Underground space protection evaluation system and method based on bim and multi-source data fusion

Abstract

The application discloses an underground space protection evaluation system based on BIM and multi-source data fusion, which comprises a data input layer, an evaluation engine layer and a result output layer which are sequentially arranged for data transmission and processing, wherein the data input layer is configured to construct a three-dimensional BIM model, an attack mode and a damage path associated rule base based on a terrorist attack parameter base and real-time environment data flow fused by multi-source sensor data, wherein the three-dimensional BIM model is integrated with geological survey data, topographic mapping, structure parameters and civil air defense engineering data through a space topology modeling technology; the evaluation engine layer is configured to receive the three-dimensional BIM model, the associated rule base and the real-time environment data flow of the data input layer, generate key damage propagation simulation results and vulnerability analysis results through explosion dynamic response simulation and space topology network analysis; and the result output layer is configured to generate a three-dimensional space classification map, dynamic design adjustment suggestions and a model self-checking report based on the damage propagation simulation results and the vulnerability analysis results of the evaluation engine layer.

Description

Technical Field

[0001] This invention relates to the field of underground space design, and specifically to an underground space protection assessment system and method based on BIM and multi-source data fusion. Background Technology

[0002] Traditional underground space design relies on static specifications, which cannot dynamically respond to the randomness and diverse nature of terrorist attacks. Existing assessment models lack self-checking capabilities, require manual input of threat scenarios, and are difficult to cover complex attack types (such as compound explosions + biological and chemical attacks). They also lack prediction of secondary disasters and have not established a chain reaction model of "attack → structural damage → personnel evacuation failure". Summary of the Invention

[0003] In view of the technical defects and drawbacks existing in the prior art, embodiments of the present invention provide an underground space protection assessment system and method based on BIM and multi-source data fusion to overcome or at least partially solve the above problems. The specific solution is as follows:

[0004] As a first aspect of the present invention, an underground space protection assessment system based on BIM and multi-source data fusion is provided, comprising a data input layer, an assessment engine layer, and a result output layer for sequential data transmission and processing, wherein:

[0005] The data input layer is configured to construct a 3D BIM model, an association rule base based on the attack patterns and destruction paths of a terrorist attack parameter database, and a real-time environmental data stream fused from multi-source sensor data. The 3D BIM model integrates geological survey data, topographic mapping, structural parameters, and civil defense engineering data through spatial topology modeling technology. The evaluation engine layer is configured to receive the 3D BIM model, association rule base, and real-time environmental data stream from the data input layer, and generate key damage propagation simulation results and vulnerability analysis results through explosion dynamic response simulation and spatial topology network analysis.

[0006] The output layer is configured to generate a three-dimensional spatial hierarchy map, dynamic design adjustment suggestions, and model self-assessment report based on the damage propagation simulation results and vulnerability analysis results of the evaluation engine layer.

[0007] Furthermore, the data input layer includes:

[0008] The BIM modeling module uses Revit parametric 3D modeling software. It takes geological survey data, topographic mapping data, structural data, and civil defense engineering data as input and integrates them using spatial topology modeling technology to generate a networked 3D model. Material properties and load constraints are then assigned, ultimately producing a 3D BIM model with structural attributes. The connection relationships between beam, column, and slab components form the basic topology for explosion propagation path analysis. The terrorist attack parameter library module uses historical events, weapon characteristics, and real-time intelligence from a global terrorist attack database based on the data input layer. It establishes a rule base for associating attack patterns with destruction paths using Bayesian network classification, building a quantitative parameter library containing multiple attack patterns for subsequent destruction propagation simulation.

[0009] The multi-source sensor data fusion module is used to integrate environmental sensors, structural sensors, and security sensors via the Modbus protocol to acquire environmental data, structural health monitoring data, toxic gas concentration data, and population density distribution data. It sets a critical data acquisition cycle of 30 seconds and a non-critical data acquisition cycle of 5 minutes. Using an appropriate fusion algorithm, it fuses the multi-source sensor data to build a data platform and realizes real-time environmental data stream output for dynamically updating the boundary conditions of the damage propagation simulation.

[0010] Furthermore, geological survey data includes three-dimensional geological models, borehole data, and geological structure data; topographic mapping data includes topographic maps, site elevations, and road layout data; structural data includes material strength, component dimensions, and connection methods; civil defense engineering data includes evacuation routes and protective door location data; building components include beams, columns, and slabs; and material properties include elastic modulus and Poisson's ratio.

[0011] Furthermore;

[0012] Historical events, including explosive yield, blast radius, type of poison gas, and method of attack;

[0013] Weapon characteristics, including new attack methods and weapon updates;

[0014] Real-time intelligence, including new attack methods and weapons updates.

[0015] Furthermore, the appropriate fusion algorithm includes one of weighted averaging, Kalman filtering, Bayesian estimation, and DS evidence reasoning algorithm.

[0016] Furthermore, the evaluation engine layer includes:

[0017] The destruction propagation simulation module receives the topological data of the 3D BIM model from the data input layer, the association rule base of attack modes and destruction paths, and real-time environmental data streams. Using Auto-DYN simulation software and the SPH algorithm, it simulates the propagation laws of shock wave diffusion, debris collapse, and secondary reflection, generating shock wave pressure cloud maps at different times within the building. It also obtains pressure time-history curves and toxic gas diffusion concentration lines at different distances from the blast center. Inputting the obtained shock wave pressure time-history curve data, it analyzes the structural plastic deformation using LS-Dyna simulation software to obtain the strain, stress, and displacement distribution characteristics of key structures within the building, i.e., residual bearing capacity indicators, used to assess residual bearing capacity. The vulnerability analysis module calculates the frequency of node occurrence in all shortest paths based on betweenness centrality, quantifies its impact on network connectivity, identifies key nodes in the network (such as evacuation route intersections), forms a set of key nodes and their importance, and obtains a conditional probability table using a Bayesian network algorithm. This simulates the failure process of associated structures, quantifies the cascading effects of key node failure on adjacent components, and outputs the probability results of cascading failure risks.

[0018] Furthermore, the method for generating the Bayesian network conditional probability table in the vulnerability analysis module includes: using the component strain data output by the failure propagation simulation module as the parent node, and calculating the failure conditional probability of the child node using the Markov chain Monte Carlo method.

[0019] Furthermore, the result output layer includes: a space protection level map generation module, used to classify the damage propagation simulation results (residual bearing capacity of key structures, peak shock wave pressure, toxic gas concentration, etc.) and vulnerability analysis results (node ​​importance of key nodes and their cascading failure risk probability) into two major indicators: structural safety and evacuation accessibility. A multi-dimensional evaluation index system is established, and the entropy-weighted TOPSIS method is used to synthesize the two major indicators. The vulnerability index of each spatial unit is calculated, and K-means clustering is applied to cluster the structural safety indicators and evacuation accessibility indicators to generate a three-dimensional space protection level map. The three-dimensional space protection level map includes: red zones, A comprehensive index > 0.8 indicates critical structural failure and obstructed evacuation, prohibiting entry; a yellow zone, with a comprehensive index of 0.5 < index ≤ 0.8, indicates a significant risk of single system failure, restricting access; a blue zone, with an index ≤ 0.5, indicates multiple redundancies and is a safe area. Simultaneously, in the 3D BIM model, corresponding semi-transparent protection level color blocks (e.g., 60% transparency for red zone, 40% for yellow) are superimposed on different zones to achieve visualization, ultimately forming a 3D spatial protection classification map composed of three zones: red (representing prohibited entry), yellow (representing restricted access), and blue (representing safe areas).

[0020] The dynamic design suggestion module is used to optimize reinforcement schemes based on the 3D spatial protection classification map and the multi-objective genetic algorithm NSGA-II to form a structural reinforcement location map. Combining the 3D spatial protection classification map and real-time population density distribution data, it considers congestion weights, dynamically updates node congestion weights, and calculates the shortest and least congested safe evacuation routes based on the Dijkstra algorithm, outputting a dynamic planning scheme for evacuation routes. The self-assessment report generation module, based on the simulation analysis results of the evaluation engine layer, randomly perturbs the model parameters through Monte Carlo simulation, calculates the structural reliability index, and outputs a stability report including confidence verification.

[0021] As a second aspect of the present invention, a method for underground space protection assessment based on BIM and multi-source data fusion is characterized in that the method includes:

[0022] Step 1: Construct a 3D BIM model, an association rule base based on a terrorist attack parameter database for attack patterns and destruction paths, and a real-time environmental data stream fused from multi-source sensor data through the data input layer. The 3D BIM model integrates geological survey data, topographic mapping, structural parameters, and civil defense engineering data through spatial topology modeling technology. Step 2: Receive the 3D BIM model, association rule base, and real-time environmental data stream from the data input layer through the evaluation engine layer. Generate key damage propagation simulation results and vulnerability analysis results through explosion dynamic response simulation and spatial topology network analysis.

[0023] Step 3: Based on the damage propagation simulation results and vulnerability analysis results of the evaluation engine layer, the results output layer generates a three-dimensional spatial hierarchy map, dynamic design adjustment suggestions, and model self-assessment report.

[0024] Further, step 1 includes:

[0025] Using Revit parametric 3D modeling software, geological survey data, topographic mapping data, structural data, and civil defense engineering data are input. Spatial topology modeling technology is used to fuse these data to generate a networked 3D model, assigning material properties and load constraints. This ultimately produces a 3D BIM model with structural attributes, where the connection relationships between beam, column, and slab components form the basic topology for explosion propagation path analysis. Based on historical events, weapon characteristics, and real-time intelligence from a global terrorist attack database at the data input layer, a rule base for linking attack patterns and destruction paths is established using Bayesian network classification. A quantitative parameter library containing multiple attack patterns is also built for subsequent destruction propagation simulation.

[0026] By integrating environmental sensors, structural sensors, and security sensors via the Modbus protocol, environmental data, structural health monitoring data, toxic gas concentration data, and population density distribution data are acquired. A critical data acquisition cycle of 30 seconds and a non-critical data acquisition cycle of 5 minutes are set. An appropriate fusion algorithm is used to fuse multi-source sensor data, construct a data platform, and realize real-time environmental data stream output for dynamically updating the boundary conditions of damage propagation simulation.

[0027] Further, step 2 includes:

[0028] The system receives 3D BIM model topology data, attack mode and destruction path association rule base, and real-time environmental data stream from the data input layer. Using Auto-DYN simulation software and the SPH algorithm, it simulates the propagation of shock waves, debris fall, and secondary reflections, generating shock wave pressure cloud maps of the building interior at different times. It also obtains pressure time-history curves and toxic gas diffusion concentration lines at different blast center distances. Inputting the obtained shock wave pressure time-history curve data, it analyzes structural plastic deformation using LS-Dyna simulation software to obtain the strain, stress, and displacement distribution characteristics of key structures within the building, i.e., residual bearing capacity indices, used to assess residual bearing capacity. Based on betweenness centrality, it calculates the frequency of node occurrence in all shortest paths, quantifies its impact on network connectivity, identifies key nodes in the network (such as evacuation route intersections), and forms a set of key nodes and their importance. Using a Bayesian network algorithm, it obtains a conditional probability table, simulates the failure process of associated structures, quantifies the cascading effects of key node failure on adjacent components, and outputs the probability results of cascading failure risks.

[0029] Furthermore, step 3 includes:

[0030] The simulation results of damage propagation (residual bearing capacity of key structures, peak shock wave pressure, toxic gas concentration, etc.) and the vulnerability analysis results (nodal importance of key nodes and their probability of cascading failure) are respectively categorized into two major indicators: structural safety and evacuation accessibility. A multi-dimensional evaluation index system is constructed, and the entropy weight-TOPSIS method is used to integrate the two major indicators to calculate the vulnerability index of each spatial unit. K-means clustering is applied to cluster the structural safety index and the evacuation accessibility index to generate a three-dimensional spatial protection classification map. The three-dimensional spatial protection classification map includes: red zones, corresponding to comprehensive index > A red zone with a critical index of 0.8 indicates that the structure is on the verge of collapse and evacuation is obstructed, and entry is prohibited. A yellow zone, with a comprehensive index of 0.5 < index ≤ 0.8, indicates a significant risk of single system failure and restricted access. A blue zone, with an index ≤ 0.5, indicates a safe area with multiple layers of protection redundancy. In the 3D BIM model, semi-transparent protection level color blocks corresponding to different zones are superimposed (e.g., 60% transparency for the red zone and 40% transparency for the yellow zone) to achieve a visualization effect. The final result is a 3D spatial protection classification map consisting of three zones: the red zone (representing prohibited entry), the yellow zone (representing restricted access), and the blue zone (representing a safe area).

[0031] Based on the 3D spatial protection classification map, the reinforcement scheme is optimized using the multi-objective genetic algorithm NSGA-II to form a structural reinforcement location map. Combining the 3D spatial protection classification map and real-time population density distribution data, considering congestion weights, the node congestion weights are dynamically updated. Based on Dijkstra's algorithm, the shortest and least congested safe evacuation routes are calculated, and the dynamic planning scheme of the evacuation routes is output. The self-audit report generation module, based on the simulation analysis results of the evaluation engine layer, randomly perturbs the model parameters through Monte Carlo simulation, calculates the structural reliability index, and outputs a stability report including confidence verification.

[0032] The present invention has the following beneficial effects:

[0033] 1. By using spatial topology modeling technology to integrate geological data, structural parameters and civil defense facilities in three dimensions, a networked BIM model is constructed, which accurately maps the connection relationship between beam and column nodes and the explosion propagation path. Compared with the traditional two-dimensional drawing evaluation model, the accuracy of spatial association rule recognition is improved by 40%.

[0034] 2. The DS evidence reasoning algorithm is used to integrate environmental, structural and security sensor data streams to achieve real-time dynamic correction of simulation boundary conditions (such as the attenuation factor correction of temperature on the elastic modulus of materials) in 30 seconds, so that the error between the explosion shock wave SPH simulation results and the measured strain data is controlled within 5%.

[0035] 3. Based on the conditional probability table of Bayesian network, a rule base for the association between attack patterns and destruction paths is constructed. Combined with the Markov chain Monte Carlo method to calculate the cascading failure probability, it can identify 78% of secondary disaster risk nodes that are ignored by traditional methods.

[0036] 4. An entropy weight-TOPSIS method is applied to construct a dual-polarity index evaluation system (negative standardization of shock wave pressure and positive standardization of bearing capacity). Combined with the NSGA-II algorithm, the Pareto optimal solution set of structural reinforcement and evacuation path is generated, which improves the overall effectiveness of the protection scheme by 25% and shortens the emergency response time by 60%.

[0037] 5. The model parameters are verified by random perturbation through Monte Carlo simulation, and the structural reliability index (β≥2.5) with a confidence level >95% is output to ensure that the evaluation results meet the requirements of the first-level protection level of the "Design Code for Civil Air Defense Engineering" (GB50225). Attached Figure Description

[0038] Figure 1 This invention provides a technical roadmap for underground space protection assessment based on BIM and multi-source data fusion. Detailed Implementation

[0039] 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 a part of the present invention, and not all of the 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.

[0040] refer to Figure 1 As shown in the figure, an underground space protection assessment system based on BIM and multi-source data fusion provided by this invention includes a data input layer, an assessment engine layer, and a result output layer for sequential data transmission and processing, wherein:

[0041] The data input layer is configured to construct a 3D BIM model, an association rule base based on the attack patterns and destruction paths of a terrorist attack parameter database, and a real-time environmental data stream fused from multi-source sensor data. The 3D BIM model integrates geological survey data, topographic mapping, structural parameters, and civil defense engineering data through spatial topology modeling technology. The evaluation engine layer is configured to receive the 3D BIM model, association rule base, and real-time environmental data stream from the data input layer, and generate key damage propagation simulation results and vulnerability analysis results through explosion dynamic response simulation and spatial topology network analysis.

[0042] The output layer is configured to generate a three-dimensional spatial hierarchy map, dynamic design adjustment suggestions, and model self-assessment report based on the damage propagation simulation results and vulnerability analysis results of the evaluation engine layer.

[0043] This invention can advance the attack risk assessment from "post-event response" to the "design stage", reducing the transformation cost by more than 60%; the self-checking algorithm ensures the reliability of the model, with a false judgment rate of <3% (compared to 15% of traditional manual assessment), and is applicable to deep underground spaces such as subways, civil defense projects, and strategic energy reserves. It can be integrated into BIM design software (such as Revit plugin) to achieve seamless integration.

[0044] In some embodiments, the data input layer includes:

[0045] The BIM modeling module uses Revit parametric 3D modeling software, accepting input data such as geological surveys (3D geological model, borehole data, geological structures), topographic mapping (topographic map, site elevation, road layout), structural data (material strength, component dimensions, connection methods), and civil defense engineering data (evacuation routes, protective door locations). Employing spatial topology modeling technology, it connects building components (beams, columns, slabs) through nodes to form a network structure, assigning material properties (elastic modulus, Poisson's ratio) and load constraints, ultimately generating a 3D BIM model with structural attributes. This supports subsequent explosive structural dynamic response analysis and extraction of building spatial topological relationships. The terrorist attack parameter library module, based on historical event parameters from the Global Terrorist Attack Database (GTD) in the data input layer, establishes a rule base for association between attack patterns and destruction paths using Bayesian network classification. This rule base is output to the evaluation engine layer as the basis for selecting explosion parameters.

[0046] This invention utilizes web crawlers to extract open-source intelligence, such as the Global Terrorist Attacks Database (GTD) and news websites, to obtain historical event parameters from the GTD database. These parameters include historical events (explosive yield, blast radius, type of poison gas, attack method), weapon characteristics (TNT yield coefficient, chemical weapon proliferation parameters), and real-time intelligence (new attack methods, weapon update dynamics). A Bayesian network classification method is employed to identify the correlation between attack types and destruction paths (e.g., explosion → shock wave → structural collapse). Ultimately, a rule base containing eight attack modes (explosion, hijacking, facility attack, etc.) is built to support subsequent simulation of destruction propagation.

[0047] The multi-source sensor data fusion module integrates environmental sensors (temperature, humidity), structural sensors (FBG fiber optic strain data), and security sensors (toxic gas concentration, crowd density) via the Modbus protocol. It uses the DS evidence reasoning algorithm to generate real-time data streams, with key data in 30-second cycles directly input into the evaluation engine layer for dynamic boundary condition updates.

[0048] This invention acquires environmental parameters (temperature, humidity, wind speed) via Modbus protocol transmission, uses FBG fiber optic sensors to monitor structural health (strain, displacement), employs electrochemical sensors to collect toxic gas concentration data in real time, and uses infrared thermal imagers (such as the FLIR A700 series) to capture crowd gathering hotspots. It also uses floor tile pressure sensors to infer crowd density distribution. A 30-second critical data acquisition cycle and a 5-minute non-critical data acquisition cycle are set. Appropriate fusion algorithms (such as weighted average, Kalman filtering, Bayesian estimation, and DS evidence inference) are used to fuse multi-source sensor data, constructing a data platform to achieve real-time environmental data stream output for dynamically updating boundary conditions in damage propagation simulations.

[0049] In some embodiments, the evaluation engine layer includes:

[0050] The damage propagation simulation module receives the topological data of the 3D BIM model from the data input layer, the association rule base of attack modes and damage paths, and the real-time environmental data stream. It simulates the propagation path of shock waves in the structural network through the SPH algorithm. The real-time environmental data stream dynamically corrects the simulation boundary conditions (the correction coefficient of wind speed on the toxic gas diffusion path and the attenuation factor of temperature on the elastic modulus of the material) and outputs the residual bearing capacity index of key components.

[0051] The explosion parameters in the structural and terrorist attack parameter library of this invention include basic data such as structural parameters (component stiffness, node constraints), explosion parameters (charge shape, equivalent, detonation distance, incident angle), and environmental parameters (temperature, humidity, wind speed). Using Auto-DYN simulation software and the SPH algorithm, the propagation laws of shock wave diffusion, debris fall, and secondary reflection are simulated. This ultimately generates shock wave pressure cloud maps of the building interior at different times, obtaining pressure time-history curves and concentration lines of toxic gas diffusion at different blast center distances. Inputting the obtained shock wave pressure time-history curve data, the structural plastic deformation is analyzed using LS-Dyna simulation software, yielding strain, stress, and displacement distribution characteristics of key structures (column bases, nodes) within the building, used to assess residual bearing capacity. The vulnerability analysis module, based on component damage data output by the damage propagation simulation module, abstracts the building plan as a spatial topological network. It identifies key nodes using a betweenness centrality algorithm and calculates cascading failure risk values ​​based on a Bayesian network conditional probability table, forming a two-dimensional evaluation index (structural performance index, system risk value) transmitted from the evaluation engine layer to the result output layer.

[0052] This invention abstracts the planar structure of deep underground spaces into a spatial topological network. Based on betweenness centrality, it calculates the frequency of a node's occurrence in all shortest paths, quantifies its impact on network connectivity, identifies key nodes in the network (such as intersections of evacuation routes), and forms a set of key nodes and their importance. Then, it obtains a conditional probability table (CPT) through a Bayesian network algorithm, simulates the failure process of associated structures, quantifies the cascading impact of key node failure on adjacent components, and outputs the probability results of cascading failure risks, such as the failure of a load-bearing column leading to the collapse of 30% of the area.

[0053] The method for generating the Bayesian network conditional probability table in the vulnerability analysis module includes: using the component strain data output by the failure propagation simulation module as the parent node, and calculating the failure conditional probability of the child node (adjacent component) using the Markov chain Monte Carlo method.

[0054] In some embodiments, the result output layer includes: a space protection level map generation module, which calculates the space unit vulnerability index using the entropy weight-TOPSIS method based on the damage propagation simulation results and vulnerability analysis results transmitted from the evaluation engine layer, wherein the weighted values ​​of the structural safety index (weight 0.6) and the evacuation accessibility index (weight 0.4) are used to generate a three-dimensional space protection level map through K-means clustering;

[0055] This invention categorizes the results of damage propagation simulation (residual bearing capacity of key structures, peak shock wave pressure, toxic gas concentration, etc.) and vulnerability analysis (nodal importance of key nodes and their probability of cascading failure) into two aspects: structural safety and evacuation accessibility. A multi-dimensional evaluation index system is established, and the entropy weight-TOPSIS method is used to integrate the two major indicators to calculate the vulnerability index of each spatial unit. K-means clustering is applied to divide the index into three levels: Red zone (comprehensive index > 0.8): critical structural failure and obstructed evacuation; Yellow zone (0.5 < index ≤ 0.8): significant risk of single system failure; Blue zone (index ≤ 0.5): multiple protection redundancies. Simultaneously, semi-transparent protection level color blocks (60% transparency for red, 40% for yellow) are overlaid in the 3D BIM model to achieve visualization. The final result is a 3D spatial hierarchy map composed of three zones: red (no entry), yellow (restricted access), and blue (safe zone). The dynamic adjustment suggestion module is used to input the key risk areas of the three-dimensional spatial protection classification map and the real-time population density data (from the inversion results of the paving stone pressure sensor in the data input layer) into the NSGA-II multi-objective optimization algorithm to generate a dynamic scheme that includes structural reinforcement parameters and evacuation path weights.

[0056] The system utilizes a 3D spatial protection grading map and employs the multi-objective genetic algorithm NSGA-II to optimize reinforcement schemes, generating a structural reinforcement location map. Combining the 3D spatial protection grading map with real-time population density distribution data, and considering congestion weights, it dynamically updates node congestion weights. Based on Dijkstra's algorithm, it calculates the shortest and least congested safe evacuation routes, outputting a dynamic planning scheme for the evacuation routes. The self-assessment report generation module, based on the simulation analysis results of the evaluation engine layer, uses Monte Carlo simulation to randomly perturb model parameters, calculates structural reliability indices, and outputs a stability report including confidence level verification.

[0057] Based on the Monte Carlo simulation method, random samples on the order of 10^4 were generated by changing parameters such as the coefficient of variation of material strength (±10%) and random disturbance factor (e.g., explosion location offset ±2m) in the model. The model confidence report was obtained by calculating the structural reliability index β=Φ-1(1-pƒ). For example, the reliability β≥3.0 at a 95% confidence level was used to verify the stability of the model output.

[0058] As a second aspect of the present invention, a method for underground space protection assessment based on BIM and multi-source data fusion is provided, the method comprising:

[0059] Step 1: Construct a 3D BIM model, an association rule base based on a terrorist attack parameter database for attack patterns and destruction paths, and a real-time environmental data stream fused from multi-source sensor data through the data input layer. The 3D BIM model integrates geological survey data, topographic mapping, structural parameters, and civil defense engineering data through spatial topology modeling technology. Step 2: Receive the 3D BIM model, association rule base, and real-time environmental data stream from the data input layer through the evaluation engine layer. Generate key damage propagation simulation results and vulnerability analysis results through explosion dynamic response simulation and spatial topology network analysis.

[0060] Step 3: Based on the damage propagation simulation results and vulnerability analysis results of the evaluation engine layer, the results output layer generates a three-dimensional spatial hierarchy map, dynamic design adjustment suggestions, and model self-assessment report.

[0061] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A BIM-based underground space protection assessment system based on multi-source data fusion, characterized in that, It includes a data input layer, an evaluation engine layer, and a result output layer, which sequentially transmit and process data. The data input layer is configured to construct a 3D BIM model, an association rule base based on the attack mode and destruction path of the terrorist attack parameter database, and a real-time environmental data stream fused from multi-source sensor data; the evaluation engine layer is configured to receive the 3D BIM model, association rule base, and real-time environmental data stream from the data input layer, and generate key destruction propagation simulation results and vulnerability analysis results through explosion dynamic response simulation and spatial topology network analysis. The output layer is configured to generate a three-dimensional spatial hierarchy map, dynamic design adjustment suggestions, and model self-assessment report based on the damage propagation simulation results and vulnerability analysis results of the evaluation engine layer. The data input layer includes: The BIM modeling module uses Revit parametric 3D modeling software. It takes geological survey data, topographic mapping data, structural data, and civil defense engineering data as input and integrates them using spatial topology modeling technology to generate a networked 3D model. Material properties and load constraints are then assigned, ultimately producing a 3D BIM model with structural attributes. The connection relationships between beam, column, and slab components form the basic topology for explosion propagation path analysis. The terrorist attack parameter library module uses historical events, weapon characteristics, and real-time intelligence from a global terrorist attack database based on the data input layer. It establishes a rule base for associating attack patterns with destruction paths using Bayesian network classification, building a quantitative parameter library containing multiple attack patterns for subsequent destruction propagation simulation. The multi-source sensor data fusion module is used to integrate environmental sensors, structural sensors, and security sensors via the Modbus protocol to acquire environmental data, structural health monitoring data, toxic gas concentration data, and population density distribution data. It sets a critical data acquisition cycle of 30 seconds and a non-critical data acquisition cycle of 5 minutes. Using an appropriate fusion algorithm, it fuses the multi-source sensor data to build a data platform and realizes real-time environmental data stream output for dynamically updating the boundary conditions of the damage propagation simulation. The evaluation engine layer includes: The destruction propagation simulation module receives the topological data of the 3D BIM model from the data input layer, the association rule base of attack modes and destruction paths, and real-time environmental data streams. Using Auto-DYN simulation software and the SPH algorithm, it simulates the propagation laws of shock wave diffusion, debris collapse, and secondary reflection, generating shock wave pressure cloud maps at different times within the building. It also obtains pressure time-history curves and toxic gas diffusion concentration lines at different distances from the blast center. Inputting the obtained shock wave pressure time-history curve data, it analyzes the structural plastic deformation using LS-Dyna simulation software to obtain the strain, stress, and displacement distribution characteristics of key structures within the building, i.e., residual bearing capacity indicators, used to assess residual bearing capacity. The vulnerability analysis module calculates the frequency of node occurrence in all shortest paths based on betweenness centrality, quantifies its impact on network connectivity, identifies key nodes in the network, forms a set of key nodes and their importance, and obtains a conditional probability table using a Bayesian network algorithm. This simulates the failure process of associated structures, quantifies the cascading effects of key node failure on adjacent components, and outputs the probability results of cascading failure risks.

2. The underground space protection assessment system based on BIM and multi-source data fusion as described in claim 1, characterized in that, Geological survey data includes three-dimensional geological models, borehole data, and geological structure data; topographic mapping data includes topographic maps, site elevations, and road layout data; structural data includes material strength, component dimensions, and connection methods; civil defense engineering data includes evacuation routes and the location of protective doors; building components include beams, columns, and slabs; and material properties include elastic modulus and Poisson's ratio.

3. The underground space protection assessment system based on BIM and multi-source data fusion as described in claim 1, characterized in that, The appropriate fusion algorithm includes one of the following: weighted average, Kalman filtering, Bayesian estimation, and DS evidence reasoning algorithm.

4. The underground space protection assessment system based on BIM and multi-source data fusion according to claim 1, characterized in that, The output layer includes a spatial protection level map generation module, which categorizes the damage propagation simulation results and vulnerability analysis results into two major indicators: structural safety and evacuation accessibility. A multi-dimensional evaluation index system is established, and the entropy weight-TOPSIS method is used to synthesize the two major indicators. The vulnerability index of each spatial unit is calculated, and K-means clustering is applied to cluster the structural safety and evacuation accessibility indicators to generate a three-dimensional spatial protection level map. The three-dimensional spatial protection level map includes: a red zone (corresponding to a comprehensive index > 0.8), indicating critical structural failure and obstructed evacuation, prohibiting entry; a yellow zone (corresponding to a comprehensive index of 0.5 < index ≤ 0.8), indicating a significant risk of single system failure, restricting passage; and a blue zone (corresponding to an index ≤ 0.5), indicating multiple protection redundancies, a safe area. Simultaneously, in the three-dimensional BIM model, corresponding semi-transparent protection level color blocks are superimposed on different zones to achieve a visualization effect, ultimately forming a three-dimensional spatial protection level map composed of three zones: the red zone (representing prohibited entry), the yellow zone (representing restricted passage), and the blue zone (representing a safe area). The dynamic design suggestion module is used to optimize reinforcement schemes based on the 3D spatial protection classification map and the multi-objective genetic algorithm NSGA-II to form a structural reinforcement location map. Combining the 3D spatial protection classification map and real-time population density distribution data, it considers congestion weights, dynamically updates node congestion weights, and calculates the shortest and least congested safe evacuation routes based on the Dijkstra algorithm, outputting a dynamic planning scheme for evacuation routes. The self-assessment report generation module, based on the simulation analysis results of the evaluation engine layer, randomly perturbs the model parameters through Monte Carlo simulation, calculates the structural reliability index, and outputs a stability report including confidence verification.

5. A method for underground space protection assessment based on BIM and multi-source data fusion, characterized in that, The method includes: Step 1: Construct a 3D BIM model, an association rule base based on the terrorist attack parameter database for attack patterns and destruction paths, and a real-time environmental data stream fused from multi-source sensor data through the data input layer; Step 2: Receive the 3D BIM model, association rule base, and real-time environmental data stream from the data input layer through the evaluation engine layer, and generate key damage propagation simulation results and vulnerability analysis results through explosion dynamic response simulation and spatial topology network analysis. Step 3: Based on the damage propagation simulation results and vulnerability analysis results of the evaluation engine layer, the result output layer generates a three-dimensional spatial hierarchy map, dynamic design adjustment suggestions, and model self-assessment report. Step 1 includes: Using Revit parametric 3D modeling software, geological survey data, topographic mapping data, structural data, and civil defense engineering data were input. Spatial topology modeling technology was used to fuse these data to generate a networked 3D model, assigning material properties and load constraints. This resulted in a 3D BIM model with structural attributes, where the connection relationships between beam, column, and slab components constituted the basic topology for explosion propagation path analysis. Based on historical events, weapon characteristics, and real-time intelligence from a global terrorist attack database at the data input layer, a rule base for linking attack patterns and destruction paths was established using Bayesian network classification. A quantitative parameter library containing multiple attack patterns was also built for subsequent destruction propagation simulation. By integrating environmental sensors, structural sensors, and security sensors through the Modbus protocol, environmental data, structural health monitoring data, toxic gas concentration data, and population density distribution data are acquired. A critical data acquisition cycle of 30 seconds and a non-critical data acquisition cycle of 5 minutes are set. An appropriate fusion algorithm is used to fuse multi-source sensor data to build a data platform and realize real-time environmental data stream output for dynamically updating the boundary conditions of damage propagation simulation. Step 2 includes: The system receives 3D BIM model topology data, attack mode and destruction path association rule base, and real-time environmental data stream from the data input layer. Using Auto-DYN simulation software and the SPH algorithm, it simulates the propagation of shock waves, debris fall, and secondary reflections, generating shock wave pressure cloud maps of the building interior at different times. It also obtains pressure time-history curves and toxic gas diffusion concentration lines at different blast center distances. Inputting the obtained shock wave pressure time-history curve data, the system analyzes structural plastic deformation using LS-Dyna simulation software to obtain the strain, stress, and displacement distribution characteristics of key structures within the building, i.e., residual bearing capacity indicators, used to assess residual bearing capacity. Based on betweenness centrality, the system calculates the frequency of node occurrence in all shortest paths, quantifies its impact on network connectivity, identifies key nodes in the network, and forms a set of key nodes and their importance. Using a Bayesian network algorithm, a conditional probability table is obtained to simulate the failure process of associated structures, quantifies the cascading effects of key node failure on adjacent components, and outputs the probability results of cascading failure risks.

6. The underground space protection assessment method based on BIM and multi-source data fusion according to claim 5, characterized in that, Step 3 includes: The results of damage propagation simulation and vulnerability analysis are categorized into two major indicators: structural safety and evacuation accessibility. A multi-dimensional evaluation index system is established, and the entropy weight-TOPSIS method is used to integrate the two major indicators. The vulnerability index of each spatial unit is calculated, and K-means clustering is applied to cluster the structural safety index and evacuation accessibility index to generate a three-dimensional spatial protection classification map. The three-dimensional spatial protection classification map includes: red zone, corresponding to a comprehensive index > 0.8, indicating that the structure is at critical failure and evacuation is blocked, and entry is prohibited; yellow zone, corresponding to a comprehensive index of 0.5 < index ≤ 0.8, indicating that the risk of single system failure is significant, and passage is restricted; blue zone, corresponding to an index ≤ 0.5, indicating that there are multiple protection redundancies and it is a safe area. At the same time, in the three-dimensional BIM model, the corresponding semi-transparent protection level color blocks are superimposed on different zones to achieve a visualization effect. Finally, a three-dimensional spatial protection classification map is formed, consisting of three zones: red zone (representing prohibited entry), yellow zone (representing restricted passage), and blue zone (representing safe area). Based on the 3D spatial protection classification map, the reinforcement scheme is optimized using the multi-objective genetic algorithm NSGA-II to form a structural reinforcement location map. Combining the 3D spatial protection classification map and real-time population density distribution data, considering congestion weights, the node congestion weights are dynamically updated. Based on Dijkstra's algorithm, the shortest and least congested safe evacuation routes are calculated, and the dynamic planning scheme of the evacuation routes is output. The self-audit report generation module, based on the simulation analysis results of the evaluation engine layer, randomly perturbs the model parameters through Monte Carlo simulation, calculates the structural reliability index, and outputs a stability report including confidence verification.

Images