes the disease type node and the trend type node as child nodes, the disease type node takes the trend type node as child nodes, and the measure type node takes the trend type node as child nodes; (2) Each disease type node has only one trend type node as a child node; (3) Each trend node has one and only one disease node as a father node; (4) Each disease node has at least one cause node as a father node; (5) All cause father nodes of all disease nodes are father nodes of trend child nodes of all disease nodes at the same time; (6) Each trend class node has at least one measure class node as a parent node.
In some embodiments, the S1016 determining the intrinsic association between the different types of diseases includes: (1) Establishing association between two corresponding disease nodes, taking the disease node representing the influence factor as a father node, taking the disease node representing the influenced factor as a child node, and considering the value of the state variable of the father node in the probability distribution of the state variable of the child node; (2) Multiple diseases are caused by the same cause, and nodes representing the same disease cause are combined; (3) And treating various diseases by the same measure, and combining nodes representing the same treatment measure.
In some embodiments, the S102 setting the bayesian network node state variable type and value range comprises: (1) For disease nodes, setting a state variable value range as { "Y: appearance "," N: no "}; if the number of the existing disease cases is large and the data is complete, grading according to the prior art standard; (2) For the cause type node, setting a state variable value field as { "present", "absent" }; if the number of the existing disease cases is large and the data is complete, expanding the value range to be { "S: severe "," E: general "," L: lighter "and" N: there is no "}; (3) For the trend class node, the state variable value domain is set to { "B: improvement "," W; a deterioration "}; if the number of the existing disease cases is large and the data is complete, expanding the value range of the existing disease cases into { "B: improvement "," W; worsened "," N; no apparent change "}; (4) For the measure class node, the state variable value domain is set to { "Y: taking an index of "" N: no "} is taken.
In some embodiments, the S201 inputting the basic feature data of the tunnel structure defect includes: diagnosing a node state variable value domain of the Bayesian network according to the tunnel structure disease, judging a node variable value according to an actual measurement condition, and inputting basic characteristic data of the tunnel structure disease; and/or the step S202 of deducing the disease cause according to the Bayesian network comprises the following steps: calculating posterior probability of all cause parent nodes of the diseases and deducing the cause of the diseases; and/or, the step S203 of verifying whether the inferred disease cause occurs includes: and (4) verifying whether the inferred disease cause occurs or not by using a monitoring or detecting means on site.
In some embodiments, the S204 predicting the tunnel structure disease development trend includes: s2041: retrieving a trend class child node set S (N _ QS) of the target disease class node; s2042: searching a disease sub-node set S (N _ BH) of a target disease node, and further searching a trend sub-node set S (M _ QS) of the target disease node; s2043: and calculating the state variable probability distribution of the trend type sub-node sets S (N _ QS) and S (M _ QS) to be used as the quantitative evaluation of the tunnel structure disease development trend under the condition that no measures are taken currently.
In some embodiments, the S205 recommending an abatement measure according to the bayesian network includes: s2051: searching a trend type sub-node set S (N _ QS) of a target disease type node according to the tunnel structure disease diagnosis Bayesian network; s2052: retrieving all action class parent set S (N _ CS) of trend class child set S (N _ QS) i ) (ii) a S2053: setting a state variable of a certain measure parent node as 'taken', observing the state variable probability distribution change of the trend child node N _ QS, and taking the measure corresponding to the corresponding measure parent node as a recommended measure when the probability of the state variable 'improvement' of the trend child node N _ QS is obviously increased; s2054: repeating S2053, traversing the measure class father node set S (N _ CS) i ) All measures in (1) are like parent nodes.
The second aspect of the present invention further provides a device for rapidly and intelligently diagnosing a structural defect of an operating tunnel, comprising: at least one memory for storing at least one program; at least one processor; when the at least one program is executed by the at least one processor, the at least one processor is caused to carry out the diagnostic method according to the first aspect of the invention. The computer software is used for realizing the work which needs experts to complete originally through the computer, the cause of the tunnel structure diseases can be rapidly deduced through simple condition input, and the dependence of tunnel structure disease diagnosis on expert resources is reduced.
The third aspect of the present invention further provides a computer readable storage medium in which a program executable by a processor is stored, the program executable by the processor being adapted to perform the diagnostic method according to the first aspect of the present invention when executed by the processor.
Compared with the prior art, the invention has the beneficial effects that:
(1) The Bayesian network structure for diagnosing the tunnel structure diseases comprehensively considers the 'disease-cause-measure-trend' node, comprehensively considers the relevant factors of the tunnel structure diseases, and is more practical.
(2) The causal association among different diseases is considered, and the development rules and treatment measures of different diseases can be considered uniformly.
(3) Disease development trend and treatment measures are brought into the Bayesian network structure, disease development trend can be directly predicted, and treatment measures can be automatically recommended.
(4) Compared with the traditional method of judging treatment measures by depending on experts, the diagnosis efficiency is improved, and the diagnosis time of diseases is calculated by days, usually about 7 days, by using the traditional diagnosis method. By using the method, the disease diagnosis time of each time including the recommendation of treatment measures can be reduced to be within 1 hour, and the disease diagnosis speed is improved by more than 10 times.
(5) The dependence on expert resources is reduced, and the cost is reduced.
It should be understood that the implementation of any embodiment of the present invention is not intended to achieve or achieve many or all of the above-described benefits simultaneously.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It should be apparent that the drawings in the following description are merely exemplary, and that other embodiments can be derived from the drawings provided by those of ordinary skill in the art without inventive effort.
The structures, ratios, sizes, and the like shown in the present specification are only used for matching with the contents disclosed in the specification, so that those skilled in the art will understand and read the present invention, and do not limit the conditions for implementing the present invention, so that the present invention has no technical essence, and any modifications of the structures, changes of the ratio relationships, or adjustments of the sizes, should still fall within the scope covered by the technical contents disclosed in the present invention without affecting the efficacy and the achievable purpose of the present invention.
FIG. 1 is a conceptual network diagram of disease-cause-measure-trend according to a preferred embodiment;
FIG. 2 is an exemplary schematic diagram (in part) of a tunnel fault diagnosis Bayesian network in accordance with a preferred embodiment;
FIG. 3 is a flow chart of a preferred embodiment of a Bayesian network for diagnosing tunnel structure diseases.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the present invention is described in further detail below with reference to the embodiments and the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
In the description of the present invention, the terms "comprises/comprising," consists of, \8230; "consists of, \8230;" or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a product, apparatus, process, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such product, apparatus, process, or method as desired. Without further limitation, an element defined by the phrases "comprising/including" \8230; "comprising 8230;" \8230; and \8230; "comprises;" does not exclude the presence of additional like elements in a product, device, process, or method comprising the element.
It is to be understood that, unless otherwise expressly specified or limited, the terms "disposed," "mounted," "connected," "secured," and the like are intended to be inclusive and mean, for example, that any suitable arrangement may be utilized and that any suitable connection, whether permanent or removable, or integral, may be utilized; can be mechanically or electrically connected; they may be directly connected or indirectly connected through intervening media, or may be connected through the use of two elements or the interaction of two elements. The specific meanings of the above terms in the present invention can be understood according to specific situations by those of ordinary skill in the art.
It will be further understood that the terms "upper," "lower," "front," "rear," "left," "right," "top," "bottom," "inner," "outer," "center," and the like are used herein to indicate an orientation or positional relationship based on that shown in the drawings for ease of description and simplicity of description, but do not indicate or imply that the device, component, or structure so referred to must have a particular orientation, be constructed or operated in a particular orientation, and should not be construed as limiting the invention.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the present invention, "a plurality" means two or more unless specifically defined otherwise.
In the description of the embodiments of the present invention, the related terms are understood as follows:
operating a tunnel: a tunnel that has been built and put into operation;
tunnel structure disease: phenomena affecting the safety and durability of the tunnel structure, including cracks, water leakage, peeling, looseness, bulging, corrosion, deformation, and the like;
and (3) diagnosing tunnel structure diseases: analyzing the type and cause of the tunnel structure diseases, grading the tunnel structure disease degree and giving disease treatment measure suggestions;
bayesian Network (BN): the belief network (belief network) is an extension of the Bayes method, is one of the most effective theoretical models in the field of uncertain knowledge expression and reasoning at present, is proposed by Pearl in 1988 and can be represented by a Directed Acyclic Graph (DAG), the DAG is composed of nodes representing variables and Directed edges connecting the nodes, the nodes represent random variables, the Directed edges among the nodes represent the interdependency systems among the nodes (the father node points to the child nodes), the conditional probability is used for expressing the relation strength, and the prior probability without the father node is used for expressing information.
The Bayesian network reasoning method is introduced into the disease diagnosis of the urban rail transit operation tunnel structure, the core is the Bayesian network for diagnosis and reasoning, the network is constructed on the basis of expert knowledge according to a large amount of actual case statistical data, the work which needs an expert to complete originally is realized through a computer, the cause of the tunnel structure disease can be deduced through simple condition input, meanwhile, the disease development trend can be directly predicted, and a recommended treatment scheme is given, so that the dependence of the tunnel structure disease diagnosis on expert resources is reduced, and the speed and the efficiency of the disease diagnosis are improved.
The following describes the implementation of the present invention in detail with reference to preferred embodiments.
The first aspect of the invention firstly provides a method for rapidly and intelligently diagnosing the structural defects of the operating tunnel based on the Bayesian network, and the technical scheme comprises two main technical links: generating: establishing an operation tunnel structure disease diagnosis Bayesian network; (II) use: and (5) carrying out operation tunnel structure disease diagnosis by using the Bayesian network.
And (I) a generation link.
A generation link of a tunnel structure disease diagnosis Bayesian network is that the tunnel structure disease diagnosis Bayesian network is established by combining expert knowledge assistance on the basis of inspecting the characteristics of the tunnel structure diseases of urban rail transit. This link mainly includes:
s101: determining a tunnel structure disease diagnosis Bayesian network structure;
on the basis of inspecting the characteristics of the tunnel structure diseases, the method combines expert knowledge to assist, establishes a tunnel structure disease diagnosis Bayesian network structure, and determines nodes forming the tunnel structure disease diagnosis Bayesian network and the incidence relation among the nodes. The Bayesian network adopts a directed acyclic graph representation, wherein nodes are elements relevant to diagnosis and treatment of tunnel structure diseases, and edges in the network are associations between the elements.
In some embodiments, the method specifically comprises the following steps:
s1011: determining a disease type set considered in a tunnel structure disease diagnosis Bayesian network according to actual requirements of tunnel structure disease diagnosis and treatment;
for example, in the operation process of the shield tunnel in the soft soil area, various diseases such as water leakage, segment staggering, uneven settlement, convergence deformation and the like occur, the diseases need to be considered by the bayesian network, and the disease types are { "water leakage", "segment staggering", "uneven settlement", "convergence deformation" }.
S1012: aiming at a single disease type, determining a disease cause set and a measure set according to expert knowledge;
for example, the causes of water leakage are { "seam open", "water stop failure", "structural cracking" } and the measures are { "grouting water stop", "embedding and sealing", "rapid plugging" }.
S1013: aiming at a single disease, dividing nodes related to the disease into four types of causes, diseases, trends and measures, and representing a Bayesian network structure by adopting a directed acyclic graph formed by disease-cause-measure-trend, as shown in FIG. 1, and generating nodes of corresponding types according to the disease cause set and the elements in the measure set determined in the step S1012;
aiming at a single disease, the Bayesian network has only one disease class node and one trend class node, and the number of the cause class node and the measure class node is not less than 1. The bayesian network shown in fig. 1 is the simplest bayesian network, and includes four nodes, namely "cause a", "disease B", "trend C", and "measure D". Under actual conditions, the same disease can be caused by different causes, and different measures can be adopted for treatment. For such cases, the invention sets cause nodes and measure nodes corresponding to the elements of the two kinds of sets one by one in the bayesian network according to the disease cause set and the measure set determined in step S1012.
For example, for the disease type "uneven settlement" of the tunnel structure, the cause set is { "interval stratum settlement", "sandy soil foundation vibration liquefaction", "ground stacking" }, so the cause A1 is set in the disease type "interval tunnel settlement" bayesian network: interval formation subsidence "," cause A2: sand foundation vibration liquefaction and cause A3: and 3 causal nodes are stacked on the ground.
In addition, the measure set of the disease type 'interval tunnel settlement' is { 'intra-tunnel micro-disturbance double-liquid grouting', 'ground grouting' }, so that the measure D1 is set in the disease type 'interval tunnel settlement' Bayesian network: in-tunnel micro-disturbance double-liquid grouting 'and' measure D2: ground grouting "2 measure nodes.
The method can be directly used for predicting the development trend of the diseases and recommending the treatment measures by directly incorporating the development trend of the diseases and the treatment measures into the Bayesian network.
S1014: establishing an association between the nodes generated in step S1013;
the cause type node and the measure type node are used as root nodes, the trend type node is a leaf node, the cause type node takes the disease type node and the trend type node as child nodes, the disease type node takes the trend type node as child nodes, and the measure type node takes the trend type node as child nodes.
The foregoing bayesian network obeys the following criteria:
(1) Each disease node has one and only one trend node as a child node;
(2) Each trend node has one and only one disease node as a father node;
(3) Each disease node has at least one cause node as a father node;
(4) All cause father nodes of all disease nodes are father nodes of trend child nodes of all disease nodes at the same time;
(5) Each trend class node has at least one measure class node as a parent node.
S1015: and repeating S1012-S1014, traversing all the disease types in the disease type set, and establishing an independent disease Bayesian network for each disease type.
S1016: and determining the causal association between the Bayesian networks of the diseases in the disease type set according to expert knowledge. And (3) considering the internal association among different types of diseases, and associating a plurality of single disease Bayesian networks. By incorporating the internal correlation between diseases into the Bayesian network, the cause and treatment measures of the diseases can be more reasonably found.
Three types of associations among the Bayesian networks of the independent diseases need to be considered:
(1) The method comprises the steps that causal relationships exist between different types of diseases, association is established between two corresponding disease nodes, the disease nodes representing influence factors can be used as father nodes according to influence rule association between different disease types, the disease nodes representing influenced factors can be used as child nodes, and the probability distribution of state variables of the child nodes considers the state variable values of the father nodes;
for example, uneven settlement of the tunnel may cause dislocation between shield segments, so the "disease B1-uneven settlement" serves as a parent node of the "disease B2-segment dislocation".
(2) Multiple diseases are caused by the same cause, and nodes representing the same disease cause are combined;
for example, since ground stacking may cause uneven settlement or convergence and deformation of a tunnel, the cause A2 of the parent node "cause B1-uneven settlement" and the cause An-ground stacking "of the parent node" cause B3-convergence and deformation "are combined.
(3) Treating various diseases by the same measure, and combining nodes representing the same treatment measure;
for example, the micro-disturbance double-liquid grouting in the tunnel can treat both uneven settlement and convergence deformation of the tunnel, so that the measure D2-ground grouting of the parent node of the trend D1-uneven settlement and the measure Dn-ground grouting of the parent node of the trend D3-convergence deformation are combined.
According to the steps, a Bayesian network structure for tunnel diseases is formed as shown in FIG. 2.
S102: setting the type and value range of the state variable of the Bayesian network node;
and setting the type and value range of the state variable of each node according to expert knowledge. For simplicity, the state variables of each node are set to be discrete, and the value range is set mainly according to the diagnosis requirement and the integrity of the existing disease case data.
In some embodiments, specifically:
(1) For the disease node, the state variable value range is set to be { "Y: appearance "," N: no "}; when the number of the existing disease cases is large and the data is complete, grading can be carried out by referring to the prior art standard.
(2) For the trend class node, the state variable value domain is set to be { "B: improvement "," W; a deterioration "}; when the number of the existing disease cases is large and the data is complete, the value range can be expanded to be { "B: improvement "," W; worsened "," N; no significant change "}.
(3) For the cause type node, the state variable value domain is set to be { "present", "absent" }; when the number of the existing disease cases is large and the data is complete, the value range can be expanded to be { "S: severe "," E: general "," L: lighter "and" N: there is no "}.
(4) For the measure class node, the state variable value field is set to { "Y: take "," N: no "} is taken.
S103: determining the probability distribution of state variables of the Bayesian network nodes;
the probability distribution of the node state variable has two determining modes:
the method I comprises the following steps: determined according to expert knowledge. Compiling a state variable probability distribution table to be set into a questionnaire and sending the questionnaire to a field expert, carrying out statistical analysis according to the result fed back by the expert and determining the state variable probability distribution value of each node. Wherein, the root node probability distribution questionnaire needs to fill in the prior probability of the root node, as shown in table 1; the child node probability distribution table is filled with the conditional probability when the parent node state variable value is determined, as shown in table 2.
TABLE 1 root node prior probability distribution settings (examples in part)
Table 2 child node conditional probability distribution settings (examples in part)
The second method comprises the following steps: and according to the collected tunnel structure disease control cases, counting the occurrence frequency of each node and the value frequency of different state variables, and calculating the probability distribution of the state variables.
(1) And (5) counting the prior probability of the root node.
For example, the total number N of the tunnel structure defect cases is collected 0 Wherein the number of disease cases with 'ground stacking' is m 0 And the rest is not subjected to ground stacking. According to the statistical result, the probability of the state variable value Y of the cause node cause A3-ground stacking is set as m 0 /N 0 The state variable takes the value "N"has a probability of 1-m 0 /N。
(2) And (5) counting the conditional probability of the child nodes.
Firstly, screening a disease case associated with a current node from a disease case library;
secondly, the screened disease cases are subjected to value combination s (x) according to all father node state variables 1 ，x 2 ，……,x n ) Grouping is carried out, and the number N(s) of disease cases in each group is counted. Wherein x is 1 ，x 2 ，……,x n A parent node state variable of the current node;
thirdly, for each group of disease cases, according to the current node state variable y i Dividing values, and respectively counting the occurrence times of each state variable and recording as m (s, y) i ）；
Finally, in m (s, y) i ) Dividing by the number N(s) of the disease cases in the group to obtain a state variable combination s (x) at the father node 1 ，x 2 ，……,x n ) When the current node state variable takes on the value y i Conditional probability p (y) of i |x 1 ，x 2 ，……,x n ）。
And (II) using links.
The technical process of the using link of the Bayesian network for diagnosing the tunnel structure diseases is shown in FIG. 3. The link comprises the following steps:
s201: inputting basic characteristic data of tunnel structure diseases;
and diagnosing a Bayesian network node variable value domain according to the tunnel structure disease, judging the node variable value according to the actual measurement condition, and inputting basic characteristic data of the tunnel structure disease. For example, for the bayesian network shown in fig. 3, monitoring shows that the curvature radius is smaller than 15000m due to the tunnel settlement in a certain section, and the state variable of "disease B1-uneven settlement" is determined to take the value of "Y".
S202: inferring disease cause according to a bayesian network;
and calculating the posterior probability of all cause parent nodes of a certain disease, and deducing the cause of the disease. For example, if the tunnel in a certain section has uneven settlement, the state variable value of "disease B1-uneven settlement" is set to Y, the probability that the state variable value of "cause A1-regional stratum settlement" of the cause type parent node is calculated to be Y is greatly increased, and it is determined that "regional stratum settlement" is a possible cause of the disease.
It should be noted that the posterior probability can be regarded as a conditional probability, that is, a probability given a value of a state variable of a certain related factor. Here, the posterior probability of the parent node "cause a" refers to the conditional probability of occurrence of the cause a when the state variable of the "disease B" takes a value of Y.
S203: verifying whether the inferred disease cause occurs;
and (5) verifying whether the inferred disease cause occurs by using a monitoring or detecting means on site. For example, when a tunnel in a certain section sinks, it is inferred from step S202 that regional stratum sinking is a possible cause of a disease, and by examining the same-region ground subsidence monitoring data in the same time period, it can be confirmed that the state variable of the cause of regional stratum sinking is Y (when the regional ground subsidence time-course curve is approximately synchronized with the regional tunnel subsidence curve) or N (when the regional ground subsidence time-course curve is not synchronized with the regional tunnel subsidence curve, the correlation is not strong).
If the verification result is 'no', modifying and setting the verified value of the disease cause node state variable, and returning to the step S202 to re-infer the disease cause. For example, when a tunnel in a certain section sinks, the probability of stratum sinking of a disease region is estimated to be the largest for the first time, but the cause is verified not to occur, the value of the node 'cause A1-regional stratum sinking' is set to be N, the posterior probabilities of other disease cause nodes are calculated again until the estimated disease cause nodes are verified to happen exactly through experience.
S204: if the verification result is yes, predicting the development trend of the tunnel structure diseases under the condition of not taking measures;
the correlation between diseases is directly reflected in the Bayesian network, so that the secondary diseases of the target diseases can be reflected when the disease development trend is predicted in the step, and the disease development trend can be predicted more comprehensively and reasonably.
In some embodiments, the method specifically comprises:
s2041: retrieving a trend class child node set S (N _ QS) of the target disease class node;
for example, if the disease type of the disease case to be diagnosed in the bayesian network shown in fig. 3 is "uneven settlement", the trend type sub-node of the disease type node "disease B1-uneven settlement" is searched in the bayesian network, and the node set { "trend C1-uneven settlement" is obtained.
S2042: and searching a disease type sub-node set S (N _ BH) of the target disease type node, and further searching a trend type sub-node set S (M _ QS) of the target disease type node.
S2043: and calculating the state variable probability distribution of the trend type sub-node sets S (N _ QS) and S (M _ QS) to be used as the quantitative evaluation of the tunnel structure disease development trend under the condition that no measures are taken currently.
S205: and recommending treatment measures based on the tunnel structure disease development trend and according to the Bayesian network.
The method directly introduces the treatment measures in the Bayesian network for the tunnel diseases, and can recommend the treatment measures by the Bayesian network.
Specifically, whether or not to take a treatment measure is generally determined according to the development tendency of a disease. If the probability of deterioration in the disease development trend is high, the step S205 needs to be performed; otherwise, the process does not need to enter S205, and only the enhanced monitoring is required under actual conditions.
In some embodiments, the method specifically comprises:
s2051: and searching a trend type sub-node set S (N _ QS) of the target disease type node according to the tunnel structure disease diagnosis Bayesian network.
S2052: retrieving all measure class parent set S (N _ CS) of trend class child set S (N _ QS) i ）；
For example, in the bayesian network shown in fig. 2, all measure-type parent nodes of the node set { "trend C1-uneven settlement" } are retrieved to obtain the set { "measure D1 intra-tunnel micro-disturbance grouting", "measure D2 ground grouting" }.
S2053: setting a state variable of a certain measure parent node as 'taken', observing the state variable probability distribution change of the trend child node N _ QS, and taking the measure corresponding to the corresponding measure parent node as a recommended measure when the probability of the state variable 'improvement' of the trend child node N _ QS is obviously increased.
S2054: repeating S2053, traversing the measure class father node set S (N _ CS) i ) All measures in (1) class parent nodes.
The intelligent diagnosis method greatly improves the diagnosis efficiency on the premise of ensuring the diagnosis accuracy, can reduce the disease diagnosis time of each time including the recommendation of treatment measures to within 1 hour, and improves the disease diagnosis speed by more than 10 times. Specific advantageous effects are not repeated here, which can be obviously derived from the above specific embodiments.
The second aspect of the present invention further provides a device for rapidly and intelligently diagnosing a structural defect of an operating tunnel, comprising: at least one memory for storing at least one program; at least one processor; when the at least one program is executed by the at least one processor, the at least one processor is caused to carry out the diagnostic method according to the first aspect of the invention. The work which needs an expert to complete originally is realized by the computer by means of computer software, the cause of the tunnel structure disease can be rapidly deduced by simple condition input, and the dependence of tunnel structure disease diagnosis on expert resources is reduced.
The third aspect of the present invention further provides a computer readable storage medium in which a program executable by a processor is stored, the program executable by the processor being adapted to perform the diagnostic method according to the first aspect of the present invention when executed by the processor.
It will be readily appreciated by those skilled in the art that the above-described preferred embodiments may be freely combined, superimposed, without conflict.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (9)
1. A method for rapidly and intelligently diagnosing operating tunnel structural diseases based on a Bayesian network is characterized by comprising the following steps:
s10: establishing a tunnel structure disease diagnosis Bayesian network, comprising:
s101: determining a tunnel structure disease diagnosis Bayesian network structure;
s102: setting the type and value range of the state variable of the Bayesian network node;
s103: determining the probability distribution of state variables of the Bayesian network nodes;
s20: the method for diagnosing the tunnel structure diseases based on the Bayesian network comprises the following steps:
s201: inputting basic characteristic data of tunnel structure diseases;
s202: inferring disease cause according to a bayesian network;
s203: verifying whether the inferred disease cause occurs; if the verification result is 'no', modifying the value of the disease cause node state variable, returning to S202 to re-infer the disease cause until the inferred disease cause happens right after verification;
s204: if the verification result is yes, predicting the development trend of the tunnel structure diseases under the condition of not taking measures;
s205: recommending treatment measures based on the tunnel structure disease development trend and through a Bayesian network; the method comprises the following steps:
s2051: searching a trend type sub-node set S (N _ QS) of a target disease type node according to the tunnel structure disease diagnosis Bayesian network;
s2052: retrieving all action class parent set S (N _ CS) of trend class child set S (N _ QS) i ）；
S2053: setting a state variable of a father node of a certain measure class as 'taken', observing the probability distribution change of the state variable of the N _ QS of the trend class child node, and taking the measure corresponding to the father node of the corresponding measure class as a recommended measure when the probability of 'improvement' of the state variable of the N _ QS of the trend class child node is obviously increased;
s2054: repeating S2053, traversing the measure class father node set S (N _ CS) i ) All measures in (1) are like parent nodes.
2. The diagnostic method of claim 1, wherein:
the step S101 of determining the Bayesian network structure for diagnosing the tunnel structure diseases comprises the following steps:
s1011: determining a disease type set considered in a tunnel structure disease diagnosis Bayesian network;
s1012: aiming at a single disease type, determining a disease cause set and a measure set;
s1013: aiming at a single disease, dividing nodes related to the disease into four types of causes, diseases, trends and measures, representing a Bayesian network structure by adopting a directed acyclic graph formed by disease-cause-measure-trend, and generating corresponding nodes according to a determined disease cause set and a determined measure set;
s1014: establishing association between the generated nodes;
s1015: repeating S1012-S1014, traversing all the disease types in the disease type set, and establishing an independent disease Bayesian network for each disease type;
s1016: and considering the internal association among different types of diseases, and determining the causal association among the Bayesian networks of the diseases in the disease type set.
3. The diagnostic method of claim 2, wherein:
the S1014 establishing an association between the generated nodes includes:
(1) The cause type node and the measure type node are used as root nodes, the trend type node is a leaf node, the cause type node takes the disease type node and the trend type node as child nodes, the disease type node takes the trend type node as child nodes, and the measure type node takes the trend type node as child nodes;
(2) Each disease node has one and only one trend node as a child node;
(3) Each trend node has one and only one disease node as a father node;
(4) Each disease node has at least one cause node as a father node;
(5) All cause type father nodes of all disease type nodes are father nodes of the trend type child nodes of all disease type nodes at the same time;
(6) Each trend class node has at least one measure class node as a parent node.
4. The diagnostic method of claim 2, wherein:
the internal association between different types of diseases in S1016 includes:
(1) Establishing association between two corresponding disease nodes, taking the disease node representing the influence factor as a father node, taking the disease node representing the influenced factor as a child node, and considering the value of the state variable of the father node in the probability distribution of the state variable of the child node;
(2) Multiple diseases are caused by the same cause, and nodes representing the same disease cause are combined;
(3) And treating various diseases by the same measure, and combining nodes representing the same treatment measure.
5. The diagnostic method of claim 1, wherein:
the step S102 of setting the type and value range of the state variable of the bayesian network node includes:
(1) For the disease node, the state variable value range is set to be { "Y: appearance "," N: no "}; if the number of the existing disease cases is large and the data is complete, grading is carried out;
(2) For the cause type node, setting a state variable value field as { "present", "absent" }; if the number of the existing disease cases is large and the data is complete, expanding the value range to be { "S: severe "," E: general "," L: lighter "and" N: there is no "};
(3) For the trend class node, the state variable value domain is set to be { "B: improvement "," W; aggravation "}; if the number of the existing disease cases is large and the data is complete, expanding the value range of the existing disease cases into { "B: improvement "," W; worsened "," N; no apparent change "};
(4) For the measure class node, the state variable value domain is set to { "Y: take "," N: no "} is taken.
6. The diagnostic method of claim 1, wherein:
the step S201 of inputting the basic feature data of the tunnel structure defect includes:
diagnosing a node state variable value domain of the Bayesian network according to the tunnel structure disease, judging a node variable value, and inputting basic characteristic data of the tunnel structure disease;
and/or the step S202 of deducing the disease cause according to the Bayesian network comprises the following steps:
calculating posterior probability of all cause father nodes of the diseases and deducing cause of the diseases;
and/or, the step S203 of verifying whether the inferred disease cause occurs includes:
and (5) verifying whether the inferred disease cause occurs by using a monitoring or detecting means on site.
7. The diagnostic method of claim 1, wherein:
the step S204 of predicting the development trend of the tunnel structure diseases under the condition of not taking measures comprises the following steps:
s2041: searching a trend sub-node set S (N _ QS) of the target disease node;
s2042: searching a disease sub-node set S (N _ BH) of a target disease node, and further searching a trend sub-node set S (M _ QS) of the target disease node;
s2043: and calculating the state variable probability distribution of the trend type sub-node sets S (N _ QS) and S (M _ QS) to be used as the quantitative evaluation of the tunnel structure disease development trend under the condition that no measures are taken currently.
8. The utility model provides a quick intelligent diagnosis device of operation tunnel structure disease which characterized in that includes:
at least one memory for storing at least one program;
at least one processor;
the at least one program, when executed by the at least one processor, causes the at least one processor to implement the diagnostic method of any one of claims 1 to 7.
9. A computer-readable storage medium in which a program executable by a processor is stored, wherein the program executable by the processor is for performing the diagnostic method of any one of claims 1 to 7 when executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211326292.9A CN115392797B (en) | 2022-10-27 | 2022-10-27 | Operating tunnel structure disease rapid intelligent diagnosis method based on Bayesian network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211326292.9A CN115392797B (en) | 2022-10-27 | 2022-10-27 | Operating tunnel structure disease rapid intelligent diagnosis method based on Bayesian network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115392797A CN115392797A (en) | 2022-11-25 |
| CN115392797B true CN115392797B (en) | 2023-04-07 |
Family
ID=84129250
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211326292.9A Active CN115392797B (en) | 2022-10-27 | 2022-10-27 | Operating tunnel structure disease rapid intelligent diagnosis method based on Bayesian network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115392797B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118134055B (en) * | 2024-05-06 | 2024-07-19 | 北京城建设计发展集团股份有限公司 | Tunnel disease occurrence probability prediction method considering disease cross influence |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017218686A2 (en) * | 2016-06-14 | 2017-12-21 | TUPL, Inc. | Fixed line resource management |
| CN111985804A (en) * | 2020-08-18 | 2020-11-24 | 华中科技大学 | Shield approaching existing tunnel safety evaluation method based on data mining and data fusion |
| CN113536678A (en) * | 2021-07-19 | 2021-10-22 | 中国人民解放军国防科技大学 | XSS risk analysis method and device based on Bayesian network and STRIDE model |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6076083A (en) * | 1995-08-20 | 2000-06-13 | Baker; Michelle | Diagnostic system utilizing a Bayesian network model having link weights updated experimentally |
| CN109145951B (en) * | 2018-07-23 | 2022-04-12 | 昆明理工大学 | Bayesian network-based tunnel lining structure service condition evaluation method |
| CN110059963A (en) * | 2019-04-20 | 2019-07-26 | 北京交通大学 | A kind of tunnel risk evaluating method based on fuzzy polymorphism Bayesian network |
| CN110991855A (en) * | 2019-11-28 | 2020-04-10 | 西南交通大学 | Bayesian network-based high-speed railway risk assessment method |
| CN111523759A (en) * | 2020-03-18 | 2020-08-11 | 北京理工大学 | Dynamic risk assessment method and device for gas pipeline |
| CN111598352A (en) * | 2020-05-25 | 2020-08-28 | 哈尔滨工业大学 | Concrete beam type bridge comprehensive evaluation method based on Bayesian network |
| CN113269960A (en) * | 2021-05-17 | 2021-08-17 | 郑州航空工业管理学院 | Method for analyzing accident risk of highway tunnel group traffic |
| CN115049124A (en) * | 2022-06-08 | 2022-09-13 | 重庆科技学院 | Deep and long tunnel water inrush prediction method based on Bayesian network |
-
2022
- 2022-10-27 CN CN202211326292.9A patent/CN115392797B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017218686A2 (en) * | 2016-06-14 | 2017-12-21 | TUPL, Inc. | Fixed line resource management |
| CN111985804A (en) * | 2020-08-18 | 2020-11-24 | 华中科技大学 | Shield approaching existing tunnel safety evaluation method based on data mining and data fusion |
| CN113536678A (en) * | 2021-07-19 | 2021-10-22 | 中国人民解放军国防科技大学 | XSS risk analysis method and device based on Bayesian network and STRIDE model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115392797A (en) | 2022-11-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Najafi et al. | Pipeline condition prediction using neural network models | |
| Van Der Gaag | Bayesian belief networks: odds and ends | |
| CN115392797B (en) | Operating tunnel structure disease rapid intelligent diagnosis method based on Bayesian network | |
| Johari et al. | Modelling the mechanical behaviour of unsaturated soils using a genetic algorithm-based neural network | |
| CN111598352A (en) | Concrete beam type bridge comprehensive evaluation method based on Bayesian network | |
| JP2006011849A (en) | Data compressing apparatus and method, data analyzing apparatus and method and data management system | |
| Ahmad et al. | A step forward towards a comprehensive framework for assessing liquefaction land damage vulnerability: Exploration from historical data | |
| CN117034143B (en) | Distributed system fault diagnosis method and device based on machine learning | |
| Najjaran et al. | Fuzzy expert system to assess corrosion of cast/ductile iron pipes from backfill properties | |
| CN104268375A (en) | Ship electric power station fault diagnosing method based on knowledge petri network | |
| CN113324754B (en) | Gear transmission system fault identification method based on multi-receptive-field-map convolution | |
| Coupé et al. | Using sensitivity analysis for efficient quantification of a belief network | |
| WO2022226843A1 (en) | System for predicting acute pancreatitis-induced organ failure, and computer device | |
| CN115049124A (en) | Deep and long tunnel water inrush prediction method based on Bayesian network | |
| CN116485029A (en) | Method, device, equipment and medium for detecting accuracy of dynamic data in oilfield development | |
| CN114580150A (en) | Submarine pipeline reliability evaluation method based on mesopic fuzzy Bayesian network | |
| CN116633639B (en) | Network intrusion detection method based on unsupervised and supervised fusion reinforcement learning | |
| CN116383723A (en) | Debris flow data anomaly identification method, computer equipment and medium | |
| US7739213B1 (en) | Method for developing complex probabilistic models | |
| CN112307918A (en) | Diagnosis method for transformer direct-current magnetic biasing based on fuzzy neural network | |
| Aldosari et al. | Generative adversarial neural network and genetic algorithms to predict oil and gas pipeline defect lengths | |
| CN117874643B (en) | Rotor fault Bayesian network diagnosis method and system based on small data set | |
| CN118335360B (en) | Interactive clinical decision support system and method based on large model knowledge enhancement | |
| Sheremetov et al. | Design of fuzzy knowledge based systems with multi-set confidence models | |
| Battefeld et al. | Formalizing cognitive biases in medical diagnostic reasoning |
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 |