Fuzzy Hierarchical Evaluation Method for Spatiotemporal Zoning of Subway Disaster Chain Evolution

Abstract

The present invention relates to the technical field of disaster management, and particularly to a fuzzy hierarchical evaluation method for the spatio-temporal zoning of subway disaster chain evolution. The method includes: dividing the target area into multiple spatial grid areas, recording disaster events in each grid area at different time slices, and extracting disaster spatio-temporal characteristics from the disaster events at different time slices; constructing a disaster chain network, generating a disaster chain spatio-temporal diffusion map and risk zoning based on the analysis of the disaster spatio-temporal characteristics; identifying disaster aggregation hotspots within the risk zoning, and predicting the spatio-temporal information of key disasters through a time prediction model; introducing fuzzy logic, comprehensively evaluating the disaster risk according to the spatio-temporal information of the key disasters, and determining high-risk areas; after determining the high-risk areas, conducting a secondary evaluation of the disaster risk in combination with the subway structure and evacuation capacity.

Description

Technical Field

[0001] The present invention relates to the technical field of disaster management, and particularly to a fuzzy hierarchical evaluation method for spatio-temporal zoning of subway disaster chain evolution. Background Art

[0002] A disaster chain refers to a chain reaction phenomenon of a series of secondary disasters triggered by a primary disaster, forming a spatio-temporally correlated disaster sequence.

[0003] Through the analysis of the disaster chain, disasters can be scientifically managed and warned. Disasters can be controlled by dealing with the disaster-causing points and cutting off the propagation paths.

[0004] The subway system is a special system. When studying the subway disaster chain, it is necessary to consider both the underground system (tunnel structure, track, etc.) and the above-ground system (ground buildings, ground traffic, etc.). Most of the existing studies focus on the research of the underground space. For example, the Chinese patent with the publication number CN 118365209 A uses the analytic hierarchy process (AHP) as the research method. This method has the defects of strong subjectivity (easy to introduce human errors) and sensitivity to the 1-9 scale method (the subjectivity of scale selection will amplify the uncertainty of the results). At the same time, this method has great difficulty in dealing with fuzzy scenarios and cannot adapt to the analysis under complex scenarios. Because the disaster accident data often has incompleteness or ambiguity, this incompleteness and ambiguity lead to a decrease in the credibility of the scoring value and the distortion of the calculation of the transfer probability for optimizing the disaster chain. Summary of the Invention

[0005] The present invention considers the above-ground space and the underground space of the subway system, combines the two to form a three-dimensional space, divides the three-dimensional space into multiple three-dimensional grids, analyzes the spatio-temporal characteristics of disasters in each three-dimensional grid, predicts key disasters using a time prediction model, and combines fuzzy logic and a risk value calculation model to evaluate the disaster risk multiple times.

[0006] The technical solution proposed by the present invention is: a fuzzy hierarchical evaluation method for spatio-temporal zoning of subway disaster chain evolution, and the method includes:

[0007] Dividing the target area into multiple spatial grid areas, recording disaster events in different time slices in each grid area, and extracting the spatio-temporal characteristics of disasters from the disaster events in different time slices;

[0008] Constructing a disaster chain network, generating a spatio-temporal diffusion map and a risk zoning of the disaster chain based on the analysis of the spatio-temporal characteristics of disasters;

[0009] Identifying the disaster aggregation hotspots in the risk zoning, and predicting the spatio-temporal information of key disasters through a time prediction model;

[0010] Introduce fuzzy logic to comprehensively evaluate disaster risks based on the spatio-temporal information of key disasters and determine high-risk areas;

[0011] After determining the high-risk areas, conduct a secondary evaluation of the disaster risks in combination with the subway structure and evacuation capacity.

[0012] Preferably, divide the target area into multiple spatial grid areas, record the disaster events in each grid area at different time slices, and extract the spatio-temporal characteristics of disasters from the disaster events in different time slices, including:

[0013] Obtain the underground data and ground data of the target area;

[0014] Unify the coordinate system and time stamp, align the underground data and ground data spatio-temporally, and construct an underground-ground three-dimensional space;

[0015] Divide the underground-ground space into multiple three-dimensional grids; specifically: segment the target area horizontally and layer it vertically to form multiple three-dimensional grids;

[0016] Obtain the historical disaster data in each three-dimensional grid, and extract the type of disaster event, the number of occurrences of the disaster event, and the occurrence time point of the disaster event from the historical disaster data;

[0017] Calculate the disaster density in the three-dimensional grid area , where represents the number of disasters occurring in the time period of the th grid area, represents the time span, represents the number of three-dimensional grids.

[0018] Preferably, the construction of the disaster chain network includes:

[0019] Obtain the accident data in the target area from the database and extract the disaster-causing factors from the accident data; the accident data includes leakage, settlement, and fire; the disaster-causing factors include rising groundwater level and structural cracks;

[0020] Determine the causal relationship between the disaster-causing factors through the expert knowledge base and construct the disaster propagation path, that is, the disaster chain;

[0021] Use the disaster-causing factors as nodes and the causal relationship between the disaster-causing factors as edges to construct the disaster chain topology graph, that is, the disaster chain network.

[0022] Preferably, the construction of the disaster chain network further includes:

[0023] Model the disaster chain network, including:

[0024] A mechanical - hydrological coupling model is constructed based on the modified Mohr - Coulomb constitutive equation and Darcy's law to simulate soil deformation and groundwater seepage during disasters;

[0025] The mechanical - hydrological model is ; where, represents the effective cohesion, represents the effective internal friction angle, represents the permeability coefficient, represents the pore water pressure; represents the hydraulic head; represents the stress tensor;

[0026] A thermal - structural coupling model is constructed based on the heat conduction equation and the thermal stress equation to simulate heat conduction during disasters;

[0027] The thermal - structural coupling model is ; where, represents heat, represents the thermal diffusivity, represents the heat gradient, represents the coefficient of thermal expansion, represents the elastic modulus; represents the thermal stress;

[0028] A disaster propagation model between various nodes of the disaster chain is established through Bayesian networks or dynamic system equations, specifically including:

[0029] Calculating the probability of disaster propagation between nodes based on Bayesian networks ;

[0030] Describing the evolution process of disaster intensity over time based on differential equations, i.e., dynamic system equations: , where, represents the disaster intensity, respectively represent the diffusion rate and the attenuation coefficient.

[0031] Preferably, based on the analysis of the spatio - temporal characteristics of disasters, a spatio - temporal diffusion map and risk zoning of the disaster chain are generated, including:

[0032] Let the three - dimensional grid of the target area be where represents the indices of the X, Y, and Z axes of the th three - dimensional grid;

[0033] The time slice of each three - dimensional grid is ;

[0034] Obtain the monitoring data of each three - dimensional grid and perform normalization processing; the monitoring data includes settlement and leakage rate;

[0035] Extract the spatio-temporal characteristics of disasters, including:

[0036] Obtain the disaster intensity , where represents the weight coefficient; respectively represent the disaster occurrence probability, disaster diffusion speed, and disaster duration after normalization;

[0037] Obtain the cumulative disaster intensity ; where represents the disaster attenuation coefficient;

[0038] If the disaster intensity of adjacent three-dimensional grids is , then trigger diffusion: ; where represents the diffusion threshold, represents the diffusion efficiency, represents the number of interval grids between two three-dimensional grids;

[0039] Map the cumulative intensity of each three-dimensional grid to a color gradient, and display the risk level of the cumulative disaster intensity through colors;

[0040] Use time as the X-axis and disaster intensity as the Y-axis to display the diffusion process of the disaster intensity;

[0041] Conduct risk zoning through the clustering algorithm K-means++, including:

[0042] Obtain the risk values of all three-dimensional grids, specifically:

[0043] Use the maximum cumulative disaster intensity of each three-dimensional grid as its risk value, that is, the risk value ; represents the disaster occurrence time period within the preset time span ;

[0044] Determine the optimal number of clusters through the elbow method ;

[0045] Establish a hierarchical classification function ; where represents the central value of the cluster; represents the three-dimensional space of the target area;

[0046] Set risk level labels, and the real-time level labels include high risk, medium risk, and low risk;

[0047] When , the three-dimensional grid Divided into high-risk zones; when the three-dimensional grid is divided into medium-risk zones; when the three-dimensional grid is divided into low-risk zones;

[0048] Perform dilation or erosion operations on the clustering results to smooth the boundaries of each risk zone.

[0049] Preferably, identifying the disaster aggregation hotspots in the risk zones includes:

[0050] Identifying the disaster aggregation hotspots in the risk areas through the local spatial autocorrelation analysis algorithm, including:

[0051] Calculating the local Moran's I index in the risk area ; where represents the average disaster intensity value; represents the variance of the global disaster intensity value; represents the spatial adjacency weight; represents the number of three-dimensional grids, represents the disaster intensity value of the

[0052] th adjacent three-dimensional grid;

[0053] Calculating the kernel density of the significantly aggregated area;

[0054] ;

[0055] Where: represents the number of disaster events, represents the Gaussian kernel function; represents the bandwidth; represents the coordinates of the center point of the area, represents the coordinates of the point to be evaluated;

[0056] If the kernel density value is greater than the density threshold, then judge that the aggregated area is a hotspot, otherwise it is a cold spot;

[0057] Predicting the spatio-temporal information of key disasters through the time prediction model, including:

[0058] Constructing a spatio-temporal prediction model for key disasters, including:

[0059] Constructing a spatio-temporal autoregressive moving average model STARMA to predict the disaster intensity at a future time point, that is:

[0060] ;

[0061] Among them, respectively represent the disaster intensity, spatial weight matrix, autoregressive coefficient, and moving average coefficient at the position within the aggregation area at time , represents the order of time lag; and respectively represent the random error term and the moving error term;

[0062] Construct a long short-term memory network (LSTM) and train the LSTM with historical disaster data; use the trained LSTM to predict the type of disaster that will occur at the hot spot at the next time point, that is, the type of key disaster;

[0063] Use a pre-trained spatio-temporal graph convolutional network (ST-GCN) to predict the disaster intensity that will occur at the hot spot within the next time period, that is, the intensity of the key disaster;

[0064] Perform Kriging interpolation on the prediction results to generate a continuous risk surface to reflect the risk distribution range;

[0065] Overlay the prediction results with the hot spot area to identify potential future hot spots.

[0066] Preferably, introduce fuzzy logic to comprehensively evaluate the disaster risk based on the spatio-temporal information of the key disaster and determine the high-risk area, including:

[0067] Obtain the spatio-temporal information of the key disaster within the hot spot area and construct the characteristic variables of the key disaster; the spatio-temporal information of the key disaster includes the disaster duration, diffusion speed, hot spot kernel density, and tunnel deformation amount;

[0068] Construct "IF-THEN" fuzzy rules;

[0069] Based on the fuzzy rules, use the Mamdani fuzzy inference algorithm to output a fuzzy conclusion after inputting the characteristic variables of the key disaster;

[0070] Through the MAX operation, aggregate all the output fuzzy conclusions to form a fuzzy set;

[0071] Through the center of gravity method (COG) defuzzification, calculate the center of gravity position of the fuzzy set;

[0072] Select the conclusion corresponding to the value of the center of gravity position as the output conclusion;

[0073] Calculate the comprehensive risk value ,

[0074] Among them, represents the The value of the centroid position of the fuzzy rule output; Indicates the rule weight; Indicates the number of fuzzy rules;

[0075] Set the high-risk threshold , if then the output hot spot is a high-risk area.

[0076] Preferably, after determining the high-risk area, combined with the subway structure and evacuation capacity, a secondary evaluation of the disaster risk is carried out, including:

[0077] Obtain a list of high-risk areas, including the longitude and latitude of the high-risk area, the type of disaster and the comprehensive risk value;

[0078] Obtain subway structure data, including the length of the subway tunnel, the depth of the subway tunnel, the support material, and the support structure;

[0079] Obtain subway evacuation capacity data, including the location and number of subway entrances and exits, the location and number of emergency exits, the distance from each emergency exit to the high-risk area, and the maximum passenger flow capacity of the subway;

[0080] Normalize the subway structure data and the subway evacuation capacity data;

[0081] Obtain the structural safety factor , where respectively represent the structural strength, the structural cross-sectional area, the safety redundancy factor, and the disaster acting force;

[0082] Obtain the evacuation capacity coefficient ;

[0083] Actual evacuation time , where represents the maximum number of people to be evacuated in the area; respectively represent the number of people that can pass through per second per unit width of the emergency exit, the width of a single emergency exit, and the number of available emergency exits; represents the evacuation attenuation coefficient; represents the minimum evacuation time, ;

[0084] Modify the risk formula to obtain the secondary evaluation score ; where represents the structural weight and the evacuation weight.

[0085] The present invention also provides an electronic device, including a processor and a memory and a communication module connected to the processor, and the electronic device is used to execute the fuzzy hierarchical evaluation method for the spatio-temporal partitioning of the subway disaster chain evolution.

[0086] The present invention also provides a computer-readable storage medium storing a computer program, which is executed by a processor to implement the fuzzy hierarchical evaluation method for spatio-temporal partitioning of subway disaster chain evolution.

[0087] Advantages of the present invention:

[0088] 1. For the subway scenario that includes both ground and underground structures, the present invention takes into account physical coupling (mechanical and hydrological interactions) and system coupling (causal relationships between disaster events), constructs a mechanical-hydrological and thermal-structural model, and determines the nodes (disasters) and edges (causal relationships) of the disaster chain network through the output of the model. By analyzing the disaster chain network, key nodes can be found and targeted prevention and control can be carried out for the key nodes.

[0089] 2. The present invention divides the subway space into multiple three-dimensional grids, identifies high-risk areas by analyzing the spatio-temporal characteristics of disasters in each three-dimensional grid; further analyzes the high-risk areas to determine disaster aggregation hotspots, introduces fuzzy logic to preliminarily evaluate the disaster risks of the disaster aggregation points, can effectively integrate the uncertainties in spatio-temporal information, and generate a more practical high-risk area division. Finally, combined with the different structural characteristics and evacuation capabilities of each subway tunnel, the disaster risks are re-evaluated. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 is a flowchart of the fuzzy hierarchical evaluation method for spatio-temporal partitioning of subway disaster chain evolution according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0091] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious deformations. The basic principles defined in the following description can be applied to other implementation schemes, deformation schemes, improvement schemes, equivalent schemes and other technical schemes that do not deviate from the spirit and scope of the present invention.

[0092] It can be understood that the term "one" should be understood as "at least one" or "one or more". That is, in one embodiment, the number of an element can be one, while in other embodiments, the number of the element can be multiple. The term "one" cannot be understood as a limitation on the number.

[0093] Embodiment 1

[0094] Refer to Figure 1 , the technical solution provided by the present invention is: a fuzzy hierarchical evaluation method for spatio-temporal partitioning of subway disaster chain evolution, including the following steps:

[0095] Step 1: Divide the target area into multiple spatial grid regions. Each grid region records disaster events in different time slices, and extract the spatio-temporal characteristics of disasters from the disaster events in different time slices. This step includes the following sub-steps:

[0096] Obtain the underground data and ground data of the target area;

[0097] Unify the coordinate system and time stamp, align the underground data and ground data spatio-temporally, and construct an underground-ground three-dimensional space;

[0098] Divide the underground-ground space into multiple three-dimensional grids; specifically: segment the target area horizontally and layer it vertically to form multiple three-dimensional grids;

[0099] Obtain the historical disaster data within each three-dimensional grid, and extract the type of disaster event, the number of occurrences of the disaster event, and the occurrence time point of the disaster event from the historical disaster data;

[0100] Calculate the disaster density within the three-dimensional grid region , where represents the th time period of the th grid region, represents the time span, represents the number of three-dimensional grids.

[0101] Step 2: Construct a disaster chain network, and generate a spatio-temporal diffusion map and risk zoning of the disaster chain based on the analysis of the spatio-temporal characteristics of disasters.

[0102] The steps for constructing the disaster chain network are as follows:

[0103] Obtain the accident data within the target area from the database, and extract the disaster-causing factors from the accident data; the accident data includes leakage, settlement, and fire; the disaster-causing factors include rising groundwater level and structural cracks;

[0104] Determine the causal relationship between the disaster-causing factors through the expert knowledge base, and construct a disaster propagation path, that is, a disaster chain;

[0105] Construct a disaster chain topology graph, that is, a disaster chain network, with the disaster-causing factors as nodes and the causal relationship between the disaster-causing factors as edges.

[0106] Determine the high-risk paths through the analysis of the disaster chain topology graph (such as the PageRank algorithm). For example: "heavy rain - rising groundwater level - softening of soil mass - ground collapse" may be a key path. Identify the key nodes in the network (such as "softening of soil mass"), and strengthening them can significantly reduce the global risk.

[0107] Among them, based on the analysis of the spatio-temporal characteristics of disasters, the steps for generating the spatio-temporal diffusion map and risk zoning of the disaster chain are as follows:

[0108] Let the three-dimensional grid of the target area be , where represents the indices of the X, Y, and Z axes of the -th three-dimensional grid;

[0109] The time slice of each three-dimensional grid is ;

[0110] Obtain the monitoring data of each three-dimensional grid and perform normalization processing; the monitoring data includes the settlement amount and the leakage rate;

[0111] Extract the spatio-temporal characteristics of disasters, including:

[0112] Obtain the disaster intensity , where represents the weight coefficient; respectively represent the disaster occurrence probability, disaster diffusion speed, and disaster duration after normalization processing;

[0113] Obtain the cumulative disaster intensity ; where represents the disaster attenuation coefficient;

[0114] If the disaster intensity of adjacent three-dimensional grids is , then diffusion is triggered: ; where represents the diffusion threshold, represents the diffusion efficiency, represents the number of intermediate grids between two three-dimensional grids;

[0115] Map the cumulative intensity of each three-dimensional grid to a color gradient, and display the risk level of the cumulative disaster intensity through colors;

[0116] Take time as the X-axis and disaster intensity as the Y-axis to display the process of the diffusion of disaster intensity;

[0117] Perform risk zoning through the clustering algorithm K-means++, including:

[0118] Obtain the risk values of all three-dimensional grids, specifically:

[0119] Take the maximum cumulative disaster intensity of each three-dimensional grid as its risk value, that is, the risk value ; represents the disaster occurrence time period within the preset time span ;

[0120] Determine the optimal number of clusters by the elbow method ;

[0121] Establish a hierarchical classification function ; Among them, represents the central value of the cluster; represents the three-dimensional space of the target area;

[0122] Set risk level labels, and the real-time level labels include high risk, medium risk and low risk;

[0123] When , divide the three-dimensional grid into high-risk partitions; when , divide the three-dimensional grid into medium-risk partitions; when , divide the three-dimensional grid into low-risk partitions;

[0124] Perform dilation or erosion operations on the clustering results to smooth the boundaries of each risk partition.

[0125] Step 3: Identify the disaster aggregation hotspots within the risk partitions, and predict the spatio-temporal information of key disasters through a time prediction model.

[0126] Among them, the steps to identify the disaster aggregation hotspots within the risk partitions are as follows:

[0127] Identify the disaster aggregation hotspots within the risk area through the local spatial autocorrelation analysis algorithm, including:

[0128] Calculate the local Moran's I index within the risk area ; Among them, represents the average disaster intensity value; represents the variance of the global disaster intensity value; represents the spatial adjacency weight; represents the number of three-dimensional grids, represents the th disaster intensity value of the adjacent three-dimensional grid;

[0129] Calculate the p-value through permutation test or Z-value. If p < 0.05, it is considered that the spatial autocorrelation is significant, and thus the areas with significant disaster aggregation within the risk area are screened out;

[0130] Calculate the kernel density of the significantly aggregated area;

[0131] ;

[0132] Among them: represents the number of disaster events, represents the Gaussian kernel function; represents the bandwidth; represents the coordinates of the center point of the area, represents the coordinates of the point to be evaluated;

[0133] If the kernel density value is greater than the density threshold, then determine that the aggregation area is a hot spot, otherwise it is a cold spot;

[0134] Among them, through the time prediction model, predicting the spatio-temporal information of key disasters includes the following steps:

[0135] Construct a spatio-temporal prediction model for key disasters, specifically:

[0136] Construct a spatio-temporal autoregressive moving average model STARMA to predict the disaster intensity at a future time point, that is:

[0137] ; among them, respectively represent the disaster intensity, spatial weight matrix, autoregressive coefficient, and moving average coefficient at the position in time the autoregressive coefficient, moving average coefficient, represents the time lag order; and respectively represent the random error term and the moving error term;

[0138] Construct a long short-term memory network LSTM and train the long short-term memory network with historical disaster data; use the trained LSTM to predict the type of disaster that will occur at the hot spot at a future time point, that is, the type of key disaster;

[0139] Use the pre-trained spatio-temporal graph convolutional network ST-GCN to predict the disaster intensity that will occur at the hot spot within a future time period, that is, the intensity of the key disaster;

[0140] Perform Kriging interpolation on the prediction results to generate a continuous risk surface to reflect the risk distribution range;

[0141] Overlay the prediction results with the hot spot area to identify potential future hot spots.

[0142] Step 4: Introduce fuzzy logic to comprehensively evaluate the disaster risk based on the spatio-temporal information of key disasters and determine the high-risk areas. It includes the following sub-steps:

[0143] Obtain the spatio-temporal information of key disasters within the hot spot area and construct key disaster characteristic variables; the spatio-temporal information of the key disasters includes the disaster duration, diffusion speed, hot spot kernel density, and tunnel deformation amount;

[0144] Construct "IF-THEN" fuzzy rules;

[0145] Based on fuzzy rules, using the Mamdani fuzzy inference algorithm, after inputting the key disaster feature variables, a fuzzy conclusion is output;

[0146] Through the MAX operation, all the output fuzzy conclusions are aggregated to form a fuzzy set;

[0147] Through the center of gravity method COG for defuzzification, the center of gravity position of the fuzzy set is calculated;

[0148] Select the conclusion corresponding to the value of the center of gravity position as the output conclusion;

[0149] Calculate the comprehensive risk value , where represents the value of the center of gravity position output by the th fuzzy rule; represents the rule weight; represents the number of fuzzy rules;

[0150] Set the high-risk threshold , if then the output hot spot is a high-risk area.

[0151] For example, when there is leakage in the subway tunnel, the three-dimensional grid A (50, 60, 2) of the subway tunnel is the hot spot area. The triangular membership function is used to determine the level of membership. The boundary of the fuzzy set is defined by the vertex of the triangle (which can be obtained based on historical data analysis).

[0152] It is calculated that the disaster intensity is 75, the membership degree is "high", and the weight is 0.8; the diffusion speed is 2.5 m / h, the membership degree is "high", and the weight is 0.7.

[0153] Fuzzy rule 1: IF hot spot area AND disaster intensity is high THEN risk is high (rule weight is 0.9);

[0154] Fuzzy rule 2: IF diffusion speed is fast THEN risk is high (rule weight is 0.7);

[0155] After defuzzification, it is calculated by the center of gravity method that , if the high-risk threshold is 0.75, and 0.82 > 0.75, then it is judged that this hot spot is a high-risk area.

[0156] Step 5. After determining the high-risk area, combined with the subway structure and evacuation capacity, conduct a secondary evaluation of the disaster risk. It includes the following sub-steps:

[0157] Obtain the list of high-risk areas, including the longitude and latitude, disaster type, and comprehensive risk value of the high-risk areas;

[0158] Obtain the subway structure data, including the length of the subway tunnel, the depth of the subway tunnel, the support material, and the support structure;

[0159] Obtain subway evacuation capacity data, including the location and quantity of subway entrances and exits, the location and quantity of emergency exits, the distance from each emergency exit to high-risk areas, and the maximum passenger capacity of the subway;

[0160] Normalize the subway structure data and subway evacuation capacity data;

[0161] Obtain the structural safety factor , where respectively represent structural strength, structural cross-sectional area, safety redundancy factor (1.2 - 1.5), and disaster acting force;

[0162] Obtain the evacuation capacity coefficient ;

[0163] Actual evacuation time , where represents the maximum number of people to be evacuated in the area; respectively represent the number of people that can pass through per second per unit width of the emergency exit, the width of a single emergency exit, and the number of available emergency exits; represents the evacuation attenuation coefficient; represents the minimum evacuation time, ;

[0164] Modify the risk formula to obtain the secondary evaluation score ' ; where represents the structural weight and evacuation weight.

[0165] For example, when there is leakage in the subway tunnel, calculated by the centroid method .

[0166] The structural parameters are that the thickness of the support structure is 0.4m, the strength of the concrete support structure is 50MPa, and the water pressure of the disaster is 200kN;

[0167] The evacuation parameters are , , , , ;

[0168] Then , minutes; ; , and this area is of low risk.

[0169] It can be seen that the original risk is high, but the structural safety is strong. Even when the evacuation capacity is insufficient (much less than 1), the comprehensive risk assessment is of low risk after secondary evaluation.

[0170] Example 2:

[0171] When constructing a disaster chain and a disaster chain network for the subway system, the spatio-temporal correlation of disasters in the ground and underground spaces needs to be considered. When studying its spatio-temporal correlation, the abstract causal relationship needs to be quantified. For this purpose, based on Example 1, we propose the following technical solutions:

[0172] Model the disaster chain network, specifically:

[0173] Construct a mechanical-hydrological coupling model based on the modified Mohr-Coulomb constitutive equation and Darcy's law to simulate soil deformation and groundwater seepage during disasters;

[0174] The mechanical-hydrological model is ; where represents the effective cohesion, represents the effective internal friction angle, represents the permeability coefficient, represents the pore water pressure; represents the water head height; represents the stress tensor;

[0175] Construct a thermal-structural coupling model based on the heat conduction equation and the thermal stress equation to simulate heat conduction during disasters;

[0176] The thermal-structural coupling model is ; where represents the heat, represents the thermal diffusivity, represents the heat gradient; represents the coefficient of thermal expansion, represents the elastic modulus; represents the thermal stress;

[0177] Establish a disaster propagation model among the nodes of the disaster chain through a Bayesian network or a dynamic system equation, specifically including:

[0178] Calculate the probability of disaster propagation between nodes based on the Bayesian network ;

[0179] Describe the evolution process of the disaster intensity over time based on a differential equation, i.e., a dynamic system equation: , where represents the disaster intensity, respectively represent the diffusion rate and the attenuation coefficient.

[0180] For example, the input of the mechanical-hydrological model: the rainfall intensity is 120 mm / h, and the rainfall duration is 6 hours, resulting in a 2 m rise in the groundwater level.

[0181] Model output: Pore water pressure is 60 kPa, predicted settlement is 12 mm.

[0182] The disaster chain is mapped as groundwater level rise - soil softening - tunnel settlement.

[0183] Thermo - structural model input: Fire temperature is 800 °C, duration is 30 minutes, resulting in a 40% decrease in concrete strength. The model output of thermal stress is 50 MPa, exceeding the tensile strength of the lining by 45 MPa. The disaster chain is fire high temperature - lining cracking - tunnel collapse.

[0184] The mechanical - hydrological and thermo - structural models provide physical mechanism support for the disaster chain, transforming abstract causal relationships into quantitative parameters (such as probability, intensity). The model output directly defines the nodes and edges of the disaster chain network. Through the topological analysis of the disaster chain network, the mapping from local physical processes to the global risk network is realized.

[0185] Meanwhile, improve the accuracy of disaster prediction, assist in formulating targeted prevention and control strategies, and adopt remedial measures (such as strengthening the drainage system, adding fire isolation layers) for key nodes (e.g., soil softening) to reduce the operation risk of the subway.

[0186] The present invention also provides an electronic device, including a processor, a memory connected to the processor, and a communication module. The electronic device is used to execute the fuzzy hierarchical evaluation method for the spatio - temporal partitioning of subway disaster chain evolution.

[0187] The present invention also provides a computer - readable storage medium. The computer - readable storage medium stores a computer program, and the computer program is executed by the processor to implement the fuzzy hierarchical evaluation method for the spatio - temporal partitioning of subway disaster chain evolution.

[0188] Embodiments disclosed by the present invention. The processes described above with reference to the flowcharts can be implemented as computer software programs. Embodiments disclosed by the present invention include a computer program product, which includes a computer program carried on a computer-readable medium. The computer program contains program codes for performing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), the above-mentioned functions defined in the methods of the present invention are executed. It should be noted that the computer-readable medium in the present invention can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. A computer-readable storage medium can, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium can include, but are not limited to: an electrical connection having one or more wire segments, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or combined with an instruction execution system, apparatus, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program codes. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, and this computer-readable medium can send, propagate, or transmit a program for use by or combined with an instruction execution system, apparatus, or device. The program codes contained on the computer-readable medium can be transmitted by any suitable medium, including but not limited to: wireless segments, wire segments, optical cables, RF, etc., or any suitable combination of the above.

[0189] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0190] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functions and structural principles of the present invention have been demonstrated and illustrated in the embodiments. Without departing from the said principles, any changes or modifications can be made to the embodiments of the present invention.

Claims

1. A fuzzy hierarchical evaluation method for the spatio-temporal zoning of the evolution of subway disaster chains, characterized in that, The method includes: Dividing the target area into multiple spatial grid areas, each grid area recording disaster events in different time slices, and extracting disaster spatio-temporal features from the disaster events in different time slices; Constructing a disaster chain network, generating a disaster chain spatio-temporal diffusion map and risk zoning based on the analysis of disaster spatio-temporal features. The construction of the disaster chain network also includes: Modeling the disaster chain network, including: Constructing a mechanical-hydrological coupling model based on the modified Mohr-Coulomb constitutive equation and Darcy's law to simulate soil deformation and groundwater seepage during disasters; The mechanical-hydrological model is ; where represents the effective cohesive force, represents the effective internal friction angle, represents the permeability coefficient, represents the pore water pressure; represents the water head height; represents the stress tensor; Constructing a thermal-structural coupling model based on the heat conduction equation and thermal stress equation to simulate heat conduction during disasters; The thermo-structural coupling model is ; where represents heat; represents the thermal diffusivity, represents the heat gradient; represents the coefficient of thermal expansion, represents the elastic modulus; represents the thermal stress; Establishing a disaster propagation model between each node of the disaster chain through a Bayesian network or a dynamic system equation, specifically including: Calculating the probability of disaster propagation between nodes based on Bayesian network ; Based on differential equations, namely dynamic system equations, to describe the evolution process of disaster intensity over time: , where represents the disaster intensity, represent the diffusion rate and the attenuation coefficient respectively; Identifying disaster aggregation hotspots within the risk zoning, and predicting the spatio-temporal information of key disasters through a time prediction model. The prediction of the spatio-temporal information of key disasters through the time prediction model includes: Constructing a spatio-temporal prediction model for key disasters, including: Constructing a spatio-temporal autoregressive moving average model STARMA to predict the disaster intensity at a future time point, that is: ; among which, respectively represent the disaster intensity, spatial weight matrix, autoregressive coefficient, and moving average coefficient at the position at time ; represents the time lag order; and respectively represent the random error term and the moving error term; Constructing a long short-term memory network LSTM and training the LSTM with historical disaster data; using the trained LSTM to predict the type of disaster occurring at the hotspot at a future time point, that is, the type of key disaster; Using a pre-trained spatio-temporal graph convolutional network ST-GCN to predict the disaster intensity of disasters occurring at the hotspot within a future time period, that is, the intensity of key disasters; Performing Kriging interpolation on the prediction results to generate a continuous risk surface to reflect the risk distribution range; Overlaying the prediction results with the hotspot area to identify potential future hotspots; Introducing fuzzy logic to comprehensively evaluate the disaster risk based on the spatio-temporal information of key disasters and determine high-risk areas; After determining the high-risk areas, combining the subway structure and evacuation capacity to conduct a secondary evaluation of the disaster risk.

2. The fuzzy hierarchical evaluation method for the spatio-temporal partitioning of the subway disaster chain evolution according to claim 1, wherein, The division of the target area into multiple spatial grid areas, each grid area recording disaster events in different time slices, and the extraction of disaster spatio-temporal features from the disaster events in different time slices include: Obtaining underground data and ground data of the target area; Unifying the coordinate system and time stamp, aligning the underground data and ground data spatio-temporally, and constructing an underground-ground three-dimensional space; Dividing the underground-ground space into multiple three-dimensional grids; specifically: segmenting the target area horizontally and layering it vertically to form multiple three-dimensional grids; Obtaining historical disaster data within each three-dimensional grid, and extracting the type of disaster event, the occurrence frequency of the disaster event, and the occurrence time point of the disaster event from the historical disaster data; Calculate the disaster density within the three-dimensional grid area , where represents the number of disasters that occurred during the time period of the th grid area, represents the time span, represents the number of three-dimensional grids.

3. The fuzzy hierarchical evaluation method for spatio-temporal zoning of subway disaster chain evolution according to claim 2, characterized in that, The construction of the disaster chain network includes: Obtaining accident data within the target area from the database, and extracting disaster-causing factors from the accident data; the accident data includes leakage, settlement, and fire; the disaster-causing factors include rising groundwater level, structural cracks; Determining the causal relationship between disaster-causing factors through an expert knowledge base, and constructing a disaster propagation path, that is, a disaster chain; Taking the disaster-causing factors as nodes and the causal relationships between the disaster-causing factors as edges, a disaster chain topology graph, i.e., a disaster chain network, is constructed.

4. The fuzzy hierarchical evaluation method for spatio-temporal zoning of subway disaster chain evolution according to claim 3, wherein, Based on the analysis of the spatio-temporal characteristics of disasters, generating a spatio-temporal diffusion map and risk zoning of the disaster chain, including: Let the three-dimensional grid of the target area be , where represents the indices of the X, Y, and Z axes of the th three-dimensional grid; The time slice of each three-dimensional grid is ; Obtaining the monitoring data of each three-dimensional grid and performing normalization processing; the monitoring data includes the settlement amount and the leakage rate. Extracting the spatio-temporal characteristics of disasters, including: Obtain the disaster intensity , where represents the weight coefficient; respectively represent the disaster occurrence probability, disaster diffusion speed, and disaster duration after normalization processing; Obtain the cumulative intensity of disasters ; where represents the disaster attenuation coefficient; If the disaster intensity of adjacent three-dimensional grids is , then diffusion is triggered: ; where represents the diffusion threshold, represents the diffusion efficiency, represents the number of spaced grids between two three-dimensional grids; Mapping the cumulative intensity of each three-dimensional grid into a color gradient, and displaying the risk level of the disaster cumulative intensity through colors. Taking time as the X-axis and the disaster intensity as the Y-axis, showing the process of the diffusion of the disaster intensity. Performing risk zoning through the clustering algorithm K-means++, including: Obtaining the risk values of all three-dimensional grids, specifically: Taking the maximum cumulative intensity of disasters for each three-dimensional grid as its risk value, i.e., the risk value ; indicating the time period when the disaster occurs within a preset time span ; Determine the optimal number of clusters by the elbow method ; Establish a grading function ; among which, represents the central value of the cluster; represents the three-dimensional space of the target area; Setting risk level labels, and the real-time level labels include high risk, medium risk, and low risk. When the three-dimensional grid is divided into high-risk zones; when the three-dimensional grid is divided into medium-risk zones; when the three-dimensional grid is divided into low-risk zones; Performing dilation or erosion operations on the clustering results to smooth the boundaries of each risk zone.

5. The fuzzy hierarchical evaluation method for spatio-temporal zoning of subway disaster chain evolution according to claim 4, characterized in that, Identifying the disaster aggregation hotspots within the risk zones, including: Identifying the disaster aggregation hotspots within the risk areas through the local spatial autocorrelation analysis algorithm, including: Calculate the local Moran's I index within the risk area ; where represents the average disaster intensity value; represents the variance of the global disaster intensity value; represents the spatial adjacency weight; represents the number of three-dimensional grids, represents the disaster intensity value of the th adjacent three-dimensional grid; Calculating the p-value through permutation test or Z-value. If p < 0.05, it is considered that the spatial autocorrelation is significant, and thus the areas with significant disaster aggregation within the risk areas are screened out. Calculating the kernel density of the significant aggregation areas. ; wherein, represents the number of disaster events, represents the Gaussian kernel function; represents the bandwidth; represents the coordinates of the center point of the area, represents the coordinates of the point to be evaluated; If the kernel density value is greater than the density threshold, then the aggregation area is judged as a hotspot, otherwise it is a cold spot.

6. The fuzzy hierarchical evaluation method for spatio-temporal zoning of subway disaster chain evolution according to claim 5, wherein, Introducing fuzzy logic to comprehensively evaluate the disaster risk based on the spatio-temporal information of key disasters and determining the high-risk areas, including: Obtaining the spatio-temporal information of key disasters within the hotspot areas and constructing characteristic variables of key disasters; the spatio-temporal information of the key disasters includes the disaster duration, diffusion speed, hotspot kernel density, and tunnel deformation amount. Constructing "IF-THEN" fuzzy rules. Based on the fuzzy rules, using the Mamdani fuzzy inference algorithm to output fuzzy conclusions after inputting the characteristic variables of key disasters. Aggregating all the output fuzzy conclusions through the MAX operation to form a fuzzy set. Defuzzifying through the center of gravity method COG to calculate the center of gravity position of the fuzzy set. Selecting the conclusion corresponding to the value of the center of gravity position as the output conclusion. Calculate the comprehensive risk value , where represents the value of the center of gravity of the output of the fuzzy rule; represents the rule weight; represents the number of fuzzy rules; Set a high-risk threshold If then the output hotspot is a high-risk area.

7. The fuzzy hierarchical evaluation method for spatio-temporal zoning of subway disaster chain evolution according to claim 6, characterized in that After determining the high-risk areas, combining the subway structure and evacuation capacity to conduct a secondary evaluation of the disaster risk, including: Obtaining a list of high-risk areas, including the longitude and latitude of high-risk areas, disaster types, and comprehensive risk values. Obtaining subway structure data, including the length of subway tunnels, the depth of subway tunnels, support materials, and support structures. Obtaining subway evacuation capacity data, including the location and number of subway entrances and exits, the location and number of emergency exits, the distance from each emergency exit to the high-risk areas, and the maximum passenger flow capacity of the subway. Performing normalization processing on the subway structure data and subway evacuation capacity data. Obtain the structural safety factor , where respectively represent the structural strength, the structural cross-sectional area, the safety redundancy factor, and the disaster acting force; Obtain the dredging capacity coefficient ; Actual evacuation time , where represents the maximum number of people to be evacuated in the area; represents the evacuation distance, i.e., the distance from the platform to the nearest emergency exit; respectively represent the walking speed, the width of a single emergency exit, and the number of available emergency exits; represents the evacuation attenuation coefficient; represents the minimum evacuation time, ; Obtain the secondary evaluation score by modifying the risk formula ; among which, represents the structure weight and the guidance weight.

8. An electronic device, comprising a processor, a memory connected to the processor, and a communication module, characterized in that, The electronic device is used to execute the fuzzy hierarchical evaluation method for the spatio-temporal zoning of subway disaster chain evolution described in any one of claims 1-7 above.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the fuzzy hierarchical evaluation method for the spatio-temporal zoning of subway disaster chain evolution described in any one of claims 1-7 above.

Images