CN113052968A - Knowledge graph construction method of three-dimensional structure geological model - Google Patents

Abstract

The invention discloses a method for constructing a knowledge graph of a three-dimensional structure geological model, which comprises the following steps: s1, establishing a knowledge graph mode layer; the method comprises the following substeps: s11, establishing a domain body on the basis of the top layer body; s12, designing a mode layer by using a modular architecture containing a plurality of sub-ontologies; s2, constructing a geological model knowledge graph data layer, recognizing input data of the structural modeling under the constraint of the mode layer, then forming triple fact expression, and finally forming an information network. The invention provides a method for efficiently organizing related information of a tectonic geological model, which is beneficial to combining the advantages of a computer in storage, calculation and three-dimensional vision and the advantages of human beings in abstract thinking, reasoning and learning so that more effective information can be utilized for tectonic modeling; the sharing and the communication of the related data among different organizations are facilitated.

Description

Knowledge graph construction method of three-dimensional structure geological model

Technical Field

The invention relates to a knowledge graph construction method of a three-dimensional structure geological model.

Background

The three-dimensional structure geological model is based on the result of seismic horizon interpretation, and forms a three-dimensional block geological framework model taking a geological surface (a general name of a section and a layer) as an inner boundary through space surface fitting, thereby providing a basic constraint framework for the sequence modeling, the lithology modeling and the oil reservoir modeling under the subsequent structural constraint; providing and constructing an initial model for seismic data inversion; in addition, the three-dimensional structure geological model is also an important supporting technology for formulating and optimizing an exploration and development scheme and well position deployment, and plays an important role in oil and gas resource exploration.

The process of building a tectonic geological model requires the integration of large amounts of geological, well logging, geophysical data and various interpretation results or conceptual models, all of which are currently stored in a distributed manner in a variety of ways. Most of the existing modeling processes are completed by geologists to integrate the knowledge and express the knowledge in a form that can be utilized by computers (namely, formal representation). However, with the rapid development of high-precision seismic exploration technology, data used for building a construction model increasingly show the characteristics of mass, multiple sources and heterogeneity, and it is increasingly difficult to organize and characterize information completely by manpower. This results in that the computer can only use partial data when building the model, and the utilization rate of the semi-structured and unstructured geological structure interpretation knowledge result is low, which directly affects the effectiveness of the modeling method and the quality of the model. For example, Lemon et al (2003) propose a method for constructing a simple model using only drilling data to directly generate a geological surface (Lemon a M, and Jones N l. building soluble models from wells and user-defined cross-sections J].Computers&Geosciens 2003,29(5), 547-555.; perrin et al (2013) use seismic horizon interpretation data for reconstructing a geological surface and geological event sequences for guiding mutual tailoring of geological surfaces in structural modeling (Perrin M, and Jean-

R.Shared earth modeling:knowledge driven solutions for building and managing subsurface 3D geological models[M]Edit Technip, 2013); wu (2017) then uses seismic stratigraphic feature to guideReconstruction of model interpolation (Wu X M. building 3D surface models forming to discrete structures and structural features [ J.].Geophysics,2017,82(3):IM21-IM30)。

If a three-dimensional tectonic geological model with high precision is built, and the time and manpower expenditure is controlled within a reasonable range, the method must be started from the organization and characterization method of knowledge. An efficient, accurate and flexible mode is found to transmit available information to a computer as much as possible, and theoretically, the efficiency, stability and accuracy of modeling can be improved remarkably.

The knowledge graph is proposed from 2012 and developed rapidly to date, and becomes one of the hot problems in the field of artificial intelligence, and obtains better landing effect in a series of practical applications, thereby generating huge social and economic benefits. A knowledge graph is a structured semantic knowledge base that describes entities and their interrelationships in the physical world in symbolic form. The entities are connected with each other through relationships to form a relationship network, the entities or concepts are represented by nodes, and the relationships are represented by connecting lines among the nodes. The basic composition unit of the knowledge graph is the fact expressed by such triples as (entity 1, relation, entity 2), (entity, attribute and attribute value), and the fact is also the basic operation unit of the knowledge graph. In the field of geosciences, existing research on a knowledge graph construction method mainly aims at unstructured text document data. For the fields of earth science, such as many elements, complex element relations and massive big data, the knowledge graph can well organize the geological data and provide the basis for further knowledge mining, but the knowledge graph construction research aiming at other types of geological data is still lacked. Compared with the field in which the existing knowledge graph is successfully constructed, the constructed geological model research has the characteristics of multi-source heterogeneity of data, uncertainty of data and priori knowledge, mutual influence of multiple elements and the like, and a set of new technical route and a technical framework suitable for the construction of the constructed geological model knowledge graph need to be established.

Disclosure of Invention

The invention aims to overcome the defects of the prior art, provides a method for efficiently organizing related information of a constructed geological model, and is beneficial to combining the advantages of a computer in storage, calculation and three-dimensional vision and the advantages of human beings in abstract thinking, reasoning and learning, so that the construction modeling can utilize more effective information.

The purpose of the invention is realized by the following technical scheme: the method for constructing the knowledge graph of the three-dimensional structure geological model comprises the following steps:

s1, establishing a knowledge graph mode layer as a constraint of knowledge graph entities, attributes and relationship types; the method specifically comprises the following substeps:

s1, establishing a knowledge graph mode layer; the method specifically comprises the following substeps:

s11, establishing a domain body on the basis of the top layer body;

s12, designing a mode layer by using a modular architecture containing a plurality of sub-ontologies;

s2, constructing a geological model knowledge graph data layer, recognizing input data of the structural modeling under the constraint of the mode layer, then forming triple fact expression, and finally forming an information network.

Further, the specific implementation method of step S11 is as follows: categorizing the ontology according to the prevalence of the knowledge characterized by the ontology: top-level ontologies, also called upper-level ontologies or base ontologies, describe common general concepts unrelated to the field, including space and time; a domain ontology describing concepts and relationships between concepts in a particular domain; the DOLCE is selected as a top layer body for constructing the model body, and a domain body in a knowledge graph mode layer is designed according to the DOLCE;

two basic ontological classes of DOLCE are persistence and continuum;

in constructing the ontology of the geological model, the geological object is a persistent physical entity;

a typical representation of a continuum entity is a geological process;

in DOLCE, the relationship between persistence and continuants is involved: a persistence being present in time by participating in a continuum; the geological objects and geological processes inherit this relationship, and any geological object is associated with at least one geological process;

the DOLCE also includes the following ontology classes:

a geological feature, which refers to a class of entities that are parasitic to a geological object, the ontology class of the geological feature being a feature;

regarding the geological structure as being composed of a plurality of geological objects, wherein the body categories corresponding to geological structure entities are arbitrary sums, and a simple collection of different types of entities is represented in DOLCE;

yet another important basic category in DOLCE is attributes, which are entities that can be perceived or measured, attribute class entities must depend on other entities and exist or disappear simultaneously with the entities they depend on;

the attitude is a physical attribute entity depending on the geological object, and is used for describing the spatial state of the geological object, and the attitude entity is related to a geological space region entity of which the ontology type is a physical region:

in DOLCE, abstract attributes are attributes which are neither time nor physical, and the entities in the class are geological relations in a geological model; the geological relationship depends on two geological objects for describing their geological contact relationship.

Further, the specific implementation method of step S12 is as follows: the geological model body is constructed by 9 modules in total: geological features, geological objects, geological structures, geological processes, geological time, attitude, geological relationships, geological time periods, and geological spatial regions; each module corresponds to a sub-ontology, and the sub-ontology contains a group of related concepts in the form of UML classes.

Further, the step S2 includes the following sub-steps:

s21, detecting whether the geological interfaces intersect by adopting a trend surface + bounding box collision method, if so, executing a step S22, otherwise, ending the operation;

s22, reasoning according to the interface intersection condition to obtain other geological objects: two interfaces are intersected to obtain a line entity, every two of the three interfaces are intersected to obtain a vertex entity, and the interfaces are intersected to form a closed body entity;

s23, obtaining a space geometric topological relation of the geological objects according to the shared boundary among the geological objects;

s24, obtaining an entity geological contact relation according to the type of the geological curved surface to which the interface belongs;

and S25, deducing the time relation generated by the geological object according to the time meaning of the geological contact relation.

Further, the specific implementation method of step S21 is as follows:

when judging whether the interfaces are intersected, firstly judging whether the trend surfaces are intersected or not, if so, further detecting whether the bounding boxes are intersected or not, and if not, judging that the interfaces are not intersected; the trend surface is calculated by a Random Sample Consensus algorithm;

when the bounding box is used for describing interface data, an octree partitioning algorithm is adopted for realizing: using a maximum bounding box as a root node of the octree, uniformly dividing all the discrete points into 8 regions by using the X, Y, Z direction as a uniform dividing plane, and calculating the bounding boxes respectively as middle nodes of the layer 1 in the octree; then, each area in the layer 1 is evenly divided into 8 areas; sequentially subdividing until points contained in all intermediate nodes are less than or equal to octadivisions; the intersection detection firstly judges whether the root nodes of the two interfaces are intersected or not, and judges whether the root nodes are intersected with the child nodes of the root nodes or not if the root nodes are intersected with the child nodes of the root nodes or not, and the detection is finished until the leaf nodes are detected or the non-intersection is detected.

The invention has the beneficial effects that:

1. the invention provides a method for efficiently organizing related information of a tectonic geological model, which is beneficial to combining the advantages of a computer in storage, calculation and three-dimensional vision and the advantages of human beings in abstract thinking, reasoning and learning so that more effective information can be utilized for tectonic modeling;

2. the invention extends the application of the knowledge graph in the field of geoscience, and the same knowledge graph construction idea can also be used for organizing seismic data, logging data and the like;

3. the invention provides a formalized characterization method of knowledge contained in a structural geological model, which is beneficial to sharing and communication of related data among different mechanisms;

4. the invention provides a cognitive method for constructing geological model data, and the process of constructing a knowledge graph is a process of knowledge mining in the data.

Drawings

FIG. 1 is a technical framework diagram of the construction of a three-dimensional tectonic geological model knowledge map;

FIG. 2 is a flow chart of a method of constructing a knowledge base of a three-dimensional tectonic geological model according to the invention;

FIG. 3 illustrates ontology classes and hierarchies defined in DOLCE;

FIG. 4 is a UML package representation of a body module;

FIG. 5 is geological interface point cloud data, (a) is a trend surface; (b) - (f) is an octree-based hierarchical bounding box;

FIG. 6 is the raw data of the geological model constructed according to the present embodiment;

FIG. 7 is a knowledge graph of the geological model constructed according to the present embodiment.

Detailed Description

The knowledge graph can be logically divided into a mode layer and a data layer, and the data layer mainly stores facts formed by entities, attributes and relations in specific examples. The mode layer is arranged above the data layer, and mainly stores prior domain knowledge, and a series of factual expressions of the data layer are specified through the form of an ontology. The ontology and the entities constrained by the ontology and their interrelationships can therefore be considered to constitute a knowledge-graph. In the traditional knowledge graph construction process, two ways of constructing the knowledge graph are formed according to whether a data model (resource mode) is determined firstly or example data is collected firstly: top down and bottom up. The top-down construction mode is that a mode layer of the knowledge graph is determined according to human understanding and cognition on problems, and specific example data are filled according to the mode layer; the bottom-up construction mode is that specific data are collected according to a fact triple mode, and then the data are sorted, analyzed, summarized and summarized to form a mode layer. Because the knowledge map of the geological model not only contains general knowledge, but also has the specific knowledge of different work areas, and in addition, the knowledge of human beings on the underground geological structure is insufficient, the research requirements are difficult to meet by only adopting any one of the construction modes. The invention provides a knowledge graph construction mode combining top-down and bottom-up aiming at the characteristics of a constructed geological model. At the initial stage of establishing the knowledge map, establishing a structural geological model body according to the prior knowledge of human about geological structure distribution and form, forming an initial mode layer, identifying structural element entities, relations and attributes of the model under the guidance of the initial mode layer, and gradually establishing a data layer. With the continuous accumulation of the data volume of the data layer, frequent substructure analysis can be performed. At this stage, we may find the absence of instances, relationships, and flaws in the data layer of the knowledge-graph due to the uncertainty of the input data, and may also find the imperfections of the original schema layer, and new knowledge may not be included in the original schema layer, at which time the schema layer needs to be revised and supplemented. The process of constructing a knowledge-graph in a top-down and bottom-up combination is essentially an iterative update process. The specific schematic diagram is shown in fig. 1.

The invention designs a knowledge graph framework of a three-dimensional structure geological model and a construction method. Firstly, designing a knowledge graph construction technical framework combining top-down and bottom-up according to the characteristic that prior knowledge and input data of a constructed geological model have uncertainty. Secondly, a framework for constructing a geological model knowledge graph mode layer is provided, a modular ontology library form is adopted to represent a mode layer ontology, and a top-layer ontology DOLCE is used as a constraint in the process of establishing the ontology, so that the normalization of constructing the model ontology is enhanced. And finally, establishing a knowledge graph of the actual constructed geological model under the guidance of the mode layer.

The technical scheme of the invention is further explained by combining the attached drawings.

As shown in fig. 2, the method for constructing a knowledge graph of a three-dimensional geological model of the invention comprises the following steps:

s1, establishing a knowledge graph mode layer; as constraints on knowledge graph entities, attributes and relationship types;

the purpose of establishing the knowledge-graph mode layer is to determine the scope of knowledge of the constructed geological model which can be described by the knowledge graph. Specifically, an ontology of the construction model is established as a constraint of knowledge-graph entities, attributes and relationship types. Considering that the concept related to the geological model basically covers the whole geological domain of the structure, including geological objects, geological processes, geological time, geological positions and other categories, it is difficult to directly establish a single huge ontology. To this end, the present invention will solve this problem in two ways:

s11, establishing a domain body on the basis of the top layer body; the use of top-level ontologies enables us to leverage defined general concepts, such as objects, procedures, properties, etc., and can inherit axioms in top-level ontologies as constraints in domain ontologies.

The specific implementation method of step S11 is:

categorizing the ontology according to the prevalence of the knowledge characterized by the ontology: top-level ontologies, also called upper-level ontologies or base ontologies, describe common general concepts unrelated to the field, including space and time; the domain ontology is a professional ontology and describes concepts and relations among the concepts in a specific domain; the method selects DOLCE (Descriptive Ontology for Linguistic and Cognitive Engineering description type Ontology) as a top-layer Ontology for constructing a model Ontology, and designs a domain Ontology in a knowledge map mode layer according to the DOLCE; figure 3 illustrates some basic ontology classes in DOLCE.

Two basic ontology classes of DOLCE are continuant (endrun) and continuum (Perdurant);

in constructing the ontology of the Geological Model (GM), the Geological Object (GO) is a persistent class entity; a geological object, such as a rock mass, if segmented by faults or partially eroded, retains its identity unless it disappears completely. The geological object may be assigned to the ontology class "physical object" (POB). Here, we denote CA (x, y) as the relationship "ontology class of entity x is y":

typical representatives of continuous physical entities are Geological Processes (GP), such as folds, fractures, deposits, erosion, etc.; geological processes may be assigned to the ontology class "process" (PRO):

in DOLCE, the relationship between persistence and continuants is involved: a persistence being present in time by participating in a continuum; geological objects and geological processes therefore also inherit this relationship, any geological object being related to at least one geological process, denoted PC (x, y, t) by "object x participates in process y during time t" relationship:

the DOLCE also includes the following ontology classes:

geological Features (GF) in geological models refer to a class of entities that are parasitic to geological objects, such as holes in the geologic body and the interior of a geological surface, bumps on a geological surface, and the like. The ontology class of geological features is feature (F); "entity x depends on entity y" relationship is represented by I (x, y):

in addition to geological objects, we also consider the concept of Geological Structures (GS). A geological structure is considered to be composed of a plurality of geological objects, for example a fractured stratum is composed of a fault plane and a plurality of bodies. The ontology class corresponding to the geological structure entity is an Arbitrary Sum (AS), and a pure collection of different classes of entities is represented in the DOLCE; by K (x, y) is meant "x is composed of y":

yet another important basic category in DOLCE is attribute (Q), which is the entity that can be perceived or measured, e.g. size of the object, time of occurrence of the event. Attribute class entities must depend on other entities and exist or disappear simultaneously with the entities on which they depend; each attribute entity is also associated with a "region" (R) entity, representing a domain of attribute values. Geologic Time (GT) is a class of "time attribute" (TQ) entities that depend on geologic processes, and is associated with a geologic time period (GTI) entity whose ontology class is "time region" (TR):

similarly, the attitude (O) is a physical attribute (PQ) entity depending on the geological object, and is used to describe the spatial state of the geological object, and the attitude entity is related to a Geological Spatial Region (GSR) entity whose ontology class is Physical Region (PR):

in DOLCE, abstract Attributes (AQ) refer to attributes that are neither temporal nor physical, and this class of entities is the Geological Relationships (GR) in the geological model; the geological relationship depends on two geological objects for describing their geological contact relationship.

S12, designing a mode layer by using a modular architecture containing a plurality of sub-ontologies; the plurality of sub-ontologies are combined to form an ontology library of the construction model. The modular architecture can simplify the modeling and maintenance of the ontology and ensure future expandability; the specific implementation method comprises the following steps: the geological model body is constructed by 9 modules in total: geological features (GeoFeature), geological objects (GeoObject), geological structures (GeoStructure), geological processes (GeoProcess), geological time (GeoTime), birth (Occurrence), geological relationships (georelationship), geological time period (GeoTimeInterval), and geospatial regions (GeoSpaceRegion); they cover important information of interest to general experts in geological models. We represent the structure of the geological model ontology in the form of a UML package, where geological objects are the most core modules. The relationships between modules indicate that one module may invoke concepts in other modules (as shown in FIG. 4).

Each module corresponds to a sub-ontology, and the sub-ontology contains a group of related concepts in the form of UML classes. For example, a geological object module will have multiple classes, including body (body), interface (interface), line (edge), and vertex (vertex); a class will have multiple attributes such as name, type, whether there are holes, etc. of the interface class and will be associated with other classes. Some modules may reuse existing ontologies, for example, the geologic structure and geologic features module uses the Structural Geology Ontology proposed by Babaie (2006) (Babaie H A. design a Modular Architecture for the Structural geomology Ontology [ M ]. Geoinformatics: Data to knowledge.2006), the geologic time and geologic time period module may use the geologic time formalized characterization proposed by Perrin et al (2011) (Perrin M, Mastella L S, Morel O, et al. Geological time formalization: an improved Structural model for the description time details and the for the third correlation [ J ]. Earth Science information, 2011,4(2): 81-96).

S2, constructing a geological model knowledge graph data layer, recognizing input data of the structural modeling under the constraint of the mode layer, then forming triple fact expression, and finally forming an information network.

The knowledge graph mode layer is used for representing domain knowledge, and the knowledge graph data layer is used for describing information contained in specific example data. We have mentioned previously that the information in the data layer is primarily represented by the properties and relationships of the geological objects. Due to the diversity of geological objects and relations, the data layer of the construction model knowledge graph is regarded as a heterogeneous information network, and nodes and edges of the network can represent different types of entities and relations. Specifically, the nodes are the aforementioned volumes, interfaces, edges and vertices in the model, and the edges may represent spatial relationships (describing spatial geometric topological relationships and geological contact relationships between objects), temporal relationships (describing geological chronological relationships generated between geological objects) and composition relationships (describing boundaries describing target objects surrounded by source objects) of the geological objects according to application requirements. The knowledge graph data layer is established by recognizing input data for constructing a model under the constraint of the mode layer, then forming triple factual expression, and finally forming an information network. The essence of the method is a process of extracting semantic information from newly input data under the guidance of a priori knowledge. Here, there are two ways to establish a data layer network: computer auto-extraction and expert input. Due to data uncertainty, particularly seismic signal reflection clutter caused by fractured zones near faults and the absence of structural interpretation data, automatic extraction faces the challenges of ambiguity and uncertainty reasoning. Relying entirely on expert input faces the labor and time overhead of manually entering hundreds of nodes and edges. Therefore, the knowledge graph data layer is established by automatically extracting the preliminary result and then manually editing and correcting.

Step S2 includes the following substeps:

s21, detecting whether the geological interfaces intersect by adopting a trend surface + bounding box collision method, if so, executing a step S22, otherwise, ending the operation;

the specific implementation method comprises the following steps:

when judging whether the interfaces are intersected, firstly judging whether the trend surfaces are intersected or not, if so, further detecting whether the bounding boxes are intersected or not, and if not, judging that the interfaces are not intersected; the trend surface is calculated by Random Sample Consensus algorithm (RANSAC), as shown in FIG. 5 (a);

when the bounding box is used for describing interface data, an octree partitioning algorithm is adopted for realizing: using a maximum bounding box as a root node of the octree, as shown in fig. 5(b), dividing all the points into 8 regions uniformly by using the X, Y, Z direction as a uniform plane, and calculating bounding boxes as middle nodes of layer 1 in the octree, as shown in fig. 5 (c); then, each area in the layer 1 is evenly divided into 8 areas; the nodes are sequentially divided until points contained in all the intermediate nodes are equal to or less than octants, as shown in fig. 5(d) - (f); the intersection detection firstly judges whether the root nodes of the two interfaces are intersected or not, and judges whether the root nodes are intersected with the child nodes of the root nodes or not if the root nodes are intersected with the child nodes of the root nodes or not, and the detection is finished until the leaf nodes are detected or the non-intersection is detected.

S22, reasoning according to the interface intersection condition to obtain other geological objects: two interfaces are intersected to obtain a line entity, every two of the three interfaces are intersected to obtain a vertex entity, and the interfaces are intersected to form a closed body entity;

s23, obtaining a space geometric topological relation of the geological objects according to the shared boundary among the geological objects;

s24, obtaining an entity geological contact relation according to the type of the geological curved surface to which the interface belongs;

and S25, deducing the time relation generated by the geological object according to the time meaning of the geological contact relation.

Experimental test environment: the experimental computer in this embodiment is configured as an Inter i5-8500 CPU, 8GRAM, 64-bit Windows 10 operating system; the software used was MATLAB R2016 and Gephi 0.8.1 beta. The actual tectonic geological model data comes from a certain oil and gas production work area in the Sichuan basin. The underlying data contained 15 sections, constructed primarily as 4 formation levels (T1, T2, T3, T4), 5 reverse faults (F1, F2, F3, F4, F5), with the F1 fault not passing through the entire model, as shown in fig. 6. The corresponding knowledge map contains 15 geologic bodies, 29 geologic surfaces and 102 boundary lines, as shown in fig. 7.

It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.

Claims (5)

1. The method for constructing the knowledge graph of the three-dimensional structure geological model is characterized by comprising the following steps of:

s1, establishing a knowledge graph mode layer; the method specifically comprises the following substeps:

s11, establishing a domain body on the basis of the top layer body;

s12, designing a mode layer by using a modular architecture containing a plurality of sub-ontologies;

s2, constructing a geological model knowledge graph data layer, recognizing input data of the structural modeling under the constraint of the mode layer, then forming triple fact expression, and finally forming an information network.

2. The method for constructing a knowledge graph of a three-dimensional geological model according to claim 1, wherein the step S11 is implemented by: categorizing the ontology according to the prevalence of the knowledge characterized by the ontology: top-level ontologies, also called upper-level ontologies or base ontologies, describe common general concepts unrelated to the field, including space and time; a domain ontology describing concepts and relationships between concepts in a particular domain; the DOLCE is selected as a top layer body for constructing the model body, and a domain body in a knowledge graph mode layer is designed according to the DOLCE;

two basic ontological classes of DOLCE are persistence and continuum;

in constructing the ontology of the geological model, the geological object is a persistent physical entity;

a typical representation of a continuum entity is a geological process;

in DOLCE, the relationship between persistence and continuants is involved: a persistence being present in time by participating in a continuum; the geological objects and geological processes inherit this relationship, and any geological object is associated with at least one geological process;

the DOLCE also includes the following ontology classes:

a geological feature, which refers to a class of entities that are parasitic to a geological object, the ontology class of the geological feature being a feature;

regarding the geological structure as being composed of a plurality of geological objects, wherein the body categories corresponding to geological structure entities are arbitrary sums, and a simple collection of different types of entities is represented in DOLCE;

yet another important basic category in DOLCE is attributes, which are entities that can be perceived or measured, attribute class entities must depend on other entities and exist or disappear simultaneously with the entities they depend on;

the attitude is a physical attribute entity depending on the geological object, and is used for describing the spatial state of the geological object, and the attitude entity is related to a geological space region entity of which the ontology type is a physical region:

in DOLCE, abstract attributes are attributes which are neither time nor physical, and the entities in the class are geological relations in a geological model; the geological relationship depends on two geological objects for describing their geological contact relationship.

3. The method for constructing a knowledge graph of a three-dimensional geological model according to claim 1, wherein the step S12 is implemented by: the geological model body is constructed by 9 modules in total: geological features, geological objects, geological structures, geological processes, geological time, attitude, geological relationships, geological time periods, and geological spatial regions; each module corresponds to a sub-ontology, and the sub-ontology contains a group of related concepts in the form of UML classes.

4. The method of constructing a knowledge graph of a three-dimensional geological model according to claim 1, wherein said step S2 comprises the substeps of:

s21, detecting whether the geological interfaces intersect by adopting a trend surface + bounding box collision method, if so, executing a step S22, otherwise, ending the operation;

s22, reasoning according to the interface intersection condition to obtain other geological objects: two interfaces are intersected to obtain a line entity, every two of the three interfaces are intersected to obtain a vertex entity, and the interfaces are intersected to form a closed body entity;

s23, obtaining a space geometric topological relation of the geological objects according to the shared boundary among the geological objects;

s24, obtaining an entity geological contact relation according to the type of the geological curved surface to which the interface belongs;

and S25, deducing the time relation generated by the geological object according to the time meaning of the geological contact relation.

5. The method for constructing a knowledge graph of a three-dimensional geological model according to claim 4, wherein the step S21 is implemented by the following steps:

when judging whether the interfaces are intersected, firstly judging whether the trend surfaces are intersected or not, if so, further detecting whether the bounding boxes are intersected or not, and if not, judging that the interfaces are not intersected; the trend surface is calculated by a Random Sample Consensus algorithm;

when the bounding box is used for describing interface data, an octree partitioning algorithm is adopted for realizing: using a maximum bounding box as a root node of the octree, uniformly dividing all the discrete points into 8 regions by using the X, Y, Z direction as a uniform dividing plane, and calculating the bounding boxes respectively as middle nodes of the layer 1 in the octree; then, each area in the layer 1 is evenly divided into 8 areas; sequentially subdividing until points contained in all intermediate nodes are less than or equal to octadivisions; the intersection detection firstly judges whether the root nodes of the two interfaces are intersected or not, and judges whether the root nodes are intersected with the child nodes of the root nodes or not if the root nodes are intersected with the child nodes of the root nodes or not, and the detection is finished until the leaf nodes are detected or the non-intersection is detected.

Images