Commercial building fire coupling risk assessment method and device

Abstract

The invention provides a commercial building fire coupling risk assessment method and device. The method comprises the following steps: establishing a multi-level interpretation structure model according to accident characteristics, historical data and related technical standards of a commercial building fire; constructing an initial static Bayesian network through risk factor mapping according to the multi-level interpretation structure model; based on commercial building fire accident historical data, optimizing the initial static Bayesian network by using an N-K model to obtain a dynamic Bayesian network model; and according to the dynamic Bayesian network model, determining the occurrence probability of the commercial building fire risk. The accuracy of evaluating the occurrence probability of the fire risk of the commercial building can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fire risk assessment, and particularly to a method and device for coupling risk assessment of commercial building fires. Background Art

[0002] In the modern urbanization process, as an important carrier of urban economic activities, commercial buildings carry multiple functions such as shopping, dining, entertainment, and office work. Their internal structures are complex and changeable, the space layout is dense, and the personnel flow is frequent. This high degree of functionality and density make commercial buildings face more severe challenges in the event of a fire compared to other types of buildings. Once a fire occurs, due to factors such as intricate internal passageways, numerous combustibles, and difficult personnel evacuation in the building, it is extremely easy for the fire to spread rapidly, causing significant casualties and property losses.

[0003] In view of the particularity and severity of the fire risk in commercial buildings, it is particularly important to conduct scientific and accurate fire risk assessments. Fire risk assessment can not only identify potential fire hazards in advance, provide a scientific basis for building fire prevention design, fire protection facility configuration, and emergency plan formulation, but also guide effective rescue operations in the event of a fire, minimizing losses to the greatest extent.

[0004] However, the currently applied fire risk assessment methods have low assessment accuracy and are difficult to meet the actual needs of commercial building fire risk management. Summary of the Invention

[0005] The present invention provides a method and device for coupling risk assessment of commercial building fires to solve the defect of low accuracy in fire risk assessment in the prior art.

[0006] A method for coupling risk assessment of commercial building fires includes:

[0007] S1. Establish a multi-level interpretive structural model according to the accident characteristics of commercial building fires, historical data of commercial building fires, and relevant technical standards for commercial building fire risk assessment;

[0008] S2. Construct an initial static Bayesian network through risk factor mapping according to the multi-level interpretive structural model;

[0009] S3. Optimize the initial static Bayesian network based on historical data of commercial building fire accidents using the N-K model to obtain a dynamic Bayesian network model;

[0010] S4. Determine the occurrence probability of commercial building fire risk according to the dynamic Bayesian network model.

[0011] According to the method for coupling risk assessment of commercial building fires provided by the present invention, the S1 includes:

[0012] S11. Establish a commercial building fire risk assessment system according to the accident characteristics, historical data of commercial building fires, and relevant technical standards for commercial building fire risk assessment;

[0013] S12. Establish a multi-level interpretive structural model according to the commercial building fire risk assessment system;

[0014] Among them, the commercial building fire risk assessment system includes: an objective layer, a criterion layer, and an index layer;

[0015] Multiple criterion factors corresponding to the objective layer and multiple index factors corresponding to each criterion factor;

[0016] The objective layer includes: the building fire hazard level;

[0017] The criterion layer includes multiple criterion factors corresponding to the building fire hazard level, including: electrical factors, fire protection systems, building environment, and fire protection management;

[0018] The index factors corresponding to the electrical factors include: power transformation and distribution devices, electrical circuits, electrical equipment, grounding systems, and electrical protection devices;

[0019] The index factors corresponding to the fire protection systems include: automatic fire alarm systems, automatic fire extinguishing systems, smoke prevention and exhaust systems, fire water supply and fire hydrant systems, emergency lighting and evacuation indication signs, fire extinguishing and rescue equipment, and fire protection power supplies;

[0020] The index factors corresponding to the building environment include: building layout, fire resistance rating, fire prevention, smoke prevention zones, fire separation measures, and safe evacuation;

[0021] The index factors corresponding to the fire protection management include: emergency plans and drills, fire prevention inspections and hidden danger rectification, fire and electricity use management systems, fire safety education and training, and fire safety systems.

[0022] According to the commercial building fire coupling risk assessment method provided by the present invention, the S12 includes:

[0023] (1) Construct a directed adjacency matrix F for the index factors according to whether the index factors affect each other, F = [f ij n×n ; where f ij represents the influence relationship of risk f i on f j , and n is the number of index factors;

[0024] (2) Perform Boolean operations on the directed adjacency matrix F to obtain a connectivity matrix M, and the calculation formula is:​

[0025]

[0026] Where r = 1, 2, 3, ..., m; I is the identity matrix of the same order as F;

[0027] (3) Decompose the connectivity matrix to obtain the reachable set R(f i ) and the antecedent set Q(f i ): The reachable set R(f i ) represents all risk index sets that can be reached directly or indirectly starting from the risk index f i ; The antecedent set Q(f i ) represents all risk index sets that can reach the risk index f i directly or indirectly;

[0028] (4) Determine the highest set U(f i ) according to the intersection of the reachable set R(f i ) and the antecedent set Q(f i );

[0029] (5) Use the highest set U(f i ) as the core risk index set, delete the rows and columns corresponding to the core risk index set to form a new connectivity matrix, and execute steps (3)-(4) again until all risk indexes are assigned to the corresponding levels, so as to obtain the hierarchical division result;

[0030] (6) Construct a multi-level interpretive structure model of the commercial building fire risk factors according to the reachable set R(f i ), the antecedent set Q(f i ), the highest set U(f i ), and the hierarchical division result.

[0031] According to the commercial building fire coupling risk assessment method provided by the present invention, the S3 includes:

[0032] S31. Use the N-K model to analyze the historical data of the commercial building fire accidents to determine the coupling risk types between different criterion factors;

[0033] S32. Determine the coupling degree of each coupling risk type;

[0034] S33. Optimize the initial static Bayesian network based on the coupling degree of each coupling risk type and the time series characteristics corresponding to each criterion factor to obtain a dynamic Bayesian network model.

[0035] According to the commercial building fire coupling risk assessment method provided by the present invention, the coupling risk types include: two-factor coupling, three-factor coupling, and four-factor coupling;

[0036] The two-factor coupling includes: electricity-system, electricity-building, electricity-management, system-building, system-management, building-management;

[0037] The three-factor coupling includes: electricity-system-building, electricity-system-management, electricity-building-management, system-building-management;

[0038] The four-factor coupling includes: electricity-system-building-management.

[0039] According to the commercial building fire coupling risk assessment method provided by the present invention, the coupling degree is calculated by the following formula:

[0040]

[0041] In the formula, a, b, c, and d respectively represent the electrical factor, the fire protection system, the building environment, and system management, and P h,i,j,k represents the probability of coupling occurrence when the electrical factor is in the h state, the fire protection system is in the i state, the building environment is in the j state, and the fire protection management is in the k state. The risk factors in the criterion layer are defined as two states, 0 representing that the risk factor does not occur, and 1 representing that the risk factor occurs.

[0042] According to the commercial building fire coupling risk assessment method provided by the present invention, the dynamic Bayesian network model is as follows:

[0043]

[0044] In the formula, represents the static Bayesian network, represents the transfer network of each index factor in the time dimension, is the i-th index in the initial state, is the i-th index at the t-th moment, is the parent node of; is the parent node of, T represents the entire time span corresponding to the dynamic Bayesian network, t represents a certain time point on the time span, and N represents the number of index factors.

[0045] According to the commercial building fire coupling risk assessment method provided by the present invention, the S4 includes:

[0046] S41. Calculate the prior probability of each coupling risk and the prior probability of the factor at the index layer respectively based on the coupling degree of each coupling risk type and historical data.

[0047] S42. Establish a conditional probability table according to the prior probability of each coupling risk and the prior probability of the factor at the index layer.

[0048] S43. Based on the conditional probability table, use the dynamic Bayesian network model for forward reasoning to determine the occurrence probability of the commercial building fire risk.

[0049] The fire risk level of the building is divided into four levels according to the occurrence probability of the fire risk. Among them, when 0 ≤ P < 0.2, it is the first-level risk, and it can operate normally and be maintained daily; when 0.2 ≤ P < 0.4, it is the second-level risk, and it needs to be monitored and maintained intensively; when 0.4 ≤ P < 0.7, it is the third-level risk, and it should be used under control and repaired within a time limit; when 0.7 ≤ P < 1, it is the fourth-level risk, and it should be stopped using and repaired immediately.

[0050] According to the commercial building fire coupling risk assessment method provided by the present invention, the prior probability of the factor at the index layer is determined in the following manner:

[0051] (1). According to the magnitude of the occurrence probability of each index factor, use the LEC evaluation method to divide the risk of each event into different risk levels. The levels include: VL level, L level, M level, H level, VH level. Among them, the risk of each event is jointly determined by the index factor and the criterion factor.

[0052] (2). Construct a corresponding membership function for each of the risk levels.

[0053]

[0054] Among them, . . . . are the membership degrees corresponding to the VL level, L level, M level, H level, and VH level respectively, is a fuzzy number;

[0055] (3). Apply the judgment results of experts on the occurrence probability of each risk factor to the corresponding membership function to obtain fuzzy numbers; and summarize the fuzzy numbers of all experts to obtain the final overall fuzzy number of the same factor at the index layer.

[0056] (4). Convert the final overall fuzzy number of the same factor at the index layer into a fuzzy possibility score.

[0057] (5) Convert the fuzzy possibility scores of factors at the same index layer into prior probabilities.

[0058] The present invention also provides a commercial building fire coupling risk assessment device, comprising:

[0059] A building unit, configured to establish a multi-level interpretive structural model according to the accident characteristics of commercial building fires, historical data of commercial building fires, and relevant technical standards for commercial building fire risk assessment;

[0060] A construction unit, configured to construct an initial static Bayesian network through risk factor mapping according to the multi-level interpretive structural model;

[0061] An optimization unit, configured to optimize the initial static Bayesian network based on historical data of commercial building fire accidents using the N-K model to obtain a dynamic Bayesian network model;

[0062] A determination unit, configured to determine the occurrence probability of commercial building fire risk according to the dynamic Bayesian network model.

[0063] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the program, it implements the commercial building fire coupling risk assessment method as described in any one of the above.

[0064] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the commercial building fire coupling risk assessment method as described in any one of the above.

[0065] The present invention also provides a computer program product, comprising a computer program, and when the computer program is executed by a processor, it implements the commercial building fire coupling risk assessment method as described in any one of the above.

[0066] The commercial building fire coupling risk assessment method and device provided by the present invention, through historical data of commercial building fire accidents, using the N-K model, optimize the initial static Bayesian network to obtain a dynamic Bayesian network model, and use the dynamic Bayesian network model to determine the occurrence probability of commercial building fire risk, which can effectively improve the accuracy of the occurrence probability of fire risk. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 is a schematic flowchart of the commercial building fire coupling risk assessment method provided by the present invention;

[0068] Figure 2 is a schematic diagram of the commercial building fire risk index system provided by the present invention;

[0069] Figure 3 is the multi-level interpretive structural model provided by the present invention;

[0070] Figure 4 is the schematic diagram of the static Bayesian network structure for the fire risk of commercial buildings provided by the present invention;

[0071] Figure 5 is the schematic diagram of the coupling risk evolution mechanism provided by the present invention;

[0072] Figure 6 is the schematic diagram of the dynamic Bayesian network structure for the coupled fire risk of commercial buildings provided by the present invention;

[0073] Figure 7 is the block diagram of the structure of the device for assessing the coupled fire risk of commercial buildings provided by the present invention;

[0074] Figure 8 is the schematic diagram of the structure of the electronic device provided by the present invention. Detailed implementation manners

[0075] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0076] Currently, there are the following difficulties in the technology for assessing the fire risk of commercial buildings: First, in the fire risk assessment, electrical factors should be regarded as an important assessment index because it is directly related to the safety and reliability of the internal electrical system of the building. In the existing fire risk assessment system for commercial buildings, the electrical factor index is relatively single. The risk evaluation index of electrical factors should be multiple and complex, and various factors need to be comprehensively considered to fully evaluate the safety performance of the electrical system; Second, the coupling effect between risks is the main driving force for risk evolution, while the existing fire risk assessment methods do not fully consider the interdependent relationship between various risk factors, and cannot identify the dependence relationship between the accident sequence and various risk factors and quantitatively evaluate their coupling relationship. In view of the deficiencies of the existing technology, the present invention provides a method for assessing the coupled fire risk of commercial buildings, Figure 1 is the schematic diagram of the process of the method for assessing the coupled fire risk of commercial buildings, as Figure 1 shown, this method includes:

[0077] S1. According to the accident characteristics, historical data of commercial building fires and relevant technical standards for commercial building fire risk assessment, establish a multi-level interpretive structural model.

[0078] Specifically, the accident characteristics of commercial building fires include: cause of fire, fire development characteristics, response of fire protection systems, and building structure. The historical data of commercial building fires include: historical fire accident investigation reports and building fire protection assessment reports. Among them, the historical fire accident investigation report is a document that details and analyzes past commercial building fire accidents. These reports usually contain information such as the time and location of the fire, cause of fire, fire development, response of fire protection systems, casualties, property damage, and investigation conclusions after the accident. By analyzing historical fire accident investigation reports, key risk factors leading to fires can be identified, such as electrical failures, illegal operations, and failures of fire protection systems. The building fire protection assessment report is a document formed after conducting fire safety inspections and assessments of commercial buildings. These reports usually contain information such as the fire protection system design of the building, status of fire protection facilities, implementation of fire protection management systems, fire hazard investigation and rectification, etc. Through the building fire protection assessment report, the current fire protection situation of commercial buildings can be comprehensively understood, including the completeness, effectiveness, and compliance of fire protection facilities, etc. The commercial building fire risk assessment standard refers to a series of specifications, standards, and guidelines formulated by authoritative agencies or industry organizations in the field of commercial building fire risk assessment to guide risk assessment practices. These technical standards usually cover multiple aspects such as building fire prevention design, fire protection system configuration, fire risk assessment methods, data collection and analysis, etc.

[0079] The following explains step S1, which specifically includes the following two steps:

[0080] S11. Establish a commercial building fire risk assessment system based on the accident characteristics, historical data, and relevant technical standards of commercial building fires; S12. Establish a multi-level interpretive structural model based on the commercial building fire risk assessment system; among them, the commercial building fire risk assessment system includes: target layer, criterion layer, and index layer.

[0081] Specifically, the present invention first identifies the key risk factors leading to fires based on the accident characteristics (such as cause of fire, fire development characteristics, response of fire protection systems, building structure, etc.) and historical data (such as historical fire accident investigation reports, building fire protection assessment reports, etc.) of commercial building fires. During the identification process, relevant technical standards such as building fire prevention codes and fire protection system design codes also need to be referred to ensure that the identified risk factors meet industry standards and practical requirements. Then, the identified risk factors are classified and sorted to construct a commercial building fire risk assessment system including a target layer, a criterion layer, and an index layer. Then, based on the commercial building fire risk system, the evolution relationship between the factors in the index layer is analyzed through fire accident case studies to establish a multi-level interpretive structural model.

[0082] Based on accident characteristics and historical data, this application identifies and sorts out the key fire risk factors in commercial buildings, and then summarizes and classifies these fire risk factors in combination with relevant standards, so as to construct a fire risk assessment system for commercial buildings. As Figure 2 shown, among them, the commercial building fire assessment system includes: a first-level risk indicator (i.e., the target layer): the fire danger level of the building; four second-level risk indicators (i.e., the criterion layer): electrical factors, fire protection systems, building environment, and fire protection management, as well as third-level risk indicators (indicator factors) corresponding to each second-level risk. Among them, the five indicator factors corresponding to electrical factors are transformer and distribution devices, electrical lines, electrical equipment, grounding systems, and electrical protection devices; the seven indicator factors corresponding to the fire protection system are: automatic fire alarm system, automatic fire extinguishing system, smoke prevention and exhaust system, fire water supply and fire hydrant system, emergency lighting and evacuation indication signs, fire extinguishing and rescue equipment, and fire protection power supply; the five indicator factors corresponding to the building environment are: building layout, fire resistance rating, fire prevention, smoke prevention zoning, fire separation measures, and safe evacuation; the five indicator factors corresponding to fire protection management are: emergency plans and drills, fire prevention inspections and hidden danger rectification, fire and electricity use management systems, fire safety education and training, and fire safety systems. Based on whether there is a direct impact between all indicator factors, a multi-level interpretive structural model is constructed, as Figure 3 shown, where level one {S3, S7, S8, S9, S 11 , S 15 , S 17} is the direct occurrence factor of the risk system, levels two to five {S2, S4, S5, S 10 , S 13 , S 14 , S 16 , S 18 , S 20}, {S1, S6, S 12}, {S 19}, {S 21} are the intermediate transition factors of the risk system, and level six {S 22} is the fundamental source factor of the risk system.

[0083] The commercial building fire coupling risk assessment method provided by the present invention establishes a multi-level interpretive structural model through the accident characteristics, historical data of commercial building fires, and commercial building fire risk assessment standards to establish a commercial building fire risk assessment system, making the established assessment system more diverse and complex, so as to comprehensively consider various factors to comprehensively evaluate the safety performance of the electrical system, and effectively improving the accuracy of the probability of commercial building fire risk occurrence.

[0084] S2. According to the multi-level interpretive structural model, through risk factor mapping, an initial static Bayesian network is constructed.

[0085] Specifically, in the multi-level interpretive structural model, each factor is converted into a node, and the relationship between each factor is converted into a directed edge. The "commercial building fire risk level" is defined as the risk top event I, and an initial static Bayesian network is constructed, as Figure 4 shown, where Y indicates the occurrence of the risk and N indicates the non-occurrence of the risk.

[0086] S3. Based on the historical data of commercial building fire accidents, using the N-K model, the initial static Bayesian network is optimized to obtain a dynamic Bayesian network model.

[0087] Specifically, during the operation of commercial buildings, adverse impacts on electrical factors, fire protection systems, building environment, and fire management will impact the risk system. However, a single risk factor often cannot break through the safety threshold of the risk system. When multiple factors are coupled together, the risk level expands, causing damage to the risk system and ultimately leading to accidents. The risk coupling types can be divided into three categories: single-factor coupling, two-factor coupling, and multi-factor coupling, as Figure 5 shown. Since the initial static Bayesian network cannot estimate the non-linear synergistic effect between risk factors, and the N-K model, as a tool for analyzing the interactions of complex adaptive systems, can, through historical data, identify which combinations of risk factors will cause an exponential increase in the probability of fire accidents. For example, the probability of a fire caused by the electrical circuit of the third-level risk index is 10%; the probability of a fire caused by the fire alarm system is 3%; but when the electrical circuit and the fire alarm system occur simultaneously, the probability of a fire is 40%, rather than 13%. Therefore, through the N-K model, the accuracy of predicting the probability of fire risk occurrence can be improved.

[0088] In addition, since the static Bayesian network can only analyze the risk relationship at the current moment and cannot predict the probability of fire risk occurrence based on real-time data, it will seriously underestimate the probability of a fire occurring. In this application, by combining the historical data of commercial building fire accidents, the initial static Bayesian network is optimized, so that the obtained dynamic Bayesian network model can know the probability of fire risk occurrence at each moment and predict the probability of fire risk occurrence at a future moment, thereby improving the accuracy and time range of predicting the probability of fire risk occurrence. At the same time, combined with the N-K model, the N-K model is used to model the interactions between the internal variables of the fire system, revealing the coupling relationship between different risk factors, thereby improving the prediction accuracy of the probability of fire risk occurrence.

[0089] S4. According to the dynamic Bayesian network model, determine the occurrence probability of the commercial building fire risk.

[0090] Specifically, the occurrence probability of the fire risk in commercial buildings can be used to determine the fire risk level of the building, and corresponding risk management measures can be proposed based on the fire risk level.

[0091] The commercial building fire coupling risk assessment method provided by the present invention uses the historical data of commercial building fire accidents, utilizes the N-K model to optimize the initial static Bayesian network to obtain a dynamic Bayesian network model, and uses the dynamic Bayesian network model to determine the occurrence probability of the commercial building fire risk, which can effectively improve the accuracy of the fire risk occurrence probability.

[0092] Furthermore, the following introduces how to establish a multi-level interpretive structural model, which specifically includes the following steps:

[0093] (1) According to whether the index factors affect each other, construct a directed adjacency matrix F for the index factors, that is, F = [f ij n×n , where f ij represents the influence relationship of risk f i on f j , and n is the number of index factors.

[0094] Specifically, based on fire accident cases, analyze the evolutionary relationship of the index factors of the fire assessment system. During the occurrence process of a fire accident, if the risk index f i has a direct impact on the occurrence of f j , then assign f ij as 1. For example, aging or short circuit of the line may directly damage electrical equipment. At this time, the electrical line S2 has a direct impact on the electrical equipment S3, then f 23 (i≠j) is assigned as 1. While the failure of the substation equipment will cause the electrical line to be overloaded, and then cause damage to the electrical equipment. At this time, the substation equipment S1 has no direct impact on the electrical equipment S3, then f 13 is assigned as 0; the same risk index (i = j) is assigned as 0, and construct the directed adjacency matrix F of the third-level index, that is, F = [f ij n×n .

[0095] (2) Perform Boolean operations on the directed adjacency matrix F to obtain the connectivity matrix M, and the calculation formula is:

[0096]

[0097] In the formula, r = 1, 2, 3,..., m, m is a non-zero constant; I is the identity matrix of the same order as F;

[0098] (3) Decompose the connectivity matrix to obtain the reachable set R(f i ) and the antecedent set Q(f​​i )); among them, the reachable set R(f i ) represents the set of all risk indicators that can be reached directly or indirectly starting from the risk indicator f i , that is, in the corresponding row of an element in the connectivity matrix, the set of elements containing 1; the antecedent set Q(f i ) represents the set of all risk indicators that can reach the risk indicator f i directly or indirectly, that is, in the corresponding column of an element in the connectivity matrix, the set of elements containing 1;

[0099] (4) Take the intersection of the reachable set R(f i ) and the antecedent set Q(f i ) as the highest set U(f i );

[0100] (5) Take the highest set U(f i ) as the core risk indicator set, remove the corresponding rows and columns in this core risk indicator set to form a new connectivity matrix, and execute steps (3)-(4) again until all risk indicators are assigned to the corresponding levels, so as to obtain the hierarchical division result, as shown in Table 1:

[0101] Table 1:

[0102] (6) According to the reachable set R(f i ), the antecedent set Q(f i ), the highest set U(f i ), and the hierarchical division result, construct a multi-level interpretive structure model of the fire risk factors of commercial buildings.

[0103] The method provided by the present invention constructs a directed adjacency matrix F and performs Boolean operations to obtain a connectivity matrix M, and then decomposes the connectivity matrix to obtain a reachable set and an antecedent set to construct a multi-level interpretive structure model, so that the constructed multi-level interpretive structure model improves the accuracy and efficiency of the fire risk assessment of commercial buildings.

[0104] Further, step S2 is introduced in detail below:

[0105] First, according to the hierarchical relationship among the target layer, criterion layer, and index factors, each risk factor in the multi-level interpretive structural model is mapped to a node in the Bayesian network. Based on the risk factor association relationships in the multi-level interpretive structural model, directed edges are established in the Bayesian network. The direction of the edges represents the causal or conditional dependence relationships among the risk factors; historical data related to each risk factor is collected, including the occurrence frequencies of each factor in different states and the joint occurrence frequencies among the factors; based on the collected data, the conditional probabilities of each node in different states of its parent nodes are estimated. These conditional probabilities will form the conditional probability table, which is used for the inference calculation of the Bayesian network. All nodes, directed edges, and conditional probability tables are integrated together to form the structure of the static Bayesian network.

[0106] Further, the following is a detailed introduction to step S3, which specifically includes the following solutions:

[0107] First, use the N-K model to analyze the historical data of commercial building fire accidents to determine the coupling risk types among different criterion factors; second, determine the coupling degree of each coupling risk type; finally, based on the coupling degree of the coupling risk types, the coupling disaster-causing characteristics of commercial building fire risks, and the time-series characteristics of the index factors, optimize the initial static Bayesian network to obtain the dynamic Bayesian network model.

[0108] Specifically, the historical data of commercial building fire accidents includes: information such as the time, location, cause, loss situation, operation status of fire-fighting facilities, and building usage nature of the fire. Process these historical data of commercial building fire accidents to obtain the encoded historical data; take the encoded historical data as input and substitute it into the N-K model to obtain the commercial building fire accident model; then, substitute all criterion factors (electrical factors, fire-fighting system, building environment, and fire-fighting management) into the commercial building fire accident model to analyze the coupling relationships among different criterion factors, so as to obtain the coupling risk types among different criterion factors. For example: The implementation of the fire and electricity use management system is not strict, there is illegal hot work in the building, combustibles are ignited, and at the same time, the fire automatic alarm system fails, which causes the fire to spread, and then a fire accident occurs. This accident is caused by the coupling of the fire-fighting system and fire-fighting management. In this case, the failure of the criterion factor (fire-fighting system) and the lack of the criterion factor (fire-fighting management) act together to cause the fire. Then, a two-factor coupling (system - management) is formed between the fire-fighting system and fire-fighting management. Then, calculate the coupling degree of each coupling risk type. Finally, incorporate the calculated coupling degree and the determined disaster-causing characteristics (such as fire spread speed, smoke diffusion characteristics, etc.) into the static Bayesian network to adjust the parameters of the static Bayesian network, and then optimize the static Bayesian network with adjusted parameters based on the time-series characteristics of all index factors to obtain the dynamic Bayesian network model. The dynamic Bayesian network model is as follows:

[0109]

[0110] In the formula, B0 represents the initial state of the fire index factor of commercial buildings (i.e., the static Bayesian network), and B → represents the transfer network of each index factor in the time dimension. is the i-th index in the initial state. is the i-th index at time t. is 's parent node; is 's parent node. T represents the entire time span corresponding to the dynamic Bayesian network, t represents a certain time point on the time span, and N represents the number of index factors.

[0111] In the embodiments of the present invention, the coupled risk types include three categories, namely: two-factor coupling, three-factor coupling, and four-factor coupling.

[0112] Among them, the two-factor coupling includes: electrical-system O 21（a,b） , electrical-building O 22（a,c） , electrical-management O 23（a,d） , system-building O 24（b,c） , system-management O 25（b,d） , building-management O 26（c,d） ; the three-factor coupling includes: electrical-system-building O 31（a,b,c） , electrical-system-management O 32（a,b,d） , electrical-building-management O 33（a,c,d） , system-building-management O 34（b,c,d） ; the four-factor coupling includes: electrical-system-building-management O 4（a,b,c,d） .

[0113] Taking the four-factor coupling as an example below, it explains how to obtain the coupling degree of each coupled risk type. Specifically, refer to the following formula:

[0114]

[0115] In the formula, a, b, c, and d respectively represent the electrical factor, the fire protection system, the building environment, and the fire protection management. P h,i,j,k represents the probability of coupling occurrence when the electrical factor is in state h, the fire protection system is in state i, the building environment is in state j, and the fire protection management is in state k. The risk factors in the criterion layer are defined as two states, that is, 0 represents that the risk factor does not occur, and 1 represents that the risk factor occurs. For example, P .1.. represents that the fire protection system has a risk, and the other factors can either occur or not occur. Then, the coupling degrees of various coupled risk types are shown in Table 2:

[0116] Table 2

[0117] Furthermore, the following introduces how to determine the occurrence probability of the fire risk in commercial buildings according to the dynamic Bayesian network model, specifically including the following solutions:

[0118] First, based on the coupling degree of each coupled risk type and historical data, calculate the prior probability of each coupled risk and the prior probability of the index layer factors respectively.

[0119] Specifically, according to the coupling degree of each coupled risk type and historical data, obtain the prior probability of each type of coupled risk and the prior probability of the index layer risk factors respectively. Among them, the coupling degree of different coupled risk types is calculated based on the number of risk couplings. The greater the coupling degree, the greater the risk of this type and the greater the probability of simultaneous occurrence. Then, the coupling degree can represent the prior probability of different coupled risk types. The prior probability of the index layer factors is obtained based on the LEC evaluation method.

[0120] Second, establish a conditional probability table according to the prior probability of each coupled risk and the prior probability of the index layer factors.

[0121] Specifically, in a Bayesian network, a non-root node refers to a node that has at least one parent node, that is, it has one or more incoming edges and is not the top-level node in the network. Non-root nodes represent factors or observations in the network that depend on other variables, and their states are directly affected by their parent nodes. The conditional probability can be calculated by the following formula:

[0122]

[0123] Among them, P(S i ), P(S j ) are prior probabilities, representing people's understanding and estimation under the existing knowledge and experience system. P(S i |S j ) is the posterior probability, representing people's new understanding after obtaining new information (knowing that S j occurs).

[0124] Taking the electrical circuit factor as an example, the state of the emergency plan and drill S 18 is directly affected by the fire safety system S 22 and the fire safety education and training S 21 . The expert judgment results of the conditional probability distribution of the emergency plan and drill S 18 are shown in Table 3.

[0125] Table 3

[0126]

[0127] Then, defuzzify the expert judgment results to obtain the contingency plan and drill S 18 The conditional probability values are shown in Table 4. In the table Indicates that the factor occurs Indicates that the factor does not occur

[0128] Table 4

[0129]

[0130] Finally, based on the above conditional probability table, use the dynamic Bayesian network model for forward reasoning to determine the occurrence probability of the commercial building fire risk

[0131] Specifically, since each factor in the risk system has two states, that is, 0 represents that the risk factor does not occur, and 1 represents that the risk factor occurs, then the state transition matrix can be obtained

[0132]

[0133] Among them, p is the probability that the system factor changes from the normal state to the failure state. When the value of p is equal to The following conditional probabilities can be obtained

[0134]

[0135] According to the prior probabilities, conditional probabilities and transfer networks of each risk factor, construct a dynamic Bayesian network model considering the risk coupling effect, as shown in Figure 6 The building fire risk level provided by the embodiment of the present invention is divided into four levels according to the occurrence probability of the fire risk. When 0 ≤ P < 0.2, it is risk level one and can operate normally and perform daily maintenance; when 0.2 ≤ P < 0.4, it is risk level two and needs to be monitored and maintained; when 0.4 ≤ P < 0.7, it is risk level three and should be controlled for use and repaired within a time limit; when 0.7 ≤ P < 1, it is risk level four and should be stopped using and repaired immediately

[0136] Furthermore, the calculation method of the prior probabilities of the risk factors in the above index layer is as follows

[0137] Since the commercial space fire risk system is a complex network system involving multiple risk factors and having dynamics and fuzziness, the Delphi method is used to select relevant experts in this field to judge the failure possibility of basic events

[0138] (1) According to the probability of occurrence of each indicator factor, the LEC evaluation method is used to classify the risk of each event into different risk levels; the levels include: very high (VH), high (H), medium (M), low (L) and very low (VL). Among them, the risk of each event is determined based on the indicator factor and the criterion factor;

[0139] (2) For each risk level, construct a corresponding membership function;

[0140] Specifically, different membership functions are constructed according to the different levels of experts' judgment on basic events. For example, formula (1) corresponds to the VL level, formula (2) corresponds to the L level, formula (3) corresponds to the M level, formula (4) corresponds to the H level, and formula (5) corresponds to the VH level. Through these formulas, the experts' judgment level on each basic event is fuzzy processed and converted into the corresponding fuzzy numbers.

[0141]

[0142]

[0143]

[0144]

[0145]

[0146] in, , , , , They are the membership degrees corresponding to VL level, L level, M level, H level, and VH level, respectively. is a fuzzy number.

[0147] (3) Apply the experts’ judgment results on the probability of occurrence of each risk factor to the corresponding membership function to obtain the fuzzy number; and then summarize the fuzzy numbers of all experts to obtain the final overall fuzzy number of all experts for the same basic event.

[0148] Specifically, the judgments of each expert are summarized into fuzzy numbers to obtain the final overall fuzzy number of all experts for the same basic event. The calculation formula is as follows:

[0149]

[0150] Among them, is the total fuzzy number of the ith basic event, is the fuzzy number of the j-th expert’s judgment result on the ith basic event, o is the total number of basic events evaluated, and q is the total number of experts involved in the judgment.

[0151] (4) Convert the overall fuzzy number into a fuzzy possibility score, where F represents the processed fuzzy possibility score:

[0152]

[0153] (5) Convert the fuzzy possibility score into an exact failure probability P F :

[0154]

[0155] For example: The probability judgment results of five experts on the risk of electrical circuit S2 are H, VH, H, M, and VH respectively, and the corresponding fuzzy numbers are (0.6, 0.75, 0.9), (0.8, 0.9, 1), (0.6, 0.75, 0.9), (0.3, 0.5, 0.7), (0.8, 0.9, 1). Then the overall fuzzy number is (0.62, 0.76, 0.9). After defuzzifying the overall fuzzy number, the probability of the risk of electrical circuit S2 can be obtained as 0.0282.

[0156] Finally, a comprehensive introduction to the commercial building fire coupling risk assessment method provided by the present invention is given, including the following steps:

[0157] S1. Based on the accident characteristics, historical data, and relevant technical standards of commercial building fires, starting from the main disaster-causing factors of commercial buildings, identify the key fire risk factors of commercial buildings and establish a three-level assessment index system for commercial building fire risks;

[0158] S2. According to the commercial building fire risk index system, establish a multi-level interpretive structural model by analyzing the evolution relationship between the third-level fire risk index factors through fire accident case analysis;

[0159] S3. According to the multi-level interpretive structural model, obtain the initial static Bayesian network through risk factor mapping;

[0160] S4. Based on the historical data of commercial building fire accidents, use the N-K model to analyze the coupling disaster-causing characteristics of the second-level risk index factors and obtain the coupling degrees of different coupling risk types;

[0161] S5. Based on the coupling disaster-causing characteristics of commercial building fire risks and the time-series characteristics of risk factors, improve the initial static Bayesian network and establish a dynamic Bayesian network model for commercial building fire coupling risks;

[0162] S6. Based on the coupling degrees and historical data of coupling risks, calculate the prior probabilities of coupling risks and the third-level risk index factors respectively and establish a conditional probability table;

[0163] S7. Based on the conditional probability table, use the dynamic Bayesian model to infer the occurrence probability of the commercial building fire risk in the forward direction, determine the building fire risk level, and propose corresponding risk management measures.

[0164] In summary, the present invention can achieve an objective and accurate dynamic assessment of the commercial building fire coupling risk, identify the key factors affecting the occurrence of the fire, and propose targeted control measures, effectively improving the building safety.

[0165] The commercial building fire coupling risk assessment device provided by the present invention will be described below. The commercial building fire coupling risk assessment device described below can be mutually referred to the commercial building fire coupling risk assessment method described above.

[0166] As Figure 7 shown, the commercial building fire coupling risk assessment device provided by the embodiment of the present invention includes:

[0167] A building unit 701, configured to establish a multi-level interpretive structural model according to the accident characteristics of the commercial building fire, the historical data of the commercial building fire, and the relevant technical standards for the commercial building fire risk assessment;

[0168] A construction unit 702, configured to construct an initial static Bayesian network through risk factor mapping according to the multi-level interpretive structural model;

[0169] An optimization unit 703, configured to optimize the initial static Bayesian network based on the historical data of the commercial building fire accidents by using the N-K model to obtain a dynamic Bayesian network model;

[0170] A determination unit 704, configured to determine the occurrence probability of the commercial building fire risk according to the dynamic Bayesian network model.

[0171] Figure 8 An entity structure diagram of an electronic device is exemplified. As Figure 8 shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840. Among them, the processor 810, the communication interface 820, and the memory 830 complete the communication with each other through the communication bus 840. The processor 810 can call the logical instructions in the memory 830 to execute the commercial building fire coupling risk assessment method, and the method includes:

[0172] S1. Establish a multi-level interpretive structural model according to the accident characteristics of the commercial building fire, the historical data of the commercial building fire, and the relevant technical standards for the commercial building fire risk assessment;

[0173] S2. According to the multi-level interpretive structural model, through risk factor mapping, construct an initial static Bayesian network;

[0174] S3. Based on the historical data of commercial building fire accidents, use the N-K model to optimize the initial static Bayesian network to obtain a dynamic Bayesian network model;

[0175] S4. According to the dynamic Bayesian network model, determine the occurrence probability of commercial building fire risks.

[0176] In addition, when the logical instructions in the above-mentioned memory 830 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0177] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the commercial building fire coupling risk assessment method provided by the above-mentioned various methods. The method includes:

[0178] S1. According to the accident characteristics of commercial building fires, the historical data of commercial building fires, and the relevant technical standards for commercial building fire risk assessment, establish a multi-level interpretive structural model;

[0179] S2. According to the multi-level interpretive structural model, through risk factor mapping, construct an initial static Bayesian network;

[0180] S3. Based on the historical data of commercial building fire accidents, use the N-K model to optimize the initial static Bayesian network to obtain a dynamic Bayesian network model;

[0181] S4. According to the dynamic Bayesian network model, determine the occurrence probability of commercial building fire risks.

[0182] On the other hand, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it realizes the commercial building fire coupling risk assessment method provided by the above-mentioned various methods. The method includes:

[0183] S1. Establish a multi-level interpretive structural model according to the accident characteristics of commercial building fires, the historical data of commercial building fires, and the relevant technical standards for commercial building fire risk assessment;

[0184] S2. Construct an initial static Bayesian network through risk factor mapping according to the multi-level interpretive structural model;

[0185] S3. Optimize the initial static Bayesian network based on the historical data of commercial building fire accidents by using the N-K model to obtain a dynamic Bayesian network model;

[0186] S4. Determine the occurrence probability of the commercial building fire risk according to the dynamic Bayesian network model.

[0187] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0188] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0189] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.

Claims

1. A method for assessing the coupled risks of commercial building fires, characterized in that, Including: S1. Establish a multi-level interpretive structural model based on the accident characteristics of commercial building fires, historical data of commercial building fires, and relevant technical standards for commercial building fire risk assessment; S2. Construct an initial static Bayesian network through risk factor mapping according to the multi-level interpretive structural model; S3. Optimize the initial static Bayesian network based on historical data of commercial building fire accidents using the N-K model to obtain a dynamic Bayesian network model; S4. Determine the occurrence probability of commercial building fire risk according to the dynamic Bayesian network model.

2. The commercial building fire coupling risk assessment method according to claim 1, characterized in that The S1 includes: S11. Establish a commercial building fire risk assessment system based on the accident characteristics of commercial building fires, historical data, and relevant technical standards for commercial building fire risk assessment; S12. Establish a multi-level interpretive structural model according to the commercial building fire risk assessment system; Among them, the commercial building fire risk assessment system includes: an objective layer, a criterion layer, and an index layer; Multiple criterion factors corresponding to the objective layer and multiple index factors corresponding to each criterion factor; The objective layer includes: the building fire hazard level; The criterion layer includes multiple criterion factors corresponding to the building fire hazard level, including: electrical factors, fire protection systems, building environment, and fire protection management; The index factors corresponding to the electrical factors include: power transformation and distribution devices, electrical circuits, electrical equipment, grounding systems, and electrical protection devices; The index factors corresponding to the fire protection system include: fire automatic alarm systems, automatic fire extinguishing systems, smoke prevention and exhaust systems, fire water supply and fire hydrant systems, emergency lighting and evacuation indication signs, fire extinguishing and rescue equipment, and fire protection power supplies; The index factors corresponding to the building environment include: building layout, fire resistance rating, fire prevention, smoke prevention zones, fire separation measures, and safe evacuation; The index factors corresponding to the fire protection management include: emergency plans and drills, fire prevention inspections and hidden danger rectification, fire and electricity use management systems, fire safety education and training, and fire safety systems.

3. The commercial building fire coupling risk assessment method according to claim 2, wherein The S12 includes: (1) Construct a directed adjacency matrix F for the said index factors according to whether the said index factors affect each other, F = [f ij n×n ; where f ij represents the influence relationship of risk f i on f j , and n is the number of index factors;​ (2) Perform Boolean operations on the directed adjacency matrix F to obtain a connectivity matrix M, and the calculation formula is: In the formula, r = 1, 2, 3,..., m; I is the identity matrix of the same order as F; (3) Decompose the connectivity matrix to obtain the reachable set $R(f$ i ) and the antecedent set $Q(f$ i ): The reachable set $R(f$ i ) represents the set of all risk indicators that can be reached directly or indirectly starting from the risk indicator $f$ i ; The antecedent set $Q(f$ i ) represents the set of all risk indicators that can reach the risk indicator $f$ i directly or indirectly; (4) Determine the highest set U(f i ) based on the intersection of the reachable set R(f i ) and the antecedent set Q(f i ); (5) Using the highest set U(f i ) as the core risk index set, removing the rows and columns corresponding to the core risk index set to form a new connectivity matrix, and then repeating steps (3)-(4) until all risk indicators are assigned to their corresponding levels, thereby obtaining the hierarchical division result; (6) Based on the reachable set R(f i ) and the antecedent set Q(f i ), the highest set U(f i ), and the hierarchical division results, construct a multi-level interpretive structural model of the fire risk factors in commercial buildings.

4. The commercial building fire coupling risk assessment method according to claim 1, wherein The S3 includes: S31. Use the N-K model to analyze the historical data of commercial building fire accidents to determine the coupling risk types between different criterion factors; S32. Determine the coupling degree of each coupling risk type; S33. Optimize the initial static Bayesian network based on the coupling degree of each coupling risk type and the time series characteristics corresponding to each criterion factor to obtain a dynamic Bayesian network model.

5. The commercial building fire coupling risk assessment method according to claim 4, characterized in that, The coupling risk types include: two-factor coupling, three-factor coupling, and four-factor coupling; The two-factor coupling includes: electrical-system, electrical-building, electrical-management, system-building, system-management, building-management; The three-factor coupling includes: electrical-system-building, electrical-system-management, electrical-building-management, system-building-management; The four-factor coupling includes: electricity - system - building - management.

6. The method for evaluating the coupled risk of commercial building fires according to claim 4, wherein The coupling degree is calculated using the following formula: Wherein, a, b, c, and d respectively represent electrical factors, fire protection systems, building environments, and system management, and P h,i,j,k represents the probability of coupling occurrence when the electrical factor is in the h state, the fire protection system is in the i state, the building environment is in the j state, and the fire management is in the k state. The risk factors at the criterion layer are defined as two states, where 0 represents that the risk factor does not occur, and 1 represents that the risk factor occurs.

7. The commercial building fire coupling risk assessment method according to claim 2, characterized in that, The dynamic Bayesian network model is as follows: In the formula, represents a static Bayesian network, represents the transfer network of each index factor in the time dimension, is the i-th index in the initial state, is the i-th index at time t, is 's parent node; is 's parent node, T represents the entire time span corresponding to the dynamic Bayesian network, t represents a certain time point on the time span, and N represents the number of index factors.

8. The commercial building fire coupling risk assessment method according to claim 5, wherein The S4 includes: S41. Based on the coupling degree and historical data of each coupling risk type, calculate the prior probability of each coupling risk and the prior probability of the index layer factors respectively; S42. Establish a conditional probability table according to the prior probability of each coupling risk and the prior probability of the index layer factors; S43. Based on the conditional probability table, use the dynamic Bayesian network model for forward reasoning to determine the occurrence probability of the commercial building fire risk; The commercial building fire risk level is divided into four levels according to the occurrence probability of the fire risk. Among them, when 0 ≤ P < 0.2, it is risk level one, and it can operate normally and perform daily maintenance; when 0.2 ≤ P < 0.4, it is risk level two, and it needs to be monitored and maintained intensively; when 0.4 ≤ P < 0.7, it is risk level three, and it should be controlled for use and repaired within a time limit; when 0.7 ≤ P < 1, it is risk level four, and it should be stopped from being used and repaired immediately.

9. The commercial building fire coupling risk assessment method according to claim 8, wherein The prior probability of the index layer factors is determined according to the following method: (1). According to the magnitude of the occurrence probability of each index factor, use the LEC evaluation method to classify the risks of each event to obtain different risk levels; These levels include: VL level, L level, M level, H level, VH level; among them, the risk of each event is jointly determined by the index factor and the criterion factor; (2). Construct a corresponding membership function for each of the risk levels; Among them, , , , , are the membership degrees corresponding to the VL level, L level, M level, H level, and VH level respectively, is a fuzzy number; (3). Apply the judgment results of experts on the occurrence probability of each risk factor to the corresponding membership function to obtain fuzzy numbers; and summarize the fuzzy numbers of all experts to obtain the final overall fuzzy number of the same index layer factor; (4). Convert the final overall fuzzy number of the same index layer factor into a fuzzy possibility score; (5). Convert the fuzzy possibility score of the same index layer factor into a prior probability.

10. A commercial building fire coupling risk assessment device, characterized in that, It includes: A building unit, which is used to establish a multi-level interpretive structural model according to the accident characteristics of commercial building fires, the historical data of commercial building fires, and the relevant technical standards for commercial building fire risk assessment; A construction unit, which is used to construct an initial static Bayesian network through risk factor mapping according to the multi-level interpretive structural model; An optimization unit, which is used to optimize the initial static Bayesian network based on the historical data of commercial building fire accidents using the N-K model to obtain a dynamic Bayesian network model; A determination unit, which is used to determine the occurrence probability of the commercial building fire risk according to the dynamic Bayesian network model.

Images