Arch dam abutment adverse structural plane processing measure recommendation method based on knowledge graph and terminal equipment

Abstract

The invention provides an arch dam abutment unfavorable structural plane processing measure recommendation method based on a knowledge graph and terminal equipment, and the method comprises the following steps: constructing an arch dam abutment unfavorable structural plane geological information knowledge graph ontology model, extracting geological entity information related to an unfavorable structural plane, and perfecting a geological information data layer; constructing an arch dam abutment unfavorable structural plane processing measure knowledge graph ontology model, extracting processing measure entity information related to an unfavorable structural plane, and perfecting a processing measure data layer; establishing a corresponding relationship between the geological information of the unfavorable structural plane and the treatment measures to form a knowledge graph model of the treatment measures of the unfavorable structural plane of the arch dam abutment; on the basis of a reasoning method of distributed representation learning, knowledge graph completion is carried out; and inputting geological survey information of the unfavorable structural plane to be processed into the atlas, and recommending corresponding processing measures. According to the method, the geological survey information and the treatment measures of the unfavorable structural plane of the arch dam abutment are subjected to structured knowledge expression, and the treatment measures of the unfavorable structural plane can be quickly adjusted along with continuous improvement of the geological information.

Description

Technical Field

[0001] The present invention relates to the field of geotechnical engineering, and in particular to a method and terminal equipment for recommending treatment measures for unfavorable structural surfaces of arch dam abutments based on a knowledge graph. Background Art

[0002] Structural surfaces are an important part of the structural characteristics of rock mass, and the engineering mechanical properties of rock mass are often controlled by structural surfaces. Structural surfaces can be divided into faults, dislocation zones, weak zones, joints and fissures according to their absolute scale. Structural surfaces of different scales affect regional mountain stability, engineering rock mass stability, etc. to varying degrees.

[0003] The unfavorable structural surface of the dam abutment refers to the structural surface that exists in the abutment area and has an adverse effect on the stability of the dam. During the construction of high arch dams, unfavorable structural surfaces (faults, weak zones, etc.) will have an adverse effect on the foundation stiffness and stability of the dam abutment, such as the stability of the slider in the abutment resistance body, deformation of the arch end caused by faults and weak zones, and uncoordinated force transmission. In addition, at all stages of the arch dam design, attention should be paid to the stability analysis of the arch dam, especially in terms of the stability of the arch seat of the arch dam. Most safety accidents of arch dams are caused by the instability or excessive deformation of the abutment rock mass, and rarely by stress problems of the arch dam itself. Therefore, it is very important to deal with and analyze the unfavorable structural surfaces of the arch dam abutment.

[0004] In traditional engineering design work, there are often three ways to deal with unfavorable structural surfaces: 1. Based on the experience of designers, preliminary measures are given to deal with unfavorable structural surfaces; 2. Geomechanics model tests are used to simulate the structural characteristics of engineering structures, rock masses, faults, weak zones, etc., and to intuitively track the cracking and damage of the overall structure; 3. Numerical calculation methods are used to simulate the stability of the dam shoulder rock mass and the reinforcement measures based on the finite element method of elastic-plastic mechanics, fracture mechanics and damage mechanics. However, these methods still have the following shortcomings:

[0005] 1. There are many uncertainties in the method based on manual experience. In particular, as the foundation surface is excavated, the geological information may differ from that in the early survey and design stage, resulting in an inability to respond quickly. At the same time, the geological conditions faced by different projects vary greatly, and the experience of designers is difficult to meet such complex and changing situations.

[0006] 2. The geomechanical model test method is limited by the test conditions and cannot conduct a comprehensive and detailed analysis of possible reinforcement measures. It is also limited by the scale effect, which limits the applicability of the test results in actual engineering.

[0007] 3. The results of numerical simulation methods are affected by factors such as boundary condition settings, parameter selection, and long calculation time, and are also difficult to meet the needs of fast real-time response.

[0008] Knowledge graph is a technical method that uses a graph model to describe knowledge and model the relationship between all things in the world. It uses nodes to represent semantic symbols and edges to represent the relationship between semantics. Knowledge graph is called knowledge domain visualization or knowledge domain mapping map in the library and information industry. It is a series of various graphics that show the development process and structural relationship of knowledge. It uses visualization technology to describe knowledge resources and their carriers, and mine, analyze, construct, draw and display knowledge and their mutual connections.

[0009] Therefore, there is an urgent need for a method to recommend treatment measures for unfavorable structural surfaces of arch dam shoulders based on knowledge graphs, which can express previous research results, design personnel experience, geological survey data and other information in a structured knowledge form. In engineering design, treatment measures for unfavorable structural surfaces can be quickly given based on the latest geological survey data, providing a basis for subsequent calculation and analysis. Summary of the invention

[0010] The purpose of the present invention is to provide a method and terminal device for recommending treatment measures for unfavorable structural surfaces of arch dam abutments based on knowledge graphs in response to the above-mentioned deficiencies in the prior art.

[0011] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0012] A method for predicting adverse structural surface treatment of an arch dam abutment comprises the following steps:

[0013] S1. Constructing the knowledge graph ontology model G1 of geological information of adverse structural surface of arch dam abutment;

[0014] The expression of the geological information knowledge graph ontology model G1 of the arch dam abutment unfavorable structural surface is:

[0015] G1=(C geo , R geo , P geo , I geo );

[0016] Among them, C geo Represents concepts related to geological information; R geo Indicates the spatial position relationship between the unfavorable structural surface and the dam abutment; P geo Represents the specific attributes of various geological information related concepts; I geo Represents a specific instance of geological information;

[0017] S2, extracting geological entity information related to the unfavorable structural surface, and improving the data layer of the geological information knowledge graph ontology model G1;

[0018] S3, construct the knowledge graph ontology model G2 of treatment measures for unfavorable structural surfaces of arch dam abutments;

[0019] The expression of the knowledge graph ontology model G2 of the treatment measures for the adverse structural surface of the arch dam abutment is:

[0020] G2=(C rein , R rein , P rein , I rein );

[0021] Among them, C rein Represents the concept related to treatment measures; R rein Indicates the corresponding relationship between treatment measures and unfavorable structural surface; P rein Represents the specific attributes of the concepts related to each treatment measure; I rein Indicates a specific instance of a treatment measure;

[0022] S4, extracting treatment measures for unfavorable structural surfaces, and improving the data layer of the treatment measure knowledge graph ontology model G2;

[0023] S5, establish a corresponding relationship between the treatment measures and the unfavorable structural surface, add a corresponding treatment measure node in the geological information knowledge graph ontology model G1, couple the geological information knowledge graph ontology model G1 and the treatment measure knowledge graph ontology model G2 to form a knowledge graph model of treatment measures for the unfavorable structural surface of the arch dam abutment;

[0024] S6. Using a distributed representation learning-based reasoning algorithm to improve the entity relationship of the knowledge graph model of the treatment measures for the adverse structural surface of the arch dam abutment;

[0025] S7. Before and after the excavation of the arch dam foundation surface, the geological information of the unfavorable structural surface to be treated is input into the knowledge graph model of treatment measures for the unfavorable structural surface of the arch dam abutment to predict treatment measures that match the geological information.

[0026] The present invention expresses the geological information of the unfavorable structural surface of the arch dam shoulder and the corresponding treatment measures in a structured knowledge manner, transforming the traditional design method relying on experience into a design based on knowledge-driven. Moreover, the knowledge graph can adjust the treatment measures quickly and in real time as the geological information is continuously improved.

[0027] Further, in step S2, by analyzing the geological survey data in the early stage of the project, the geological entity information related to the unfavorable structural surface is extracted, and the geological entity information with the same meaning is fused through the K-means clustering algorithm, and the fused first entity information is stored in the Neo4j graph database; wherein, the input of the K-means clustering algorithm is all unfavorable structural surfaces confirmed by geological exploration, including the spatial position and rock mass parameters of each unfavorable structural surface, and the output is the merged geological entity information (merging the unfavorable structural surfaces with similar spatial positions and similar rock mass parameters);

[0028] In step S4, the entity information of treatment measures related to the adverse structural surface is extracted by analyzing the design data and publicly published literature, and the entity information of treatment measures with the same meaning is fused through the K-means clustering algorithm, and the fused second entity information is stored in the Neo4j graph database; wherein, the input of the K-means clustering algorithm is the adverse structural surface treatment measures extracted from the design data and publicly published literature, and the output is the merged treatment measures (that is, the treatment measures with different names but actually expressing the same meaning are merged and expressed with a unified name).

[0029] Further, in step S6, the TransE algorithm is used to predict and improve the triple information of the knowledge graph of the treatment measures for the adverse structural surface of the arch dam abutment. The specific implementation process includes:

[0030] The TransE algorithm is used to define the distance function; the head entity and the tail entity in the knowledge graph model of the treatment measures for the adverse structural surface of the arch dam abutment are randomly replaced to generate a damaged triple; the distance function and the damaged triple are added to the loss function of the knowledge graph model of the treatment measures for the adverse structural surface of the arch dam abutment to perform knowledge reasoning training. The training set contains both correct triples (positive samples) and damaged triples (negative samples). These negative samples are generated by randomly replacing the head entity or the tail entity on the basis of the original positive samples. The positive samples and negative samples in the training set are constructed in a 1:1 ratio. Only positive samples are included in the test set. The data volume of the training set and the test set is divided in a ratio of 80%:20%. In each training iteration, the loss function value of the model (TransE) is calculated. When the loss function value no longer decreases significantly in multiple consecutive training iterations, the model can be considered to have converged. A maximum number of iterations can also be set, for example, 100 times. If the model has converged before the maximum number of iterations is reached, the training can be stopped early; otherwise, the training will automatically stop when the maximum number of iterations is reached.

[0031] The distance function calculation formula is:

[0032]

[0033] Among them, h, r, and t represent the vector representation of the head entity, relation, and tail entity respectively. represents the L2 norm;

[0034] The calculation formula for the destroyed triples is:

[0035] S ( ' h,l,t) ={(h′,l,t)|h′∈E}∪{(h,l,t′)|t′∈E};

[0036] Among them, (h′,l,t) and (h,l,t′) indicate that the head entity or the tail entity is replaced by a random entity as a control group, and E represents the entity set;

[0037] The calculation formula of the loss function is:

[0038]

[0039] Among them, (h′, l, t′) indicates that the head entity or the tail entity is replaced by a random entity, S represents the undamaged triple, γ represents the interval in the loss function, which is equivalent to the interval between a correct triple and an incorrect triple. The larger the γ, the stricter the correction for word embedding. [x] + If it is greater than 0, it takes the original value; if it is less than 0, it takes 0.

[0040] As an inventive concept, the present invention also provides a terminal device, comprising:

[0041] one or more processors;

[0042] A memory having one or more programs stored thereon, which, when executed by the one or more processors, enables the one or more processors to implement the steps of the above method of the present invention.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] (1) The present invention expresses the geological information of the unfavorable structural surface of the arch dam abutment and the corresponding treatment measures in a structured knowledge form, thereby transforming the traditional design method from experience-based to knowledge-driven design. It can clearly reveal the relationship and influence between the geological information and the treatment measures, enabling personnel to more accurately understand and treat the unfavorable structural surface of the arch dam abutment, and providing a more scientific and effective solution for the treatment of the unfavorable structural surface of the arch dam abutment.

[0045] (2) The knowledge graph of treatment measures for unfavorable structural surfaces of arch dam abutments of the present invention is an intelligent knowledge base, which contains all possible information and treatment measures on unfavorable structural surfaces of arch dam abutments. It can be updated and adjusted in real time as geological information is continuously improved, providing a more comprehensive and scientific solution for the treatment of unfavorable structural surfaces of arch dam abutments, thereby ensuring the accuracy and effectiveness of the treatment measures, and ensuring the smooth progress of the project and the safety of its quality.

[0046] (3) The present invention can be implemented through computer software, which has high intelligence and automation, can easily obtain and process geological information, and automatically generate corresponding treatment measures, which not only improves the efficiency and quality of the design, but also reduces the impact of human factors on the design, making the results more objective and fair. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is an example diagram of the geological information knowledge map of the unfavorable structural surface of the arch dam abutment according to an embodiment of the present invention;

[0048] Figure 2 This is an example diagram of the knowledge graph of treatment measures for unfavorable structural surfaces of arch dam abutments according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] Example 1

[0050] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. It should be understood by those skilled in the art that these implementations are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0051] like Figure 1 As shown, an embodiment of the present invention provides a method for predicting adverse structural surface treatment of an arch dam abutment, comprising the following steps:

[0052] S1: Construct the knowledge graph ontology model G1 of the geological information of the adverse structural surface of the arch dam shoulder to express the geological information in a structured knowledge manner.

[0053] The expression of the knowledge graph ontology model G1 of the unfavorable structural surface of the arch dam abutment is:

[0054] G1=(C geo , R geo , P geo , I geo );

[0055] Among them, C geo Represents concepts, including concepts related to geological information, such as lithology, faults, weak structural surfaces, etc.; R geoIt indicates the spatial relationship between the unfavorable structural surface and the dam abutment, including spatial relationship, correlation relationship, causal relationship, etc.; P geo Represents attributes, i.e., specific attributes of various geological information-related concepts, such as lithology grade, fault orientation, etc.; I geo Represents a specific instance of geological information.

[0056] This embodiment 1 constructs a knowledge graph of geological information of arch dam abutments by combining top-down and bottom-up methods. That is, the model layer defines conceptual entities and their attributes, hierarchical semantic relationships and constraint rules from top to bottom to construct an accurate and well-structured conceptual system architecture. In this embodiment 1, this system architecture is mainly reflected in the characterization of the relationship between the unfavorable structural surface and the abutment through spatial coordinates, and the coupling of the parameters of the unfavorable structural surface (such as strike, width, etc.) with the rock performance parameters (such as strength, rock grade, etc.), and the characterization of factors that may affect subsequent treatment measures through semantic relationship definitions.

[0057] S2. By analyzing the geological survey data in the early stage of the project, the geological entity information related to the unfavorable structural surface is extracted to improve the data layer of the geological information knowledge graph ontology model G1 of the unfavorable structural surface of the arch dam shoulder.

[0058] like Figure 1 As shown, the unfavorable structural surface geological information knowledge map constructed in this embodiment 1 focuses on displaying the geological conditions of the arch dam abutment in a structured form. The impact of the unfavorable structural surface of the dam abutment on the safety and stability of the arch dam is mainly reflected in the direction of the unfavorable structural surface, the position from the dam abutment, and the rock mass parameters around the unfavorable structural surface.

[0059] In addition, this embodiment 1 uses the K-means clustering algorithm to merge the geological information entity information with the same meaning, and stores the merged first entity information in the Neo4j graph database.

[0060] Specifically, the K-means clustering algorithm is a simple and classic distance-based clustering algorithm, which uses distance as an evaluation index of similarity, and believes that the closer the distance between two objects, the greater their similarity. The algorithm believes that a cluster is composed of objects that are close in distance, so obtaining compact and independent clusters is the ultimate goal. In this embodiment 1, the input of the K-means clustering algorithm is all unfavorable structural surfaces confirmed by geological exploration, including the spatial position of each unfavorable structural surface and rock mass parameters, and the output is the merged geological entity information (merging unfavorable structural surfaces with similar spatial positions and similar rock mass parameters.

[0061] Specifically, Neo4j is an open source NoSQL graph database that can transform data into nodes, relationships, and attributes. It is based on graph theory, stores data as nodes and relationships, and allows high-performance retrieval and query of these structures.

[0062] S3. Construct the knowledge graph ontology model G2 of treatment measures for unfavorable structural surfaces of arch dam abutments, and express the treatment measures in a structured knowledge form.

[0063] The expression of the knowledge graph ontology model G2 of the treatment measures for the unfavorable structural surface of the arch dam abutment is:

[0064] G2=(C rein , R rein , P rein , I rein );

[0065] Among them, C rein Represents concepts, including concepts related to treatment measures, such as grouting, cut-and-fill replacement, footing, etc.; R rein Represents the relationship, mainly the corresponding relationship between the treatment measures and the unfavorable structural surface; P rein Represents attributes, i.e., the specific attributes of the concepts related to each treatment measure, such as the length of the shear hole, the cross-sectional dimensions, etc.; I rein Indicates a specific instance of a treatment measure.

[0066] S4. Based on design data and publicly published literature, first extract the entity information of treatment measures related to the unfavorable structural surface, that is, what treatment measures are available for the unfavorable structural surface, and improve the data layer of the knowledge graph ontology model G2 of the treatment measures of the unfavorable structural surface of the arch dam abutment. And through the K-means clustering algorithm, merge the entity information of treatment measures with the same meaning, and store the merged second entity information in the Neo4j graph database.

[0067] In this embodiment 1, the input of the K-means clustering algorithm is the adverse structural surface treatment measures extracted from the design data and publicly published literature, and the output is the merged treatment measures (that is, the treatment measures with different names but actually expressing the same meaning are merged and expressed with a unified name).

[0068] S5. Establish a corresponding relationship between the treatment measures and the unfavorable structural surface, add corresponding treatment measure nodes to the constructed geological information knowledge graph of the unfavorable structural surface, couple the geological information knowledge graph and the treatment measure knowledge graph to form a knowledge graph model of treatment measures for the unfavorable structural surface of the arch dam shoulder, such as Figure 2 shown.

[0069] S6: Use the reasoning algorithm based on distributed representation learning to improve the entity relationships of the constructed knowledge graph model.

[0070] It should be noted that the reasoning method based on distributed representation learning is to find a mapping function that maps symbolic representations to vector space for numerical representation, capturing the implicit associations between entities and relations.

[0071] This embodiment 1 adopts the TransE algorithm, i.e., the transfer distance algorithm, which transforms the problem of measuring the rationality of triples in the vectorized knowledge graph into the problem of measuring the distance between the head entity and the tail entity. It regards the entities and relations in the knowledge graph as two matrices. The entity matrix structure is (ne*d1), where ne represents the number of entities, d1 represents the dimension of each entity vector, and each row in the matrix represents the word vector of an entity; while the relationship matrix structure is (nr*d2), where nr represents the number of relationships, and d2 represents the dimension of each relationship vector. In the TransE algorithm, both entities and relationships are represented as vectors. For a specific relationship (head, relation, tail), the vector representation of the relationship is interpreted as a transfer vector from the vector of the head entity to the vector of the tail entity. By continuously adjusting h, l and t (the vectors of head, relation and tail), (h+l) is made as equal to t as possible, i.e., h+l=t. In other words, if a triple in the knowledge graph is established, its entities and relationships need to satisfy the relationship head+relation=tail.

[0072] Specifically, this embodiment 1 uses the TransE algorithm to define a distance function; randomly replaces the head entity in the knowledge graph model of treatment measures for adverse structural surfaces of arch dam abutments, and the tail entity generates a destroyed triplet; adds the distance function and the destroyed triplet to the loss function in the knowledge graph model of treatment measures for adverse structural surfaces of arch dam abutments to perform knowledge reasoning training.

[0073] Given a triple set S, each triple is represented as (h, l, t), h and t belong to the entity set E, and the corresponding embedding is represented as (h, l, t), satisfying h+l=t, where h represents the vector representation of the head entity, l represents the vector representation of the relationship, and t represents the vector representation of the tail entity.

[0074] Define the distance function d(h+l,t) to measure the distance between h+l and t:

[0075]

[0076] Here, 2 represents the L2 norm.

[0077] At the same time, this embodiment 1 also considers the damaged triplets, that is, the triplets that do not exist in the real data set. Since these triplets are wrong, it is expected that the model can increase the distance between its predicted tail entity vector and the real tail entity vector, so that the loss function value of the damaged triplets is larger.

[0078] Destroyed triple S ( ' h,l,t) The calculation formula is:

[0079] S′ (h,l,t) ={(h′,l,t)|h′∈E}∪{(h,l,t′)|t′∈E};

[0080] Among them, (h′,l,t) and (h,l,t′) indicate that the head entity or the tail entity is replaced by a random entity as a control group.

[0081] When training a knowledge graph model, we hope that the model can correctly predict the correct triple relationship, so the original triple loss function should be smaller, while the destroyed triple loss function should be larger.

[0082] Specifically, the TransE algorithm adds the head entity and relation vectors to obtain the predicted vector of the tail entity. If this predicted vector is close to the true tail entity vector, then the triple relationship is correct, that is, the loss function value of the original triple should be small.

[0083] Add the distance function and the destroyed triples into the loss function of the TransE algorithm The loss function The calculation formula is:

[0084]

[0085] Among them, (h′, l, t′) indicates that the head entity or the tail entity is replaced by a random entity, S represents the undamaged triple, γ represents the interval in the loss function, which is equivalent to the interval between a correct triple and an incorrect triple. The larger the γ, the stricter the correction for word embedding. [x] + If it is greater than 0, it takes the original value; if it is less than 0, it takes 0.

[0086] Therefore, when training the knowledge graph model, this embodiment 1 optimizes the model by minimizing the loss function of the original triples and maximizing the loss function of the destroyed triples. This allows the model to more accurately predict the correct triple relationship and avoid predicting the wrong binary relationship.

[0087] Step 7: Before and after the excavation of the arch dam foundation surface, the geological information of the unfavorable structural surface to be treated is input into the knowledge graph model of treatment measures for the unfavorable structural surface of the arch dam abutment to predict the treatment measures that match the geological information.

[0088] Specifically, before the excavation of the arch dam foundation surface, by inputting the preliminary geological survey information, the knowledge map of treatment measures for the unfavorable structural surface of the arch dam abutment can be retrieved, queried and inferred, and the key unfavorable structural surface that affects the overall stability of the arch dam-foundation can be identified, and preliminary treatment measures can be recommended. Designers and geological surveyors can optimize the design plan of the arch dam foundation surface based on the given suggestions, and can repeat this step to continuously adjust the design plan.

[0089] Specifically, during the excavation of the arch dam foundation surface, as geological survey information becomes richer, more detailed geological information will be stored in the knowledge graph as new knowledge. Similarly, the treatment measures for unfavorable structural surfaces will also be updated to match the actual geological conditions in real time.

[0090] Example 2

[0091] Embodiment 2 of the present invention provides a terminal device corresponding to the above-mentioned embodiment 1. The terminal device may be a processing device for a client, such as a mobile phone, a laptop computer, a tablet computer, a desktop computer, etc., to execute the method of the above-mentioned embodiment.

[0092] The terminal device of the second embodiment includes a memory, a processor, and a computer program stored in the memory; the processor executes the computer program in the memory to implement the steps of the method of the first embodiment above.

[0093] In some implementations, the memory may be a high-speed random access memory (RAM), and may also include a non-volatile memory (non volatile memory), such as at least one disk memory.

[0094] In some other implementations, the processor may be a central processing unit (CPU), a digital signal processor (DSP), or other general-purpose processors of various types, which are not limited herein.

[0095] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, which all belong to the protection scope of the present invention.

[0096] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0097] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for recommending treatment measures for unfavorable structural surfaces of arch dam abutments based on knowledge graph, characterized in that: The following steps are involved: S1. Constructing the knowledge graph ontology model G1 of geological information of adverse structural surface of arch dam abutment; The expression of the geological information knowledge graph ontology model G1 of the arch dam abutment unfavorable structural surface is: G1=(C geo ,R geo ,P geo ,I geo ); Among them, C geo Represents concepts related to geological information; R geo Indicates the spatial position relationship between the unfavorable structural surface and the dam abutment; P geo Represents the specific attributes of various geological information related concepts; I geo Represents a specific instance of geological information; Construct the knowledge graph ontology model G2 of treatment measures for unfavorable structural surfaces of arch dam abutments; The expression of the knowledge graph ontology model G2 of the treatment measures for the adverse structural surface of the arch dam abutment is: G2=(C rein ,R rein ,P rein ,I rein ); Among them, C rein Represents the concept related to treatment measures; R rein Indicates the corresponding relationship between treatment measures and unfavorable structural surface; P rein Represents the specific attributes of the concepts related to each treatment measure; I rein Indicates a specific instance of a treatment measure; S2, extracting geological entity information related to the unfavorable structural surface, and improving the data layer of the geological information knowledge graph ontology model G1; Extract treatment measures for unfavorable structural surfaces and improve the data layer of the knowledge graph ontology model G2 of the treatment measures; S3. Establish the corresponding relationship between treatment measures and unfavorable structural surfaces, add corresponding treatment measure nodes in the geological information knowledge graph ontology model G1 after the data layer is improved, and couple the geological information knowledge graph ontology model G1 after adding the treatment measure nodes and the treatment measure knowledge graph ontology model G2 after the data layer is improved to form a knowledge graph model of treatment measures for unfavorable structural surfaces of arch dam abutments; S4. Using a distributed representation learning-based reasoning algorithm to improve the entity relationship of the knowledge graph model of the treatment measures for the adverse structural surface of the arch dam abutment; S5. Before and after the excavation of the arch dam foundation surface, the geological information of the unfavorable structural surface to be treated is input into the knowledge graph model of treatment measures for the unfavorable structural surface of the arch dam shoulder after the entity relationship is improved, so as to obtain treatment measures matching the geological information.

2. The method for recommending treatment measures for unfavorable structural surfaces of arch dam abutments based on knowledge graph according to claim 1 is characterized in that: In step S2, by analyzing the geological survey data in the early stage of the project, the "geological entity and relationship" information related to the unfavorable structural surface is extracted, and the geological entity information with the same meaning is fused through the K-means clustering algorithm, and the "geological entity and relationship" information obtained after fusion is stored in the Neo4j graph database, that is, the geological information knowledge graph ontology model G1 after the data layer is improved is obtained; wherein, the input of the K-means clustering algorithm is all unfavorable structural surfaces confirmed by geological exploration, including the spatial position of each unfavorable structural surface and rock mass parameter information, and the output is the merged geological entity information; In step S2, the "treatment measure entity and relationship" information related to the unfavorable structural surface is extracted, and the treatment measure entity information with the same meaning is fused through the K-means clustering algorithm, and the "treatment measure entity and relationship" information obtained after the fusion is stored in the Neo4j graph database, that is, the geological information knowledge graph ontology model G2 after the data layer is improved; wherein, the input of the K-means clustering algorithm is the extracted unfavorable structural surface treatment measures, and the output is the merged treatment measures.

3. The method for recommending treatment measures for unfavorable structural surfaces of arch dam abutments based on knowledge graph according to claim 1 is characterized in that: In step S4, the TransE algorithm is used to supplement and improve the triple information of the knowledge graph of treatment measures for the adverse structural surface of the arch dam abutment. The specific implementation process includes: Use the TransE algorithm to define a distance function; randomly replace the head entity or the tail entity in the knowledge graph model of treatment measures for adverse structural surfaces of the arch dam abutment to generate a damaged triplet; add the distance function and the damaged triplet to the loss function of the knowledge graph model of treatment measures for adverse structural surfaces of the arch dam abutment to perform knowledge reasoning training; The distance function calculation formula is: Among them, h, l, and t represent the vector representation of the head entity, relationship, and tail entity respectively. represents the L2 norm; The expression of the destroyed triple is: S′ (h，l，t) ={(h′,l,t)|h′∈E}∪{(h,l,t′)|t′∈E}; Wherein, (h′, l, t) and (h, l, t′) represent that the head entity or the tail entity is replaced by a random entity as a control group, and E represents the entity set; the calculation formula of the loss function is: Where (h′, l, t′) indicates that the head entity or the tail entity is replaced by a random entity, h+l indicates the addition of vector h to vector l. If the triple is correct, then h+l=t, h′+l indicates the addition of vector h′ to vector l, S indicates the uncorrupted triple, γ indicates the interval in the loss function, that is, the interval between the correct triple and the wrong triple, [x] + If it is greater than 0, it takes the original value; if it is less than 0, it takes 0.

4. A terminal device, characterized in that: include: one or more processors; A memory having one or more programs stored thereon, which, when the one or more programs are executed by the one or more processors, enables the one or more processors to implement the steps of the method according to any one of claims 1 to 3.

Images