Intelligent Analysis Method, Medium and Equipment for Mine Disasters Based on Knowledge Graph

Abstract

The present application discloses an intelligent analysis method, medium and device for mine disasters based on a knowledge graph, relating to the technical field of knowledge graphs. The method includes: obtaining historical multi-source heterogeneous data of a mine, integrating the historical multi-source heterogeneous data, and constructing a knowledge graph, where the nodes of the knowledge graph are used to represent the entities of the mine, and the edges of the knowledge graph are used to represent the mutual relationships between the entities; in response to inputting the input data obtained in real time into the knowledge graph, identifying key entities; extracting the mutual relationships between the key entities, and adding the mutual relationships between the key entities to the knowledge graph to obtain an updated knowledge graph; extracting variables in the input data, and determining at least one mine disaster rule based on the variables; and determining the mine disaster with the highest occurrence probability in the current mine according to the updated knowledge graph and at least one mine disaster rule. The present application can effectively save resources while ensuring the accuracy of mine disaster analysis.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of knowledge graphs, and in particular, to an intelligent analysis method, medium, and device for mine disasters based on a knowledge graph. Background Art

[0002] At present, the mine operation environment is complex. If a mine accident occurs, it will not only cause casualties, but also damage equipment and production stagnation. Therefore, mine disaster analysis is not only related to the safety of miners and the economic interests of enterprises, but also of great significance to the sustainable development of the industry and the fulfillment of social responsibilities.

[0003] With the development of technology, currently, mine disasters are usually analyzed by means of the Internet of Things. For example, by collecting data from various sensors installed in the mine and integrating the data, and inputting it into a preset prediction model, the prediction of mine disasters can be realized. This mine disaster analysis method can effectively achieve the prediction of mine disasters. However, the disadvantage is that the prediction model is constructed based on historical data. Since the mine environment is changing in real time and dynamically, if the prediction model is reconstructed every once in a while according to the obtained mine data, it will consume a large amount of human and material resources.

[0004] Based on this, there is an urgent need for a method that can ensure the accuracy of mine disaster analysis while effectively saving resources. Summary of the Invention

[0005] The embodiments of the present application provide an intelligent analysis method, medium, and device for mine disasters based on a knowledge graph, which are used to ensure the accuracy of mine disaster analysis while effectively saving resources.

[0006] To achieve the above object, the embodiments of the present application adopt the following technical solutions:

[0007] In a first aspect, an intelligent analysis method for mine disasters based on a knowledge graph is provided. The method includes:

[0008] Obtain historical multi-source heterogeneous data of the mine, and integrate the historical multi-source heterogeneous data to construct a knowledge graph. The nodes of the knowledge graph are used to represent the entities of the mine, and the edges of the knowledge graph are used to represent the mutual relationships between the entities;

[0009] In response to inputting the input data obtained in real time into the knowledge graph, identify key entities;

[0010] Extract the mutual relationships between the key entities, and add the mutual relationships between the key entities to the knowledge graph to obtain an updated knowledge graph;

[0011] Extract the variables from the input data and determine at least one mine disaster rule based on the variables;

[0012] Based on the updated knowledge graph and at least one of the mine disaster rules, determine the mine disaster with the highest probability of occurrence in the current mine.

[0013] In a possible implementation manner of the first aspect, the extracting the mutual relationships between the key entities and adding the mutual relationships between the key entities to the knowledge graph to obtain an updated knowledge graph includes:

[0014] Obtain the context information of the key entities in the knowledge graph and the attributes of the key entities;

[0015] Based on the context information and the attributes of the key entities, identify the mutual relationships between the key entities;

[0016] Add the mutual relationships between the key entities to the knowledge graph in the form of edges to obtain an updated knowledge graph.

[0017] In another possible implementation manner of the first aspect, the identifying the mutual relationships between the key entities based on the context information and the attributes of the key entities includes:

[0018] Based on the context information and the attributes of the key entities, obtain the associated entities of the key entities within the same preset time period;

[0019] Construct a relationship grid of the associated entities, and based on the attributes and the relationship grid of the associated entities, determine the relationships between the key entities.

[0020] In another possible implementation manner of the first aspect, the constructing the relationship grid of the associated entities and determining the relationships between the key entities based on the attributes and the relationship grid of the associated entities includes:

[0021] According to the context information, obtain the direct relationships between the associated entities, and construct the relationship grid of the associated entities according to the direct relationships, where in the relationship grid, the associated entities are nodes and the direct relationships are edges;

[0022] Compare the attributes of the key entities and the attributes of the associated entities, and determine that there is a direct relationship between the key entity and the associated entity when the attribute values of the key entity and the associated entity in the same time period are in an inclusion relationship;

[0023] When there is an intermediate entity between the associated entity and the key entity, determine whether there is a direct relationship or an indirect relationship between the associated entity and the key entity. Among them, if the intermediate entity has a direct relationship with both the associated entity and the key entity, but there is no direct relationship between the associated entity and the key entity, then there is an indirect relationship between the associated entity and the key entity. If the intermediate entity has a direct relationship with both the associated entity and the key entity, and there is a direct relationship between the associated entity and the key entity, then there is a direct relationship between the associated entity and the key entity;

[0024] According to the direct relationship and / or indirect relationship between the associated entity and the key entity, determine the similarity index and anomaly index between the associated entity and the key entity, and according to the similarity index and anomaly index between the associated entity and the key entity, determine the relationship between the key entities.

[0025] In another possible implementation manner of the first aspect, the method further includes:

[0026] According to the relationship between the key entities, determine the edges of the second knowledge graph, and construct the second knowledge graph according to the edges of the second knowledge graph and the key entities, where the nodes in the second knowledge graph are the key entities, the edges of the second knowledge graph are the relationships between the key entities, and the relationships between the key entities include direct relationships and indirect relationships;

[0027] The determining the edges of the second knowledge graph according to the relationship between the key entities includes:

[0028] Calculate the similarity of the attribute values of the key entities connected to each edge in the same time period, and determine the weight of each edge;

[0029] Count the occurrence frequencies of the direct relationships and the indirect relationships between the key entities corresponding to each edge, and determine the length of each edge according to the occurrence frequency and the weight of each edge.

[0030] In another possible implementation manner of the first aspect, the determining at least one mine disaster rule based on the variable includes:

[0031] Use an association rule learning model to identify the association rules between the variables;

[0032] According to the association rules between the variables, determine the abnormal variables;

[0033] Determine the events related to the abnormal variables, and screen the events to obtain the mine disaster rules;

[0034] Among them, determining the abnormal variable according to the association rule between variables includes:

[0035] When the numerical relationship between the variables does not conform to the association rule, determining the variable as an abnormal variable.

[0036] In another possible implementation manner of the first aspect, determining the event related to the abnormal variable and screening the event to obtain the mine disaster rule includes:

[0037] Obtaining historical disaster records;

[0038] Performing correlation analysis on the historical disaster records and the abnormal variable to determine the event related to the abnormal variable in the historical disaster records;

[0039] Performing regression analysis on the event related to the abnormal variable to determine the disaster event related to the abnormal variable, where the disaster event and the abnormal variable are used to characterize the mine disaster rule.

[0040] In another possible implementation manner of the first aspect, determining the mine disaster with the highest occurrence probability in the current mine according to the updated knowledge graph and at least one of the mine disaster rules includes:

[0041] According to the updated knowledge graph, obtaining context information related to the input data, where the context information includes multiple target abnormal variables;

[0042] Calculating the conditional probability of each mine disaster rule occurring under the same context condition, where the same context condition means that the target abnormal variables are the same;

[0043] Taking the mine disaster with the highest conditional probability as the mine disaster with the highest occurrence probability in the current mine.

[0044] In a second aspect, the present application provides a machine-readable storage medium, on which instructions are stored, and the instructions are used to cause a machine to execute the above-mentioned intelligent analysis method for mine disasters based on a knowledge graph.

[0045] In a third aspect, the present application provides an electronic device, including:

[0046] A memory configured to store instructions; and

[0047] A processor configured to call the instructions from the memory and capable of implementing the above-mentioned intelligent analysis method for mine disasters based on a knowledge graph when executing the instructions.

[0048] Through the above technical solutions, the historical multi-source heterogeneous data of the mine is first obtained and integrated, which can comprehensively reflect the actual situation of the mine. By constructing a knowledge graph, the entities in the mine and the relationships between entities are visualized and structured, facilitating the understanding of the complexity of the mine environment. Through the structured relationships of the knowledge graph, key entities and their mutual relationships can be quickly identified based on the input data, contributing to the timely discovery of potential risks. In addition, by obtaining the input data in real time and inputting it into the knowledge graph to identify key entities, the dynamic changing environment can be quickly adapted to, and by extracting the mutual relationships between key entities, the knowledge graph is updated, effectively reducing the need for frequent reconstruction of prediction models and saving human and material resources. Extracting variables from the input data and determining mine disaster rules based on the variables can quickly generate prediction rules for the current environment, which is more flexible and adaptable than the existing models based on historical data.

[0049] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. Brief Description of the Drawings

[0050] Figure 1 It is a schematic flow chart of an intelligent analysis method for mine disasters based on a knowledge graph provided by an embodiment of the present application;

[0051] Figure 2 It is a schematic diagram of a knowledge graph provided by an embodiment of the present application;

[0052] Figure 3 It is a schematic diagram of another knowledge graph provided by an embodiment of the present application;

[0053] Figure 4 It is a schematic diagram of a relationship grid provided by an embodiment of the present application. Detailed Description of the Embodiments

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. It should be understood that the specific implementation manners described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0055] It should be noted that if there are directional indications (such as up, down, left, right, front, back...) involved in the embodiments of the present application, the directional indications are only used to explain the relative positional relationships and movement conditions between components in a specific posture (as shown in the drawings). If the specific posture changes, the directional indications will also change accordingly.

[0056] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, such descriptions of "first", "second", etc. are only for descriptive purposes and should not be construed as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions between various embodiments may be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0057] Figure 1 Schematically shows a schematic flow diagram of an intelligent analysis method for mine disasters based on a knowledge graph according to an embodiment of the present application. As Figure 1 shown, an embodiment of the present application provides an intelligent analysis method for mine disasters based on a knowledge graph, and the method may include the following steps.

[0058] S110. Obtain the historical multi-source heterogeneous data of the mine, integrate the historical multi-source heterogeneous data, and construct a knowledge graph. The historical multi-source heterogeneous data includes sensor data, historical disaster records, and geological data. The nodes of the knowledge graph are used to represent the entities of the mine, and the edges of the knowledge graph are used to represent the mutual relationships between the entities;

[0059] S120. In response to inputting the input data obtained in real time into the knowledge graph, identify the key entities;

[0060] S130. Extract the mutual relationships between the key entities, and add the mutual relationships between the key entities to the knowledge graph to obtain an updated knowledge graph;

[0061] S140. Extract the variables in the input data, and determine at least one mine disaster rule based on the variables;

[0062] S150. Determine the mine disaster with the highest occurrence probability in the current mine according to the updated knowledge graph and at least one mine disaster rule.

[0063] The historical multi-source heterogeneous data of the mine includes sensor data (such as rainfall, pressure, water level height, etc.), historical disaster records (such as water disasters, collapse events, etc.), and geological data (such as rock stratum structure, mine depth, etc.). In this embodiment, the historical multi-source heterogeneous data is integrated by means of data cleaning to ensure the consistency and availability of the data. For example, since the sensor data may come from different devices and have different units and formats, it is necessary to unify it into a standard format.

[0064] When constructing a knowledge graph, nodes are used to represent various entities in a mine, such as sensors, disaster events, and geological features, etc., while edges represent the relationships between entities, such as "rainfall affects water level height" or "the association between historical disaster records and current conditions". The knowledge graph can effectively reflect the complex environment and potential risks of the mine, such as Figure 2 as shown

[0065] Input the real-time acquired input data into the knowledge graph to identify key entities. The input data can be real-time monitoring data from sensors or other relevant environmental data. The process of identifying key entities includes analyzing the input data and extracting variables related to mine safety. For example, if it is real-time monitored that the rainfall reaches a preset threshold, the pressure is higher than the preset pressure threshold, and the water level height rises abnormally, then the variables of rainfall, pressure, and water level height are marked as key entities.

[0066] Extract the mutual relationships between key entities and add them to the knowledge graph to obtain an updated knowledge graph. The mutual relationships between key entities, such as the mutual influence between rainfall, pressure, and water level height, can use correlation analysis to identify the mutual relationships between key entities and add them to the knowledge graph in the form of edges. For example, if it is found that heavy rainfall and high pressure usually accompany the rise of the water level height, then an edge of "heavy rainfall and high pressure cause the water level height to rise" can be added to the knowledge graph. The updated knowledge graph can more comprehensively reflect the current state of the mine.

[0067] After that, extract the variables in the input data and determine at least one mine disaster rule based on the variables. In this embodiment, first identify the key variables in the input data, such as rainfall, pressure, and water level height, etc. Then, the relationships between the variables can be analyzed to determine potential mine disaster rules. For example, if it is found that "when the rainfall reaches a preset threshold and the pressure is higher than the preset pressure threshold, the water level height is often abnormal", then it can be defined as a mine disaster rule.

[0068] Based on the updated knowledge graph and at least one mine disaster rule, determine the mine disaster with the highest occurrence probability in the current mine. According to the updated knowledge graph and at least one mine disaster rule, the conditional probability of each mine disaster rule occurring can be calculated. For example, if the current rainfall is 3mm and the pressure is 100kPa, then the occurrence probability of water damage under this condition can be calculated. By comparing the conditional probabilities of different disaster rules, select the mine disaster with the highest probability as the current risk.

[0069] In summary, by integrating historical multi-source heterogeneous data, identifying key entities in real time, extracting interrelationships, and determining mine disaster rules based on variables, accurate prediction of mine disasters is ultimately achieved. This not only improves the efficiency of mine safety management but also provides a safer working environment for miners and reduces potential accident risks.

[0070] In this embodiment, historical multi-source heterogeneous data of the mine is first obtained and integrated, which can comprehensively reflect the actual situation of the mine. By constructing a knowledge graph, the entities and relationships between entities in the mine are visualized and structured, facilitating the understanding of the complexity of the mine environment. Through the structured relationships in the knowledge graph, key entities and their interrelationships can be quickly identified based on the input data, helping to detect potential risks in a timely manner. In addition, input data is obtained in real time and input into the knowledge graph to identify key entities, enabling rapid adaptation to a dynamically changing environment. By extracting the interrelationships between key entities and updating the knowledge graph, the need for frequent reconstruction of the prediction model is effectively reduced, saving human and material resources. Extracting variables from the input data and determining mine disaster rules based on the variables can quickly generate prediction rules for the current environment, which is more flexible and adaptable than existing models based on historical data.

[0071] In one implementation manner of this embodiment, the interrelationships between key entities are extracted and added to the knowledge graph to obtain an updated knowledge graph, including the following steps:

[0072] S210. Obtain the context information of the key entity in the knowledge graph and the attributes of the key entity;

[0073] S220. Based on the context information and the attributes of the key entity, identify the interrelationships between the key entities;

[0074] S230. Add the interrelationships between the key entities to the knowledge graph in the form of edges to obtain an updated knowledge graph.

[0075] The context information of the key entity in the knowledge graph includes the state of the key entity in a specific environment, such as real-time data on rainfall, pressure, and water level height. The context information can provide the current situation of the key entity and help understand its role in the mine environment. The attributes of the key entity include its basic characteristics, such as the type, location, and historical records of the sensor. For example, a rainfall sensor may be located at a specific depth in the mine, and the recorded rainfall range is from 0 mm to 10 mm. At this time, all context information related to the key entity can be extracted from the knowledge graph to form a comprehensive context view.

[0076] Based on the obtained context information and the attributes of key entities, identify the relationships between key entities. Specifically, data mining can be used to analyze the mutual influence between key entities. For example, if the pressure continues to rise while the water level height also shows abnormalities, it can be inferred that there is a certain relationship between the pressure and the water level height. By establishing a mathematical model and using statistical analysis methods, the strength and direction of the relationship between the pressure and the water level height can be quantified. For example, if it is found that "for every 1 unit increase in pressure, the water level height increases by 2 units", this discovery can be used as the relationship between key entities.

[0077] Add the relationships between the identified key entities to the knowledge graph in the form of edges to obtain an updated knowledge graph. For example, if the relationship "high pressure causes the water level height to rise" is identified, add an edge from the pressure node to the water level height node in the knowledge graph and label the strength and nature of this relationship. In this way, the knowledge graph can be continuously updated to form a dynamic relationship network that can reflect the changes in the mine environment in real time.

[0078] The process of extracting the relationships between key entities and updating the knowledge graph in this embodiment can effectively improve the intelligent level of mine safety management. By obtaining context information and key entity attributes, identifying relationships, and adding them to the knowledge graph in the form of edges, a dynamic and comprehensive knowledge network is formed. This not only enhances the understanding of the mine environment but also provides important support for real-time risk assessment and decision-making, helping to effectively reduce potential safety risks and ensure the safety of miners.

[0079] In one implementation manner of this embodiment, based on the context information and the attributes of key entities, identifying the relationships between key entities includes the following steps:

[0080] S310. Based on the context information and the attributes of key entities, obtain the associated entities of the key entities within the same preset time period;

[0081] S320. Construct a relationship grid of the associated entities, and based on the attributes of the associated entities and the relationship grid, determine the relationships between key entities.

[0082] Based on the context information and the attributes of the key entities, associated entities within the same preset time period are obtained. That is to say, other entities related to the key entities within a specific time range are extracted from the knowledge graph. For example, if the key entity is a rainfall sensor and its context information shows the rainfall data recorded by the sensor in the past hour, then other associated entities within the same time period need to be identified, such as pressure sensors, water level sensors, etc. For example, if the rainfall recorded by the rainfall sensor is 3 mm and the pressure recorded by the pressure sensor within the same time period is 100 kPa, then there is a potential association between these two entities. Specifically, through timestamp matching, the associated entities of the key entities within the same preset time period can be obtained to ensure the relevance of the extracted associated entities within the same time period.

[0083] After that, based on the obtained associated entities, a relationship grid is constructed. The relationship grid is a graphical data structure that can display the connections and interactions between different entities. For example, if the data of the rainfall sensor, pressure sensor, and water level sensor within the same time period are all extracted, then a grid can be constructed with nodes representing the sensors and edges representing the relationships between the entities. By analyzing the attributes of these edges, such as the correlation strength, influence degree, etc., the relationships between the key entities can be further determined. Suppose when the rainfall increases, the pressure also increases correspondingly, and the water level height may increase accordingly. The above relationships can be quantified through statistical models (such as regression analysis models). Finally, based on the analysis results of the relationship grid, the specific relationships between the key entities can be clarified, providing an important basis for subsequent risk assessment.

[0084] The process of this embodiment for identifying the mutual relationships between entities based on context information and key entity attributes can effectively improve the understanding of the mine environment. By obtaining associated entities within the same time period and constructing a relationship grid, the connections and influences between different entities can be clearly shown. It not only effectively enhances the accuracy of data analysis but also facilitates real-time monitoring and risk warning, helping managers to identify potential safety hazards in a timely manner, thus ensuring the safety of miners and the stable operation of the mine.

[0085] In one implementation manner of this embodiment, constructing a relationship grid of associated entities and determining the relationships between key entities based on the attributes of the associated entities and the relationship grid includes the following steps:

[0086] S410. According to the context information, obtain the direct relationships between the associated entities, and construct a relationship grid of the associated entities based on the direct relationships. In the relationship grid, the associated entities are nodes and the direct relationships are edges;

[0087] S420. Compare the attributes of the key entity and the associated entity, and determine that there is a direct relationship between the key entity and the associated entity when the attribute values of the key entity and the associated entity in the same time period are in an inclusion relationship;

[0088] S430. When there is an intermediate entity between the associated entity and the key entity, determine that there is a direct relationship or an indirect relationship between the associated entity and the key entity. Among them, if the intermediate entity has a direct relationship with both the associated entity and the key entity, but there is no direct relationship between the associated entity and the key entity, then there is an indirect relationship between the associated entity and the key entity. If the intermediate entity has a direct relationship with both the associated entity and the key entity, and there is a direct relationship between the associated entity and the key entity, then there is a direct relationship between the associated entity and the key entity;

[0089] S440. According to the direct relationship and / or indirect relationship between the associated entity and the key entity, determine the similarity index and anomaly index between the associated entity and the key entity, and determine the relationship between the key entities according to the similarity index and anomaly index between the associated entity and the key entity.

[0090] Obtain the direct relationship between the associated entities according to the context information and construct a relationship grid. Specifically, first identify the direct relationships between all associated entities. For example, the relationship between a rainfall sensor and a pressure sensor. By analyzing the data of the sensors in the same time period, the mutual influence between the associated entities can be determined. For example, if the data of the rainfall sensor shows an increase in rainfall and the data of the pressure sensor shows a corresponding change in pressure, it can be considered that there is a direct relationship between the two sensors, and a relationship grid can be constructed according to the direct relationship, where the nodes represent the associated entities and the edges represent the direct relationships between the associated entities. The relationship grid can visually display the mutual connections between the associated entities.

[0091] Compare the attributes of the key entity and the associated entity, and determine the direct relationship between the key entity and the associated entity when the attribute values are in an inclusion relationship. Specifically, assume that the key entity is a rainfall sensor with the attribute "rainfall range: 0mm to 10mm", and the associated entity is a pressure sensor with the attribute "working environment: rainfall range: 3mm to 8mm". In this case, the attribute value of the key entity includes the attribute value of the associated entity, so it can be judged that there is a direct relationship between the two, and at this time, the interaction between the key entity and the associated entity can be further confirmed.

[0092] In the case where there is an intermediate entity between the associated entity and the key entity, determine the direct or indirect relationship between the associated entity and the key entity. If there is an intermediate entity that has a direct relationship with both the key entity and the associated entity, but there is no direct relationship between the associated entity and the key entity, then it can be determined that there is an indirect relationship between the associated entity and the key entity. For example, if there is no direct relationship between a rainfall sensor and a pressure sensor, but both have a direct relationship with a humidity sensor, then it can be considered that there is an indirect relationship between the associated entity and the key entity. On the other hand, if the intermediate entity has a direct relationship with both the key entity and the associated entity, and there is also a direct relationship between them, then it can be confirmed that there is a direct relationship between the associated entity and the key entity.

[0093] Based on the direct and / or indirect relationship between the associated entity and the key entity, determine the similarity index and the anomaly index, and based on the similarity index and the anomaly index, determine the relationship between the key entities. The similarity index can be obtained by calculating the similarity of attribute values. For example, methods such as cosine similarity or Euclidean distance can be used to measure the similarity degree between two entities. At the same time, the anomaly index can be identified by analyzing the deviation points in the historical data.

[0094] For example, when the readings of sensor A are significantly higher or lower than the normal range, it can be marked as an anomaly. Through the similarity index and the anomaly index, the relationship between the key entities can be judged more accurately. For example, if the similarity index between the rainfall sensor and the pressure sensor is high and there are no abnormal situations, then it can be inferred that there is a strong mutual influence relationship between the rainfall sensor and the pressure sensor.

[0095] This embodiment not only enhances the accuracy of data analysis through the processes of constructing the relationship grid of associated entities, comparing attributes, analyzing the influence of intermediate entities, and determining similarity and anomaly indices, but also facilitates real-time monitoring and risk assessment, helps to identify potential safety hazards in a timely manner, and thus ensures the safe and stable operation of the mine.

[0096] In one implementation manner of this embodiment, the following steps are further included:

[0097] S510. Determine the edges of the second knowledge graph according to the relationship between the key entities, and construct the second knowledge graph according to the edges of the second knowledge graph and the key entities, where the nodes in the second knowledge graph are the key entities, the edges of the second knowledge graph are the relationships between the key entities, and the relationships between the key entities include direct relationships and indirect relationships;

[0098] Determining the edges of the second knowledge graph according to the relationship between the key entities includes:

[0099] S520. Calculate the similarity of the attribute values of the key entities connected to each edge within the same time period, and determine the weight of each edge;

[0100] S530. Count the occurrence frequencies of the direct and indirect relationships between the key entities corresponding to each edge, and determine the length of each edge according to the occurrence frequency and the weight of each edge.

[0101] Determine the edges of the second knowledge graph according to the relationships between the key entities, and construct the second knowledge graph. In this process, first determine the relationships between the key entities. A direct relationship refers to the direct interaction or influence between two key entities. For example, a rainfall sensor directly affects the working state of a water pump. An indirect relationship is a relationship that indirectly affects through other entities. For example, a rainfall sensor indirectly affects the reading of a pressure sensor by affecting environmental conditions such as water level height. After determining the relationships between the key entities, the key entities can be used as nodes, and the relationships between the key entities can be used as edges to construct the second knowledge graph. The second knowledge graph can clearly show the mutual connections between key entities.

[0102] After that, calculate the similarity of the attribute values of the key entities connected to each edge within the same time period, and determine the weight of each edge. Specifically, first, it is necessary to collect the attribute data of each key entity within a specific time period, such as rainfall, pressure, etc. Then, similarity calculation methods such as cosine similarity or Euclidean distance can be used to evaluate the similarity degree between the attributes of two key entities. Suppose the rainfall of key entity A is 3mm and the rainfall of key entity B is 5mm. Using Euclidean distance calculation, the similarity between the two can be obtained. The higher the similarity, the closer the two entities are in terms of attributes, so a higher weight can be assigned to the edge between the two key entities to quantify the relationship strength between the key entities.

[0103] Finally, count the occurrence frequencies of the direct and indirect relationships between the key entities corresponding to each edge, and determine the length of each edge according to the occurrence frequency and the weight of each edge. For example, it can be counted how many times the direct relationship between the rainfall sensor and the pressure sensor occurs, and whether there is an indirect relationship between the rainfall sensor and the pressure sensor through other entities (such as a water pump). Through statistics, the direct relationship frequency and indirect relationship frequency of each edge can be obtained.

[0104] Next, combine the occurrence frequency of each edge with the weight of each edge, and comprehensively consider the influence of the two to determine the length of each edge. The length of the edge can reflect the strength and stability of the relationship. The shorter the length, the closer the relationship.

[0105] In this embodiment, by determining the relationships between key entities, calculating similarities, and counting relationship frequencies, a structured second knowledge graph can be constructed. By clarifying the strength and nature of the relationships, potential risks can be better identified, thereby optimizing the management process and improving overall efficiency.

[0106] In one implementation of this embodiment, at least one mine disaster rule is determined based on variables, including the following steps:

[0107] S610. Use an association rule learning model to identify the association rules between variables;

[0108] S620. Determine the abnormal variables according to the association rules between variables;

[0109] S630. Determine the events related to the abnormal variables, and screen the events to obtain the mine disaster rules;

[0110] Among them, determining the abnormal variables according to the association rules between variables includes:

[0111] S640. When the numerical relationship between variables does not conform to the association rules, determine the variable as an abnormal variable.

[0112] Association rule learning is a data mining technique aimed at discovering the relationships between variables in a dataset. In this embodiment, the key variables may include rainfall, pressure, water level height, etc. Through the association rule learning model, the co-occurrence of key variables in historical data can be analyzed. Specifically, when implementing, the support and confidence thresholds are first set. The support represents the frequency of a certain rule appearing in the data, while the confidence measures the reliability of the rule. For example, if in 80% of the time, when the pressure is higher than 100 kPa, the rainfall is also higher than 3 mm, then the rule "high pressure means high rainfall" can be formed.

[0113] After that, according to the identified association rules, the abnormal variables can be determined. Suppose at a certain moment, the pressure is 120 kPa, and according to the previous association rule, when the pressure is high, the rainfall should be above 3 mm. If the rainfall is only 2 mm at this time, it means that this rainfall value does not conform to the association rule and may be abnormal. Specifically, a threshold can be set to judge whether a variable is abnormal. For example, if the relationship between rainfall and pressure does not meet the expectation within a certain range, it is marked as an abnormal variable, which can effectively identify the abnormal situations of potential mine disasters, so as to take timely measures to prevent risks.

[0114] Identify events related to abnormal variables and screen the events to obtain mine disaster rules. First, historical events related to abnormal variables can be analyzed, and events that frequently occur when the abnormal variables appear can be determined. For example, if events of increased pressure often accompany abnormal rainfall in historical data, then "pressure increases when rainfall is abnormal" can be regarded as a potential mine disaster rule. By screening the events, the most risky rules can be refined for subsequent monitoring and early warning. When implementing specifically, statistical analysis methods can be used to calculate the frequency and correlation of event occurrences, so as to ensure that the selected rules are representative and reliable enough.

[0115] In the process of refining and determining abnormal variables, first, a numerical relationship model between variables needs to be established and compared with association rules. When the variable values actually observed do not conform to the relationship expected by the model, the variable can be determined to be abnormal. For example, assume that under normal circumstances, the relationship between rainfall and pressure is linear, and for every 20 kPa increase in pressure, the rainfall should increase by 1 mm. If the actual observed pressure is 40 kPa, but the rainfall shows 0.5 mm, which is significantly lower than expected, then this rainfall value can be marked as abnormal. The above method can effectively identify potential risks related to mine safety and take preventive measures in a timely manner.

[0116] This implementation method can effectively construct mine disaster rules by using an association rule learning model to identify the relationships between variables, determine abnormal variables and their related events, and effectively improve the ability to identify potential risks. By discovering and dealing with abnormal situations in a timely manner, the occurrence probability of mine disasters can be significantly reduced, ensuring the safety of miners and the stable operation of the mine.

[0117] In one implementation manner of this embodiment, identifying events related to abnormal variables and screening the events to obtain mine disaster rules includes the following steps:

[0118] S710. Obtain historical disaster records, where the historical disaster records include event types, occurrence events, and influence ranges;

[0119] S720. Conduct an association analysis on the historical disaster records and abnormal variables to determine the events related to the abnormal variables in the historical disaster records;

[0120] S730. Conduct a regression analysis on the events related to the abnormal variables to determine the disaster events related to the abnormal variables. The disaster events and the abnormal variables are used to characterize the mine disaster rules.

[0121] In this embodiment, historical disaster records are first obtained, including information such as event types, occurrence events, and affected ranges. Event types may include water disasters, collapses, fires, etc., and the occurrence event refers to the specific situation of the disaster occurrence, such as a certain event of excessive water level height. The affected range describes the degree of impact of the disaster on the mine and the surrounding environment, such as which mining areas are affected and how much economic loss is caused. During specific implementation, a database management system can be used to organize the historical disaster records into a structured data set for subsequent analysis.

[0122] By performing correlation analysis on historical disaster records and abnormal variables, events related to the abnormal variables are determined. For example, assume that when the water level height is abnormal, there are often records of water disasters. Methods such as chi-square test or correlation coefficient analysis can be used to evaluate the correlation degree between the abnormal variable and various events. If the occurrence frequency of a certain event increases significantly when the abnormal variable appears, it can be regarded as an event related to the abnormal variable. During specific implementation, the relationship between different events and the abnormal variable can be visually displayed to facilitate the rapid identification of high-risk events.

[0123] Perform regression analysis on the events related to the abnormal variable to determine the disaster events related to the abnormal variable. Methods such as linear regression or logistic regression can be used to perform regression analysis on the events related to the abnormal variable to determine the disaster events related to the abnormal variable. During specific implementation, first, the abnormal variable is used as the independent variable, and the historical disaster event is used as the dependent variable to construct a regression model. By fitting the regression model, the degree of impact of the abnormal variable on the occurrence of the disaster event can be evaluated. For example, assume that in the regression analysis, it is found that for every one-unit increase in the water level height, the probability of a water disaster increases significantly, indicating that the water level height is an important risk factor.

[0124] Through regression analysis in this embodiment, not only can it be confirmed whether there is a significant relationship between the abnormal variable and the disaster event, but also the relationship between the abnormal variable and the disaster event can be quantified, providing a scientific basis for formulating mine disaster rules.

[0125] By obtaining historical disaster records, performing correlation analysis and regression analysis in this implementation method, mine disaster events related to abnormal variables can be systematically identified, providing important support for mine safety management, being able to timely discover potential risks, and formulate corresponding preventive measures. By establishing scientific mine disaster rules, the probability of accidents can be effectively reduced, ensuring the safety of miners and the stable operation of the mine.

[0126] In one implementation manner of this embodiment, according to the updated knowledge graph and at least one mine disaster rule, the mine disaster with the highest occurrence probability in the current mine is determined, including the following steps:

[0127] S810. Obtain context information related to the input data according to the updated knowledge graph. The context information includes multiple target anomaly variables.

[0128] S820. Calculate the conditional probability of each mine disaster rule occurring under the same context condition, where the same context condition means that the target anomaly variables are the same.

[0129] S830. Take the mine disaster with the maximum conditional probability as the mine disaster with the highest occurrence probability in the current mine.

[0130] According to the updated knowledge graph, obtain context information related to the input data. The context information related to the input data includes multiple target anomaly variables. The knowledge graph is a graphical representation constructed by integrating different data sources, which can clearly show the relationships between variables and their contexts.

[0131] Specifically, first, it is necessary to analyze the operating status of the current mine and collect real-time data related to mine safety, such as water level height, rainfall, pressure, etc. Then input the input data into the knowledge graph to identify the current target anomaly variables. In specific implementation, data mining techniques can be used to match the input data with the context information in the knowledge graph. For example, if it is detected that the water level height has increased abnormally, it is taken as a target anomaly variable. This implementation method can ensure that the obtained context information is up-to-date and relevant, making the risk assessment more accurate.

[0132] Calculate the conditional probability of each mine disaster rule occurring under the same context condition. The conditional probability refers to the probability of an event occurring under specific conditions, which can reflect the occurrence possibilities of various mine disaster rules when the target anomaly variables are the same.

[0133] Give an example to illustrate the calculation of the conditional probability of each mine disaster rule occurring under the same context condition:

[0134] Suppose there are the following anomaly variables and mine disaster rules:

[0135] Anomaly variables: water level height, high; rainfall, high; Mine disaster rules: Disaster A: fire; Disaster B: collapse; Disaster C: water disaster.

[0136] First, determine the context condition: Select the target anomaly variable, such as "high rainfall"; Second, in the historical records, count the number of times each disaster occurs when the rainfall is high. For example, the number of times of fire occurrence: 0 times; the number of times of collapse occurrence: 2 times; the number of times of water disaster occurrence: 8 times; Calculate the conditional probability of various disasters when the rainfall is high. Assume that there are a total of 10 records of high rainfall.

[0137] The conditional probability is calculated as follows:

[0138] P1 (Fire) = 0, P2 (Collapse) = 0.2, P3 (Water Hazard) = 0.8.

[0139] Under this context condition, the conditional probability of water hazard is the highest (0.8). Therefore, it can be considered that under the condition of high rainfall, the most likely mine disaster to occur in the current mine is water hazard.

[0140] By obtaining context information, calculating conditional probabilities, and identifying the most likely mine disasters to occur, this embodiment can effectively improve mine safety, contribute to timely identification of potential risks, and can significantly reduce the probability of mine accidents by establishing a mine disaster early warning mechanism based on data analysis, ensuring the safety of miners and the stable operation of the mine.

[0141] For ease of understanding, the embodiments of this application are illustrated by examples below:

[0142] Suppose in a certain mine, the following key variables are monitored: rainfall; pressure; water level height; historical disaster records (such as: water hazard, collapse).

[0143] Refer to Figure 3 , the sensor data (such as rainfall, pressure, water level height) and historical disaster records collected are integrated into the knowledge graph. In the knowledge graph, nodes represent entities (such as rainfall, pressure, water level height, disaster type), and edges represent the relationships between key entities.

[0144] Obtain the context information of the current rainfall, pressure, and water level height, and identify the relationship that "high rainfall and high pressure may cause the water level height to increase", and add it to the knowledge graph. Based on the relationships between entities, construct the Figure 4 relationship grid shown, and analyze the direct and indirect relationships.

[0145] If through analysis, it is found that when the rainfall exceeds 3mm and the pressure exceeds 80kPa, the water level height is often abnormal. Under this condition, if the water level height exceeds the safety threshold, it is marked as an abnormal variable. Conduct a correlation analysis between the water level height and historical disaster records, and find that a water hazard has occurred under this condition. And determine the mine disaster rule as "high rainfall, high pressure, abnormal water level height may cause water hazard".

[0146] As Figure 3 shown, the nodes are rainfall, pressure, water level height, water hazard (disaster type), historical disaster records, and the edges are the influence of rainfall and pressure on the water level height, the causal relationship between the water level height and water hazard, and the correlation between historical disaster records and water hazard.

[0147] An embodiment of the present application also provides a machine-readable storage medium, on which instructions are stored, and the instructions are used to cause a machine to execute the above-mentioned intelligent analysis method for mine disasters based on a knowledge graph.

[0148] An embodiment of the present application also provides an electronic device, including:

[0149] A memory configured to store instructions; and

[0150] A processor configured to call instructions from the memory and, when executing the instructions, be able to implement the above-mentioned intelligent analysis method for mine disasters based on a knowledge graph.

[0151] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0152] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0153] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0154] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide for implementing the functions in the processFigure 1 one or more processes and / or blocks Figure 1 steps of the functions specified in one or more blocks

[0155] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0156] The memory may include non-permanent memory in the computer-readable medium, in the form of random access memory (RAM) and / or non-volatile memory such as read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0157] Computer-readable media includes permanent and non-permanent, removable and non-removable media implemented by any method or technology for information storage. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0158] It should also be noted that the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising a..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0159] The above are only embodiments of the present application and are not used to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for intelligent analysis of mine disasters based on knowledge graph, characterized in that: include: Acquire historical multi-source heterogeneous data of the mine, integrate the historical multi-source heterogeneous data, and construct a knowledge graph, wherein the nodes of the knowledge graph are used to represent entities of the mine, and the edges of the knowledge graph are used to represent the relationships between the entities; In response to inputting the input data acquired in real time into the knowledge graph, identifying key entities; Obtaining context information of the key entity in the knowledge graph and attributes of the key entity; Based on the context information and the attributes of the key entity, acquiring the associated entities of the key entity within the same preset time period; Constructing a relationship grid of the associated entities, and determining the mutual relationships between the key entities based on the attributes of the associated entities and the relationship grid, wherein the mutual relationships include direct relationships and indirect relationships; Adding the mutual relationships between the key entities to the knowledge graph in the form of edges to obtain an updated knowledge graph; extracting variables from the input data and determining at least one mine hazard rule based on the variables; Determining the mine disaster with the highest probability of occurring in the current mine according to the updated knowledge graph and at least one of the mine disaster rules; After obtaining the updated knowledge graph, it also includes: According to the mutual relationships between the key entities, the edges of the second knowledge graph are determined, and according to the edges of the second knowledge graph and the key entities, the second knowledge graph is constructed, wherein the weight of each edge of the second knowledge graph is determined by calculating the similarity of the attribute values ​​of the key entities connected to each edge within the same time period, and the length of each edge is determined by counting the occurrence frequency of the direct relationship and the indirect relationship between the key entities corresponding to each edge, and based on the occurrence frequency and the weight of each edge.

2. The method according to claim 1, characterized in that The step of constructing a relationship grid of the associated entities and determining the mutual relationships between the key entities based on the attributes of the associated entities and the relationship grid includes: According to the context information, a direct relationship between related entities is acquired, and a relationship grid of the related entities is constructed according to the direct relationship, wherein in the relationship grid, the related entities are nodes and the direct relationship is an edge; Comparing the attribute of the key entity with the attribute of the associated entity, and determining that there is a direct relationship between the key entity and the associated entity when the attribute values ​​of the key entity and the associated entity in the same time period are in a containment relationship; In the case where there is an intermediate entity between the associated entity and the key entity, determining whether there is a direct relationship or an indirect relationship between the associated entity and the key entity, wherein if the intermediate entity has a direct relationship with both the associated entity and the key entity, but there is no direct relationship between the associated entity and the key entity, then there is an indirect relationship between the associated entity and the key entity; if the intermediate entity has a direct relationship with both the associated entity and the key entity, and there is a direct relationship between the associated entity and the key entity, then there is a direct relationship between the associated entity and the key entity; Based on the direct relationship and / or indirect relationship between the associated entity and the key entity, the similarity index and the abnormality index between the associated entity and the key entity are determined, and based on the similarity index and the abnormality index between the associated entity and the key entity, the mutual relationship between the key entities is determined.

3. The method according to claim 1, characterized in that The determining at least one mine hazard rule based on the variable comprises: identifying association rules between the variables using an association rule learning model; Determining abnormal variables according to the association rules between the variables; Determining events related to the abnormal variables, and screening the events to obtain mine disaster rules; Wherein, determining abnormal variables according to the association rules between the variables includes: When the numerical relationship between the variables does not conform to the association rule, the variables are determined to be abnormal variables.

4. The method according to claim 3, characterized in that The determining of events related to the abnormal variables and screening the events to obtain mine disaster rules includes: Access historical disaster records; Performing correlation analysis on the historical disaster records and the abnormal variables to determine events in the historical disaster records that are related to the abnormal variables; Regression analysis is performed on events related to the abnormal variables to determine disaster events related to the abnormal variables, and the disaster events and the abnormal variables are used to characterize the mine disaster rules.

5. The method according to claim 1, characterized in that The determining, according to the updated knowledge graph and at least one of the mine disaster rules, the mine disaster with the highest probability of occurring in the current mine comprises: According to the updated knowledge graph, context information related to the input data is obtained, wherein the context information includes a plurality of target abnormal variables; Calculate the conditional probability of each mine disaster rule occurring under the same context condition, where the same context condition means that the target abnormal variable is the same; The mine disaster with the largest conditional probability is regarded as the mine disaster with the highest probability of occurrence in the current mine.

6. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores instructions for enabling a machine to execute the knowledge graph-based intelligent analysis method for mine disasters according to any one of claims 1 to 5.

7. An electronic device, characterized in that: include: a memory configured to store instructions; as well as A processor is configured to call the instruction from the memory and to implement the knowledge graph-based intelligent analysis method for mine disasters according to any one of claims 1 to 5 when executing the instruction.

Images