A building fire evacuation analysis method and system based on BIM

Abstract

The embodiment of the application discloses a building fire evacuation analysis method and system based on BIM. The technical scheme provided by the embodiment of the application can significantly improve the environmental adaptability of the evacuation path, improve the fire evacuation path planning accuracy, ensure that the evacuation path can respond to the fire evolution, always avoid the high-temperature area, the dense smoke area, the structure dangerous area and the crowded node, greatly reduce the path guidance failure risk caused by the change of disaster situation, and improve the success rate of escape.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a BIM-based method and system for analyzing building fire evacuation. Background Technology

[0002] Currently, in fire prevention and control scenarios, BIM (Building Information Modeling) models are frequently used as the dynamic decision-making center for fire evacuation. By integrating information such as building spatial topology and sensor data, a high-precision digital twin environment is constructed. When a fire occurs, sensor information can detect the fire's location, which is then imported into the BIM model, enabling a visual representation of the fire alarm. Furthermore, by combining the building spatial topology of the BIM model, fire evacuation routes can be constructed in real time, allowing for fire evacuation guidance based on these routes.

[0003] However, current fire evacuation route construction schemes based on BIM models use static route planning, which cannot dynamically respond to the evolution of a fire, resulting in significant errors in the planned routes and affecting the effectiveness of fire evacuation guidance. Summary of the Invention

[0004] This application provides a BIM-based building fire evacuation analysis method and system, which can dynamically plan fire evacuation routes according to the fire evolution process, improve the accuracy of fire evacuation route planning, and solve the problem of fire evacuation route planning error.

[0005] In a first aspect, embodiments of this application provide a BIM-based building fire evacuation analysis method, including:

[0006] The BIM model of the corresponding monitoring area collects temperature data, smoke data and flame monitoring data at each path monitoring location. Based on the temperature data, smoke data and flame monitoring data, fire stage analysis is performed to generate fire stage determination results. The fire stages include predefined initial stage, development stage and decay stage.

[0007] The BIM model of the corresponding monitoring area collects crowd density information at each path monitoring location. Based on the temperature data and the flame monitoring data, the structural damage information of the corresponding path monitoring location is calculated. Based on the fire stage determination result, smoke cost weight, structural damage cost weight and crowd density cost weight are configured. Based on the smoke data, the structural damage information, the crowd density information, the smoke cost weight, the structural damage cost weight and the crowd density cost weight, the path cost of the corresponding path monitoring location is calculated. The path cost represents the access difficulty of the corresponding path monitoring location.

[0008] Based on the path cost, a real-time evacuation route is planned, and the real-time evacuation route is imported into a BIM model for visualization.

[0009] Furthermore, the formula for calculating the path cost is expressed as follows:

[0010]

[0011] Wherein, Loss represents path cost, L_smoke represents the smoke data, w_smoke represents the smoke cost weight, L_structure represents the structural damage information, w_structure represents the structural damage cost weight, L_density represents the crowd density information, and w_density represents the crowd density cost weight.

[0012] Further, the step of planning real-time evacuation routes based on the path cost includes:

[0013] Based on the monitoring locations of each path, the distance between the corresponding road segments is determined, and the A* algorithm is used to plan real-time evacuation routes according to the road segment distance and the path cost.

[0014] Furthermore, the step of planning real-time evacuation routes using the A* algorithm based on the road segment distance and the path cost includes:

[0015] Based on the distance of the road segments, the A* algorithm is used to plan a set number of candidate paths with the shortest distance;

[0016] The total cost of each candidate path is calculated based on the path cost, and the candidate path with the minimum total cost is selected as the real-time evacuation path.

[0017] Furthermore, after importing the real-time evacuation routes into the BIM model for visualization, the process also includes:

[0018] If a change in the fire stage is detected, the real-time evacuation routes are updated.

[0019] If the path monitoring location corresponding to the real-time evacuation path is detected to be in an abnormal state, the real-time evacuation path is updated. The abnormal state indicates that the path cost at the current path monitoring location has reached a set cost threshold.

[0020] Furthermore, after calculating the structural damage information for the corresponding path monitoring location based on the temperature data and the flame monitoring data, the method further includes:

[0021] Collect monitoring image information of the corresponding path monitoring location, identify specified image features based on the monitoring image information, generate correction coefficients based on the specified image features, and correct the structural damage information based on the correction coefficients.

[0022] Furthermore, the step of performing fire stage analysis based on the temperature data, the smoke data, and the flame monitoring data to generate a fire stage determination result includes:

[0023] The temperature data, smoke data, and flame monitoring data are input into a pre-built LSTM network. Based on the LSTM network, the probability distribution information of the fire stage for a specified period in the future is predicted. Based on the probability distribution information of the fire stage, a fire stage determination result is generated.

[0024] In a second aspect, embodiments of this application provide a BIM-based building fire evacuation analysis system, comprising:

[0025] The stage determination module is used to collect temperature data, smoke data and flame monitoring data at each path monitoring location in the BIM model of the corresponding monitoring area, perform fire stage analysis based on the temperature data, smoke data and flame monitoring data, and generate fire stage determination results. The fire stages include predefined initial stage, development stage and decay stage.

[0026] The cost calculation module is used to collect crowd density information at each path monitoring location from the BIM model of the corresponding monitoring area, calculate the structural damage information at the corresponding path monitoring location based on the temperature data and the flame monitoring data, configure smoke cost weight, structural damage cost weight and crowd density cost weight according to the fire stage determination result, and calculate the path cost at the corresponding path monitoring location based on the smoke data, the structural damage information, the crowd density information, the smoke cost weight, the structural damage cost weight and the crowd density cost weight. The path cost represents the access difficulty of the corresponding path monitoring location.

[0027] The path generation module is used to plan real-time evacuation routes based on the path costs and import the real-time evacuation routes into the BIM model for visualization.

[0028] In a third aspect, embodiments of this application provide an electronic device, including:

[0029] Memory and one or more processors;

[0030] The memory is used to store one or more programs;

[0031] When the one or more programs are executed by the one or more processors, the one or more processors implement the BIM-based building fire evacuation analysis method as described in the first aspect.

[0032] In a fourth aspect, embodiments of this application provide a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform the BIM-based building fire evacuation analysis method as described in the first aspect.

[0033] This application embodiment collects temperature, smoke, and flame monitoring data at each path monitoring location using a BIM model of the corresponding monitoring area. Based on these data, a fire stage analysis is performed to generate a fire stage determination result, which includes a predefined initial stage, development stage, and decay stage. The BIM model of the corresponding monitoring area also collects crowd density information at each path monitoring location. Based on the temperature and flame monitoring data, structural damage information at each path monitoring location is calculated. Smoke cost weights, structural damage cost weights, and crowd density cost weights are configured according to the fire stage determination result. Path costs at each path monitoring location are calculated based on the smoke data, structural damage information, crowd density information, smoke cost weights, structural damage cost weights, and crowd density cost weights. Path costs characterize the accessibility of the corresponding path monitoring location. Real-time evacuation routes are planned based on the path costs and imported into the BIM model for visualization. By employing the aforementioned technical means, adjusting path costs according to different fire stages in the fire evolution process, and planning real-time evacuation routes based on path costs, the environmental adaptability of evacuation routes can be significantly improved, the accuracy of fire evacuation route planning can be enhanced, and the evacuation routes can be ensured to respond to fire evolution, always avoiding high-temperature zones, dense smoke zones, structurally hazardous zones, and congested nodes. This significantly reduces the risk of path guidance errors caused by changes in the disaster situation and improves the escape success rate. Attached Figure Description

[0034] Figure 1 This is a flowchart of a BIM-based building fire evacuation analysis method provided in Embodiment 1 of this application;

[0035] Figure 2 This is a flowchart of the dynamic path planning in Embodiment 1 of this application;

[0036] Figure 3 This is a flowchart illustrating the generation of real-time evacuation routes in Embodiment 1 of this application;

[0037] Figure 4 This is a flowchart of the real-time evacuation path update process in Embodiment 1 of this application;

[0038] Figure 5This is a schematic diagram of a BIM-based building fire evacuation analysis system provided in Embodiment 2 of this application;

[0039] Figure 6 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of this application. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0041] Example 1:

[0042] Figure 1 A flowchart of a BIM-based building fire evacuation analysis method according to Embodiment 1 of this application is provided. The BIM-based building fire evacuation analysis method provided in this embodiment can be executed by a BIM-based building fire evacuation analysis device. This BIM-based building fire evacuation analysis device can be implemented through software and / or hardware. The device can consist of two or more physical entities, or it can consist of a single physical entity. Generally, the BIM-based building fire evacuation analysis device can be a backend server, a computer, or other computing device.

[0043] The following description uses a server as the primary entity executing a BIM-based building fire evacuation analysis method. (Refer to...) Figure 1 The BIM-based building fire evacuation analysis method specifically includes:

[0044] S110. The BIM model of the corresponding monitoring area collects temperature data, smoke data and flame monitoring data at each path monitoring location. Based on the temperature data, smoke data and flame monitoring data, fire stage analysis is performed to generate fire stage judgment results. The fire stages include the predefined initial stage, development stage and decay stage.

[0045] When performing building fire evacuation analysis, this application uses relevant sensors installed in the monitored building to detect fire alarms. When a fire alarm is detected or the fire situation meets the set conditions, it triggers the dynamic generation of real-time evacuation routes.

[0046] Specifically, the raw data (temperature, smoke concentration, flame presence / intensity) collected by sensors at various path monitoring locations in the BIM model of the corresponding monitoring area is transformed into a macroscopic understanding of the current fire state (initial, developing, decaying). This allows for the dynamic generation of real-time evacuation routes based on the different fire stage assessments. Flame monitoring data can be acquired through image and infrared methods, with flame image area representing its intensity. This application does not impose fixed restrictions on the acquisition methods of temperature data, smoke data, and flame monitoring data, which will not be elaborated upon here.

[0047] During data collection, temperature sensors (thermocouples, infrared), smoke detectors (photoelectric, ionization), and flame detectors (ultraviolet, infrared, visual sensor recognition, etc.) deployed at critical path monitoring locations in the building (such as corridor intersections, stairwells, room entrances, etc.) continuously transmit data to the backend server via the network. This data is aligned and calibrated in terms of timestamps and spatial location (correlated with spatial coordinates in the BIM model) to eliminate sensor errors and environmental interference, thus enabling the determination of the fire stage based on the collected sensor data.

[0048] In the initial stage of a fire, the temperature in a specific area (near the fire source) may exceed the ambient temperature threshold (e.g., 60°C), but the rise is gradual (i.e., the rate of temperature change is below the threshold), the smoke concentration reaches the alarm threshold but the diffusion range is limited (i.e., the amount or area of ​​smoke detected is below the threshold), and the flame monitoring data (e.g., thermal radiation intensity) is below the threshold or is only detected in the corresponding area.

[0049] The significant characteristics of the fire development stage are a rapid increase in temperature (e.g., temperature rise rate > 5°C / min), a rapid increase in smoke concentration (concentration change rate reaches the corresponding threshold) and an expansion of the diffusion range (e.g., the number of sensors triggering alarms reaches the set threshold), a significant enhancement of flame monitoring data (intensity and area reach the corresponding threshold), and these phenomena show a clear trend of spatial spread (i.e., the number of sensors detecting the corresponding data and the area increase).

[0050] During the fire decay phase, the fire weakens due to the reduction or suppression of combustibles. Fire characteristics include a slow decline or stabilization at a high temperature after reaching its peak, a decrease in smoke concentration (but less new smoke), and a reduction in flame intensity or range.

[0051] Based on the characteristics of temperature, smoke, and flame monitoring data at different stages, and by pre-setting different thresholds or value ranges, the current stage of the fire can be determined by comparing the indicator data. By fusing multi-source sensor data in real time, the current macroscopic evolution stage of the fire (initial, development, decay) can be accurately identified, providing a core basis for the dynamic weighting of subsequent path costs.

[0052] Optionally, fire stage analysis is performed based on temperature data, smoke data, and flame monitoring data to generate fire stage determination results, including:

[0053] Temperature data, smoke data, and flame monitoring data are input into a pre-built LSTM network. The LSTM network is used to predict the probability distribution of fire stages for a specified period in the future, and the fire stage determination result is generated based on the probability distribution of fire stages.

[0054] This application introduces an LSTM network into fire stage determination to achieve accurate and predictive fire stage determination.

[0055] During the LSTM network training phase, a historical dataset covering the entire fire cycle is first constructed, comprising a three-dimensional time-series data stream containing temperature, smoke, and flame monitoring data, with a uniform sampling frequency of 1Hz. The collected data stream undergoes timestamp alignment (to address clock skew from multiple sources), outlier repair (e.g., cubic spline interpolation to correct momentary loss due to flame obstruction), and normalization (temperature mapped to 0-1 based on a 0-1000°C range; smoke concentration linearly scaled to 0-1 by 0-100% obs / m; flame monitoring data also mapped to 0-1 by area). Then, in the feature engineering phase, key dynamic indicators are extracted, such as the values ​​of temperature, smoke, and flame monitoring data, as well as their rates of change and accelerations. Finally, according to relevant fire safety standards, different feature data are categorized into corresponding fire stage labels (initial, development, and decay). The LSTM network employs a three-layer 128-unit hidden layer. The input layer receives a 9-dimensional vector (original temperature / smoke / flame values ​​+ 6 derived features) over a 60-second time window. The output layer is a softmax-activated ternary classification (early / developing / declining). During training, an Adam optimizer with Nesterov momentum (learning rate 0.001) is used, with a weighted cross-entropy loss function (to address phase imbalance) and an early stopping strategy (patience=10) to prevent overfitting.

[0056] Subsequently, during LSTM network deployment, real-time sensor data streams (temperature data, smoke data, and flame monitoring data) undergo the same preprocessing and feature derivation (features such as rate of change and acceleration of change) via edge computing nodes, forming a sliding time window (lasting 60 seconds, updated every second). The LSTM network processes this window data through forward propagation: the input gate filters key features (such as abrupt temperature gradients), the forget gate clears expired states (such as low-temperature records from 5 minutes ago), the memory gate updates cell states (accumulating evidence of fire evolution), and the output gate generates hidden states that are passed to the next time step. The final hidden states are mapped to a probability distribution through a fully connected layer. For example, the output vector [0.18, 0.75, 0.07] represents an initial probability of 18%, a development probability of 75%, and a decay probability of 7% within the next 30 seconds. The system takes the stage corresponding to the highest probability as the judgment result. A confidence threshold (>60%) can also be set to avoid false positives due to low certainty. If the highest probability does not reach the threshold, the judgment result from the previous cycle is maintained. This result interface is used for subsequent weight configuration.

[0057] By using LSTM to deeply model the temporal dynamics of multi-source sensors, the system can accurately predict the stage transition (such as the initial stage → development stage) within a short period of time after a fire occurs, respond in advance to changes in the fire's intensity, and provide timely, accurate, and predictive stage determination results for the dynamic adjustment of evacuation routes.

[0058] S120. Collect crowd density information at each path monitoring location using the BIM model of the corresponding monitoring area. Calculate structural damage information at the corresponding path monitoring location based on temperature data and flame monitoring data. Configure smoke cost weight, structural damage cost weight, and crowd density cost weight based on the fire stage determination results. Calculate the path cost at the corresponding path monitoring location based on smoke data, structural damage information, crowd density information, smoke cost weight, structural damage cost weight, and crowd density cost weight. The path cost represents the accessibility difficulty at the corresponding path monitoring location.

[0059] Furthermore, based on the fire stage determination results mentioned above, accurate path cost calculations can be performed using the BIM model. In the BIM model, path cost represents a comprehensive, quantified accessibility score for each monitoring location (node ​​or segment) of an evacuation route (such as corridors, staircases, and doors). By dynamically integrating three core risk factors affecting accessibility (smoke, structural damage to the building, and crowd density) and assigning them weights matching the fire stage, accurate path cost calculations can be achieved.

[0060] For smoke data, the smoke concentration values ​​collected by the sensor can be used directly and normalized (range 0-1).

[0061] Crowd density information is obtained by analyzing surveillance video to detect targets and determine the number of people in different areas, thereby estimating the real-time number of people in a specific area (corresponding to the monitoring location along the path), calculating the number of people per unit area or the degree of congestion, and normalizing the data.

[0062] Structural damage information for buildings can be obtained using temperature data (especially high temperature) and flame monitoring data (direct burning). Furthermore, the BIM model pre-stores or associates material information (concrete grade, steel type, wood type, etc.) and key thermodynamic parameters (such as coefficient of thermal expansion, strength reduction factor, and fire resistance limit) of building components (walls, columns, floors, doors).

[0063] Simplified empirical formulas are used to calculate structural damage information. For example, the strength of concrete decreases significantly at high temperatures, steel softens and buckles, and areas directly burned by flames experience accelerated damage. Therefore...

[0064] Based on the material and temperature data collected at the corresponding path monitoring location, a pre-defined mapping table of temperature, material, and material reduction factor can be consulted to determine the normalized material reduction factor λ. Furthermore, according to actual needs, this application also introduces flame data of flame area and duration to correct the material reduction factor λ. The larger the flame data, the greater the impact on the material reduction factor λ, and the larger the value of the material reduction factor λ. Finally, a material reduction factor λ reflecting the impact of the fire on that location (e.g., 0 indicates safety, 1 indicates complete failure), i.e., structural damage information, is output and normalized.

[0065] Based on the fire stage determination results obtained in the previous step, the server calls the predefined weight configuration rules to configure the smoke cost weight, structural damage cost weight, and crowd density cost weight for the corresponding stage.

[0066] For example, in the initial stages, the fire area is small and the structural risk is low, but smoke can be generated rapidly. Weights might prioritize crowd density (to avoid panic and congestion) and smoke (high initial inhalation risk), with lower weights for structural damage. For example: w_smoke=0.5, w_density=0.4, w_structure=0.1;

[0067] During the development phase, the fire spreads rapidly, and the risks of high temperatures and structural failure increase dramatically, with smoke billowing everywhere. Weights are heavily weighted towards structural damage and smoke, while the weight of population density is relatively reduced (but still represents congestion risk). For example: w_smoke=0.4, w_density=0.2, w_structure=0.4;

[0068] During the decay phase, the fire intensifies, but the structure may have been severely damaged (ember embers, high temperatures) or partially collapsed, and smoke may still linger. Weighting focuses on structural damage (assessing residual risk) and smoke (residual hazards), while the weighting of population density increases (as people may still gather in the later stages of evacuation). For example: w_smoke=0.3, w_density=0.2, w_structure=0.5.

[0069] Based on the aforementioned smoke data, structural damage information, crowd density information, smoke cost weight, structural damage cost weight, and crowd density cost weight, the corresponding path cost can be calculated according to the established calculation rules. The formula for calculating the path cost is as follows:

[0070]

[0071] Where Loss represents path cost, L_smoke represents smoke data, w_smoke represents smoke cost weight, L_structure represents structural damage information, w_structure represents structural damage cost weight, L_density represents crowd density information, and w_density represents crowd density cost weight.

[0072] A higher path cost indicates greater difficulty / risk in traversing that location. This cost can be updated in real time to reflect changes brought about by the dynamic evolution of the fire and the movement of people. Path costs can be calculated using adaptive calculation rules based on the above information, according to actual needs. This application does not impose fixed restrictions on specific calculation rules, which will not be elaborated upon here.

[0073] By calculating structural damage risk in real time, dynamically integrating population density information, and intelligently adjusting the weights of various risk factors according to the fire stage, a multi-dimensional dynamic cost field that accurately reflects the real-time access difficulty and risk distribution inside the building can be generated, thus providing highly contextualized input for path planning.

[0074] Optionally, after calculating the structural damage information at the corresponding path monitoring location based on temperature data and flame monitoring data, the method further includes:

[0075] Collect monitoring image information of the corresponding path monitoring location, identify specified image features based on the monitoring image information, generate correction coefficients based on the specified image features, and correct structural damage information based on the correction coefficients.

[0076] This application, building upon existing temperature and flame data-driven structural damage calculations, introduces monitoring image correction to enhance the accuracy of structural damage analysis. First, real-time video streams are acquired using high-definition visible light cameras and thermal imagers pre-positioned at monitoring locations (such as corridor corners, stair landings, and near load-bearing components). Then, physical features directly related to structural damage are extracted using a computer vision model. For example, macroscopic deformations of beams and columns (such as bending angles and displacements) are identified by comparing the background subtraction method with the baseline contour of the BIM model. A semantic segmentation network is used to segment the spalling areas on the concrete surface and calculate the area percentage. Based on these visual features, a machine learning model identifies the mapping relationship between these features and measured structural damage information (e.g., a 10% increase in concrete spalling area corresponds to a 15% reduction in building strength), ultimately outputting a corresponding dynamic correction coefficient. This correction coefficient is then used to correct the structural damage information. By capturing building physical deformation and material degradation phenomena in real time through visual features, the limitations of indirect inference from pure sensor models can be overcome, upgrading structural damage assessment from "probabilistic prediction" to "empirical calibration." This avoids misjudgments caused by temperature measurement deviations, flame obstruction, or material anomalies, achieving higher-precision structural damage analysis.

[0077] S130. Plan real-time evacuation routes based on path costs, and import the real-time evacuation routes into the BIM model for visualization.

[0078] Furthermore, referring to Figure 3 Based on the determined path costs, this application combines the BIM model of the current monitoring area to perform dynamic path planning. Specifically, using real-time updated path cost data, the optimal or feasible real-time evacuation path is dynamically calculated for evacuees or groups within a precise spatial topology network constructed from the BIM model (nodes represent path monitoring locations, edges represent paths / corridors / staircases), and the results are presented intuitively on the BIM visualization platform. For example, the path with the minimum total path cost is planned using a path planning algorithm as the real-time evacuation path, thereby ensuring that the real-time evacuation path has low passage difficulty.

[0079] Optionally, real-time evacuation routes are planned based on path costs, including:

[0080] Based on the monitoring locations of each path, the distance between the corresponding road segments is determined, and the A* algorithm is used to plan real-time evacuation routes according to the road segment distance and path cost.

[0081] In the implementation of the A* algorithm for real-time evacuation route planning, this application first converts the BIM model into a topological path mesh, with nodes corresponding to path monitoring locations (room exits, corridor intersections), and edges associated with precise geometric distances (extracted from the BIM spatial database). Then, based on the A* algorithm, it searches for the N shortest paths and selects the path with the minimum cumulative total cost as the real-time evacuation route. There are many implementations of the A* algorithm for optimal path search, which will not be elaborated upon here.

[0082] Furthermore, based on actual needs, the A* algorithm can be used to plan the minimum cost path, the minimum distance path, and the optimal alternative path that combines cost and distance. These different paths are then visualized on the BIM model using different colors for users to select from.

[0083] Optionally, the server can also identify high-risk road segments based on path cost to further alert users to the difficulty of traversing the corresponding road segments and improve evacuation safety.

[0084] Optionally, refer to Figure 3 This application uses the A* algorithm to plan real-time evacuation routes based on road segment distance and path cost, including:

[0085] S1301. Based on the road segment distance, use the A* algorithm to plan a set number of candidate paths with the shortest distance;

[0086] S1302. Calculate the total cost of each candidate path based on the path cost, and select the candidate path with the minimum total cost as the real-time evacuation path.

[0087] In the implementation of real-time evacuation route planning, candidate route generation based on geometric distance is performed first. The algorithm converts the BIM model into a topology map, where nodes represent the monitoring locations of key evacuation points, and physical passages between nodes (corridor sections, stairs, connecting corridors) are abstracted as edges. The distance of each edge is accurately extracted through the BIM model (e.g., corridor AB = 15.2m, vertical span of stairs BC = 4.5m). Subsequently, the K-shortest path A-star algorithm is used for multi-path search. The number of target paths is set during initialization (e.g., 5 paths). In the first round, the pure geometric distance is used as the cost function to search for the first shortest path (e.g., a direct path from the main corridor, 42m). Next, the path deviation engine is activated, the node sequence of the first path is frozen, and its key decision nodes (such as the corridor fork Node_X) are selected in sequence. The parent node of Node_X is forcibly switched to point to the adjacent unexplored node (such as replacing the original parent node Node_P with Node_Q). Based on the new parent node, the A-star search is re-executed to generate a topologically heterogeneous suboptimal path (such as detouring 47m around the west passage). This process is repeated until a set number of candidate paths (such as 5) are obtained. The spatial distribution diversity of the paths is ensured by the minimum node difference rate constraint (>30%) (such as including the atrium duplex path, the external corridor path, etc.). Finally, the set of candidate paths with increasing distances but all satisfying basic accessibility is output.

[0088] For each candidate path, the path cost at each monitoring location along that path is accumulated to obtain the total cost of different candidate paths. The candidate path with the lowest total cost is then selected as the real-time evacuation path. By combining distance-based initial screening with cost-based final selection in path planning, fire environmental risks and population distribution are dynamically integrated while ensuring physical accessibility of the path. This ensures that evacuation paths always balance efficiency and safety, thereby avoiding the systemic failure problem of traditional single-path planning under sudden risks.

[0089] Optionally, refer to Figure 4 After importing real-time evacuation routes into the BIM model for visualization, it also includes:

[0090] S1303. If a change in the stage of a fire is detected, update the real-time evacuation routes.

[0091] S1304. If the path monitoring location corresponding to the real-time evacuation path is detected to be in an abnormal state, the real-time evacuation path shall be updated. The abnormal state indicates that the path cost of the current path monitoring location has reached the set cost threshold.

[0092] After completing the BIM visualization of the initial path, this application also continuously runs dynamic path updates. Specifically, when the fire stage analysis detects a stage transition (such as from the initial stage to the development stage), the system immediately triggers a global reassessment, and then recalculates the path cost of all path nodes based on the new stage weight (such as increasing the structural damage weight from 0.1 to 0.4), and re-plans the real-time evacuation path based on the path cost.

[0093] Furthermore, when the real-time path cost at any monitored location reaches a set cost threshold, the server locates the high-risk node and initiates local path optimization. For example, a bidirectional A-star search is used to backtrack / look ahead from the anomaly point to both ends, bypassing the danger zone to generate the shortest detour segment (e.g., if the path cost of structural damage to the 3rd-floor east stairwell reaches the set cost threshold, personnel are guided to detour via the adjacent conference room). This allows for a local path update to be completed in a short time and pushed to users for display. For example, a flashing warning icon is pushed to the user terminal display screen on the affected floors.

[0094] By using an update mechanism that drives global updates during fire phases and triggers local optimizations at high-risk nodes, the system achieves high spatial accuracy in risk avoidance while ensuring real-time performance, enabling evacuation routes to dynamically adapt to different fire conditions.

[0095] Optionally, based on BIM visualization, the server can also overlay escape route indicators in real time via the AR module of a fire helmet or mobile terminal. For example, the device's camera can capture real-world images, and combined with SLAM technology, the user's location can be determined. The real-time evacuation route's 3D coordinates for the current floor can be obtained from the BIM model, generating directional arrows (green) and risk area warning boxes (red semi-transparent) overlaid on the real-world view. When a change in the fire stage triggers a route update, the AR arrows dynamically switch directions (e.g., a red cross is highlighted in the original corridor, guiding the user to a new exit), and the voice module simultaneously broadcasts turning instructions ("Turn left to enter the safety passage"), thereby solving the problem of getting lost due to insufficient visibility in smoke.

[0096] Optionally, the server can also dynamically convert the path data format according to the terminal type, push the whole building BIM risk heat map and path topology to the fire command center; send simplified vector paths (JSON coordinate point sequence + turning instructions) to the mobile phones of evacuees; and output directional icons and distances ("Safety exit ahead 50 meters ↑") to the embedded displays in the building to ensure reliable guidance in high-noise environments.

[0097] Optionally, after the fire stage is determined by the LSTM network, real-time sensor data can be used as state input to simulate the risk evolution of different path choices over the next 10 minutes (e.g., choosing corridor A may cause crowd gathering → triggering crush risk); thereby outputting the long-term survival probability score of each path, dynamically correcting the current path cost calculation (original cost × reciprocal of survival probability), and avoiding the algorithm from getting trapped in local optima.

[0098] Optionally, when the structural damage value of a path node on a certain layer reaches a threshold, a vertical risk transmission analysis can be initiated. For example, by extracting the load-bearing components connected to the upper and lower layers through the BIM model (e.g., a damaged column on the 3rd floor is connected to the floor on the 4th floor), the probability of damage transmission is calculated (the risk value of the foundation is increased in the area directly above the damaged column); the path cost of the corresponding node on the upper layer is updated, thereby achieving three-dimensional risk avoidance.

[0099] The above-mentioned methods involve collecting temperature, smoke, and flame monitoring data at each path monitoring location using the BIM model of the corresponding monitoring area, performing fire stage analysis based on the temperature, smoke, and flame monitoring data, and generating fire stage determination results. Fire stages include predefined initial, development, and decay stages. The BIM model of the corresponding monitoring area also collects crowd density information at each path monitoring location. Based on the temperature and flame monitoring data, structural damage information at the corresponding path monitoring location is calculated. Smoke cost weights, structural damage cost weights, and crowd density cost weights are configured according to the fire stage determination results. Path costs for the corresponding path monitoring locations are calculated based on smoke data, structural damage information, crowd density information, smoke cost weights, structural damage cost weights, and crowd density cost weights. Path costs characterize the accessibility of the corresponding path monitoring location. Real-time evacuation routes are planned based on the path costs and imported into the BIM model for visualization. By employing the aforementioned technical means, adjusting path costs according to different fire stages in the fire evolution process, and planning real-time evacuation routes based on path costs, the environmental adaptability of evacuation routes can be significantly improved, the accuracy of fire evacuation route planning can be enhanced, and the evacuation routes can be ensured to respond to fire evolution, always avoiding high-temperature zones, dense smoke zones, structurally hazardous zones, and congested nodes. This significantly reduces the risk of path guidance errors caused by changes in the disaster situation and improves the escape success rate.

[0100] Example 2:

[0101] Based on the above embodiments, Figure 5 This is a structural schematic diagram of a BIM-based building fire evacuation analysis system provided in Embodiment 2 of this application. (Reference) Figure 5 The BIM-based building fire evacuation analysis system provided in this embodiment specifically includes: a stage determination module 21, a cost calculation module 22, and a path generation module 23.

[0102] Among them, the stage determination module 21 is used to collect temperature data, smoke data and flame monitoring data of each path monitoring location from the BIM model of the corresponding monitoring area, perform fire stage analysis based on the temperature data, smoke data and flame monitoring data, and generate fire stage determination results. The fire stages include the predefined initial stage, development stage and decay stage.

[0103] The cost calculation module 22 is used to collect crowd density information at each path monitoring location from the BIM model of the corresponding monitoring area, calculate the structural damage information at the corresponding path monitoring location based on temperature data and flame monitoring data, configure smoke cost weight, structural damage cost weight and crowd density cost weight according to the fire stage judgment result, and calculate the path cost at the corresponding path monitoring location based on smoke data, structural damage information, crowd density information, smoke cost weight, structural damage cost weight and crowd density cost weight. The path cost represents the passage difficulty of the corresponding path monitoring location.

[0104] The path generation module 23 is used to plan real-time evacuation paths based on path costs and import the real-time evacuation paths into the BIM model for visualization.

[0105] Specifically, after calculating the structural damage information at the corresponding path monitoring location based on temperature data and flame monitoring data, the process also includes:

[0106] Collect monitoring image information of the corresponding path monitoring location, identify specified image features based on the monitoring image information, generate correction coefficients based on the specified image features, and correct structural damage information based on the correction coefficients.

[0107] Fire stage analysis is performed based on temperature data, smoke data, and flame monitoring data to generate fire stage determination results, including:

[0108] Temperature data, smoke data, and flame monitoring data are input into a pre-built LSTM network. The LSTM network is used to predict the probability distribution of fire stages for a specified period in the future, and the fire stage determination result is generated based on the probability distribution of fire stages.

[0109] Specifically, the formula for calculating path cost is as follows:

[0110]

[0111] Where Loss represents path cost, L_smoke represents smoke data, w_smoke represents smoke cost weight, L_structure represents structural damage information, w_structure represents structural damage cost weight, L_density represents crowd density information, and w_density represents crowd density cost weight.

[0112] Specifically, real-time evacuation routes are planned based on path costs, including:

[0113] Based on the monitoring locations of each path, the distance between the corresponding road segments is determined, and the A* algorithm is used to plan real-time evacuation routes according to the road segment distance and path cost.

[0114] Based on road segment distance and path cost, the A* algorithm is used to plan real-time evacuation routes, including:

[0115] Based on the distance between road segments, the A* algorithm is used to plan a set number of candidate paths with the shortest distance;

[0116] The total cost of each candidate path is calculated based on the path cost, and the candidate path with the minimum total cost is selected as the real-time evacuation path.

[0117] After importing real-time evacuation routes into the BIM model for visualization, the following is also included:

[0118] If a change in the stage of a fire is detected, the real-time evacuation routes will be updated.

[0119] If an abnormal state is detected at the path monitoring location corresponding to the real-time evacuation route, the real-time evacuation route will be updated. The abnormal state indicates that the path cost at the current path monitoring location has reached the set cost threshold.

[0120] The above-mentioned methods involve collecting temperature, smoke, and flame monitoring data at each path monitoring location using the BIM model of the corresponding monitoring area, performing fire stage analysis based on the temperature, smoke, and flame monitoring data, and generating fire stage determination results. Fire stages include predefined initial, development, and decay stages. The BIM model of the corresponding monitoring area also collects crowd density information at each path monitoring location. Based on the temperature and flame monitoring data, structural damage information at the corresponding path monitoring location is calculated. Smoke cost weights, structural damage cost weights, and crowd density cost weights are configured according to the fire stage determination results. Path costs for the corresponding path monitoring locations are calculated based on smoke data, structural damage information, crowd density information, smoke cost weights, structural damage cost weights, and crowd density cost weights. Path costs characterize the accessibility of the corresponding path monitoring location. Real-time evacuation routes are planned based on the path costs and imported into the BIM model for visualization. By employing the aforementioned technical means, adjusting path costs according to different fire stages in the fire evolution process, and planning real-time evacuation routes based on path costs, the environmental adaptability of evacuation routes can be significantly improved, the accuracy of fire evacuation route planning can be enhanced, and the evacuation routes can be ensured to respond to fire evolution, always avoiding high-temperature zones, dense smoke zones, structurally hazardous zones, and congested nodes. This significantly reduces the risk of path guidance errors caused by changes in the disaster situation and improves the escape success rate.

[0121] The BIM-based building fire evacuation analysis system provided in Embodiment 2 of this application can be used to execute the BIM-based building fire evacuation analysis method provided in Embodiment 1 above, and has corresponding functions and beneficial effects.

[0122] Example 3:

[0123] This application provides an electronic device in embodiment three, referring to... Figure 6 The electronic device includes a processor 31, a memory 32, a communication module 33, an input device 34, and an output device 35. The electronic device may have one or more processors and one or more memories. The processor, memory, communication module, input device, and output device of the electronic device can be connected via a bus or other means.

[0124] Memory, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the BIM-based building fire evacuation analysis method described in any embodiment of this application (e.g., the stage determination module, stage determination module, and path generation module in a BIM-based building fire evacuation analysis system). Memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on device usage, etc. Furthermore, memory may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0125] The communication module is used for data transmission.

[0126] The processor executes various functional applications and data processing of the device by running software programs, instructions, and modules stored in memory, thereby realizing the above-mentioned BIM-based building fire evacuation analysis method.

[0127] Input devices can be used to receive input numerical or character information, and to generate key signal inputs related to user settings and function control of the device. Output devices may include display devices such as displays.

[0128] The electronic equipment provided above can be used to execute the BIM-based building fire evacuation analysis method provided in Embodiment 1 above, and has the corresponding functions and beneficial effects.

[0129] Example 4:

[0130] This application embodiment also provides a storage medium containing computer-executable instructions. When executed by a computer processor, the computer-executable instructions are used to execute a BIM-based building fire evacuation analysis method. This BIM-based building fire evacuation analysis method includes: collecting temperature data, smoke data, and flame monitoring data at each path monitoring location from the BIM model of the corresponding monitoring area; performing fire stage analysis based on the temperature data, smoke data, and flame monitoring data to generate a fire stage determination result, where the fire stages include a predefined initial stage, development stage, and decay stage; collecting crowd density information at each path monitoring location from the BIM model of the corresponding monitoring area; calculating structural damage information at the corresponding path monitoring location based on the temperature data and flame monitoring data; configuring smoke cost weights, structural damage cost weights, and crowd density cost weights based on the fire stage determination result; calculating the path cost at the corresponding path monitoring location based on the smoke data, structural damage information, crowd density information, smoke cost weights, structural damage cost weights, and crowd density cost weights; the path cost characterizes the accessibility difficulty of the corresponding path monitoring location; planning real-time evacuation routes based on the path costs; and importing the real-time evacuation routes into the BIM model for visualization.

[0131] Storage medium – any type of memory device or storage device. The term “storage medium” is intended to include: mounting media, such as CD-ROM, floppy disk, or magnetic tape devices; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (e.g., hard disk or optical storage); registers or other similar types of memory elements, etc. Storage medium may also include other types of memory or combinations thereof. Furthermore, storage medium may reside in a first computer system in which the program is executed, or it may reside in a different second computer system connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term “storage medium” can include two or more storage media residing in different locations (e.g., in different computer systems connected via a network). Storage medium may store program instructions (e.g., specifically implemented as a computer program) executable by one or more processors.

[0132] Of course, the computer-executable instructions provided in the embodiments of this application are not limited to the BIM-based building fire evacuation analysis method described above, but can also execute related operations in the BIM-based building fire evacuation analysis method provided in any embodiment of this application.

[0133] The BIM-based building fire evacuation analysis system, storage medium, and electronic equipment provided in the above embodiments can execute the BIM-based building fire evacuation analysis method provided in any embodiment of this application. For technical details not described in detail in the above embodiments, please refer to the BIM-based building fire evacuation analysis method provided in any embodiment of this application.

[0134] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of this application, the scope of which is determined by the scope of the claims.

Claims

1. A BIM-based building fire evacuation analysis method, characterized by, include: The BIM model of the corresponding monitoring area collects temperature data, smoke data and flame monitoring data at each path monitoring location. Based on the temperature data, smoke data and flame monitoring data, fire stage analysis is performed to generate fire stage determination results. The fire stages include predefined initial stage, development stage and decay stage. The BIM model of the corresponding monitoring area collects crowd density information at each path monitoring location. Based on the temperature data and flame monitoring data, structural damage information at the corresponding path monitoring location is calculated. Smoke cost weights, structural damage cost weights, and crowd density cost weights are configured according to the fire stage determination result. Specifically, predefined weight configuration rules are used to configure the smoke cost weights, structural damage cost weights, and crowd density cost weights for each stage. In the weight configuration rules, in the initial stage of the fire stage, the crowd density cost weights and smoke cost weights are configured higher, while the structural damage cost weight is configured lower. In the development stage of the fire stage, the smoke cost weights and structural damage weights are configured higher, while the crowd density cost weight is configured lower. In the decay stage of the fire stage, the smoke cost weights and structural damage weights are configured higher, while the crowd density cost weight is configured lower. Based on the smoke data, structural damage information, crowd density information, smoke cost weights, structural damage cost weights, and crowd density cost weights, the path cost for the corresponding path monitoring location is calculated. The path cost represents the accessibility difficulty of the corresponding path monitoring location. Based on the path cost, a real-time evacuation route is planned, and the real-time evacuation route is imported into a BIM model for visualization.

2. The BIM-based building fire evacuation analysis method of claim 1, wherein, The formula for calculating the path cost is as follows: Wherein, Loss represents path cost, L_smoke represents the smoke data, w_smoke represents the smoke cost weight, L_structure represents the structural damage information, w_structure represents the structural damage cost weight, L_density represents the crowd density information, and w_density represents the crowd density cost weight.

3. The BIM-based building fire evacuation analysis method of claim 1, wherein, The step of planning real-time evacuation routes based on the path cost includes: Based on the monitoring locations of each path, the distance between the corresponding road segments is determined, and the A* algorithm is used to plan real-time evacuation routes according to the road segment distance and the path cost.

4. The BIM-based building fire evacuation analysis method of claim 3, wherein, The step of planning real-time evacuation routes using the A* algorithm based on the road segment distance and the path cost includes: Based on the distance of the road segments, the A* algorithm is used to plan a set number of candidate paths with the shortest distance; The total cost of each candidate path is calculated based on the path cost, and the candidate path with the minimum total cost is selected as the real-time evacuation path.

5. The BIM-based building fire evacuation analysis method of claim 1, wherein, After importing the real-time evacuation routes into the BIM model for visualization, the method further includes: If a change in the fire stage is detected, the real-time evacuation routes are updated. If the path monitoring location corresponding to the real-time evacuation path is detected to be in an abnormal state, the real-time evacuation path is updated. The abnormal state indicates that the path cost at the current path monitoring location has reached a set cost threshold.

6. The BIM-based building fire evacuation analysis method of claim 1, wherein, After calculating the structural damage information for the corresponding path monitoring location based on the temperature data and the flame monitoring data, the method further includes: Collect monitoring image information of the corresponding path monitoring location, identify specified image features based on the monitoring image information, generate correction coefficients based on the specified image features, and correct the structural damage information based on the correction coefficients.

7. The BIM-based building fire evacuation analysis method of claim 1, wherein, The step of analyzing the fire stage based on the temperature data, the smoke data, and the flame monitoring data to generate a fire stage determination result includes: The temperature data, smoke data, and flame monitoring data are input into a pre-built LSTM network. Based on the LSTM network, the probability distribution information of the fire stage for a specified period in the future is predicted. Based on the probability distribution information of the fire stage, a fire stage determination result is generated.

8. A BIM-based building fire evacuation analysis system, characterized in that, include: The stage determination module is used to collect temperature data, smoke data and flame monitoring data at each path monitoring location in the BIM model of the corresponding monitoring area, perform fire stage analysis based on the temperature data, smoke data and flame monitoring data, and generate fire stage determination results. The fire stages include predefined initial stage, development stage and decay stage. The cost calculation module is used to collect crowd density information at each path monitoring location from the BIM model of the corresponding monitoring area, calculate structural damage information at the corresponding path monitoring location based on the temperature data and the flame monitoring data, and configure smoke cost weight, structural damage cost weight, and crowd density cost weight according to the fire stage determination result. Specifically, predefined weight configuration rules are used to configure the smoke cost weight, structural damage cost weight, and crowd density cost weight for the corresponding stage. In the weight configuration rules, in the initial stage of the fire stage, the crowd density cost weight and smoke cost weight are configured higher, and the structural damage cost weight is configured lower; in the development stage of the fire stage, the smoke cost weight and structural damage cost weight are configured higher, and the crowd density cost weight is configured lower; in the decay stage of the fire stage, the smoke cost weight and structural damage cost weight are configured higher, and the crowd density cost weight is configured lower. Based on the smoke data, the structural damage information, the crowd density information, the smoke cost weight, the structural damage cost weight, and the crowd density cost weight, the path cost for the corresponding path monitoring location is calculated. The path cost represents the accessibility difficulty of the corresponding path monitoring location. The path generation module is used to plan real-time evacuation routes based on the path costs and import the real-time evacuation routes into the BIM model for visualization.

9. An electronic device, characterized in that, include: Memory and one or more processors; The memory is used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the BIM-based building fire evacuation analysis method as described in any one of claims 1-7.

10. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the BIM-based building fire evacuation analysis method as described in any one of claims 1-7.

Images