Digital twin-driven dynamic topology network intelligent fire-fighting collaboration method and system

Abstract

The invention is suitable for the field of fire extinguishing and automatic control, and provides a digital twin-driven dynamic topology network intelligent fire-fighting cooperation method and system, and the method comprises the steps: building a digital twin model of a fire-fighting monitoring site, carrying out the real-time monitoring of data through various devices, and enabling the model to be dynamically updated; establishing a dynamic topology network, deploying communication nodes and optimizing topology; and then analyzing a monitoring scene, defining task priorities in combination with resource and environment information, allocating tasks and scheduling communication nodes, and finally visually outputting a result. According to the system, task allocation, resource scheduling and communication optimization are incorporated into a unified mathematical framework, the defect that a traditional fire-fighting decision is split is overcome, and the rescue efficiency is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of fire fighting and automatic control, and particularly relates to a digital twin driven dynamic topology network intelligent fire fighting collaboration method and system. BACKGROUND

[0002] With the rapid development of urban construction, complex buildings such as high-rise buildings and underground spaces have increased significantly, and the risk and harm degree of fire have increased significantly. As a new technology, digital twin technology can realize real-time monitoring, simulation analysis and optimization control of physical systems by constructing a virtual model of the physical system. Dynamic topology network technology can automatically adjust the network structure according to the dynamic changes of the nodes to ensure the stability and reliability of the communication. Both of them provide new technical support for fire fighting collaboration.

[0003] Currently, in fire rescue, traditional decision-making methods are mainly relied on, including task allocation, resource scheduling and communication support. Among them, task allocation mainly depends on artificial experience, and the carrying capacity of the communication link is often ignored in the resource scheduling process. The communication deployment is not coordinated with the task and resource, and each link works independently, which leads to the disadvantages of fragmented decision-making, delayed rescue orders, resource mismatch, waste of manpower and material resources, increased safety risks of rescue personnel, and serious impact on the efficiency of fire fighting and rescue. SUMMARY

[0004] The purpose of the present application is to provide a digital twin driven dynamic topology network intelligent fire fighting collaboration method, which aims to solve the technical problems existing in the prior art identified in the background.

[0005] The present application is implemented in the following way, A digital twin driven dynamic topology network intelligent fire fighting collaboration method, the main steps of which are: Step 1: Construct a digital twin model of the fire monitoring site. When there is no fire, real-time monitoring of various data in the scene is carried out through temperature sensors, anemometers, gas leak imagers, laser methane remote sensors and other equipment. When a fire occurs, the development trend of the fire can also be reflected in real time. The model can be continuously updated and corrected according to the data collected on site.

[0006] The geometric structure of the fire building or scene to be monitored is obtained by unmanned aerial vehicle oblique photography or three-dimensional laser scanning, and environmental data such as door and window positions, wall materials, evacuation channels, water source positions, building spacing, etc. are added. For oil and gas stations, which are flammable and explosive scenes with gas leakage, the material and valve position information of the combustible gas pipeline should also be added, as well as the information of historical leakage points and high-risk areas (such as valve interfaces and flange connections). A three-dimensional space framework is constructed.

[0007] Deploy a comprehensive monitoring and early warning network, monitor the area layout of temperature sensors, smoke sensors, combustible / toxic gas component sensors, wind speed and direction instruments; in particular, deploy laser methane remote sensing instruments in flammable and explosive areas; deploy gas leak imaging instruments in open spaces or high places to achieve visualized gas leak areas; deploy gas leak acoustic imaging instruments in closed or semi-closed areas; all monitoring and early warning equipment and sensors transmit data back to the space architecture through low-power networks.

[0008] Based on the three-dimensional space architecture, the scene space is decomposed into grid units, each grid unit is associated with material properties and ventilation parameters to form a calculable space topology. A "hidden danger-fire" correlation prediction model is established. If a fire has occurred, an intelligent fire spread prediction model is established based on aerodynamics and machine learning algorithms.

[0009] The specific logic of the "hidden danger-fire" correlation prediction is as follows: When the concentration of flammable gas reaches 10%-25% of the lower explosive limit, it is considered as a first-level warning, triggering local area alert and dispatching unmanned aerial vehicles for close review; When the concentration of flammable gas reaches 25%-50% of the lower explosive limit, it is considered as a second-level warning, and the regional valves are closed, the exhaust system is started, and personnel are notified to evacuate; When the temperature abnormally rises (such as exceeding 80℃ and continuously rising) or the imaging instrument detects a spark, combined with gas concentration data, it is determined as "fire precursor". It is considered as a third-level warning, and the fire extinguishing preparation work is directly started.

[0010] The specific logic of the intelligent fire spread prediction model is as follows: Solve the temperature field, velocity field and smoke concentration field at the initial time according to the continuity equation and momentum equation of Newton's second law in fluid dynamics to generate a physical characteristic matrix. Collect historical fire data and extract spatial and temporal features such as temperature gradient and spread speed. The physical characteristic matrix and historical data are used as input to train a machine learning fusion model.

[0011] Input real-time monitoring data into the model, update the grid state through physical equations, and then correct the prediction results by the machine learning model to output the fire spread path and dangerous area in the future period of time.

[0012] Step 2: Establish a dynamic topology network and deploy fire communication nodes. Use adaptive routing algorithm to optimize the topology network in real time.

[0013] Establish a dynamic topology network and optimize the topology network through adaptive routing algorithm; Deploy high-power base stations, backup power supplies, and anti-interference wires in areas such as building entrances and fire wells that are structurally stable and not easily damaged. They act as signal relays for the network hub.

[0014] When the device discovers a leak point or temperature anomaly, the dynamic topology network automatically marks the location as a "potential fire source" and dispatches the nearest fire-fighting robot carrying a portable detector to confirm; if it is confirmed that a fire has occurred, a drone is deployed to dynamically hover over the high-risk block according to the intelligent fire spread prediction model, providing communication relay, blind area danger alarm and other functions. A fire-fighting robot equipped with a multi-frequency communication module is deployed to access the spatial architecture model to obtain a three-dimensional map of the fire scene and go to the severe fire area to perform fire extinguishing operations such as throwing fire extinguishing bombs; Step 3: The data twin model analyzes the monitoring scene in real time, combines the deployment of fire-fighting resources and the current environmental information, defines the task priority, allocates warning processing tasks or fire extinguishing tasks, and dynamically solves the scheduling of fire-fighting communication nodes.

[0015] For different monitoring scenes, the task priority is slightly different. The first-level tasks mainly include fire source suppression, life rescue, key facility protection, etc. The second-level tasks mainly include obstacle removal, environment monitoring, relay deployment, containment of combustible gas leakage, and temperature reduction. The third-level tasks mainly include high-risk area isolation, fire extinguishing agent supply, etc.

[0016] The beneficial effects of the present application are: The present application solves the problem of fragmented traditional fire-fighting decision-making by constructing a fire scene digital twin model that can be updated and corrected in real time, optimizing communication with a dynamic topology network and adaptive routing algorithm, and integrating task allocation, resource scheduling and communication optimization into a unified mathematical framework. The "hidden danger-fire" correlation prediction of devices such as gas leak imagers realizes the upgrade from passive fire extinguishing to active prevention. The intelligent fire spread prediction model combines fluid dynamics and machine learning algorithms to accurately output fire spread paths and dangerous areas. The dynamic priority algorithm can adjust task weights according to the fire stage. The task allocation model and communication node scheduling model achieve efficient resource matching and reliable communication through constraint conditions. The system can dynamically adjust strategies and visualize the output results, effectively improving the scientificity, timeliness and resource utilization efficiency of fire-fighting and rescue, and reducing the risk of personnel casualties and property losses. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 The present application is a whole method flowchart; Figure 2 The present application is a system structure schematic diagram. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0019] As Figure 1 shown, the present application provides a digital twin driven dynamic topology network intelligent fire fighting collaboration method, the specific steps are as follows: Step 1: Construct a digital twin model of the fire monitoring site: The model, when there is no fire, monitors various data in the scene in real time with the help of temperature sensors, anemometers, gas leak imagers, laser methane remote sensors, etc.; When a fire occurs, it can reflect the development of the fire in real time and can be updated and corrected continuously according to the collection of on-site data. The specific construction process is as follows: 1. Construct a high-precision three-dimensional space architecture: Obtain the geometric structure of the fire building or scene to be monitored through unmanned aerial photography or three-dimensional laser scanning, add environmental data such as door and window positions, wall materials, evacuation channels, water source positions, building spacing, etc.; For oil and gas stations and other flammable and explosive scenes with gas leaks, you also need to add information such as the material of the flammable gas pipeline, valve position, and historical leak points and high-risk area information (such as valve interfaces and flange connections).

[0020] 2. Deploy a comprehensive monitoring and early warning network: Place temperature sensors, smoke sensors, flammable / toxic gas component sensors, and anemometers in the monitoring area; Deploy laser methane remote sensors in flammable and explosive areas, gas leak imagers in open spaces or high places to visualize gas leak areas, and gas leak acoustic imagers in closed or semi-closed areas; All monitoring and early warning devices and sensors transmit data back to the space architecture through a low-power network.

[0021] 3. Construct a prediction model: Based on the above three-dimensional space architecture, the scene space is divided into grid cells, each grid cell is associated with material properties and ventilation parameters, forming a calculable spatial topology; Construct a "hidden danger-fire" correlation prediction model; If a fire has occurred, combine aerodynamics and machine learning algorithms to establish an intelligent fire spread prediction model.

[0022] The logic of the "hidden danger-fire" correlation prediction is as follows: When the concentration of flammable gas reaches 10%-25% of the lower explosive limit, a first-level warning is triggered, local area alert is started, and a drone is dispatched for closer review; When the concentration of flammable gas reaches 25%-50% of the lower explosive limit, a second-level warning is triggered, regional valves are closed, the exhaust system is started, and personnel are notified to evacuate; When the temperature abnormally rises (such as exceeding 80℃ and continuously rising) or the imager captures sparks, combined with gas concentration data, it is determined to be a "fire precursor", triggering a third-level warning, and directly starting fire extinguishing preparations.

[0023] The logic of the intelligent fire spread prediction model is as follows: According to the continuity equation and the momentum equation of Newton's second law in fluid dynamics, the temperature field, velocity field and smoke concentration field at the initial time are solved to generate the physical characteristic matrix; historical fire data are collected to extract the temperature gradient, spread speed and other space-time characteristics; the physical characteristic matrix and historical data are taken as inputs to train the machine learning fusion model.

[0024] The transfer equation set of the machine learning fusion model is as follows: ; ; ; ; ; ; In the above equation set, is the input feature at time t, specifically the real-time temperature, smoke concentration, wind speed and other data of the sensor; is the hidden state at time t, specifically the current fire trend analysis result; is the cell state at time t, specifically the stored fire development information such as temperature change law and spread mode; is the activation value of the forget gate, input gate and output gate, wherein the forget gate is specifically the "trust degree" of the historical fire state; is the weight matrix; is the bias term; is the activation function. The real-time monitoring data is input into the model, the grid state is updated through the physical equation, and then the prediction result is corrected by the machine learning model to output the fire spread path and dangerous area in the future period of time.

[0025] Step 2: Establish a dynamic topology network and deploy a fire communication node: An adaptive routing algorithm is used to optimize the topology network in real time, and the specific process is as follows:

[0026] 1. Optimize the topology network: establish a dynamic topology network and optimize it through an adaptive routing algorithm, and the path selection cost function is as follows: ; ; In the above formula, is the hop count; is the available bandwidth, with the unit of (bit per second); ​​​For transmission delay, unit (millisecond); For weight coefficient.

[0027] 2. Deploy network hubs: In areas where structures are stable and not easily destroyed, such as building entrances, fire wells, etc., deploy high-power base stations, backup power supplies, and anti-interference wires to serve as network hubs for signal relay.

[0028] 3. Schedule fire-fighting equipment: When the equipment detects a leak or temperature anomaly, the dynamic topology network automatically marks the location as a "potential fire source" and dispatches the nearest fire-fighting robot carrying a portable detector to confirm. If a fire is confirmed, deploy a drone to dynamically hover over high-risk blocks based on intelligent fire spread prediction models, providing communication relay, blind area danger alert, and other functions. At the same time, deploy fire-fighting robots equipped with multi-frequency communication modules to access the spatial architecture model to obtain a three-dimensional map of the fire scene and execute fire extinguishing operations such as throwing fire extinguishing bombs in the severe fire area.

[0029] Step 3: Real-time analysis, task definition, and dynamic solution: The data twin model performs real-time analysis on the monitored scene, combines fire-fighting resource deployment and current environmental information, defines task priority, allocates warning processing tasks or fire extinguishing tasks, schedules fire-fighting communication nodes for dynamic solution, and visualizes the results. The specific process is as follows: 1. Define task priority: Task priority differs for different monitoring scenarios. First-level tasks include fire source suppression, life rescue, key facility protection, etc.; second-level tasks include obstacle removal, environment monitoring, relay deployment, containment of combustible gas leaks, and temperature reduction for explosion prevention; third-level tasks include high-risk area isolation, fire extinguishing agent supply, etc. Dynamic priority definition algorithm is adopted: ; In the above formula, is the dynamic priority, is the task urgency (determined by fire spread speed and trapped personnel risk), is the time sensitivity (trapped personnel tolerance time), is the matching degree of fire-fighting resources. Adjusted dynamically by the digital twin model, such as increasing the proportion of fire extinguishing in the initial fire, and increasing the proportion of life rescue in the medium-term fire.

[0030] 2. Task allocation model: ; In the above formula, is the task execution cost, where Total number of tasks; Total number of fire-fighting resources (deployed drones and fire-fighting robots, etc.); Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks; Total number of tasks;

[0031] The constraints of this model include: To avoid misuse of fire-fighting resources, only one resource is allowed to execute each task, i.e. The load of fire-fighting resources when executing tasks should not exceed the maximum carrying capacity of the type of fire-fighting resource, i.e. In the above formula, represents the load of the task on the fire-fighting resource, such as the number of fire extinguishing bullets that a fire-fighting robot can carry.

[0032] If it is the early warning stage, the carried items of the fire-fighting robot are changed to handheld detection equipment.

[0033] Special tasks specify special fire-fighting resources for execution, such as drones cannot carry heavy demolition tools, i.e. In the above formula, if is executable, otherwise

[0034] This model pursues the balance between cost and value, and considers the execution cost and priority value of tasks to avoid ignoring high-priority tasks in pursuit of low cost. In addition, through load restriction and resource compatibility constraints, the feasibility of the task allocation scheme is ensured.

[0035] ​​​​​​​​3. Scheduling communication nodes: The communication node scheduling model is: ; In the above formula, is the task communication quality, where represents the task required communication bandwidth executed by the resource ; represents the communication link reliability of the task-resource pair , and the calculation method is the product of the reliability of all nodes on the path, and the mathematical formula is: , where in the mathematical formula is the damage / failure probability of the communication node predicted by digital twinning during task execution, for example, the failure probability of a node in a high-temperature area with a serious fire is and will increase with the increase of high-temperature time; is the node activation cost, where is the total number of communication nodes; is the communication decision variable, 1 represents activating the communication node , and 0 represents that the communication node is not activated; is the activation cost of node ; is the cost sensitivity coefficient, which is dynamically adjusted by digital twinning according to the remaining power of the fire-fighting resource, and usually increases when the power is low.

[0036] The constraints of this model include: The communication reliability requirement of the primary important task (such as life rescue) is greater than or equal to 95%, i.e.: ; The number of connections of a single node should not exceed the maximum support number , where represents whether the task-resource pair communicates through node ; is the maximum number of connections of node .

[0037] The task-resource pair must ensure that there is at least one valid communication path, i.e.: ; A high reliability threshold is set for the primary task (such as life rescue) to ensure stable transmission of critical task instructions and avoid rescue failure due to communication failure. At the same time, the cost sensitivity coefficient is dynamically adjusted according to the power of the fire-fighting resource, and the activation of high-cost nodes is reduced when the power is low, achieving a balance between communication demand and resource consumption.

[0038] 4. Dynamic solution: Obtain digital twin data, including: task location , task type , task priority , task risk level , fire-fighting resource location , remaining power , fire-fighting resource load , communication node location , communication node damage / failure probability .

[0039] Task allocation and communication scheduling are performed in a double-layer nested manner: the upper layer solves task allocation using integer linear programming for a given communication information node ; the lower layer optimizes communication node activation and path according to the task allocation result of the upper layer .

[0040] System dynamic adjustment: in the hidden danger stage and the initial stage of fire, increase the cost coefficient to minimize the cost; in the serious stage of fire, allow more communication nodes to be used to ensure the reliability of fire extinguishing task execution; when the digital twin predicts that the probability of failure of a node , switch to the standby node; when the power of the fire-fighting resource , consider returning and reassigning tasks.

[0041] As shown in Figure 2 , the application further provides a digital twin driven dynamic topology network intelligent fire fighting collaboration system, which is applied to any one of the above-mentioned digital twin driven dynamic topology network intelligent fire fighting collaboration methods, and the system comprises: A digital twin modeling module is used to construct a virtual mapping of a real scene and realize digital expression of a physical space.

[0042] A multi-source data fusion module is used to obtain multi-dimensional monitored scene data and update a digital twin model in real time.

[0043] A dynamic topology optimization module is used for communication allocation and deployment of fire-fighting resources, real-time adjustment of fire-fighting network topology based on field changes, and optimization of resource collaboration path.

[0044] A real-time communication control module is used to ensure instruction transmission and state feedback of various communication nodes and terminals in the topology network.

[0045] An intelligent decision reasoning module is used for hidden danger detection and early warning, fire spread prediction and reasoning, and task allocation result visualization generation.

[0046] Any technical features in the above-described embodiments can be combined in any manner, and for the sake of brevity, not all possible combinations are described, but it is understood that the scope of the present specification includes all possible combinations.

[0047] The above-described embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for those skilled in the art, some modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

[0048] The above-described embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for those skilled in the art, some modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims. The above-described embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for those skilled in the art, some modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

Claims

1. A digital twin-driven dynamic topology network intelligent firefighting collaboration method, characterized by: The method comprises: Build a digital twin model of the fire monitoring scene, and monitor scene data in real time through temperature sensors, anemometers, gas leak imagers, and laser methane telemeters. When a fire occurs, analyze the fire development trend in real time, and update and modify the digital twin model based on the monitoring scene data collected on site. Establish a dynamic topology network, deploy fire communication nodes, and use adaptive routing algorithms to optimize the topology network in real time; The digital twin model is used to analyze monitoring scene data in real time. Combined with the currently deployed fire-fighting resources and current environmental information, task priorities are defined, tasks are assigned, early warning processing tasks and fire-fighting tasks are allocated, and fire-fighting communication nodes are dynamically solved and dispatched.

2. The method according to claim 1, characterized in that The construction of a digital twin model of the fire monitoring site specifically includes: Obtain the geometric structure of the fire monitoring scene, add environmental data, build a three-dimensional spatial architecture, and deploy a monitoring and early warning network; Based on the three-dimensional spatial architecture, the fire monitoring scene space is decomposed into grid units. Each grid unit is associated with material properties and ventilation parameters to build a hidden danger-fire correlation prediction model. If a fire already exists, an intelligent fire spread prediction model is established by combining aerodynamics and machine learning algorithms.

3. The method according to claim 2, characterized in that The implementation logic of the hidden danger-fire association prediction model is: When the concentration of combustible gas reaches 10%-25% of the lower explosion limit, a level 1 warning is triggered and a drone is dispatched for review; When the concentration of combustible gas reaches 25%-50%, a secondary warning is triggered, the valve is closed, exhaust is started, and personnel are evacuated; When the temperature continues to rise above 80°C or sparks are detected, a level 3 warning is triggered and fire extinguishing preparations are initiated.

4. The method according to claim 3, characterized in that The intelligent fire spread prediction model implementation logic is as follows: The temperature field, velocity field and smoke concentration field at the initial moment of fire occurrence are solved according to the continuity equation in fluid dynamics and the momentum equation of Newton's second law to generate a physical characteristic matrix; The intelligent fire spread prediction model is trained with historical fire data, and real-time monitoring scene data is input after training. The grid state is updated and the prediction results are corrected through physical equations to output the fire spread path and danger zone in the future period.

5. The method according to claim 4, characterized in that The establishment of a dynamic topology network, deployment of fire communication nodes, and real-time optimization of the topology network using an adaptive routing algorithm specifically include: Establish a dynamic topology network, and the path selection cost function is: ； Where, is the number of hops; is the available bandwidth, in units of ; is the transmission delay, in units of ; is the weight coefficient; When a regional leak or temperature anomaly is detected, the potential fire source is marked and a firefighting robot is dispatched for confirmation. After the fire is confirmed, a drone is deployed to dynamically hover over the fire area, access the three-dimensional spatial architecture to obtain a three-dimensional map of the fire area, and dispatch the firefighting robot to perform fire-fighting operations.

6. The method according to claim 5, characterized in that The implementation logic of the execution task allocation is: Establish a task allocation model: ； Where, is the task execution cost, where Indicates the total number of tasks; is the total number of firefighting resources; Indicates Firefighting resources to execute The comprehensive cost of the task, including consumption costs , that is, distance Energy consumption coefficient The product of risk cost , that is, the mission risk level and the vulnerability of firefighting resources Product of time cost , that is, the estimated execution time of the task and task urgency The product of For the The urgency of the task; is the task decision variable, ; The overall goal is to maximize the task value, where For the task priority; The task completion status is defined as 1 for completion and 0 for incomplete. is the value magnification factor; The task allocation model has the following constraints: Only one firefighting resource is assigned to each mission; The load of firefighting resources when performing tasks shall not exceed the maximum carrying capacity of the firefighting resources of the corresponding model; When special missions are identified, specific special firefighting resources are assigned to perform them.

7. The method according to claim 6, characterized in that The implementation logic of the dispatching fire communication node is: Establish a communication node scheduling model: ； Where, is the task communication quality, where Indicates a task By resource The communication bandwidth required for execution; Represents a task-resource pair The reliability of the communication link, , which is the product of the reliabilities of all nodes on the path, Communication nodes predicted for digital twins Probability of damage / failure during mission execution; is the node activation cost, where is the total number of communication nodes; is the communication decision variable, , 1 means activating the communication node ,0 indicates that the communication node is not activated; For nodes activation costs; is the cost sensitivity coefficient, which is dynamically adjusted by the digital twin based on the remaining power of firefighting resources; The communication node scheduling model has the following constraints: For Level 1 important missions, the communication reliability requirement is greater than or equal to 95%; There must be at least one valid communication path between a task and a resource pair.

8. The method according to claim 7, characterized in that The specific steps of the dynamic solution include: Acquire digital twin data; Perform double-layer nested solution for task allocation and communication node scheduling: the upper layer uses integer linear programming to solve task allocation , the lower layer uses heuristic search to optimize the activation of communication nodes With path ; Dynamically adjust the strategy according to the fire stage: determine the current stage of the fire event, including the initial stage and the severe stage. For the initial stage, increase the cost coefficient, and for the severe stage, increase the communication nodes. When the digital twin predicts the failure probability of any node Switch to the backup node when the firefighting resource power Return and reallocate tasks.

Images