Dynamic three-dimensional escape route generation method based on fire point

Abstract

The present application relates to the technical field of escape route analysis, and discloses a dynamic three-dimensional escape route generation method based on a fire point, comprising: acquiring temperature data, smoke data, train state data and a three-dimensional space structure in a target site in real time, analyzing the preprocessed temperature data, smoke data and train data to obtain three-dimensional coordinates of the fire point and a fire situation-train coupling risk factor, and the target site being a subway station; capturing flow field changes through a three-dimensional transient wind flow field intensity atlas, and finally constructing a dynamic risk avoidance traffic atlas and generating an escape route in combination with regional smoke condition adaptive weight values and exit dynamic evacuation efficiency values, wherein the route generation process is real-time adaptive to dynamic changes of the flow field, smoke and personnel distribution, which ensures route safety and also takes into account evacuation efficiency, and effectively avoids route failure caused by flow field or smoke mutation.

Description

Technical Field

[0001] This invention relates to the field of escape route analysis technology, specifically to a method for generating dynamic three-dimensional escape routes based on the point of ignition. Background Technology

[0002] Currently, in many places, once a fire is detected, escape routes are usually generated by combining sensor networks and AI algorithms. The fire point is monitored in real time by temperature and smoke sensors deployed in multiple areas, and the shortest safe path is calculated using an improved path planning algorithm. The dynamic escape routes are then presented through a 3D visualization platform.

[0003] However, the above-mentioned method of generating escape routes still has the following drawbacks in subway stations: After the fire starts, a large amount of high-temperature smoke will quickly form in the subway station. Although the existing sensors can initially capture the distribution of smoke, when the train enters the station, the strong piston wind generated by the train will instantly change the direction of smoke diffusion. Since the sensors are mostly fixedly deployed on the top of the platform and the side walls of the passage, they cannot quickly identify this unsteady flow field characteristic. This will cause the previously generated escape routes to fail, which is not conducive to the evacuation of people. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for generating dynamic three-dimensional escape routes based on the ignition point of a fire, thus solving the aforementioned problems.

[0005] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0006] A dynamic 3D escape route generation method based on the ignition point of a fire includes:

[0007] Step S1: Real-time acquisition of temperature data, smoke data, train status data and three-dimensional spatial structure in the target location; analysis of preprocessed temperature data, smoke data and train data to obtain the three-dimensional coordinates of the ignition point and the fire-train coupling risk factor; the target location is a subway station.

[0008] Step S2: Based on train status data, analyze the airflow change pattern of the target location to obtain a three-dimensional transient wind flow field intensity map covering the entire space of the target location;

[0009] Step S3: Combine the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map to generate a dynamic smoke cloud evolution body in three-dimensional space.

[0010] Step S4: Analyze the dynamic smoke cloud evolution body and the three-dimensional spatial structure, assign a channel safety resilience coefficient to each three-dimensional space, and generate a dynamic risk avoidance passage map.

[0011] Step S5: Obtain the three-dimensional coordinates of the people to be evacuated in real time, analyze the three-dimensional coordinates and dynamic evacuation and passage map, and generate a three-dimensional escape route.

[0012] Furthermore, the preprocessed temperature data, smoke data, and train data were analyzed to obtain the three-dimensional coordinates of the ignition point and the fire-train coupling risk factors, including:

[0013] Dynamic spatiotemporal domain division is performed on the preprocessed temperature data, smoke data, and train status data. The area centered on the real-time position of the train is used as the dynamic interference zone, and the remaining areas are used as the static reference zone, generating a dynamic-static domain data partitioning matrix.

[0014] Based on the dynamic-static domain data partitioning matrix, the train disturbance correction value of temperature data in the dynamic interference zone and the natural diffusion gradient value of temperature data in the static reference zone are calculated respectively.

[0015] The train disturbance correction value is fused with the natural diffusion gradient value to generate a zoned temperature feature vector;

[0016] Based on the temperature feature vector of each zone, the smoke data is analyzed to generate a smoke source tracing weight map.

[0017] Furthermore, the preprocessed temperature data, smoke data, and train data were analyzed to obtain the three-dimensional coordinates of the ignition point and the fire-train coupling risk factor, which also included:

[0018] The relationship between running speed and acceleration in the train status data is analyzed, and the impact intensity of the train on the airflow in the dynamic interference zone during the train's entry and exit from the station is calculated to obtain the train aerodynamic disturbance coefficient.

[0019] Using the extreme points of the zone temperature feature vector as initial coordinates, combined with the probability direction of the smoke source tracing weight map, and analyzing the spatial connectivity characteristics of each region in the three-dimensional spatial structure, the three-dimensional coordinates of the ignition point and the corresponding initial fire intensity are obtained.

[0020] The initial fire intensity is combined with the train aerodynamic disturbance coefficient to obtain the dynamic spread risk value;

[0021] Analyze the spatial conflict probability between the ignition point and the train's trajectory to generate a fire-train coupling risk factor.

[0022] Furthermore, based on train status data, the airflow variation patterns at the target location are analyzed to obtain a three-dimensional transient wind flow field intensity map covering the entire space of the target location, including:

[0023] Based on the dynamic-static domain data partitioning matrix, the spatial range of the dynamic interference zone and the static reference zone is analyzed, and a domain-related airflow analysis table is generated.

[0024] Based on the domain-related airflow analysis table, the train wake wind speed characteristics in the dynamic interference zone are calculated, and the airflow attenuation coefficient in the static reference zone is analyzed. The train wake wind speed characteristics and airflow attenuation coefficient are integrated to obtain the set of airflow characteristic parameters for each zone.

[0025] Obtain the parameters of the fixed ventilation facilities in the three-dimensional spatial structure, and spatially constrain the set of characteristic parameters of the airflow in the partition based on the fixed ventilation facility parameters to obtain the structurally constrained airflow vector field;

[0026] Based on the real-time position and speed of the train, the airflow influence range is defined according to the speed with the train position as the center, and the airflow disturbance intensity of each grid within the airflow influence range is calculated.

[0027] Based on the reference wind speed of the structurally constrained airflow vector field in each grid, the airflow disturbance intensity of each grid is calculated with the reference wind speed to form a three-dimensional transient airflow field intensity map covering the entire space.

[0028] Furthermore, based on the fixed ventilation facility parameters, spatial constraints are applied to the zonal airflow characteristic parameter set to obtain the structurally constrained airflow vector field, including:

[0029] The parameters of fixed ventilation facilities are analyzed, and a ventilation facility area constraint table is generated based on the spatial range covered by the fixed ventilation facilities.

[0030] Spatially associate the ventilation facility area constraint table with the zone airflow characteristic parameter set, and adjust the airflow intensity and direction for areas affected by different airflows to generate a spatially adapted airflow correction set.

[0031] Based on the structural characteristics of the target site in three-dimensional space, the spatially adapted airflow correction set is optimized to obtain the structurally constrained airflow vector field.

[0032] Furthermore, by combining the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map, a dynamic smoke cloud evolution body in three-dimensional space is generated, including:

[0033] Using the three-dimensional coordinates of the ignition point as the origin, the energy of fire spread is analyzed by combining the fire-train coupling risk factor and dynamic diffusion risk value, and a dynamic correlation value between fire and train is generated.

[0034] Based on the three-dimensional transient wind flow field intensity map, the dynamic correlation value between fire and train is transiently corrected to generate a dynamic wind field adaptive diffusion set.

[0035] The system acquires the number of people in the target location in real time, analyzes the degree of obstruction of smoke diffusion by the area where people gather based on the number of people, and generates a people density obstruction coefficient.

[0036] The dynamic wind field adaptive diffusion set is corrected based on the personnel density hindrance coefficient to obtain the dual-disturbance diffusion set for people and vehicles.

[0037] Furthermore, by combining the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map, a dynamic smoke cloud evolution in three-dimensional space is generated, which also includes:

[0038] Based on the human-vehicle dual-perturbation corrected diffusion set, the correlation between the concentration change and diffusion rate of smoke at each spatial point is analyzed, and the dual-perturbation coupled concentration feedback coefficient is generated.

[0039] Based on the three-dimensional transient wind flow field intensity map and the dual-perturbation coupled concentration feedback coefficient, the smoke distribution, concentration and diffusion area at each time node are analyzed to generate a trend diffusion optimization set;

[0040] By analyzing the trend diffusion optimization set, the double-perturbation coupling concentration feedback coefficient, and the personnel density hindrance coefficient, a dynamic smoke cloud evolution body in three-dimensional space is generated.

[0041] Furthermore, the dynamic smoke cloud evolution body and three-dimensional spatial structure are analyzed, and a channel safety resilience coefficient is assigned to each three-dimensional space to generate a dynamic risk avoidance and passage map, including:

[0042] Analyze the three-dimensional spatial structure of the target site, analyze the spatial relationship between the physical properties of each passage and the train trajectory, and generate a passage-train spatial relationship table;

[0043] Based on the dynamic smoke cloud evolution, the concentration and diffusion rate of smoke in the channel at each time point are calculated. Combined with the channel-train spatial correlation table, the actual threat level of smoke to the channel is calculated, and the smoke-train channel threat weight value is generated.

[0044] Analyze the evacuation function characteristics of channels in a three-dimensional spatial structure and generate the channel evacuation fault tolerance coefficient;

[0045] The safety resilience coefficient of the passage is obtained by integrating the passage-train spatial association table, the smoke-vehicle passage threat weight value, and the passage evacuation fault tolerance coefficient.

[0046] Furthermore, the dynamic smoke cloud evolution body and three-dimensional spatial structure are analyzed, and a channel safety resilience coefficient is assigned to each three-dimensional space to generate a dynamic risk avoidance and passage map. This also includes:

[0047] Based on the dynamic smoke cloud evolution and train status data, the channel safety resilience coefficient is dynamically corrected in two dimensions to generate a dual-disturbance dynamic safety resilience coefficient.

[0048] Based on the three-dimensional spatial structure, the dynamic safety resilience coefficient of the dual-disturbance system is marked to each channel to generate a dynamic risk avoidance and passage map.

[0049] Furthermore, the three-dimensional coordinates and dynamic hazard avoidance and passage maps are analyzed to generate three-dimensional escape routes, including:

[0050] The three-dimensional coordinates are correlated with the dynamic risk avoidance and passage map, and combined with the dynamic smoke cloud evolution body, the regional smoke situation adaptation adjustment of the dual-disturbance dynamic safety resilience coefficient is carried out to generate regional smoke situation adaptation weight value.

[0051] Real-time acquisition of the number of safety exits at the target location and the number of people at each safety exit;

[0052] Based on the number of people, the number of safety exits, and the three-dimensional coordinates of the people to be evacuated, the dynamic evacuation efficiency value of each safety exit is calculated.

[0053] Using regional smoke condition matching weight as the core constraint and exit dynamic evacuation efficiency as the optimization target, a path search is performed in the channel of the dynamic risk avoidance passage map to generate a three-dimensional escape route with safety and efficiency.

[0054] In summary, the present invention has the following main beneficial effects:

[0055] By dynamically dividing the spatiotemporal domain, a dynamic interference zone and a static reference zone are defined with the real-time position of the train as the center, generating a dynamic-static domain data partition matrix. Then, the train's aerodynamic disturbance coefficient is calculated by combining the train's speed and acceleration to understand the impact intensity of the piston wind on the airflow. Based on this matrix, a domain-related airflow analysis table is generated to specifically calculate the train wake wind speed characteristics in the dynamic interference zone and the airflow attenuation coefficient in the static reference zone. At the same time, fixed ventilation facility parameters are incorporated to form a structurally constrained airflow vector field. Finally, the spherical airflow influence range is defined according to the real-time position and speed of the train, the grid disturbance intensity is calculated, and a three-dimensional transient wind flow field intensity map is generated. The entire process realizes real-time tracking of the instantaneous changes in the flow field caused by the piston wind. Compared with traditional sensors that can only initially capture smoke distribution, this solution can analyze the dynamics of the flow field in the entire space and avoid the failure of escape routes due to flow field identification errors.

[0056] A smoke diffusion analysis system with triple coupling of fire, train, and personnel was constructed. Taking the three-dimensional coordinates of the ignition point as the origin, the fire-train coupling risk factor and dynamic diffusion risk value were integrated to generate a dynamic correlation value between fire and train, thus understanding the comprehensive energy of fire diffusion. The correlation value was then transiently corrected by combining the synthetic wind speed from the three-dimensional transient wind flow field intensity map, generating a dynamic wind field-adapted diffusion set. At the same time, a personnel density hindrance coefficient was introduced to correct the diffusion value based on the real-time number of people in the grid, taking into account the hindrance effect of personnel gathering on smoke. Furthermore, the dual-perturbation coupling concentration feedback coefficient was combined with the flow field guidance direction and time slice division to generate a trend diffusion optimization set that marks smoke concentration, diffusion direction, and regional attributes. Finally, a dynamic smoke cloud evolution body was obtained, which adapts in real time to the dual interference of piston wind and personnel gathering, and dynamically updates the smoke concentration and diffusion direction. Compared with the static smoke monitoring of traditional solutions, this system can more accurately understand the dynamic distribution of smoke.

[0057] By calculating the smoke concentration and diffusion rate in the passageway, and combining the passageway-train spatial correlation table, a smoke-train passageway threat weight value is generated. At the same time, the passageway evacuation tolerance coefficient and fire protection attributes are incorporated to obtain the passageway safety resilience coefficient. The dual-dimensional dynamic safety resilience coefficient is generated by dynamically correcting the smoke coverage ratio and train speed. Finally, a dynamic risk avoidance and passage map is marked with green, yellow and red colors. Furthermore, the regional smoke situation adaptation weight value is calculated with the three-dimensional coordinates of the evacuees as the center, combined with the dynamic smoke cloud evolution body. At the same time, the exit distance, congestion and number are combined to calculate the exit dynamic evacuation efficiency value. With the regional smoke situation adaptation weight value as the core constraint and the exit evacuation efficiency value as the optimization target, the path with the largest comprehensive weight of the passageway is searched. The route generation process adapts to the dynamic changes of the flow field, smoke and personnel distribution in real time, which ensures that the route is always in a low-risk area and takes into account evacuation efficiency, effectively avoiding route failure caused by sudden changes in the flow field or smoke. Attached Figure Description

[0058] Figure 1 This is a flowchart illustrating the steps of the dynamic three-dimensional escape route generation method based on the ignition point of the present invention. Detailed Implementation

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] refer to Figure 1 A method for generating dynamic three-dimensional escape routes based on the ignition point of a fire includes:

[0061] Step S1: Real-time acquisition of temperature data, smoke data, train status data, and three-dimensional spatial structure within the target location. Train status data includes: maximum train cross-section, train car height, entry and exit time, real-time train position (three-dimensional coordinates), running speed, acceleration, and running direction. The pre-processed temperature data, smoke data, and train data are analyzed to obtain the three-dimensional coordinates of the ignition point and the fire-train coupling risk factor. The target location is a subway station.

[0062] Step S2: Based on train status data, analyze the airflow change pattern of the target location to obtain a three-dimensional transient wind flow field intensity map covering the entire space of the target location;

[0063] Step S3: Combine the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map to generate a dynamic smoke cloud evolution body in three-dimensional space.

[0064] Step S4: Analyze the dynamic smoke cloud evolution body and the three-dimensional spatial structure, assign a channel safety resilience coefficient to each three-dimensional space, and generate a dynamic risk avoidance passage map.

[0065] Step S5: Obtain the three-dimensional coordinates of the people to be evacuated in real time, analyze the three-dimensional coordinates and dynamic evacuation and passage map, and generate a three-dimensional escape route.

[0066] By acquiring real-time train status data and combining it with temperature, smoke data, and three-dimensional spatial structure, the system first accurately locates the ignition point and calculates the fire-train coupling risk factor. Then, based on train data analysis of airflow changes, it generates a three-dimensional transient wind flow field intensity map covering the entire space, effectively capturing the unsteady flow field characteristics caused by piston wind. Furthermore, by combining the three-dimensional transient wind flow field intensity map with ignition point information, it generates a dynamic smoke cloud evolution body, assigns a safety resilience coefficient to the passage, and constructs a dynamic hazard avoidance and passage map. Finally, it generates a three-dimensional escape route adapted to the flow field changes based on the three-dimensional coordinates of the people to be evacuated, avoiding route failure due to sudden changes in smoke diffusion direction, and facilitating the efficient evacuation of people in the subway station.

[0067] In one embodiment, the preprocessed temperature data, smoke data, and train data are analyzed to obtain the three-dimensional coordinates of the ignition point and the fire-train coupling risk factor, including:

[0068] Dynamic spatiotemporal domain partitioning is performed on preprocessed temperature data, smoke data, and train status data. The area centered on the real-time position of the train is designated as the dynamic interference zone, and the remaining areas are designated as the static reference zone. A dynamic-static domain data partitioning matrix is ​​generated. Specifically, this includes: interpolating the preprocessed temperature and smoke data using a 0.5m × 0.5m × 0.5m spatial grid to obtain the temperature value and smoke concentration for each grid; using the train's three-dimensional coordinates as the center; and defining a cylindrical dynamic interference zone spatial range with a horizontal radius of 1.2 times the train's maximum cross-sectional diameter and a vertical radius of 1.2 times the carriage height; and sampling data every 5 seconds. The frequency is divided into time slices, and each spatial grid is combined with its corresponding time slice to form a spatiotemporal unit of spatial grid-time slice. A unique index is assigned to each spatiotemporal unit. Each unit contains temperature, smoke, and train data for that grid during that time period. For each spatiotemporal unit, if the coordinates of the spatiotemporal unit are within the spatial range of the dynamic interference zone, then the region is a dynamic interference zone. If the coordinates of the spatiotemporal unit are not within the spatial range of the dynamic interference zone, then the region is a static reference zone. A dynamic-static domain data partitioning matrix is ​​generated with the spatiotemporal unit index as the row and the region type (dynamic interference zone, static reference zone) as the two columns. The corresponding region is filled with 1, and the other column is filled with 0.

[0069] Based on the dynamic-static domain data partitioning matrix, the train disturbance correction value of temperature data in the dynamic interference zone and the natural diffusion gradient value of temperature data in the static reference zone are calculated respectively. Specifically, for the dynamic interference zone, for the spatiotemporal cell marked as 1 in the dynamic-static domain data partitioning matrix, the average temperature of all spatiotemporal cells is subtracted from the temperature value of the spatiotemporal cell to obtain the train disturbance correction value; while for the static reference zone, for the spatiotemporal cell marked as 0 in the dynamic-static domain data partitioning matrix, the temperature values ​​of the six adjacent cells (front, back, left, right, up, down) around the cell in each spatiotemporal cell are selected, the temperature difference between the temperature value and the current spatiotemporal cell is calculated, and then divided by the cell spacing to obtain the gradient in each direction. The gradients in each direction are added together and the average value is taken as the natural diffusion gradient value.

[0070] The train disturbance correction value and the natural diffusion gradient value are fused to generate a zone temperature feature vector. Specifically, the train disturbance correction value and the natural diffusion gradient value are normalized to the range of 0-1 respectively; the train disturbance correction value and the natural diffusion gradient value of each spatiotemporal unit are arranged in ascending order according to the index of the spatiotemporal unit to form a one-dimensional array, which is the zone temperature feature vector.

[0071] Based on the zoned temperature feature vector, the smoke data is analyzed to generate a smoke source weighting map. Specifically, this includes: normalizing the train disturbance correction value and smoke concentration corresponding to the spatiotemporal unit in the zoned temperature feature vector to the range of 0-1; setting the wake weight according to train speed (0.6 when train speed ≥ 5 m / s, otherwise 0.4), multiplying the normalized train disturbance correction value by the wake weight + multiplying the normalized smoke concentration by (1 minus the wake weight) to obtain the unit smoke weight for the dynamic interference zone; and considering natural diffusion in the zoned temperature feature vector. The gradient value is weighted as follows: when the natural diffusion gradient value is ≥2℃ / m, the weight is 0.7; when the natural diffusion gradient value is <2℃ / m, the weight is 0.3. Three adjacent spatiotemporal cells are selected, and the ratio of the natural diffusion gradient value of each adjacent spatiotemporal cell to the gradient of the current cell is calculated to obtain the gradient proportion. For each adjacent cell, the smoke concentration of the adjacent spatiotemporal cell is multiplied by the gradient proportion and then by the weight to obtain the initial smoke weight. The initial smoke weights of the three adjacent cells are then added together to obtain the cell smoke weight of the static reference area. The smoke weights of all cells are arranged in ascending order of spatiotemporal cell index into a two-dimensional grid diagram, i.e., the smoke source weight diagram.

[0072] By defining a cylindrical dynamic interference zone centered on the real-time position of the train, with a cross-sectional diameter 1.2 times the train's maximum diameter and a carriage height 1.2 times the train's height, and designating the remaining area as a static reference zone, a spatiotemporal unit and dynamic-static domain data partitioning matrix are constructed using 5-second time slices. For the dynamic interference zone, the interference of piston wind on temperature data is proposed by calculating the train disturbance correction value. For the static reference zone, the natural diffusion gradient value is calculated to restore the temperature diffusion law without train influence. This approach compensates for the deficiency of fixed sensors in quickly identifying unsteady flow fields, allowing temperature data to truly reflect the state after the superposition of fire and train interference.

[0073] By setting wake weights in the dynamic interference zone and train speed, the normalized train disturbance correction value and smoke concentration are integrated according to weights to reflect the impact of piston wind on smoke. The static reference zone sets weights based on the natural diffusion gradient value, and calculates weights by combining the gradient ratio of adjacent units and smoke concentration to restore the smoke distribution under natural diffusion. This allows the smoke source tracing weight map to match the unsteady flow field changes caused by piston wind in real time, enabling accurate tracing of the source of smoke and avoiding the failure of subsequent escape routes due to sudden changes in the direction of smoke diffusion.

[0074] In one embodiment, the preprocessed temperature data, smoke data, and train data are analyzed to obtain the three-dimensional coordinates of the ignition point and the fire-train coupling risk factor, and the analysis also includes:

[0075] The relationship between train speed and acceleration in the train status data is analyzed, and the impact intensity of the train on the airflow in the dynamic interference zone during the train's entry and exit from the station is calculated to obtain the train aerodynamic disturbance coefficient. Specifically, this includes: multiplying the real-time train speed by the acceleration and then dividing by the cross-sectional area of ​​the dynamic interference zone (π × horizontal radius R²) to obtain the airflow impact intensity per unit area; obtaining the maximum impact intensity of the train entering the station at full speed; dividing the airflow impact intensity per unit area by the maximum impact intensity, and normalizing the calculation result to the range of 0-1, which is the train aerodynamic disturbance coefficient.

[0076] Using the extreme points of the partitioned temperature feature vector as initial coordinates, combined with the probability direction of the smoke source weighting map, and analyzing the spatial connectivity characteristics of each region in the three-dimensional spatial structure, the three-dimensional coordinates of the ignition point and the corresponding initial fire intensity are obtained. Specifically, this includes: traversing the temperature values ​​of all spatiotemporal units in the partitioned temperature feature vector one by one, and using the spatial grid coordinates of the spatiotemporal units corresponding to the first three maximum temperature values ​​as candidate initial coordinates; for the spatial grid coordinates of the first candidate coordinate, in the smoke source weighting map, using the direction of the unit with the largest smoke weight within the 3×3×3 cube surrounding it as the direction of the maximum probability of fire source spread; treating each grid in the three-dimensional space of the subway station as a node, and determining whether the node belongs to a static reference area or a dynamic reference area. Dynamic interference zone; Construct a 3×3×3 cube region centered on the candidate coordinates. Check if there is at least one node within the cube region that simultaneously satisfies a temperature exceeding 60℃ and a smoke weight (unit smoke weight of the dynamic interference zone or static reference zone) > 0.8, and if this node can form a continuous path with the candidate coordinates, then it is considered connected; If the first candidate coordinate is not connected, then check the second and third candidate coordinates in turn; If all three are not connected, an alarm is triggered, and manual intervention is required; Obtain the historical maximum fire temperature of the target location; For the finally connected candidate coordinates, divide its temperature value by the historical maximum fire temperature, and normalize the calculation result to 0-1 to obtain the initial fire intensity. This candidate coordinate is the three-dimensional coordinate of the ignition point;

[0077] The initial fire intensity and the train aerodynamic disturbance coefficient are combined to obtain a dynamic diffusion risk value. Specifically, the initial fire intensity is weighted at 0.6 and the train aerodynamic disturbance coefficient is weighted at 0.4. The dynamic diffusion risk value (range 0-1, with higher values ​​indicating greater risk) is obtained by multiplying the initial fire intensity by 0.6 and the train aerodynamic disturbance coefficient by 0.4. The initial fire intensity directly reflects the diffusion dynamics of the fire source itself. Higher temperatures and stronger fires result in faster natural diffusion rates of smoke and fire, making it the fundamental factor determining diffusion risk. Therefore, its weight is relatively high at 0.6. The train aerodynamic disturbance coefficient, on the other hand, is an external environmental disturbance that diffuses only through airflow entering and leaving the station. It is an external influence and cannot be separated from the fire itself to determine the diffusion trend. Therefore, its weight is relatively low at 0.4.

[0078] The spatial conflict probability between the fire ignition point and the train trajectory is analyzed to generate a fire-train coupling risk factor. Specifically, this involves: using the three-dimensional coordinates of the fire ignition point as the center, setting a spherical influence radius according to the initial fire intensity (5m for intensity 0.1-0.3, 8m for 0.4-0.6, and 12m for 0.7-1.0) to form a three-dimensional spatial influence domain; calculating the total trajectory length (the sum of distances between adjacent coordinate points) based on the continuously updated position coordinates of the train during its entry and exit from the station; comparing the total trajectory length with the three-dimensional spatial influence domain segment by segment, counting the lengths of line segments within the three-dimensional spatial influence domain, and dividing the line segment lengths by the total trajectory length to obtain the spatial conflict probability (range 0-1); multiplying the spatial conflict probability and the dynamic diffusion risk value by 0.5 respectively, adding them together, and normalizing the calculation result to 0-1 to obtain the fire-train coupling risk factor.

[0079] By calculating the train aerodynamic disturbance coefficient, the impact intensity of piston wind on airflow is understood. By combining the extreme values ​​of the zone temperature characteristic vector and the smoke source weight map, connectivity is verified in a 3×3×3 cubic region, the ignition point is accurately located and the initial fire intensity is calculated. Then, the dynamic fire-train coupling risk factor is generated to understand the superposition effect of fire source diffusion and train disturbance, thus realizing the dynamic quantification of the interaction between fire and train under unsteady flow field.

[0080] In one embodiment, based on train status data, the airflow variation pattern of the target location is analyzed to obtain a three-dimensional transient wind flow field intensity map covering the entire space of the target location, including:

[0081] Based on the dynamic-static domain data partitioning matrix, the spatial extent of the dynamic interference zone and the static reference zone is analyzed, and a domain-related airflow analysis table is generated. Specifically, this includes: extracting the spatiotemporal cells marked 1 from the dynamic-static domain data partitioning matrix, selecting the minimum and maximum values ​​of their three-dimensional coordinates (x, y, z), using the minimum starting coordinate as the starting boundary, and the maximum starting coordinate + grid spacing as the ending boundary to determine the spatial extent of the dynamic interference zone; extracting the spatiotemporal cells marked 0 from the dynamic-static domain data partitioning matrix, selecting the minimum and maximum values ​​of their x, y, z coordinates, using the minimum starting coordinate as the starting boundary, and the maximum starting coordinate + grid spacing as the ending boundary to determine the spatial extent of the static reference zone; and generating a table containing the spatial extent of the dynamic interference zone, the spatial extent of the static reference zone, and the corresponding indexes of adjacent spatiotemporal cells, i.e., the domain-related airflow analysis table.

[0082] Based on the domain-related airflow analysis table, the train wake velocity characteristics in the dynamic interference zone are calculated, and the airflow attenuation coefficient in the static reference zone is analyzed. The train wake velocity characteristics and airflow attenuation coefficient are integrated to obtain a set of zoned airflow characteristic parameters. Specifically, this includes: finding the real-time train speed corresponding to the spatial range of the dynamic interference zone from the domain-related airflow analysis table, multiplying the train speed by 0.3 to obtain the wake velocity of the dynamic unit, and calculating the average wake velocity of the dynamic unit in all spatiotemporal units of the dynamic interference zone to obtain the train wake velocity characteristics; for the spatial range of the static reference zone in the domain-related airflow analysis table, calculating the distance from the starting boundary to the ending boundary, calculating the attenuated wind speed of each spatiotemporal unit according to the rule of wind speed attenuation of 0.1 for every 1m distance, and calculating the average wind speed of all spatiotemporal units in all static reference zones to obtain the airflow attenuation coefficient; classifying the wake velocity characteristics of the dynamic zone and the airflow attenuation coefficient of the static zone according to the dynamic interference zone and the static reference zone to form a set of zoned airflow characteristic parameters.

[0083] Obtain the parameters of the fixed ventilation facilities in the three-dimensional spatial structure, and spatially constrain the set of characteristic parameters of the airflow in the partition based on the fixed ventilation facility parameters to obtain the structurally constrained airflow vector field;

[0084] The parameters of fixed ventilation facilities include: equipment ID, wind speed, power, area, direction and three-dimensional coordinates of the air outlet of the air duct, etc.

[0085] Based on the real-time position and speed of the train, the airflow influence range is defined with the train position as the center and the speed as the criterion. The airflow disturbance intensity of each grid within the airflow influence range is calculated. Specifically, the following steps are taken: with the real-time three-dimensional coordinates of the train as the center, the initial disturbance radius is set to 3m. The radius is increased by 1m for every 5m / s increase in speed (for example, when the speed is 12m / s, 12÷5=2.4, rounded to the nearest integer 2, the radius is 3+2=5m), thus defining a spherical airflow influence range. For the grid within the spherical airflow influence range of 0.5m, the straight-line distance from the grid center to the train center is calculated. The airflow disturbance intensity of each grid within the spherical airflow influence range is obtained by multiplying the train speed by 0.1 × (0.9 raised to the power of the straight-line distance).

[0086] Based on the reference wind speed of the structurally constrained airflow vector field in each grid, the airflow disturbance intensity of each grid is calculated with the reference wind speed to form a three-dimensional transient wind flow field intensity map covering the entire space. Specifically, this includes: based on the reference wind speed of the structurally constrained airflow vector field in each grid, for each grid, if it is within the coverage area of ​​the spherical airflow influence range, the airflow disturbance intensity is added to the reference value in the same direction and subtracted in the opposite direction to obtain the composite wind speed; if it is not within the coverage area of ​​the spherical airflow influence range, the reference wind speed is directly used as the composite wind speed; the composite wind speed of all grids is mapped to three-dimensional coordinates according to the three-dimensional spatial structure of the subway station, and the mapping is performed according to the rule that the higher the wind speed, the darker the color, and the corresponding color is marked on the position of each grid in the three dimensions to form a three-dimensional transient wind flow field intensity map covering the entire space.

[0087] By determining the spatial range of two regions based on a dynamic-static domain partitioning matrix, the train wake wind speed characteristics in the dynamic interference zone and the airflow attenuation coefficient in the static reference zone are calculated in a targeted manner. Then, a structurally constrained airflow vector field is formed by combining the parameters of fixed ventilation facilities. At the same time, the influence range of the spherical airflow is dynamically delineated according to the train speed and the grid disturbance intensity is calculated. Finally, a three-dimensional transient wind flow field intensity map covering the entire space is generated by synthesizing wind speed. This accurately captures the instantaneous change of airflow caused by piston wind, making up for the shortcomings of fixed sensors that can only conduct preliminary monitoring and cannot reflect unsteady flow fields in real time. This facilitates the generation of subsequent escape routes, avoids the failure of escape routes due to inaccurate flow field identification, and ensures the safety of crowd evacuation.

[0088] In one embodiment, spatial constraints are applied to the set of characteristic airflow parameters for a given zone based on fixed ventilation facility parameters to obtain a structurally constrained airflow vector field, including:

[0089] The parameters of the fixed ventilation system are analyzed, and a ventilation system area constraint table is generated based on the spatial range covered by the fixed ventilation system. Specifically, this includes aligning the major axis of the air outlet of the fixed ventilation system parameters with the three-dimensional coordinate system of the target site. If the major axis is parallel to the positive x-axis (towards the platform), the airflow direction is defined as the positive x-axis direction; if the major axis is parallel to the negative y-axis (towards the passage), it is defined as the negative y-axis direction. A reference wind speed is determined based on the fan power (e.g., 2kW corresponds to 1.2m / s, 3kW corresponds to 1.8m / s). / s); Using the three-dimensional coordinates of the air duct outlet as the center, set the coverage radius according to the power (e.g., 2kW corresponds to 3m, 3kW corresponds to 5m, and the radius increases by 2m for every 1kW increase in power); For all grids within the target area, calculate the straight-line distance from the center of each grid to the air outlet. If the straight-line distance is ≤ the coverage radius of the corresponding power, then the grid is included in the coverage range of the ventilation facility; otherwise, it is not included in the coverage range of the ventilation facility; Generate a ventilation facility area constraint table by treating the facility ID, airflow direction, fan power, reference wind speed, and coverage range as 5 columns;

[0090] The ventilation facility area constraint table is spatially correlated with the zone airflow characteristic parameter set. For areas affected by different airflows, the airflow intensity and direction are corrected to generate a spatially adapted airflow correction set. Specifically, for the grid of the dynamic interference zone, the train wake wind speed characteristic of the zone airflow characteristic parameter set grid is added to the reference wind speed of the corresponding grid of the ventilation facility area constraint table to obtain the corrected airflow intensity, and the direction is the corresponding guiding direction. For the grid of the static reference zone, the airflow attenuation coefficient of the zone airflow characteristic parameter set grid is added to the reference wind speed of the corresponding grid of the ventilation facility area constraint table to obtain the corrected airflow intensity, and the direction is the corresponding guiding direction. The grids are sorted in ascending order of coordinates, and the coordinates, corrected intensity, direction, region (dynamic interference zone or static reference zone), and reference wind speed of each grid are integrated to generate a spatially adapted airflow correction set.

[0091] Based on the structural characteristics of the target site's three-dimensional space, the spatially adapted airflow correction set is optimized to obtain a structurally constrained airflow vector field. Specifically, for the structure of the target site's three-dimensional space, if the area is a narrow channel, the corrected intensity = original intensity × (5 ÷ channel area); if the area is a closed area, the corrected intensity = original intensity × area ratio (compared to the total area); where the direction remains unchanged, the coordinates, corrected intensity, direction, region (dynamic interference area or static reference area), and reference wind speed of each grid are integrated to obtain the structurally constrained airflow vector field.

[0092] By analyzing the parameters of fixed ventilation facilities, the direction of airflow, the reference wind speed and the coverage area are determined, and a ventilation facility area constraint table is generated. Then, the airflow is modified according to the differences between dynamic interference area and static reference area. Finally, the airflow intensity is optimized for spatial structures such as narrow passages and enclosed areas to obtain the structurally constrained airflow vector field. This process fully integrates the function of ventilation facilities and spatial characteristics, avoids the inability to effectively capture smoke diffusion due to flow field identification deviation, ensures the safety of escape routes, and enables people to evacuate safely.

[0093] In one embodiment, the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map are combined to calculate and generate a dynamic smoke cloud evolution body in three-dimensional space, including:

[0094] Using the three-dimensional coordinates of the ignition point as the origin, the energy of fire spread is analyzed by combining the fire-train coupling risk factor and the dynamic diffusion risk value, generating a fire-train dynamic correlation value. Specifically, the dynamic diffusion risk value is weighted at 0.6, and the fire-train coupling risk factor is weighted at 0.4. The coupling risk factor is multiplied by 0.4, and the dynamic diffusion risk value is multiplied by 0.6 to obtain the fire-train dynamic correlation value. The fire-train dynamic correlation value can quantify the comprehensive energy level of fire spread. Among them, the dynamic diffusion risk value is determined by the initial fire intensity, which directly reflects the diffusion power of the fire source itself and is the fundamental energy source of smoke diffusion. Without a fire, there is no basis for diffusion, and its driving role in the evolution of smoke clouds is decisive, so its weight is relatively high at 0.6. On the other hand, the fire-train coupling risk factor is the spatial conflict superposition between the train trajectory and the fire influence domain. It is essentially an external environmental interference and depends on the fire itself for its existence, so its weight is relatively low at 0.4.

[0095] Based on the three-dimensional transient wind flow field intensity map, the dynamic correlation value between fire and train is transiently corrected to generate a dynamic wind field adapted diffusion set. Specifically, this includes: using the composite wind speed of each grid in the three-dimensional transient wind flow field intensity map as the correction basis; setting the weight of the dynamic correlation value between fire and train to 0.7 and the weight of the composite wind speed to 0.3; obtaining the maximum wind speed at the target location; dividing the composite wind speed of each grid by the maximum wind speed to obtain the normalized wind speed; multiplying the dynamic correlation value between fire and train by 0.7 and the dynamic correlation value between fire and train by the normalized wind speed by 0.3 to obtain the corrected diffusion value for each grid; and integrating all corrected diffusion values ​​in ascending order of grid coordinates to generate a dynamic wind field adapted diffusion set.

[0096] The system acquires the number of people in the target area in real time, analyzes the degree of obstruction of smoke diffusion by the crowd gathering area based on the number of people, and generates a crowd density obstruction coefficient. Specifically, the system includes: dividing the target area into 0.5m × 0.5m grids and acquiring the real-time number of people in each grid; setting the minimum standing area per person in the crowd gathering area to 0.25㎡, dividing 0.25㎡ by the horizontal area of ​​a single grid to obtain the saturated number of people in each grid; dividing the real-time number of people in each grid by the saturated number of people in each grid to obtain the normalized density, ranging from 0 to 1; setting the base obstruction coefficient to 0.1; and adding 0.1 to the normalized density × 0.7 to generate a crowd density obstruction coefficient ranging from 0.1 to 0.8.

[0097] The dynamic wind field adaptation diffusion set is corrected based on the personnel density hindrance coefficient to obtain the personnel and vehicle dual disturbance correction diffusion set. Specifically, the process involves multiplying the corrected diffusion value of each grid in the dynamic wind field adaptation diffusion set with the personnel density hindrance coefficient of the corresponding grid according to the 0.5m×0.5m grid coordinates to obtain the single grid dual disturbance correction diffusion value. The dual disturbance correction diffusion values ​​of all grids are then integrated in ascending order of grid coordinates to generate the personnel and vehicle dual disturbance correction diffusion set.

[0098] By using the coordinates of the ignition point as the origin, the dynamic diffusion risk value and the fire-train coupling risk factor are integrated. The diffusion value is then transiently corrected by combining the synthetic wind speed of the three-dimensional transient wind flow field to adapt to the characteristics of the unsteady flow field. In addition, a personnel density resistance coefficient is introduced to correct the diffusion value according to the real-time number of people in the grid. This takes into account the resistance effect of personnel gathering on smoke, and realizes dynamic and accurate analysis of smoke diffusion under the dual interference of piston wind and personnel gathering, so as to avoid the failure of escape routes due to sudden changes in the direction of smoke diffusion.

[0099] In one embodiment, the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map are combined to calculate and generate a dynamic smoke cloud evolution body in three-dimensional space, which also includes:

[0100] Based on the human-vehicle dual-disturbance corrected diffusion set, the correlation between smoke concentration change and diffusion rate at each spatial point is analyzed to generate a dual-disturbance coupled concentration feedback coefficient. Specifically, this includes: for the dual-disturbance corrected diffusion value of each grid in the human-vehicle dual-disturbance corrected diffusion set, finding the smoke concentration of the corresponding grid, subtracting the smoke concentration 5 seconds ago from the current smoke concentration to obtain the smoke change; dividing the smoke change by the dual-disturbance corrected diffusion value to obtain the single-grid concentration-rate correlation value; finding the largest single-grid concentration-rate correlation value among all grids, dividing the single-grid concentration-rate correlation value by the largest single-grid concentration-rate correlation value, and normalizing the calculation result to between 0 and 1 to obtain the dual-disturbance coupled concentration feedback coefficient.

[0101] Based on the three-dimensional transient wind flow field intensity map and the dual-perturbation coupled concentration feedback coefficient, the smoke distribution, concentration, and diffusion area at each time node are analyzed to generate a trend diffusion optimization set. Specifically, the process includes: dividing the time nodes into 5-second time slices; for each node's 0.5m×0.5m×0.5m grid, using the direction of the synthetic wind velocity of the grid in the three-dimensional transient wind flow field intensity map as the smoke diffusion direction; multiplying the dual-perturbation coupled concentration feedback coefficient by the smoke concentration of that grid to obtain the corrected concentration; if the corrected concentration > 0.2, it is marked as a diffusion area; if the corrected concentration ≤ 0.2, it is not marked; and integrating the three-dimensional coordinates, corrected concentration, diffusion direction, and regional attributes (whether it is a diffusion area) of each grid in ascending order of time nodes to generate a trend diffusion optimization set.

[0102] The trend diffusion optimization set, the dual-perturbation coupling concentration feedback coefficient, and the personnel density hindrance coefficient are analyzed to generate a dynamic smoke cloud evolution body in three-dimensional space. Specifically, for each 0.5m×0.5m×0.5m grid, the corrected concentration of the trend diffusion optimization set for that grid is multiplied by the personnel density hindrance coefficient of the corresponding grid to obtain the final smoke concentration. The diffusion direction and regional attributes (whether it is a diffusion area) of the grid in the trend diffusion optimization set are retained, and the three-dimensional coordinates, final smoke concentration, diffusion direction, and regional attributes of all grids are integrated in ascending order of time nodes to generate a dynamic smoke cloud evolution body.

[0103] The direction of smoke diffusion is determined by a three-dimensional transient wind flow field intensity map. Nodes are divided according to a 5-second time slice. The concentration is corrected by a feedback coefficient and the diffusion area is marked to form a trend diffusion optimization set. Finally, the personnel density retardation coefficient is incorporated to obtain the final smoke concentration. The dynamic smoke cloud evolution body is generated. This process adapts to the instantaneous changes in smoke diffusion under unsteady flow field in real time, making up for the shortcomings of fixed sensors that can only initially capture the distribution and cannot dynamically track the diffusion trend.

[0104] In one embodiment, the dynamic smoke cloud evolution body and three-dimensional spatial structure are analyzed, and a channel safety resilience coefficient is assigned to each three-dimensional space to generate a dynamic risk avoidance and passage map, including:

[0105] Analyze the three-dimensional spatial structure of the target site, analyze the spatial relationship between the physical properties of each passage and the train trajectory, and generate a passage-train spatial relationship table, specifically including: obtaining the passage ID and fire protection attributes;

[0106] The three-dimensional space of the target site is divided into 0.5m×0.5m×0.5m grids. For the passages in the three-dimensional space of the target site, the coordinates and fire protection attributes (whether it is a fire protection passage) of each passage are obtained. The straight-line distance between the center grid of the passage and each coordinate point of the train track is calculated, and the minimum straight-line distance is taken as the spatial association distance. A passage-train spatial association table is generated with passage ID, fire protection attribute and spatial association distance as columns.

[0107] Based on the dynamic smoke cloud evolution model, the concentration and diffusion rate of smoke in the channel at each time point are calculated. Combined with the channel-train spatial correlation table, the actual threat level of smoke to the channel is calculated, generating a smoke-train channel threat weight value. Specifically, this includes: finding the final smoke concentration of all grids in the channel from the dynamic smoke cloud evolution model at 5-second time points, and taking the average of the final smoke concentrations of all grids as the channel smoke concentration; subtracting the final smoke concentration of the previous 5 seconds from the final smoke concentration of the current node and dividing by 5 seconds to obtain the diffusion rate; and finding the spatial correlation distance of the corresponding grid from the channel-train spatial correlation table based on this node; normalizing the channel smoke concentration, diffusion rate, and spatial correlation distance to 0-1 respectively; setting the channel smoke concentration weight to 0.6, the diffusion rate weight to 0.3, and the spatial correlation distance weight to 0.1; and multiplying the channel smoke concentration, diffusion rate, and spatial correlation distance by their respective weights and then summing them to obtain the smoke-train channel threat weight value.

[0108] The evacuation function characteristics of the passage in the three-dimensional spatial structure are analyzed to generate the passage evacuation fault tolerance coefficient. Specifically, this includes: obtaining the number of independent paths and the number of diversion nodes from both ends of the passage to the safety exit; if there is only 1 independent path, assign a value of 0.3; if there are ≥2 independent paths, assign a value of 0.7; if there are 0 diversion nodes, assign a value of 0.2; if there are ≥1 diversion nodes, assign a value of 0.8; multiply the assigned values ​​of the number of independent paths and the number of diversion nodes by 0.5 respectively and then add them together to obtain the passage evacuation fault tolerance coefficient.

[0109] The passage safety resilience coefficient is obtained by integrating the passage-train spatial association table, the smoke-vehicle passage threat weight value, and the passage evacuation tolerance coefficient. Specifically, this involves finding the fire protection attribute of the passage from the passage-train spatial association table. If it is a fireproof passage, a value of 0.9 is assigned; otherwise, a value of 0.4 is assigned. The smoke-vehicle passage threat weight value is added to the fire protection attribute value of the passage and then divided by 2 to obtain the mean value. The mean value is then added to the passage evacuation tolerance coefficient and divided by 2 to obtain the passage safety resilience coefficient.

[0110] In one embodiment, the dynamic smoke cloud evolution body is analyzed in relation to the three-dimensional spatial structure, and a channel safety resilience coefficient is assigned to each three-dimensional space to generate a dynamic risk avoidance and passage map. The method also includes:

[0111] The channel safety resilience coefficient is dynamically corrected in two dimensions based on dynamic smoke cloud evolution data and train status data, generating a dual-disturbance dynamic safety resilience coefficient. Specifically, this involves: dividing the number of grids marked as diffusion areas in the dynamic smoke cloud evolution by the total number of grids in the channel to obtain the smoke coverage percentage; if the smoke coverage percentage is ≤30%, 1 is used as the smoke cloud correction term; if 30% < smoke coverage percentage ≤60%, 0.9 is used as the smoke cloud correction term; if the smoke coverage percentage is >60%, 0.8 is used as the smoke cloud correction term; dividing the train speed by the spatial correlation distance to obtain the ratio; normalizing the ratio to 0-1 to obtain the normalized ratio; subtracting the normalized ratio from 1 to obtain the train correction term; multiplying the channel safety resilience coefficient by the smoke cloud correction term and then by the train correction term, and normalizing the result to 0-1 to obtain the dual-disturbance dynamic safety resilience coefficient.

[0112] Based on the three-dimensional spatial structure, the dynamic safety resilience coefficient of dual disturbances is labeled to each channel to generate a dynamic hazard avoidance and passage map. Specifically, this includes: obtaining the three-dimensional coordinates and channel ID of each channel in the three-dimensional spatial structure; establishing a channel ID-grid coordinate mapping table to ensure that each channel's grid corresponds to a unique coordinate; then matching the dynamic safety resilience coefficient of each channel to all grid coordinates of that channel in the mapping table, with each grid coordinate labeled with the corresponding dynamic safety resilience coefficient; setting visual rules for the dynamic safety resilience coefficient of dual disturbances: green for a coefficient of 0.8-1.0, yellow for a coefficient of 0.5-0.7, and red for a coefficient of <0.5; and finally, positioning and labeling according to grid coordinates to generate a dynamic hazard avoidance and passage map.

[0113] This technical solution obtains the smoke concentration and diffusion rate of the passage at each time point through a dynamic smoke cloud evolution body, combines the spatial correlation distance of the passage-train spatial correlation table, generates an evacuation tolerance coefficient based on the number of independent paths and diversion nodes in the passage, and incorporates fire protection attribute assignment for fusion calculation. It comprehensively considers the dynamic changes of smoke caused by piston wind, the spatial influence of train and passage, and the passage's own evacuation capacity, which can more realistically reflect the safety level of the passage under unsteady flow field and avoid misjudgment of passage safety due to a single evaluation dimension.

[0114] By generating a smoke cloud correction term based on the smoke coverage ratio of the dynamic smoke cloud evolution body, and combining the train speed with the spatial correlation distance to generate a train correction term, the safety resilience coefficient of the passage is dynamically adjusted to ensure that the coefficient matches the dual interference of smoke diffusion and train piston wind in real time. At the same time, through passage ID-grid coordinate mapping, the dual-disturbance dynamic safety resilience coefficient is marked with green, yellow and red colors to generate a dynamic hazard avoidance passage map, allowing evacuees to quickly identify safe passages and avoid route failures caused by sudden changes in smoke direction or failure to update train influence.

[0115] In one embodiment, the three-dimensional coordinates and dynamic evacuation and passage map are analyzed to generate a three-dimensional escape route, including:

[0116] The system correlates three-dimensional coordinates with a dynamic evacuation and passage map, and combines this with a dynamic smoke cloud evolution model to adjust the regional smoke situation adaptation weight value for the dual-disturbance dynamic safety resilience coefficient. Specifically, this involves: defining a 0.5m × 0.5m × 0.5m cube region centered on the three-dimensional coordinates of the personnel to be evacuated; determining all grids corresponding to this region in the dynamic evacuation and passage map; extracting the number of grids with a final smoke concentration > 0.3 from the dynamic smoke cloud evolution model; dividing this number by the total number of grids in the region to obtain the smoke situation percentage; if the smoke situation percentage is ≤ 20%, the smoke situation coefficient is 1; if 20% < smoke situation percentage ≤ 40%, the smoke situation coefficient is 0.9; and if the smoke situation percentage is > 40%, the smoke situation coefficient is 0.8. The system also calculates the mean of the dual-disturbance dynamic safety resilience coefficient for all grids in the region, multiplies the mean by the smoke situation coefficient, and normalizes it to 0-1 to obtain the regional smoke situation adaptation weight value.

[0117] Real-time acquisition of the number of safety exits at the target location and the number of people at each safety exit;

[0118] Based on the number of people, the number of safety exits, and the three-dimensional coordinates of the people to be evacuated, the dynamic evacuation efficiency value of each safety exit is calculated. Specifically, this includes: calculating the difference between the three-dimensional coordinates of the people to be evacuated and the three-dimensional coordinates of each safety exit (the difference in each dimension); squaring each of the three dimensions and summing them; calculating the square root of the sum to obtain the straight-line distance from the three-dimensional position of the person to be evacuated to the three-dimensional position of the safety exit; normalizing the straight-line distance to 0-1; subtracting the normalized straight-line distance from 1 to obtain the distance fit coefficient; dividing the current number of people at each safety exit by the maximum capacity to obtain the congestion level; subtracting the congestion level from 1 to obtain the congestion fit coefficient; taking the reciprocal of the number of safety exits to obtain the exit allocation weight; multiplying the distance fit coefficient by the congestion fit coefficient and then by the exit allocation weight, and normalizing the product to 0-1 to obtain the dynamic evacuation efficiency value of each safety exit.

[0119] Using regional smoke condition adaptation weight as the core constraint and exit dynamic evacuation efficiency as the optimization objective, path search is performed in the channels of the dynamic evacuation passage map to generate a three-dimensional escape route with optimal safety and efficiency. Specifically, this involves: multiplying the dual-disturbance dynamic safety resilience coefficient of each channel in the dynamic evacuation passage map by the regional smoke condition adaptation weight and the exit dynamic evacuation efficiency, and then adding the two products to obtain the channel comprehensive weight; starting from the grid where the evacuees are located, searching for all paths to each safe exit, adding the channel comprehensive weights of each path to obtain the total channel comprehensive weight, and taking the path with the largest total weight as the three-dimensional escape route with optimal safety and efficiency.

[0120] By defining a cubic area centered on the three-dimensional coordinates of the people to be evacuated, and combining the real-time smoke concentration extracted from the dynamic smoke cloud evolution body to calculate the proportion of smoke, the dynamic safety resilience coefficient of the dual-disturbance system is adapted and adjusted. The generated regional smoke adaptation weight value can match the instantaneous changes in smoke diffusion in real time. The distance from the people to be evacuated to the exit, the congestion of the exit, and the number of exits are considered to calculate the dynamic evacuation efficiency value of the exit. Finally, the optimal path is searched through the comprehensive weight of the passage to ensure that the three-dimensional escape route adapts to the dynamic smoke and takes into account the evacuation efficiency. This solves the problem of route failure caused by sudden changes in the direction of smoke diffusion and facilitates the safe and efficient evacuation of people in subway stations.

[0121] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for generating dynamic three-dimensional escape routes based on the ignition point of a fire, characterized in that, include: Step S1: Real-time acquisition of temperature data, smoke data, train status data and three-dimensional spatial structure within the target location; analysis of preprocessed temperature data, smoke data and train data to obtain the three-dimensional coordinates of the ignition point and the fire-train coupling risk factor, including: dynamic spatiotemporal domain division of preprocessed temperature data, smoke data and train status data, taking the area centered on the real-time position of the train as the dynamic interference zone and the remaining areas as the static reference zone, generating a dynamic-static domain data division matrix; Based on the dynamic-static domain data partitioning matrix, the train disturbance correction value of temperature data in the dynamic interference zone and the natural diffusion gradient value of temperature data in the static reference zone are calculated respectively. The train disturbance correction value is fused with the natural diffusion gradient value to generate a zoned temperature feature vector; Based on the temperature feature vector of each zone, the smoke data is analyzed to generate a smoke source weighting map; The relationship between running speed and acceleration in the train status data is analyzed, and the impact intensity of the train on the airflow in the dynamic interference zone during the train's entry and exit from the station is calculated to obtain the train aerodynamic disturbance coefficient. Using the extreme points of the zone temperature feature vector as initial coordinates, combined with the probability direction of the smoke source tracing weight map, and analyzing the spatial connectivity characteristics of each region in the three-dimensional spatial structure, the three-dimensional coordinates of the ignition point and the corresponding initial fire intensity are obtained. The initial fire intensity is combined with the train aerodynamic disturbance coefficient to obtain the dynamic spread risk value; Analyze the spatial conflict probability between the fire ignition point and the train's trajectory to generate a fire-train coupling risk factor. The target location is a subway station; Step S2: Based on train status data, analyze the airflow change pattern of the target location to obtain a three-dimensional transient wind flow field intensity map covering the entire space of the target location; Step S3: Combine the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map to generate a dynamic smoke cloud evolution body in three-dimensional space. Step S4: Analyze the dynamic smoke cloud evolution body and the three-dimensional spatial structure, assign a channel safety resilience coefficient to each three-dimensional space, and generate a dynamic risk avoidance passage map. Step S5: Obtain the three-dimensional coordinates of the people to be evacuated in real time, analyze the three-dimensional coordinates and dynamic evacuation and passage map, and generate a three-dimensional escape route.

2. The method for generating dynamic three-dimensional escape routes based on the ignition point according to claim 1, characterized in that, Based on train status data, the airflow variation patterns at the target location are analyzed to obtain a three-dimensional transient wind flow field intensity map covering the entire space of the target location, including: Based on the dynamic-static domain data partitioning matrix, the spatial range of the dynamic interference zone and the static reference zone is analyzed, and a domain-related airflow analysis table is generated. Based on the domain-related airflow analysis table, the train wake wind speed characteristics in the dynamic interference zone are calculated, and the airflow attenuation coefficient in the static reference zone is analyzed. The train wake wind speed characteristics and airflow attenuation coefficient are integrated to obtain the set of airflow characteristic parameters for each zone. Obtain the parameters of the fixed ventilation facilities in the three-dimensional spatial structure, and spatially constrain the set of characteristic parameters of the airflow in the partition based on the fixed ventilation facility parameters to obtain the structurally constrained airflow vector field; Based on the real-time position and speed of the train, the airflow influence range is defined according to the speed with the train position as the center, and the airflow disturbance intensity of each grid within the airflow influence range is calculated. Based on the reference wind speed of the structurally constrained airflow vector field in each grid, the airflow disturbance intensity of each grid is calculated with the reference wind speed to form a three-dimensional transient airflow field intensity map covering the entire space.

3. The method for generating dynamic three-dimensional escape routes based on the ignition point according to claim 2, characterized in that, Spatial constraints are applied to the set of characteristic airflow parameters in different zones based on fixed ventilation facility parameters to obtain a structurally constrained airflow vector field, including: The parameters of fixed ventilation facilities are analyzed, and a ventilation facility area constraint table is generated based on the spatial range covered by the fixed ventilation facilities. Spatially associate the ventilation facility area constraint table with the zone airflow characteristic parameter set, and adjust the airflow intensity and direction for areas affected by different airflows to generate a spatially adapted airflow correction set. Based on the structural characteristics of the target site in three-dimensional space, the spatially adapted airflow correction set is optimized to obtain the structurally constrained airflow vector field.

4. The method for generating dynamic three-dimensional escape routes based on the ignition point according to claim 2, characterized in that, By combining the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map, a dynamic smoke cloud evolution in three-dimensional space is generated, including: Using the three-dimensional coordinates of the ignition point as the origin, the energy of fire spread is analyzed by combining the fire-train coupling risk factor and dynamic diffusion risk value, and a dynamic correlation value between fire and train is generated. Based on the three-dimensional transient wind flow field intensity map, the dynamic correlation value between fire and train is transiently corrected to generate a dynamic wind field adaptive diffusion set. The system acquires the number of people in the target location in real time, analyzes the degree of obstruction of smoke diffusion by the area where people gather based on the number of people, and generates a people density obstruction coefficient. The dynamic wind field adaptive diffusion set is corrected based on the personnel density hindrance coefficient to obtain the dual-disturbance diffusion set for people and vehicles.

5. The method for generating dynamic three-dimensional escape routes based on the ignition point according to claim 4, characterized in that, The system combines the three-dimensional coordinates of the ignition point, the fire-train coupling risk factor, and the three-dimensional transient wind flow field intensity map to generate a dynamic smoke cloud evolution in three-dimensional space. This also includes: Based on the human-vehicle dual-perturbation corrected diffusion set, the correlation between the concentration change and diffusion rate of smoke at each spatial point is analyzed, and the dual-perturbation coupled concentration feedback coefficient is generated. Based on the three-dimensional transient wind flow field intensity map and the dual-perturbation coupled concentration feedback coefficient, the smoke distribution, concentration and diffusion area at each time node are analyzed to generate a trend diffusion optimization set; By analyzing the trend diffusion optimization set, the double-perturbation coupling concentration feedback coefficient, and the personnel density hindrance coefficient, a dynamic smoke cloud evolution body in three-dimensional space is generated.

6. The method for generating dynamic three-dimensional escape routes based on the ignition point according to claim 5, characterized in that, By analyzing the dynamic evolution of smoke clouds and their three-dimensional spatial structure, a channel safety resilience coefficient is assigned to each three-dimensional space, generating a dynamic risk avoidance and passage map, including: Analyze the three-dimensional spatial structure of the target site, analyze the spatial relationship between the physical properties of each passage and the train trajectory, and generate a passage-train spatial relationship table; Based on the dynamic smoke cloud evolution, the concentration and diffusion rate of smoke in the channel at each time point are calculated. Combined with the channel-train spatial correlation table, the actual threat level of smoke to the channel is calculated, and the smoke-train channel threat weight value is generated. Analyze the evacuation function characteristics of channels in a three-dimensional spatial structure and generate the channel evacuation fault tolerance coefficient; The safety resilience coefficient of the passage is obtained by integrating the passage-train spatial association table, the smoke-vehicle passage threat weight value, and the passage evacuation fault tolerance coefficient.

7. The method for generating dynamic three-dimensional escape routes based on the ignition point according to claim 6, characterized in that, The dynamic smoke cloud evolution body is analyzed in relation to the three-dimensional spatial structure. A channel safety resilience coefficient is assigned to each three-dimensional space, generating a dynamic risk avoidance and passage map. This also includes: Based on the dynamic smoke cloud evolution and train status data, the channel safety resilience coefficient is dynamically corrected in two dimensions to generate a dual-disturbance dynamic safety resilience coefficient. Based on the three-dimensional spatial structure, the dynamic safety resilience coefficient of the dual-disturbance system is marked to each channel to generate a dynamic risk avoidance and passage map.

8. The method for generating dynamic three-dimensional escape routes based on the ignition point according to claim 7, characterized in that, Analyzing the three-dimensional coordinates and dynamic hazard avoidance map, three-dimensional escape routes are generated, including: The three-dimensional coordinates are correlated with the dynamic risk avoidance and passage map, and combined with the dynamic smoke cloud evolution body, the regional smoke situation adaptation adjustment of the dual-disturbance dynamic safety resilience coefficient is carried out to generate regional smoke situation adaptation weight value. Real-time acquisition of the number of safety exits at the target location and the number of people at each safety exit; Based on the number of people, the number of safety exits, and the three-dimensional coordinates of the people to be evacuated, the dynamic evacuation efficiency value of each safety exit is calculated. Using regional smoke condition matching weight as the core constraint and exit dynamic evacuation efficiency as the optimization target, a path search is performed in the channel of the dynamic risk avoidance passage map to generate a three-dimensional escape route with safety and efficiency.

Images