Coal mine gas explosion accident reason analysis method and system

Abstract

The embodiment of the invention provides a coal mine gas explosion accident reason analysis method and system, and belongs to the technical field of gas explosion accident prevention. The method comprises the following steps: acquiring report information of historical gas explosion accidents, and performing structured processing on the report information based on a predetermined reason classification framework to obtain structured information; constructing a corresponding accident three-dimensional network model based on the structured information, and determining attribute information of each node in the accident three-dimensional network model; analyzing the occurrence reason of the corresponding gas explosion accident based on the attribute information of each node to obtain the analysis result of each analysis index; and constructing a corresponding explosion accident self-inspection prevention list based on the analysis result of each analysis index, and pushing the explosion accident self-inspection prevention list to a user side. According to the scheme, real-time updating and accurate implementation of accident prevention measures are achieved, and efficient and feasible technical support is provided for coal mine safety management.

Description

Technical Field

[0001] The present invention relates to the technical field of preventing gas explosion accidents, and particularly relates to a method for analyzing the causes of coal mine gas explosion accidents and a system for analyzing the causes of coal mine gas explosion accidents. Background Art

[0002] With the expansion of coal mine production scale and the complication of production conditions, gas explosion accidents have become one of the most serious accident types in coal mine safety management. The occurrence of such accidents is usually caused by the interaction of multiple factors, involving aspects such as gas accumulation, poor ventilation, illegal operation, equipment failure, etc. In order to effectively prevent gas explosion accidents, accurately identifying and analyzing their causes has become the core task of coal mine safety management. However, the existing gas explosion accident analysis methods have many deficiencies in dealing with complex accident cause networks.

[0003] Traditional analysis methods mostly rely on statistical means to identify risk factors by counting the accident occurrence frequency and simply classifying accident causes. Although this method can reveal the accident occurrence rules in some cases, due to the failure to consider the complex mutual influence and non - linear relationship between accident causes, its analysis results are mostly qualitative and lack in - depth excavation of accident causes. Therefore, traditional methods cannot provide precise quantitative support for actual safety management and decision - making, and have great limitations.

[0004] In addition, although accident causation models can describe the accident occurrence process through causal chains, with the increase in the number of accident samples, traditional graph model analysis methods become long and complex, making it difficult to effectively handle large amounts of data. The calculation and display of graph models are also easily affected by manual analysis, resulting in poor reliability and accuracy of analysis results. When dealing with diversified accident causes, graph models often cannot comprehensively display the multi - level and complex relationships between accident causes, resulting in insufficient depth and breadth of analysis results and being difficult to meet actual needs.

[0005] At the same time, although Bayesian networks and fault tree analysis can, to a certain extent, reveal the potential associations of accident causes, they rely on a large amount of training data and have high computational complexity. When facing large - scale and complex accident data, there are still limitations in the number of nodes and edges, making it difficult to comprehensively cover all possible accident influencing factors. Especially in coal mine accident analysis, the limitations of Bayesian networks and fault trees make it impossible to fully consider the deep - level connections between accident causes, and the analysis results may not be precise enough to provide practical and effective support for decision - makers.

[0006] Therefore, although the existing methods for analyzing coal mine gas explosion accidents provide a framework for identifying accident causes to a certain extent, when faced with large-scale and complex multi-dimensional accident data, there are still problems such as insufficient analysis depth, limited processing capacity, and insufficient quantitative data analysis. To better address these challenges, a new method that can perform both qualitative and quantitative analysis is needed to help more comprehensively reveal the relationships between accident causes, thereby providing more scientific and accurate decision-making support for coal mine safety management. Summary of the Invention

[0007] The purpose of the embodiments of the present invention is to provide a method and system for analyzing the causes of coal mine gas explosion accidents, so as to at least solve the problems of insufficient analysis depth, limited processing capacity, and insufficient quantitative data analysis that still exist in the existing solutions when faced with large-scale and complex multi-dimensional accident data.

[0008] To achieve the above purpose, the first aspect of the present invention provides a method for analyzing the causes of coal mine gas explosion accidents. The method includes: collecting the report information of historical gas explosion accidents, and performing structured processing on the report information based on a pre-determined cause classification framework to obtain structured information; constructing a corresponding accident three-dimensional network model based on the structured information, and determining the attribute information of each node in the accident three-dimensional network model; performing analysis on the causes of the corresponding gas explosion accidents based on the attribute information of each node to obtain the analysis results of each analysis index; constructing a corresponding self-inspection and prevention list for explosion accidents based on the analysis results of each analysis index, and pushing the self-inspection and prevention list for explosion accidents to the user terminal.

[0009] Optionally, the pre-determined cause classification framework includes multiple levels of index types, and each level of index type includes multiple safety indicators; the performing structured processing on the report information based on the pre-determined cause classification framework to obtain structured information includes: performing preprocessing on the report information, and performing clustering processing on the preprocessed report information based on each safety indicator to obtain data sets under each safety indicator; performing structured processing on each data set to obtain the corresponding structured information.

[0010] Optionally, the constructing a corresponding accident three-dimensional network model based on the structured information includes: based on the structured information, determining the cause-effect relationship and time sequence of each accident event, so as to construct a corresponding accident causal link based on the cause-effect relationship and time sequence; combining the accident causal links of each accident time to construct a corresponding accident three-dimensional network model; wherein, a node in the accident three-dimensional network model represents a structured accident event, and an edge represents the causal relationship between the two corresponding accident events connected.

[0011] Optionally, determining the attribute information of each node in the accident three-dimensional network model includes: determining the ordered chain of each accident based on the accident three-dimensional network model to obtain the corresponding accident cause network; constructing the corresponding adjacency matrix based on the accident cause network, and determining the corresponding attribute information of each node based on the adjacency matrix; where the attribute information includes: degree centrality, which is used to represent the centrality degree of each node; geodesic distance, which is used to represent the shortest path length along the network edge between two nodes; shortest path, which is used to represent the path with the fewest hops between two nodes; clustering coefficient, which is used to represent the link degree between each node and its neighbor nodes; betweenness centrality, which is used to represent the importance degree of each node in the accident three-dimensional network model.

[0012] Optionally, the determination rule of the degree centrality is:

[0013]

[0014] Wherein, and are respectively the in-degree and out-degree of the i-th node; is the degree centrality of the i-th node; the determination rule of the geodesic distance is:

[0015]

[0016] Wherein, gd(u,v) represents the geodesic distance between node u and node v; i and j represent the intermediate nodes from node u to node v; m ui , m ij , m jv are respectively the indicators of whether there is a connection relationship between node u and node i, between node i and node j, and between node j and node v. If there is a connection relationship, it is 1, otherwise it is 0.

[0017] Optionally, the determination rule of the clustering coefficient is:

[0018]

[0019] Wherein, C i is the clustering coefficient of node i; K v is the number of nodes directly connected to node i; CAM i,j is the indicator of whether there is a connection relationship between node i and node j. If there is a connection relationship, it is 1, otherwise it is 0; the determination rule of the betweenness centrality is:

[0020]

[0021] Wherein, N jl (i) is the number of data with the shortest path passing through node i; N jlis the shortest path between node i and node l.

[0022] Optionally, each analysis index includes any one or more of the accident cause network node degree, accident cause network cumulative degree, accident cause network cohesion measure, and accident cause network centrality degree.

[0023] Optionally, constructing a corresponding self-inspection and prevention list for explosion accidents based on the analysis results of each analysis index and pushing the self-inspection and prevention list for explosion accidents to the user side includes: performing importance ranking of accident causes based on the analysis results of each analysis index; determining the attention targets and attention priorities for self-inspection and prevention of explosion accidents based on the importance ranking results of accident causes; generating a corresponding self-inspection and prevention list for explosion accidents based on the attention targets and attention priorities for self-inspection and prevention of explosion accidents, and pushing the self-inspection and prevention list for explosion accidents to the user side.

[0024] In a second aspect of the present invention, a system for analyzing the causes of coal mine gas explosion accidents is provided. The system includes: a collection unit, configured to collect report information of historical gas explosion accidents and perform structured processing on the report information based on a pre-determined cause classification framework to obtain structured information; a model construction unit, configured to construct a corresponding three-dimensional accident network model based on the structured information and determine the attribute information of each node in the three-dimensional accident network model; an analysis unit, configured to analyze the occurrence causes of corresponding gas explosion accidents based on the attribute information of each node to obtain the analysis results of each analysis index;

[0025] an output unit, configured to construct a corresponding self-inspection and prevention form for explosion accidents based on the analysis results of each analysis index and push the self-inspection and prevention list for explosion accidents to the user side.

[0026] On the other hand, the present invention provides a computer-readable storage medium, on which instructions are stored, and when the instructions run on a computer, the computer is caused to execute the above-mentioned method for analyzing the causes of coal mine gas explosion accidents.

[0027] Through the above technical solution, the solution of the present invention collects historical gas explosion accident report information and performs structured processing based on a predetermined cause classification framework, successfully transforming a large amount of unordered accident data into a standardized information format. This process not only improves the operability and analysis efficiency of accident data, but also provides an accurate data basis for subsequent accident cause analysis. By constructing a three-dimensional network model and determining the attribute information of each node, the complex causal relationships between accident causes can be comprehensively reflected, and the root causes of accidents can be deeply explored through multi-dimensional analysis indicators. This method combines qualitative and quantitative analysis, breaking through the limitation of traditional methods that can only perform simple statistical analysis, thereby enhancing the depth and accuracy of accident cause analysis. The self-check and prevention list constructed based on the analysis results can be sorted according to the importance of accident causes, helping coal mining enterprises accurately formulate safety management measures and avoid the recurrence of accidents. Finally, by pushing the self-check and prevention list to the user side, the solution realizes the real-time update and accurate implementation of accident prevention measures, providing efficient and feasible technical support for coal mine safety management.

[0028] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent specific embodiments section. Brief Description of the Drawings

[0029] The drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. They are used together with the following specific embodiments to explain the embodiments of the present invention, but do not constitute a limitation to the embodiments of the present invention. In the drawings:

[0030] Figure 1 is a flowchart of the steps of a method for analyzing the causes of coal mine gas explosion accidents provided by an embodiment of the present invention;

[0031] Figure 2 is a schematic diagram of a classification framework for influencing factors of coal mine accidents provided by an embodiment of the present invention;

[0032] Figure 3 is a schematic diagram of the adjacency matrix and its causal network model corresponding to the example in Table 1 provided by an embodiment of the present invention;

[0033] Figure 4 is a schematic diagram of topological indices of a gas explosion accident cause network provided by an embodiment of the present invention;

[0034] Figure 5 is a system structure diagram of a system for analyzing the causes of coal mine gas explosion accidents provided by an embodiment of the present invention. Detailed Description of the Embodiments

[0035] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for the purpose of illustration and explanation of the present invention, and are not intended to limit the present invention.

[0036] Figure 1 It is a method flow chart of a method for analyzing the causes of coal mine gas explosion accidents provided by an embodiment of the present invention. As Figure 1 shown, an embodiment of the present invention provides a method for analyzing the causes of coal mine gas explosion accidents, and the method includes:

[0037] Step S10: Collect the report information of historical gas explosion accidents, and perform structured processing on the report information based on a pre-determined cause classification framework to obtain structured information.

[0038] Specifically, the pre-determined cause classification framework includes multiple layers of index types, and each layer of index type includes multiple safety indicators; performing structured processing on the report information based on the pre-determined cause classification framework to obtain structured information includes: performing pre-processing on the report information, and performing clustering processing on the pre-processed report information based on each safety indicator to obtain data sets under each safety indicator; performing structured processing on each data set to obtain corresponding structured information.

[0039] In the embodiments of the present invention, the adopted cause classification framework consists of multiple layers of index types, and each layer of index includes multiple safety indicators. This framework can systematically classify the causes of gas explosion accidents, and refine each type of cause into multiple specific safety indicators, thus ensuring comprehensive coverage and refined processing of accident data. As Figure 2 , in a possible implementation manner, a classification framework for influencing factors of coal mine accidents based on 24Model is obtained, including 4 layers and 42 indicators, as a standardized classification framework for subsequent search for the causes of coal mine gas explosion accidents.

[0040] In actual operation, first, the report information is preprocessed to remove noise data and irrelevant information, ensuring the quality of the data relied on for analysis. After that, based on each safety indicator, the preprocessed accident report information is clustered. In the clustering process, similar accident cases are grouped into one category according to the performance of each report information under different safety indicators, thus forming data sets for various safety indicators. These data sets cover different aspects of the accident, helping to deeply explore the impact of each safety indicator on the occurrence of gas explosion accidents. By performing further structured processing on these data sets, the corresponding structured information is finally obtained. This structured information not only standardizes the presentation form of accident report data but also enables subsequent analysis work to be carried out on a standardized basis, thus ensuring the efficiency and accuracy of the analysis. This method solves the problems of chaotic data and unsystematic analysis in the process of accident report information processing and provides a reliable data basis for subsequent accident cause modeling and analysis.

[0041] Based on the solution of the present invention, the availability of accident data is greatly improved through structured processing, enabling the data to more clearly and accurately reflect the causes of accidents. The classification framework of multi-level indicator types ensures the comprehensiveness and in-depthness of data analysis. Through preprocessing and clustering, the sorting efficiency of accident information is further improved, laying a solid foundation for subsequent quantitative analysis and risk assessment. This structured processing not only optimizes the data processing process but also provides more scientific and systematic data support for safety management personnel, enhancing the accuracy and effectiveness of safety management.

[0042] Step S20: Construct a corresponding three-dimensional accident network model based on the structured information and determine the attribute information of each node in the three-dimensional accident network model.

[0043] Specifically, the constructing of the corresponding three-dimensional accident network model based on the structured information includes: based on the structured information, determining the cause-and-effect relationship and time sequence of each accident event, so as to construct a corresponding accident causal link based on the cause-and-effect relationship and time sequence; combining the accident causal links of each accident time to construct a corresponding three-dimensional accident network model; wherein, a node in the three-dimensional accident network model represents a structured accident event, and an edge represents the causal relationship between the two corresponding connected accident events.

[0044] In the embodiments of the present invention, first, according to the structured information obtained after preprocessing, the causal relationships between various accident events are determined. These causal relationships include not only the direct connections between events but also the sequential dependencies in the time series. For example, a certain safety event may be a prerequisite for another accident or a direct consequence caused by it. Through the confirmation of these causal relationships, the relationships between each event and other events are clearly characterized, helping to reveal the potential mechanisms of accident occurrence. Then, combined with the time sequence of event occurrence, these causal relationships are further integrated into accident causal chains. Each chain represents the occurrence process of an accident event and its causal connections with related events before and after. Through the analysis of multiple accident cases, the accident causal chains occurring in different time periods can be combined to construct an overall three-dimensional accident network model. In this model, each node represents a specific accident event, and the edges between nodes represent the causal relationships between different accident events. These edges not only reflect the direct relationships between events but also can reflect the complex interaction between events.

[0045] Further, determining the attribute information of each node in the three-dimensional accident network model includes: determining the ordered chains of each accident based on the three-dimensional accident network model to obtain the corresponding accident cause network; constructing the corresponding adjacency matrix based on the accident cause network, and determining the corresponding attribute information of each node based on the adjacency matrix; where each piece of attribute information includes: degree centrality, which is used to represent the degree of centrality of each node; geodesic distance, which is used to represent the shortest path length along the network edge between two nodes; shortest path, which is used to represent the path with the least number of hops between two nodes; clustering coefficient, which is used to represent the degree of connection between each node and its neighbor nodes; betweenness centrality, which is used to represent the importance degree of each node in the three-dimensional accident network model.

[0046] In the embodiments of the present invention, by analyzing the causal relationships between each node in the three-dimensional accident network model, the ordered chains of each accident event are determined. These ordered chains not only show the occurrence order of accident events but also reveal the internal connections between them. Based on these ordered chains, an accident cause network is constructed, and an adjacency matrix is further formed through this network. The adjacency matrix is a commonly used data structure in network analysis. By clarifying whether there is a connection relationship (i.e., causal relationship) between each pair of nodes, it provides a basis for subsequent calculation of attribute information.

[0047] With the support of the adjacency matrix, various attribute information of each node was further determined. The calculation of node attribute information involves multiple network analysis metrics, including: degree centrality, geodesic distance, shortest path, clustering coefficient, and betweenness centrality, etc. These metrics evaluate the position and importance of each node in the network through different mathematical methods. Degree centrality is used to measure the degree of connection of a node, reflecting the centrality of the node in the entire network; geodesic distance represents the shortest path length between nodes along the network edges and is an important metric for measuring the closeness of the connection between nodes; the shortest path provides the most direct path between nodes, facilitating the judgment of the directness between nodes; the clustering coefficient reflects the closeness of the relationship between a node and its neighbors and is usually used to measure whether a certain node tends to form a cluster; betweenness centrality is used to measure the criticality of a node in the entire network, especially when information or resources flow between nodes, and it is the frequency of their flow through the node.

[0048] In a possible implementation, the construction of the coal mine gas explosion accident cause network is a fundamental link in exploring the accident occurrence mechanism and process. This paper gives its formal definition as: character

[0049] MAUAN=(V h ,V c ,E)

[0050] Where: V h is the accident node, representing the danger at the system layer of the gas explosion accident; V c ={V i ,i = 1,2,...,N a non-empty finite set of causal nodes, representing the specific causal factors of Vh, where N is the number of accident cause nodes V h identified, and V i is the i-th cause of V h ; E={<V i ,V j >,<V k ,V h} is a non-empty finite set of causal relationships, representing the causal edges between causal factors and between causal factors and accidents, where V i ,V j ,V k ,V h ,i,j = 1,2,...,N. <V i ,V j > represents the causal relationship between accident cause nodes V i and V j , and <V k ,V h > represents the causal relationship between accident cause node V k and accident node V hThe causality. For an ordered pair of chains <V a , V b >, <V c , V d >, the following sequence relationship should be satisfied:

[0051] 1) In the accident, the ordered chain <V a , V b > only represents the sequential structure of event V a causing event V b . It does not allow <V b , V a > to appear simultaneously. Here, V b can represent the cause or result of the accident.

[0052] 2) Due to the chain effect, in the relationship of <V a , V b >, <V b , V c >, V a is only an indirect cause rather than a direct cause of V b .

[0053] After determining the rules for the causal ordered set of accident causes, a network model of accident causes can be established by constructing an adjacency matrix, which can be represented by an N×N adjacency matrix MAM, as shown below.

[0054]

[0055] Among them, i represents the cause of each accident, and j represents other causes or accident results that lead to this cause. When MAM ij = 1, it means that there is a directed edge from cause node i to j in the accident. The causal network matrix of coal mine gas explosion accidents can be constructed through the relevant relationships between different nodes. Taking two cases of gas explosion accidents as an example, the set of causal ordered chains identified in the accidents is shown in Table 1, where ACC is the result of the accident, namely the gas explosion accident.

[0056] Table 1 Ordered set of accident causes in two accident cases

[0057]

[0058] Taking the set of accident causal sequences shown in Table 1 as an example, the adjacency matrix and the corresponding causal network model composed of the causes and results of the gas explosion accident are as Figure 3 shown.

[0059] Specifically, the determination rules for the attribute information corresponding to each node are as follows:

[0060] 1) Degree centrality: Degree centrality is used to measure the relationship between nodes and edges in a network model. The greater the degree of a node, the higher the degree centrality of that node, indicating the importance of a certain accident cause in the causal network model. The number of edges k i connected to node V i is the degree of node V i . In a directed network, the degree centrality of a node is divided into two parts: in-degree and out-degree. The in-degree and out-degree are the number of edges flowing into and out of the node respectively, and their sum is also the total degree distribution. Denote the in-degree, out-degree, and total degree by and respectively, as shown in (1.2)-(1.4). Among them, h, i, j are all network nodes. If there is a directed event chain between two nodes, then m ij is 1, otherwise it is 0.

[0061]

[0062] 2) Geodesic distance and shortest path of the network: If there is at least one path between a pair of nodes (u, v) in the causal network model, it is called the geodesic distance gd(u, v). The shortest path (sg) refers to the path with the fewest number of edges between two vertices in a graph. This kind of path shows the most direct connection way from the starting point to the end point. The average path length L s is an index to measure the overall connectivity of the network graph. It calculates the average value of the shortest paths between all possible nodes in the graph, which is crucial for analyzing the efficiency and density of the network.

[0063] The defining functions of the above three network distance statistics are shown in (1.5)-(1.7), where u, v are a pair of nodes and u≠v, i, j are a pair of nodes connecting u, v.

[0064] gd(u, v) = ∑ i,j∈E (m ui + m ij + m jv )(1.5)

[0065]

[0066] 3) Clustering coefficient: When observing the network structure, the nodes in the network show a significant clustering tendency. This tendency is not just random, but indicates an inherent attraction between nodes, making certain specific types of nodes more likely to form tight clusters with each other. The clustering coefficient is used to reflect the degree of connection between a node and its neighbor nodes in the causal network model. Suppose node Vi in the network model is directly connected to kv nodes. Then the maximum number of possible edges between the kv nodes is kv(kv - 1) / 2. The clustering coefficient Ci of node Vi is defined as shown in 1.8

[0067]

[0068] 4) Betweenness centrality: Judging importance solely based on the degree value of a node has certain limitations. Some nodes may have a small degree, but they may be the connection points between two node communities and play an important role in the network. For such nodes, we use the betweenness centrality of the node to judge its importance in the network model. The betweenness centrality of node Vi is based on the number N of the shortest paths between all pairs of nodes passing through node Vi jl (i) and the number N jl of the shortest paths between two nodes. Then the betweenness B i of the node can be expressed as:

[0069] B i = ∑ i≠j≠l [N jl (i) / N jl (1.9).

[0070] Step S30: Analyze the causes of corresponding gas explosion accidents based on the attribute information of each node to obtain the analysis results of each analysis index.

[0071] Specifically, according to the obtained specific analysis results, Figure 4 the basic characteristic indexes of the gas explosion accident cause network analysis are shown. Each analysis index includes any one or more of the following: the degree of the accident cause network nodes, the cumulative degree of the accident cause network, the cohesion measure of the accident cause network, and the centrality degree of the accident cause network.

[0072] In a possible real-time manner, the analysis rules of each analysis index include:

[0073] 1) Conduct analysis of the degree of accident cause network nodes: Figure 4(a) shows the total degree value distribution of 145 nodes in the gas explosion accident cause network, with an average total degree of 9.56. This indicates that each accident cause in the accident is related to at least ten other safety events in the network. Limited by the display size, the accident nodes with higher total degree values are SH01 (habitually violating procedures and rules), PB03 (wishful thinking), SA01 (poor awareness of observing regulations), IA26 (operating in gentle breeze), and SA02 (poor ability to detect hazards). Among them, operating in gentle breeze (IA26) has the largest in-degree value of 29, which means that there are 29 possible paths leading to the occurrence of operating in gentle breeze in the gas explosion accident. Just like the preconditions required to trigger a gas explosion, the accumulation of gas concentration and the appearance of an ignition source are often two important conditions for inducing accidents under the condition of sufficient oxygen. Compared with other accident causes with smaller in-degree values, multiple connected paths will inevitably make it more difficult to prevent gas accumulation.

[0074] 2) Conduct an analysis of the cumulative degree distribution of the accident cause network: As Figure 4 (d) shows, the cumulative degree distribution of the gas explosion accident cause network model is approximately fitted to a power-law distribution function in the double-logarithmic coordinate system, which also shows that the accident network has the characteristics of a scale-free network. The fitting function is y = 1.2714x -0.7027 ,R 2 = 0.8179. The total degree values of 50% of the accident cause nodes are less than 7.5. Only a few nodes, as central nodes, have a large number of connecting edges, while most nodes have a small number of connecting edges. Facing various causal relationships or safety events, safety personnel need to focus on preventing the occurrence of these accident-causing factors with large degree values to reduce the accumulation of the domino effect between various causal relationships and accident events.

[0075] 3) Conduct an analysis of the cohesion measure of the accident cause network: The average clustering coefficient of the gas explosion accident cause network is 0.108, and the average path length is 3.675. The cause network shows a relatively short average path length, and the clustering coefficient is much larger than that of a random network of the same scale (0.056). Therefore, it is considered that the gas explosion accident cause network has small-world characteristics. The cases of gas inspectors working without a license (GA17), non-standard coal mining methods (GA25), incorrect opening of closed areas (GA29), and incorrect adjustment of the ventilation system (GA16) are as high as 0.45, 0.40, 0.38, and 0.38 respectively. The factors with larger clustering coefficients are also strongly associated with their neighbor factors. Once the central node shows abnormalities, it is easy to cause risks to the neighbor nodes. Controlling the nodes with larger clustering coefficients is an important way to reduce the network connectivity.

[0076] 4) Conduct network centrality measurement analysis of accident causes: The network centrality measurement analysis (including node betweenness centrality and edge betweenness centrality) is a common strategy to understand the influence and status of nodes in a complex network. The present invention proposes three network centrality indicators for quantitatively measuring the influence of nodes, including degree centrality (DC), betweenness centrality (BC), and eigenvector centrality (EC), so as to determine the core cause factors and key cause links of gas explosion accidents.

[0077] Step S40: Construct a corresponding self-check and prevention list for explosion accidents based on the analysis results of each analysis indicator, and push the self-check and prevention list for explosion accidents to the user side.

[0078] Specifically, based on the analysis results of each analysis indicator, perform importance ranking of accident causes; based on the importance ranking results of accident causes, determine the attention targets and attention priorities for self-check and prevention of explosion accidents; generate a corresponding self-check and prevention list for explosion accidents based on the attention targets and attention priorities for self-check and prevention of explosion accidents, and push the self-check and prevention list for explosion accidents to the user side.

[0079] In the embodiment of the present invention, through the evaluation of the analysis indicators of each accident cause, a comprehensive analysis result of the accident cause is formed. These indicators include degree centrality, clustering coefficient, shortest path, geodesic distance, betweenness centrality, etc., which can help accurately judge the importance and influence of each accident factor in the accident occurrence process. Based on these analysis results, importance ranking of accident causes is carried out, clarifying which factors play a decisive role in the accident and which factors, although important, have a smaller impact.

[0080] According to the importance ranking of accident causes, the attention targets and priorities for self-check and prevention of explosion accidents are further determined. In this process, the importance ranking provides a clear guiding direction for subsequent decision-making. The most critical factors in the accident will be given priority consideration and treatment, while the factors with smaller influence can be arranged for later management and improvement measures according to the actual situation. Through this priority division, it can ensure that coal mining enterprises, under the condition of limited resources, concentrate the limited time and energy on the most dangerous and urgent safety hazards, maximizing the safety prevention effect.

[0081] According to the determined attention targets and priorities, a corresponding self-check and prevention list for explosion accidents is generated. This list not only covers all relevant safety hazards and problems, but also details the prevention and control measures and treatment suggestions for each problem. The generation method of the self-check list follows a scientific logical order to ensure that each measure can be carried out in an orderly manner, thereby effectively preventing accidents. By pushing this self-check and prevention list to the user side, the safety management personnel of coal mining enterprises can receive the latest safety guidance in real time and conduct daily inspections and management according to the content of the list.

[0082] Based on the solution of the present invention, through the quantitative analysis and priority ranking of accident causes, the blindness and randomness that may exist in traditional safety inspections are avoided, making safety management measures more precise and targeted. The function of generating and pushing the accident self-inspection and prevention list enables safety management personnel to obtain the latest information and measures for accident prevention in a timely manner, thereby ensuring the continuous progress of safety management work and enabling adjustments at any time according to the changing safety conditions. In addition, the system not only improves the efficiency of accident self-inspection, but also strengthens the execution of accident prevention work, reduces the risk of accidents, and ultimately provides a more scientific and refined safety management solution for coal mining enterprises.

[0083] Figure 5 It is the system structure diagram of the coal mine gas explosion accident cause analysis system provided by an embodiment of the present invention. As Figure 5 shown, an embodiment of the present invention provides a coal mine gas explosion accident cause analysis system, the system includes: a collection unit, configured to collect the report information of historical gas explosion accidents, and perform structured processing on the report information based on a pre-determined cause classification framework to obtain structured information; a model construction unit, configured to construct a corresponding accident three-dimensional network model based on the structured information, and determine the attribute information of each node in the accident three-dimensional network model; an analysis unit, configured to analyze the occurrence causes of the corresponding gas explosion accidents based on the attribute information of each node to obtain the analysis results of each analysis index; an output unit, configured to construct a corresponding explosion accident self-inspection and prevention list based on the analysis results of each analysis index, and push the explosion accident self-inspection and prevention list to the user side.

[0084] An embodiment of the present invention also provides a computer-readable storage medium, on which instructions are stored, and when running on a computer, the computer is made to execute the above-mentioned coal mine gas explosion accident cause analysis method.

[0085] Those skilled in the art can understand that all or part of the steps in the method for implementing the above embodiments can be completed by instructing relevant hardware through a program, and the program is stored in a storage medium, including several instructions for enabling a single-chip microcomputer, a chip or a processor to execute all or part of the steps of the method described in various embodiments of the present invention. And the foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks or optical disks that can store program codes.

[0086] The optional embodiments of the present invention have been described in detail above in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention. Additionally, it should be noted that, in the above specific embodiments, the various specific technical features described can be combined in any appropriate manner without conflict. To avoid unnecessary repetition, the embodiments of the present invention will not separately describe various possible combination methods.

[0087] In addition, any combination can be made among the various different embodiments of the present invention, as long as it does not violate the idea of the embodiments of the present invention, and it should also be regarded as the content disclosed by the embodiments of the present invention.

Claims

1. A method for analyzing the causes of coal mine gas explosion accidents, characterized in that, The method includes: Collecting the report information of historical gas explosion accidents, and performing structured processing on the report information based on a pre-determined cause classification framework to obtain structured information; Constructing a corresponding three-dimensional accident network model based on the structured information, and determining the attribute information of each node in the three-dimensional accident network model; Performing analysis on the causes of corresponding gas explosion accidents based on the attribute information of each node to obtain the analysis results of each analysis index; Constructing a corresponding self-inspection and prevention list for explosion accidents based on the analysis results of each analysis index, and pushing the self-inspection and prevention list for explosion accidents to the user side.

2. The method according to claim 1, wherein The pre-determined cause classification framework includes multiple layers of index types, and each layer of index type includes multiple safety indicators; The performing structured processing on the report information based on the pre-determined cause classification framework to obtain structured information includes: Performing preprocessing on the report information, and performing clustering processing on the preprocessed report information based on each safety indicator to obtain data sets under each safety indicator; Performing structured processing on each data set to obtain corresponding structured information.

3. The method according to claim 1, characterized in that, The constructing a corresponding three-dimensional accident network model based on the structured information includes: Based on the structured information, determining the cause and effect relationship and time sequence of each accident event, and constructing a corresponding accident causal link based on the cause and effect relationship and time sequence; Combining the accident causal links of each accident time to construct a corresponding three-dimensional accident network model; wherein, A node in the three-dimensional accident network model represents a structured accident event, and an edge represents the causal relationship between the corresponding two connected accident events.

4. The method according to claim 1, wherein The determining the attribute information of each node in the three-dimensional accident network model includes: Determining the ordered chain of each accident based on the three-dimensional accident network model to obtain a corresponding accident cause network; Constructing a corresponding adjacency matrix based on the accident cause network, and determining the respective attribute information corresponding to each node based on the adjacency matrix; wherein, The respective attribute information includes: Degree centrality, which is used to represent the centrality degree of each node; Geodesic distance, which is used to represent the shortest path length along the network edge between two nodes; Shortest path, which is used to represent the path with the fewest hops between two nodes; Clustering coefficient, which is used to represent the link degree between each node and its neighbor nodes; Betweenness centrality, which is used to represent the importance degree of each node in the three-dimensional accident network model.

5. The method according to claim 4, characterized in that, The determination rule of the degree centrality is: wherein, and are the in-degree and out-degree of the i-th node, respectively; is the degree centrality of the i-th node; The determination rule of the geodesic distance is: wherein, gd(u,v) represents the geodesic distance between node u and node v; i and j represent the intermediate nodes between node u and node v; m ui , m ij , m jv are respectively indicators of whether there is a connection relationship between node u and node i, between node i and node j, and between node j and node v. A connection relationship exists as 1, and otherwise as 0.

6. The method according to claim 4, wherein The determination rule of the clustering coefficient is: Among them, C i is the clustering coefficient of node i; K v is the number of nodes directly connected to node i; CAM i,j It indicates whether there is a connection relationship between node i and node j. If there is a connection relationship, it is 1; otherwise, it is 0. The determination rule of the betweenness centrality is: Among them, N jl (i) is the number of data whose shortest path passes through node i; N jl is the shortest path between node i and node l.

7. The method according to claim 1, characterized in that Each analysis index includes: Any one or more of the accident cause network node degree, accident cause network cumulative degree, accident cause network cohesion measure, and accident cause network centrality degree.

8. The method according to claim 1, characterized in that, The constructing a corresponding self-inspection and prevention list for explosion accidents based on the analysis results of each analysis index, and pushing the self-inspection and prevention list for explosion accidents to the user side includes: Performing importance ranking of accident causes based on the analysis results of each analysis index; Based on the sorting result of the importance of accident causes, determine the attention targets and attention priorities for self-inspection and prevention of explosion accidents; Generate a corresponding self-inspection and prevention list for explosion accidents based on the attention targets and attention priorities for self-inspection and prevention of explosion accidents, and push the self-inspection and prevention list for explosion accidents to the user side.

9. A system for analyzing the causes of coal mine gas explosion accidents, characterized in that, The system includes: A collection unit, configured to collect the report information of historical gas explosion accidents, and perform structured processing on the report information based on a pre-determined cause classification framework to obtain structured information; A model construction unit, configured to construct a corresponding accident three-dimensional network model based on the structured information, and determine the attribute information of each node in the accident three-dimensional network model; An analysis unit, configured to analyze the causes of corresponding gas explosion accidents based on the attribute information of each node, and obtain the analysis results of each analysis index; An output unit, configured to construct a corresponding self-inspection and prevention list for explosion accidents based on the analysis results of each analysis index, and push the self-inspection and prevention list for explosion accidents to the user side.

10. A computer-readable storage medium, characterized in that, Instructions are stored on the computer-readable storage medium, and when running on a computer, the computer is caused to execute the method for analyzing the causes of coal mine gas explosion accidents according to any one of claims 1-8.

Images