Intelligent early warning system for urban fire

Abstract

The invention relates to the technical field of intelligent fire early warning, in particular to an urban fire intelligent early warning system which comprises an urban structure intelligent analysis module, a fire dynamic early warning simulation module, a fire prevention strategy optimization module and an environment influence intelligent analysis module. According to the method, the urban graph model is generated by using the geographic positions, materials, fire data and building age data of the urban buildings, the relevance and characteristics between the buildings are deeply analyzed, and the fire diffusion influence is simulated through the graph convolutional network, so that the prediction precision of the fire propagation path and speed is remarkably improved; the evacuation difficulty and risk level of a building under multiple fire source points are calculated in an auxiliary manner, a basis is provided for scientific planning of evacuation routes and fireproof isolation areas, emergency response and resource allocation are optimized, meteorological, population dynamics and holiday activity data analysis are integrated, the effect of environmental factors in fire early warning is enhanced, the accuracy and timeliness of early warning are ensured, and the early warning efficiency is improved. And the overall efficiency of an urban fire prevention system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent fire early warning, and in particular to an intelligent urban fire early warning system. Background Art

[0002] The field of intelligent fire warning technology involves the development and application of advanced algorithms and sensor systems to detect fires early and quickly notify relevant personnel or automatically initiate response measures. This field combines fire science, Internet of Things technology, machine learning and big data analysis. The purpose is to improve the speed and accuracy of fire detection by analyzing environmental data, smoke composition, temperature changes and other information. The system includes a variety of sensors, such as smoke sensors, temperature sensors and flame monitors. The sensors can generate real-time data for fire analysis and warning. The intelligent fire warning system also integrates mobile communications and cloud computing technologies to achieve cross-regional data sharing and coordinated response.

[0003] Among them, the urban fire intelligent early warning system involves the deployment of a comprehensive intelligent system in the urban environment. The system monitors environmental data and potential fire sources at multiple key points in the city in real time to warn residents and fire departments of fire risks in advance. The system uses advanced algorithms to analyze data and identify early signs of fire, such as abnormal temperature rise or increased smoke. The purpose of the system is to enhance the city's emergency response capabilities, reduce casualties and property losses caused by fires, and optimize the allocation and application of firefighting resources through intelligent data processing and response mechanisms.

[0004] Although existing intelligent fire warning technologies can monitor environmental data in real time and detect signs of fire through sensors, the technology focuses on monitoring a single building or a small area, and fails to fully refer to the complex interactions between buildings and environmental factors at the urban scale, resulting in blind spots in fire prevention and response strategies at the city level. For example, data processing in existing systems is limited to information obtained directly from sensors, and lacks in-depth analysis of connectivity between buildings and the overall urban environmental dynamics, which limits the effectiveness of warnings in large-scale fire scenarios. Existing technologies also show deficiencies in assessing the impact of changes in environmental factors on fires, such as failing to effectively predict the impact of changes in traffic flow during holidays or special events on fire risks, which leads to irrational allocation of fire prevention resources and increases the overall risk of the system. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an intelligent urban fire early warning system.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: The urban fire intelligent early warning system comprises:

[0007] The urban structure intelligent analysis module allocates nodes to buildings based on the geographical location, materials, fire data and building age data of urban buildings, takes the channel type and distance between buildings as the attributes of the edges, generates an urban graph model, and generates a node and edge feature matrix according to the node and edge attributes of the urban graph model;

[0008] The fire dynamic warning simulation module inputs the node and edge feature matrix into the graph convolutional network, analyzes the impact of building characteristics and interconnectivity on fire spread, obtains a fire spread influence score, and uses the fire spread influence score to analyze the propagation path and speed of differentiated fire source points to generate a fire spread prediction model;

[0009] The fire prevention strategy optimization module uses the fire spread prediction model to calculate the evacuation difficulty and fire risk level of each building under differentiated fire source points, generates a building risk level assessment result, optimizes the evacuation route and fire isolation zone according to the building risk level assessment result, and generates a fire isolation zone planning map;

[0010] The environmental impact intelligent analysis module captures meteorological condition data, urban population dynamics and holiday activity data based on the fire isolation zone planning map, analyzes the correlation between environmental factors and fire occurrence, obtains the environmental factor impact score, adjusts the fire spread prediction model based on the environmental factor impact score, and constructs an adjusted fire risk warning plan.

[0011] As a further solution of the present invention, the steps of obtaining the node and edge feature matrices are specifically as follows:

[0012] Based on the data of urban building location, materials, fire data and building age, key data are extracted from geographic information and building databases and organized into a unified format to obtain a set of urban building attributes;

[0013] Using the urban building attribute set, a graph node is assigned to each building, and the channel information between buildings is encoded as a graph edge, and a graph structure model reflecting the real-time urban layout is constructed using a graph theory method to obtain a city layout map;

[0014] Based on the city layout graph, the attribute information of nodes and edges is extracted using the formula:

[0015] M F =P N ·k N +P E ·k E

[0016] Perform attribute weighting and merging, calculate the integrated feature scores of nodes and edges, and obtain the node and edge feature matrix, where P N Represents the attribute set of the node, P Erepresents the attribute set of the edge, k N and k E are the attribute weight coefficients of nodes and edges, M F Represents the integrated feature scores of nodes and edges.

[0017] As a further solution of the present invention, the steps for obtaining the fire propagation influence score are specifically as follows:

[0018] Input the node and edge feature matrix into the graph convolutional network to encode and convert the features, optimize the nonlinear representation of the data through iterative processing of the network layer, and generate feature data after graph convolution processing;

[0019] Analyze the feature data after the graph convolution processing, using the formula:

[0020] I fire,b =∑(k b ·F node,b )+∑(j b ·F edge,b )

[0021] The characteristics of nodes and edges are weighted and summarized to evaluate the potential contribution of each building and connection to fire spread and obtain the fire impact score, where I fire,b represents the fire impact score, k b and j b is the weight coefficient of node and edge characteristics, F node,b and F edge,b Represent the feature vectors of nodes and edges respectively;

[0022] According to the fire impact score, the fire spread potential of multiple buildings in the entire urban network is integrated to generate a fire spread impact score.

[0023] As a further solution of the present invention, the steps of obtaining the fire spread prediction model are specifically as follows:

[0024] Integrate the fire propagation influence score, simulate differentiated fire source points, simulate the fire spread path through graph theory algorithm, and generate the fire source point spread path data;

[0025] By analyzing the diffusion path data of the fire source point, the formula is applied:

[0026]

[0027] Calculate the fire spread speed of the differentiated path and obtain speed evaluation data, where V c Indicates the speed of fire spread, d c is the total length of the path, a c is the acceleration factor, L c is the complexity index of the path, tc is the predicted diffusion time, ∈ c is the time error correction factor;

[0028] Combining the speed assessment data with the fire spread influence score, the model is optimized through multivariate regression analysis, the data is integrated, and the fire spread trend of the network is predicted to generate a fire spread prediction model.

[0029] As a further solution of the present invention, the steps for obtaining the building risk level assessment result are specifically as follows:

[0030] Based on the fire spread prediction model, the location of the differentiated fire source and the associated building characteristics are input, the evacuation difficulty of each building is calculated, and preliminary evacuation difficulty data is obtained;

[0031] Based on the preliminary evacuation difficulty data, combined with the original safety record and structural characteristics of each building, the fire risk level is calculated to obtain a preliminary fire risk score;

[0032] Combining the above preliminary fire risk score and the difficulty of evacuation for each building, the risk level is calculated and assessed, applying the formula:

[0033] F d =α d ·R d +β d ·S d

[0034] The building risk level assessment results are obtained, where R d represents the preliminary fire risk score, S d represents the structural parameters of the building, α d and β d is the weight parameter adjusted based on the original data, F d Indicates the building risk level assessment result.

[0035] As a further solution of the present invention, the steps for obtaining the fire isolation zone planning map are specifically as follows:

[0036] Evaluate the risk level assessment results of the building and the surrounding environment, capture data on the structural safety of the building and previous evacuation routes, identify potential fire isolation areas, and generate an overview of preliminary fire isolation needs;

[0037] Based on the above preliminary overview of fire separation requirements, refine the layout and evacuation requirements of the building and apply the formula:

[0038]

[0039] Construct the layout of fire isolation zones, where A hrepresents the recommended area of ​​the fire separation zone, β h is the safety parameter, S h is the building area, γ h is the environmental adjustment factor;

[0040] Combining the fire separation zone layout with real-time building occupancy, apply the formula:

[0041] P h =δ h ·A h +∈ h ·A h

[0042] Generate a fire isolation zone planning map, where P h Represents the effectiveness of the quarantine area planning map, δ h and ∈ h are adjustment coefficient and risk assessment coefficient, A h Indicates the recommended area of ​​the fire separation zone.

[0043] As a further solution of the present invention, the steps for obtaining the environmental factor impact score are specifically as follows:

[0044] According to the fire isolation zone planning map, meteorological condition data, urban population dynamics and holiday activity data are captured, and integrated into a single environmental data set through data processing to generate a processed environmental data set;

[0045] The processed environmental data set is statistically analyzed using the formula:

[0046]

[0047] Calculate the contribution of multiple data factors to fire risk and generate statistical analysis results, where w 1g 、w 2g 、w 3g Represents the weight of the differentiated data factor, D 1g , D 2g , D 3g Represents the corresponding data value, N g Represents the total number of data points, Y g Indicates the results of statistical analysis;

[0048] Using the statistical analysis results, combined with the original fire records and trend analysis of the region, the scoring model is used to optimize the formula:

[0049]

[0050] Get the environmental factor impact score, where k 1g , k 2g is the adjustment coefficient, Sg Indicates the environmental factor impact score, Y g Represents the statistical analysis results.

[0051] As a further solution of the present invention, the steps for obtaining the adjusted fire risk warning scheme are specifically as follows:

[0052] Comparing the environmental factor impact score with the original fire data set, sorting the environmental factor data set, screening variables associated with fire frequency, processing the original fire data, determining the correlation between the environmental score and the fire frequency through comparative analysis, and generating correlation analysis results;

[0053] Combining the correlation analysis results with the current meteorological data, the formula is adopted:

[0054] F v =k 1v ·R+k 2v exp(M)

[0055] Adjust the fire spread prediction model parameters to generate adjusted model parameters, where k 1v and k 2v They are adjustment coefficients, M represents the real-time meteorological data impact coefficient, F v represents the adjusted model parameters;

[0056] The adjusted model parameters are used to reassess the fire risk in the area, and an adjusted fire risk warning plan is generated based on the model prediction results.

[0057] Compared with the prior art, the advantages and positive effects of the present invention are:

[0058] In the present invention, by generating a city graph model based on the geographical location, materials, fire data and building age of urban buildings, the correlation and characteristics between buildings are deeply analyzed, and the dynamic impact of fire spread is further simulated using a graph convolutional network, effectively improving the prediction accuracy of the fire propagation path and speed, and using an algorithm to calculate the evacuation difficulty and risk level of each building under differentiated fire sources, which helps to formulate more scientific evacuation routes and fire isolation zones, greatly optimizing emergency response strategies and resource allocation. Comprehensive analysis of environmental factors such as meteorological conditions, urban population dynamics and holiday activity data improves the contribution of environmental factors to fire warning, making fire risk warning more accurate and timely, and improving the overall effectiveness of urban fire prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a system flow chart of the present invention;

[0060] Figure 2 It is a flow chart of node and edge feature matrix in the present invention;

[0061] Figure 3 A flow chart for scoring the fire propagation influence in the present invention;

[0062] Figure 4 It is a flow chart of the fire spread prediction model in the present invention;

[0063] Figure 5 It is a flow chart of the building risk level assessment results in the present invention;

[0064] Figure 6 It is a flow chart of the fire isolation zone planning diagram in the present invention;

[0065] Figure 7 A flow chart for scoring the impact of environmental factors in the present invention;

[0066] Figure 8 It is a flow chart of the fire risk warning scheme after adjustment in the present invention. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0068] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0069] See also Figure 1 , the urban fire intelligent early warning system includes:

[0070] The urban structure intelligent analysis module allocates nodes to buildings based on the geographical location, materials, fire data and building age data of urban buildings, takes the channel type and distance between buildings as the attributes of the edges, generates an urban graph model, and generates node and edge feature matrices based on the node and edge attributes of the urban graph model;

[0071] The fire dynamic warning simulation module inputs the node and edge feature matrix into the graph convolutional network, analyzes the impact of building characteristics and interconnectivity on fire spread, obtains the fire spread influence score, and uses the fire spread influence score to analyze the propagation path and speed of differentiated fire source points to generate a fire spread prediction model;

[0072] The fire prevention strategy optimization module uses the fire spread prediction model to calculate the evacuation difficulty and fire risk level of each building under different fire source points, generate building risk level assessment results, optimize evacuation routes and fire isolation zones based on the building risk level assessment results, and generate fire isolation zone planning maps;

[0073] The intelligent environmental impact analysis module captures meteorological conditions data, urban population dynamics and holiday activity data based on the fire isolation zone planning map, analyzes the correlation between environmental factors and fire occurrence, obtains the environmental factor impact score, adjusts the fire spread prediction model based on the environmental factor impact score, and constructs an adjusted fire risk warning plan.

[0074] The urban graph model includes node type, node distribution, and edge connection mode. The node and edge feature matrix includes node data set, edge data set, and feature encoding. The fire spread influence score includes structural sensitivity, connection responsiveness, and propagation probability. The fire spread prediction model includes a spread speed model and an impact range prediction. The building risk level assessment results include fire susceptibility, evacuation complexity, and risk classification. The fire isolation zone planning map includes isolation safety zones, priority evacuation routes, and safety exit layouts. The environmental factor impact score includes a climate change score, a population density score, and an activity intensity score. The adjusted fire risk warning plan includes warning level updates, prevention strategy adjustments, and emergency response plans.

[0075] See also Figure 2 , the steps to obtain the node and edge feature matrices are as follows:

[0076] Based on the data of urban building location, materials, fire data and building age, key data are extracted from geographic information and building databases and organized into a unified format to obtain a set of urban building attributes;

[0077] Using the urban building attribute set, we assign graph nodes to each building, encode the channel information between buildings as graph edges, and use graph theory methods to construct a graph structure model that reflects the real-time urban layout to obtain the urban layout map.

[0078] Based on the city layout graph, extract the attribute information of nodes and edges using the formula:

[0079] M F =P N ·k N +P E ·kE

[0080] Perform attribute weighting and merging, calculate the integrated feature scores of nodes and edges, and obtain the node and edge feature matrix, where P N Represents the attribute set of the node, P E represents the attribute set of the edge, k N and k E are the attribute weight coefficients of nodes and edges, M F Represents the integrated feature scores of nodes and edges.

[0081] Integrate the geographic coordinates and construction material data of urban buildings, extracted from geographic information (GIS), according to the formula:

[0082]

[0083] Calculate the straight-line distance between two locations, where x1 and y1 represent the longitude and latitude of the first location, x2 and y2 represent the longitude and latitude of the second location, and Da represents the straight-line distance between the two locations;

[0084] This formula is the Euclidean distance formula used to calculate the straight-line distance between two points on Earth. For example, if there are two locations, one at longitude 100 degrees and latitude 20 degrees (x1=100, y1=20), and the other at longitude 101 degrees and latitude 21 degrees (x2=101, y2=21), the straight-line distance between them can be calculated by plugging the values ​​into the formula:

[0085]

[0086] This result indicates that there is approximately 1.414 units of straight-line distance between the two locations.

[0087] Using the city building attribute set, assign graph nodes to each building, and encode the channel information between buildings as graph edges according to the formula:

[0088] Ca=n a ·(n a -1) / 2

[0089] Calculate the maximum number of connections, where n a represents the total number of nodes in the graph, and Ca represents the maximum number of connections;

[0090] This formula is used to calculate the maximum number of edges in a complete graph, that is, there is an edge between any two nodes in the graph. Assuming there are 5 buildings in the city, the total number of nodes n a =5, substitute it into the formula to calculate the maximum number of connections:

[0091] Ca=5·(5-1) / 2=5·4 / 2=10

[0092] This result shows that if every building is directly connected to other buildings, there can be a total of 10 differentiated connection methods, which helps to explain the complexity of the network in a fully connected scenario.

[0093] Extract the attribute information of nodes and edges from the city layout graph using the formula:

[0094] M F =P N ·k N +P E ·k E

[0095] The attributes are weighted and combined to calculate an integrated feature score for each node and edge.

[0096] In this formula, P N and P E represent the attribute vectors of nodes and edges respectively, and k N and k E is the corresponding weight coefficient. Suppose a node has an attribute value of 10, the corresponding weight is 0.4, and the edge attribute value is 20, the corresponding weight is 0.6. Insert the value for calculation:

[0097] M F =10·0.4+20·0.6=4+12=16

[0098] The calculation result 16 represents the integrated feature score of the node and edge under a given weight. This score helps quantify the importance or influence of each element in the graph, and is then used for further network analysis or optimization tasks.

[0099] See also Figure 3 , the steps to obtain the fire spread influence score are as follows:

[0100] Input node and edge feature matrices into the graph convolutional network to encode and transform features, optimize the nonlinear representation of data through iterative processing of the network layer, and generate feature data after graph convolution processing;

[0101] Analyze the feature data after graph convolution processing, using the formula:

[0102] I fire,b =∑(k b ·F node,b )+∑(j b ·F edge,b )

[0103] The characteristics of nodes and edges are weighted and summarized to evaluate the potential contribution of each building and connection to fire spread and obtain the fire impact score, where I fire,b represents the fire impact score, kb and j b is the weight coefficient of node and edge characteristics, F node,b and F edge,b Represent the feature vectors of nodes and edges respectively;

[0104] According to the fire impact score, the fire spread potential of multiple buildings in the entire urban network is integrated to generate a fire spread impact score.

[0105] Input node and edge feature matrices into the graph convolutional network to encode and transform features, optimize the nonlinear representation of data through iterative processing of the network layer, and generate feature data after graph convolution processing;

[0106] Analyze the feature data after graph convolution processing, using the formula:

[0107] I fire,b =∑(k b ·F node,b )+∑(j b ·F edge,b )

[0108] The characteristics of nodes and edges are weighted and summarized to evaluate the potential contribution of each building and connection to fire spread and obtain the fire impact score, where I fire,b represents the fire impact score, k b and j b is the weight coefficient of node and edge characteristics, F node,b and F edge,b Represent the feature vectors of nodes and edges respectively;

[0109] According to the fire impact score, the fire spread potential of multiple buildings in the entire urban network is integrated to generate a fire spread impact score.

[0110] Input the feature matrix into the graph convolutional network according to the formula:

[0111] E b =W b ·X b +b b

[0112] Calculate the encoding and transformation of features, where E b Represents the encoded feature output, W b is the weight matrix, X b is the input feature matrix, b b is the bias term;

[0113] This formula represents the operation of each layer in the graph convolutional network. The input feature matrix is ​​multiplied by the weight matrix and then the bias term is added to generate a new feature representation. Set the input feature matrix X bA sample of is [1, 2, 3], the weight matrix W b is [[1, 0], [0, 1], [1, 1]], bias b b is [1, 1], the calculation process is as follows:

[0114] E b =[[1,0],[0,1],[1,1]]·[1,2,3]+[1,1]

[0115] =[1·1+0·2+1·3,0·1+1·2+1·3]+[1,1]

[0116] =[4,6]+[1,1]=[5,7]

[0117] This result shows that the new feature representation is [5, 7], which represents the output after the input features are transformed through the graph convolutional network layer.

[0118] Analyze the feature data after graph convolution, according to the formula:

[0119] I fire,b =∑(k b ·F node,b )+∑(j b ·F edge,b )

[0120] Calculate the fire influence of each node and edge, where I fire,b represents the fire spread impact score, k b and j b is the weight coefficient of node and edge characteristics, F node,b and F edge,b Represent the feature data of nodes and edges respectively;

[0121] This formula is used to evaluate the influence on fire propagation based on the characteristics of nodes and edges. Assuming that a node eigenvalue is 10 and its weight is 0.5, and an edge eigenvalue is 20 and its weight is 0.3, the calculation process is as follows:

[0122] I fire,b =∑(0.5·10)+∑(0.3·20)=5+6=11

[0123] This calculation result 11 means that the total influence score of the node and edge on fire spread combined with their weights is 11, which helps to identify the critical paths and critical buildings for fire spread.

[0124] Fire spread impact scoring formula:

[0125] S final,b =α b I fire,b

[0126] In this formula, S final,b represents the fire spread influence score, α b is the influence rating coefficient adjusted according to the city layout and building materials. fire,b is the fire impact score calculated previously, and set the known I fire,b is 11, adjustment coefficient α b is 1.1, and the calculation process is as follows:

[0127] S final,b =1.1·11=12.1

[0128] The results show that after referring to the urban layout and building material characteristics, the adjusted fire propagation influence score is 12.1, which helps to ultimately determine the criticality and priority of buildings in fire simulations.

[0129] See also Figure 4 , the specific steps for obtaining the fire spread prediction model are:

[0130] Integrate the fire spread influence score, simulate the differentiated fire source points, simulate the fire spread path through graph theory algorithm, and generate the fire source point diffusion path data;

[0131] By analyzing the diffusion path data of the fire source point, the formula is applied:

[0132]

[0133] Calculate the fire spread speed of the differentiated path and obtain speed evaluation data, where V c Indicates the speed of fire spread, d c is the total length of the path, a c is the acceleration factor, L c is the complexity index of the path, t c is the predicted diffusion time, ∈ c is the time error correction factor;

[0134] Combining the speed assessment data with the fire spread influence score, the model is optimized through multivariate regression analysis, the data is integrated, and the fire spread trend of the network is predicted to generate a fire spread prediction model.

[0135] Simulate the fire spread path according to the formula:

[0136] P c =G(V c ,E c )

[0137] Calculate the diffusion simulation of the differentiation starting point, where P c represents the path simulation result, G represents the graph theory algorithm, Vc and E c Represent the characteristic data of the vertices and edges in the graph respectively;

[0138] This formula is used to simulate the fire spread path of differentiated fire source points. Graph theory algorithms are used to analyze the interactions between nodes (vertices) and connections (edges) to predict the fire spread path. If there are 5 vertices and 4 edges in a specific example, the vertices represent differentiated building locations and the edges represent the connection channels between buildings. The algorithm will calculate the diffusion route starting from the specified starting point.

[0139] Fire spread rate formula:

[0140]

[0141] This formula calculates a speed score for each simulated fire spread path, taking into account the path length, path complexity and error correction of the prediction time. For example, if the total length of a path is 500 meters, the acceleration factor is 0.05, the path complexity index is 10, the prediction time is 50 minutes, and the time error correction factor is 2 minutes, the calculation is as follows:

[0142]

[0143] This calculated result of 9.62 m / min represents the fire spread rate along this path, providing a quantitative speed assessment that helps further risk management and emergency response strategy formulation.

[0144] The fire spread prediction model ensures that each step from data input to model output closely corresponds to the final output fire spread prediction model, avoiding any redundant operations, and using multivariate regression analysis to integrate the fire spread speed and influence score according to the formula:

[0145] M c =β c ·V c +γ c ·S final,b

[0146] Calculate the fire spread prediction of the entire network, where M c represents the fire spread prediction model, β c and γ c is the regression coefficient, V c is the fire spread speed, S final,b is the fire spread impact score;

[0147] This formula combines the fire spread rate and fire impact score to predict the fire spread pattern in the entire city network through regression analysis. For example, if the fire spread rate score is 9.62 and the impact score is 200, the regression coefficient βc is 0.3, γ c is 0.7, and the calculation process is as follows:

[0148] M c =0.3·9.62+0.7·200=2.886+140=142.886

[0149] This calculated result of 142.886 represents the predicted fire spread pattern score throughout the urban network, which is used to guide the optimization of emergency management and fire prevention measures.

[0150] See also Figure 5 , the specific steps for obtaining the building risk level assessment results are as follows:

[0151] Based on the fire spread prediction model, the location of the differentiated fire source and the characteristics of the associated buildings are input to calculate the evacuation difficulty of each building and obtain preliminary evacuation difficulty data;

[0152] Based on the preliminary evacuation difficulty data, combined with the original safety record and structural characteristics of each building, the fire risk level is calculated to obtain a preliminary fire risk score;

[0153] Combining the preliminary fire risk score and the evacuation difficulty of each building, the risk level is calculated and assessed, applying the formula:

[0154] F d =α d ·R d +β d ·S d

[0155] The building risk level assessment results are obtained, where R d represents the preliminary fire risk score, S d represents the structural parameters of the building, α d and β d is the weight parameter adjusted based on the original data, F d Indicates the building risk level assessment result.

[0156] According to the correlation data of fire source and building characteristics, the evacuation difficulty index of each building is calculated. This calculation depends on the basic output of the model according to the formula:

[0157] D d =f(S d ,F d )

[0158] Among them, S d Represents the structural parameters of the building, F d Represents the influence of fire;

[0159] Dd It represents the evacuation difficulty index, which is based on the building structure parameter S d and fire source influence F d Calculated by function f, for example, if the building structure parameter S d is 5, the fire source influence F d If is 2, then set f to be the weighted sum:

[0160] D d =0.6×S d +0.4×F d =0.6×5+0.4×2=3+0.8=3.8

[0161] This result shows that the evacuation difficulty score of the building under the influence of the target fire source point is 3.8.

[0162] Based on the evacuation difficulty index, combined with the historical fire data of the building H d , calculate the preliminary fire risk score R of the building d , according to the formula:

[0163]

[0164] Perform calculations to produce a preliminary fire risk score for each building;

[0165] Formula R d represents the preliminary fire risk score and is the evacuation difficulty index D d With the historical fire data of the building H d The mean value of D d =3.8, historical fire data of buildings H d If is 4, the calculation process is as follows:

[0166]

[0167] The results showed that the building’s preliminary fire risk score was 3.9, which was used for further analysis and evaluation.

[0168] Integrate the preliminary fire risk score R of the building in the previous step d With safety parameter S d , calculate the building risk level assessment result, the formula is expressed as:

[0169] F d =α d ·R d +β d ·S d

[0170] Among them, α d and β d is the adjustment weight parameter, Fd Represents the building risk level assessment result, which is composed of the preliminary fire risk score R d and the structural parameters S of the building d After comprehensive consideration, the weight parameter α is set d = 0.7 and β d =0.3, R d =3.9, the structural parameter of the building S d =2, the calculation process is as follows:

[0171] F d =0.7·3.9+0.3·2=2.73+0.6=3.33

[0172] This result of 3.33 indicates that after referring to the safety parameters, the risk level assessment result of the building is 3.33, which can be used to formulate specific fire prevention strategies and emergency preparedness measures.

[0173] See also Figure 6 , the specific steps for obtaining the fire isolation zone planning map are:

[0174] Evaluate the building risk level assessment results and surrounding environment, capture data on the building's structural safety and previous evacuation routes, identify potential fire isolation areas, and generate an overview of preliminary fire isolation needs;

[0175] Based on the preliminary overview of fire separation requirements, refine the layout and evacuation requirements of the building and apply the formula:

[0176]

[0177] Construct the layout of fire isolation zones, where A h represents the recommended area of ​​the fire separation zone, β h is the safety parameter, S h is the building area, γ h is the environmental adjustment factor;

[0178] Combining the fire separation zone layout with real-time building occupancy, apply the formula:

[0179] P h =δ h ·A h +∈ h ·A h

[0180] Generate a fire isolation zone planning map, where P h Represents the effectiveness of the quarantine area planning map, δ h and ∈ h are adjustment coefficient and risk assessment coefficient, A h Indicates the recommended area of ​​the fire separation zone.

[0181] The recommended area formula for fire separation zones is:

[0182]

[0183] A h is the recommended area of ​​the fire isolation zone, in square meters, S h Obtained through field measurement, for example, the actual floor area of ​​a building is 5,000 square meters;

[0184] β h With γ h The safety parameter and environmental adjustment factor, obtained through historical data analysis and risk assessment model calculation, were set to 1.2 and 0.8, based on previous fire safety planning cases in similar areas;

[0185] Set S h =5000 square meters, β h =1.2,γ h =0.8;

[0186] Calculate A h :

[0187]

[0188] This calculation indicated that the recommended fire isolation zone area was 7,500 square meters, which was based on the actual floor space of the building and the zone safety parameters to ensure that there was enough space for effective isolation in the event of a fire.

[0189] Validity formula of quarantine area planning map:

[0190] P h =δ h ·A h +∈ h ·A h

[0191] P h is the effectiveness score of the final plan, A h is the recommended area of ​​the fire separation zone obtained from the previous step, which is 7500 square meters;

[0192] δ h and ∈ h It is the adjustment factor and risk assessment factor, which are set based on expert consultation and safety regulations, for example, set to 0.6 and 0.4 to ensure a reasonable balance between area and risk;

[0193] Setting δ h =0.6,∈ h =0.4, R h =7500;

[0194] Calculate P h :

[0195] P h =0.6·7500+0.4·7500=4500+3000=7500

[0196] The results show that with comprehensive reference to the fire isolation zone area and regional risk assessment, the final fire isolation zone planning effectiveness score is 7500, reflecting the overall quality and safety of the plan.

[0197] See also Figure 7 , the specific steps for obtaining the environmental factor impact score are:

[0198] According to the fire isolation zone planning map, meteorological condition data, urban population dynamics and holiday activity data are captured and integrated into a single environmental data set through data processing to generate a processed environmental data set;

[0199] The processed environmental data set was statistically analyzed using the formula:

[0200]

[0201] Calculate the contribution of multiple data factors to fire risk and generate statistical analysis results, where w 1g 、w 2g 、w 3g Represents the weight of the differentiated data factor, D 1g , D 2g , D 3g Represents the corresponding data value, N g Represents the total number of data points, Y g Indicates the results of statistical analysis;

[0202] Using the statistical analysis results, combined with the original fire records and trend analysis of the region, the scoring model is used to optimize the formula:

[0203]

[0204] Get the environmental factor impact score, where k 1g , k 2g is the adjustment coefficient, S g Indicates the environmental factor impact score, Y g Represents the statistical analysis results.

[0205] Integrate meteorological conditions, urban population dynamics and holiday activity data into an environmental data set using the formula:

[0206]

[0207] Get the processed environmental data set, where v gi Represents the importance coefficient of the i-th data, d gi represents the actual observed value of the i-th data, and n represents the total number of data points;

[0208] In actual situations, the importance coefficients of meteorological data, population dynamics, and holiday activity data are assumed to be 0.3, 0.5, and 0.2, respectively. The meteorological data observation value is assumed to be 70, the population dynamics observation value is assumed to be 50, and the holiday activity observation value is assumed to be 30. The calculation formula is as follows:

[0209] M g =(0.3×70)+(0.5×50)+(0.2×30)=21+25+6=52

[0210] This calculation process showed that the score of the preliminary environmental model was 52, which reflected the comprehensive environmental status and provided a basic data model for risk association analysis in the next step.

[0211] Statistical analysis result formula:

[0212]

[0213] This formula is used to calculate the contribution of each environmental factor to the fire risk, where w 1g 、w 2g 、w 3g They represent the weight coefficients of meteorological conditions, population dynamics, and holiday activities, respectively. 1g , D 2g , D 3g Indicates the corresponding data value, N g is the total number of data points, setting the meteorological weight to 0.4, population dynamics to 0.4, holiday activities to 0.2, the total number of data points to 3, and the actual data values ​​to 70, 50, and 30 respectively. The calculation is as follows:

[0214]

[0215] The results show that under given weights and data conditions, the average contribution of environmental factors to fire risk is 18, and this value will be used to further evaluate the impact score of environmental factors.

[0216] The formula for environmental factors to influence the score is:

[0217]

[0218] This formula is used to calculate the final environmental factor impact score S g , where k 1g and k 2g are the adjustment coefficients, Yg For statistical analysis results, set the adjustment coefficient k 1g =0.5, k 2g =0.3, and from the above calculation, Y g =18, the calculation is as follows:

[0219] S g =0.5·log(18)+0.3·18 2

[0220] S g =0.5·2.89+0.3·324

[0221] S g =1.445+97.2=98.645

[0222] The results show that the final environmental factor impact score is 98.645, with comprehensive reference to the adjustment coefficient and environmental contribution. This score reflects the overall impact of environmental factors on fire risk and is used for further risk management and prevention and control decisions.

[0223] See also Figure 8 , the specific steps for obtaining the adjusted fire risk warning plan are:

[0224] Compare the environmental factor impact scores with the original fire data set, organize the environmental factor data set, screen the variables associated with the fire frequency, process the original fire data, determine the correlation between the environmental score and the fire frequency through comparative analysis, and generate correlation analysis results;

[0225] Combining the correlation analysis results with the current meteorological data, the formula is adopted:

[0226] F v =k 1v ·R+k 2v exp(M)

[0227] Adjust the fire spread prediction model parameters to generate adjusted model parameters, where k 1v and k 2v They are adjustment coefficients, M represents the real-time meteorological data impact coefficient, F v represents the adjusted model parameters;

[0228] The fire risk in the area is reassessed using the adjusted model parameters, and an adjusted fire risk warning plan is generated based on the model prediction results.

[0229] Environmental factor impact scores and historical fire datasets were used for preliminary comparisons according to the formula:

[0230]

[0231] Calculate the correlation between environmental score and fire frequency, where S v Represents the environmental factor impact score, F v represents the historical fire incidence rate, N v Represents the total number of events in the dataset;

[0232] The reference environmental impact score is 30 points. The fire frequencies of the five records in the historical fire data set are 10%, 20%, 15%, 25% and 30%, which are converted into numerical values ​​of 0.1, 0.2, 0.15, 0.25 and 0.3. The set values ​​are based on actual monitoring and records. The R is calculated. v :

[0233]

[0234] The results show that the average correlation score obtained by referring to the combined impact of the environmental score and historical fire data is 6, which indicates that there is a certain correlation between the environmental score and the fire incidence.

[0235] The adjusted model parameter formula is:

[0236] F v =k 1v ·R+k 2v exp(M)

[0237] This formula is used to adjust the parameters of the fire spread prediction model, where k 1v and k 2v are adjustment coefficients, R is the correlation analysis result obtained from the previous step, M is the current meteorological data influence coefficient, and k is set 1v = 0.8 and k 2v =0.5, and from the above calculation R v The value of is 6, and the influence coefficient of M is set to 3. The calculation is as follows:

[0238] F v =0.8·6+0.5·exp(3)

[0239] F v =4.8+0.5·20.085

[0240] F v =4.8+10.0425=14.8425

[0241] The results show that the adjusted fire spread prediction model parameter is 14.8425, combining the exponential impact of the environmental relevance score and real-time meteorological data, which will be used to optimize the response and accuracy of the early warning system and improve the effectiveness of fire risk early warning.

[0242] Adjusted fire risk warning scheme formula:

[0243] P v =α v ·F v +β v ·G

[0244] This formula is used to construct a fire risk early warning scheme, where α v and β v is the adjustment factor, F v is the adjusted fire spread prediction model parameter, G represents the comprehensive impact of risk factors, and the adjustment coefficient α is set v =0.7,β v = 0.3, and from the previous calculation, the value of F is 14.8425, and the value of G is set to 5, the calculation is as follows:

[0245] P v =0.7·14.8425+0.3·5

[0246] P v =10.38975+1.5=11.88975

[0247] The results show that combined with the influence of the adjusted model parameters and other risk factors, the final fire risk warning scheme is scored 11.88975. The score can assist related departments to respond in a timely manner, optimize fire prevention and response measures, and ensure community safety.

[0248] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. The urban fire intelligent early warning system is characterized by: The system comprises: The urban structure intelligent analysis module allocates nodes to buildings based on the geographical location, materials, fire data and building age data of urban buildings, takes the channel type and distance between buildings as the attributes of the edges, generates an urban graph model, and generates a node and edge feature matrix according to the node and edge attributes of the urban graph model; The fire dynamic warning simulation module inputs the node and edge feature matrix into the graph convolutional network, analyzes the impact of building characteristics and interconnectivity on fire spread, obtains a fire spread influence score, and uses the fire spread influence score to analyze the propagation path and speed of differentiated fire source points to generate a fire spread prediction model; The fire prevention strategy optimization module uses the fire spread prediction model to calculate the evacuation difficulty and fire risk level of each building under differentiated fire source points, generates a building risk level assessment result, optimizes the evacuation route and fire isolation zone according to the building risk level assessment result, and generates a fire isolation zone planning map; The environmental impact intelligent analysis module captures meteorological condition data, urban population dynamics and holiday activity data based on the fire isolation zone planning map, analyzes the correlation between environmental factors and fire occurrence, obtains the environmental factor impact score, adjusts the fire spread prediction model based on the environmental factor impact score, and constructs an adjusted fire risk warning plan.

2. The urban fire intelligent early warning system according to claim 1 is characterized in that: The steps for obtaining the node and edge feature matrices are specifically as follows: Based on the data of urban building location, materials, fire data and building age, key data are extracted from geographic information and building databases and organized into a unified format to obtain a set of urban building attributes; Using the urban building attribute set, a graph node is assigned to each building, and the channel information between buildings is encoded as a graph edge, and a graph structure model reflecting the real-time urban layout is constructed using a graph theory method to obtain a city layout map; Based on the city layout graph, the attribute information of nodes and edges is extracted using the formula: M F =P N ·k N +P E ·k E Perform attribute weighting and merging, calculate the integrated feature scores of nodes and edges, and obtain the node and edge feature matrix, where P N Represents the attribute set of the node, P E represents the attribute set of the edge, k N and k E are the attribute weight coefficients of nodes and edges, M F Represents the integrated feature scores of nodes and edges.

3. The urban fire intelligent early warning system according to claim 2 is characterized in that: The steps for obtaining the fire spread influence score are as follows: Input the node and edge feature matrix into the graph convolutional network to encode and convert the features, optimize the nonlinear representation of the data through iterative processing of the network layer, and generate feature data after graph convolution processing; Analyze the feature data after the graph convolution processing, using the formula: I fire,b =∑(k b ·F node,b )+∑(j b ·F edge,b ) The characteristics of nodes and edges are weighted and summarized to evaluate the potential contribution of each building and connection to fire spread and obtain the fire impact score, where I fire,b represents the fire impact score, k b and j b is the weight coefficient of node and edge characteristics, F node,b and F edge,b Represent the feature vectors of nodes and edges respectively; According to the fire impact score, the fire spread potential of multiple buildings in the entire urban network is integrated to generate a fire spread impact score.

4. The urban fire intelligent early warning system according to claim 3 is characterized in that: The steps for obtaining the fire spread prediction model are specifically as follows: Integrate the fire propagation influence score, simulate differentiated fire source points, simulate the fire spread path through graph theory algorithm, and generate the fire source point spread path data; By analyzing the diffusion path data of the fire source point, the formula is applied: Calculate the fire spread speed of the differentiated path and obtain speed evaluation data, where V c Indicates the speed of fire spread, d c is the total length of the path, a c is the acceleration factor, L c is the complexity index of the path, t c is the predicted diffusion time, ∈ c is the time error correction factor; Combining the speed assessment data with the fire spread influence score, the model is optimized through multivariate regression analysis, the data is integrated, and the fire spread trend of the network is predicted to generate a fire spread prediction model.

5. The urban fire intelligent early warning system according to claim 4 is characterized in that: The steps for obtaining the building risk level assessment result are specifically as follows: Based on the fire spread prediction model, the location of the differentiated fire source and the associated building characteristics are input, the evacuation difficulty of each building is calculated, and preliminary evacuation difficulty data is obtained; Based on the preliminary evacuation difficulty data, combined with the original safety record and structural characteristics of each building, the fire risk level is calculated to obtain a preliminary fire risk score; Combining the above preliminary fire risk score and the difficulty of evacuation for each building, the risk level is calculated and assessed, applying the formula: F d =a d ·R d +b d ·S d The building risk level assessment results are obtained, where R d represents the preliminary fire risk score, S d represents the structural parameters of the building, α d and β d is the weight parameter adjusted based on the original data, F d Indicates the building risk level assessment result.

6. The urban fire intelligent early warning system according to claim 5 is characterized in that: The steps for obtaining the fire isolation zone planning map are specifically as follows: Evaluate the risk level assessment results of the building and the surrounding environment, capture data on the structural safety of the building and previous evacuation routes, identify potential fire isolation areas, and generate an overview of preliminary fire isolation needs; Based on the above preliminary overview of fire separation requirements, refine the layout and evacuation requirements of the building and apply the formula: Construct the layout of fire isolation zones, where A h represents the recommended area of ​​the fire separation zone, β h is the safety parameter, S h is the building area, γ h is the environmental adjustment factor; Combining the fire separation zone layout with real-time building occupancy, apply the formula: P h =d h ·A h +∈ h ·A h Generate a fire isolation zone planning map, where P h Represents the effectiveness of the quarantine area planning map, δ h and ∈ h are adjustment coefficient and risk assessment coefficient, A h Indicates the recommended area of ​​the fire separation zone.

7. The urban fire intelligent early warning system according to claim 6 is characterized in that: The steps for obtaining the environmental factor impact score are specifically as follows: According to the fire isolation zone planning map, meteorological condition data, urban population dynamics and holiday activity data are captured, and integrated into a single environmental data set through data processing to generate a processed environmental data set; The processed environmental data set is statistically analyzed using the formula: Calculate the contribution of multiple data factors to fire risk and generate statistical analysis results, where w 1g 、w 2g 、w 3g Represents the weight of the differentiated data factor, D 1g , D 2g , D 3g Represents the corresponding data value, N g Represents the total number of data points, Y g Indicates the results of statistical analysis; Using the statistical analysis results, combined with the original fire records and trend analysis of the region, the scoring model is used to optimize the formula: Get the environmental factor impact score, where k 1g , k 2g is the adjustment coefficient, S g Indicates the environmental factor impact score, Y g Represents the statistical analysis results.

8. The urban fire intelligent early warning system according to claim 7 is characterized in that: The steps for obtaining the adjusted fire risk warning plan are specifically as follows: Comparing the environmental factor impact score with the original fire data set, sorting the environmental factor data set, screening variables associated with fire frequency, processing the original fire data, determining the correlation between the environmental score and the fire frequency through comparative analysis, and generating correlation analysis results; Combining the correlation analysis results with the current meteorological data, the formula is adopted: F v =k 1v ·R+k 2v ·exp(M) Adjust the fire spread prediction model parameters to generate adjusted model parameters, where k 1v and k 2v They are adjustment coefficients, M represents the real-time meteorological data impact coefficient, F v represents the adjusted model parameters; The adjusted model parameters are used to reassess the fire risk in the area, and an adjusted fire risk warning plan is generated based on the model prediction results.

Images