Methods, devices, electronic equipment and storage media for analyzing the causes of power accidents

Abstract

This application discloses a method, apparatus, electronic device, and storage medium for analyzing the causes of power accidents. The method includes: acquiring an accident report of a power accident and extracting multiple causal events from the accident report; determining the relationship between the multiple causal events based on their order of occurrence and frequency, and constructing an accident causal network; analyzing the importance of network nodes through the topological properties of the accident causal network, and calculating the causal risk of the power accident using cascading failure theory, generating risk management recommendations. This effectively identifies the unique causes of each type of accident and facilitates the identification of the causes that play a significant role in directly triggering the accident. Therefore, it solves the problems of related technologies failing to identify the unique key causes of each type of accident, hindering the identification of key causes from the perspective of the direct causes of the accident, and failing to fully consider the dynamic process of the domino effect, resulting in insufficient simulation of the actual occurrence process of the accident.

Description

Technical Field

[0001] This application relates to the field of power risk management technology, and in particular to a method, apparatus, electronic device and storage medium for power accident causation analysis. Background Technology

[0002] The safe operation of power systems is crucial to social stability, economic development, and daily life. Due to the complex operating environment, power systems present numerous hazards during production and maintenance. Data shows that multiple power-related accidents occur annually, causing significant personal injury and economic losses, posing challenges to risk control for enterprises. Therefore, providing targeted recommendations for the prevention of power-related accidents is becoming increasingly important.

[0003] Analysis of accident investigation reports reveals that many accidents are caused by the interaction of multiple factors. According to Heinrich's domino theory, the successive occurrence of multiple causal events ultimately leads to an accident; this process can be described as an accident causation chain. Statistical analysis of accident occurrences shows that some causal events appear in many causation chains, while others occur only once or twice. This indicates that some factors play a more significant role in directly triggering accidents. Therefore, to provide more targeted recommendations for risk management, it is necessary to identify key causal events through accident causation analysis. If companies can take measures to proactively manage key causal events, they can potentially reduce risks in power operations.

[0004] Traditional accident causation analysis employs accident causation theory, a method to explore the patterns of accident occurrence and development, and to reveal the essence of accidents. It typically categorizes causal events into types such as human, material, environmental, and managerial factors, considering each type as a whole within a hierarchical framework. Typical methods include Heinrich's domino causation chain, Bird and Loftus causation chains, and the Reason cheese model. In recent years, more systematic accident analysis methods have been developed, such as Accip, HFACS, and STAMP. The shortcomings of these methods are: with the rapid development of power systems, the types of causal events are increasing, and their relationships are becoming more complex. Most methods only perform qualitative analysis on single or a few accidents, making it difficult to quantitatively rank key causes from a large number of past accident reports.

[0005] In recent years, incident causation analysis based on complex networks has received increasing attention. It often focuses on identifying key causes through network topology analysis, thereby providing targeted risk reduction recommendations. These methods analyze incident reports, using nodes to represent events and edges to represent relationships between events, to construct incident causation networks containing multiple causation chains. Then, topology analysis is used to identify key nodes in the network. Incident causation networks are generally divided into two categories: directed weighted networks and directed unweighted networks. Here, direction indicates the order of occurrence, and weight indicates the frequency of co-occurrence.

[0006] Existing causation analysis methods based on complex networks are mostly focused on railway systems, with limited application in power systems. Power-related personal injury accidents have their own unique characteristics. First, the causative events of power-related personal injury accidents vary significantly depending on the accident type. Many existing studies have constructed a unified network encompassing all causes, which leads to the interaction of causative events across different accident types, making it difficult to identify the key causes specific to each type of accident. Second, the causation chain of power-related personal injury accidents is a process in which multiple causes occur successively and ultimately lead to the accident. It includes two types of nodes: causative nodes and accident nodes. Many studies have constructed homogeneous networks containing the same type of nodes without distinguishing between paths between causative nodes and paths directly leading to the accident. This is not conducive to determining key causes from the perspective of the direct causes of the accident. Third, existing causation risk quantification methods do not fully consider the dynamic process of the domino effect and lack sufficient simulation of the actual accident occurrence process. Summary of the Invention

[0007] This application provides a method, apparatus, electronic device, and storage medium for analyzing the causes of power accidents, which solves the problems of related technologies being unable to discover the key causes unique to each type of accident, being unfavorable for determining key causes from the perspective of the direct causes of accidents, and failing to fully consider the dynamic process of the domino effect, resulting in insufficient simulation of the actual occurrence process of accidents.

[0008] The first aspect of this application provides a method for analyzing the causes of power accidents, comprising the following steps: obtaining an accident report of a power accident and extracting multiple causal events from the accident report; determining the relationship between the multiple causal events based on their order of occurrence and frequency of occurrence, and constructing an accident causal network; analyzing the importance of network nodes through the topological properties of the accident causal network, and calculating the causal risk of the power accident using cascading failure theory, and generating risk management recommendations.

[0009] Optionally, in one embodiment of this application, the step of determining the relationship between the causative events based on their sequential occurrence and frequency, and constructing an accident causation network, includes:

[0010] The direction of the edges of the accident causation network is determined according to the order in which the multiple causative events occur; the weight of the accident causation network is determined according to the number of times the multiple causative events occur; the accident causation network is constructed using the multiple causative events as nodes, based on the direction of the edges and the weights.

[0011] Optionally, in one embodiment of this application, the step of analyzing the importance of network nodes using the topological properties of the accident causation network includes: calculating the degree distribution, clustering coefficient, and shortest path length of the accident causation network; determining whether the accident causation network is a scale-free network; when the accident causation network is a scale-free network, calculating the betweenness factor of the accident causation network with a preset weight value; determining whether the number of parallel nodes with the same betweenness factor is less than a preset threshold; if it is less, using the ranking result corresponding to the preset weight value as the importance of the network nodes; otherwise, adjusting the size of the preset value until the number of parallel nodes with the same betweenness factor is less than the preset threshold.

[0012] Optionally, in one embodiment of this application, the calculation of the causal risk of a power accident using cascading failure theory includes: calculating the node state, initial dangerous load, safety threshold, and propagation intensity between nodes in the accident causal network at the initial moment; when a node in the accident causal network changes to a dangerous state, starting from the dangerous node, traversing all outgoing edges of the starting node, calculating the dangerous load value of each node pointed to by the dangerous node when an outgoing edge exists, otherwise, the dangerous propagation terminates; determining whether the dangerous load value of each node is greater than the safety threshold, if it is, the node greater than the safety threshold changes from a safe state to a dangerous state, otherwise, the dangerous propagation terminates; if the dangerous load value of multiple nodes pointed to by the dangerous node exceeds the safety threshold, the edge with the strongest propagation intensity is selected for dangerous propagation, the corresponding node state changes to a dangerous state, starting from the dangerous node, traversing again until the dangerous propagation terminates; calculating the probability of the occurrence of the causal chain leading to the accident, the risk value of the causal node in each type of accident, and the risk value of the causal node in all accidents to obtain the causal risk of the power accident.

[0013] A second aspect of this application provides a power accident causation analysis device, comprising: an extraction module for acquiring an accident report of a power accident and extracting multiple causative events from the accident report; a construction module for determining the relationship between the multiple causative events based on their order of occurrence and frequency of occurrence, and constructing an accident causation network; and a generation module for analyzing the importance of network nodes through the topological properties of the accident causation network, calculating the causative risk of the power accident using cascading failure theory, and generating risk management recommendations.

[0014] Optionally, in one embodiment of this application, the construction module includes: a first determining unit, configured to determine the direction of the accident causation network edge according to the chronological order of the occurrence of the plurality of causative events;

[0015] The second determining unit is used to determine the weights of the accident causation network based on the number of times the plurality of causative events occur; the network construction unit is used to construct the accident causation network using the plurality of causative events as nodes, based on the direction of the edges and the weights.

[0016] Optionally, in one embodiment of this application, the generation module includes: a first judgment unit, configured to calculate the degree distribution, clustering coefficient, and shortest path length of the accident causation network, and determine whether the accident causation network is a scale-free network; and a second judgment unit, configured to, when the accident causation network is a scale-free network, calculate the betweenness of the accident causation network with a preset weight value, and determine whether the number of parallel nodes with the same betweenness is less than a preset threshold. If it is less, the ranking result corresponding to the preset weight value is used as the importance of the network nodes; otherwise, the size of the preset value is adjusted until the number of parallel nodes with the same betweenness is less than the preset threshold.

[0017] Optionally, in one embodiment of this application, the generation module includes: a parameter unit, used to calculate the node state, initial hazard load, safety threshold, and propagation intensity between nodes of each node in the accident causation network at the initial time; a conversion unit, used to, when the state of a node in the accident causation network changes to a hazard state, take the hazard node as the starting node, traverse all outgoing edges of the starting node, calculate the hazard load value of each node pointed to by the hazard node if an outgoing edge exists, otherwise, the hazard propagation terminates; and a third judgment unit, used to judge whether the current hazard load value of each node is greater than the safety threshold. The system employs a full threshold mechanism. When a threshold is exceeded, nodes exceeding the safety threshold are transitioned from a safe state to a dangerous state; otherwise, the danger propagation terminates. A selection unit is used to select the edge with the strongest propagation intensity if multiple nodes pointed to by a dangerous node have dangerous load values ​​exceeding the safety threshold. The corresponding node state transitions to a dangerous state, and the process continues until the danger propagation terminates, starting with the dangerous node. A calculation unit is used to calculate the probability of the causal chain leading to the accident, the risk value of the causal node in each type of accident, and the risk value of the causal node in all accidents, thus obtaining the causal risk of the power accident.

[0018] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to perform the power accident causation analysis method as described in the above embodiments.

[0019] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to perform the power accident causation analysis method as described in the above embodiments.

[0020] Therefore, the embodiments of this application have the following beneficial effects:

[0021] This method involves acquiring accident reports of power outages and extracting multiple causative events from these reports. The relationships between these events are determined based on their chronological order and frequency of occurrence, constructing an accident causative network. The importance of network nodes is analyzed using the topological properties of this network, and the cascading failure theory is applied to calculate the causative risk of the power outage, generating risk management recommendations. This approach effectively identifies the unique causes of each type of accident and facilitates the identification of the most significant causes directly triggering the accident. Therefore, it addresses the shortcomings of existing technologies, such as the inability to identify the unique key causes of each type of accident, the difficulty in determining key causes from the perspective of the direct causes of accidents, and the insufficient consideration of the dynamic process of domino effects and inadequate simulation of the actual accident occurrence process.

[0022] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0023] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0024] Figure 1 This is a flowchart of a power accident causation analysis method provided according to an embodiment of this application;

[0025] Figure 2 This is a schematic diagram illustrating the execution logic of a power accident cause analysis method according to an embodiment of this application;

[0026] Figure 3 This is a schematic diagram of an accident causation network construction process according to an embodiment of this application;

[0027] Figure 4 This is a schematic diagram of an accident causation network constructed for an electric shock accident according to an embodiment of this application;

[0028] Figure 5 This is a schematic diagram of an accident causation network constructed for a fall from height accident according to an embodiment of this application;

[0029] Figure 6 This is a schematic diagram of an accident causation network constructed for a mechanical injury accident according to an embodiment of this application;

[0030] Figure 7 This is a schematic diagram of an accident causation network constructed for a collapse accident according to an embodiment of this application;

[0031] Figure 8 This is a schematic diagram of an accident causation network constructed for an object impact accident according to an embodiment of this application;

[0032] Figure 9 This is a schematic diagram of an accident causation network constructed for electric shock, fall from height, mechanical injury, collapse, and object strike accidents according to an embodiment of this application;

[0033] Figure 10 This is a schematic diagram of a network topology characteristic analysis process according to an embodiment of this application;

[0034] Figure 11 This is a schematic diagram of the network topology sorting result of an electric shock accident when w is 1 in the improved betweenness factor provided according to an embodiment of this application;

[0035] Figure 12 This is a schematic diagram of the network topology sorting result of a fall-from-height accident when w is 1 in the improved betweenness factor provided according to an embodiment of this application;

[0036] Figure 13 This is a schematic diagram of the network topology sorting result of mechanical injury accidents when w is 1 in the improved betweenness factor provided according to an embodiment of this application;

[0037] Figure 14 This is a schematic diagram of the network topology sorting result of a collapse accident when w is 1 in the improved betweenness factor provided according to an embodiment of this application;

[0038] Figure 15 This is a schematic diagram of the network topology sorting result of an object strike incident when w is 1 in the improved betweenness factor provided according to an embodiment of this application;

[0039] Figure 16 This is a schematic diagram illustrating the change of the improved betweenness factor when w changes in an electric shock accident, according to an embodiment of this application.

[0040] Figure 17 This is a schematic diagram of the network topology sorting result of an electric shock accident when w is 0.98 in the improved betweenness factor provided according to an embodiment of this application;

[0041] Figure 18 This is a schematic diagram of the network topology sorting result for a fall-from-height accident when w is 0.98 in the improved betweenness factor provided according to an embodiment of this application;

[0042] Figure 19This is a schematic diagram of the network topology sorting result of mechanical injury accidents when w is 0.98 in the improved betweenness factor provided according to an embodiment of this application;

[0043] Figure 20 This is a schematic diagram of the network topology sorting result of a collapse accident when w is 0.98 in the improved betweenness factor provided according to an embodiment of this application;

[0044] Figure 21 This is a schematic diagram of the network topology sorting result of an object impact incident when w is 0.98 in the improved betweenness factor provided according to an embodiment of this application;

[0045] Figure 22 This is a schematic diagram of a causative risk quantification process according to an embodiment of this application;

[0046] Figure 23 This is an example diagram of a power accident cause analysis device according to an embodiment of this application;

[0047] Figure 24 A schematic diagram of the structure of the electronic device provided in the application embodiment.

[0048] Explanation of reference numerals in the attached diagram: Extraction module-100, Construction module-200, Generation module-300, Memory-2401, Processor-2402, Communication interface-2403. Detailed Implementation

[0049] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0050] The following describes, with reference to the accompanying drawings, a method, apparatus, electronic device, and storage medium for analyzing the causal factors of power accidents according to embodiments of this application. Addressing the problems mentioned in the background art, this application provides a method for analyzing the causal factors of power accidents. In this method, an accident report of a power accident is obtained, and multiple causal events are extracted from the accident report. The relationship between the multiple causal events is determined based on their order of occurrence and frequency of occurrence, constructing an accident causal network. The importance of network nodes is analyzed through the topological properties of the accident causal network, and the causal risk of the power accident is calculated using cascading failure theory, generating risk management recommendations. This effectively identifies the unique causes of each type of accident, facilitating the identification of causes that play a significant role in directly triggering the accident. Furthermore, considering both the severity of consequences and the frequency of cause exposure, it is convenient for enterprise application. This solves the problems of related technologies failing to identify the unique key causes of each type of accident, hindering the identification of key causes from the perspective of the direct causes of the accident, and failing to fully consider the dynamic process of the domino effect, resulting in insufficient simulation of the actual occurrence process of the accident. The symbols and meanings of the accident causal events involved in the embodiments of this application are explained in Table 1.

[0051] Table 1. Symbols and Meanings of Accident Causes

[0052]

[0053] Specifically, Figure 1 This is a flowchart of a power accident causation analysis method provided in an embodiment of this application.

[0054] like Figure 1 As shown, the method for analyzing the causes of power accidents includes the following steps:

[0055] In step S101, an accident report of a power accident is obtained, and multiple causal events in the accident report are extracted.

[0056] It is understandable that accidents such as fires, explosions, electromechanical accidents, and traffic accidents each have their own causes and conditions, but also share some commonalities. These can be summarized as follows: unsafe acts by people (such as low work proficiency, excessive fatigue from overtime work, etc.), unsafe acts by equipment (such as defects in mechanical design, aging equipment, etc.), and poor working environments (such as poor ventilation and the presence of toxic or harmful gases). Most of these causes stem from management deficiencies. Therefore, in the embodiments of this application, the causal events are first extracted from accident reports according to the four elements of people, equipment, environment, and management, constructing an accident causal network to effectively identify the unique causes of each type of accident, thereby facilitating subsequent operations, such as... Figure 2 As shown, the network construction process will be described in detail below.

[0057] In step S102, the relationship between multiple causative events is determined based on their order of occurrence and frequency, and an accident causation network is constructed.

[0058] After extracting the causative events from the aforementioned accident reports, a causative factor network is constructed based on the accident type to identify the unique causes for each type of accident. The process for constructing the causative factor network is as follows: Figure 3 As shown.

[0059] Specifically, in one embodiment of this application, the relationship between multiple causative events is determined based on their order of occurrence and frequency of occurrence, and an accident causation network is constructed. This includes: determining the direction of the edges of the accident causation network based on the order of occurrence of the multiple causative events; determining the weights of the accident causation network based on the frequency of occurrence of the multiple causative events; and constructing the accident causation network using the multiple causative events as nodes, based on the direction and weight of the edges.

[0060] It should be noted that by using multiple causal events as nodes, an adjacency matrix can be obtained based on the direction and weight of the edges, as shown in the following formula, to further construct the accident causation network.

[0061]

[0062] Among them, a ij =1 indicates that event i has a direct impact on event j, w ij Indicates the number of times it affects the system.

[0063] It should be noted that, in the embodiments of this application, five accident causative factor networks were constructed according to the accident type. Specifically, electrical personal injury accidents were divided into five categories: electric shock, falls from heights, mechanical injuries, collapses, and being struck by objects. Accident causative networks were constructed for each of the above accident types. Based on experimental results, such as... Figures 4-9 As shown, this method can effectively identify the unique causes of each type of accident.

[0064] In step S103, the importance of network nodes is analyzed by the topological properties of the accident causation network, and the causation risk of power accidents is calculated using the cascading failure theory to generate risk management recommendations.

[0065] After the accident causation network is constructed as described above, the specific causes of each type of accident are obtained. However, in order to distinguish between the paths that cause accidents and the paths between causes, and to identify the causes that play a significant role in directly causing accidents, embodiments of this application further utilize the topological properties of the accident causation network to analyze the importance of network nodes.

[0066] Optionally, in one embodiment of this application, the importance of network nodes is analyzed using the topological properties of the incident-causing network, including: calculating the degree distribution, clustering coefficient, and shortest path length of the incident-causing network to determine whether the incident-causing network is a scale-free network; when the incident-causing network is a scale-free network, calculating the betweenness factor of the incident-causing network with preset weights, determining whether the number of parallel nodes with the same betweenness factor is less than a preset threshold; if it is less, using the ranking result corresponding to the preset weight as the importance of the network nodes; otherwise, adjusting the preset value until the number of parallel nodes with the same betweenness factor is less than the preset threshold. The above network topology characteristic analysis process is as follows: Figure 10 As shown.

[0067] Specifically, the network topology characteristic analysis steps in this application embodiment are as follows:

[0068] (1) Calculate the degree distribution, clustering coefficient and shortest path length to determine whether the network is a scale-free network or a small-world network;

[0069] (2) If the scale-free property is satisfied, it means that some nodes need to be protected. If the small-world property is satisfied, it means that the network is prone to chain effects. In this case, it is necessary to continue topology analysis and risk quantification.

[0070] (3) Calculate the improved betweenness factor when w=1 to better identify the nodes that play a key role in the risk propagation that directly causes the accident:

[0071]

[0072] Where w is the weight, CE is the set of causal nodes, and AC is the set of incident nodes. It is the number of shortest paths between nodes s and t that pass through node i, g st It represents the total number of shortest paths between nodes s and t. The weight w reflects the degree of bias towards the shortest path that directly caused the accident.

[0073] (4) Determine if there are many parallel nodes at this time. If not, use the sorting when w=1 as the result.

[0074] (5) If there are many nodes with the same number of nodes, the value of w needs to be reduced appropriately. Specifically, observe how the improved betweenness value of the 10 nodes with larger betweenness values ​​changes with w, so that it is as close to 1 as possible and small changes around it will not cause a big change in the ranking.

[0075] It should be noted that, in the process of selecting the value of w in the improved betweenness factor in the embodiments of this application, when w is 1, according to experimental results, such as... Figures 11-15 As shown, it can be observed that there are many instances of nodes side by side in the experimental results.

[0076] also, Figure 16 The changes in the improved betweenness factor as w changes in electric shock accidents are shown. It can be seen that the ranking changes relatively slowly with the change of w, and small changes in w will not cause a big change in the ranking.

[0077] It should be noted that the w value selected in this embodiment is 0.98. Based on experimental results, as... Figures 17-21 As shown, the sorting results have improved discrimination compared to when w=1.

[0078] Table 2 shows the ranking results of the improved betweenness coefficient under several weights. The ranking results are close when w=1 and w=0.98, and close when w=0.5 and w=0.0. Using the ranking result when w=0.98 to approximate w=1 satisfies the research objective and can also distinguish the rankings.

[0079] Table 2 Ranking results of improved betweenness under different weights

[0080] betweenness Improved betweenness w = 0.98 Improved betweenness w = 1 Improved betweenness w = 0 A1 H2, H7, H8, H6, H5 H7, H2, H14, H13, H8 H7, H2, H14, H13, H8 H2, H7, H8, H6, H1 A2 O1,H2,H21,H5,H1 H2,O1,M6,H1,H16 H2,O1,M6,H10,H16 O1,H2,H21,H5,H1 A3 H10,O13,H3,H33,O12 O13,H10,H3,H29,O12 O13,H10,H29,H3,O14 H10,O13,H3,H33,O12 A4 O10,M6,H35,H2,H5 H2,O10,H10,M6,H29 H2, H10, O10, H29, M6 O10,M6,H35,H2,H5 A5 H1,M6,H5,O14,H24 O11,O14,H10,O18,H24 O11,O14,H10,O18,H24 H1,M6,H5,H24,O1 A H10,H2,H5,O11,O1 H10,H2,O11,O1,H5 H10,O11,H2,O1,M6 H10,H2,H5,O11,O1

[0081] Understandably, the network topology analysis described above distinguishes between the paths that directly cause accidents and the paths between the causes, facilitating the identification of the causative factors that play a significant role in directly causing accidents. However, in enterprises, the calculation of causative risk often needs to consider factors such as exposure frequency and accident consequences. Therefore, this application's embodiment uses a causative risk quantification method based on cascading failure theory to fully consider the dynamic process of the domino effect and simulate the actual occurrence of accidents.

[0082] Optionally, in one embodiment of this application, the causative risk of a power accident is calculated using the cascading failure theory, i.e., the causative risk is quantified. The specific execution process is as follows: Figure 22 As shown, the process includes: calculating the node state, initial hazardous load, safety threshold, and propagation strength between nodes in the accident causation network at the initial moment; when a node in the accident causation network changes to a hazardous state, starting from the hazardous node, traversing all outgoing edges of the starting node, calculating the hazardous load value of each node pointed to by the hazardous node if an outgoing edge exists, otherwise, the hazardous propagation terminates; determining whether the hazardous load value of each node is greater than the safety threshold, if it is, the node with the higher safety threshold changes from a safe state to a hazardous state, otherwise, the hazardous propagation terminates; if multiple nodes pointed to by the hazardous node have hazardous load values ​​exceeding the safety threshold, the edge with the strongest propagation strength is selected for hazardous propagation, and the corresponding node state changes to a hazardous state, starting from the hazardous node again, traversing again until the hazardous propagation terminates; calculating the probability of the occurrence of the causation chain leading to the accident, the risk value of the causation node in each type of accident, and the risk value of the causation node in all accidents, to obtain the causation risk of the power accident.

[0083] Specifically, this application's embodiments assume that in the same type of accident, the causal chain satisfies the Markov property. Therefore, the causal risk quantification steps are as follows:

[0084] (1) When t = 0, calculate the node state and initial dangerous load L of each node in the network. i (0), Safety threshold C i and the propagation intensity I between nodes ij .

[0085] C i =(1+α)L i (0) (3)

[0086] The initial hazardous load of each node is defined as the betweenness factor, and α is the tolerance factor. The larger the tolerance factor, the more difficult it is for the node to transition from a safe state to a hazardous state.

[0087]

[0088] Among them, w ij Let s be the weights from node i to node j. j V represents the strength of node j. i Let i be the set of nodes in the next level that are in a safe state.

[0089] (2) When a node in the network changes to a dangerous state, t = t + 1. At this time, the node that caused the danger is taken as the starting node, and the node state s = 1.

[0090] (3) Traverse all outgoing edges of the initial dangerous node. If an outgoing edge exists, calculate the dangerous load value L of each node pointed to by the dangerous node according to the following formula. j (t); if there is no outgoing edge, the propagation of danger terminates.

[0091]

[0092] (4) Determine whether the dangerous load value of each node being pointed to exceeds the safety threshold. If it does, the node changes from a safe state to a dangerous state; otherwise, the dangerous propagation terminates.

[0093] (5) If the dangerous load values ​​of multiple nodes pointed to by the dangerous node exceed the safety threshold, the edge with the strongest propagation strength is selected for dangerous propagation, and the corresponding node state will also change to dangerous state.

[0094] (6) Return to step (2) until the spread of danger is terminated.

[0095] (7) Calculate the probability of the causal chain that leads to the accident.

[0096]

[0097] Where j and k represent the starting and ending nodes of an edge, respectively, and C ia This represents the set of nodes in the causal chain from causal node i to incident node a.

[0098] (8) Calculate the risk value of the causative node in each type of accident.

[0099] R i =S a ×E i ×P ia (7)

[0100] Among them, S a Indicates the severity of the accident, based on casualties and economic losses; E i This indicates the exposure frequency of the causative event, calculated by counting the average number of occurrences per year.

[0101] (9) Calculate the risk value of the causative node in all accidents.

[0102]

[0103] Where j is the subnetwork number.

[0104] Table 3 presents the experimental results of the causal risk quantification operation. The nodes with higher risks are H10, H2, and O11, and the results are basically consistent with the improved betweenness factor. It can serve as a supplement to provide more reference for enterprises.

[0105] Table 3. Experimental results data for causal risk quantification.

[0106]

[0107]

[0108] Based on the above analysis, especially the results of the improved betweenness factor when w=0.98, more targeted risk management recommendations can be given. Overall, the key causes to be controlled are H10 (entering a hazardous area), H2 (improper use of protective equipment), and M6 (lack of on-site supervision). Each type of accident has its own key control targets. In electric shock accidents, these are H7 (failure to verify, disconnect, and ground voltage) and H14 (bare contact with live parts); in falls from heights, these are O1 (insufficient protective measures) and H2 (improper use of protective equipment); in mechanical injury accidents, these are H29 (cleanup errors) and O13 (sudden equipment startup); in collapse accidents, these are O10 (design flaws) and H2 (improper use of protective equipment); and in falling object accidents, these are O11 (facility instability) and O14 (falling heavy objects).

[0109] Understandably, by starting with the domino effect of an accident, calculating the probability of the causal chain, and comprehensively considering the severity of the consequences and the frequency of causal exposure, the dynamic process of the domino effect can be fully taken into account, and the actual occurrence process of the accident can be simulated. This can provide more targeted suggestions for the risk management of power companies and facilitate their application.

[0110] According to the power accident causation analysis method proposed in this application, an accident report of a power accident is obtained, and multiple causative events are extracted from the accident report. The relationship between the multiple causative events is determined based on their order of occurrence and frequency, and an accident causation network is constructed. The importance of network nodes is analyzed through the topological properties of the accident causation network, and the causative risk of the power accident is calculated using the cascading failure theory to generate risk management suggestions. Thus, by extracting causes from accident reports, analyzing relationships, constructing an accident causation network, and quantitatively identifying key causative events of past accidents, more targeted risk control suggestions can be provided to enterprises.

[0111] Next, the power accident cause analysis apparatus proposed according to the embodiments of this application is described with reference to the accompanying drawings.

[0112] Figure 23 This is a block diagram of a power accident cause analysis device according to an embodiment of this application.

[0113] like Figure 23 As shown, the power accident cause analysis device 10 includes: an extraction module 100, a construction module 200, and a generation module 300.

[0114] The extraction module 100 is used to obtain accident reports of power accidents and extract multiple causative events from the accident reports; the construction module 200 is used to determine the relationship between multiple causative events based on their order of occurrence and frequency of occurrence, and construct an accident causative network; the generation module 300 is used to analyze the importance of network nodes by improving betweenness and topological properties of the accident causative network, and to calculate the causative risk of power accidents using cascading failure theory, and generate risk management recommendations.

[0115] Optionally, in one embodiment of this application, the construction module 200 includes: a first determining unit, configured to determine the direction of the accident causation network edges according to the order in which multiple causative events occur; a second determining unit, configured to determine the weight of the accident causation network according to the number of times the multiple causative events occur; and a network construction unit, configured to construct the accident causation network using the multiple causative events as nodes, according to the direction and weight of the edges.

[0116] Optionally, in one embodiment of this application, the generation module 300 includes: a first judgment unit, used to calculate the degree distribution, clustering coefficient, and shortest path length of the accident causation network, and to determine whether the accident causation network is a scale-free network; and a second judgment unit, used to calculate the betweenness of the accident causation network when the weight is a preset value, and to determine whether the number of parallel nodes with the same betweenness is less than a preset threshold. If it is less than a preset threshold, the ranking result corresponding to the preset weight is used as the importance of the network nodes; otherwise, the size of the preset value is adjusted until the number of parallel nodes with the same betweenness is less than the preset threshold.

[0117] Optionally, in one embodiment of this application, the generation module 300 includes: a parameter unit, used to calculate the node state, initial hazard load, safety threshold, and propagation intensity between nodes in the accident causation network at the initial time; a conversion unit, used to, when the state of a node in the accident causation network changes to a hazard state, take the hazard node as the starting node, traverse all outgoing edges of the starting node, calculate the hazard load value of each node pointed to by the hazard node if an outgoing edge exists, otherwise, hazard propagation terminates; and a third judgment unit, used to judge whether the current hazard load value of each node is greater than the safety threshold. The system has several functions: First, if the value of a node exceeds the safety threshold, the node in the safe state changes to a dangerous state; otherwise, the danger propagation terminates. Second, if the dangerous node points to multiple nodes whose dangerous load values ​​exceed the safety threshold, the system selects the edge with the strongest propagation strength to propagate the danger, and the corresponding node changes to a dangerous state. The system then iterates again from the dangerous node until the danger propagation terminates. Third, the system calculates the probability of the causal chain leading to the accident, the risk value of the causal node in each type of accident, and the risk value of the causal node in all accidents, thus obtaining the causal risk of the power accident.

[0118] It should be noted that the foregoing explanation of the embodiment of the power accident cause analysis method also applies to the power accident cause analysis device of this embodiment, and will not be repeated here.

[0119] According to the power accident causation analysis device proposed in the embodiments of this application, the device acquires the accident report of the power accident and extracts multiple causative events from the accident report; determines the relationship between the multiple causative events based on their order of occurrence and frequency of occurrence, and constructs an accident causation network; analyzes the importance of network nodes through the topological properties of the accident causation network, and calculates the causative risk of the power accident using the cascading failure theory, generating risk management suggestions. This effectively identifies the unique causes of each type of accident, facilitates the identification of the causes that play a significant role in directly causing the accident, and comprehensively considers the severity of the consequences and the frequency of cause exposure, making it convenient for enterprises to apply.

[0120] Figure 24A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0121] The memory 2401, the processor 2402, and the computer program stored on the memory 2401 and capable of running on the processor 2402.

[0122] When the processor 2402 executes the program, it implements the power accident cause analysis method provided in the above embodiments.

[0123] Furthermore, electronic devices also include:

[0124] Communication interface 2403 is used for communication between memory 2401 and processor 2402.

[0125] The memory 2401 is used to store computer programs that can run on the processor 2402.

[0126] The memory 2401 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0127] If the memory 2401, processor 2402, and communication interface 2403 are implemented independently, then the communication interface 2403, memory 2401, and processor 2402 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 24 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0128] Optionally, in a specific implementation, if the memory 2401, processor 2402, and communication interface 2403 are integrated on a single chip, then the memory 2401, processor 2402, and communication interface 2403 can communicate with each other through an internal interface.

[0129] The processor 2402 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0130] This embodiment also provides a computer-readable storage medium storing a computer program thereon, characterized in that the program, when executed by a processor, implements the above-described power accident cause analysis method.

[0131] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0132] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0133] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0134] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0135] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

Claims

1. A method for analyzing the causes of power accidents, characterized in that, Includes the following steps: Obtain accident reports of power outages and extract multiple causal events from the accident reports; The relationship between the multiple causative events is determined based on their order of occurrence and frequency of occurrence, and an accident causation network is constructed. The importance of network nodes is analyzed by examining the topological properties of the accident causation network, and the causal risk of the power accident is calculated using the cascading failure theory to generate risk management recommendations. The analysis of network node importance using the topological properties of the accident causation network includes: Calculate the degree distribution, clustering coefficient, and shortest path length of the accident causation network to determine whether the accident causation network is a scale-free network; When the accident causation network is a scale-free network, the betweenness of the accident causation network is calculated when the weight is a preset value. It is then determined whether the number of parallel nodes with the same betweenness is less than a preset threshold. If it is less, the ranking result corresponding to the preset weight is used as the importance of the network nodes. Otherwise, the size of the preset value is adjusted until the number of parallel nodes with the same betweenness is less than the preset threshold. The calculation of causal risk for power accidents using cascading failure theory includes: When calculating the initial time, the node state, initial hazard load, safety threshold, and propagation strength between nodes in the accident causation network are calculated. When a node in the accident causation network changes to a dangerous state, the dangerous node is taken as the starting node, and all outgoing edges of the starting node are traversed. If there are outgoing edges, the danger load value of each node pointed to by the dangerous node is calculated; otherwise, the danger propagation terminates. Determine whether the current dangerous load value of each node is greater than the safety threshold. If it is greater, the node that is greater than the safety threshold will change from a safe state to a dangerous state; otherwise, the dangerous propagation will terminate. If multiple nodes pointed to by a dangerous node have dangerous load values ​​exceeding the safety threshold, the edge with the strongest propagation strength is selected for dangerous propagation, and the corresponding node state is changed to dangerous state. Starting from the dangerous node, the process is repeated until dangerous propagation ends. The causal risk of power accidents is obtained by calculating the probability of occurrence of the causal chain leading to the accident, the risk value of the causal node in each type of accident, and the risk value of the causal node in all accidents.

2. The method according to claim 1, characterized in that, The step of determining the relationship between the multiple causative events based on their chronological order and frequency of occurrence, and constructing an accident causation network, includes: The direction of the accident causation network edge is determined based on the chronological order of the occurrence of the multiple causative events. The weights of the accident causation network are determined based on the frequency of occurrence of the multiple causative events. Using the multiple causal events as nodes, the accident causation network is constructed according to the direction of the edges and the weights.

3. A device for analyzing the causes of power accidents, characterized in that, include: An extraction module is used to obtain accident reports of power accidents and extract multiple causal events from the accident reports; A construction module is used to determine the relationship between the multiple causative events based on their order of occurrence and frequency, and to construct an accident causation network. The generation module is used to analyze the importance of network nodes through the topological properties of the accident causation network, calculate the causation risk of the power accident using the cascading failure theory, and generate risk management recommendations. The generation module includes: The first judgment unit is used to calculate the degree distribution, clustering coefficient and shortest path length of the accident cause network, and to determine whether the accident cause network is a scale-free network. The second judgment unit is used to calculate the betweenness of the accident cause network when the weight is a preset value when the accident cause network is a scale-free network, and to determine whether the number of parallel nodes with the same betweenness is less than a preset threshold. If it is less than the preset threshold, the ranking result corresponding to the preset weight is used as the importance of the network nodes. Otherwise, the size of the preset value is adjusted until the number of parallel nodes with the same betweenness is less than the preset threshold. The generation module includes: The parameter unit is used to calculate the node state, initial hazard load, safety threshold, and propagation strength between nodes in the accident causation network at the initial time. The conversion unit is used to, when the state of a node in the accident causation network changes to a dangerous state, take the dangerous node as the starting node, traverse all outgoing edges of the starting node, calculate the danger load value of each node pointed to by the dangerous node when outgoing edges exist, otherwise, the danger propagation terminates. The third judgment unit is used to determine whether the current dangerous load value of each node is greater than the safety threshold. If it is greater, the node that is greater than the safety threshold will change from a safe state to a dangerous state; otherwise, the dangerous propagation will terminate. The selection unit is used to select the edge with the strongest propagation strength to propagate danger if the danger load value of multiple nodes pointed to by the danger node exceeds the safety threshold. The corresponding node state is changed to danger state. The dangerous node is used as the starting node to traverse again until the danger propagation ends. The calculation unit is used to calculate the probability of the occurrence of the causal chain leading to the accident, the risk value of the causal node in each type of accident, and the risk value of the causal node in all accidents, so as to obtain the causal risk of the power accident.

4. The apparatus according to claim 3, characterized in that, The building module includes: The first determining unit is used to determine the direction of the accident causation network edge according to the chronological order of the occurrence of the plurality of causative events; The second determining unit is used to determine the weights of the accident causation network based on the number of times the plurality of causative events occur; A network construction unit is used to construct the accident causation network using the plurality of causative events as nodes, according to the direction of the edges and the weights.

5. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the power accident causation analysis method as described in any one of claims 1-2.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the power accident causation analysis method as described in any one of claims 1-2.

Images