Tailing dam accident path identification method based on mixed causal knowledge graph and related device

Abstract

The invention belongs to the field of tailing pond safety risk identification and accident analysis, and discloses a tailing dam accident path identification method based on a mixed causal knowledge graph and a related device, and the method comprises the steps: building a mixed causal logic model through employing an obtained tailing dam accident initial hazard source; solving a set of accident cause links of the tailing dam by utilizing the mixed causal logic model; in the set of the tailing dam accident cause links, the logic relation of the tailing dam accident and the accident cause links are used as nodes and edges of a knowledge graph respectively, and a Python + Neo4j method is used for constructing a tailing dam accident knowledge graph; and coding and classifying the tailing dam accident knowledge graph by using a BERT-GCN-SVM model to realize tailing dam accident path identification. The mine tailing dam accident path identification method is improved and optimized based on the thought of the knowledge graph, a set of mine tailing dam accident chain identification method is formed, and the mine tailing dam accident path identification method can be used as a new tool for mine tailing dam accident directional control and disaster situation deduction.

Description

Technical Field

[0001] The present invention belongs to the field of tailings dam safety risk identification and accident analysis, and specifically relates to a tailings dam accident path identification method and related devices based on a hybrid causal knowledge graph. Background Art

[0002] A large amount of tailings generated by mineral processing operations are usually stored in tailings ponds. As production operations continue, the tailings pond will accumulate potential energy as the slope gradually rises in altitude. The structure between the sand and gravel of the tailings is loose, and when the water content is too high, it is easy to cause fluidity and structural damage, and there is a risk of dam break after a large amount of precipitation. In recent years, the production capacity of mines has gradually increased, and the large volume, high potential energy, and high risk characteristics of tailings ponds have become more and more obvious. There are many tailings ponds at present. Among all tailings ponds, abnormal ponds account for about 40%, and most of them have the risk of dam break. Tailings ponds are long-term large-scale projects with complex causes of accidents. It is difficult to judge and predict their evolution process and the mechanism of action of factors with a single monitoring method. Therefore, how to effectively identify the disaster path of tailings dam accidents, assess the risk of tailings dam accidents, and then support risk investigation and transformation decisions is an important and urgent issue facing the current engineering and academic communities. Summary of the invention

[0003] In order to solve the problems existing in the prior art, the purpose of the present invention is to provide a tailings dam accident path identification method and related devices based on a hybrid causal knowledge graph. Based on the idea of ​​knowledge graph, the present invention improves and optimizes the tailings dam accident path identification method to form a set of tailings dam accident chain identification methods, which can be used as a new tool for directional prevention and control of tailings dam accidents and disaster deduction.

[0004] To achieve the above purpose, the technical solution adopted by the present invention is as follows: The tailings dam accident path identification method based on hybrid causal knowledge graph includes: A mixed causal logic model is constructed using the acquired initial hazard sources of tailings dam accidents; Using the hybrid causal logic model, solving the set of causal links of tailings dam accidents; In the set of tailings dam accident causal links, the logical relationship of tailings dam accidents and accident causal links are used as nodes and edges of the knowledge graph respectively, and the knowledge graph of tailings dam accidents is constructed using Python+Neo4j method; The tailings dam accident knowledge graph is encoded and classified using the BERT-GCN-SVM model to achieve tailings dam accident path identification.

[0005] Preferably, the process of obtaining the initial hazard source of the tailings dam accident includes: Analyze historical accident data, take the cause of the accident as part of the initial hazard source set, and distinguish the life cycle of the tailings pond to form an initial hazard source set including human factors, technology, environment and management. Specifically: The initial hazard sources of tailings dam accidents are analyzed from the tailings dam construction stage, tailings pond operation stage and tailings pond closure stage; During the construction phase of tailings dams, the environmental factors that affect the safety of tailings dams mainly include hydrological and meteorological conditions, earthquakes, the stability of the mountains around the dam, and the physical properties of the tailings; technical factors mainly exist in the initial dam construction process, in order to save costs, various technical indicators are not achieved, and inappropriate structural measures are taken; management factors mainly include laws and regulations, safety investment, safety evaluation and safety supervision; During the operation phase of the tailings dam, the environmental factors that affect the safety of the tailings dam mainly include hydrological and meteorological conditions and earthquakes; technical factors are mainly caused by equipment failure or inappropriate structure; management factors mainly include safety training, safety investment, safety evaluation and safety supervision; During the closure stage of the tailings pond, the environmental factors that affect the safety of the tailings pond mainly include extreme weather and instability of the mountains around the pond; the technical factors are mainly those caused by equipment failure or inappropriate structure; the management factors mainly include imperfect closure regulations, insufficient investment in closure safety, inadequate safety evaluation, insufficient safety supervision and inadequate emergency management.

[0006] Preferably, the hybrid causal logic model includes an event sequence diagram layer, a fault tree layer and a Bayesian network layer, the event sequence diagram layer as the top layer, the fault tree layer as the middle layer, and the Bayesian network layer as the third layer, wherein the event sequence diagram layer is used to describe risk logic and calculate accident probability, the fault tree layer is used to find the causes of key events, and the Bayesian network layer is used to evaluate the impact of variability and correlation events.

[0007] Preferably, when using the hybrid causal logic model to solve the set of tailings dam accident causal links, the risk source that may occur in the tailings dam is used as the starting event, and different results are evolved through the hybrid causal logic model to obtain different accident causal links of the tailings dam. The different accident causal links of the tailings dam constitute the set of tailings dam accident causal links.

[0008] Preferably, in the set of tailings dam accident causal links, the logical relationship of the tailings dam accident and the accident causal link are used as nodes and edges of the knowledge graph respectively, and the process of constructing the tailings dam accident knowledge graph using the Python+Neo4j method includes: The Python+Neo4j method is used to transform the node-to-node logical relationship obtained from the hybrid causal logic model into a tailings dam accident knowledge graph based on hybrid causal logic. The entity construction of the tailings dam accident revolves around the mechanism of the accident, and the entities are set according to the human factors, technology, environment and management categories, and the relationships are set according to the cause and effect, time, inclusion and condition categories.

[0009] Preferably, in the BERT-GCN-SVM model: Using the pre-trained language model BERT to encode text can learn the semantic information of words; The graph convolutional neural network (GCN) is used to encode and model the subgraph of each path in the tailings dam accident knowledge graph. The SVM model is used as a classifier to classify the combination of node semantic information and path subgraph in the tailings dam accident knowledge graph, and output labels and probabilities for path prediction and identification of new entities.

[0010] The present invention also provides a tailings dam accident path identification system based on a hybrid causal knowledge graph, which is used to implement the tailings dam accident path identification method based on a hybrid causal knowledge graph of the present invention, including: Hybrid causal logic model modeling unit: used to construct a hybrid causal logic model using the acquired initial hazard source of tailings dam accidents; Calculation module: used to solve the set of causal links of tailings dam accidents by using the hybrid causal logic model; Knowledge graph construction module: used to take the logical relationship of tailings dam accidents and accident causation links in the set of tailings dam accident causation links as nodes and edges of the knowledge graph respectively, and use Python+Neo4j method to construct the tailings dam accident knowledge graph; Classification unit: used to encode and classify the tailings dam accident knowledge graph using the BERT-GCN-SVM model to achieve tailings dam accident path identification.

[0011] The present invention also provides an electronic device, comprising: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the tailings dam accident path identification method based on the hybrid causal knowledge graph as described above in the present invention.

[0012] The present invention also provides a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the tailings dam accident path identification method based on a hybrid causal knowledge graph as described above in the present invention is implemented.

[0013] The present invention also provides a computer program product, which includes computer instructions, and the computer instructions instruct the computer to execute the tailings dam accident path identification method based on the hybrid causal knowledge graph as described above.

[0014] The present invention has the following beneficial effects: Based on the extensive accident data obtained from the survey, the present invention uses a hybrid causal logic model to construct an accident chain set, and on this basis, establishes an accident knowledge graph, which is visualized and analyzed using Neo4j. The BERT-GCN-SVM model is used to accurately and comprehensively identify the causal paths in the tailings dam accident knowledge graph, fully detect possible risks, greatly optimize the supporting theory for tailings dam accident prevention and deduction, and provide better generalization performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flowchart of the steps of the tailings dam accident path identification method based on the hybrid causal knowledge graph provided by the present invention; Figure 2 It is a tailings dam accident event sequence diagram model provided by the present invention; Figure 3 It is a flow chart for establishing a hybrid causal logic model provided by the present invention; Figure 4 The invention provides a method for constructing a knowledge graph of tailings dam accidents; Figure 5 It is the result of constructing the knowledge graph of tailings dam accidents provided by the present invention; Figure 6 This is the BERT-GCN-SVM model architecture used in this invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] The present invention provides a tailings dam accident path identification method based on a hybrid causal knowledge graph. Firstly, a hybrid causal logic model of tailings dam accidents is constructed based on methods such as event sequence diagram, fault tree and Bayesian network to obtain a set of accident causal links; then, according to the tailings dam accident causal links, a Python+Neo4j method is used to form a tailings dam accident knowledge graph; finally, based on the formed knowledge graph, a BERT-GCN-SVM deep learning model is used to effectively identify key causal chains in the tailings dam accident knowledge graph.

[0018] For details, see Figure 1 The present invention constructs an accident cause knowledge graph based on a hybrid causal logic model, and uses natural language-knowledge graph multimodal information optimization identification. The tailings dam accident path identification method based on the hybrid causal knowledge graph of the present invention specifically includes the following steps: Step 1: Identify the initial hazard sources of tailings dam accidents from the four aspects of human factors, technology, environment and management. Specifically, analyze historical accident data, take the causes of accidents as part of the initial hazard source set, and form an initial hazard source set based on different stages in the life cycle of the tailings pond, including human factors, technology, environment and management; define typical accident patterns as risk scenarios; extract risk information based on accident data, define risk evolution as the people and objects involved in the accident, including risk groups, accident types, and form a set of basic elements of the accident chain network together with hazard sources.

[0019] Step 2: Construct a hybrid causal logic model based on the initial hazard source of the tailings dam accident. Among them, the event sequence layer of the hybrid causal logic model is used to describe the risk logic and calculate the accident probability, the fault tree layer is used to find the causes of key events, and the Bayesian network layer is used to evaluate the impact of variability and correlation events; Specifically: Based on the basic element set of the accident chain network finally formed in step 1, the deterministic causal relationship and the uncertain causal relationship are identified respectively: Identify the deterministic causal relationship, for each risk scenario, use the event sequence diagram and fault tree analysis method to establish the causal logic relationship from the initial hazard source to the accident result, and determine the failure mode and safety barrier; Identify the uncertain causal relationship, for human factors and organizational aspects, use the Bayesian network to establish the impact model, and determine the relationship between the influencing factors; Combine the deterministic causal relationship and the uncertain causal relationship, and establish a hybrid causal model for the tailings dam accident on the above basis.

[0020] Step 3: Use the hybrid causal logic model constructed in step 2 to solve the set of causal links of the tailings dam accident; specifically, it includes: using the hybrid causal logic model to evolve different results, and then obtain different sets of accident causal links, and quantify the probability of occurrence of related accident causal links through the quantitative analysis principle of the event sequence diagram model. Among them, the evolution process of the tailings dam accident from the initial event to the final result is described by the event sequence layer in the hybrid causal model, the fault tree layer is used to analyze and evaluate the potential causes of the event, and the Bayesian network layer is used to represent the uncertainty relationship between events; on this basis, according to the principle of quantitative analysis of the event sequence diagram model, the probability of occurrence of related accident causal links is quantified.

[0021] Step 4: The tailings dam accident causal links and accident logical relationships in the tailings dam accident causal link set are used as nodes and edges of the knowledge graph, and the knowledge graph is constructed using Python+Neo4j; specifically, the following steps are used: using the Python+Neo4j method to convert the nodes and node logical relationships obtained by the hybrid causal logic model into a tailings dam accident knowledge graph, and to express the knowledge structure and semantic information of the tailings dam accident in the form of entities and relationships. The entity construction of the tailings dam accident revolves around the mechanism of the accident, and the entities are set in the categories of "human factors, technology, environment, and management", and the relationships are set in the categories of "causality, time, inclusion, and conditions". The Neo4j graph database is used to realize the construction and visualization of the knowledge graph.

[0022] Step 5. Encode and classify the knowledge graph constructed in step 4 using the BERT-GCN-SVM model to identify the path of the tailings dam accident. Specifically, the following steps are used: the text is encoded using the pre-trained large language model BERT to learn the semantic information of words; the subgraph of each path of the knowledge graph is encoded and modeled using the graph convolutional network (GCN). For example, there is usually a clear correlation between the initial dormant hidden dangers and the secondary threat hidden dangers and the final active accidents, and this relationship can be captured and utilized by the graph convolutional network (GCN); the support vector machine (SVM) model is used as a classifier to classify the combination of node semantic information and path subgraphs, output their labels and probabilities, and the trained support vector machine (SVM) model is used for path prediction and identification of new entities.

[0023] The present invention provides the above-mentioned tailings dam accident path identification method based on hybrid causal knowledge graph, which is applicable to the field of tailings pond safety risk identification and accident analysis, and is particularly applicable to tailings dam accident cause analysis and risk assessment. Example like Figure 1 As shown, this embodiment takes a gold mine tailings pond rainy season risk identification project in Shaanxi, China as a reference, and provides a tailings dam accident path identification method based on a hybrid causal knowledge graph, including the following steps: Step 1: Identify the initial hazard sources of tailings dam accidents from four aspects: human factors, technology, environment, and management: Among them, according to the statistics of typical tailings dam accidents that occurred from 2001 to 2020, it was analyzed that there are four main types of causes of tailings dam accidents, namely: seepage damage 50.72%, surge overtopping 34.78%, slope instability 8.70%, and earthquake disasters 5.80%. Among them, the danger sources of tailings dams can be analyzed from several different periods, namely, problems left over from the design period, hidden dangers during operation, and hidden dangers in the closure stage: Tailings dam construction stage: The environmental factors that affect the safety of tailings ponds mainly include hydrological and meteorological conditions, earthquakes, and the stability of the mountains around the pond; technical factors mainly exist in the initial dam construction process, when in order to save costs, various technical indicators are not met and inappropriate structural measures are taken; management factors mainly include laws and regulations, safety investment, safety evaluation and safety supervision.

[0024] Tailings dam operation stage: The environmental factors affecting the safety of tailings dams mainly include hydrological and meteorological conditions and earthquakes; technical factors are mainly caused by equipment failure or inappropriate structure; management factors mainly include safety training, safety investment, safety evaluation and safety supervision.

[0025] Tailings dam closure stage: The environmental factors that affect the safety of tailings dams mainly include extreme weather and instability of the mountains around the dam; technical factors are mainly caused by equipment failure or inappropriate structure; management factors mainly include imperfect closure regulations, insufficient investment in closure safety, inadequate safety evaluation, insufficient safety supervision and inadequate emergency management.

[0026] Based on the above, the set of hazard sources and trigger mechanisms is sorted out, as shown in Table 1: Table 1

[0027] Through the initial hazard source analysis table, according to the risk characteristics of the three periods of tailings pond construction, the initial hazard sources are obtained from the four perspectives of cause, technology, environment and management as the basis for constructing the mixed causal logic model.

[0028] Step 2: Build a hybrid causal logic model: The event sequence diagram layer describes the risk logic and calculates the accident probability, the fault tree layer finds the causes of key events, and the Bayesian network layer evaluates the impact of variability and correlation events. The event sequence diagram is the top layer, the fault tree of key events in the event sequence diagram is the middle layer, and the Bayesian network model of human factors is the third layer.

[0029] First, construct the time series graph nodes. The node references are shown in Table 2: Table 2

[0030] Combined with the specific situation of the embodiment, a sequence diagram model of dam overtopping accident is constructed, such as Figure 2 shown.

[0031] On this basis, the Bayesian network layer is used to identify non-deterministic causality. The Bayesian network parameters are derived from the human factors parameters identified in the accident database and time series diagram, including work concentration, insufficient qualifications of construction units, lack of safety knowledge and skills assessment, decision-making ability, experience, work concentration, self-study, communication, training or drills, panic, stress, and lack of concentration.

[0032] On this basis, a hybrid causal logic model is constructed. The model structure is as follows: Figure 3 shown.

[0033] Step 3: Use the constructed hybrid causal logic model to solve the set of causal links of tailings dam accidents.

[0034] On the basis of the hybrid causal logic model constructed in step 2, the possible risk source of the tailings dam is taken as the starting event, different results are evolved, and then different accident cause links are obtained.

[0035] Combined with the specific situation of the embodiment, the cause link of the tailings dam accident under the rainfall situation is evolved, and the evolution results are shown in Table 3: Table 3

[0036] According to the probability of occurrence of the accident cause chain, the accident paths with serial numbers 23 and 24 have higher probability of occurrence, which are 0.029 and 0.138 respectively. Both paths involve events of whether the flood drainage facilities meet the requirements. This shows that in rainy weather, flood drainage facilities are prone to failure. The probability of occurrence of the accident paths with serial numbers 32 and 33 is 9.465×10 -16 5.550×10 -25 , where the accident path No. 33 has the lowest probability of occurrence. Both paths involve the occurrence of piping, indicating that the risk of seepage damage to the tailings dam is relatively low.

[0037] Step 4: Based on the hybrid causal logic model, the logical relationship and accident cause link of the tailings dam accident are used as the nodes and edges of the tailings dam accident knowledge graph, and Python+Neo4j is used to build the tailings dam accident knowledge graph: The tailings dam accident knowledge graph consists of nodes (entities) and edges (relationships between entities). It is a structured semantic network knowledge base that describes concepts, entities and their relationships in the objective world in the form of triples. Its representation is: (1) in, It is the entity set in the knowledge graph of tailings dam accidents, representing the entities in the knowledge graph of tailings dam accidents. different entities, It is a collection of relations in the knowledge graph, representing the relationships between entities. Different kinds of relationships. Represents a set of triples, where a triple sequence is generally composed of Represents the type of "entity-relationship-entity" and is used to describe the relationship between entities.

[0038] This embodiment adopts a top-down model layer construction method, using the hybrid causal model constructed in step 3, and uses the Python+Neo4j method to convert the node-node logical relationship obtained by the hybrid causal logic model into a tailings dam accident knowledge graph based on hybrid causal logic, expressing the knowledge structure and semantic information of the tailings dam accident in the form of entities and relationships.

[0039] The process of building the knowledge graph of tailings dam accidents is as follows: Figure 4 As shown: After determining the knowledge graph of tailings dam accidents, Neo4j was used for implementation and visualization.

[0040] Use Cypher language to create nodes and relationship edges one by one in the visual interface and import them into the Neo4j graph database.

[0041] First, establish entity nodes: For each entity related to the tailings dam accident, such as tailings dam, personnel, and environmental factors, establish corresponding entity nodes in the graph database in the form of nodes. Check whether there is already a node of the same type and name. If so, get the existing node; if not, create a new node and add it to the graph database.

[0042] Establish relationship edges between entity nodes to describe the connection between them. Use relationship edges to represent complex associations such as causal relationships, temporal relationships, inclusion relationships, and conditional relationships in tailings dam accident knowledge. Ensure that the establishment of relationship edges can clearly reflect the association relationships of actual tailings dam accidents.

[0043] In combination with the specific situation of the embodiment, a knowledge graph of tailings dam accidents is constructed. The graph contains 89 nodes and 175 relationships. Different types of entities are distinguished by nodes of different colors. Figure 5 shown.

[0044] Step 5: Use the BERT-GCN-SVM model to encode and classify the obtained tailings dam accident knowledge graph to realize tailings dam accident path identification.

[0045] The BERT-GCN-SVM model structure is as follows Figure 6 As shown, it is used to process and analyze the knowledge graph information of tailings dam accidents: First, the pre-trained language model BERT is used to encode the text, which can learn the semantic information of words. For example, the word "the infiltration line is too high" may contain the semantics of "tailings dam danger", thereby enhancing the representation ability of text information. Specifically, the text description of each node in the knowledge graph is converted into a 64-dimensional semantic vector with a length of less than 256 characters. Then, the BERT model is used to obtain the last hidden state of the text, and the output of the CLS (i.e. classification) tag is extracted as the feature representation of the text to generate a comprehensive and semantically rich vector.

[0046] Subsequently, GCN (graph convolutional neural network) is used to encode and model the subgraph of each path in the knowledge graph. For example, the clear correlation between the initial dormant hidden dangers and secondary threat hidden dangers and the final active accidents is captured; the path subgraphs in the knowledge graph are encoded and each subgraph is converted into a 32-dimensional network vector.

[0047] The SVM (support vector machine) model is used as a classifier to classify the combination of node semantic information and path subgraphs in the knowledge graph, and its labels and probabilities are output for path prediction and identification of new entities.

[0048] According to the specific information of the embodiment, a new node "structure collapse" is input into the BERT-GCN-SVM model, and the relevant path identification is performed on the knowledge graph to obtain the tailings dam disaster path identification results as shown in Table 4: Table 4

[0049] The present invention also provides a system for implementing the tailings dam accident path identification method based on a hybrid causal knowledge graph provided in any of the above embodiments, comprising: A first data set acquisition unit: used to acquire a first data set, wherein the first data set includes a change value of a back gate threshold voltage, a change value of a back gate subthreshold swing, and a change value of a back gate source leakage current before and after the VOCs to be identified pass through a single electrostatic wire to form a nanowire semiconductor; A second data set acquisition unit: used to acquire a second data set, wherein the second data set includes a change value of a junction gate threshold voltage, a change value of a junction gate subthreshold swing, and a change value of a junction gate source leakage current before and after the VOCs to be identified pass through a single electrostatic nanowire semiconductor; Identification module: used to perform Stacking model fusion on the first data set and the second data set to obtain the final prediction result of the type of VOCs to be identified; when performing Stacking model fusion, the base model adopts a table deep learning model, and the meta-model adopts a table deep learning model.

[0050] The embodiments of the present invention also provide corresponding electronic devices and computer-readable storage media for implementing the solutions provided by the embodiments of the present invention.

[0051] Among them, the device includes a memory and a processor, the memory is used to store instructions or codes, and the processor is used to execute the instructions or codes, so that the device executes the tailings dam accident path identification method based on a hybrid causal knowledge graph described in any embodiment of the present application.

[0052] The storage medium stores a computer program, wherein the computer program, when executed by the processor, implements the tailings dam accident path identification method based on a hybrid causal knowledge graph as described in any embodiment of the present application.

[0053] Another object of an embodiment of the present invention is to provide a computer program product, which includes computer instructions, and the computer instructions instruct a computer to execute the tailings dam accident path identification method based on a hybrid causal knowledge graph.

[0054] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0055] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0056] The present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, readable storage media, optical storage, etc.) containing computer-usable program codes.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A tailings dam accident path identification method based on a hybrid causal knowledge graph, characterized in that: include: A mixed causal logic model is constructed using the acquired initial hazard sources of tailings dam accidents; Using the hybrid causal logic model, solving the set of causal links of tailings dam accidents; In the set of tailings dam accident causal links, the logical relationship of tailings dam accidents and accident causal links are used as nodes and edges of the knowledge graph respectively, and the knowledge graph of tailings dam accidents is constructed using Python+Neo4j method; The tailings dam accident knowledge graph is encoded and classified using the BERT-GCN-SVM model to achieve tailings dam accident path identification.

2. The tailings dam accident path identification method based on hybrid causal knowledge graph according to claim 1 is characterized in that: The process of obtaining the initial hazard source of the tailings dam accident includes: Analyze historical accident data, take the cause of the accident as part of the initial hazard source set, and distinguish the life cycle of the tailings pond to form an initial hazard source set including human factors, technology, environment and management. Specifically: The initial hazard sources of tailings dam accidents are analyzed from the tailings dam construction stage, tailings pond operation stage and tailings pond closure stage; During the construction phase of tailings dams, the environmental factors that affect the safety of tailings dams mainly include hydrological and meteorological conditions, earthquakes, the stability of the mountains around the dam, and the physical properties of the tailings; technical factors mainly exist in the initial dam construction process, in order to save costs, various technical indicators are not achieved, and inappropriate structural measures are taken; management factors mainly include laws and regulations, safety investment, safety evaluation and safety supervision; During the operation phase of the tailings dam, the environmental factors that affect the safety of the tailings dam mainly include hydrological and meteorological conditions and earthquakes; technical factors are mainly caused by equipment failure or inappropriate structure; management factors mainly include safety training, safety investment, safety evaluation and safety supervision; During the closure stage of the tailings pond, the environmental factors that affect the safety of the tailings pond mainly include extreme weather and instability of the mountains around the pond; the technical factors are mainly those caused by equipment failure or inappropriate structure; the management factors mainly include imperfect closure regulations, insufficient investment in closure safety, inadequate safety evaluation, insufficient safety supervision and inadequate emergency management.

3. The tailings dam accident path identification method based on hybrid causal knowledge graph according to claim 1 is characterized in that: The hybrid causal logic model includes an event sequence diagram layer, a fault tree layer and a Bayesian network layer, wherein the event sequence diagram layer is the top layer, the fault tree layer is the middle layer, and the Bayesian network layer is the third layer, wherein the event sequence diagram layer is used to describe risk logic and calculate accident probability, the fault tree layer is used to find the causes of key events, and the Bayesian network layer is used to evaluate the impact of variability and correlation events.

4. The tailings dam accident path identification method based on hybrid causal knowledge graph according to claim 1 is characterized in that: When solving the set of tailings dam accident causal links using the hybrid causal logic model, the risk source that may occur in the tailings dam is taken as the starting event, and different results are evolved through the hybrid causal logic model to obtain different accident causal links of the tailings dam. The different accident causal links of the tailings dam constitute the set of tailings dam accident causal links.

5. The tailings dam accident path identification method based on hybrid causal knowledge graph according to claim 1 is characterized in that: In the set of tailings dam accident cause links, the logical relationship of tailings dam accidents and the accident cause links are used as nodes and edges of the knowledge graph respectively. The process of constructing the tailings dam accident knowledge graph using Python+Neo4j method includes: The Python+Neo4j method is used to transform the node-to-node logical relationship obtained from the hybrid causal logic model into a tailings dam accident knowledge graph based on hybrid causal logic. The entity construction of the tailings dam accident revolves around the mechanism of the accident, and the entities are set according to the human factors, technology, environment and management categories, and the relationships are set according to the cause and effect, time, inclusion and condition categories.

6. The tailings dam accident path identification method based on hybrid causal knowledge graph according to claim 1 is characterized in that: In the BERT-GCN-SVM model: Using the pre-trained language model BERT to encode text can learn the semantic information of words; The graph convolutional neural network (GCN) is used to encode and model the subgraph of each path in the tailings dam accident knowledge graph. The SVM model is used as a classifier to classify the combination of node semantic information and path subgraph in the tailings dam accident knowledge graph, and output labels and probabilities for path prediction and identification of new entities.

7. Tailings dam accident path identification system based on hybrid causal knowledge graph, characterized by: include: Hybrid causal logic model modeling unit: used to construct a hybrid causal logic model using the acquired initial hazard source of tailings dam accidents; Calculation module: used to solve the set of causal links of tailings dam accidents by using the hybrid causal logic model; Knowledge graph construction module: used to take the logical relationship of tailings dam accidents and accident causation links in the set of tailings dam accident causation links as nodes and edges of the knowledge graph respectively, and use Python+Neo4j method to construct the tailings dam accident knowledge graph; Classification unit: used to encode and classify the tailings dam accident knowledge graph using the BERT-GCN-SVM model to achieve tailings dam accident path identification.

8. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the tailings dam accident path identification method based on a hybrid causal knowledge graph as described in any one of claims 1 to 6.

9. A storage medium, characterized in that: A computer program is stored thereon, wherein when the computer program is executed by a processor, the method for identifying tailings dam accident paths based on a hybrid causal knowledge graph as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising computer instructions, characterized in that: The computer instructions instruct the computer to execute the tailings dam accident path identification method based on a hybrid causal knowledge graph as described in any one of claims 1-6.

Images