Regional geology automatic analysis method and system fusing multi-source knowledge relationship constraint

Abstract

The invention relates to the technical field of oil exploration data information, in particular to a regional geology automatic analysis method and system fusing multi-source knowledge relationship constraints, and the method comprises the following steps: S1, constructing the constraints of a multi-source knowledge relationship: obtaining regional geology data; preprocessing the acquired regional geological data; quantitatively expressing the preprocessed geological data by using structured knowledge; fusing the data of the knowledge quantitative representation into the same knowledge graph framework to obtain a multi-source knowledge graph framework; constructing a regional geological data pattern recognition model, obtaining an implicit relationship among the geological data, and adding the obtained implicit relationship among the geological data to a multi-source knowledge graph framework to obtain a multi-source knowledge graph; s2, fusing constraints of the multi-source knowledge relationship established in the step S1, and analyzing regional geologic features; quantitative analysis of geological features is achieved, and interconnection and intercommunication of geological data in the oil-gas exploration and development process are achieved.

Description

Technical Field

[0001] The present invention relates to the field of petroleum exploration data information technology, and in particular to a regional geological automatic analysis method and system integrating multi-source knowledge relationship constraints. Background Art

[0002] With the continuous development of informatization in oil and gas exploration and development, various types of data, formulas, charts, documents and other data information involved in the exploration and development process are scattered among various disciplines. The production, scientific research and management work data information between each discipline are independent of each other and interdependent, forming the business process of the entire industrial chain of oil exploration and development. It is difficult to extract relevant information and knowledge elements between various disciplines and to build related services, which leads to the current situation of difficulty in regional geological analysis during oil and gas exploration and development.

[0003] In existing technologies, technicians often build a hierarchical regional geological analysis ontology framework based on the business needs of existing professional databases, application software, and core exploration and development business models in the exploration and development field. Using business models and data models as a reference system, they implement automated analysis based on these models. However, further research revealed that regional geological analysis in the current oil and gas exploration and development process is completed by dynamic business activities under multiple business operations. Existing technologies are unable to fully describe these dynamic business activities, making it impossible to automatically analyze regional geological conditions during oil and gas exploration and development.

[0004] Therefore, there is an urgent need to provide a regional geological automatic analysis method and system that integrates multi-source knowledge relationship constraints, so as to achieve quantitative analysis of geological characteristics and realize the interconnection and interoperability of geological data in the process of oil and gas exploration and development compared with existing technologies. Summary of the Invention

[0005] The present invention solves the technical problems existing in the prior art and provides a regional geological automatic analysis method and system integrating multi-source knowledge relationship constraints.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] The automatic regional geological analysis method integrating multi-source knowledge relationship constraints includes the following steps:

[0008] S1. Constructing constraints on multi-source knowledge relationships, specifically including the following steps:

[0009] S101. Acquire regional geological data;

[0010] S102, preprocessing the acquired regional geological data;

[0011] S103, quantitatively representing the pre-processed geological data using structured knowledge;

[0012] S104, integrating the data processed in step S103 into the same knowledge graph framework to obtain a multi-source knowledge graph framework;

[0013] S105. Construct a regional geological data pattern recognition model to obtain implicit relationships between geological data, and add the implicit relationships between the obtained geological data to the multi-source knowledge graph framework obtained in S104 to obtain a multi-source knowledge graph, which is a constraint of the constructed multi-source knowledge relationship.

[0014] S2. Integrate the constraints of the multi-source knowledge relationships established in step S1 to analyze the regional geological characteristics.

[0015] Furthermore, S102 specifically includes the following steps:

[0016] S1021. Perform data redundancy check and data quality check;

[0017] S1022. Construct a geological ontology and unify geological terminology based on the constructed geological ontology;

[0018] S1023, unifying the format of the data processed in step S1022;

[0019] S1024, labeling and classifying the data processed in step S1023, and labeling different types of data according to the classification in the geological ontology;

[0020] S1025. Perform data association on the data processed in step S1024 to build potential connections and relationship strengths between data from different data sources.

[0021] Furthermore, the geological ontology includes a plurality of geological common concepts and their attributes and the relationships between the geological common concepts, and the geological common concepts and their attributes are all standard terms.

[0022] Furthermore, the specific method for unifying geological terminology based on geological ontology in S1022 is as follows:

[0023] (1) Using natural language processing technology to scan the data processed in step S1021 and extract geological terms from the data;

[0024] (2) Compare the extracted geological terms with the standard terms in the geological ontology, identify inconsistent or erroneous terms in the data, and replace the inconsistent or erroneous data to make the terms in the data consistent with the terms in the geological ontology.

[0025] Furthermore, the specific replacement method in step (2) of S1022 is: when an inconsistent or erroneous term is identified, the first replacement method is executed, and the second replacement method is executed at the same time.

[0026] Furthermore, the first replacement method is: using the relationship between geological ontology standard terms to understand the contextual semantics of inconsistent or erroneous terms in the data, thereby determining the most suitable standard terms for the inconsistent or erroneous terms in the data, and replacing the inconsistent or erroneous terms in the data with the determined most suitable standard terms.

[0027] Furthermore, the second replacement method is: for data in different data sources that are subsequently processed, when there is a term identical to the term replaced by the first replacement method, the term is replaced by the most suitable standard term determined by the first replacement method.

[0028] Furthermore, S1025 specifically includes the following steps:

[0029] (1) Using random forest algorithm to mine and analyze the potential connections between data from different data sources;

[0030] (2) Use the Pearson correlation coefficient to quantify the strength of the relationship between different data sources.

[0031] Furthermore, the specific method of performing data redundancy check in S1021 is: verifying and removing duplicate data through unique constraint judgment.

[0032] Furthermore, the specific method of performing data quality check in S1021 is: identifying erroneous data through data quality rules, and filtering missing data records through data integrity check.

[0033] Furthermore, the specific method of step S103 is: using a graph embedding algorithm and a graph self-attention network to learn the structural features in the geological data, thereby mapping the entities and the relationships between entities in the geological data to a low-dimensional space.

[0034] Furthermore, mapping entities and relationships between entities in geological data into low-dimensional space is specifically expressed by the following formula:

[0035] E i =f(V,D,R)

[0036] In the above formula, E i represents the embedding vector of the i-th entity in the low-dimensional space, V represents the set of entities in the graph, D represents the set of relationships between entities, and R represents the set of various relationship types between entities.

[0037] Furthermore, S104 specifically includes the following steps:

[0038] S1041. Obtain a knowledge graph framework;

[0039] S1042. Standardize the temporal and spatial attributes of geological data;

[0040] S1043. Use ontology alignment technology to identify the correspondence between different source entities and geological ontology concepts;

[0041] S1044. Determine the optimal mapping relationship between the entity and the geological ontology using the cosine similarity semantic similarity measurement method;

[0042] S1045. Using an ontology reasoning engine to ensure that the semantics of the mapped data remain consistent, thereby correcting the mapping relationship between the geological data and the corresponding concepts of the geological ontology, and forming a mapping relationship set between the geological data entities and the corresponding concepts of the geological ontology;

[0043] S1046. Combine the mapping relationship set between the geological data entities and the corresponding concepts of the geological ontology obtained in S1045 with the knowledge graph framework obtained in S1041 to obtain a multi-source knowledge graph framework.

[0044] Furthermore, the implicit relationship between the geological data in step S105 is specifically expressed by the following formula:

[0045] R ij =g(E i ,E j )

[0046] In the above formula, E i represents the embedding vector of the i-th entity in the low-dimensional space, E j represents the embedding vector of the j-th entity in the low-dimensional space, R ij represents the relationship between the i-th entity and the j-th entity, and g represents the function used to calculate the attribute relationship between the two entities.

[0047] Furthermore, S2 specifically includes the following steps:

[0048] S201, using community detection algorithms to identify patterns and groups in geological features;

[0049] S202, using a clustering algorithm to perform cluster analysis on the patterns and groups identified in step S201, and extract key features from the data;

[0050] S203, using a machine learning model, taking the key features extracted in S202 as input, and patterns and groups in the geological features as labels, to classify and predict the geological features and build an analysis model;

[0051] S204. Optimize the analysis model.

[0052] Furthermore, the regional geological automatic analysis method also includes step S3, and the specific content of step S3 is: applying the regional geological characteristics analysis results, and displaying the analysis results through a visual interface in the form of maps, charts and reports.

[0053] A system using any of the above-mentioned regional geological automatic analysis methods integrating multi-source knowledge relationship constraints includes a first module, a second module, and a third module connected in sequence, wherein the first module is used to execute the content in step S1, the second module is used to execute the content in step S2, and the third module is used to execute the content in step S3.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] (1) The present invention automatically analyzes regional geology by integrating multi-source knowledge relationship constraints, namely multi-source knowledge graphs, thereby achieving the interconnection, results sharing, experience inheritance, and joint collaboration of regional geological data, and providing technical support for solving the current difficulties in analyzing regional geology during oil and gas exploration and development.

[0056] (2) The present invention obtains a multi-source knowledge graph by constructing a geological ontology and a machine learning algorithm, making the obtained multi-source knowledge graph more accurate and quantifying the relationship between different source entities, thereby facilitating the quantitative analysis of geological characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0058] The technical solution of the present invention will be clearly described below in conjunction with the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0059] like Figure 1 As shown, the present invention provides a regional geological automatic analysis method integrating multi-source knowledge relationship constraints, comprising the following steps:

[0060] S1. Constructing constraints on multi-source knowledge relationships, specifically including the following steps:

[0061] S101. Acquire regional geological data. Regional geological data includes entities from multiple different sources, including but not limited to geological maps, drilling data, seismic data, remote sensing images, mineral databases, and historical geological reports.

[0062] S102: Preprocess the acquired regional address data, specifically including the following steps:

[0063] S1021. Perform data redundancy check and data quality check. The specific method for data redundancy check is: verify and remove duplicate data through unique constraint judgment; the specific method for data quality check is: identify erroneous data through data quality rules, and filter missing data records through data integrity check.

[0064] S1022. Construct a geological ontology and unify geological terminology based on the constructed geological ontology.

[0065] The geological ontology includes common geological concepts such as structure, sedimentary pattern, lithology, mineralogy, and their attributes, as well as the relationships between the above common concepts. Common geological concepts and their attributes are all standard terms.

[0066] The specific method for unifying geological terms according to geological ontology is:

[0067] (1) Using natural language processing technology to scan the data processed in step S1021 and extract geological terms from the data;

[0068] (2) Compare the extracted geological terms with the standard terms in the geological ontology, identify inconsistent or erroneous terms in the data, and replace the inconsistent or erroneous data to make the terms in the data consistent with the terms in the geological ontology.

[0069] The specific replacement method is: when inconsistent or incorrect terms are identified, the first replacement method is executed, and the second replacement method is executed at the same time. The first replacement method is: using the relationship between the standard terms of the geological ontology to understand the contextual semantics of the inconsistent or incorrect terms in the identified data, thereby determining the most suitable standard terms for the inconsistent or incorrect terms in the data, and replacing the inconsistent or incorrect terms in the data with the determined most suitable standard terms. The second replacement method is: replacing the above-mentioned identical terms in the data of different data sources for subsequent processing, so that the different names for the same geological structure in different data sources are unified.

[0070] S1023. Unify the format of the data processed in step S1022, and convert all data into the same format to facilitate subsequent processing, for example, converting the coordinate systems of all geological maps into a universal geographic coordinate system.

[0071] S1024. Label and classify the data processed in step S1023. Label different types of data according to the classification in the geological ontology. For example, classify the drilling data according to rock type, age, and stratum. That is, label the drilling data into three categories: the first category is rock type, the second category is age, and the third category is stratum.

[0072] S1025. Perform data association on the data processed in step S1024 to build potential connections and relationship strengths between data from different data sources. Specifically:

[0073] (1) Use the random forest algorithm to mine and analyze the potential connections between data from different data sources.

[0074] (2) Use the Pearson correlation coefficient to quantify the strength of the relationship between different data sources.

[0075] S103. Convert complex, multi-source geological data into a structured quantitative representation of knowledge. The specific method is to use graph embedding algorithms and graph attention networks to learn the structural features of geological data, thereby mapping the entities and relationships in the geological data into a low-dimensional space. The specific conversion is performed using the following formula:

[0076] E i =f(V,D,R)

[0077] In the above formula, E i represents the embedding vector of the i-th entity in the low-dimensional space, V represents the set of entities in the graph, D represents the set of relationships between entities, and R represents the set of various relationship types between entities.

[0078] S104: Fusing the data processed in step S103 and unifying them into the same knowledge graph framework, specifically including the following steps:

[0079] S1041. Obtain a knowledge graph framework;

[0080] S1042. Standardize the temporal and spatial attributes of geological data to ensure the consistency of all geological data in time series and spatial position. Specifically, the temporal data is standardized using a unified time resolution, and the spatial data is standardized using standard deviation normalization.

[0081] S1043. Use ontology alignment technology to identify the correspondence between different source entities and geological ontology concepts;

[0082] S1044. Using a cosine similarity semantic similarity measurement method to determine the optimal mapping relationship between the entity and the geological ontology;

[0083] S1045. Using an ontology reasoning engine to ensure that the semantics of the mapped data remain consistent, thereby correcting the mapping relationship between the geological data and the corresponding concepts of the geological ontology, and forming a mapping relationship set between the geological data entities and the corresponding concepts of the geological ontology;

[0084] S1046. Combine the mapping relationship set between the geological data entities obtained in 1045 and the corresponding concepts of the geological ontology with the knowledge graph framework obtained in S1041 to obtain a unified multi-source knowledge graph framework that integrates geological data from different sources.

[0085] S105. Construct a regional geological data pattern recognition model, obtain the implicit relationship between geological data entities, and construct a multi-source knowledge graph.

[0086] Specifically, a regional geological data pattern recognition model is constructed by utilizing machine learning algorithms and artificial intelligence technologies, and is expressed as follows:

[0087] R ij =g(E i ,E j )

[0088] In the above formula, E i represents the embedding vector of the i-th entity in the low-dimensional space, E j represents the embedding vector of the j-th entity in the low-dimensional space, R ij represents the relationship between the i-th entity and the j-th entity, and g represents the function used to calculate the attribute relationship between the two entities.

[0089] The regional geological data pattern recognition model is mainly used to quantify the relationship strength or type between two entities, and to obtain a multi-source knowledge graph by adding the quantified relationship strength or type between the entities to the multi-source knowledge graph framework obtained in S104.

[0090] The multi-source knowledge graph is the constraint of the constructed multi-source knowledge relationship.

[0091] S2: Integrate the multi-source knowledge graph established in step S1 to analyze the regional geological characteristics, which specifically includes the following steps:

[0092] S201. Identify patterns and groups in geological features using a community detection algorithm.

[0093] S202: Using a clustering algorithm, perform cluster analysis on the patterns and groups identified in step S201 to extract key features from the data.

[0094] S203. Use a machine learning model, take the key features extracted in S202 as input, and use the patterns and groups in the geological features as labels to classify and predict the geological features and build an analysis model.

[0095] S204. Optimize the analysis model, specifically: use the existing expert interpretation results as input information for the machine learning model, and continuously optimize the analysis model through iteration; thereby ensuring the accuracy and reliability of the analysis results, and thus achieving quantitative analysis of regional geological characteristics by integrating multi-source knowledge graphs.

[0096] S3. Apply the results of regional geological analysis and present them through visual interfaces, including maps, charts, and reports. Provide researchers with interpretations and recommendations to support regional geological and exploration deployment decisions.

[0097] This technology visualizes the data required for exploration deployment, including exploration project distribution maps throughout the entire exploration and drilling lifecycle, 3D seismic interpretation profiles, drilling, logging, well logging, oil testing results, and core sampling and analysis data. This provides strong support for exploration deployment and decision-making in the Tahe region of Northwest China.

[0098] The present invention also provides a regional geological automatic analysis system that integrates multi-source knowledge relationship constraints, including a first module, a second module and a third module. The first module, the second module and the third module are connected in sequence. The first module is used to execute the content described in step S1, the second module is used to execute the content described in step S2, and the third module is used to execute the content described in step S3.

[0099] The present invention automatically analyzes regional geology by integrating multi-source knowledge relationship constraints, namely multi-source knowledge graphs, thereby achieving the interconnection, results sharing, experience inheritance, and joint collaboration of regional geological data, and providing technical support for solving the current difficulties in analyzing regional geology during oil and gas exploration and development.

[0100] The present invention obtains a multi-source knowledge graph by constructing a geological ontology and a machine learning algorithm, making the obtained multi-source knowledge graph more accurate and quantifying the relationship between different source entities, thereby facilitating the quantitative analysis of geological characteristics.

[0101] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions of the technical solution of the present invention by ordinary technicians in this field do not deviate from the essence and scope of the technical solution of the present invention.

Claims

1. A regional geological automatic analysis method integrating multi-source knowledge relationship constraints, characterized by: The following steps are involved: S1. Constructing constraints on multi-source knowledge relationships, specifically including the following steps: S101. Acquire regional geological data; S102, preprocessing the acquired regional geological data; S103, quantitatively representing the pre-processed geological data using structured knowledge; S104, integrating the data processed in step S103 into the same knowledge graph framework to obtain a multi-source knowledge graph framework; S105. Construct a regional geological data pattern recognition model to obtain implicit relationships between geological data, and add the implicit relationships between the obtained geological data to the multi-source knowledge graph framework obtained in S104 to obtain a multi-source knowledge graph, which is a constraint of the constructed multi-source knowledge relationship. S2. Integrate the constraints of the multi-source knowledge relationships established in step S1 to analyze the regional geological characteristics.

2. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 1 is characterized in that: S102 specifically includes the following steps: S1021. Perform data redundancy check and data quality check; S1022. Construct a geological ontology and unify geological terminology based on the constructed geological ontology; S1023, unifying the format of the data processed in step S1022; S1024, labeling and classifying the data processed in step S1023, and labeling different types of data according to the classification in the geological ontology; S1025. Perform data association on the data processed in step S1024 to build potential connections and relationship strengths between data from different data sources.

3. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 2 is characterized in that: The geological ontology includes a plurality of geological common concepts and their attributes and the relationships between the geological common concepts. The geological common concepts and their attributes are all standard terms.

4. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 3 is characterized in that: The specific method for unifying geological terminology based on geological ontology in S1022 is as follows: (1) Using natural language processing technology to scan the data processed in step S1021 and extract geological terms from the data; (2) Compare the extracted geological terms with the standard terms in the geological ontology, identify inconsistent or erroneous terms in the data, and replace the inconsistent or erroneous data to make the terms in the data consistent with the terms in the geological ontology.

5. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 4 is characterized in that: The specific replacement method in step (2) of S1022 is: when inconsistent or incorrect terms are identified, the first replacement method is executed and the second replacement method is executed at the same time.

6. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 5 is characterized in that: The first replacement method is: using the relationship between geological ontology standard terms to understand the contextual semantics of inconsistent or erroneous terms in the data, thereby determining the most suitable standard terms for the inconsistent or erroneous terms in the data, and replacing the inconsistent or erroneous terms in the data with the determined most suitable standard terms.

7. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 6 is characterized in that: The second replacement method is: for data in different data sources that are subsequently processed, when there is a term identical to the term replaced by the first replacement method, the term is replaced by the most suitable standard term determined by the first replacement method.

8. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 2 is characterized in that: S1025 specifically includes the following steps: (1) Using random forest algorithm to mine and analyze the potential connections between data from different data sources; (2) Use the Pearson correlation coefficient to quantify the strength of the relationship between different data sources.

9. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 2 is characterized in that: The specific method of checking data redundancy in S1021 is: verifying and removing duplicate data through unique constraint judgment.

10. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 2 is characterized in that: The specific method for S1021 to perform data quality check is: identifying erroneous data through data quality rules, and filtering missing data records through data integrity check.

11. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 1 is characterized in that: The specific method of step S103 is: using a graph embedding algorithm and a graph self-attention network to learn the structural features in the geological data, thereby mapping the entities and the relationships between entities in the geological data into a low-dimensional space.

12. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 11 is characterized in that: Mapping entities and relationships in geological data to low-dimensional space is specifically expressed by the following formula: E i =f(V,D,R) In the above formula, E i represents the embedding vector of the i-th entity in the low-dimensional space, V represents the set of entities in the graph, D represents the set of relationships between entities, and R represents the set of various relationship types between entities.

13. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 1 is characterized in that: S104 specifically includes the following steps: S1041. Obtain a knowledge graph framework; S1042. Standardize the temporal and spatial attributes of geological data; S1043. Use ontology alignment technology to identify the correspondence between different source entities and geological ontology concepts; S1044. Determine the optimal mapping relationship between the entity and the geological ontology using the cosine similarity semantic similarity measurement method; S1045. Using an ontology reasoning engine to ensure that the semantics of the mapped data remain consistent, thereby correcting the mapping relationship between the geological data and the corresponding concepts of the geological ontology, and forming a mapping relationship set between the geological data entities and the corresponding concepts of the geological ontology; S1046. Combine the mapping relationship set between the geological data entities and the corresponding concepts of the geological ontology obtained in S1045 with the knowledge graph framework obtained in S1041 to obtain a multi-source knowledge graph framework.

14. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 1 is characterized in that: The implicit relationship between the geological data in step S105 is specifically expressed by the following formula: R ij =g(E i ,E j ) In the above formula, E i represents the embedding vector of the i-th entity in the low-dimensional space, E j represents the embedding vector of the jth entity in the low-dimensional space, R ij represents the relationship between the i-th entity and the j-th entity, and g represents the function used to calculate the attribute relationship between the two entities.

15. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 1 is characterized in that: S2 specifically includes the following steps: S201, using community detection algorithms to identify patterns and groups in geological features; S202, using a clustering algorithm to perform cluster analysis on the patterns and groups identified in step S201, and extract key features from the data; S203, using a machine learning model, taking the key features extracted in S202 as input, and patterns and groups in the geological features as labels, to classify and predict the geological features and build an analysis model; S204. Optimize the analysis model.

16. The regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to claim 1 is characterized in that: The regional geological automatic analysis method further includes step S3, the specific content of which is: applying the regional geological characteristics analysis results, and displaying the analysis results through a visual interface in the form of maps, charts and reports.

17. A system using the regional geological automatic analysis method integrating multi-source knowledge relationship constraints according to any one of claims 1 to 16, characterized in that: It includes a first module, a second module and a third module connected in sequence, wherein the first module is used to execute the content in step S1, the second module is used to execute the content in step S2, and the third module is used to execute the content in step S3.

Images