Dam safety anomaly diagnosis method based on knowledge graph and Bayesian network

Abstract

The present invention relates to a dam safety anomaly diagnosis method based on a knowledge graph and a Bayesian network. The method comprises: obtaining at least one dam abnormality performance input by a user; based on the dam abnormality performance, diagnosing the dam abnormality problem that may exist in the dam through the constructed dam safety diagnosis Bayesian network; the construction of the Bayesian network comprises: collecting dam abnormality cases, dam abnormality problems and dam abnormality performances, and correlating and corresponding them to construct a dam safety diagnosis knowledge graph; based on the correlation between each dam abnormality problem and dam abnormality performance in the dam safety diagnosis knowledge graph, forming a dam safety diagnosis Bayesian network structure; based on the number of dam abnormality cases associated with dam abnormality problems and dam abnormality performance in the dam safety diagnosis knowledge graph, determining the prior probability of each dam abnormality problem node and the conditional probability table of each dam abnormality performance node in the dam safety diagnosis Bayesian network structure.

Description

Technical Field

[0001] The present invention relates to a dam safety anomaly diagnosis method based on knowledge graph and Bayesian network, which is applicable to the field of dam safety management. Background Art

[0002] my country has a vast territory, numerous rivers, a wide terrain disparity, and a huge river drop. It is rich in hydropower resources and is truly a country with abundant water resources. At present, my country has built more than 98,000 reservoirs and dams of various types, with a total storage capacity of more than 930 billion cubic meters and a hydropower installed capacity of 390 million kilowatts. Water conservancy and hydropower facilities play an increasingly important role in ensuring social and economic stability and people's living standards. Most of these dams were built from the 1950s to the 1970s. Due to the constraints of technology, equipment, and economic conditions at the time, the design requirements and construction standards were generally low, and the overall quality was poor. With the increase of dam age, the dam structure gradually ages. In addition, due to the defects of the dam body itself, environmental changes, improper operation, and poor monitoring, a considerable number of dams have hidden dangers such as material aging and deterioration, structural property decay, and foundation leakage damage, which seriously affect the performance of the project benefits and may pose a risk of dam collapse.

[0003] Dam damage and collapse is a gradual process from quantitative change to qualitative change. Discovering potential problems of dams in advance and managing them safely is the most effective way to prevent dam collapse accidents. In the practice of dam safety management, experts with rich engineering experience and professional knowledge can often quickly discover abnormal performance of dams, accurately analyze possible hidden dangers, and give corresponding engineering disposal measures. However, some front-line dam safety managers do not have this ability due to lack of engineering experience. Summary of the invention

[0004] The technical problem to be solved by the present invention is: in view of the above-mentioned problems, a dam safety anomaly diagnosis method based on knowledge graph and Bayesian network is provided.

[0005] The technical solution adopted by the present invention is: a dam safety anomaly diagnosis method based on knowledge graph and Bayesian network, comprising:

[0006] Obtain at least one abnormal performance of a dam entered by a user;

[0007] Based on the abnormal performance of the dam, the possible abnormal problems of the dam are diagnosed through the constructed dam safety diagnosis Bayesian network;

[0008] The construction of the dam safety diagnosis Bayesian network includes:

[0009] Collect dam abnormal cases, dam abnormal problems and dam abnormal performance;

[0010] Correlate and correspond dam abnormal cases, dam abnormal problems and dam abnormal performances to build a dam safety diagnosis knowledge graph;

[0011] Based on the correlation between abnormal dam problems and abnormal dam performance in the dam safety diagnosis knowledge graph, a Bayesian network structure for dam safety diagnosis is formed;

[0012] Based on the number of dam abnormality cases associated with dam abnormal problems and dam abnormal performances in the dam safety diagnosis knowledge graph, the prior probability of each dam abnormality problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure are determined.

[0013] The above-mentioned dam safety diagnosis Bayesian network structure is formed based on the correlation between various dam abnormal problems and dam abnormal performances in the dam safety diagnosis knowledge graph, including:

[0014] Each dam abnormality problem is used as the root node of the dam safety diagnosis Bayesian network, and each dam abnormality performance is used as the child node of the dam safety diagnosis Bayesian network;

[0015] Determine whether there is a second-degree correlation between various dam abnormal problems and various dam abnormal manifestations in the dam safety diagnosis knowledge graph. If so, draw the directed edges between the corresponding root nodes and child nodes in the dam safety diagnosis Bayesian network.

[0016] The method of determining the prior probability of each dam abnormal problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure based on the number of abnormal cases associated with the dam abnormal problems and the dam abnormal performance in the dam safety diagnosis knowledge graph includes:

[0017] Query dam abnormal problem nodes A i The number of dam anomaly cases associated with the dam safety diagnosis knowledge graph a i ;

[0018] Based on the number of abnormal cases of dams a i and the number of abnormal cases of all dams in the dam safety diagnosis knowledge graph w , determine the abnormal problem nodes of the dam A i The prior probability P ( A i ).

[0019] The number of abnormal cases based on the dam a i and the number of abnormal cases of all dams in the dam safety diagnosis knowledge graphw , determine the abnormal problem nodes of the dam A i The prior probability P ( A i ),include:

[0020] ,in, is the probability of an abnormal problem occurring in a dam.

[0021] The method of determining the prior probability of each dam abnormal problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure based on the number of abnormal cases associated with the dam abnormal problems and the dam abnormal performance in the dam safety diagnosis knowledge graph includes:

[0022] Query the abnormal performance node of the dam B j The number of dam anomaly cases associated with the dam safety diagnosis knowledge graph b j ;

[0023] Querying dam abnormal problem nodes in the dam safety diagnosis knowledge graph A i Node with abnormal performance of dam B j Number of dam anomaly cases between f ij ;

[0024] Based on the number of abnormal cases of dams a i 、Number of abnormal dam cases b j 、Number of abnormal dam cases f ij and the number of abnormal cases of dams w , determine the abnormal problem nodes of the dam A i Dam abnormal performance nodes under the premise of occurrence and non-occurrence B j The probability of something happening or not happening.

[0025] The number of abnormal cases based on the dam a i 、Number of abnormal dam cases b j 、Number of abnormal dam cases f ij and the number of abnormal cases of dams w , determine the abnormal problem nodes of the dam A i Dam abnormal performance nodes under the premise of occurrence and non-occurrenceB j The probability of occurrence and non-occurrence, including:

[0026] 1) Dam abnormal problem nodes Under the premise of the occurrence of abnormal performance nodes of the dam Probability of occurrence ;

[0027] 2) Dam abnormal problem nodes The dam abnormal performance node does not occur under the premise of Probability of occurrence ;

[0028] 3) Dam abnormal problem nodes Under the premise of the occurrence of abnormal performance nodes of the dam Probability of not happening ;

[0029] 4) Dam abnormal problem nodes The dam abnormal performance node does not occur under the premise of Probability of not happening .

[0030] Also includes:

[0031] In the dam safety diagnosis knowledge graph, taking the abnormal dam performance as the starting point, the dam abnormal cases and dam abnormal problems associated with it in the second-degree relationship are queried, and the reasoning path between the abnormal dam performance and the abnormal dam problems is displayed in a sub-graph visualization way.

[0032] A method for constructing a Bayesian network for dam safety diagnosis, comprising:

[0033] Collect dam abnormal cases, dam abnormal problems and dam abnormal performance;

[0034] Correlate and correspond dam abnormal cases, dam abnormal problems and dam abnormal performances to build a dam safety diagnosis knowledge graph;

[0035] Based on the correlation between abnormal dam problems and abnormal dam performance in the dam safety diagnosis knowledge graph, a Bayesian network structure for dam safety diagnosis is formed;

[0036] Based on the number of dam abnormality cases associated with dam abnormal problems and dam abnormal performances in the dam safety diagnosis knowledge graph, the prior probability of each dam abnormality problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure are determined.

[0037] A storage medium stores a computer program executable by a processor, wherein the steps of the method are implemented when the computer program is executed.

[0038] A data processing device comprises a memory and a processor, wherein a computer program executable by the processor is stored in the memory, and the steps of the method are implemented when the computer program is executed.

[0039] The beneficial effects of the present invention are as follows: the present invention constructs a dam safety diagnosis knowledge graph by associating and corresponding dam abnormal cases, dam abnormal problems, and dam abnormal performances, parses and reconstructs the network relationship of the dam safety diagnosis knowledge graph, obtains the association relationship between each abnormal problem and abnormal performance, forms a dam safety diagnosis Bayesian network structure, and based on the number of dam abnormal cases associated with dam abnormal problems and dam abnormal performances, determines the prior probability of each dam abnormal problem node and the conditional probability table of each dam abnormal performance node, and completes the construction of the dam safety diagnosis Bayesian network.

[0040] For abnormal performance of the dam, the present invention can diagnose possible abnormal problems of the dam through the constructed dam safety diagnosis Bayesian network, thereby assisting dam safety management personnel in diagnosing abnormal problems with the help of artificial intelligence methods, and visually displaying the diagnostic reasoning path through the dam safety diagnosis knowledge graph. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flowchart of the dam safety abnormality diagnosis method in the embodiment.

[0042] Figure 2 This is the dam safety diagnosis knowledge graph in the embodiment.

[0043] Figure 3 This is the Bayesian network structure for dam safety diagnosis in the embodiment.

[0044] Figure 4 It is the conditional probability table of the dam abnormal performance node in the embodiment.

[0045] Figure 5 It is a schematic diagram of the display interface of the dam safety abnormality diagnosis result in the embodiment.

[0046] Figure 6 It is the reasoning path of dam safety abnormality diagnosis in the embodiment. DETAILED DESCRIPTION

[0047] Example 1: Figure 1 As shown, this embodiment is a dam safety anomaly diagnosis method based on knowledge graph and Bayesian network, which specifically includes the following steps:

[0048] S1. Obtain at least one abnormal performance of a dam entered by a user.

[0049] S2. Based on the abnormal performance of the dam, the possible abnormal dam problems of the dam are diagnosed through the constructed dam safety diagnosis Bayesian network.

[0050] S3. Based on the abnormal performance of the dam and combined with the constructed dam safety diagnosis knowledge graph, the reasoning path between the abnormal performance of the dam and the abnormal problems of the dam is displayed in a visual way.

[0051] The method for constructing a dam safety diagnosis knowledge graph in this embodiment includes the following steps:

[0052] In response to the needs of dam safety abnormality diagnosis, various typical abnormal problems and abnormal manifestations of dams are identified, and dam abnormality cases are widely collected to form structured data required for building a knowledge graph.

[0053] (1) Constructing a knowledge base of abnormal dam problems and abnormal dam performance. By means of a questionnaire survey, experts in the field of dam safety were asked to collect information on various abnormal dam problems and abnormal dam performances that might occur, thus forming a knowledge base of abnormal dam problems and abnormal dam performances.

[0054] The knowledge base of abnormal dam problems includes but is not limited to various abnormal dam problems such as "abnormal dam leakage" and "dam cracks"; the knowledge base of abnormal dam manifestations includes but is not limited to the obvious manifestations of abnormal dam problems such as "large water volume in the measuring weir" and "large tensile stress at the dam heel".

[0055] (2) Construct a dam anomaly case database. Extract historical dam anomaly cases at home and abroad from professional reports, professional books, academic papers, various web pages and other media related to dam safety management to form a dam anomaly case database.

[0056] In this example, case extraction includes two methods: manual extraction and automatic extraction. Automatic extraction includes various extraction methods such as long short-term memory network and UIE model.

[0057] B. Associate dam abnormal cases, dam abnormal problems, and dam abnormal performances to build a dam safety diagnosis knowledge graph.

[0058] (1) Extracting triples. Analyze each dam abnormality case in step S1, break down the abnormal problem and abnormal performance to which the case belongs, and correspond them with the abnormal problems and abnormal performance in the dam abnormal problem knowledge base and the dam abnormal performance knowledge base, forming a large number of triples in the form of (abnormal case → abnormal problem) and (abnormal case → abnormal performance).

[0059] (2) Construct a knowledge graph for dam safety diagnosis. Use the py2neo algorithm package to build an interface between Python and the Neo4j database, and programmatically import the extracted triples into the Neo4j database in batches to complete the construction of the knowledge graph for dam safety diagnosis. The knowledge graph for dam safety diagnosis is as follows: Figure 2 shown.

[0060] In this embodiment, the dam safety diagnosis Bayesian network is constructed based on the dam safety diagnosis knowledge graph, and its construction method includes:

[0061] Ⅰ. Analyze and reconstruct the network relationship of the dam safety diagnosis knowledge graph to obtain the correlation between various abnormal problems and abnormal manifestations, and form a Bayesian network structure for dam safety diagnosis. The local Bayesian network is shown in the figure below: Figure 3 shown.

[0062] (1) Extracting dam safety diagnosis Bayesian network nodes. Extract each element in the dam abnormal problem knowledge base to form the root node (abnormal problem node) of the dam safety diagnosis Bayesian network. A i ; Extract each element from the dam abnormal performance knowledge base to form a sub-node (abnormal performance node) of the dam safety diagnosis Bayesian network B j .

[0063] (2) Construct the directed edges of the dam safety diagnosis Bayesian network. For each abnormal problem node in the dam safety diagnosis Bayesian network A i , analyze it and each abnormal performance node in turn B j Whether there is a second-degree association relationship in the dam safety diagnosis knowledge graph, if the number of existing relationships is greater than 1, a directed edge is drawn from the abnormal problem node to the corresponding abnormal performance node. The cypher formula for judging the directed edge is as follows:

[0064] (1)

[0065] Ⅱ. Based on the abnormal cases associated with abnormal problems and abnormal performances in the dam safety diagnosis knowledge graph, the parameters of the dam safety diagnosis Bayesian network are learned, and the prior probability of the abnormal problem node and the conditional probability table of the abnormal performance node are set respectively to complete the construction of the Bayesian network.

[0066] (1) Calculate the prior probability of abnormal problem nodes. For each abnormal problem node in the dam safety diagnosis Bayesian network A i , query the number of dam anomaly cases associated with it in the dam safety diagnosis knowledge graph through cypher commandsa i ,as follows:

[0067] (2)

[0068] Count all the abnormal dam cases in the dam safety diagnosis knowledge graph w , the cypher command is as follows:

[0069] (3)

[0070] Calculate each abnormal problem node A i The ratio of the number of associated dam anomaly cases to the total number of dam anomaly cases in the knowledge graph is as follows:

[0071] (4)

[0072] Furthermore, considering the probability of abnormal problems in an ordinary dam , you can get each abnormal problem node A i The prior probability P ( A i ),as follows:

[0073] (5)

[0074] (2) Calculate the conditional probability table of abnormal performance nodes. In the dam safety diagnosis Bayesian network, for abnormal problem nodes with directed edges Nodes with abnormal performance ,definition , Respectively represent abnormal problem nodes Happening and not happening, , Respectively represent abnormal performance nodes If it happens or not, you need to determine the abnormal performance node The conditional probability table is as follows Figure 4 As shown, the parameters that need to be solved are , , and .

[0075] First, query the abnormal performance nodes in the dam safety diagnosis knowledge graph Number of associated dam hazard cases b j , the cypher command is as follows:

[0076] (6)

[0077] Query hidden danger problem nodes in the dam safety diagnosis knowledge graph Nodes with abnormal performance Number of dam hazard cases between f ij , the cypher command is as follows:

[0078] (7)

[0079] Indicates abnormal problem node The abnormal performance node occurs under the premise The probability of occurrence is calculated as follows:

[0080] (8)

[0081] in, f ij Diagnose potential problems in the knowledge graph for dam safety Nodes with abnormal performance The number of dam hazard cases between a i Diagnose potential problems in the knowledge graph for dam safety Number of associated dam anomaly cases.

[0082] Indicates abnormal problem node Abnormal performance node without occurrence The probability of occurrence is calculated as follows:

[0083] (9)

[0084] in, b j Diagnose abnormal nodes in the knowledge graph for dam safety The number of associated dam anomaly cases, It is the total number of dam anomaly cases in the dam safety diagnosis knowledge graph.

[0085] and About 1 complement, so It is expressed as:

[0086] (10)

[0087] and About 1 complement, so It is expressed as:

[0088] (11)

[0089] In this embodiment, step S2 includes: activating the abnormal performance node of the current dam in the dam safety diagnosis Bayesian network, and then each node in the dam safety diagnosis knowledge graph will be adaptively adjusted by probability calculation and other methods, and the abnormal problem node with a high probability after adjustment will be output in the form of "abnormal problem + probability value", which is the result of dam safety abnormal diagnosis, such as Figure 5 shown.

[0090] In this embodiment, step S3 includes: in the constructed dam safety diagnosis knowledge graph, taking the dam abnormal performance as the starting point, querying the dam abnormal cases and dam abnormal problems associated with it in the second-degree relationship, and displaying the reasoning path between the dam abnormal performance and the dam abnormal problems in a sub-graph visualization manner, such as Figure 6 shown.

[0091] Embodiment 2: This embodiment is a storage medium on which a computer program that can be executed by a processor is stored. When the computer program is executed, the steps of the dam safety anomaly diagnosis method based on knowledge graph and Bayesian network in Embodiment 1 are implemented.

[0092] Embodiment 3: This embodiment is a data processing device having a memory and a processor. The memory stores a computer program that can be executed by the processor. When the computer program is executed, the steps of the dam safety anomaly diagnosis method based on knowledge graph and Bayesian network in Embodiment 1 are implemented.

[0093] Embodiment 4: This embodiment is a method for constructing a Bayesian network for dam safety diagnosis, which specifically includes the following steps:

[0094] Collect dam abnormal cases, dam abnormal problems and dam abnormal performance;

[0095] Correlate and correspond dam abnormal cases, dam abnormal problems and dam abnormal performances to build a dam safety diagnosis knowledge graph;

[0096] Based on the correlation between abnormal dam problems and abnormal dam performance in the dam safety diagnosis knowledge graph, a Bayesian network structure for dam safety diagnosis is formed;

[0097] Based on the number of dam abnormality cases associated with dam abnormal problems and dam abnormal performances in the dam safety diagnosis knowledge graph, the prior probability of each dam abnormality problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure are determined.

[0098] Embodiment 5: This embodiment is a storage medium on which a computer program that can be executed by a processor is stored. When the computer program is executed, the steps of the method for constructing a Bayesian network for dam safety diagnosis in Embodiment 4 are implemented.

[0099] Embodiment 6: This embodiment is a data processing device having a memory and a processor. The memory stores a computer program that can be executed by the processor. When the computer program is executed, the steps of the method for constructing a Bayesian network for dam safety diagnosis in Embodiment 4 are implemented.

Claims

1. A dam safety anomaly diagnosis method based on knowledge graph and Bayesian network, characterized in that: include: Obtain at least one abnormal performance of a dam entered by a user; Based on the abnormal performance of the dam, the possible abnormal problems of the dam are diagnosed through the constructed dam safety diagnosis Bayesian network; The construction of the dam safety diagnosis Bayesian network includes: Collect dam abnormal cases, dam abnormal problems and dam abnormal performance; Correlate and correspond dam abnormal cases, dam abnormal problems and dam abnormal performance to build a dam safety diagnosis knowledge graph; Based on the correlation between abnormal dam problems and abnormal dam performance in the dam safety diagnosis knowledge graph, a Bayesian network structure for dam safety diagnosis is formed; Based on the number of dam abnormality cases associated with dam abnormality problems and dam abnormality performance in the dam safety diagnosis knowledge graph, determine the prior probability of each dam abnormality problem node and the conditional probability table of each dam abnormality performance node in the dam safety diagnosis Bayesian network structure; The method of determining the prior probability of each dam abnormal problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure based on the number of abnormal cases associated with the dam abnormal problems and the dam abnormal performance in the dam safety diagnosis knowledge graph includes: Query dam abnormal problem node A i The number of dam anomaly cases associated with the dam safety diagnosis knowledge graph a i ; Based on the number of abnormal cases of damsa i And the number of all dam abnormal cases w in the dam safety diagnosis knowledge graph, determine the dam abnormal problem node A i The prior probability P(A i ); The method of determining the prior probability of each dam abnormal problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure based on the number of abnormal cases associated with the dam abnormal problems and the dam abnormal performance in the dam safety diagnosis knowledge graph includes: Query dam abnormal performance node B j The number of dam anomaly cases associated with the dam safety diagnosis knowledge graph b j ; Query dam abnormal problem node A in the dam safety diagnosis knowledge graph i Node B with abnormal performance of dam j The number of abnormal cases of dams between ij ; Based on the number of abnormal cases of damsa i 、Number of abnormal dam casesb j 、Number of abnormal dam cases f ij and the number of dam abnormality cases w, determine the dam abnormality problem node A i Dam abnormal performance node B under the premise of occurrence and non-occurrence j The probability of something happening or not happening.

2. The dam safety anomaly diagnosis method based on knowledge graph and Bayesian network according to claim 1 is characterized in that: The Bayesian network structure for dam safety diagnosis is formed based on the correlation between various dam abnormal problems and dam abnormal performances in the dam safety diagnosis knowledge graph, including: Each dam abnormality problem is used as the root node of the dam safety diagnosis Bayesian network, and each dam abnormality performance is used as the child node of the dam safety diagnosis Bayesian network; Determine whether there is a second-degree correlation between various dam abnormal problems and various dam abnormal manifestations in the dam safety diagnosis knowledge graph. If so, draw the directed edges between the corresponding root nodes and child nodes in the dam safety diagnosis Bayesian network.

3. The dam safety anomaly diagnosis method based on knowledge graph and Bayesian network according to claim 1 is characterized in that: The number of abnormal cases based on dam a i And the number of all dam abnormal cases w in the dam safety diagnosis knowledge graph, determine the dam abnormal problem node A i The prior probability P(A i ),include: Among them, β is the probability of a dam having abnormal problems.

4. The dam safety anomaly diagnosis method based on knowledge graph and Bayesian network according to claim 1 is characterized in that: The number of abnormal cases based on dam a i 、Number of abnormal dam casesb j 、Number of abnormal dam cases f ij and the number of dam abnormality cases w, determine the dam abnormality problem node A i Dam abnormal performance node B under the premise of occurrence and non-occurrence j The probability of occurrence and non-occurrence, including: 1) Dam abnormal problem node A i Under the premise of the occurrence of abnormal performance of dam node B j Probability of occurrence 2) Dam abnormal problem node A i Under the premise that the dam does not abnormally behave node B j Probability of occurrence 3) Dam abnormal problem node A i Under the premise of the occurrence of abnormal performance of dam node B j Probability of not happening 4) Dam abnormal problem node A i Under the premise that the dam does not abnormally behave node B j Probability of not happening 5. The dam safety anomaly diagnosis method based on knowledge graph and Bayesian network according to claim 1 is characterized in that: Also includes: In the dam safety diagnosis knowledge graph, taking the abnormal dam performance as the starting point, the dam abnormal cases and dam abnormal problems associated with it in the second-degree relationship are queried, and the reasoning path between the abnormal dam performance and the abnormal dam problems is displayed in a sub-graph visualization way.

6. A method for constructing a Bayesian network for dam safety diagnosis, characterized in that: include: Collect dam abnormal cases, dam abnormal problems and dam abnormal performance; Correlate and correspond dam abnormal cases, dam abnormal problems and dam abnormal performance to build a dam safety diagnosis knowledge graph; Based on the correlation between abnormal dam problems and abnormal dam performance in the dam safety diagnosis knowledge graph, a Bayesian network structure for dam safety diagnosis is formed; Based on the number of dam abnormality cases associated with dam abnormality problems and dam abnormality performance in the dam safety diagnosis knowledge graph, determine the prior probability of each dam abnormality problem node and the conditional probability table of each dam abnormality performance node in the dam safety diagnosis Bayesian network structure; The method of determining the prior probability of each dam abnormal problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure based on the number of abnormal cases associated with the dam abnormal problems and the dam abnormal performance in the dam safety diagnosis knowledge graph includes: Query dam abnormal problem node A i The number of dam anomaly cases associated with the dam safety diagnosis knowledge graph a i ; Based on the number of abnormal cases of damsa i And the number of all dam abnormal cases w in the dam safety diagnosis knowledge graph, determine the dam abnormal problem node A i The prior probability P(A i ); The method of determining the prior probability of each dam abnormal problem node and the conditional probability table of each dam abnormal performance node in the dam safety diagnosis Bayesian network structure based on the number of abnormal cases associated with the dam abnormal problems and the dam abnormal performance in the dam safety diagnosis knowledge graph includes: Query dam abnormal performance node B j The number of dam anomaly cases associated with the dam safety diagnosis knowledge graph b j ; Query dam abnormal problem node A in the dam safety diagnosis knowledge graph i Node B with abnormal performance of dam j The number of abnormal cases of dams between ij ; Based on the number of abnormal cases of damsa i 、Number of abnormal dam casesb j 、Number of abnormal dam cases f ij and the number of dam abnormality cases w, determine the dam abnormality problem node A i Dam abnormal performance node B under the premise of occurrence and non-occurrence j The probability of something happening or not happening.

7. A storage medium having stored thereon a computer program executable by a processor, characterized in that: When the computer program is executed, the steps of the method according to any one of claims 1 to 6 are implemented.

8. A data processing device comprising a memory and a processor, wherein the memory stores a computer program executable by the processor, wherein: When the computer program is executed, the steps of the method according to any one of claims 1 to 6 are implemented.

Images