Intelligent decision-making method for regional attack and defense integration under the threat of explosion shock

Abstract

The present invention discloses a regional attack and defense integrated intelligent decision-making method under the threat of explosion impact, comprising the following steps: constructing an explosion damage effect database and an explosion impact structural mechanical response database to obtain an attack and defense integrated intelligent decision-making database; constructing an explosion impact structural mechanical response digital object driven intelligent model; constructing a scheduling, protection, and attack regional network flow model; obtaining regional network flow from the regional network flow model; constructing a regional key facility repair scheduling decision model; constructing a regional weak point and vulnerability assessment model; constructing a regional protection plan decision model; constructing a regional attack plan decision model; according to the decision requirements and on-site decision parameters obtained by real-time communication, calling the corresponding decision model to make a decision, and obtaining an attack and defense integrated intelligent decision result. The intelligent decision-making method of the present invention has comprehensive and accurate decision-making.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent decision-making, and in particular to an integrated intelligent decision-making method for regional attack and defense under the threat of explosion impact. Background Art

[0002] Regional attack and defense integrated intelligent decision-making refers to a comprehensive intelligent decision-making process based on the comprehensive and accurate assessment of the structural mechanical response of regional explosion impact and the network loss of related key facilities, for the emergency repair scheduling of regional key facilities, the selection of overall protection plans for urban areas, and the decision-making of effective strike plans against the opponent. The regional attack and defense integrated intelligent decision-making method under the threat of explosion impact is the basis for regional emergency repair after the explosion, regional protection before the explosion, and regional strike during the explosion. It is the premise for the construction of regional emergency response system, urban engineering safety and protection system, and efficient strike planning decision-making system.

[0003] The structural mechanical response of regional explosion impact refers to the structural mechanical response of the facilities in the region when the region is impacted by the explosion. These responses specifically include stress, strain, displacement and dynamic acceleration. The regional facility network refers to a network of facilities composed of multiple key facilities and related key facilities, which are interconnected and work together. The emergency repair scheduling decision refers to the decision of an orderly regional emergency repair action plan made in real time for the key facilities that play a major role in the region in order to quickly restore the overall function of the region after the regional explosion. This decision clarifies the priority order and responsibility allocation of emergency repair tasks. The protection plan decision refers to the construction of the optimal protection plan based on protection effectiveness and protection cost before the regional explosion. The protection plan mainly includes protection targets, protection means and protection costs. The strike plan decision refers to the construction of the optimal strike plan based on strike effectiveness and strike cost for the struck area. The strike plan mainly includes strike targets, strike times, strike ammunition at each strike time and strike costs.

[0004] At present, the evaluation of the response of regional facilities under explosion impact is described in the paper "Beirut explosion 2020A case study for a large-scale urban blast simulation" (G. Valsamos, M.Larcher, F. Casadei, Beirut explosion 2020: A case study for a large-scale urban,blast simulation, Safety Science, Volume 137, 2021, 105190, ISSN 0925-7535.), which discloses a numerical simulation method for large-scale urban explosion simulation, including: using open source map tools to obtain urban area geospatial data; using urban area geospatial data to establish an urban finite element model; using numerical simulation methods to perform explosion simulation on the urban finite element model; using the results of numerical simulation to further evaluate the structural damage and casualties of urban buildings.

[0005] However, the above method relies on numerical simulation and makes many assumptions, which means that it can only evaluate the explosion impact response under specific working conditions and has weak generalization ability.

[0006] At present, the evaluation of the structural mechanical response of a single component under blast impact is described in the paper "Deep learning-based prediction of structural responses of RC slabs subjected to blast loading" (Xiao-Qing Zhou, Bing-Gui Huang, Xiao-You Wang, Yong Xia, Deeplearning-based prediction of structural responses of RC slabs subjected to blast loading, Engineering Structures, Volume 311, 2024, 118184, ISSN 0141-0296), which discloses a deep learning-based method for evaluating the structural response of reinforced concrete (RC) slabs, including: constructing a dataset using existing literature data and supplementary numerical simulation data; establishing a multi-layer perceptron neural network (MLP) model to predict the maximum displacement of reinforced concrete (RC) slabs under blast; establishing a one-dimensional convolutional neural network (1D-CNN) model to predict the failure mode of reinforced concrete (RC) slabs; and performing permutation feature importance analysis to determine the influence of input features on the prediction results.

[0007] However, the MLP and 1D-CNN network models used in this method are limited by the simple model structure and can only evaluate the maximum displacement and corresponding failure mode of a single component, and cannot carry out high-precision evaluation of the regional structural mechanical response under the impact of explosion. It can be seen that there is a lack of a high-precision evaluation method for the regional structural mechanical response under the impact of explosion in the field.

[0008] At present, there are methods for evaluating the indestructibility of associated critical facility networks, such as the paper "A method for evaluating the indestructibility of associated critical facility networks based on mean shift" (Jin Xuyu, Zhao Sixiang. A method for evaluating the indestructibility of associated critical facility networks based on mean shift [J]. Systems Science and Mathematics, 2024, 44(07): 1931-1944.), which discloses a method for identifying vulnerable areas and evaluating network indestructibility of associated critical facility networks based on the mean shift method, including: initializing parameters, and using damage circles and scanning circles to obtain regional damage data sets; calculating the maximum flow comprehensive index of the network system at this time through the damage model, and determining the component weights; obtaining key areas and calculating indestructibility through the mean shift method; and verifying the effectiveness of the algorithm through numerical experiments.

[0009] However, the regional damage dataset of the above method is constructed through the damage attenuation function, and the damage attenuation function takes the damage radius as a single independent variable, which cannot achieve accurate assessment of the loss of the associated critical facility network under the explosion impact.

[0010] In addition, the above method realizes the indestructibility assessment of the entire network by weighted summation of the total traffic of each sub-network, so that when facing targets in different regions, the weights of each sub-network need to be recalculated, which in turn affects the adaptability of the loss assessment method, and thus cannot achieve a comprehensive assessment of the overall facility network loss under the impact of explosion in any region.

[0011] At present, the evaluation of the emergency repair value of engineering projects is described in the paper "Evaluation Model of Earthquake Emergency Repair and Construction Value Based on Two-stage Fuzzy Entropy Weight" (Wang Fengshan, Huang Yu, Liu Meng. Evaluation Model of Earthquake Emergency Repair and Construction Value Based on Two-stage Fuzzy Entropy Weight [J]. Journal of China Safety Science, 2010, 20(11): 66-71. DOI: 10.16265 / j.cnki.issn1003-3033.2010.11.009.), which discloses a method for evaluating the emergency repair value of engineering projects after an earthquake based on two-stage fuzzy entropy weight, including: constructing a value feature system of post-earthquake engineering projects; realizing the measurement feature analysis of value indicators by distinguishing different types of value indicators; obtaining value parameters by quantitatively evaluating each value indicator; standardizing the value parameters and calculating the value indicator weights using the information entropy weight theory; and using the fuzzy set theory to calculate the value coefficients of the engineering projects under various value criteria to achieve value evaluation.

[0012] However, the above method obtains the emergency repair value coefficient of each engineering project through the two-stage fuzzy entropy weight method, thereby obtaining the emergency repair order of each engineering project, but ignores the impact of the spatial location of the emergency repair facilities and the emergency repair team on the emergency repair plan, and the evaluation is not comprehensive enough.

[0013] At present, the decision-making on the emergency repair scheduling of regional key facilities is described in the paper "Multi-objective joint optimization of maintenance resource allocation and task scheduling" (Liu Shengyu, Qi Xiaogang, Liu Lifang. Multi-objective joint optimization of maintenance resource allocation and task scheduling [J]. Acta Ordnance, 2024, 45(07): 2442-2450.), which discloses an optimization decision-making method for realizing resource allocation and task scheduling in complex sites based on a multi-objective joint optimization model, including complex factors such as multi-center, open, multi-repair state, time window restriction, non-traversal path and capacity constraint in emergency repair scheduling, and establishes the research goal of maximizing maintenance benefits and minimizing risk costs; by introducing a dynamic replenishment strategy, combining regional risks and maintenance costs, constructing a multi-objective optimization mathematical model, and dynamically allocating resources; using the optimized multi-objective artificial bee colony (MOABC) algorithm, especially the MOABC-MMHS algorithm, to effectively handle extreme situations and achieve efficient solutions.

[0014] However, the above methods supplement data through random generation to obtain the importance of facilities, which affects the objectivity and accuracy of resource allocation and task scheduling decisions.

[0015] At present, the design of facility protection schemes is such as the Chinese invention patent "A research method for the protection of oil storage facilities" (application number: CN202310644969.1, publication date: 2023-09-29). This patent uses vertical steel oil tanks as the application scenario, and is aimed at the protection of oil storage facilities and tank bottom corrosion problems. Through the macro and micro multi-scale structural design of the material, the anti-corrosion performance of the oil tank is improved. Similarly, the Chinese invention patent "Method, device, equipment and medium for determining the protection range of line fire protection facilities" (application number: CN202210999049.7, publication date: 2022-10-21) provides a method for determining the protection range of line fire protection facilities, which determines the protection range by calculating the movement time of high-temperature objects and the influence of wind speed on their horizontal acceleration.

[0016] However, the above-mentioned public protection decision-making technologies provide valuable applications in their respective technical fields, but to form a protection plan decision-making method under multi-dimensional explosions, a more comprehensive and result-oriented decision-making method is needed.

[0017] In addition, the existing work has not evaluated the impact of multi-dimensional strike events on the effectiveness of protection schemes; the protection effectiveness evaluation is only for protective clothing ("A fire-proof, nuclear radiation-proof, neutron-ray-proof, biological and chemical-proof split protective clothing", application number: CN202122451234.6, publication date: 2022-04-05), ships ("A method for evaluating the explosion protection effectiveness of anti-ship missiles", application number: CN202211568672.3, publication date: 2023-04-18), There are many single-target areas, such as facility security ("A substation security protection system based on multi-level intelligence", application number: CN202210203419.1, publication date: 2022-07-22), cyber warfare ("Method and device for determining a protection scheme for an attack path", application number: CN202180001195.X, publication date: 2021-08-06), etc., and there is a lack of research on the protection effectiveness of regional targets such as building complexes and system function cascades against explosion impact firepower.

[0018] At present, most technologies for optimizing strike plans focus on multi-target strike plans in a joint strike environment, and explore the impact of the surrounding environment on the strike effect. The Chinese invention patent application "Method for Establishing a Real-time Joint Strike Optimization Model" (application number: CN202110290487.1, publication date: 2021.08.06) discloses a method for establishing a real-time joint strike optimization model, including: obtaining a data set for establishing a joint strike optimization model, the data set includes weather environment data, target feature data, interception equipment data, and the occlusion relationship between the target and the interception equipment, and the interception equipment includes laser equipment, radio equipment, and flexible net equipment; according to the data set, the spatial dimension constraints, time dimension constraints, resource dimension constraints, weather dimension constraints, and environmental dimension constraints of each interception equipment are established; according to these constraints, an interception weight factor is established, and a joint strike optimization model is established according to the interception weight factor.

[0019] However, while this method improves the feasibility of the strike plan results in terms of model building, it mainly focuses on a single type of target and lacks the optimal decision for a strike plan that is universally applicable to the entire regional system.

[0020] The Chinese invention patent application "Method for Rapid Assessment of the Feasibility of Multi-target Strike Missions" (application number: CN202310364835.4, publication date: 2023.04.07) discloses a method for rapid assessment of the feasibility of multi-target strike missions, comprising the following steps: establishing a feasibility analysis and assessment constraint model for multi-target strike missions based on a strike plan list and a force resource list; constructing a feasibility assessment model for multi-target strike missions based on the established feasibility analysis and assessment constraint model for multi-target strike missions with the minimum number of excess resources as the optimization goal; solving the feasibility assessment model for multi-target strike missions based on an improved genetic algorithm to obtain a conclusion on mission feasibility.

[0021] However, since this method has a lot of assumptions, such as resource availability and target reachability, it may not be consistent with the actual situation. In addition, the so-called multi-target only evaluates the feasibility of the strike mission from a quantitative perspective, and still lacks the strike plan decision-making technology that considers the overall system function damage and cascading effects.

[0022] In summary, in the existing technology, the generalization ability of the evaluation of regional structural mechanical response under explosion impact is not strong, and the evaluation accuracy is not high; under regional explosions, the loss assessment of the associated key facility network is not accurate and comprehensive enough; the decision-making of regional emergency repair scheduling after the explosion, the decision-making of regional protection plan before the explosion, and the decision-making of regional explosion plan are not comprehensive and accurate; ultimately, the regional attack and defense integrated intelligent decision-making under the threat of explosion is also incomplete and has low accuracy. Summary of the invention

[0023] The purpose of the present invention is to provide an intelligent decision-making method for regional attack and defense under the threat of explosion impact, which makes comprehensive and accurate decisions.

[0024] The technical solution to achieve the purpose of the present invention is:

[0025] An integrated attack and defense intelligent decision-making method for a region under the threat of explosion impact comprises the following steps:

[0026] Construction of an attack-defense integrated intelligent decision-making database: Collect multi-dimensional and multi-target explosion damage effect data, integrate protection technology data collected from multiple channels, obtain numerical simulation explosion data, build an explosion damage effect database and an explosion impact structural mechanical response database, and obtain an attack-defense integrated intelligent decision-making database;

[0027] Construction of a digital-physics-driven intelligent model of the mechanical response of an explosion impact structure: constructing a neural network of the mechanical response of an explosion impact structure, and using the mechanical response database of the explosion impact structure to train the neural network of the mechanical response of the explosion impact structure to obtain a digital-physics-driven intelligent model of the mechanical response of an explosion impact structure;

[0028] Regional network model construction: Use the GIS information system to collect dispatch, protection, and strike regional information in real time, build the corresponding associated key facility network, combine the corresponding regional associated network traffic data set, define network traffic model parameters and variables, and build the dispatch, protection, and strike regional network traffic model;

[0029] Acquisition of regional network traffic: Acquisition of regional network traffic by the regional network traffic model, including regional network traffic before the explosion, regional network traffic after the explosion, and real-time regional network traffic for emergency repair and recovery;

[0030] Construction of regional emergency repair scheduling decision model: Obtain the emergency repair path plan according to the facility emergency repair priority, combine the real-time regional function recovery index, build a multi-objective optimization function, and obtain the regional key facility emergency repair scheduling decision model;

[0031] Construction of regional weak point and vulnerability assessment model: Monte Carlo method is used to design random strikes and construct regional explosion damage circle; mean shift method is used to obtain regional weak point and regional vulnerability assessment results in combination with the regional explosion damage circle, and regional weak point and vulnerability assessment model is constructed;

[0032] Construction of regional protection plan decision model: According to the type and vulnerability of regional weak point facilities, quantify the protection measures of key nodes of the system, obtain the quantitative protection measures of the weak points of the protection area, and establish a multi-objective constraint and multi-standard analysis model based on the cost-effectiveness ratio of each protection plan to form the optimal protection plan and construct a regional protection plan decision model;

[0033] Construction of regional strike plan decision model: According to the type of facilities at the weak points in the strike area and the regional strike effectiveness index, the strike facilities are obtained, the cost of the strike plan is estimated, and the optimal ammunition use plan is planned according to the ammunition performance, target type and protection level. According to the best cost-effectiveness ratio of the weak point facilities under the set strike index, the optimal strike plan is dynamically selected to build a regional strike plan decision model;

[0034] Intelligent decision-making for integrated offense and defense under the threat of explosion impact: Based on the decision-making requirements and on-site decision-making parameters obtained through real-time communication, the corresponding decision-making model is called to make decisions and obtain intelligent decision-making results for integrated offense and defense.

[0035] Furthermore, the intelligent decision database construction step includes:

[0036] Damage effect data collection: Collect explosion damage effect data covering various force field types and various damage modes, including experimental data, historical accident data and computer simulation data;

[0037] Protection technology data collection: Collect protection technology data from test data, historical data and simulation data;

[0038] Numerical simulation explosion data acquisition: by designing numerical simulation explosion input data, batch numerical simulation is performed to obtain numerical simulation explosion data;

[0039] Database construction: Build an integrated attack and defense intelligent decision-making database consisting of an explosion damage effect database and an explosion impact structural mechanics response database.

[0040] Furthermore, the steps of constructing the digital-physical driven intelligent model include:

[0041] Data set division: dividing the explosion impact structural mechanical response database into a training data set and a verification data set;

[0042] Construction of neural network for mechanical response of explosion structures: Construction of a graph neural network that embeds physical information to predict the mechanical response of explosion structures;

[0043] Acquisition of a mathematical-physics-driven intelligent model of the mechanical response of a structure to an explosion impact: using the training data set to train the graph neural network in which the physical information is embedded, and using the verification data set to optimize the graph neural network in which the physical information is embedded, to obtain a mathematical-physics-driven intelligent model of the mechanical response of a structure to an explosion impact.

[0044] Furthermore, the regional network model building step includes:

[0045] Regional information collection: regional information is collected in real time through the GIS information system. The regional information includes the dispatching, protection, and strike regional function types, dispatching, protection, and strike facility function characteristics, and emergency repair dispatching data;

[0046] Construction of the network of associated key facilities: According to the regional functional type and the functional characteristics of the facilities, the regional functional system, the key facilities of each functional system, and the associated key facilities of the regional functional system are determined, and the network of associated key facilities is constructed according to the flow attributes between the facilities;

[0047] Construction of regional correlation network flow data set: Quantify the facility working parameters in the functional characteristics of the facilities to obtain the key facility flows and construct the regional correlation network flow data set;

[0048] Regional network model construction: define network traffic model parameters and variables, map the regional associated network traffic data set to the associated key facility network, and obtain a regional network traffic model, including a scheduling regional network traffic model, a protection regional network traffic model, and a strike regional network traffic model.

[0049] Furthermore, the regional network traffic acquisition step includes:

[0050] Acquisition of regional network traffic before the explosion: acquiring regional network traffic before the explosion by using the regional network traffic model;

[0051] Acquisition of regional network traffic after the explosion: By defining the facility function loss function, the function loss of key facilities is obtained, and combined with the regional network traffic model, the regional network traffic after the explosion is obtained;

[0052] Acquisition of real-time regional network traffic for emergency repair and recovery: Acquire the actual functional loss of facilities after the explosion, obtain the real-time traffic of each key facility for emergency repair and recovery, and take the sum of the real-time traffic of the underlying demand node facilities as the real-time regional network traffic for emergency repair and recovery;

[0053] Acquisition of regional network traffic: The regional network traffic is composed of the regional network traffic before the explosion, the regional network traffic after the explosion, and the real-time regional network traffic for emergency repair and recovery.

[0054] Furthermore, the steps of constructing the regional emergency repair dispatch decision model include:

[0055] Acquisition of emergency repair priority: Quantify the functional loss of key facilities, the flow attributes of key facilities, and the emergency repair time limit of key facilities. Through data processing, construct a classification index of emergency repair priority of key facilities, and combine the two-stage fuzzy entropy weight method to obtain the comprehensive fuzzy evaluation emergency repair priority of each key facility;

[0056] Construction of regional key facility emergency repair scheduling model: The emergency repair path plan is obtained according to the facility emergency repair priority, and a multi-objective optimization function is constructed in combination with real-time regional function recovery indicators. The resource satisfaction, comprehensive facility emergency repair and non-repetitive emergency repair are used as constraints to obtain the regional key facility emergency repair scheduling decision model.

[0057] Furthermore, the steps of constructing the regional weak point and vulnerability assessment model include:

[0058] Acquisition of regional explosion damage circle: Monte Carlo method is used to design random strikes with different explosion equivalents, explosion locations, and strike ranges in the region, obtain the functional loss of each key facility, update the network traffic of key facilities, and construct a regional explosion damage circle;

[0059] Vulnerability assessment model construction: The mean shift method is used to continuously update the regional explosion damage circle, so as to obtain regional weak points and regional vulnerability assessment results, and build a regional weak point and vulnerability assessment model.

[0060] Furthermore, the step of constructing the regional protection scheme decision model includes: constructing a regional explosion damage circle,

[0061] Quantification of protection measures for weak points in the protection area: Quantify the protection measures for key nodes of the system according to the type of facilities at the weak points in the protection area and the regional protection effectiveness index, and obtain the quantitative protection measures for the weak points in the protection area;

[0062] Protection scheme cost-effectiveness evaluation: Based on the types of weak point facilities in the protection area and the regional protection effectiveness indicators, generate protection schemes for key urban node functions under multiple random explosion events, and evaluate the protection effectiveness and protection cost of each protection scheme to obtain the cost-effectiveness ratio of each protection scheme;

[0063] Construction of regional protection scheme decision model: Based on the cost-effectiveness ratio of each protection scheme, the protection scheme and its corresponding protection cost of all key nodes of the comprehensive system that meet the set protection indicators under the protection measures are established, a multi-objective constraint and multi-standard analysis model is established, the optimal protection scheme is formed, and a regional protection scheme decision model is constructed.

[0064] Furthermore, the strike plan decision model construction step includes:

[0065] Calculation of strike effectiveness and cost: Estimate the cost of the strike plan based on the type of facilities at the weak points in the strike area and the regional strike effectiveness indicators;

[0066] Construction of regional strike plan decision-making model: Plan the optimal ammunition use plan based on ammunition performance, target type and protection level, and dynamically select the optimal strike plan based on the best cost-effectiveness of key infrastructure under the set strike indicators.

[0067] Furthermore, the regional attack and defense integrated intelligent decision-making steps include:

[0068] Decision demand acquisition: through real-time communication, the decision demand and on-site decision parameters are acquired, the decision demand includes the area scope and decision type; the decision type includes the regional emergency repair scheduling decision, the regional protection plan decision or the regional strike plan decision; the decision parameters include the regional emergency repair scheduling decision parameters, the regional protection plan decision parameters and the regional strike plan decision parameters;

[0069] Decision model calling: calling the decision model corresponding to the decision requirement;

[0070] Offense-defense integrated intelligent decision-making: The decision information of the called decision model within the said area is output as the result of the offense-defense integrated intelligent decision-making.

[0071] Compared with the prior art, the present invention has the following significant advantages:

[0072] 1. More comprehensive decision-making: The present invention establishes an explosion impact structural response database with random regional characteristics, random facility characteristics, and random explosion characteristics, constructs a regional facility network with upper and lower network connections by analyzing the supply and demand relationship of related key facilities, and constructs a multi-objective optimization function for emergency repair based on resource satisfaction, facility emergency repair priority, emergency repair path plan, and real-time regional function recovery index. By using a weak point and vulnerability assessment model, the regional weak point facility type, regional protection effectiveness index, and regional strike effectiveness index are obtained, thereby improving the comprehensiveness of regional structural mechanical response assessment under explosion impact, regional facility network construction, regional emergency repair scheduling decision, regional protection plan decision, and regional strike plan decision, so that the offensive and defensive decision-making factors are more comprehensive.

[0073] 2. More accurate results: The present invention embeds physical information of the mechanical response of the explosion impact structure into a graph neural network, obtains the functional loss of each key facility based on the geometric data before and after the explosion of the key facilities, constructs a classification index for the priority of facility repair, decides the optimal protection plan based on the protection effectiveness and protection cost, constructs the strike cost-effectiveness ratio based on the strike effectiveness and strike cost, and decides on the optimal strike plan, thereby improving the accuracy of regional structural mechanical response assessment under explosion impact, regional facility network construction, regional emergency repair scheduling decision-making, regional protection plan decision-making, and regional strike plan decision-making, making the factors considered in offensive and defensive decisions more accurate.

[0074] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 It is the main flow chart of the regional attack and defense integrated intelligent decision-making method under the threat of explosion impact of the present invention.

[0076] Figure 2 yes Figure 1 Flowchart of the steps for building an intelligent decision-making database that integrates offense and defense.

[0077] Figure 3 Structural diagram of an example of a facility base circle design.

[0078] Figure 4 It is explosion point 1 Schematic diagram of the structure of an example of a constructed overpressure time history curve.

[0079] Figure 5 It's the explosion point 2 Schematic diagram of the structure of an example of a constructed overpressure time history curve.

[0080] Figure 6 yes Figure 1 Flowchart of the steps for constructing a mathematical-physics-driven intelligent model of the mechanical response of structures subjected to explosion impact.

[0081] Figure 7 yes Figure 1 Flowchart of the steps in building the regional network model.

[0082] Figure 8 is a structural diagram of an example regional facility network.

[0083] Fig. 9 is a structural diagram of an example network traffic model.

[0084] Fig.10 yes Figure 1 Flowchart of the steps for obtaining network traffic in the middle area.

[0085] Fig.11 is a structural diagram of an example of network traffic in the area after the explosion.

[0086] Fig.12 yes Figure 1 Flowchart of the steps for building the regional emergency repair dispatch decision model.

[0087] Fig.13 is a structural diagram of an example of real-time regional functional restoration network traffic.

[0088] Fig.14 yes Figure 1 Flowchart of the steps for constructing a regional weakness and vulnerability assessment model.

[0089] Fig.15 yes Figure 1 Flowchart of the steps for building the decision model for regional protection schemes.

[0090] Fig.16 yes Figure 1 Flowchart of the steps for building the decision-making model for the regional strike plan.

[0091] Fig.17 yes Figure 1Flowchart of the steps for intelligent decision-making for regional attack and defense integration under the threat of medium explosion impact. DETAILED DESCRIPTION

[0092] like Figure 1 As shown, the present invention provides an integrated attack and defense intelligent decision-making method for a region under the threat of explosion impact, comprising the following steps:

[0093] S10. Construction of an attack-defense integrated intelligent decision-making database: Collect multi-dimensional and multi-target explosion damage effect data, integrate protection technology data collected from multiple channels, obtain numerical simulation explosion data, build an explosion damage effect database and an explosion impact structural mechanics response database, so as to realize the construction of an attack-defense integrated intelligent decision-making database.

[0094] like Figure 2 As shown, the S10, construction of an attack-defense integrated intelligent decision-making database, includes the following steps:

[0095] S11. Damage effect data collection: Collect explosion damage effect data covering various force field types and various damage modes, including experimental data, historical accident data and computer simulation data.

[0096] By collecting data from multiple channels, including experimental data, historical accident data, and computer simulation data, which cover a variety of force field types and damage modes, we can comprehensively obtain explosion damage effect data from multi-dimensional strikes and multi-target attacks.

[0097] For example, the source of experimental data can be explosion tests, etc., the source of historical accident data can be industrial explosion records, etc., and the source of computer simulation data can be Abaqus simulation results, etc.

[0098] Force field types, such as fuel-air explosives, penetrating bombs, explosive-killing bombs, thermobaric bombs, and other ammunition types;

[0099] Target damage modes, such as spatial damage to urban core node components such as buildings, bridges, tunnel shelters, damage to heat-sensitive facilities, and different damage modes such as local damage and continuous collapse.

[0100] S12. Protection technology data collection: Collect protection technology data from test data, historical data, and simulation data.

[0101] The protection technology data involves three-dimensional desensitization cascade structure, high temperature resistant and high resistance structure, camouflage protection technology; it also includes a variety of advanced protection materials and structural designs, such as high-performance composite materials fiber reinforced polymers, ultra-high performance concrete, etc., energy-absorbing protection structures, specially designed walls, foundations and supporting structures, the use of metal grilles and energy-absorbing layers, the establishment of protective walls and isolation barriers, etc.

[0102] S13. Acquisition of numerical simulation explosion data: by designing numerical simulation explosion input data and performing batch numerical simulation, the numerical simulation explosion data is acquired.

[0103] The step S13, acquiring numerical simulation explosion data, comprises the following steps:

[0104] S131. Design of numerical simulation explosion input data: designing regional characteristic data, facility characteristic data, and explosion characteristic data by a random number method, and combining them according to explosion cases to obtain numerical simulation explosion input data;

[0105] The step S131, designing numerical simulation explosion input data, comprises:

[0106] 1. Acquisition of regional characteristic data: Design regional size, facility type, facility distribution, and facility quantity through random number method to obtain regional characteristics;

[0107] 2. Facility characteristic data acquisition: The geometry of each facility and the material type of each facility are designed by random number method to obtain the facility characteristics;

[0108] By obtaining the facility information input data, the random facility information is obtained by arranging and transforming it to construct the facility node data.

[0109] Acquisition of facility information input data: Classify facilities according to different types of facilities, such as buildings, bridges, oil tanks, etc.; obtain key information such as reasonable materials, structural composition, distance between buildings, etc. of each type of facility according to building design specifications. Randomly design reasonable building geometry and building structure, and convert them into building location information. The schematic diagram of the facility bottom circle design is as follows Figure 3 As shown, the corresponding geometric facility node data:

[0110] ,

[0111] in, represents the numerical simulation node number, Indicates No. Node facility type, Indicates No. Node location information, Indicates No. Node type information, when , indicating that the node is a common component node of the facility. , indicating that the node is a key component node of the facility, and the node will be removed when an explosion occurs. , indicating that the node belongs to the ground; Indicates Number of numerical simulation nodes material types.

[0112] For example: [1,25,25,60,1,2] means that the node belongs to the first type of facility: power facilities; the node location: ; This node is a key component of the facility; The material type of this node is 2, concrete.

[0113] 3. Acquisition of explosion characteristic data: The explosion time, number of explosions, explosion equivalent, and explosion center are designed through the random number method to obtain the explosion characteristics.

[0114] Get the exploded input by random construction:

[0115] ,

[0116] in, Indicates the number of the center node of the numerical simulation explosion. is the explosion center of numerical simulation, Indicates the number of explosions at the explosion center. Indicates the explosion center The moment the explosion occurred, Indicated in The explosive equivalent at the moment.

[0117] For example: ,

[0118] Example shows that in the area and Explosions occurred at each of the two sites. There were 2 explosions: When the first explosion occurs, the explosion equivalent is , No. When the second explosion occurs, the explosion equivalent is ; Explosion point 2 1 explosion occurred: When , an explosion occurs, and the explosion equivalent is ,Here, data expansion is used to align the data dimensions of the explosion events at two moments.,Constructing the explosion shock wave overpressure time history load:,obtain multiple explosion shock loads through the overpressure time history curve.,Key parameters for constructing the curve: explosion time, explosion peak overpressure,,explosion positive pressure time.

[0119] The explosion center distance is obtained based on the facility node and the explosion center, as shown in the formula As shown; according to the explosion equivalent and the distance from the explosion center, the proportional distance is obtained, as shown in the formula As shown; according to different proportional distances, the corresponding peak overpressure and shock wave positive pressure time are obtained by reasonably using the Sadovskyi formula, Henrych formula, etc.; the explosion time is simplified to the shock wave arrival time.

[0120] For example: The explosion input information is as follows:

[0121] ,

[0122] Assume the node position is .

[0123] ,

[0124] in, The explosion point 1 Explosive distance: , explosion point 2 Explosive distance: .

[0125] (2)

[0126] in, is the proportional distance.

[0127] At explosion point 1 : In the When the first explosion occurs, for the node Scale distance: , in When , the second explosion occurs, for the node Scale distance: At explosion point 2 : In the When , an explosion occurs, for the node Scale distance: .

[0128] For example, Henrych's peak overpressure formula:

[0129] .

[0130] Henrych positive pressure action time formula:

[0131] .

[0132] Sadovski peak overpressure formula:

[0133] .

[0134] Sadovski positive pressure action time formula:

[0135] ,

[0136] In the above empirical relationship, represents the peak overpressure, Indicates the positive pressure action time, represents the proportional distance, Indicates explosive equivalent.

[0137] At explosion point 1 : In the When the first explosion occurs, for the node Peak overpressure: , positive pressure time: , in When , the second explosion occurs, for the node Peak overpressure: , positive pressure time: At explosion point 2 : In the When , an explosion occurs, for the node Peak overpressure: , positive pressure time: .

[0138] The explosion time is set to the time when the shock wave arrives. Explosion point 1 The constructed overpressure time history curve is shown in Figure 4 As shown, explosion point 2 The constructed overpressure time history curve is shown in Figure 5 shown.

[0139] 4. Acquisition of numerical simulation explosion input data: According to the explosion case, three-dimensional data is constructed based on the above regional characteristics and facility characteristics. The above three-dimensional data is combined with the explosion characteristics to obtain the numerical simulation explosion input data:

[0140] [ ].

[0141] S132. Acquisition of numerical simulation explosion data: performing batch numerical simulation calculations on the numerical simulation explosion input data according to explosion cases, acquiring the structural mechanical responses of the corresponding explosion cases, and combining the structural mechanical responses with the numerical simulation explosion input data to obtain numerical simulation explosion data.

[0142] 1. Structural mechanics response data example: [ ],

[0143] in, Indicates the number The stress at the node, Indicates the number The strain at the node, Indicates the number The displacement of the node, For this node Dynamic response acceleration at all times. This part is also used as the training truth value.

[0144] 2. Example of numerical simulation explosion data:

[0145] [ ].

[0146] Through the acquired explosion information input data and random facility construction information, the numerical calculation input of the structural mechanical response of the explosion impact facility is designed through random combination and rationality verification. Batch numerical simulation is designed using python and numerical simulation software to quickly obtain the structural mechanical response of the facility nodes of each random explosion simulation event. The facility structural mechanical response consists of stress, strain, displacement and dynamic response acceleration. The structural mechanical responses of the explosion event facility are combined to obtain the true value data for each explosion event training.

[0147] S14. Database construction: The attack-defense integrated intelligent decision-making database consists of an explosion damage effect database and an explosion impact structural mechanics response database.

[0148] The step of constructing a database in step S14 includes:

[0149] S141. Establishment of explosion damage effect database: The damage effect data and protection technology data are pre-processed, classified and integrated, and entered into the library system to establish an explosion damage effect database.

[0150] Before entering the data into the database system, the ambiguity, differences and heterogeneity of the data in structure and semantics should be unified.

[0151] The steps of establishing the S141 explosion damage effect database include:

[0152] S1411. Data preprocessing: remove or correct outliers, standardize feature values, and normalize data;

[0153] Extract the feature vector Q=[q1, q2, …, qn] from the original data, standardize the feature values ​​so that they are on the same scale, and use Min-Max or Z-score normalization, whichever is better, depending on the type and distribution of the data;

[0154] S1412, Data feature classification and fusion: Use convolutional neural network to classify and fuse data features;

[0155] S1413. Establishment of explosion damage effect database: Data entry system, establishment of explosion damage effect database.

[0156] The present invention collects and integrates data from multiple sources and types, including historical data, experimental data and numerical simulation data, covering a variety of ammunition types and target damage effects, comprehensively evaluates physical damage, functional loss and its cascading impact on the system, and improves the accuracy of strike plans and the comprehensiveness of decision support.

[0157] S142. Establishment of explosion impact structural mechanics response database: Integrate test and historical explosion data and numerical simulation explosion data, perform data processing, obtain multiple sets of explosion impact input data and corresponding structural mechanics response data, combine the explosion impact input data and structural mechanics response data according to explosion cases, and obtain the explosion impact structural mechanics response database.

[0158] The step S142 of establishing a database of explosion impact structural mechanical responses includes:

[0159] S1421. Explosion data extraction: Classify the test and historical explosion data, and numerical simulation explosion data according to data characteristics, clean the erroneous, incomplete or inconsistent data in the above classified data, and extract the explosion data that meets the requirements.

[0160] S1422. Data conversion and combination: Format the explosion data that meets the requirements according to the data type, standardize the formatted data, and combine the standardized data according to the regional characteristics, facility characteristics, explosion characteristics, and structural mechanical responses of each explosion case to achieve data conversion and combination.

[0161] 1. Data formatting: Convert the raw data in the digital-physics-driven intelligent model into format data recognizable by the neural network, and collect data features from discrete grid nodes.

[0162] For example, the original data is reshaped into a continuous feature vector form, and the data is expanded or compressed in a certain order into a storage form recognizable by the neural network through interpolation or sampling methods.

[0163] Use graph-structured data to directly collect features from discrete grid nodes to avoid loss of accuracy during data reshaping, and inherit the spatial characteristics of these grids into the graph according to the graph data structure.

[0164] 2. Data standardization: According to different data features of the formatted data, a data standardization method is used to map the numerical ranges of different data features to similar intervals.

[0165] For example, the standardization method for the mechanical response data of explosion impact structure is shown in Table 1. is the original data, is the standardized data, parameter for The mean value, parameter for The standard deviation of .

[0166] Table 1 Standardization methods

[0167] .

[0168] The explosion impact input data is obtained by combining the test and historical explosion data, the explosion information and facility information under each explosion event in the random simulation model. The explosion impact input data and the corresponding input structural mechanics response data are integrated according to the explosion event to obtain the explosion impact structural mechanics response data. The data structure is as follows:

[0169] S1423, explosion impact input data construction: The above data obtained by combining the regional characteristics, facility characteristics, and explosion characteristics of each explosion case are used to construct the explosion impact input data according to the following data structure:

[0170] [ ],

[0171] in, Indicates the node number, Indicates No. Node facility type, Indicates No. Node location information, Indicates No. Node type information, when , indicating that the node is a common component node of the facility. , indicating that the node is a key component node of the facility, and the node will be removed when an explosion occurs. , indicating that the node belongs to the ground; Indicates No. Node material type; Indicates the number of explosions at the explosion center. Indicates the time when the explosion occurred. Indicated in At the node The explosion equivalent at the moment, For the corresponding The moment of occurrence The distance from the explosion center at node No.

[0172] S1424, structural mechanics response data construction: The above data obtained by combining the structural mechanics responses of each explosion case are combined into the following data structure: structural mechanics response parameter data:

[0173] [ , ],

[0174] in, Indicates the number The stress at the node, Indicates the number The strain at the node, Indicates the number The node displacement, For this node Dynamic response acceleration at all times. This part is also used as the training truth value.

[0175] S1425. Establishment of explosion impact structural mechanics response database: storing the above-mentioned explosion impact input data and structural mechanics response data as an explosion impact structural mechanics response database.

[0176] Explosion impact structural mechanical response data:

[0177] [ ].

[0178] The present invention ensures the diversity of the database and the robustness of the neural network by establishing an explosion impact structural response database with random regional characteristics, random facility characteristics, and random explosion characteristics, thereby realizing a generalized evaluation of the structural mechanical responses to regional characteristics, facility characteristics, and explosion characteristics under explosion impact.

[0179] S143. Database construction: Combine the explosion damage effect database and the explosion impact structural mechanics response database to realize the construction of an integrated attack and defense intelligent decision-making database.

[0180] S20. Construction of a mathematical-physical driven intelligent model of the mechanical response of an explosion impact structure: A mathematical-physical driven intelligent model of the mechanical response of an explosion impact structure is obtained by constructing a neural network of the mechanical response of an explosion impact structure and using the mechanical response database of the explosion impact structure to train the neural network of the mechanical response of an explosion impact structure.

[0181] like Figure 6 As shown, the S20, explosion impact structural mechanical response digital-physics driven intelligent model construction includes the following steps:

[0182] S21, data set division: dividing the explosion impact structural mechanical response database into a training data set and a verification data set;

[0183] S22. Construction of neural network for mechanical response of explosion structures: Construction of a graph neural network embedded with physical information for predicting mechanical response of explosion structures;

[0184] The neural network for the mechanical response of explosion structures is constructed by incorporating the key control equations in the structural mechanical response as constraints into the loss function of the graph neural network consisting of a graph neural network encoding layer, a graph neural network processing layer, and a graph neural network decoding layer.

[0185] The step S22, constructing a neural network for the mechanical response of the explosion impact structure, comprises:

[0186] S221. Graph neural network construction: construct a graph neural network including a graph neural network encoding layer, a graph neural network processing layer, and a graph neural network decoding layer;

[0187] The graph neural network encoding layer is the part that constructs the first layer of potential graph data; the graph neural network processing layer is the part that updates the graph data, and obtains the last layer of graph data through multiple iterations of the graph neural network processing layer; the graph neural network decoding layer is the part that transforms the graph data according to the model's tasks and the nature of the problem to obtain the final prediction result;

[0188] The graph neural network consists of a graph neural network encoding layer, a graph neural network processing layer, and a graph neural network decoding layer. The graph neural network encoding layer is the part that constructs the first layer of potential graph data; the graph neural network processing layer is the part that updates the graph data, and obtains the last layer of graph data through multiple iterations of the graph neural network processing layer; the graph neural network decoding layer is the part that transforms the graph data according to the model's tasks and the nature of the problem to obtain the final result.

[0189] The step S221, graph neural network construction step includes:

[0190] 1. Construction of the graph neural network encoding layer: A multi-layer fully connected neural network is used, and the ReLU activation function is introduced for nonlinear processing to construct the graph neural network encoding layer;

[0191] In the process of building a graph data structure suitable for graph neural network (GNN) processing, the core function of the input layer is to convert all the original grid node data into Mapping to a graph representation rich in high-dimensional information ,This conversion process lays a solid foundation for the subsequent graphics processing layer.

[0192] The graph neural network input layer construction includes:

[0193] Automatically establish association pairs between nodes through spatial proximity (such as radius search method, Kmeans), and then use the graph encoding layer to construct edge features , represents the connection relationship between node Mi and node Mj;

[0194] Design a multi-layer fully connected neural network architecture, aiming to give the network powerful feature capture capabilities through embedding operations in high-dimensional space;

[0195] In terms of nonlinear processing, the ReLU activation function is introduced. Its unique positive value retention property prompts the network to focus more on key features that have a significant impact on the task objectives during the learning process, thereby effectively improving the efficiency and accuracy of the training process.

[0196] Through the graph neural network encoding layer, association pairs between nodes are established and edge features are constructed to obtain the first layer of potential graph data, laying a solid foundation for subsequent graph neural network processing.

[0197] 2. Construction of graph neural network processing layer: In the graph neural network processing layer, the potential graph data obtained by the encoding layer Iterative processing is performed through multiple graph neural network (GN) blocks with the same structure to achieve in-depth analysis and feature extraction of data; each GN block iteration includes the update of two key information, namely edge information and point information update;

[0198] Regarding the GN block: the update of the side information is achieved by aggregating the starting point, end point and side information of the previous iteration. This process can be expressed as:

[0199] ,

[0200] Among them, The output of the vertex and edge iterations is Input times, Indicates The side information input at the iteration, and Respectively represent The starting point and end point information connected to this edge entered in the iteration, Indicates The side information output by the iteration. Function The side information is learned and updated through a fully connected network.

[0201] And point information The update is done by aggregating the original node information and all updated side information connected to it To achieve:

[0202] ,

[0203] Among them, The output of the th vertex and edge iteration is the input of the th, Indicates The node information entered at the iteration, Indicates the node connected to The side information set after the iteration update. Indicates The newly generated node information of the iteration. Function Point information is also updated through the fully connected network.

[0204] After each GN block iteration, the updated node information and side information Will be recombined to get Graph data for iterations Through the continuous iteration of multiple GN blocks, the model can capture the complex node and edge relationships in the graph data, and finally generate highly abstract and characterized graph data. , and feed it into the graph decoding layer to generate the result.

[0205] 3. Construction of graph neural network decoding layer: The main task of the graph decoding layer is to convert the highly abstract graph data output by the graph processing layer Mapped back to the original space with actual physical meaning. The input of the graph neural network decoding layer is the high-dimensional graph data of the last layer of the graph neural network processing layer. The node information of the multi-layer fully connected network can be restored to obtain the node information data with stress, strain, displacement, and dynamic response acceleration:

[0206] ,

[0207] in, It is the grid node information data obtained by the final iteration of the graph processing layer. are the predicted node parameters.

[0208] S222. Embedding physical information: constructing physical information constraints through explosion impact structural response control equations, adding the physical information constraints to the loss of the graph neural network, and obtaining a graph neural network embedded with physical information.

[0209] The step of embedding physical information in step S222 includes:

[0210] 1. Construction of physical information constraint loss term: Physical information constraint loss term Obtained from the residual of the elastic-plastic mechanics constraint equation under explosion impact: The dynamic equilibrium equation constraint describing the relationship between external force and stress , geometric equation constraints describing the relationship between strain and displacement :

[0211] .

[0212] The dynamic equilibrium equation for the relationship between external forces and stresses is described by the constraint loss term:

[0213] ,

[0214] in, Predicted stress tensor , is the material density, is the external force per unit mass, is the predicted dynamic acceleration component obtained from the dynamic mechanical response.

[0215] The geometric equations describing the relationship between strain and displacement are described by the constraint loss term:

[0216] ,

[0217] in, is the component of the predicted strain tensor , , represents the predicted displacement component.

[0218] 2. Physical information constraint embedding: Constructing data loss term through mean square error (MSE):

[0219] ,

[0220] in, Indicates the use of MSE method to obtain data loss, represents the number of nodes in the graph, Indicates The truth value of each node, Indicates The predicted value of a node.

[0221] By adding physical information constraints to the loss, the embedding of physical information constraints is achieved:

[0222] .

[0223] Embedding physical information is achieved by adding physical information constraints to the loss, which is mainly reflected in the loss term of the model. The loss term of the model is composed of the data loss term and the physical information constraint loss term Since the final prediction focuses on the mechanical response field of explosion impact, the physical information constraint loss term Loss of elastic-plastic mechanics constraint equations under explosion impact: Dynamic equilibrium equation constraints describing the relationship between external force and stress , geometric equation constraints describing the relationship between strain and displacement Since the material type is selected in the training input data, the constructed model can achieve material adaptation, so the physical equation describing the stress-strain relationship in the elastic-plastic mechanics constraint is not added to the constraint equation.

[0224] S23. Acquisition of a mathematical-physics-driven intelligent model of the mechanical response of an explosion impact structure: using the training data set to train the neural network of the mechanical response of an explosion structure, and using the verification data set to optimize the neural network of the mechanical response of an explosion structure, so as to obtain a mathematical-physics-driven intelligent model of the mechanical response of an explosion impact structure for evaluation.

[0225] The step S23, obtaining a digital object driven intelligent model, includes:

[0226] S231, Model training: The Adam algorithm is used to update the model parameters. The model is trained in a supervised manner. The model parameters are randomly generated at the beginning of training. The back propagation algorithm is needed to back propagate the value of the loss function and calculate the gradient of the model parameters relative to the loss value to update the model parameters. The selected back propagation algorithm is the Adam algorithm, which combines the characteristics of gradient descent and momentum optimization and has good performance and convergence speed.

[0227] The Adam algorithm uses the following formula in each iteration:

[0228] 1. Calculate the gradient :

[0229] ,

[0230] In the formula, is to find the gradient of the loss function with respect to the model parameters, Represents the loss of the model. Obviously, the size of the gradient is affected by the loss function used and the model structure.

[0231] 2. Update the first-order moment estimate:

[0232] ,

[0233] In the formula, is the updated first-order moment estimate, is the first-order moment estimate from the previous step, is the exponential decay rate of the first-order moment estimate, usually taken as 0.9.

[0234] 3. Update the second-order moment estimate:

[0235] ,

[0236] In the formula, is the updated second-order moment estimate, is the second-order moment estimate from the previous step, is the exponential decay rate of the second-order moment estimate, usually taken as 0.999.

[0237] Since the first and second moment estimates are close to zero at the beginning of the calculation, this may lead to and , and In order to correct these deviations, the Adam algorithm introduces a correction factor to correct the first-order moment estimate and the second-order moment estimate:

[0238] ,

[0239] ,

[0240] In the formula, for The first-order moment estimate after iterations is: for The second moment estimate after iterations.

[0241] 4. Use the corrected first-order and second-order moment estimates to calculate updates to the model parameters:

[0242] ,

[0243] in, are the model parameters before updating, are the updated model parameters, is the learning rate, It is a small number added to avoid the denominator being 0, usually 10e-8. This formula is the key step in updating the parameters, and each parameter will be updated in an adaptive manner. It should be noted that all parameters ( , , , ) need to be set before training and usually will not change during training.

[0244] S232. Model optimization: When performing hyperparameter tuning, you need to pay attention to a series of key hyperparameters, including the number of iterations, batch size, learning rate, and adjustment of the model structure. The key points of adjustment are as follows:

[0245] 1. The number of iterations is between 50 and 300 epochs;

[0246] 2. The batch size can be adjusted between 2 and 128;

[0247] 3. The learning rate is usually adjusted between 0.0001 and 0.01;

[0248] 4. Adjustment of the model structure is also a key step in optimization. According to the gradient calculation formula, the design of the structure affects the gradient update direction. Therefore, the model structure needs to be modified according to the model prediction results, including adjusting the number of layers in the model, the number of units in each layer, and selecting the appropriate graph network layer configuration.

[0249] S233. Model evaluation: Model evaluation is a crucial step in the training process. It not only helps us determine whether the performance of the model meets expectations, but also guides us on how to further adjust and optimize the model.

[0250] 1. Divide the original data set into a training data set and a validation data set. The validation data set does not directly participate in model training, but is used to provide feedback on the current training effect of the model.

[0251] 2. It is the prediction error of the model on the test set. The prediction results of the model will be compared point by point with the true value of the numerical calculation to calculate the relative error between them. The test set is completely independent of the original data set and is usually data that has not been seen in model training, which is conducive to further evaluating the generalization performance of the model.

[0252] In addition, the mechanical response of the facility unit: stress, strain, displacement, and acceleration of dynamic response are taken as prediction targets. In some small equivalent cases, the displacement of the facility unit is close to 0. In order to avoid the influence of the denominator, in this study, the calculation formula of the relative error is as follows:

[0253] ,

[0254] In the formula represents a node in the graph, is the node feature for numerical calculation, The node features predicted by the model.

[0255] ,

[0256] in, represents the mean relative error of the nodes, Indicates the maximum relative error of the node.

[0257] The present invention embeds the physical information of the structural mechanical response to explosion impact into a graph neural network, which can not only ensure the model's learning ability for the physical problems of the structural mechanical response under explosion impact, but also ensure the model's learning ability for three-dimensional unstructured data, thereby achieving high-precision evaluation of the regional structural mechanical response under explosion impact.

[0258] S30. Construction of regional network model: Collect dispatching, protection and strike regional information in real time through GIS information system, build corresponding related key facility network, define network traffic model parameters and variables by building corresponding regional related network traffic data set, and build dispatching, protection and strike regional network model.

[0259] like Figure 7 As shown, the steps of S30, constructing a regional network model, include:

[0260] S31. Regional information collection: regional information is collected in real time through the GIS information system. The regional information includes the functional characteristics of the dispatching, protection, and strike areas, the functional characteristics of the dispatching, protection, and strike facilities, and the emergency dispatching data.

[0261] The step S31, collecting regional information, includes:

[0262] S311. Collection of regional functional characteristics of dispatching, protection and strike: Regional functional characteristics include regional functional type and regional facility distribution.

[0263] Regional functional type refers to the collection of main functions or services undertaken in the corresponding area, such as power supply, water supply, oil / gas supply, etc.

[0264] Regional facility distribution refers to the distribution of facilities in a specific area according to certain planning, functional requirements and spatial constraints. This distribution not only involves the location selection of facilities, but also covers the relative relationship, density, scale and interaction between facilities and their surrounding environment.

[0265] S312. Collection of functional characteristics of dispatching, protection and strike facilities: The functional characteristics of dispatching, protection and strike facilities include the flow attributes of the corresponding facilities, the working parameters of the corresponding facilities, and the geometric parameters of the corresponding facilities.

[0266] The facility flow attributes describe the order of matter or energy within or between facilities, including: facility function type and corresponding flow order;

[0267] Facility operating parameters refer to the parameters that ensure the normal operation of the facility, including: the properties of materials or energy within the facility or between facilities, such as the power generation of fuel generators , fuel consumption rate ;

[0268] Facility geometry refers to the parameters of the facility's physical size, shape, and spatial layout, which are usually described by three-dimensional data, including: the facility's three-dimensional model and the facility's spatial position relationship.

[0269] S313. Emergency repair dispatch data collection: The emergency repair dispatch data consists of the geometric parameters of the facilities before and after the explosion, the emergency repair attributes of the facilities, the emergency repair time limit, the emergency repair team resource information, and the shortest path between the facilities.

[0270] 1. The geometric parameters of the facility before and after the explosion refer to the physical size, shape and spatial layout of the facility, which are usually described by three-dimensional data, including: the three-dimensional model of the facility and the spatial position relationship of the facility.

[0271] 2. The attributes of facility emergency repair include: maintenance team, emergency repair facilities, and safety facilities.

[0272] 3. Emergency repair time limit: The emergency repair time limit of the emergency repair facility is obtained based on the facility recovery time limit provided by the expert.

[0273] For example, if the facility recovery time limit is 1 hour, then the corresponding emergency repair facility repair time limit is 1 hour.

[0274] 4. Emergency repair team resource information: The emergency repair team resource information consists of the emergency repair team’s material resources and human resources, and is collected through on-site communications.

[0275] The material resources of the emergency repair team refer to the various material resources needed to restore or improve the function of facilities during the emergency repair process, such as emergency repair tools and materials; the human resources of the emergency repair team refer to all personnel involved in the emergency repair work, including technicians, managers, operators, etc.

[0276] Quantification of material resources and human resources of the repair team: obtain the materials required for the repair facilities through on-site communication. If the corresponding repair team meets the requirements, the value is 1, and if not, the value is 0, thus realizing the quantification of the material resources of the repair team; obtain the human resources required for the repair facilities through on-site communication, and quantify the human resources of the repair team according to a score of 0-10;

[0277] 5. The shortest path between facilities: The shortest path between facilities is obtained through the facility spatial location and facility attributes in the facility geometric parameters and GIS information technology.

[0278] S32. Construction of a network of associated key facilities: According to the regional function type and facility function characteristics in the regional information, determine the regional functional system, key facilities of each functional system, and key facilities associated with the regional functional system, and construct a network of associated key facilities according to the traffic flow attributes between the facilities.

[0279] The step of constructing the corresponding associated key facility network in S32 includes:

[0280] S321, determining the regional function system: obtaining the regional function system by statistics according to the regional function type in the regional information;

[0281] For example: regional functional systems: power supply system, water supply system, oil / gas supply system.

[0282] S322, determining the key facilities of each functional system: according to the functional characteristics of the regional functional system and each facility under the system, statistically obtaining the key facilities of each functional system;

[0283] For example: key facilities of the power supply system: power plants, substation equipment, distribution equipment, key production facilities;

[0284] The facility component area is determined as a key facility based on whether it belongs to a functional system;

[0285] For example: a cultural center is a facility but not a critical facility, while a power plant is both a facility and a critical facility.

[0286] S323, determining key facilities associated with the regional functional system: determining each key facility as one or more of a supply node, a transport node, and a demand node according to the facility flow attribute in the facility function characteristics, and setting a key facility that is both a system supply node and another system demand node as a key facility associated with the regional functional system;

[0287] For example: Fuel generator: a key facility associated with the power supply system and the fuel supply system, serving as both a supply node for the power supply system and a demand node for the fuel supply system.

[0288] S324, regional facility network construction: constructing an associated key facility network according to the facility flow flow sequence in the facility flow flow attribute in the facility functional characteristics.

[0289] Through the steps of determining the regional functional system, determining the key facilities of each functional system, and determining the key facilities associated with the regional functional system, the distribution of key facilities in the network is obtained. The dynamic process of interaction between facilities is revealed through the flow attributes of facility traffic in the regional information, including the origin, path, and end point of traffic. Based on the distribution of the key facilities in the network and the analysis of the flow attributes of facility traffic, a regional facility network is further constructed, such as Figure 8 shown.

[0290] From the above process, combined Figure 8 It can be seen that the regional facility network not only includes each key facility and associated key facilities as nodes, but also represents the traffic flow relationship between facilities through directed edges (or links).

[0291] S33. Construction of regional correlation network traffic data set: quantify the facility working parameters in the functional characteristics of the facility to obtain the key facility traffic and construct a regional correlation network traffic data set.

[0292] The step of constructing the regional associated network traffic data set in S33 includes:

[0293] S331. Acquisition of key facility flow: Quantify the flow of working parameters of each key facility, and obtain the key facility flow through data classification, cleaning and conversion;

[0294] 1. Network traffic quantification: obtain the network traffic quantification of key facilities according to the facility working parameters in the corresponding regional information;

[0295] For example: fuel generator: power generation: , power supply system current: 100, fuel consumption rate: , oil flow rate of oil supply system 22.5;

[0296] 2. Data combination: combining the network traffic of the key facilities according to the flow attributes of the facility traffic in the regional information;

[0297] 3. Data classification: classifying the combined key facility network flow data according to the flow attributes of each facility flow;

[0298] 4. Data cleaning: Data cleaning is achieved by deleting erroneous, incomplete or inconsistent data from the above classified data;

[0299] 5. Data conversion: Format the cleaned data according to the data type, normalize the formatted data, and implement data conversion to obtain the network traffic of key facilities.

[0300] S332. Acquisition of regional associated network traffic data set: Based on the supply and demand relationship of associated key facilities, the upper and lower layer networks are linked in the form of weights to obtain the regional associated network traffic data set.

[0301] The step of obtaining the associated network traffic data set in S332 includes:

[0302] S3321. Supply and demand relationship analysis: extract all related key facility data, analyze each node function type and corresponding node function level, and obtain the supply and demand relationship of related key facilities;

[0303] S3322. Acquisition of regional associated network traffic data set: Map the supply traffic of key facilities in the lower network to the demand traffic of its upper network in the form of weights, thereby unifying the network traffic in the entire region and obtaining the regional associated network traffic data set:

[0304] 1. The key facilities are the key facilities that link the functional networks and serve as the demand nodes of the upper network and the supply nodes of the lower network. The flow attributes of the key facilities build the corresponding relationship between the upper and lower networks.

[0305] For example, a fuel generator belongs to an associated key facility. Its upper network is the fuel supply system, and it serves as a demand node in the upper network. Its lower network is the power supply system, and it serves as a supply node in the lower network.

[0306] 2. Construction of the traffic network for the entire region: Build from the end of the regional traffic, and through the association of key facilities, step by step upward to achieve the construction of the traffic network for the entire region.

[0307] For example: The end of regional traffic refers to the end of regional traffic flow, which only serves as a traffic demand node, not a traffic supply node: key production facilities, corresponding network: power supply system, corresponding network-associated key facilities: fuel generators, normalized fuel generator power supply system network traffic: 0.6, normalized fuel generator fuel supply system network traffic: 0.8, then the weight value from the upper fuel supply system to the lower power supply system is 0.75, that is, all key facilities of the fuel supply system update the traffic with a weight of 0.75.

[0308] ,

[0309] ,

[0310] in, Represents the lower layer network ( ) supplies traffic to the upper layer network ( ), Indicates the supply flow of the associated key facility in the lower-layer network, Indicates the original demand flow of the associated key facility in the upper network. Indicates the traffic of each node in the upper network before the weight is updated, Represents the traffic of each node in the upper network after the weight is updated.

[0311] S34, regional network model construction: define network traffic model parameters and variables, map the regional associated network traffic data set to the regional facility network, obtain regional network traffic to obtain a network traffic model;

[0312] Define and introduce key network traffic model parameters and variables, map the regional associated network traffic data set to the regional facility network, obtain the regional network traffic model before the explosion, and combine the regional damage data set to obtain the regional network traffic model after the explosion;

[0313] The network traffic model consists of a regional network traffic model before the explosion and a regional network traffic model after the explosion. Key network traffic model parameters and variables are defined and introduced to comprehensively describe the characteristics of regional network traffic and the interaction between key facilities. Based on the defined network traffic model parameters and variables, the regional associated network traffic data set is mapped to the key facilities in the associated facility network to obtain a regional network traffic model before the explosion. Combined with the damage data set of the region, the network traffic model is updated to obtain a post-explosion network traffic model that includes the distribution of key facilities, functional network associations, network traffic flow, and loss of key facility functions.

[0314] The step of constructing a network traffic model in step S34 includes:

[0315] S341. Network traffic model parameter definition: define network traffic model parameters; see Table 2 for network traffic model parameter definition:

[0316] Table 2 Network traffic model parameter definition table

[0317] .

[0318] S342. Network traffic model variable definition: define network traffic model variables; the network traffic model variable definition is shown in Table 3:

[0319] Table 3 Network traffic variable parameter definition table

[0320] .

[0321] S343, network traffic model construction: according to the network traffic model parameters and variables, regional associated network traffic data set, build an explosion network traffic model, such as Fig. 9 shown.

[0322] S40, regional network traffic acquisition: regional network traffic is acquired by a regional network traffic model, wherein the regional network traffic is composed of regional network traffic before the explosion, regional network traffic after the explosion, and regional network traffic in real time for emergency repair and recovery.

[0323] like Fig.10 As shown, the steps of S40, obtaining regional network traffic, include:

[0324] S41. Acquisition of regional network traffic before the explosion: Acquisition of the total regional traffic before the explosion using the regional network traffic model.

[0325] The total traffic before the regional explosion is obtained from the network traffic model:

[0326] ,

[0327] ,

[0328] ,

[0329] ,

[0330] ,

[0331] ,

[0332] ,

[0333] ,

[0334] in, represents the total flow rate before the regional explosion, Indicates the key facility at the end of the network traffic flow. This key facility only serves as a demand node. Indicates the number of key facilities, Represents the network traffic of the underlying demand node, The network traffic and the Equation represent the network traffic of the underlying demand node. Describe the traffic balance constraints of key network facilities, describes the maximum capacity constraint, Indicates key facility nodes, Represents the supply node set, demand node set, and transportation node set. Indicates the facility node before the explosion The facility flow, It represents the total actual supply of facilities in the area before the explosion, the total actual demand of facilities in the area before the explosion, and the total actual transportation of facilities in the area before the explosion. , , Indicates key facilities before the explosion Maximum supply, key facilities before explosion Maximum demand, key facilities before explosion Maximum transport volume.

[0335] S42. Acquisition of regional network traffic after the explosion: By defining the facility function loss function, the function loss of key facilities is obtained, and combined with the network traffic model, the network traffic after the explosion is obtained.

[0336] S421. Acquisition of key facility function loss: The mechanical response of each key facility structure is acquired through the explosion impact structural mechanical response digital-physics-driven intelligent model and the real-time geometric information of the facilities after the explosion, and the geometry of the key facilities after the explosion is reconstructed. The regional facility function loss is obtained by the geometric feature recognition method and the facility function loss evaluation function, see formula .

[0337] 1. Acquisition of evaluation parameters for functional loss of emergency repair facilities:

[0338] ,

[0339] ,

[0340] ,

[0341] Among them, the top angle damage ratio , Area Damage Ratio , volume damage ratio Indicates the parameter for assessing the loss of facility function, the number of top corner damages before the explosion , the number of top corner damage after the explosion , the surface area of ​​the facility before the explosion , the damaged surface area of ​​the facility after the explosion , the volume of the facility before the explosion , the volume of facilities damaged after the explosion , which represents the geometric characteristics of the facility function loss before and after the explosion.

[0342] 2. Assessment of functional loss of emergency repair facilities:

[0343] ,

[0344] in, represents the facility function loss assessment index, , , represents the facility function loss assessment parameter, represents the facility function loss assessment weight, : Facility top angle damage ratio, : Facility surface area damage ratio, : Facility volume damage ratio, : Facility top corner damage ratio weight, : Facility surface area damage ratio weight, : Facility volume damage ratio weight.

[0345] S422. Acquisition of regional network traffic after the explosion: Update the network traffic of key facilities through the functional loss of key facilities. See formula ,Construct the post-explosion network traffic model, and obtain the regional network traffic after the explosion by mapping the regional facility network, such as Fig.11 As shown;

[0346] ,

[0347] ,

[0348] ,

[0349] ,

[0350] ,

[0351] ,

[0352] ,

[0353] ,

[0354] represents the total regional flow after the explosion, Indicates the number of explosion strikes. Indicates The underlying demand nodes after the random explosion in the sub-region of network traffic, Indicates Key facility nodes after random explosions in the sub-region of network traffic, , , Region No. Total supply, total satisfaction, and total transportation capacity after random explosions in sub-regions, , , Respectively represent The maximum supply of the supply node, the maximum satisfaction of the demand node, and the maximum capacity of the transportation node after the random explosion of the sub-region are: Describe the traffic balance constraints of key network facilities, Describes the maximum capacity constraint.

[0355] S43. Acquisition of real-time regional network traffic for emergency repair and recovery: Acquire the actual functional loss of facilities after the explosion, and obtain the real-time traffic of each key facility for emergency repair and recovery. Define the sum of the underlying demand nodes as the total real-time regional traffic for emergency repair and recovery.

[0356] Bottom-level demand nodes: key facilities that serve only as traffic demand points and not as traffic supply points, such as regional key production facilities.

[0357] ,

[0358] ,

[0359] ,

[0360] Indicates the total flow of the area restored in real time after emergency repair. Indicates the underlying demand node after emergency repair and recovery of network traffic, Indicates the real-time network traffic of the underlying demand nodes and, Indicates that key facilities before emergency repair and restoration The restored flow, is the weight corresponding to the facility function recovery level, is the network traffic of each node for function recovery weight update, represents the network traffic of each node that recovers the weight update through the combined function, It means that the network traffic of the whole area is constructed from the bottom up. Represents critical facilities obtained through the regional emergency repair network The network traffic, Indicates the network traffic of each node updated in real time during emergency repair.

[0361] S44. Acquisition of regional network traffic: The regional network traffic consists of the acquired regional network traffic before the explosion, the acquired regional network traffic after the explosion, and the real-time regional network traffic for emergency repair and recovery.

[0362] S50, construction of regional emergency repair scheduling decision model: construct a comprehensive fuzzy evaluation emergency repair priority for each key facility, obtain the emergency repair path plan according to the facility emergency repair priority, combine the real-time regional function recovery index, construct a multi-objective optimization function, and obtain the regional key facility emergency repair scheduling decision model;

[0363] like Fig.12 As shown, the step S50, building a regional emergency repair dispatch decision model, includes:

[0364] S51. Obtaining repair priority: Quantify the functional loss of key facilities, the flow attributes of key facilities, and the time limit for repair of key facilities. Through data processing, construct a classification index for the repair priority of key facilities. Combined with the two-stage fuzzy entropy weight method, obtain the comprehensive fuzzy evaluation repair priority of each key facility.

[0365] The steps of obtaining the repair priority in S51 include:

[0366] S511. Construction of priority classification indicators for critical facility repairs: Based on the critical facility recovery time limits provided by experts, the critical facility repair time limits are quantified, and through further data processing, the priority classification indicators for critical facility repairs are obtained.

[0367] The construction of the S511, priority classification index for emergency repair of key facilities includes:

[0368] S5111. Quantification of the time limit for emergency repair of key facilities: Obtain the time limit for emergency repair of the key facilities according to the recovery time limit of the key facilities provided by experts;

[0369] For example, if the critical facility recovery time limit is 1 hour, then the corresponding critical facility emergency repair time limit is also 1 hour.

[0370] S5113. Construction of priority classification indicators for emergency repair of key facilities: Based on the attributes of the key facilities, the above-processed key facility function loss, key facility flow and key facility emergency repair time limit are mapped to the corresponding facilities as priority classification indicators for emergency repair of each key facility.

[0371] S512. Obtaining the weights of the first stage of the fuzzy entropy weight method: by using the key facility repair priority classification index, calculating the entropy value of each index, and obtaining the information utility value, the weights of the first stage of the fuzzy entropy weight method are obtained.

[0372] 1. Entropy calculation:

[0373] ,

[0374] ,

[0375] in, For indicators The entropy value of is the number of key facilities, For the Key facilities The probability of an indicator, Represents the weight in the entropy value calculation formula, through the number of key facilities get.

[0376] 2. Obtaining information utility value:

[0377] ,

[0378] Indicators The information utility value.

[0379] 3. Determination of weights in the first stage of fuzzy entropy weight method: Normalize the information utility value to obtain the weight of each key facility repair priority classification index.

[0380] S513, the second stage comprehensive fuzzy evaluation of the fuzzy entropy weight method: divide the classification indicators of the emergency repair priority of the key facilities into different fuzzy sets, construct a fuzzy evaluation matrix by constructing a membership function, and finally use the weighted average method to realize the second stage comprehensive fuzzy evaluation of the fuzzy entropy weight method to obtain the fuzzy comprehensive evaluation emergency repair priority of each key facility.

[0381] 1. Fuzzy set division: The key facility repair priority classification indicators are divided into parameter fuzzy sets and fuzzy fuzzy sets according to the acquisition method. The functional loss of the repair facilities and the flow of key facilities are taken as parameter fuzzy sets, and the repair time limit of the repair facilities is taken as fuzzy fuzzy sets.

[0382] 2. Construction of fuzzy evaluation matrix: Determine the membership function for the fuzzy set through expert scoring method, so as to describe the membership of each key facility under different evaluation indicators and construct the fuzzy evaluation matrix;

[0383] Among them, the rows of the fuzzy evaluation matrix correspond to the evaluation indicators, the columns correspond to the evaluation levels, and the elements are the corresponding membership values.

[0384] 3. Fuzzy comprehensive evaluation: The fuzzy evaluation matrix and the first-stage weights of the fuzzy entropy weight method are calculated using the weighted average method to obtain the fuzzy comprehensive evaluation repair priority of each key facility.

[0385] S514, obtaining the priority of emergency repair of key facilities: setting the emergency repair priority of each key facility according to the fuzzy comprehensive evaluation as the emergency repair priority of the facility.

[0386] S52. Construction of regional key facility emergency repair scheduling model: Based on the total emergency repair path, real-time regional function recovery indicators, and key facility emergency repair priorities, a multi-objective optimization function is constructed, and resource satisfaction, comprehensive facility emergency repair, and non-repetitive emergency repair are used as constraints to obtain a regional key facility emergency repair scheduling decision model.

[0387] The step S52, constructing a regional key facility emergency repair scheduling model includes:

[0388] S521. Obtaining resource satisfaction: Obtain resource satisfaction by collecting and quantifying the material resources and human resources of each repair team in real time.

[0389] The step of obtaining resource satisfaction in step S521 includes:

[0390] S5211. Collection of material resources and human resources of the emergency repair team: Obtain material resources and human resources of the emergency repair team through on-site communications.

[0391] The material resources of the emergency repair team refer to the various material resources needed to restore or improve the function of facilities during the emergency repair process, such as emergency repair tools and materials; the human resources of the emergency repair team refer to all personnel involved in the emergency repair work, including technicians, managers, operators, etc.

[0392] S5212. Quantification of material resources and human resources of the emergency repair team: Obtain the materials required for the emergency repair facilities through on-site communication. If the corresponding emergency repair team meets the usage, the value is 1, otherwise it is 0, thus realizing the quantification of material resources of the emergency repair team; obtain the human resources required for the emergency repair facilities through on-site communication, and quantify the human resources of the emergency repair team according to a score of 0-10.

[0393] S5213, resource satisfaction acquisition: through the material resources and human resources of the repair team, if the material resource usage or human resources do not meet the requirements, the resources are not satisfied; if both the material resource usage and human resources meet the requirements, available resources are obtained according to the repair capacity of human resources;

[0394] Resource satisfaction: , Indicates material resources. If the usage is met, it is recorded as 1, and if it is not met, it is recorded as 0. It represents human resources, and uses 0-10 to represent the emergency repair capability. If the emergency repair capability is less than 5, emergency repair is not required.

[0395] For example, available resources: , None of them meet the requirements for emergency repair, and their resource satisfaction is 0.

[0396] S522, construction of emergency repair path evaluation function: Based on the facility spatial location and facility attributes in the facility geometric parameters, the shortest path between the facilities is obtained through GIS information technology, and combined with the emergency repair priority, an emergency repair path evaluation function is constructed:

[0397] ,

[0398] ,

[0399] ,

[0400] ,

[0401] ,

[0402] represents the repair path evaluation function considering the repair priority, represents the repair path evaluation set, which is obtained by the repair priority of the facility and the shortest path between the corresponding facilities. Indicates the set of facilities to be repaired. Indicates the repair group. They represent the first facility to be repaired, the first facility to be repaired, and the facility to be repaired later when the repair team arrives. Indicates the emergency repair team. Indicates the emergency repair team Repair facilities The shortest path between Indicates emergency repair facilities The repair priority, Indicates that the facility is under repair , The shortest path between Indicates that the facility is under repair The repair priority, It represents the threshold range of the repair path evaluation function, expressed as a percentage, It represents the non-repetitive constraint condition of the directed path, which is solved by the path optimization algorithm to obtain the emergency repair path plan.

[0403] S523. Real-time regional function recovery index acquisition: define a regional function recovery function, collect and quantify the recovery level of key facility emergency repair functions in real time, map the regional facility network, and obtain real-time regional function recovery indicators.

[0404] The step of obtaining the real-time regional function recovery index in S523 includes:

[0405] S5231, regional function recovery function definition: by obtaining the real-time flow of each key facility before the explosion and after emergency repair, the total regional flow before the explosion and the real-time total regional flow after emergency repair are obtained, and then the regional function recovery function is defined;

[0406] Definition of regional function recovery function: The regional function recovery function is obtained by recovering the real-time traffic of each key facility and the real-time total regional traffic through emergency repair.

[0407] ,

[0408] ,

[0409] ,

[0410] ,

[0411] ,

[0412] ,

[0413] ,

[0414] ,

[0415] in, Represents the regional functional recovery index, , , Actual supply, actual satisfaction and actual transportation capacity after regional emergency repairs. , , They represent the maximum supply of the supply node after regional emergency repair, the maximum satisfaction of the demand node, and the maximum capacity of the transportation node. Describe the traffic balance constraints of key network facilities, Describes the maximum capacity constraint.

[0416] S5232. Data collection and quantification: The recovery level of the emergency repair function of the facility is obtained in real time through on-site communication.

[0417] When the repair level The weights for different levels are set as follows: 1, 0.75, 0.5, 0.25, 0, respectively, for complete functional recovery, partial functional recovery, basic functional maintenance, limited functional recovery, and total paralysis.

[0418] S5233, real-time regional function recovery index acquisition: through the regional function recovery function, combined with the collected and quantified data, map the regional facility network, and obtain real-time regional function recovery indicators, such as Fig.13 As shown;

[0419] S524, regional emergency repair scheduling decision model construction: the emergency repair path plan, the real-time regional function recovery index and the priority of the key facility emergency repair are combined to construct a multi-objective optimization function, and the resource satisfaction, comprehensive emergency repair of key facilities and non-repeated emergency repair are used as constraints to obtain a regional key facility emergency repair scheduling model;

[0420] The mathematical expression of the regional key facility emergency repair scheduling model is shown in the following formulas:

[0421] ,

[0422] ,

[0423] ,

[0424] ,

[0425] ,

[0426] ,

[0427] ,

[0428] in, It indicates the repair journey time of each feasible repair plan. , , They represent the first facility to be repaired, the first facility to be repaired, and the facility to be repaired later in the feasible plan. Indicates the operating speed of the repair team. , They represent the repair groups in the feasible solutions. Repair facilities The shortest path between the two, the repair team in the feasible plan Repair facilities The shortest path between Represents the weight of the multi-objective optimization function, obtained by expert scoring method, It represents a multi-objective optimization function that satisfies both the maximum regional function recovery index and the shortest repair journey time. Indicates time The regional function recovery index under It represents a multi-objective optimization function that satisfies both the maximum regional function recovery index and the shortest repair distance. is the repair priority of the facility. They represent the constraints of sufficient resources, emergency repair of key facilities and no repeated emergency repair. Indicates key facilities The repair resources used, To repair resources for the region, Indicates the last moment of regional emergency repairs. Display facilities Whether it has been repaired, Display facilities The number of times repairs were performed.

[0429] Example of critical facility repair and dispatch model data:

[0430] ,

[0431] in, Indicates the key facility number, Indicates the repair priority of the facility. Indicates the number of emergency repair teams in the area. Indicates key facilities With number The distance between the repair teams, Number The emergency repair team will provide emergency assistance to key facilities Resource satisfaction, Indicates the number of times the facility has been repaired during the current period. Indicates whether the emergency repair of the facility has been completed during the current period.

[0432] For example: critical facility repair scheduling model data: It indicates that the facility numbered 1 has a repair priority of 0.8. There are two repair teams in the area. The distance between the first repair team and the facility is 100m, and the resource satisfaction is 0.7. The distance between the second repair team and the facility is 1000m, and the resource satisfaction is 0.9. The facility has not yet been repaired.

[0433] S60. Construction of regional weak point and vulnerability assessment model: Construct regional explosion damage circle through Monte Carlo design of random strikes; obtain regional weak points and regional vulnerability assessment through mean shift method to construct regional weak point and vulnerability assessment model;

[0434] like Fig.14 As shown, S60, regional weak point and vulnerability assessment model is constructed, the steps include:

[0435] S61. Acquisition of regional explosion damage circle: Through Monte Carlo design of random strikes with different explosion equivalents, explosion locations, and strike ranges in the region, the functional loss of each key facility is obtained, the network traffic of key facilities is updated, and the regional explosion damage circle is constructed;

[0436] The fixed radius of the strike range is designed to be The designed strike area is a radius of circle.

[0437] S62. Construction of vulnerability assessment model: The regional explosion damage circle is continuously updated through the mean shift method to obtain regional weak points and regional vulnerability assessment to build a vulnerability assessment model;

[0438] 1. Calculate the offset vector: For the current center point, calculate the offset vectors of all points within the bandwidth and the center point. The offset vector usually represents the vector pointing from the center point to each data point.

[0439] 2. Calculate the offset mean: Use the Gaussian kernel function to perform weighted averaging of all offset vectors to obtain a new center point position, i.e., the offset mean.

[0440] 3. Move the center point: Move the current center point to the calculated offset mean position.

[0441] 4. Repeat iteration: Repeat the above steps until the change in the center point position is less than the threshold.

[0442] 5. Clustering: As the iteration proceeds, data points will gradually gather into several clusters according to their distance to the center point and density distribution. The center point of each cluster represents a high-density gathering point in the area, which may correspond to the center of the damage circle.

[0443] 6. Evaluate weak points and vulnerability: By analyzing the density, size, location and other characteristics of each cluster, the weak points in the area are evaluated based on the area with the lowest density or the most significant damage effect, and the vulnerability is obtained based on the weak points.

[0444] S70, construction of regional protection scheme decision model: according to the type and vulnerability of regional weak point facilities, quantify the protection measures of key nodes of the system, obtain the quantitative protection measures of the weak points of the protection area, and establish a multi-objective constraint and multi-standard analysis model according to the cost-effectiveness ratio of each protection scheme, form the optimal protection scheme, and construct a regional protection scheme decision model;

[0445] like Fig.15 As shown, S70, regional protection plan decision model construction, the steps include:

[0446] S71. Quantification of protection measures for weak points in the protection area: Quantify the protection measures for key nodes of the system according to the type of facilities at the weak points in the protection area and the regional protection effectiveness index, and obtain the quantitative protection measures for the weak points in the protection area;

[0447] The step of quantifying the protection measures for weak points in the protection area in S71 includes:

[0448] S711, Confirmation of protection index: Refer to industry specifications or standards (such as ISO, ASTM, etc.) to determine the protection performance benchmark of each key node, and set specific protection index for each node according to the functional attributes of the weak points in the protection area;

[0449] Protection indicators such as explosion resistance (PSI), impact resistance, fire resistance, etc.;

[0450] S712. Quantification of protective measures: Based on the protection index of each weak point in the protection area, the type of facilities at the weak point in the protection area, and the regional protection effectiveness index, and based on the protection characteristic data recorded in the database, that is, according to the damage results of the nodes under different explosions after the introduction of protection technology, evaluate the quantitative protective measures required for the weak points in the protection area to meet the specified protection standards under the protection of the existing protective measures;

[0451] The quantitative protection measures include information such as protection area and protection location;

[0452] Utilize the protection effects that can be achieved by the existing protection technologies in the database to quantify the protection measures required for the target key nodes to achieve the protection indicators.

[0453] S72, protection scheme cost-effectiveness evaluation: based on the types of weak point facilities in the protection area and the regional protection effectiveness index, generate protection schemes for key node functions in the city under multiple random explosion events, and evaluate the protection effectiveness and protection cost of each protection scheme to obtain the cost-effectiveness ratio of each protection scheme;

[0454] The step S72, evaluating the effectiveness and cost of the protection scheme, includes:

[0455] S721. Protection plan generation: Generate multiple protection plans based on the quantitative protection measures after the multi-dimensional explosion based on the regional weak point facilities.

[0456] According to the key nodes in the target area, such as substations, water treatment plants, major transportation hubs, etc.

[0457] The step of generating a protection scheme in step S721 includes:

[0458] S7211. Formulate a protection resource allocation strategy: Formulate a protection resource allocation strategy based on the impact of regional weak point facilities in the system function multi-layer model and the set protection indicator level.

[0459] This includes the development of impact assessment models:

[0460] (71),

[0461] in, Is a node The influence of is its priority, is its dependency in the system functional layer, is its connectivity in the system.

[0462] In addition, it also includes determining the protection requirements of each node based on its protection indicator level:

[0463] (72),

[0464] in, Is a node protection needs, is the protection index level, It is the building standard grade of the node.

[0465] Protection resource allocation strategies such as giving priority to protecting key nodes and allocating more resources to areas with high protection requirements (high risks);

[0466] S7212. Obtain different protection plans: formulate protection plans for each explosion event, key system nodes, and protection resource allocation strategy, combined with quantitative protection measures;

[0467] Protection plan, including protection area, protection location, different combinations of protection measures and implementation sequence, etc.

[0468] S722. Evaluation of the effectiveness of protection schemes: Based on the multiple protection schemes, protection measures are simulated at the key nodes of the urban system function model. The protection features recorded in the explosion damage database are used again to deduce the damage to the nodes under the protection of each protection scheme. The impact on the overall system function of the city is further evaluated through the system function multi-layer model. The final protection effect is unified with the system function retention, loss reduction ratio, and recovery speed data. All feasible schemes that can achieve consistent protection effects are screened out, and the protection effectiveness of each scheme is evaluated according to the degree of impact.

[0469] The less the functions of urban systems are affected, the more effective the protection plan will be.

[0470] S723. Protection cost assessment: Determine the main cost elements in the protection plan, including material cost, construction cost and maintenance and renewal cost, use decision tree neural network to establish the relationship model between cost elements and total cost, and evaluate the cost elements that have the greatest impact on the total cost.

[0471] The protection cost assessment step S723 is specifically as follows:

[0472] S7231. Material cost estimation: Estimate the market price of protective materials, such as high-strength concrete, steel bars, explosion-proof glass, etc.;

[0473] S7232. Construction cost estimation: According to the project plan, refine the timetable and resource allocation of each work. Estimate the construction costs such as civil engineering, equipment installation, and labor costs;

[0474] S7233. Maintenance and renewal cost estimation: Based on the age and environmental conditions of the facility and the life cycle of the facility, estimate the cost of routine maintenance and regular renewal, including regular inspections, repairs, and upgrades.

[0475] S7234. Establishment of the relationship model between cost factors and total cost: Use decision tree neural network to establish the relationship model between cost factors and total cost. By recursively splitting the data set, select the feature that can minimize the cost variance as the split point, retain the feature split point that contributes the most to cost prediction, evaluate the importance of each cost factor in the decision tree, and obtain the cost factor that has the greatest impact on the total cost.

[0476] S73. Construction of regional protection scheme decision model: Based on the cost-effectiveness ratio of each protection scheme, the protection schemes and their corresponding protection costs of the key nodes of the comprehensive system that all achieve the set protection indicators under the protection measures are established, a multi-objective constraint and multi-standard analysis model is established, the optimal protection scheme is formed, and a regional protection scheme decision model is constructed.

[0477] According to the type of target (such as power supply stations, oil storage depots, command centers, transportation hubs, etc.) and protection index requirements, select appropriate protection measures, optimize emergency response mechanisms, etc., so as to improve the explosion resistance and recovery capabilities of the overall system.

[0478] The step of constructing the regional protection plan decision model in S73 includes:

[0479] S731, Multi-objective optimization: define multi-objective functions, maximize protection effectiveness and minimize protection costs, take budget constraints, implementation events, availability of materials and human resources as model constraints according to the implementation costs of different protection schemes, obtain multiple protection strategy combinations through random generation or heuristic methods, sort each solution in the solution set, identify solutions that cannot be dominated by other solutions, form a Pareto frontier, further sort the solutions on the Pareto frontier, identify the non-dominated level of each solution, and obtain the Pareto solution set;

[0480] Taking into account that the optimal protection plan does not mean finding the extreme value, that is, there is a situation where there is no upper limit on the protection budget and extreme protection of key nodes is required, multiple standard criteria are set in the optimization model to achieve the purpose of optimal decision-making on demand.

[0481] S732, multi-criteria decision-making: based on the Pareto solution set, the weight of each criterion is set according to the decision maker's goals or resources;

[0482] For example, protection effectiveness is prioritized, cost control is prioritized, etc. Ensure that the values ​​of different standards are compared on the same scale;

[0483] S733. Construction of regional protection plan decision model: Use multi-criteria decision analysis to dynamically evaluate the weighted ratio scores of each plan under different criteria, and decide on the optimal protection plan based on actual protection needs to build a regional protection plan decision model.

[0484] Through the above steps, the explosion resilience of urban integrated system functions under multiple explosion events can be systematically evaluated, and scientific protection and recovery recommendations can be provided to ensure the safety and reliability of critical infrastructure.

[0485] The core advantage of the present invention lies in its flexibility and adaptability. Through Monte Carlo simulation and multi-objective optimization algorithms, it can adjust and optimize the protection strategy in real time and effectively deal with complex and changeable explosion scenarios. In terms of technical implementation, the present invention first establishes an urban system function model and an explosion damage effect database, then constructs a multi-layer model of system functions, and quantifies the protection measures of key nodes of the system. On this basis, the present invention further evaluates the effectiveness and cost of the protection plan, determines the optimal protection plan through multi-objective optimization and multi-criteria decision analysis, and constructs a regional protection plan decision model.

[0486] The implementation of the present invention can not only improve the scientificity and accuracy of the protection plan, but also ensure the efficient use of resources through precise cost factor analysis, and can provide global and detailed protection plan decisions for regional weak point facilities, significantly improving the ability and efficiency of urban engineering safety and protection.

[0487] S80. Construction of regional strike plan decision model: According to the type of facilities at the weak points in the strike area and the regional strike effectiveness indicators, the strike facilities are obtained, the cost of the strike plan is estimated, and the optimal ammunition use plan is planned according to the ammunition performance, target type and protection level. According to the best cost-effectiveness of the weak point facilities under the set strike indicators, the optimal strike plan is dynamically selected, thereby constructing a regional strike plan decision model.

[0488] like Fig.16 As shown, the step of constructing the regional strike plan decision model in S80 includes:

[0489] S81. Calculation of strike effectiveness and cost: Estimate the cost of the strike plan based on the type of facilities at the weak points in the strike area and the regional strike effectiveness indicators.

[0490] By quantitatively reflecting the impact of strike actions on target system functions, the cost and effectiveness of strike plans can be evaluated, providing reliable basic data support for strike decision-making and budget planning;

[0491] The step S81, calculating the strike effectiveness, includes:

[0492] S811, System Function Damage Assessment: Assess system function damage based on the type of facilities at the weak points in the strike area and the regional strike effectiveness indicators;

[0493] The step of S811, system function damage assessment, includes:

[0494] S8111. Setting of strike indicators for weak point facilities: According to the types of weak point facilities in the strike area, set strike indicators for weak point facilities in the target area;

[0495] According to the identification of weak points in the target area, such as substations, water treatment plants, major transportation hubs, etc., determine the specific strike indicators for each facility according to demand, such as the percentage of power supply capacity reduction, water treatment capacity loss, traffic flow reduction, etc.;

[0496] S8112, Physical layer damage assessment: The functional damage data of the facility is based on the type of facility at the weak point of the strike area and the regional strike effectiveness.

[0497] Use multi-dimensional system status information to further conduct system function damage assessment and identify the functional and geographical damage caused by the strike event.

[0498] S812. Strike plan cost estimation: Based on the system function damage data, estimate the cost of weapons and ammunition, deployment and launch costs, logistics and support costs, and determine the cost of the strike plan.

[0499] Guided by the results of system functional damage assessment, provide accurate estimates of weapons and ammunition procurement costs to help decision makers understand the composition and main driving factors of various costs;

[0500] The step of estimating the cost of the strike plan in step S812 includes:

[0501] S8121. Weapons and ammunition cost estimation: Based on the system function damage data, determine the type and quantity of weapons and ammunition required, including conventional ammunition, guided weapons, drones and other high-tech equipment, and estimate the market price;

[0502] S8122, Deployment and launch cost estimates: Estimate deployment and launch costs, including transportation, launcher setup and operation, and labor costs, based on the type and quantity of weapons and ammunition described;

[0503] Break down the time and resource requirements for each operation step to ensure the accuracy of cost estimates.

[0504] S8123. Logistics and support cost estimation: Based on the deployment and launch costs, estimate the logistics support costs required in the execution of the strike mission, such as fuel, communications, and logistics personnel costs;

[0505] S8124. Determination of the relationship between cost elements and total cost: Use a decision tree neural network model to determine the relationship between cost elements and total cost.

[0506] By recursively splitting the data set, the feature that minimizes the cost variance is selected as the split point, and a decision tree model is established.

[0507] S8125. Identification of the most influencing factors: Based on the importance of each cost element in the decision tree, identify the cost element that has the greatest impact on the total cost.

[0508] S82. Construction of regional strike plan decision model: Plan the optimal ammunition use plan based on ammunition performance, target type and protection level, and dynamically select the optimal strike plan based on the best cost-effectiveness of key infrastructure under the set strike indicators.

[0509] Select appropriate ammunition types based on the target type (such as important bases, communication centers, energy facilities, etc.) and protection level; make matching decisions on target types and node functions to maximize the strategic and tactical benefits of the strike;

[0510] The step S82, constructing a regional strike plan decision model, includes:

[0511] S821. Ammunition use planning: Plan the explosion point and explosion method according to the type and equivalent of ammunition, combined with the target type, distribution and protection level;

[0512] The purpose of planning the ammunition use plan is to optimize the efficiency of ammunition use and ensure that the selected ammunition can effectively achieve the expected effect of the strike mission;

[0513] The steps of S821, ammunition use planning, include:

[0514] S8211. Select the type of ammunition: Select the type of ammunition to be used by referring to the destructive power, penetration ability and effectiveness against specific targets of each type of ammunition recorded in the database;

[0515] S8212. Determine the ammunition equivalent: Based on the assessment results, consider the structural strength, location and protection measures of the target, and the explosive equivalent required for each target, and determine the ammunition equivalent sufficient to achieve the expected destructive effect;

[0516] S8213. Calculate the number of ammunition required: Calculate the number of ammunition required based on the size, shape and distribution of the target;

[0517] For example, against multiple dispersed small targets, multiple small munitions may be needed; against a single large target, a small number of high-yield munitions may be used;

[0518] S8214, Explosion point analysis and explosion mode optimization: Dynamically match the ammunition strike characteristics with the target damage characteristics, determine the ammunition explosion point based on the strike index and the target structure damage response, and determine the best detonation method based on the target information extracted from the geographic information system, including target characteristics and the scene where the target is located;

[0519] By accurately calculating the explosion point and optimizing the detonation method (such as air explosion, contact detonation, delayed detonation) according to target characteristics (such as type, protection measures, dynamics) and scene characteristics (such as surface, underground, clustered, dispersed), the destructive effect can be maximized and collateral damage can be minimized.

[0520] S822. Optimal strike plan selection: Dynamically select the optimal strike plan based on the best cost-effectiveness of the weak point facilities under the set strike indicators.

[0521] Clarify the key targets of the strike and improve the strategic and tactical effectiveness of the strike;

[0522] S8221. Cost-effectiveness optimization: Taking cost-effectiveness as the objective function, maximizing the strike benefit, taking the feasibility and economy of the plan as constraints, using genetic algorithms to generate multiple weak point strike plans, and evaluating the fitness of each weak point strike plan;

[0523] S8222, Dynamic Planning and Path Optimization: Decompose the strike decision-making process into multiple stages, make dynamic planning decisions at each stage based on the current state and possible future states, analyze the possibility and effect of state transitions at each stage, and evaluate the benefits and risks of different decision paths;

[0524] S8223. Construction of regional strike plan decision model: Compare the benefits and risks of each decision path, select the path with the highest total benefit and the lowest risk, form the optimal strike plan, and obtain the regional strike plan decision model.

[0525] The present invention combines genetic algorithms and dynamic programming methods to establish a multi-objective optimization model, realizes adaptive matching of ammunition use and targets, and dynamically adjusts the strike strategy during the decision-making process, so that the strike plan can flexibly adapt to changes in the on-site environment and select the plan that maximizes benefits and minimizes risks. This adaptability greatly improves the flexibility of decision-making and the efficiency of strike operations.

[0526] Through the real-time data feedback mechanism, the decision-making system is allowed to make dynamic adjustments based on actual conditions, ensuring that in the actual field environment, decision makers can quickly adjust strategies to respond to emergencies and environmental changes, thereby minimizing the risks brought by uncertainty and optimizing the strike effect.

[0527] S90, integrated attack and defense intelligent decision-making under the threat of explosion impact: through real-time communication, the regional scope, decision type and on-site decision parameters are obtained to realize decision demand analysis, and the regional emergency repair scheduling decision model, regional protection plan decision model and regional strike plan decision model are reasonably called to make decisions, so as to realize integrated attack and defense intelligent decision-making;

[0528] like Fig.17 As shown, the steps of S90, regional attack and defense integrated intelligent decision-making under the threat of explosion impact, include:

[0529] S91. Decision demand analysis: through real-time communication, the area scope and decision type are obtained to realize decision demand analysis and on-site decision parameters; the decision parameters include regional emergency repair scheduling decision parameters, regional protection plan decision parameters, and regional strike plan decision parameters.

[0530] The regional emergency repair dispatch decision parameters include the weight corresponding to the functional recovery level of the emergency repair facility, the facility recovery time limit, and the material resources and human resources of each emergency repair team;

[0531] The decision parameters of the regional protection plan include the regional information updated after the strike, including the characteristics of the protection area and the characteristics of the protection facilities;

[0532] The decision parameters of the regional strike plan include the regional information updated after the strike, including the strike area characteristics and strike facility characteristics;

[0533] Regional scope: regional name, longitude and latitude, etc.

[0534] Decision type: Decision types are divided into emergency repair dispatch decision, protection plan decision, and strike plan decision, which are obtained according to the decision type corresponding to the corresponding area obtained in real time;

[0535] (92) Decision model invocation: reasonably invoking the regional emergency repair scheduling decision model, the regional protection plan decision model, and the regional strike plan decision model according to the decision requirements;

[0536] (93) Intelligent decision-making for integrated offense and defense: Obtain corresponding decision information through the regional emergency repair scheduling decision model, regional protection plan decision model, and regional strike plan decision model to achieve intelligent decision-making for integrated offense and defense.

Claims

1. An intelligent decision-making method for regional attack and defense under the threat of explosion impact, characterized in that: The steps include: S10. Construction of intelligent decision-making database: Collect multi-dimensional and multi-target explosion damage effect data, integrate protection technology data collected from multiple channels, obtain numerical simulation explosion data, build explosion damage effect database and explosion impact structural mechanical response database, and obtain an integrated attack and defense intelligent decision-making database; S20, constructing a digital-physics-driven intelligent model: constructing a neural network for mechanical response of an explosion impact structure, and using the explosion impact structure mechanical response database to train the neural network for mechanical response of an explosion impact structure, to obtain a digital-physics-driven intelligent model for mechanical response of an explosion impact structure; S30, regional network traffic model construction: collect dispatching, protection, and strike regional information in real time through the GIS information system, build the corresponding associated key facility network, combine the corresponding regional associated network traffic data set, define network traffic model parameters and variables, and build the dispatching, protection, and strike regional network traffic model; S40, regional network traffic acquisition: regional network traffic is acquired by the regional network traffic model, including regional network traffic before the explosion, regional network traffic after the explosion, and real-time regional network traffic for emergency repair and recovery; S50, construction of regional emergency repair scheduling decision model: construct a comprehensive fuzzy evaluation emergency repair priority for each key facility, obtain the emergency repair path plan according to the facility emergency repair priority, combine the real-time regional function recovery index, construct a multi-objective optimization function, and obtain the regional key facility emergency repair scheduling decision model; S60. Construction of regional weak point and vulnerability assessment model: using the Monte Carlo method to design random strikes and construct a regional explosion damage circle; using the mean shift method, combined with the regional explosion damage circle, to obtain regional weak point and regional vulnerability assessment results, and constructing a regional weak point and vulnerability assessment model; S70. Construction of regional protection plan decision model: According to the type and vulnerability of regional weak point facilities, quantify the protection measures of key nodes of the system, obtain the quantitative protection measures of the weak points of the protection area, and establish a multi-objective constraint and multi-standard analysis model according to the cost-effectiveness ratio of each protection plan to form the optimal protection plan and construct a regional protection plan decision model; S80. Construction of strike plan decision model: According to the type of facilities at the weak points in the strike area and the regional strike effectiveness index, the strike facilities are obtained, the cost of the strike plan is estimated, the optimal ammunition use plan is planned according to the ammunition performance, target type and protection level, and the optimal strike plan is dynamically selected according to the best cost-effectiveness ratio of the weak point facilities under the set strike index, and the regional strike plan decision model is constructed; S90, regional attack and defense integrated intelligent decision-making: Based on the decision-making requirements and on-site decision-making parameters obtained through real-time communication, the corresponding decision-making model is called to make decisions and obtain the attack and defense integrated intelligent decision-making results; The steps of constructing the S60, regional weak point and vulnerability assessment model include: S61. Acquisition of regional explosion damage circle: Monte Carlo method is used to design random strikes with different explosion equivalents, explosion locations, and explosion ranges in the region, obtain the functional loss of each key facility, update the network traffic of key facilities, and construct the regional explosion damage circle; S62. Construction of vulnerability assessment model: The mean shift method is used to continuously update the regional explosion damage circle, thereby obtaining regional weak points and regional vulnerability assessment results, and constructing a regional weak point and vulnerability assessment model.

2. The intelligent decision-making method according to claim 1, characterized in that: The step of constructing the intelligent decision database in S10 includes: S11. Damage effect data collection: Collect explosion damage effect data covering various force field types and various damage modes, including experimental data, historical accident data and computer simulation data; S12. Protection technology data collection: Collect protection technology data from test data, historical data and simulation data; S13, numerical simulation explosion data acquisition: by designing numerical simulation explosion input data, performing batch numerical simulation, and acquiring numerical simulation explosion data; S14. Database construction: Construct an attack-defense integrated intelligent decision-making database consisting of an explosion damage effect database and an explosion impact structural mechanics response database.

3. The intelligent decision-making method according to claim 2, characterized in that: The step of constructing a database in step S14 includes: S141. Establishment of explosion damage effect database: pre-processing, classifying, integrating and entering the damage effect data and protection technology data into a database system to establish an explosion damage effect database; S142, establishing a database of explosion impact structural mechanics response: integrating test and historical explosion data, and numerical simulation explosion data, performing data processing, obtaining multiple sets of explosion impact input data and corresponding structural mechanics response data, combining the explosion impact input data and structural mechanics response data according to explosion cases, and obtaining an explosion impact structural mechanics response database; S143. Construction of an attack-defense integrated intelligent decision-making database: combining the explosion damage effect database and the explosion impact structural mechanics response database to obtain an attack-defense integrated intelligent decision-making database.

4. The intelligent decision-making method according to claim 1, characterized in that: The step S20 of constructing a digital-object-driven intelligent model includes: S21, data set division: dividing the explosion impact structural mechanical response database into a training data set and a verification data set; S22. Construction of neural network for mechanical response of explosion structures: Construction of a graph neural network embedded with physical information for predicting mechanical response of explosion structures; S23. Acquisition of a mathematical-physics-driven intelligent model of the mechanical response of a structure to an explosion impact: using the training data set to train the graph neural network in which the physical information is embedded, and using the verification data set to optimize the graph neural network in which the physical information is embedded, to obtain a mathematical-physics-driven intelligent model of the mechanical response of a structure to an explosion impact.

5. The intelligent decision-making method according to claim 1, characterized in that: The step of constructing the regional network model in S30 includes: S31, regional information collection: regional information is collected in real time through the GIS information system, the regional information includes dispatching, protection, strike regional function types, dispatching, protection, strike facility function characteristics, and emergency repair dispatch data; S32, constructing a network of associated key facilities: determining the regional functional system, key facilities of each functional system, and key facilities associated with the regional functional system according to the regional functional type and the functional characteristics of the facilities, and constructing a network of associated key facilities according to the flow attributes between the facilities; S33, regional correlation network flow data set construction: quantify the facility working parameters in the facility functional characteristics to obtain key facility flows and construct a regional correlation network flow data set; S34. Regional network model construction: define network traffic model parameters and variables, map the regional associated network traffic data set to the associated key facility network, and obtain a regional network traffic model, including a scheduling regional network traffic model, a protection regional network traffic model, and a strike regional network traffic model.

6. The intelligent decision-making method according to claim 1, characterized in that: The step of obtaining regional network traffic in S40 includes: S41, obtaining the regional network traffic before the explosion: obtaining the regional network traffic before the explosion by using the regional network traffic model; S42, obtaining regional network traffic after the explosion: obtaining the function loss of key facilities by defining the facility function loss function, and obtaining the regional network traffic after the explosion by combining the regional network traffic model; S43, obtaining the real-time regional network traffic for emergency repair and recovery: obtaining the actual functional loss of the facilities after the explosion, obtaining the real-time traffic of each key facility for emergency repair and recovery, and taking the sum of the real-time traffic of the underlying demand node facilities as the real-time regional network traffic for emergency repair and recovery; S44, regional network traffic acquisition: the regional network traffic is composed of the regional network traffic before the explosion, the regional network traffic after the explosion, and the real-time regional network traffic for emergency repair and recovery.

7. The intelligent decision-making method according to claim 1, characterized in that: The step S50, constructing the regional emergency repair dispatch decision model includes: S51. Obtaining emergency repair priority: Quantify the functional loss of key facilities, the flow attributes of key facilities, and the emergency repair time limit of key facilities. Through data processing, construct a classification index of emergency repair priority of key facilities, and combine the two-stage fuzzy entropy weight method to obtain the comprehensive fuzzy evaluation emergency repair priority of each key facility; S52. Construction of regional key facility emergency repair scheduling model: The emergency repair path plan is obtained according to the facility emergency repair priority, and a multi-objective optimization function is constructed in combination with real-time regional function recovery indicators. The resource satisfaction, comprehensive facility emergency repair and non-repetitive emergency repair are used as constraints to obtain the regional key facility emergency repair scheduling decision model.

8. The intelligent decision-making method according to claim 7, characterized in that: The step of constructing the regional key facility emergency repair scheduling model in S52 includes: S521, resource satisfaction acquisition: obtain resource satisfaction by collecting and quantifying the material resources and human resources of each repair team in real time; S522, constructing a repair path evaluation function: according to the facility spatial location and facility attributes in the facility geometric parameters, the shortest path between the facilities is obtained through GIS information technology, and the repair path evaluation function is constructed in combination with the repair priority; S523, real-time regional function recovery index acquisition: define the regional function recovery function, collect and quantify the key facility repair function recovery level in real time, map the regional facility network, and obtain the real-time regional function recovery index; S524. Construction of regional emergency repair scheduling decision model: Obtain the emergency repair path plan according to the facility emergency repair priority, combine the real-time regional function recovery index, construct a multi-objective optimization function, and use the resource satisfaction, comprehensive emergency repair of key facilities and non-repetitive emergency repair as constraints to obtain the regional key facility emergency repair scheduling decision model.

9. The intelligent decision-making method according to claim 1, characterized in that: The step of constructing the regional protection plan decision model in S70 includes: S71. Quantification of protection measures for weak points in the protection area: Quantify the protection measures for key nodes of the system according to the type of facilities at the weak points in the protection area and the regional protection effectiveness index, and obtain the quantitative protection measures for the weak points in the protection area; S72, protection scheme cost-effectiveness evaluation: based on the types of facilities at the weak points of the protection area and the regional protection effectiveness index, generate protection schemes for key node functions in the city under multiple random attack events, and evaluate the protection effectiveness and protection cost of each protection scheme to obtain the cost-effectiveness ratio of each protection scheme; S73. Construction of regional protection scheme decision model: Based on the cost-effectiveness ratio of each protection scheme, the protection schemes and their corresponding protection costs of the key nodes of the comprehensive system that all achieve the set protection indicators under the protection measures are established, a multi-objective constraint and multi-standard analysis model is established, the optimal protection scheme is formed, and a regional protection scheme decision model is constructed.

10. The intelligent decision-making method according to claim 1, characterized in that: The S90, regional attack and defense integrated intelligent decision-making steps include: S91, decision demand acquisition: acquire decision demand and on-site decision parameters through real-time communication, wherein the decision demand includes regional scope and decision type; the decision type includes regional emergency repair scheduling decision, regional protection plan decision or regional strike plan decision; the decision parameters include regional emergency repair scheduling decision parameters, regional protection plan decision parameters and regional strike plan decision parameters; S92, decision model calling: calling the decision model corresponding to the decision requirement; S93, attack-defense integrated intelligent decision-making: The decision information of the called decision model within the said area is output as the result of the attack-defense integrated intelligent decision-making.

Images