Coal mine multi-source geological data dynamic diagnosis system and method based on knowledge graph

Abstract

The embodiment of the invention discloses a coal mine multi-source geological data dynamic diagnosis system and method based on a knowledge graph, and the system comprises a data collection module which is used for collecting and preprocessing multi-source heterogeneous data, and obtaining a standardized database; the data fusion module is used for determining a geological knowledge map, performing multi-source data knowledge extraction on the standardized database, performing knowledge fusion and conflict resolution, obtaining a knowledge triple and updating the geological knowledge map in real time; the intelligent diagnosis and decision module is used for identifying abnormal events in the standardized database, and performing graph traversal and evidence fusion based on the geological knowledge graph to obtain a diagnosis conclusion; a geological situation is adapted to obtain a diagnosis model, and an optimized diagnosis model is obtained through an incremental learning mechanism; and the decision application module is used for performing three-dimensional visual presentation on the diagnosis conclusion and the geological knowledge map, generating a diagnosis report, updating the geological knowledge map based on the feedback data flow and optimizing the diagnosis model. According to the invention, the accuracy and real-time performance of dynamic diagnosis can be improved.

Description

Technical Field

[0001] This application relates to the field of mine safety monitoring and intelligent decision-making technology, and includes, but is not limited to, dynamic diagnostic systems and methods for multi-source geological data in coal mines based on knowledge graphs. Background Technology

[0002] Coal mining is a typical hidden engineering project, and the complexity and uncertainty of its geological conditions severely restrict safe production. With the development of sensing technology and Internet of Things technology, modern coal mines have deployed a large number of multi-source monitoring devices to continuously collect heterogeneous geological and engineering data, including seismic exploration, drilling, gas concentration, ground stress, microseismic data, and mining progress. These data can be used to warn of potential risks and reveal key information such as geological patterns.

[0003] In existing technologies, the diagnosis of multi-source geological data in coal mines suffers from the following main deficiencies. First, because various monitoring data belong to different systems, such as ventilation systems, geological survey systems, and microseismic monitoring systems, the data formats, spatiotemporal benchmarks, and semantic standards are inconsistent, lacking effective correlation mechanisms and making it difficult to achieve collaborative analysis of cross-source information, ultimately leading to a serious data silo phenomenon. Second, the diagnostic capabilities of existing technologies are limited to the shallow perception stage, often using threshold alarms or simple statistical models to monitor single-source data, failing to reveal the deep geological causes behind abnormal events and lacking the ability to perform correlation reasoning and in-depth mining of multi-source information, resulting in diagnostic results that lack specificity and operability. Finally, the analytical models used in existing diagnostic systems are usually trained offline based on historical data, with relatively fixed parameters. They cannot dynamically adjust the diagnostic logic and early warning thresholds according to newly discovered faults, hidden collapse columns, and other newly revealed geological conditions, leading to model rigidity and a lack of context-adaptive capabilities. In the rapidly changing geological environment of mining operations, such models are prone to false alarms or missed alarms, seriously affecting the accuracy and timeliness of early warnings. In addition, some existing technologies use knowledge graph technology to semantically model mine data. However, the construction of this graph is mostly an offline and static process that is not linked with real-time mining activities and monitoring data streams. This makes it impossible to reflect the dynamic evolution of the mine's geological state. At the same time, this technology lacks a feedback mechanism between the graph reasoning results and the AI ​​diagnostic model, and cannot achieve closed-loop optimization from cognition to decision-making.

[0004] Therefore, there is an urgent need for a coal mine multi-source geological data diagnostic system that can deeply integrate multi-source data, dynamically reflect geological knowledge, and possess dynamic reasoning and adaptive diagnostic capabilities. This system would address the existing problems such as poor coordination of multi-source information, lack of specificity in diagnostic results, and low accuracy of early warnings, thereby significantly improving the accuracy and real-time performance of early warnings for coal mine geological disasters. Summary of the Invention

[0005] This application provides a knowledge graph-based dynamic diagnostic system and method for multi-source geological data in coal mines.

[0006] The technical solution of this application embodiment is implemented as follows: In a first aspect, embodiments of this application provide a dynamic diagnostic system for multi-source geological data in coal mines based on knowledge graphs. The system includes a data acquisition module, a data fusion module, an intelligent diagnosis and decision-making module, and a decision application module, wherein: The data acquisition module is used to collect multi-source heterogeneous data from the mine and preprocess the multi-source heterogeneous data to obtain a standardized database. The data fusion module is used to determine a geological knowledge graph based on the coal mine geology ontology, extract knowledge from the standardized database based on the geological knowledge graph to obtain initial knowledge triples, perform knowledge fusion and conflict resolution on the initial knowledge triples to obtain further knowledge triples, and update the geological knowledge graph in real time based on the knowledge triples and real-time data change events. The intelligent diagnosis and decision-making module is used to identify anomalies in the standardized database using an anomaly detection algorithm to obtain abnormal events, and based on the anomalies... The event and the geological knowledge graph are used to obtain a diagnostic conclusion through graph traversal and evidence fusion. Based on the diagnostic conclusion and the geological context, a model is adapted to obtain a diagnostic model adapted to the geological context. Based on real-time monitoring data, the diagnostic model is optimized through an incremental learning mechanism to obtain an optimized diagnostic model. A decision application module is used to overlay and render the diagnostic conclusion and the geological knowledge graph onto a 3D scene for 3D visualization. The diagnostic conclusion is encapsulated into a diagnostic report and pushed out. Feedback data streams are obtained based on user interactions in the 3D visualization scene or the diagnostic report, and the geological knowledge graph and the optimized diagnostic model are updated based on the feedback data streams.

[0007] The technical solution provided in this application collects multi-source heterogeneous data from the mine through a data acquisition module, achieving comprehensive collection of sensor data, business databases, and geological documents. Real-time data stream processing technology is used for immediate processing, significantly improving the completeness and timeliness of data acquisition. Preprocessing of the multi-source heterogeneous data yields a standardized database, effectively solving the problem of data silos due to inconsistent data formats and varying spatiotemporal benchmarks, providing a high-quality and reliable data foundation for subsequent analysis. In the data fusion module, a geological knowledge graph is determined based on the coal mine geology ontology, and multi-source data fusion is performed on the standardized database based on this geological knowledge graph. Knowledge extraction yields initial knowledge triples, enabling semantic-level deep fusion of heterogeneous data from multiple sources such as seismic, drilling, geophysical, and sensor data. Isolated data points are transformed into interconnected knowledge networks. Knowledge fusion and conflict resolution are performed on the initial knowledge triples to obtain new knowledge triples. Based on these knowledge triples and real-time data change events, the geological knowledge graph is updated in real-time, ensuring its timeliness and accuracy and providing reliable knowledge support for subsequent intelligent diagnosis. In the intelligent diagnosis and decision-making module, anomaly detection algorithms identify anomalies in the standardized database, identifying anomalous events. Based on these anomalous events and the geological knowledge graph... Diagnostic conclusions are derived through graph traversal and evidence fusion, enabling multi-hop correlation reasoning from anomalies to geological roots. This elevates diagnostic conclusions from simple threshold alarms to in-depth analysis encompassing complete causal chains, significantly enhancing the interpretability of the diagnostic process. Model adaptation is performed based on the diagnostic conclusions and geological context to obtain a diagnostic model suited to the specific geological conditions. Furthermore, based on real-time monitoring data, an incremental learning mechanism optimizes the diagnostic model, resulting in an optimized model. This allows the model parameters to dynamically adjust according to real-time geological conditions, significantly improving the accuracy and adaptability of the diagnostic model under complex geological conditions. In the decision-making application module, the diagnostic... The diagnostic conclusions and geological knowledge graphs are overlaid and rendered on a 3D scene for 3D visualization, presenting a complete chain of evidence from the anomaly source to the geological origin in an intuitive and visual form. This greatly improves the transparency of the system's diagnostic process. The diagnostic conclusions are packaged into a diagnostic report and pushed out. Feedback data streams are obtained based on user interactions in the 3D visualization scene or diagnostic report, and the geological knowledge graph is updated and the diagnostic model is optimized based on the feedback data streams. This achieves a complete closed loop from user interaction feedback to knowledge graph and diagnostic model updates, enabling the system to have continuous learning and self-optimization capabilities, thereby significantly improving the system's practicality and reliability in real-world applications.

[0008] Optionally, the preprocessing of the multi-source heterogeneous data to obtain a standardized database includes: removing abnormal data and repairing missing data in the multi-source heterogeneous data; converting the cleaned multi-source heterogeneous data into a standardized JSON or Avro format within the system; mapping the spatial information of the unified formatted multi-source heterogeneous data to a unified mine coordinate system, and synchronizing the timestamps to a standard time server to obtain the standardized database.

[0009] Optionally, the data fusion module includes a schema layer definition unit, a knowledge extraction unit, a knowledge fusion and conflict resolution unit, and a knowledge graph update unit. Specifically: the schema layer definition unit is used to construct a coal mine geology ontology based on professional knowledge in the coal mine geology field and an ontology language, defining entity types, hierarchical structures between entities, and entity attributes to obtain a schema layer of the geological knowledge graph; the knowledge extraction unit is used to extract entity instances, hierarchical structures between entity instances, and entity instance attributes from the standardized database based on the schema layer to obtain the initial knowledge triples. Specifically, for structured data in the standardized database, a streaming processing framework is used to perform real-time transformation on the structured data to obtain structured knowledge triples; for unstructured data in the standardized database, a natural language processing model is used to process the unstructured data. Unstructured data undergoes entity recognition and relation extraction to obtain unstructured knowledge triples. The knowledge fusion and conflict resolution unit uses a multi-dimensional feature-based entity alignment algorithm to match and merge the initial knowledge triples, resulting in a unified entity set. Based on a confidence fusion model using DS evidence theory, the unit integrates data source authority, timestamp freshness, and consistency features to calculate the confidence score for each knowledge fact in the unified entity set. Based on the confidence score, conflict decisions are made according to a preset strategy to obtain the knowledge triples. The knowledge graph update unit updates the graph database in a transactional manner based on the knowledge triples and real-time data change events, and performs real-time evolution of the geological knowledge graph. It also records state change events through an event tracing mode to obtain a geological knowledge graph version sequence.

[0010] Optionally, the intelligent diagnosis and decision-making module includes an anomaly identification unit, a map association diagnosis unit, a model adaptation unit, and a model optimization unit, wherein: the anomaly identification unit is used to identify anomalies in each category of data in the standardized database using corresponding anomaly detection algorithms to obtain anomaly events; the map association diagnosis unit is used to perform multi-hop association reasoning based on the anomaly events and the geological knowledge map through map traversal and evidence fusion to obtain a diagnostic conclusion containing a complete causal chain; the model adaptation unit is used to adapt the model based on the diagnostic conclusion and the geological context through a dynamic configuration mechanism to obtain a diagnostic model adapted to the geological context; the model optimization unit is used to obtain an incremental learning training set based on sliding window updates of the real-time monitoring data, fine-tune the model parameters of the diagnostic model using a stochastic gradient descent algorithm based on the incremental learning training set, update the diagnostic model, calculate performance indicators based on the prediction effect of the updated diagnostic model on the latest real-time monitoring data, and perform decay detection on the performance indicators. When the performance decay exceeds a preset threshold, the model is readjusted until the performance decay does not exceed the preset threshold to obtain the optimized diagnostic model.

[0011] Optionally, the graph-related diagnostic unit includes a graph traversal query subunit, an evidence fusion subunit, and a causal chain generation subunit, wherein: the graph traversal query subunit is used to perform multi-hop extended queries in the geological knowledge graph based on a predefined inference path template, with the anomalous event point as the center, to obtain diagnostic evidence associated with the anomalous event, the diagnostic evidence including entities and entity relationship sets associated with the anomalous event; the evidence fusion subunit is used to perform diagnostic hypothesis evaluation through a random forest-based classifier based on the diagnostic evidence collected from different inference paths, to obtain a preset number of multiple diagnostic hypotheses and their confidence levels; the causal chain generation subunit is used to perform causal logic organization based on the diagnostic hypotheses and associated entities to obtain a causal chain of diagnostic conclusions, the causal chain including anomalous phenomena, geological structures, and inducing mechanisms.

[0012] Optionally, the model adaptation unit includes a context feature encoding subunit, a model matching subunit, and a parameter adjustment subunit, wherein: the context feature encoding subunit is used to encode the geological scenario in the diagnostic conclusion to obtain standardized context features, which include geological structural features, rock mass characteristics, and engineering features; the model matching subunit is used to calculate the matching degree between the current geological scenario and the applicable conditions of each model in the model repository based on the standardized context features using cosine similarity, to obtain the diagnostic model with the highest matching degree; the parameter adjustment subunit is used to dynamically adjust the feature weights and reconfigure the model parameters based on the key features of the current geological scenario to obtain a diagnostic model adapted to the current geological scenario.

[0013] Secondly, embodiments of this application provide a dynamic diagnostic method for multi-source geological data in coal mines based on knowledge graphs. This method is applied to a dynamic diagnostic system for multi-source geological data in coal mines based on knowledge graphs. The system includes a data acquisition module, a data fusion module, an intelligent diagnosis and decision-making module, and a decision application module. The method includes: acquiring multi-source heterogeneous data from the mine and preprocessing the multi-source heterogeneous data to obtain a standardized database; determining a geological knowledge graph based on the coal mine geological domain ontology; extracting multi-source data knowledge from the standardized database based on the geological knowledge graph to obtain initial knowledge triples; performing knowledge fusion and conflict resolution on the initial knowledge triples to obtain further knowledge triples; and performing further analysis on the geological knowledge graph based on the knowledge triples and real-time data change events. The system performs real-time updates; it identifies anomalies in the standardized database using an anomaly detection algorithm to obtain anomaly events; based on the anomaly events and the geological knowledge graph, it obtains diagnostic conclusions through graph traversal and evidence fusion; it adapts the diagnostic conclusions to the geological context to obtain a diagnostic model adapted to the geological context, and optimizes the diagnostic model through an incremental learning mechanism based on real-time monitoring data to obtain an optimized diagnostic model; it overlays and renders the diagnostic conclusions and the geological knowledge graph onto a 3D scene for 3D visualization, encapsulates the diagnostic conclusions into a diagnostic report and pushes it out, obtains feedback data streams based on user interactions in the 3D visualization scene or the diagnostic report, and updates the geological knowledge graph and the optimized diagnostic model based on the feedback data streams.

[0014] Thirdly, embodiments of this application provide an electronic device, including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements the steps in the above-described dynamic diagnosis method for multi-source geological data of coal mines based on knowledge graphs.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps in the above-described dynamic diagnosis method for multi-source geological data in coal mines based on knowledge graphs.

[0016] The beneficial effects of the technical solutions provided in this application include at least the following: This application provides a knowledge graph-based dynamic diagnostic system and method for multi-source geological data in coal mines. The system collects multi-source heterogeneous data from the mine through a data acquisition module, achieving comprehensive collection of sensor data, business databases, and geological documents. Real-time data stream processing technology is used for immediate processing, significantly improving the completeness and timeliness of data acquisition. Preprocessing of the multi-source heterogeneous data yields a standardized database, effectively solving the problem of data silos due to inconsistent data formats and varying spatiotemporal benchmarks, providing a high-quality and reliable data foundation for subsequent analysis. In the data fusion module, a geological knowledge graph is determined based on the coal mine geology ontology, and based on the geological knowledge graph... The spectral analysis performs multi-source data knowledge extraction on a standardized database to obtain initial knowledge triples, enabling semantic-level deep fusion of heterogeneous data from sources such as seismic, drilling, geophysical, and sensor data. This transforms isolated data points into interconnected knowledge networks. Knowledge fusion and conflict resolution are then applied to the initial knowledge triples to obtain more knowledge triples. Based on these knowledge triples and real-time data change events, the geological knowledge graph is updated in real-time, ensuring its timeliness and accuracy and providing reliable knowledge support for subsequent intelligent diagnosis. In the intelligent diagnosis and decision-making module, anomaly detection algorithms identify anomalies in the standardized database, identifying anomalous events. Based on these anomalies... The event and geological knowledge graph, through graph traversal and evidence fusion, yields diagnostic conclusions, enabling multi-hop correlation reasoning from anomalies to geological roots. This elevates diagnostic conclusions from simple threshold alarms to in-depth analysis encompassing complete causal chains, significantly enhancing the interpretability of the diagnostic process. Model adaptation is performed based on the diagnostic conclusions and geological context to obtain a diagnostic model suited to the specific geological conditions. Furthermore, based on real-time monitoring data, an incremental learning mechanism optimizes the diagnostic model, resulting in an optimized model. This allows the diagnostic model to dynamically adjust its parameters according to real-time geological conditions, significantly improving its accuracy and adaptability under complex geological conditions. In decision-making applications… Within this module, diagnostic conclusions and geological knowledge graphs are overlaid and rendered onto a 3D scene for 3D visualization. This presents a complete chain of evidence from the source of the anomaly to its geological origin in an intuitive and visual form, greatly enhancing the transparency of the system's diagnostic process. Diagnostic conclusions are encapsulated into diagnostic reports and pushed out. Feedback data streams are obtained based on user interactions in the 3D visualization scene or diagnostic reports, and the geological knowledge graph is updated and the diagnostic model is optimized based on these feedback data streams. This achieves a complete closed loop from user interaction feedback to knowledge graph and diagnostic model updates, enabling the system to have continuous learning and self-optimization capabilities, thereby significantly improving the system's practicality and reliability in real-world applications. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein: Figure 1 A schematic diagram of a knowledge graph-based dynamic diagnostic system for multi-source geological data in coal mines, provided in an embodiment of this application. Figure 2 A flowchart illustrating the diagnostic process for abnormal gas concentration events provided in this application embodiment; Figure 3 A flowchart of a knowledge graph-based dynamic diagnosis method for multi-source geological data in coal mines, provided in an embodiment of this application; Figure 4 This is a schematic diagram of the hardware entity of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0020] It should be noted that the terms "first, second, and third" used in the embodiments of this application are merely to distinguish similar objects and do not represent a specific order of objects. It is understood that "first, second, and third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0021] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of this application pertain. It should also be understood that terms such as those defined in general dictionaries should be understood to have a meaning consistent with their meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0022] The embodiments of this application will be further described below with reference to the accompanying drawings.

[0023] In view of the current problems in the dynamic diagnosis of multi-source geological data in coal mines in the field of mine safety monitoring and intelligent decision-making technology, this application provides a knowledge graph-based dynamic diagnosis system and method for multi-source geological data in coal mines.

[0024] The technical solution of this application is described below, starting with the system implementation of this application.

[0025] Please refer to Figure 1 It illustrates a schematic diagram of a knowledge graph-based dynamic diagnostic system for multi-source geological data in coal mines, as provided in an embodiment of this application. Figure 1 As shown, the system includes a data acquisition module 01, a data fusion module 02, an intelligent diagnosis and decision-making module 03, and a decision application module 04. The data acquisition module 01 collects multi-source heterogeneous data from the mine and preprocesses the data to obtain a standardized database. The data fusion module 02 determines a geological knowledge graph based on the coal mine geology ontology, extracts knowledge from the standardized database using the geological knowledge graph, obtains initial knowledge triples, performs knowledge fusion and conflict resolution on the initial knowledge triples to obtain further knowledge triples, and updates the geological knowledge graph in real time based on the knowledge triples and real-time data change events. The intelligent diagnosis and decision-making module 03 identifies anomalies in the standardized database using an anomaly detection algorithm, obtains abnormal events, and updates the database based on the anomalies and conflicts. The system uses a combination of routine events and the geological knowledge graph to obtain diagnostic conclusions through graph traversal and evidence fusion. Based on these conclusions and the geological context, a model is adapted to the specific geological context to obtain a diagnostic model. Then, based on real-time monitoring data, the diagnostic model is optimized using an incremental learning mechanism to obtain an optimized diagnostic model. The decision application module 04 overlays and renders the diagnostic conclusions and the geological knowledge graph onto a 3D scene for 3D visualization. It also encapsulates the diagnostic conclusions into a diagnostic report and pushes it to the system. Feedback data streams are obtained based on user interactions in the 3D visualization scene or the diagnostic report, and the geological knowledge graph and the optimized diagnostic model are updated based on these feedback data streams.

[0026] In this embodiment, the data acquisition module 01 is used to collect multi-source heterogeneous data from the mine and preprocess the multi-source heterogeneous data to obtain a standardized database. Specifically, multi-source heterogeneous data from the coal mine production environment is continuously and in parallel collected by various data acquisition devices deployed at the mine. This includes sensor data, data recorded in the business database, and data from geological documents. The sensor data includes continuous monitoring data generated by devices such as gas sensors, wind speed sensors, ground stress sensors, and microseismic event monitoring sensors. The data recorded in the business database includes structured data recorded in databases such as geological databases, drilling record databases, and mining engineering plan databases. The data in the unstructured geological documents includes data recorded in professional geological data such as seismic interpretation reports and geological specifications. Furthermore, the collected raw, multi-source heterogeneous data undergoes standardized preprocessing through a unified governance engine. First, the raw data is cleaned and repaired, using a Z-score anomaly detection algorithm to identify and remove outliers, and interpolation algorithms to repair missing data, ensuring data integrity and quality. Next, the cleaned data is formatted uniformly, converting data from different protocols into the system's standardized JSON or Avro format using a protocol converter to achieve structural and semantic consistency. Finally, the formatted data undergoes spatiotemporal alignment, mapping the spatial information of all data to a unified mining coordinate system using a coordinate transformation algorithm, and synchronizing timestamps to a standard time server via a time synchronization service to establish a unified spatiotemporal reference. These three closely linked standardization operations result in a standardized database. Finally, the preprocessed standardized database is published to the data bus, providing a high-quality, high-standard, and reliable data foundation for subsequent knowledge graph construction and intelligent diagnostic analysis.

[0027] In this embodiment, the data fusion module 02 includes a schema layer definition unit, a knowledge extraction unit, a knowledge fusion and conflict resolution unit, and a knowledge graph update unit. The schema layer definition unit is used to construct a coal mine geology ontology based on professional knowledge in the field of coal mine geology, according to an ontology language, and to define entity types, hierarchical structures between entities, and entity attributes to obtain the schema layer of the geological knowledge graph. Specifically, entity types include engineering entities, geological entities, and monitoring entities. Engineering entities include workfaces, boreholes, and roadways; geological entities include coal seams, faults, folds, and strata; and monitoring entities include gas sensors, stress sensors, and microseismic sensors. Entity attributes include object attributes and data attributes. Object attributes include spatial relationships, geological relationships, and engineering relationships. Spatial relationships include "isLocatedIn," "isAdjacentTo," and "intersectsWith." Geological relationships include "reveals," "contains," and "belongsTo." Engineering relationships include "monitors," "affects," and "controls." Data attributes mainly describe the numerical characteristics of entities, including dip angle, gas concentration, advancing speed, and burial depth. Transitivity and symmetry constraints between entity attributes are defined using SWRL rules to ensure logical consistency in knowledge reasoning, ultimately constructing the schema layer of the geological knowledge graph.

[0028] In this embodiment, the knowledge extraction unit is used to extract entity instances, hierarchical structures between entity instances, and entity instance attributes from a standardized database based on the schema layer of a geological knowledge graph, thereby obtaining initial knowledge triples. Specifically, for structured data in the standardized database, the system uses the Apache NiFi streaming data processing framework to capture database change events in real time. A pre-configured message middleware (such as the Jinja2 template engine) is used to convert structured data (such as relational data) records into knowledge triples in real time, resulting in structured knowledge triples. For example, when a new borehole data record arrives, the system automatically instantiates the Borehole individual, parses its spatial coordinates, coal seam depth, and other attributes, and establishes an intersectsWith relationship with the corresponding CoalSeam individual. For unstructured text data in standardized databases, geological entity recognition and relation extraction are performed using natural language processing models such as the BERT-based named entity recognition model to obtain unstructured knowledge triples. The BERT-based named entity recognition model is fine-tuned on labeled coal mine geological report corpora. It adopts a relation extraction pipeline based on dependency parsing, extracts semantic relations between entities by recognizing the subject-verb-object structure, and performs spatial relation parsing through regular expression matching and a custom parser to convert the directional descriptions in the text into normalized spatial coordinate offsets, thereby establishing accurate spatial relations and finally obtaining structured and unstructured knowledge triples.

[0029] In this embodiment, within the knowledge fusion and conflict resolution unit, a multi-dimensional feature-based entity alignment algorithm comprehensively utilizes multi-dimensional features such as edit distance, word vector cosine similarity, and spatial distance to generate candidate matching pairs. This enables entity matching and merging of the initial knowledge triples, resulting in a unified entity set. Further, based on a confidence fusion model using DS evidence theory, a confidence score is calculated for each knowledge fact in the unified entity set. The model's input features include data source authority, timestamp freshness, and consistency with other facts. Based on this confidence score, when a knowledge conflict is detected, the system performs knowledge base consistency maintenance according to a preset resolution strategy, such as prioritizing drilling data over geophysical interpretation results, to ensure the accuracy and reliability of knowledge, ultimately yielding a clean and consistent knowledge triple. In an optional embodiment, the knowledge graph update unit is used to update the graph database and perform real-time evolution of the geological knowledge graph based on knowledge triples and real-time data change events. Specifically, it reduces the update latency of the geological knowledge graph through an event-driven architecture, receives real-time data change events through a Kafka message queue, encapsulates the graph database update operation into atomic transactions to ensure the ACID properties of the knowledge graph, retains all historical state change event records through an event sourcing mode, obtains the geological knowledge graph version sequence, thereby supporting the tracing and rollback of the knowledge evolution trajectory, providing complete data support for the interpretability of the diagnostic process, and ultimately realizing the real-time evolution and version management of the geological knowledge graph accompanying mining activities.

[0030] In this embodiment, the intelligent diagnosis and decision-making module 03 includes an anomaly identification unit, a map association diagnosis unit, a model adaptation unit, and a model optimization unit. The anomaly identification unit identifies anomalous events by using corresponding anomaly detection algorithms for various categories of data in the standardized database. Specifically, for sensor data such as gas concentration and wind speed in the standardized database, a dynamic threshold-based anomaly detection algorithm is used to perform real-time over-limit judgment, resulting in threshold-based anomalous events. For time-series data such as ground stress and microseismic events, a statistical process control algorithm is used to perform control chart analysis, resulting in statistically based anomalous events. All identified anomalous events are encapsulated in a structured form containing metadata such as anomaly type, location, timestamp, and severity.

[0031] In this embodiment, the graph association diagnosis unit is used to perform multi-hop association reasoning based on anomalies and geological knowledge graphs through graph traversal and evidence fusion to obtain a diagnostic conclusion containing a complete causal chain. Specifically, the graph association diagnosis unit includes a graph traversal query subunit, an evidence fusion subunit, and a causal chain generation subunit. The graph traversal query subunit is used to perform multi-hop extended queries in the geological knowledge graph based on a predefined reasoning path template, with the anomaly point as the center. The reasoning path template includes multiple reasoning chains such as geological structure association paths, engineering activity impact paths, and associated phenomenon verification paths. The multi-hop extended queries are implemented using the Gremlin graph traversal language, thereby obtaining diagnostic evidence associated with the anomaly. This diagnostic evidence includes a set of entities and entity relationships associated with the anomaly collected from different reasoning paths. The evidence fusion subunit, based on diagnostic evidence collected from different reasoning paths, organizes evidence entities and their attributes into feature vectors as model input. A random forest-based classifier is used to evaluate diagnostic hypotheses. This classifier, trained on historical cases, identifies typical disaster patterns and outputs a preset number of most probable diagnostic hypotheses and their corresponding confidence levels. The causal chain generation subunit, based on diagnostic hypotheses and associated entities, performs causal logic organization to obtain a complete causal chain diagnostic conclusion encompassing anomalies, geological structures, and triggering mechanisms.

[0032] In this embodiment, the model adaptation unit is used to adapt the model to the geological situation based on the diagnostic conclusion and the identified geological situation through a dynamic configuration mechanism, thereby obtaining a diagnostic model adapted to the geological situation. Specifically, the model adaptation unit includes a situation feature encoding subunit, a model matching subunit, and a parameter adjustment subunit. The situation feature encoding subunit is used to encode the geological situation in the diagnostic conclusion, representing the geological environment of the current working face as a standardized feature vector to obtain standardized situation features, including geological structural features, rock mass characteristics, and engineering features, such as fault density, fold curvature, RQD value, rock mass quality classification, mining depth, and coal mining methods. The model matching subunit is used to calculate the matching degree between the current geological situation and the applicable conditions of each model in the model repository based on the standardized situation features using cosine similarity. The model repository stores dedicated prediction models trained for different geological scenarios, such as fault zone gas prediction models and hard roof pressure prediction models. Based on the matching degree between the standardized situation features of the current geology and the applicable conditions of each model, the diagnostic model with the highest matching degree is selected by calculating cosine similarity. The parameter adjustment subunit is used to dynamically adjust the input feature weights of the diagnostic model based on the key characteristics of the current geological situation, and to reconfigure the model parameters to strengthen features that are strongly related to the current situation. For example, when the diagnostic result is the influence of faults, the weight of the distance feature from the fault is increased by 40%, thereby obtaining a diagnostic model that is adapted to the current geological situation.

[0033] In this embodiment, the model optimization unit continuously optimizes the diagnostic model performance based on real-time monitoring data through an incremental learning mechanism. Specifically, a sliding window update is performed on the real-time monitoring data based on a time window strategy. A circular buffer is maintained for each active working surface to store the monitoring data from the most recent N hours, resulting in an incremental learning training set. Based on the incremental learning training set, the model parameters of the diagnostic model are fine-tuned using a stochastic gradient descent algorithm to achieve online incremental learning and updating of the diagnostic model. Performance metrics such as AUC are calculated based on the prediction effect of the updated diagnostic model on the latest real-time monitoring data, and performance metric decay detection is performed. When the performance decay exceeds a preset threshold, a model retraining alarm is triggered, and the model is readjusted until the performance decay does not exceed the preset threshold, resulting in an optimized diagnostic model, thereby completing the continuous self-optimization of the diagnostic model.

[0034] In this embodiment, within the decision application module 04, an immersive 3D mine scene is constructed using the WebGL graphics library. This integrates a 3D geological model, tunnel model, and equipment layout data to form a complete virtual mine environment. Based on the evidence chain in the diagnostic conclusions, a geological knowledge graph reasoning path is overlaid and rendered within the 3D scene using highlighting, regional coloring, and dynamic flow effects. This creates a complete visual display of the evidence chain from the anomaly source to the geological root cause. For user interactions with any node in the 3D scene, a data query interface is used to retrieve associated data, displaying all original multi-source data associated with that node. This ensures the traceability and verifiability of the diagnostic process, ultimately forming a 3D visualized diagnostic scene with an overlaid knowledge graph reasoning path. Furthermore, the diagnostic conclusions are packaged into structured diagnostic reports and pushed through multiple channels. The diagnostic report template is based on the standards in the field of coal mine safety diagnosis and includes an anomaly summary, diagnostic conclusions, evidence chain, risk level, and handling recommendations. Based on the diagnostic conclusions and evidence chain, a natural language generation engine automatically synthesizes the structured diagnostic information into a conclusion description that conforms to professional terminology. A complete diagnostic report containing abnormal phenomena, related geological factors, reasoning process, risk level, and handling recommendations is generated based on the diagnostic report template. Further, according to user configuration and warning level, the packaged diagnostic report is proactively pushed to relevant users through web service interfaces, mobile push gateways, or email servers to ensure timely delivery of warning information. The system monitors user interactions in 3D visualization scenes or diagnostic reports in real time, including actions such as clicking, confirming, correcting, and text input. It records the original interaction logs to obtain initial feedback information. Based on a predefined feedback data pattern, it performs intent parsing and format conversion on the initial feedback information, transforming it into structured feedback data containing target entity identifiers, feedback types, and specific correction content. The structured feedback data is then published as feedback event messages through the system data bus, driving it to flow to the data fusion module 02 to update the geological knowledge graph content and to the intelligent diagnosis and decision-making module 03 to optimize the diagnostic model parameters. This completes the collaborative update of the geological knowledge graph and the optimized diagnostic model based on the feedback data flow, forming a complete closed loop of the perception-diagnosis-decision-optimization system.

[0035] In a specific embodiment, please refer to Figure 2This document illustrates a flowchart of a diagnostic process for abnormal gas concentration events provided in an embodiment of this application, demonstrating a complete diagnosis of an abnormal gas concentration event at gas sensor G-07 in the T101 return airway of the working face. When the sensor data rapidly increases from 0.3% to 0.8% within a short period and triggers a primary alarm, the system immediately initiates the intelligent diagnostic process. The anomaly identification unit identifies the gas concentration exceeding the limit event using a threshold detection algorithm, generating a structured description of the anomaly event containing metadata such as the anomaly type "abnormal gas concentration," the location "G-07 sensor," a timestamp, and the severity. Upon receiving an abnormal event, the map association diagnosis unit immediately performs a multi-hop association query in the geological knowledge map, centered on the G-07 sensor entity. The map traversal query subunit expands the query according to the predefined reasoning path template, locates the T101 working face entity along the installation relationship path "GasSensor→isInstalledIn→Workface", discovers the adjacent F101 fault entity along the spatial relationship path "Workface→isAdjacentTo→Fault", and finds the geostress sensor S-12 associated with the fault along the association relationship path "Fault→hasCorrelated→StressSensor". The evidence fusion subunit organizes diagnostic evidence (including entities and their relationships such as G-07 sensor, T101 working face, F101 fault, S-12 sensor, etc.) collected from these different reasoning paths into feature vectors, which are then input into a random forest-based classifier for evaluation. This classifier evaluates the hypothesis based on the historically trained "fault-activated gas outburst" pattern and calculates that the diagnostic hypothesis has the highest confidence. The causal chain generation subunit then organizes and generates a complete causal chain diagnostic conclusion: the working face advances to the F101 fault structural influence zone, inducing stress release and fracture conduction, resulting in abnormal gas outburst from the fault. Based on the "fault influence" geological situation identified in the diagnosis, the model adaptation unit initiates a situation-aware adaptive mechanism. The situation feature encoding subunit encodes the current geological environment into a standardized feature vector containing features such as fault distance and geostress change rate. The model matching subunit selects a dedicated "fault zone gas prediction model" from the model repository through cosine similarity calculation. The parameter adjustment subunit dynamically adjusts the model parameters, increases the weight of the "distance from fault" feature by 40%, and configures the corresponding online learning parameters to obtain a diagnostic model adapted to the current geological situation.

[0036] In summary, the knowledge graph-based dynamic diagnostic system for multi-source geological data in coal mines provided in this application collects multi-source heterogeneous data from the mine through a data acquisition module, achieving comprehensive collection of sensor data, business databases, and geological documents. Real-time data stream processing technology is used for immediate processing, significantly improving the completeness and timeliness of data collection. Preprocessing of the multi-source heterogeneous data yields a standardized database, effectively solving the problem of data silos due to inconsistent data formats and varying spatiotemporal benchmarks, providing a high-quality and reliable data foundation for subsequent analysis. In the data fusion module, a geological knowledge graph is determined based on the coal mine geology ontology, and based on geological... The knowledge graph extracts knowledge from multi-source data in a standardized database to obtain initial knowledge triples, enabling semantic-level deep fusion of heterogeneous data from sources such as seismic, drilling, geophysical, and sensor data. It transforms isolated data points into an interconnected knowledge network. Knowledge fusion and conflict resolution are then performed on the initial knowledge triples to obtain more knowledge triples. The geological knowledge graph is updated in real-time based on these knowledge triples and real-time data change events, ensuring its real-time performance and accuracy, and providing reliable knowledge support for subsequent intelligent diagnosis. In the intelligent diagnosis and decision-making module, anomaly detection algorithms identify anomalies in the standardized database, identifying abnormal events. Anomaly events and geological knowledge graphs, through graph traversal and evidence fusion, yield diagnostic conclusions. This enables multi-hop correlation reasoning from anomalies to geological roots, elevating diagnostic conclusions from simple threshold alarms to in-depth analysis containing complete causal chains, significantly enhancing the interpretability of the diagnostic process. Model adaptation is performed based on diagnostic conclusions and geological contexts to obtain a diagnostic model suited to the specific geological conditions. Furthermore, based on real-time monitoring data, an incremental learning mechanism optimizes the diagnostic model, resulting in an optimized model. This allows the diagnostic model to dynamically adjust its parameters according to real-time geological conditions, significantly improving its accuracy and adaptability under complex geological conditions. This technology is applied in decision-making. In this module, diagnostic conclusions and geological knowledge graphs are overlaid and rendered on a 3D scene for 3D visualization. This presents a complete chain of evidence from the source of the anomaly to its geological origin in an intuitive and visual form, greatly improving the transparency of the system's diagnostic process. Diagnostic conclusions are packaged into diagnostic reports and pushed out. Feedback data streams are obtained based on user interactions in the 3D visualization scene or diagnostic reports. The geological knowledge graph is updated and the diagnostic model is optimized based on the feedback data streams. This achieves a complete closed loop from user interaction feedback to knowledge graph and diagnostic model updates, enabling the system to have continuous learning and self-optimization capabilities, thereby significantly improving the system's practicality and reliability in real-world applications.

[0037] The above is a description of the system embodiments of this application. Based on the foregoing embodiments, the method embodiments of this application are described below.

[0038] Please refer to Figure 3It illustrates a flowchart of a knowledge graph-based dynamic diagnosis method for multi-source geological data in coal mines, as provided in an embodiment of this application. This method is applied to applications such as... Figure 1 The knowledge graph-based dynamic diagnostic system for coal mine multi-source geological data shown herein is illustrated in the system implementation example. For details not disclosed in the method embodiments, please refer to the system implementation example. The system includes a data acquisition module, a data fusion module, an intelligent diagnosis and decision-making module, and a decision application module. Figure 3 As shown, the method includes the following steps S310 to S340.

[0039] Step S310: Collect multi-source heterogeneous data from the mine and preprocess the multi-source heterogeneous data to obtain a standardized database.

[0040] Step S320: Determine a geological knowledge graph based on the ontology of coal mine geology; extract knowledge from multi-source data in the standardized database based on the geological knowledge graph to obtain initial knowledge triples; perform knowledge fusion and conflict resolution on the initial knowledge triples to obtain more knowledge triples; and update the geological knowledge graph in real time based on the knowledge triples and real-time data change events. Step S330: Anomalies are identified in the standardized database using an anomaly detection algorithm to obtain anomaly events. Based on the anomaly events and the geological knowledge graph, a diagnostic conclusion is obtained through graph traversal and evidence fusion. Based on the diagnostic conclusion and the geological context, a model is adapted to obtain a diagnostic model adapted to the geological context. Based on real-time monitoring data, the diagnostic model is optimized through an incremental learning mechanism to obtain an optimized diagnostic model.

[0041] Step S340: Overlay the diagnostic conclusion and the geological knowledge graph onto a 3D scene and perform 3D visualization; encapsulate the diagnostic conclusion into a diagnostic report and push it out; obtain feedback data stream based on the user's interactive operations in the 3D visualization scene or the diagnostic report; and update the geological knowledge graph and the optimized diagnostic model based on the feedback data stream.

[0042] It should be noted that, in the embodiments of this application, if the above-mentioned dynamic diagnosis method for multi-source geological data in coal mines based on knowledge graphs is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0043] Correspondingly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon. When executed by a processor, this computer program implements the steps in any of the knowledge graph-based dynamic diagnosis methods for multi-source geological data in coal mines described in the above embodiments. Correspondingly, embodiments of this application also provide a computer program product. When executed by a processor of an electronic device, this computer program product is used to implement the steps in any of the knowledge graph-based dynamic diagnosis methods for multi-source geological data in coal mines described in the above embodiments.

[0044] Based on the same technical concept, this application provides an electronic device for implementing the knowledge graph-based dynamic diagnosis method for multi-source geological data in coal mines described in the above-described method embodiments. Figure 4 This is a hardware entity diagram of an electronic device provided in an embodiment of this application, such as... Figure 4 As shown, the electronic device 400 includes a memory 410 and a processor 420. The memory 410 stores a computer program that can run on the processor 420. When the processor 420 executes the program, it implements the steps in any of the knowledge graph-based dynamic diagnosis methods for coal mine multi-source geological data according to the embodiments of this application.

[0045] The memory 410 is configured to store instructions and applications executable by the processor 420, and can also cache data to be processed or already processed by the processor 420 and various modules in the electronic device (e.g., image data, audio data, voice communication data and video communication data), which can be implemented by flash memory or random access memory (RAM).

[0046] When the processor 420 executes the program, it implements any of the steps of the knowledge graph-based dynamic diagnosis method for multi-source geological data in coal mines. The processor 420 typically controls the overall operation of the electronic equipment 400.

[0047] The aforementioned processor can be at least one of the following: Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that other electronic devices can also implement the functions of the aforementioned processor, and this application does not specifically limit the specific implementation.

[0048] The aforementioned computer storage media / memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM), etc.; or it can be various electronic devices that include one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0049] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0050] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0051] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0052] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0053] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0054] In addition, each functional unit in the various embodiments of this application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0055] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause the device automatic test line to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0056] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0057] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0058] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A dynamic diagnostic system for multi-source geological data in coal mines based on knowledge graphs, characterized in that, The system includes: The data acquisition module is used to collect multi-source heterogeneous data from the mine and preprocess the multi-source heterogeneous data to obtain a standardized database. The data fusion module is used to determine a geological knowledge graph based on the ontology of the coal mine geology field, extract knowledge from multi-source data in the standardized database based on the geological knowledge graph to obtain initial knowledge triples, perform knowledge fusion and conflict resolution on the initial knowledge triples to obtain knowledge triples, and update the geological knowledge graph in real time according to the knowledge triples and real-time data change events. The intelligent diagnosis and decision-making module is used to identify anomalies in the standardized database through anomaly detection algorithms to obtain abnormal events. Based on the abnormal events and the geological knowledge graph, a diagnostic conclusion is obtained through graph traversal and evidence fusion. Based on the diagnostic conclusion and the geological context, a model is adapted to obtain a diagnostic model adapted to the geological context. Based on real-time monitoring data, the diagnostic model is optimized through an incremental learning mechanism to obtain an optimized diagnostic model. The decision application module is used to overlay and render the diagnostic conclusion and the geological knowledge graph on a three-dimensional scene and present it in three-dimensional visualization, encapsulate the diagnostic conclusion into a diagnostic report and push it out, obtain feedback data stream based on the user's interactive operations in the three-dimensional visualization scene or the diagnostic report, and update the geological knowledge graph and the optimized diagnostic model based on the feedback data stream.

2. The method according to claim 1, characterized in that, The preprocessing of the multi-source heterogeneous data to obtain a standardized database includes: Abnormal data in the multi-source heterogeneous data is removed and missing data is repaired; The cleaned multi-source heterogeneous data is converted into the system's internal standardized JSON or Avro format; The spatial information of the multi-source heterogeneous data after unification is mapped to a unified mine coordinate system, and the timestamps are synchronized to a standard time server to obtain the standardized database.

3. The system according to claim 1, characterized in that, The data fusion module includes a pattern layer definition unit, a knowledge extraction unit, a knowledge fusion and conflict resolution unit, and a knowledge graph update unit, wherein: The pattern layer definition unit is used to construct an ontology of the coal mine geology field based on professional knowledge in the field of coal mine geology and according to the ontology language, and to define entity types, hierarchical structure between entities and entity attributes to obtain the pattern layer of the geological knowledge graph. The knowledge extraction unit is used to extract entity instances, hierarchical structures between entity instances, and entity instance attributes from the standardized database based on the pattern layer to obtain the initial knowledge triples. Specifically, for the structured data in the standardized database, a streaming processing framework is used to perform real-time transformation on the structured data to obtain structured knowledge triples. For the unstructured data in the standardized database, a natural language processing model is used to perform entity recognition and relation extraction on the unstructured data to obtain unstructured knowledge triples. The knowledge fusion and conflict resolution unit is used for entity alignment algorithm based on multi-dimensional features to perform entity matching and merging on the initial knowledge triples to obtain a unified entity set; according to the confidence fusion model based on DS evidence theory, the confidence score of each knowledge fact in the unified entity set is calculated by fusing data source authority, timestamp freshness and consistency features; based on the confidence score, conflict decision is made according to a preset strategy to obtain the knowledge triples. The knowledge graph update unit is used to update the graph database in a transactional manner and perform real-time evolution of the geological knowledge graph based on the knowledge triples and the real-time data change events, and record the state change events through an event tracing mode to obtain the geological knowledge graph version sequence.

4. The system according to claim 1, characterized in that, The intelligent diagnosis and decision-making module includes an anomaly identification unit, a graph association diagnosis unit, a model adaptation unit, and a model optimization unit, wherein: The anomaly identification unit is used to identify anomalies in each category of data in the standardized database using corresponding anomaly detection algorithms, thereby obtaining anomaly events. The map association diagnostic unit is used to perform multi-hop association reasoning based on the abnormal event and the geological knowledge map through map traversal and evidence fusion to obtain a diagnostic conclusion containing a complete causal chain. The model adaptation unit is used to adapt the model to the geological context through a dynamic configuration mechanism based on the diagnostic conclusions and the geological context, so as to obtain a diagnostic model adapted to the geological context. The model optimization unit is used to obtain an incremental learning training set by updating the real-time monitoring data through a sliding window. Based on the incremental learning training set, the model parameters of the diagnostic model are fine-tuned using a stochastic gradient descent algorithm to update the diagnostic model. The performance index is calculated based on the prediction effect of the updated diagnostic model on the latest real-time monitoring data, and the performance index is subjected to decay detection. When the performance decay exceeds a preset threshold, the model is readjusted until the performance decay does not exceed the preset threshold to obtain the optimized diagnostic model.

5. The system according to claim 4, characterized in that, The graph association diagnostic unit includes a graph traversal query subunit, an evidence fusion subunit, and a causal chain generation subunit, wherein: The graph traversal query subunit is used to perform multi-hop extended queries on the geological knowledge graph based on a predefined reasoning path template, with the anomalous event point as the center, to obtain diagnostic evidence associated with the anomalous event. The diagnostic evidence includes entities and entity relationship sets associated with the anomalous event. The evidence fusion subunit is used to evaluate diagnostic hypotheses based on diagnostic evidence collected from different reasoning paths, and obtain a preset number of diagnostic hypotheses and their reliability. The causal chain generation subunit is used to perform causal logic organization based on diagnostic hypotheses and related entities to obtain a causal chain of diagnostic conclusions, which includes abnormal phenomena, geological structures and inducing mechanisms.

6. The system according to claim 4, characterized in that, The model adaptation unit includes a context feature encoding subunit, a model matching subunit, and a parameter adjustment subunit, wherein: The context feature coding subunit is used to encode the geological context in the diagnostic conclusion to obtain standardized context features, which include geological structural features, rock mass characteristics and engineering features. The model matching subunit is used to calculate the matching degree between the current geological situation and the applicable conditions of each model in the model warehouse based on the standardized situation features and cosine similarity, so as to obtain the diagnostic model with the highest matching degree. The parameter adjustment subunit is used to dynamically adjust feature weights and reconfigure model parameters based on the key features of the current geological situation, so as to obtain a diagnostic model adapted to the current geological situation.

7. A dynamic diagnostic method for multi-source geological data in coal mines based on knowledge graphs, characterized in that, A knowledge graph-based dynamic diagnostic system for multi-source geological data in coal mines is provided. The system includes a data acquisition module, a data fusion module, an intelligent diagnosis and decision-making module, and a decision application module. The method includes: Collect multi-source heterogeneous data from the mining end and preprocess the multi-source heterogeneous data to obtain a standardized database; A geological knowledge graph is determined based on the ontology of the coal mine geology field. Based on the geological knowledge graph, multi-source data knowledge is extracted from the standardized database to obtain initial knowledge triples. Knowledge fusion and conflict resolution are performed on the initial knowledge triples to obtain knowledge triples. The geological knowledge graph is updated in real time according to the knowledge triples and real-time data change events. Anomalies are identified in the standardized database using an anomaly detection algorithm to obtain anomaly events. Based on the anomaly events and the geological knowledge graph, a diagnostic conclusion is obtained through graph traversal and evidence fusion. The model is then adapted to the geological context based on the diagnostic conclusion and the geological context to obtain a diagnostic model adapted to the geological context. Based on real-time monitoring data, the diagnostic model is optimized through an incremental learning mechanism to obtain an optimized diagnostic model. The diagnostic conclusion and the geological knowledge graph are overlaid and rendered on a 3D scene and presented in 3D visualization. The diagnostic conclusion is encapsulated into a diagnostic report and pushed out. Feedback data stream is obtained based on the user's interactive operations in the 3D visualization scene or the diagnostic report, and the geological knowledge graph and the optimized diagnostic model are updated based on the feedback data stream.

8. An electronic device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method of claim 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method of claim 7.

Images