CN117910673B - Digital twin firefighting control room graphic display system and method - Google Patents
Digital twin firefighting control room graphic display system and method Download PDFInfo
- Publication number
- CN117910673B CN117910673B CN202410309990.0A CN202410309990A CN117910673B CN 117910673 B CN117910673 B CN 117910673B CN 202410309990 A CN202410309990 A CN 202410309990A CN 117910673 B CN117910673 B CN 117910673B
- Authority
- CN
- China
- Prior art keywords
- fire
- data
- module
- dynamic
- scene
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 106
- 238000011161 development Methods 0.000 claims abstract description 79
- 238000013468 resource allocation Methods 0.000 claims abstract description 77
- 238000013507 mapping Methods 0.000 claims abstract description 76
- 238000005457 optimization Methods 0.000 claims abstract description 64
- 238000005516 engineering process Methods 0.000 claims abstract description 52
- 230000000007 visual effect Effects 0.000 claims abstract description 45
- 238000007637 random forest analysis Methods 0.000 claims abstract description 18
- 238000013144 data compression Methods 0.000 claims abstract description 14
- 238000012098 association analyses Methods 0.000 claims abstract description 10
- 230000007613 environmental effect Effects 0.000 claims description 79
- 238000004422 calculation algorithm Methods 0.000 claims description 63
- 238000004458 analytical method Methods 0.000 claims description 62
- 238000012502 risk assessment Methods 0.000 claims description 40
- 238000013439 planning Methods 0.000 claims description 38
- 238000010586 diagram Methods 0.000 claims description 37
- 238000009826 distribution Methods 0.000 claims description 32
- 238000012545 processing Methods 0.000 claims description 30
- 238000013461 design Methods 0.000 claims description 28
- 230000008569 process Effects 0.000 claims description 25
- 238000011156 evaluation Methods 0.000 claims description 24
- 230000002452 interceptive effect Effects 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 22
- 238000007405 data analysis Methods 0.000 claims description 20
- 230000008859 change Effects 0.000 claims description 19
- 238000013079 data visualisation Methods 0.000 claims description 17
- 238000003909 pattern recognition Methods 0.000 claims description 17
- 230000006399 behavior Effects 0.000 claims description 16
- 230000003993 interaction Effects 0.000 claims description 16
- 238000010276 construction Methods 0.000 claims description 13
- 230000007246 mechanism Effects 0.000 claims description 13
- 230000009467 reduction Effects 0.000 claims description 13
- 230000001427 coherent effect Effects 0.000 claims description 12
- 238000003066 decision tree Methods 0.000 claims description 12
- 238000006243 chemical reaction Methods 0.000 claims description 11
- 238000007619 statistical method Methods 0.000 claims description 11
- 230000010354 integration Effects 0.000 claims description 10
- 238000012706 support-vector machine Methods 0.000 claims description 10
- 230000000694 effects Effects 0.000 claims description 9
- 238000010801 machine learning Methods 0.000 claims description 9
- 238000007635 classification algorithm Methods 0.000 claims description 8
- 239000000284 extract Substances 0.000 claims description 8
- 238000000556 factor analysis Methods 0.000 claims description 8
- 238000007906 compression Methods 0.000 claims description 7
- 230000006835 compression Effects 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 7
- 238000012847 principal component analysis method Methods 0.000 claims description 6
- 238000010219 correlation analysis Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 239000013598 vector Substances 0.000 claims description 5
- 238000003672 processing method Methods 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 238000005728 strengthening Methods 0.000 claims description 3
- 230000001960 triggered effect Effects 0.000 claims description 3
- 230000004888 barrier function Effects 0.000 claims description 2
- 230000004044 response Effects 0.000 abstract description 18
- 238000012800 visualization Methods 0.000 abstract description 5
- 206010063385 Intellectualisation Diseases 0.000 abstract description 3
- 238000002347 injection Methods 0.000 abstract description 3
- 239000007924 injection Substances 0.000 abstract description 3
- 238000000513 principal component analysis Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 6
- 238000007726 management method Methods 0.000 description 6
- 238000004088 simulation Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 230000001186 cumulative effect Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 230000002265 prevention Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 239000000243 solution Substances 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000008713 feedback mechanism Effects 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 238000011867 re-evaluation Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Human Resources & Organizations (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Software Systems (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- General Health & Medical Sciences (AREA)
- Geometry (AREA)
- Computer Graphics (AREA)
- Primary Health Care (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of digital twinning and discloses a digital twinning fire control room graphic display system and a digital twinning fire control room graphic display method, wherein the system comprises a data compression mapping module, a flame trend prediction module, a scene recognition module, an association analysis module, a dynamic model optimization module, a resource allocation decision module and a visual presentation module; according to the invention, the response capability of the fire control room to complex conditions is remarkably improved through the methods of data compression mapping, flame trend prediction, scene recognition, association analysis, dynamic model optimization, resource allocation decision and visual presentation, the rescue resource allocation is optimized through accurate fire development prediction and risk recognition, the high efficiency and the intellectualization of fire rescue work are realized, the accuracy of fire prediction is enhanced by utilizing a random forest and graph injection force network, the rationality of resource allocation is improved, visual and real-time data support is provided for command decision by utilizing a three-dimensional visualization technology, and the rescue efficiency and the scientificity of command decision are remarkably improved.
Description
Technical Field
The invention relates to the technical field of digital twinning, in particular to a digital twinning firefighting control room graphic display system and method.
Background
The digital twin technical field is an advanced computer simulation technology, and by creating a virtual copy of a physical entity, the real-time simulation and monitoring of behaviors, states and processes in the real world are realized; the digital twin technology utilizes data collected by the sensor to update twin objects, so that the virtual model and a physical counterpart thereof are ensured to keep synchronous; the method is widely applied to manufacturing industry, building management, urban planning and emergency response systems, and efficiency, safety and reliability are improved through prediction simulation, performance optimization and fault prevention.
The digital twin fire control room graphic display system is innovation of the application of digital twin technology in the field of fire safety, and aims to realize efficient and accurate fire control monitoring and management by creating virtual copies of buildings or facilities; the response speed and the response efficiency of the fire safety management are improved, the fire risk is reduced, and the decision making process is supported in an emergency; through real-time simulation and visualization, the fire development trend can be predicted, firefighters are guided to take the most effective countermeasures, and personnel safety and property safety are protected.
The traditional fire control room is often insufficient in the aspects of processing and analyzing complex fire data, the fire behavior is difficult to accurately predict and effectively identify potential risks, the accuracy and the efficiency of resource allocation and rescue path planning are also often limited by manual operation and experience judgment, so that the optimal configuration of rescue resources cannot be realized, and the rescue response speed and decision quality are also difficult to ensure; the traditional method has obvious defects in the aspects of information extraction, risk assessment and decision support, so that rescue resources are unreasonably distributed, rescue response time is delayed, and loss caused by fire is increased.
Disclosure of Invention
The invention aims to solve the technical problems of providing a digital twin fire control room graphic display system and a digital twin fire control room graphic display method, which can remarkably improve the response capability of a fire control room to complex conditions, optimize rescue resource allocation through accurate fire development prediction and risk identification, realize high efficiency and intellectualization of fire rescue work, enhance the accuracy of fire prediction by utilizing a random forest and graph injection force network, improve the rationality of resource allocation, provide visual and real-time data support for command decisions by utilizing a three-dimensional visualization technology, and remarkably improve rescue efficiency and the scientificity of the command decisions.
In order to solve the technical problems, the invention provides the following technical scheme:
the digital twin fire control room graphic display system comprises a data compression mapping module, a flame trend prediction module, a scene recognition module, a correlation analysis module, a dynamic model optimization module, a resource allocation decision module and a visual presentation module;
The data compression mapping module performs linear transformation on the data based on the multidimensional fire fighting data, analyzes and reflects key components of data variability, reduces data dimension, performs nonlinear mapping on the multidimensional data by adopting a t-distribution random neighborhood embedding method, and generates a compression mapping data set;
The flame trend prediction module predicts the speed and direction of fire development by combining historical fire data with real-time environmental parameters including temperature, humidity and wind speed based on a compressed mapping data set by adopting a random forest algorithm, and generates a fire development prediction result;
The scene recognition module analyzes a data structure in a fire scene by adopting a topology data analysis method based on a fire development prediction result, extracts key topology features from a data set, recognizes potential fire risks and risk structures in the fire scene, and generates a scene risk assessment result;
The association analysis module analyzes dynamic relations among fire facilities, personnel and fire sources by adopting a graph attention network based on scene risk assessment results, codes and analyzes entities and relations in a fire scene, optimizes resource allocation and analyzes rescue paths, and generates a dynamic relation graph;
The dynamic model optimization module carries out nonlinear identification and optimization on the fire dynamic model by adopting a kernel method and Gaussian process regression based on the dynamic relation diagram, describes the dynamic development of the fire under the differential environmental conditions, and generates an optimized dynamic model;
The resource allocation decision module is used for optimizing allocation of fire resources and planning rescue paths based on the optimized dynamic model by adopting a decision support algorithm and combining real-time environment data and the fire dynamic model, and comprises fire engines and firefighters to generate a resource allocation scheme;
The visual presentation module is based on a resource allocation scheme, adopts a data visualization technology, constructs a three-dimensional model of a fire scene through a three-dimensional graph display method, captures the spatial distribution and dynamic change of fire resources, and utilizes a dynamic graph display technology to update key data indexes in real time so as to generate a comprehensive visual interface.
The following is a further optimization of the above technical solution according to the present invention:
The compressed mapping data set comprises a feature set after dimension reduction and data point coordinates after nonlinear mapping, the fire development prediction result comprises a predicted fire expansion speed and a predicted fire expansion direction, the scene risk assessment result comprises a data structure of an identified fire mode and a potential risk, the dynamic relation graph comprises relations among facilities, personnel and fire sources and a criticality score, the optimized dynamic model comprises adjusted model parameters, the resource allocation scheme comprises firefighting resources of a target place and firefighting rescue paths, and the comprehensive visual interface comprises three-dimensional display of real-time data, a dynamic diagram of fire expansion and a visual path of resource allocation.
Further optimizing: the data compression mapping module comprises a dimension reduction sub-module, a nonlinear mapping sub-module and a data integration sub-module;
The dimension reduction submodule is used for carrying out linear conversion on data characteristics based on multidimensional fire fighting data by adopting a principal component analysis method, reducing the data dimension, and generating limited dimension data, wherein the limited dimension data comprises key component scores, variable contribution rates and accumulated contribution rates;
the nonlinear mapping submodule performs nonlinear mapping on multidimensional data by adopting a t-distribution random neighborhood embedding method based on limited dimensional data, shows the distribution condition of the data in a limited dimensional space and generates nonlinear mapping data;
The data integration submodule performs normalization processing on the data after nonlinear mapping by adopting a data formatting processing technology based on nonlinear mapping data, wherein the normalization processing comprises data standardization processing, data type conversion and missing value processing, and a compressed mapping data set is generated.
Further optimizing: the flame trend prediction module comprises a historical data analysis sub-module, a real-time data analysis sub-module and a prediction model training sub-module;
The historical data analysis submodule analyzes the historical fire data based on the compressed mapping data set, extracts key characteristics and modes including the development speed, the direction and the influenced environmental parameters of the historical fire, provides a training data set for constructing a prediction model, and generates the historical fire characteristic data;
The real-time data analysis sub-module is used for acquiring real-time environmental parameters including temperature, humidity and wind speed in real time based on the historical fire characteristic data, and predicting the development trend of fire by combining the historical fire characteristic data to generate comprehensive prediction input data;
The prediction model training submodule is used for training a fire behavior prediction model by adopting a random forest algorithm based on comprehensive prediction input data, predicting the speed and direction of fire development, and performing model training and optimization by combining historical data and real-time environmental parameters to generate a fire behavior development prediction result.
Further optimizing: the scene recognition module comprises a data structure analysis sub-module, a pattern recognition sub-module and a risk assessment sub-module;
The data structure analysis submodule adopts continuous coherent group analysis based on a fire development prediction result, performs fire data structure analysis by constructing and analyzing a multi-scale topology abstract of data, including life cycle and cave characteristics, extracts key topology characteristics in the data set and generates a topology characteristic set;
the pattern recognition submodule carries out multidimensional feature classification by using a vector machine based on a topological feature set and a machine learning classification algorithm, and establishes a classification rule by using a decision tree to recognize potential fire risks in a fire scene and generate a potential fire pattern recognition result;
the risk assessment submodule adopts a risk assessment model to construct a risk assessment framework based on fire probability based on a potential fire pattern recognition result, calculates and predicts occurrence probability of multiple types of risk events by combining a statistical analysis method and a Bayesian network, assesses risk levels in a fire scene, and generates a scene risk assessment result.
Further optimizing: the association analysis module comprises a graph construction sub-module, an attention mechanism analysis sub-module and a dynamic relationship updating sub-module;
The map construction submodule adopts a graph theory algorithm based on a scene risk assessment result, utilizes an adjacency matrix and an edge list construction algorithm to encode and structure the relation among fire facilities, personnel and fire sources, captures the dynamic relation among entities in a fire scene and generates a fire relation map;
the attention mechanism analysis submodule is used for carrying out dynamic relation analysis among entities in a fire scene by adopting a graph attention network based on the fire-fighting relation graph, dynamically distributing differentiated weights to nodes in the graph by introducing an attention mechanism, and generating a key characteristic relation graph;
the dynamic relation updating sub-module updates a time attenuation model and an event-triggered algorithm by adopting a dynamic diagram updating technology based on the key characteristic relation diagram, adjusts and updates nodes and edges in the diagram in real time, reflects the real-time state of a fire scene and generates a dynamic relation diagram.
Further optimizing: the dynamic model optimization module comprises a nonlinear identification sub-module, an environmental factor analysis sub-module and a model optimization sub-module;
The nonlinear identification submodule carries out nonlinear identification on the fire dynamic model by adopting a kernel method and Gaussian process regression based on the dynamic relation diagram, solves the nonlinear problem by converting data into a multidimensional space, provides probability distribution for prediction by utilizing Gaussian process regression, and generates a nonlinear identification model result;
the environmental factor analysis submodule analyzes the influence of various environmental factors on fire dynamics based on a nonlinear identification model result, wherein the influence of various environmental factors on fire dynamics comprises temperature, humidity and wind speed, quantitatively evaluates the influence of environmental variables on fire development by using a statistical analysis method and an environmental influence model, and generates an environmental influence analysis result;
The model optimization sub-module adjusts and optimizes the dynamic model by adopting a data-driven optimization method based on the environmental impact analysis result, identifies key environmental factors of the dynamic development influence of the fire disaster by analyzing the environmental impact analysis result, adjusts parameters of the dynamic model, and generates an optimized dynamic model.
Further optimizing: the resource allocation decision module comprises a resource evaluation sub-module, a path planning sub-module and a decision generation sub-module;
The resource evaluation submodule evaluates the fire protection resources based on the optimized dynamic model by using a resource evaluation algorithm, comprises linear programming and a resource optimization model, evaluates the real-time resource condition, predicts the resource requirement and generates a resource state evaluation result;
The path planning submodule adopts a path planning algorithm to plan a rescue path according to real-time fire resources based on the resource state evaluation result, comprehensively refers to various factors including path length, traffic conditions and barriers, calculates a path from a real-time position of the resources to a fire scene, and generates a rescue path planning scheme;
The decision generation submodule adopts a decision support method to carry out resource allocation and path decision based on a rescue path planning scheme, analyzes the influence of differentiated rescue resource allocation and path selection on rescue efficiency, evaluates the differentiated resource allocation scheme and the path selection through multi-criterion decision analysis, and generates a resource allocation scheme.
Further optimizing: the visual presentation module comprises a data display sub-module, an interactive design sub-module and an interface optimization sub-module;
The data display submodule adopts a data visualization technology based on a resource allocation scheme, simulates a fire scene through a three-dimensional graph, reflects data change in real time by utilizing a dynamic graph, tracks the dynamic state of key indexes and generates data visualization display;
The interactive design submodule adopts a user interface interactive design principle based on data visual display, and comprises responsive design and user-oriented interaction strategy, so that user experience is optimized, required information is acquired and analyzed, and an interactive design interface is generated;
The interface optimization submodule is based on an interactive design interface, adopts an interface performance optimization technology, reduces interface loading time through performance optimization, improves the fluency of user interaction, and utilizes the design of a visual effect strengthening method and an animation effect to improve interface attraction so as to generate a comprehensive visual interface.
The invention also provides a graphic display method of the digital twin fire control room, which is executed based on the graphic display system of the digital twin fire control room and comprises the following steps:
S1: based on multidimensional fire fighting data, adopting a principal component analysis method to perform linear conversion on data characteristics, reducing the dimension of a data set, and reserving key component scores, variable contribution rates and accumulated contribution rates to generate limited dimension data;
S2: based on the limited dimension data, adopting a t-distribution random neighborhood embedding method to carry out nonlinear mapping on the dimension-reduced data, and carrying out data integration and formatting processing to generate a compressed mapping data set;
S3: based on the compression mapping data set, a random forest algorithm is adopted, and a fire development prediction model is trained by combining historical fire data and real-time environmental parameters, so that the development speed and direction of a fire are predicted, and a fire development prediction result is generated;
s4: based on the fire development prediction result, adopting continuous coherent group analysis, analyzing a fire data structure, extracting key topological features, and identifying potential fire risks and risk grades in a fire scene by using a support vector machine through a machine learning classification algorithm to generate a scene risk assessment result;
S5: based on the scene risk assessment result, a firefighting relation graph between firefighting facilities, personnel and fire sources is constructed by adopting a graph theory algorithm, dynamic relations among entities are analyzed through a graph attention network, a dynamic graph updating technology is applied, and nodes and edges in the graph are adjusted in real time to generate a dynamic relation graph;
S6: based on the dynamic relation diagram, adopting a Gaussian process regression and kernel method to carry out nonlinear identification and optimization on a fire dynamic model, adjusting model parameters, analyzing fire dynamic development under differential environmental conditions, and generating an optimized dynamic model;
S7: based on the optimized dynamic model, optimizing and distributing fire resources and planning rescue paths by adopting a resource evaluation algorithm and a path planning algorithm, and generating a resource allocation scheme by referring to various factors including journey, speed and environmental obstacle;
S8: based on the resource allocation scheme, a data visualization technology is adopted, a fire scene is simulated, real-time data change is analyzed, a digital twin model of the fire scene is constructed by utilizing a three-dimensional modeling technology, key indexes in the fire resource allocation scheme are extracted through a real-time dynamic data processing method, and data are updated and dynamically displayed in real time, so that a comprehensive visual interface is generated.
By adopting the technical scheme, the fire control room has ingenious conception and reasonable structure, and by integrating data processing and analysis technologies, the fire control room comprises data compression mapping, flame trend prediction, scene recognition, association analysis, dynamic model optimization, resource allocation decision making and visual presentation, and the response capability of the fire control room to complex conditions is obviously improved; the method has the core advantages that the complex fire data can be effectively managed and analyzed, and the rescue resource allocation is optimized through accurate fire development prediction and risk identification, so that the high efficiency and the intellectualization of the fire rescue work are realized; by using advanced algorithms such as random forests and graph injection meaning networks, not only the accuracy of fire prediction is enhanced, but also the rationality of resource allocation is improved, visual and real-time data support is provided for command decisions through a three-dimensional visualization technology, and the rescue efficiency and the scientificity of the command decisions are remarkably improved.
The invention will be further described with reference to the drawings and examples.
Drawings
FIG. 1 is a system flow diagram of an embodiment of the present invention;
FIG. 2 is a schematic diagram of a system frame according to an embodiment of the present invention;
FIG. 3 is a flow chart of a data compression mapping module according to an embodiment of the present invention;
FIG. 4 is a flow chart of a flame trend prediction module according to an embodiment of the present invention;
FIG. 5 is a flowchart of a scene recognition module according to an embodiment of the present invention;
FIG. 6 is a flowchart of a correlation analysis module according to an embodiment of the present invention;
FIG. 7 is a flowchart of a dynamic model optimization module in an embodiment of the invention;
FIG. 8 is a flowchart of a resource allocation decision module according to an embodiment of the present invention;
FIG. 9 is a flow chart of a visual presentation module according to an embodiment of the present invention;
FIG. 10 is a schematic diagram of steps of a method according to an embodiment of the present invention.
Detailed Description
Embodiment one: referring to fig. 1 to 2, a digital twin fire control room graphic display system includes a data compression mapping module, a flame trend prediction module, a scene recognition module, a correlation analysis module, a dynamic model optimization module, a resource allocation decision module, and a visual presentation module.
The data compression mapping module performs linear transformation on the data based on the multidimensional fire fighting data, analyzes and reflects key components of data variability, reduces data dimension, performs nonlinear mapping on the multidimensional data by adopting a t-distribution random neighborhood embedding method, and generates a compression mapping data set.
The flame trend prediction module predicts the speed and direction of fire development by combining historical fire data with real-time environmental parameters including temperature, humidity and wind speed based on the compressed mapping data set by adopting a random forest algorithm, and generates a fire development prediction result.
The scene recognition module analyzes the data structure in the fire scene by adopting a topology data analysis method based on the fire development prediction result, extracts key topology features from the data set, recognizes potential fire risks and risk structures in the fire scene, and generates a scene risk assessment result.
The association analysis module analyzes dynamic relations among fire facilities, personnel and fire sources by adopting a graph attention network based on scene risk assessment results, codes and analyzes entities and relations in a fire scene, optimizes resource allocation and analyzes rescue paths, and generates a dynamic relation graph.
The dynamic model optimization module carries out nonlinear identification and optimization on the fire dynamic model by adopting a kernel method and Gaussian process regression based on the dynamic relation diagram, describes the dynamic development of the fire under the differential environmental conditions, and generates an optimized dynamic model.
The resource allocation decision module is used for optimizing allocation of fire resources and planning rescue paths based on the optimized dynamic model by adopting a decision support algorithm and combining real-time environment data and the fire dynamic model, and comprises fire engines and firefighters to generate a resource allocation scheme.
The visual presentation module is based on a resource allocation scheme, adopts a data visualization technology, constructs a three-dimensional model of a fire scene through a three-dimensional graph display method, captures the spatial distribution and dynamic change of fire resources, and utilizes a dynamic graph display technology to update key data indexes in real time to generate a comprehensive visual interface.
The compressed mapping data set comprises a feature set after dimension reduction and data point coordinates after nonlinear mapping, the fire development prediction result comprises a predicted fire expansion speed and a predicted fire expansion direction, the scene risk assessment result comprises a recognized fire mode and a potential risk data structure, the dynamic relation graph comprises relations among facilities, personnel and fire sources and key scores, the optimized dynamic model comprises adjusted model parameters, the resource allocation scheme comprises firefighting resources of a target place and firefighting rescue paths, and the comprehensive visual interface comprises three-dimensional display of real-time data, a dynamic diagram of fire expansion and a visual path of resource allocation.
In the data compression mapping module, multidimensional fire data is processed through Principal Component Analysis (PCA) and t-distribution random neighborhood embedding (t-SNE) methods, and multidimensional data related to fire conditions, such as temperature, humidity, smoke concentration and the like, are collected and arranged to form an initial data set; PCA is applied to carry out linear transformation and dimension reduction on data, main components reflecting variability of the data are selected, complexity of the data is effectively reduced, and meanwhile, most critical information is reserved; performing nonlinear mapping on the reduced dimension data by adopting t-SNE, and further revealing the internal structure and mode of the data, so that similar data points are more tightly gathered in a low-dimension space; by processing, the generated compressed mapping data set not only reduces the complexity of the original data, but also provides a clear and easy-to-process data base for the analysis of the subsequent modules, and improves the efficiency and accuracy of the whole system for processing multidimensional data.
In the flame trend prediction module, a random forest algorithm and real-time environmental parameters are combined to predict the development trend of the fire, historical fire data are analyzed, key features are extracted, and the real-time monitored environmental parameters such as temperature, humidity and wind speed are fused; training and analyzing data by constructing a plurality of decision trees by a random forest algorithm so as to estimate the expansion speed and direction of fire; by integrating the prediction results of a plurality of decision trees, the module can improve the accuracy and the robustness of prediction; the generated fire development prediction result is not only beneficial to identifying the potential path of fire spread in advance, but also provides scientific basis for timely fire control response and resource allocation, and remarkably improves the efficiency of emergency treatment of the fire.
In a scene recognition module, a Topology Data Analysis (TDA) method is applied to deeply explore a data structure in a fire scene, a topology abstract is constructed for a data set in a prediction result by using a TDA technology, key topology features such as inherent holes and clusters in data are recognized and extracted, and the features represent potential fire modes and risk structures; by analyzing the shape and the structure of the data, the scene recognition module can reveal complex modes and risk points which are not easy to directly observe in the fire scene, and the generated scene risk assessment result provides a deeper and comprehensive view angle for fire decision; by identifying and understanding the inherent links of fire data, the pertinence and effectiveness of fire prevention and countermeasures are greatly improved.
In the association analysis module, dynamic relations among fire facilities, personnel and fire sources are analyzed through a graph attention network (GAT), entities (such as the fire facilities, the personnel and the fire sources) in a fire scene and interaction thereof are converted into a graph form, nodes represent the entities, and edges represent the relations among the entities; processing the graph by using a graph annotation force network, wherein the network optimizes the resource allocation and analysis of rescue paths by allocating different weights to each node, wherein the weights reflect the importance of the relationship among the nodes; the system can identify which fire resources are most critical to control specific fire sources, so that a dynamic relationship graph is generated, and the real-time interaction relationship between the resources and the fire in a fire scene is revealed; the generated dynamic relation diagram provides deep understanding for the fire scene, and provides accurate data support for resource optimal allocation and rescue path planning.
In the dynamic model optimization module, a kernel method and a Gaussian Process Regression (GPR) are adopted to conduct nonlinear identification and optimization on the fire dynamic model, nonlinear problems are processed by mapping data to a multidimensional space, the GPR provides probability distribution for prediction of the model, and uncertainty assessment of the prediction is increased; by combining the dynamic relation diagram, the dynamic development of the fire disaster under different environmental conditions is deeply analyzed and modeled, and the model parameters are adjusted through an optimization algorithm, so that the accuracy and adaptability of the dynamic model of the fire disaster are improved; the optimized dynamic model can describe the behavior of fire disaster under various conditions more accurately, and provides powerful scientific basis for fire control decision, especially under complex or varied environmental conditions.
In the resource allocation decision module, a decision support algorithm is used for optimizing allocation and rescue path planning of fire resources, real-time environment data and the output of a fire dynamic model are integrated, and algorithms such as linear programming are used for evaluating and allocating the fire resources (fire engines and firefighters); considering the real-time state of resources, the predicted fire development trend and the potential influence thereof on the surrounding environment, thereby determining the most effective resource allocation and rescue path; through a refined decision-making process, the generated resource allocation scheme ensures that fire resources are utilized most reasonably, and rescue operation can respond to fire events rapidly and effectively.
In a visual presentation module, a three-dimensional model of a fire scene is constructed by adopting a data visualization technology through a three-dimensional graph display method, the spatial distribution and dynamic change of fire resources are displayed in an intuitive three-dimensional form by utilizing an advanced graph rendering technology, and meanwhile, key data indexes such as fire expansion speed, direction, resource arrival time and the like are updated in real time by using a dynamic graph display technology; the comprehensive visual interface not only provides real-time and accurate scene simulation and data analysis for fire command officers, but also greatly improves the efficiency and scientificity of fire decision making, so that the command officers can make quick and intelligent decisions under emergency conditions.
Referring to fig. 2 and 3, the data compression mapping module includes a dimension reduction sub-module, a nonlinear mapping sub-module, and a data integration sub-module.
The dimension reduction submodule adopts a principal component analysis method to perform linear conversion of data characteristics based on multidimensional fire fighting data, reduces data dimension, and comprises key component scores, variable contribution rates and accumulated contribution rates to generate limited dimension data.
The nonlinear mapping submodule performs nonlinear mapping of multidimensional data by adopting a t-distribution random neighborhood embedding method based on limited dimensional data, displays the distribution condition of the data in a limited dimensional space and generates nonlinear mapping data.
The data integration submodule performs normalization processing on the data after nonlinear mapping by adopting a data formatting processing technology based on nonlinear mapping data, wherein the normalization processing comprises data standardization processing, data type conversion and missing value processing, and a compressed mapping data set is generated.
In the dimension reduction submodule, a Principal Component Analysis (PCA) method is adopted, the object is to reduce the dimension of a data set through linear conversion of data characteristics, key components of fire fighting data, such as temperature, smoke concentration and the like, are extracted, and principal components which can most reflect the variability of the data are identified by utilizing the PCA analysis data; in operation, key component scores, variable contribution rates, and cumulative contribution rates are calculated to determine which components should be retained and which can be discarded, thereby generating a reduced-dimension dataset; the method reduces the calculation resources required by processing and analysis, reserves the most important information for fire prediction, and lays a foundation for subsequent analysis.
In the nonlinear mapping submodule, a t-distribution random neighborhood embedding (t-SNE) method is applied to the data with the reduced dimensionality for nonlinear mapping, wherein the t-SNE is a powerful machine learning algorithm used for exploring and visualizing the distribution situation of high dimensionality in a low-dimensional space; processing the dimension reduction data through t-SNE, and mapping to a two-dimensional or three-dimensional space to reveal the inherent modes and structures of the data, such as clustering trend; the fire control data can be more intuitively observed and analyzed, potential fire development trend and risk points can be identified, and support is provided for decision making.
In the data integration sub-module, formatting and normalizing are carried out on nonlinear mapping data generated by the t-SNE, the data are normalized to a unified measure, the consistency and the accuracy of model analysis are ensured, and the data after nonlinear mapping are subjected to normalization, data type conversion and missing value processing so as to ensure the quality and the integrity of a data set; the generated compressed mapping data set not only optimizes the data structure, but also provides an accurate and easy-to-process data basis for subsequent fire trend prediction and scene recognition.
Referring to fig. 2 and 4, the flame trend prediction module includes a historical data analysis sub-module, a real-time data analysis sub-module, and a prediction model training sub-module.
The historical data analysis submodule analyzes the historical fire data based on the compressed mapping data set, extracts key characteristics and modes including development speed, direction and influenced environmental parameters of the historical fire, provides a training data set for constructing a prediction model, and generates the historical fire characteristic data.
The real-time data analysis sub-module is used for collecting real-time environmental parameters including temperature, humidity and wind speed in real time based on the historical fire characteristic data, and predicting the development trend of the fire by combining the historical fire characteristic data to generate comprehensive prediction input data.
The prediction model training sub-module is used for training a fire behavior prediction model by adopting a random forest algorithm based on comprehensive prediction input data, predicting the speed and direction of fire behavior development, and performing model training and optimization by combining historical data and real-time environment parameters to generate a fire behavior development prediction result.
In the historical data analysis submodule, key features and modes of the historical fire are analyzed and extracted from the compressed mapping data set, the data set optimized by a principal component analysis and t-SNE method is used as a basis, and the data mining technology is used for deeply exploring the development speed, the development direction and the relation between the development speed and the development direction of the historical fire event and environmental parameters (such as temperature, humidity and wind speed); the main factors influencing the fire behavior are identified through statistical analysis and pattern recognition technology, so that a comprehensive characteristic data set is formed, and an accurate training data set is provided for the construction and training of a fire development prediction model.
In the real-time data analysis sub-module, real-time environmental parameters are monitored and collected in real time through analysis, the data collected by the Internet of things technology and the sensor network are combined with historical fire characteristic data, and real-time data processing and stream data analysis technology is used for immediately evaluating real-time fire risks and predicting the development trend of fire; by combining the historical data and the real-time data and through Complex Event Processing (CEP) technology, comprehensive prediction input data are generated, and an immediate update and comprehensive data view is provided for dynamic prediction of fire development.
In the prediction model training submodule, a random forest algorithm improves the accuracy and stability of prediction by establishing a plurality of decision trees and summarizing prediction results, and the algorithm analyzes the interaction between historical data and real-time environmental parameters and continuously adjusts and optimizes model parameters so as to adapt to complex and changeable environmental conditions and fire behavior modes; the generated fire development prediction result not only provides scientific basis for fire command, but also enhances understanding and prediction capability of future fire behavior.
Referring to fig. 2 and 5, the scene recognition module includes a data structure analysis sub-module, a pattern recognition sub-module, and a risk assessment sub-module.
The data structure analysis sub-module is used for analyzing the fire data structure by constructing and analyzing multi-scale topology abstracts of data, including life time and cave characteristics, adopting continuous coherent group analysis based on a fire development prediction result, extracting key topology characteristics in the data set and generating a topology characteristic set.
The pattern recognition submodule carries out multidimensional feature classification by using a vector machine based on the topological feature set and a machine learning classification algorithm, builds a classification rule by using a decision tree, recognizes potential fire risks in a fire scene and generates a potential fire pattern recognition result.
The risk assessment submodule adopts a risk assessment model to construct a risk assessment framework based on fire probability based on a potential fire pattern recognition result, calculates and predicts occurrence probability of multiple types of risk events by combining a statistical analysis method and a Bayesian network, assesses risk levels in a fire scene, and generates a scene risk assessment result.
In the data structure analysis submodule, multi-scale topology abstract construction and analysis of data are carried out aiming at a fire development prediction result through continuous coherent group analysis, the method comprises the steps of extracting life time and cave characteristics of a fire-fighting data set, and key topology structures in the data are identified by utilizing continuous coherent group theory, for example, by calculating Betti numbers in a multi-dimensional data set, and how the cave quantity, connectivity and characteristics of captured data change along with the scale. The process involves constructing a series of nested subspaces, called filters, computing coherent groups of subspaces, thereby generating a topological feature set, and collecting important topological structure information in summary data sets; the topological features of the data set, which are vital to fire prediction, are accurately extracted, so that accurate basic data are provided for pattern recognition, and high correlation and accuracy of analysis results are ensured.
In the pattern recognition sub-module, a machine learning classification algorithm combining a vector machine and a decision tree is adopted to recognize potential fire risks in a fire scene, a Support Vector Machine (SVM) is utilized to process multidimensional topological features, and a feature space is converted by selecting a proper kernel function such as a Radial Basis Function (RBF) so as to improve the classification resolution; the decision tree algorithm builds a classification rule according to the classification result of the SVM, optimizes the branching conditions by using information gain and other standards, and ensures that the generated decision tree model can efficiently identify different types of potential fire risks; through accurate classification mechanism, the identification of potential fire risk under the fire scene is more accurate, provides key basis for subsequent risk assessment.
In the risk assessment submodule, a risk assessment model based on fire probability is adopted, a statistical analysis method and a Bayesian network are combined to assess risk levels in a fire scene, a Bayesian network model is constructed, a potential fire pattern recognition result is taken as an input variable, various potential risk factors and conditional probabilities are comprehensively considered, and the occurrence probability of multiple types of risk events in different scenes is estimated and calculated by using the statistical analysis method such as a conditional probability table and prior probability; the model can dynamically adjust and optimize a risk assessment framework, and integrates the accuracy of statistics and Bayesian theory; by the method, accurate assessment of different risk levels in the fire scene can be generated, and scientific and accurate basis is provided for fire decision.
Referring to fig. 2 and 6, the association analysis module includes a graph construction sub-module, an attention mechanism analysis sub-module, and a dynamic relationship update sub-module.
The diagram construction submodule adopts a graph theory algorithm based on a scene risk assessment result, utilizes an adjacency matrix and an edge list construction algorithm to encode and structure the relation among fire facilities, personnel and fire sources, captures the dynamic relation among entities in a fire scene and generates a fire fighting relation diagram.
The attention mechanism analysis submodule is used for carrying out dynamic relation analysis among entities in a fire scene by adopting a graph attention network based on the fire-fighting relation graph, dynamically distributing differentiated weights to nodes in the graph by introducing an attention mechanism, and generating a key characteristic relation graph.
The dynamic relation updating sub-module adopts a dynamic diagram updating technology to update a time attenuation model and an event-triggered algorithm based on the key characteristic relation diagram, adjusts and updates nodes and edges in the diagram in real time, reflects the real-time state of a fire scene and generates a dynamic relation diagram.
In the diagram construction submodule, the relation between fire facilities, personnel and fire sources is coded and structured through a diagram theory algorithm, and the interaction relation between entities in a fire scene is converted into diagram form expression by adopting an adjacency matrix and edge list construction algorithm; the adjacency matrix is used for representing whether a two-dimensional array of edges exist among the nodes, the edge list lists all node pairs and the relations thereof, and the dynamic relations in the fire scene are accurately captured and encoded through two data structures to generate a fire-fighting relation graph; the method not only realizes the visualization of entity relationship in the fire scene, but also lays a structural foundation for the subsequent deep analysis; the generated fire-fighting relationship graph details the interaction relationship among facilities, personnel and fire sources, and provides an intuitive and structured tool for understanding and analyzing fire scenes.
In the attention mechanism analysis submodule, based on the generated fire-fighting relation graph, deep analysis is carried out by adopting a graph attention network (GAT), and aiming at dynamic relation development research among entities in a fire scene, the GAT dynamically distributes different weights for each node in the graph by introducing an attention mechanism so as to highlight key nodes and edges; the GAT evaluates the importance of adjacent nodes of each node, and adjusts the representation of node characteristics through the learned weight parameters so as to generate a key characteristic relation diagram; the method can identify the most critical entity and interaction for the analysis of the fire scene, and optimize the focus and efficiency of the subsequent analysis; the generated key characteristic relation graph provides accurate basis for fire control decision, highlights areas and entities needing important attention, and enhances the pertinence and effect of fire scene analysis.
In the dynamic relation updating sub-module, adopting a dynamic graph updating technology to adjust and update nodes and edges in the graph in real time so as to reflect the latest state of a fire scene, and adopting a time attenuation model and an event triggering algorithm to dynamically adjust the structure in the graph; the time attenuation model automatically adjusts the weight of the nodes and the edges according to the time change so as to simulate the dynamic change of the relation between the entities; the event triggering algorithm updates the graph structure in real time according to the newly-occurring event in the fire scene, such as newly-added nodes or edges, or adjusts the attributes of the nodes and edges; the dynamic updating mechanism ensures that the fire control relation graph can timely reflect the change in the fire scene, and the generated dynamic relation graph provides a real-time updated decision support tool for fire control monitoring and emergency response, so that the response speed and accuracy of fire control management are improved.
Referring to fig. 2 and 7, the dynamic model optimization module includes a nonlinear identification sub-module, an environmental factor analysis sub-module, and a model optimization sub-module.
Based on the dynamic relation diagram, the nonlinear identification submodule carries out nonlinear identification on the fire dynamic model by adopting a kernel method and Gaussian process regression, solves the nonlinear problem by converting data into a multidimensional space, provides probability distribution for prediction by utilizing Gaussian process regression, and generates a nonlinear identification model result.
The environmental factor analysis submodule analyzes the influence of various environmental factors on fire dynamics based on the nonlinear identification model result, wherein the influence of various environmental factors on fire dynamics comprises temperature, humidity and wind speed, quantitatively evaluates the influence of environmental variables on fire development by using a statistical analysis method and an environmental influence model, and generates an environmental influence analysis result.
The model optimization sub-module adjusts and optimizes the dynamic model by adopting a data-driven optimization method based on the environmental impact analysis result, identifies key environmental factors of the dynamic development influence of the fire disaster by analyzing the environmental impact analysis result, adjusts parameters of the dynamic model, and generates an optimized dynamic model.
In the nonlinear identification submodule, nonlinear identification is carried out on a fire dynamic model through a kernel method and Gaussian process regression, the kernel method is utilized to convert fire dynamic data from an original space into a multidimensional feature space, the nonlinear problem of the original data is solved, and a proper kernel function (such as radial basis function RBF) is selected to realize nonlinear mapping of the data, so that the data relationship in a high-dimensional space can be approximated by a linear model; then, learning and predicting the converted data by adopting a Gaussian Process Regression (GPR), wherein the GPR provides a flexible probability framework, and the GPR is used for predicting the nonlinear trend of fire dynamics by defining a priori distribution on a function space; the training of the GPR model involves the optimization of kernel function parameters and the estimation of noise level, so that the accuracy and reliability of a prediction result are ensured; and generating an identification model result capable of accurately reflecting the dynamic nonlinear characteristics of the fire disaster, and providing a basis for subsequent environmental factor analysis and model optimization.
In the environmental factor analysis submodule, comprehensively considering the influence of various environmental factors such as temperature, humidity and wind speed on fire dynamics, and quantitatively evaluating the relation between each environmental variable and fire development by adopting a statistical analysis method and an environmental influence model; by collecting corresponding environmental data and using statistical methods such as correlation analysis, regression analysis and the like, the influence degree of environmental variables on key indexes such as fire development speed, diffusion direction and the like is identified and quantified; the environmental impact model provides a comprehensive evaluation framework by integrating analysis results, and outputs quantitative indexes of each environmental factor on dynamic impact of fire; the generated environmental impact analysis result reveals the change rule of fire behavior under different environmental conditions, and provides scientific basis for formulating more accurate preventive measures and coping strategies.
In the model optimization sub-module, a data-driven optimization method is adopted to adjust and optimize the dynamic model, environmental factors which are most critical to the dynamic development influence of the fire are identified by analyzing the environmental influence analysis result in detail, and parameters of the dynamic model are adjusted according to the factors; re-evaluation and adjustment of model parameters are involved, such as changing weights of variables in the model or introducing new variables to better reflect the effects of environmental factors; the optimized dynamic model can more accurately simulate the behavior of the fire under different environmental conditions, and the accuracy and the practicability of fire prediction are improved; the optimized dynamic model is generated, so that the generalization capability of the model is enhanced, and a more effective tool is provided for fire prediction and management.
Referring to fig. 2 and 8, the resource allocation decision module includes a resource evaluation sub-module, a path planning sub-module, and a decision generation sub-module.
The resource evaluation sub-module evaluates the message resources based on the optimized dynamic model by using a resource evaluation algorithm, wherein the message resources comprise linear programming and a resource optimization model, evaluates real-time resource conditions, predicts resource requirements and generates a resource state evaluation result.
The path planning sub-module adopts a path planning algorithm to plan a rescue path according to real-time fire resources based on the resource state evaluation result, comprehensively refers to various factors including path length, traffic conditions and obstacles, calculates the path from the real-time position of the resources to the fire scene, and generates a rescue path planning scheme.
The decision generation submodule is based on a rescue path planning scheme, adopts a decision support method to carry out resource allocation and path decision, analyzes the influence of differentiated rescue resource allocation and path selection on rescue efficiency, analyzes and evaluates the differentiated resource allocation scheme and the path selection through multi-criterion decision making, and generates a resource allocation scheme.
In the resource evaluation sub-module, the resource evaluation algorithm is used for evaluating the fire fighting resources based on the optimized dynamic model, the application of the linear programming and the resource optimization model is involved, and the linear programming is used for determining the optimal solution of the resource configuration so as to minimize the response time and the resource use cost, and simultaneously, the availability and the demand prediction of the fire fighting resources are considered; the resource optimization model further analyzes the resource allocation problem, and ensures the high efficiency and fairness of resource allocation by considering the specific constraint and priority of different types of fire-fighting resources (such as fire-fighting vehicles, firefighters and rescue equipment); the algorithm comprehensively evaluates the real-time condition of the resource and predicts future resource demands by collecting and processing real-time resource condition data, including resource position, state and availability information; the generated resource state evaluation result is used for specifying real-time resource distribution, predicted demand and potential shortage, providing scientific basis for decision makers, guiding optimized allocation and future planning of resources.
In the path planning submodule, a path planning algorithm is adopted to accurately plan a rescue path, and various factors including the path length, traffic conditions, obstacles and the like are considered to ensure the efficient implementation of rescue actions; calculating an optimal path from a real-time position of a resource to a fire scene by a real-time traffic data and Geographic Information System (GIS), and simultaneously considering special traffic demands and encountered obstacles of emergency rescue vehicles; the rescue path planning scheme output by the algorithm indicates the travel route of the rescue team in detail, and the arrival time and the risk point are predicted, so that the resource utilization is maximized, the rescue time is shortened, and the rescue success rate is improved.
In the decision generation sub-module, a decision support method is adopted to carry out resource allocation and path decision, influences of different rescue resource allocation and path selection on rescue efficiency are deeply analyzed, and a multi-criterion decision analysis (MCDM) method is adopted to evaluate benefits and risks of different resource allocation schemes and path selection; taking a plurality of decision-making criteria such as rescue speed, resource utilization rate, safety and the like into consideration, and identifying optimal resource allocation and rescue paths through weight allocation and priority sequencing; the generated resource allocation scheme specifies how to dynamically allocate fire resources and select the optimal rescue path according to real-time resource conditions and fire scene requirements, ensures the high efficiency and effectiveness of rescue actions, and improves the response capability to sudden fire events.
Referring to fig. 2 and 9, the visual presentation module includes a data display sub-module, an interactive design sub-module, and an interface optimization sub-module.
The data display submodule adopts a data visualization technology based on a resource allocation scheme, simulates a fire scene through a three-dimensional graph, reflects data change in real time by utilizing a dynamic graph, tracks the dynamic state of key indexes and generates data visualization display.
The interactive design submodule adopts a user interface interactive design principle based on data visual display, and comprises responsive design and user-oriented interaction strategy, so that user experience is optimized, required information is acquired and analyzed, and an interactive design interface is generated.
The interface optimization submodule is based on an interactive design interface, adopts an interface performance optimization technology, reduces interface loading time through performance optimization, improves the fluency of user interaction, and utilizes the design of a visual effect strengthening method and an animation effect to improve interface attraction so as to generate a comprehensive visual interface.
In the data display sub-module, the information based on the resource allocation scheme is presented in a three-dimensional graph and dynamic chart form through a data visualization technology, so that the simulation of the fire scene and the tracking of real-time data change are realized; adopting an advanced graphic rendering technology such as WebGL, and combining a data visualization library such as D3.js and three.js to construct a three-dimensional fire scene and a dynamic chart; the three-dimensional graph simulates the spatial layout of fire-fighting resource allocation, such as the positions of fire-fighting vehicles and rescue workers, the fire source distribution and the like, while the dynamic graph reflects the changes of key indexes such as resource utilization rate, response time, fire control progress and the like in real time; the visual representation simplifies the complex data information, helps decision makers and related personnel to quickly understand scene conditions and resource dynamics, and improves the acceptability and decision efficiency of the information; the generated data visual display provides an intuitive and interactive interface for the user, and promotes more effective resource management and scheduling decision.
In the interactive design sub-module, user experience is optimized by adopting a user interface interactive design principle, adaptability of display contents on different equipment and screen sizes is ensured through responsive design, and a user-oriented interactive strategy focuses on simplifying a user operation flow and improving information retrieval efficiency; the optimization of interface layout, the reasonable arrangement of interactive elements and the setting of dynamic feedback mechanisms, such as interactive prompt and operation confirmation, are involved; design tools such as Sketch and Adobe XD are applied to refine user interface elements and interaction flows, so that a user can intuitively acquire required information, collect feedback through user testing, and further adjust design to meet user requirements; the generated interactive design interface not only improves the operation convenience and experience of the user, but also strengthens the understanding and application effects of the data, and supports the user to make more accurate and rapid decisions.
In the interface optimization sub-module, based on an interactive design interface, the smoothness of user interaction is improved by adopting an interface performance optimization technology, the interface loading time is reduced by front-end performance optimization measures such as code compression, asynchronous loading and buffer strategies, and meanwhile, the visual attractiveness is enhanced by utilizing CSS animation and advanced graphic effects such as shading and gradual change; performance analysis tools such as Google Lighthouse and WebPageTest are used to diagnose performance bottlenecks and guide the implementation of optimization measures; in consideration of diversity of user interaction, self-adaptive layout and touch operation support are introduced, so that excellent user experience can be provided on various devices; through the technical means, the generated comprehensive visual interface ensures attractive appearance and efficient performance, provides smooth and quick-response interaction experience for users, and enhances the participation degree and satisfaction degree of the users.
Referring to fig. 10, a digital twin fire control room graphic display method is executed based on the digital twin fire control room graphic display system, and includes the following steps:
S1: based on multidimensional fire fighting data, adopting a principal component analysis method to perform linear conversion on data characteristics, reducing the dimension of a data set, and reserving key component scores, variable contribution rates and accumulated contribution rates to generate limited dimension data;
s2: based on limited dimension data, adopting a t-distribution random neighborhood embedding method to carry out nonlinear mapping on the dimension-reduced data, and carrying out data integration and formatting processing to generate a compressed mapping data set;
S3: based on the compressed mapping data set, a random forest algorithm is adopted, and a fire development prediction model is trained by combining historical fire data and real-time environmental parameters, so that the development speed and direction of a fire are predicted, and a fire development prediction result is generated;
S4: based on a fire development prediction result, adopting continuous coherent group analysis, analyzing a fire data structure, extracting key topological features, and identifying potential fire risks and risk grades in a fire scene by using a support vector machine through a machine learning classification algorithm to generate a scene risk assessment result;
s5: based on scene risk assessment results, a firefighting relation graph between firefighting facilities, personnel and fire sources is constructed by adopting a graph theory algorithm, dynamic relations among entities are analyzed through a graph attention network, a dynamic graph updating technology is applied, nodes and edges in the graph are adjusted in real time, and a dynamic relation graph is generated;
S6: based on the dynamic relation diagram, adopting a Gaussian process regression and kernel method to carry out nonlinear identification and optimization on a fire dynamic model, adjusting model parameters, analyzing the fire dynamic development under the differential environment condition, and generating an optimized dynamic model;
S7: based on the optimized dynamic model, adopting a resource evaluation algorithm and a path planning algorithm to perform optimal allocation of fire resources and planning of rescue paths, and generating a resource allocation scheme by referring to various factors including journey, speed and environmental obstacle;
s8: based on a resource allocation scheme, a data visualization technology is adopted, a fire scene is simulated, real-time data change is analyzed, a digital twin model of the fire scene is constructed by utilizing a three-dimensional modeling technology, key indexes in the fire resource allocation scheme are extracted through a real-time dynamic data processing method, real-time updating and dynamic display of data are carried out, and a comprehensive visual interface is generated.
In step S1, the multidimensional fire data is linearly converted by a Principal Component Analysis (PCA) method, so as to reduce the dimension of the data set while retaining the most critical component score, variable contribution rate and cumulative contribution rate, thereby generating a finite-dimension data set, and various parameters in the fire data set, such as multidimensional data of temperature, humidity, wind speed, etc., are standardized to eliminate dimension influence and deviation of numerical value size from analysis results. The PCA algorithm is applied to calculate covariance matrixes of the data, characteristic values and characteristic vectors are extracted, and the first main components are selected as key components according to the size ordering of the characteristic values; the scores of the key components reflect most of the variability in the original data set, while the variable contribution rates and the cumulative contribution rates account for the importance of the principal components in describing the original data structure; the generated limited dimension data set not only greatly reduces the complexity of data processing, but also retains the information which is most critical to fire prediction, and provides a refined and effective data basis for subsequent analysis.
In step S2, nonlinear mapping is carried out on the limited dimension data obtained in step S1 through a t-distribution random neighborhood embedding method (t-SNE), the purpose of further compressing the data while maintaining the structural characteristics of the original data is achieved, the similarity of data points in an original space is calculated through the t-SNE method, the data points are converted into conditional probability representation, and a data point distribution is searched in the limited dimension space, so that the conditional probability distribution of the distribution is similar to the original space; searching for an optimal mapping by optimizing the KL divergence of the two probability distributions; the nonlinear structure problem which cannot be solved by the PCA method can be effectively solved, and the data set after dimension reduction can maintain the local structure characteristics of multidimensional data in a limited dimension space; the generated compressed map data set provides a formatted and structured data input for the fire development prediction model, enhancing the ability of the model to process nonlinear data.
In step S3, training a fire development prediction model based on a compression mapping data set and combining historical fire data and real-time environmental parameters by using a random forest algorithm, wherein the random forest is an integrated learning method, and the accuracy and stability of prediction are improved by constructing a plurality of decision trees and summarizing prediction results; in the training process, a random forest algorithm constructs each decision tree by randomly selecting different data subsets and feature subsets, and the introduction of randomness reduces the overfitting risk of the model and improves generalization capability; summarizing the prediction results of all decision trees by voting or averaging and other methods to generate prediction results of the development speed and direction of fire; not only can accurately predict the development trend of fire, but also can provide scientific basis for the formulation of fire countermeasures, and remarkably improve the scheduling efficiency of fire resources and the effect of fire control.
In step S4, analyzing the fire data structure based on the fire development prediction result by means of continuous coherent group analysis and Support Vector Machine (SVM), extracting key topology features, and identifying potential fire risks and risk classes in the fire scene, wherein the continuous coherent group analysis can reveal the basic structure of the data and the existence of "holes", and the topology features are critical for understanding the intrinsic properties of the data; performing pattern recognition based on the extracted topological features by adopting an SVM-machine learning classification algorithm; the SVM separates data of different categories by constructing an optimal hyperplane, and optimizes the generalization capability of the model; by combining the structural information obtained by continuous coherent group analysis and the classification capability of the SVM, different fire modes and risk levels in a fire scene can be effectively identified and classified, and the generated scene risk assessment result has important significance for fire prevention and emergency response, and the pertinence and the effectiveness of a fire strategy are improved.
In step S5, constructing and updating a fire-fighting relation graph in real time by combining a graph-theory algorithm with a graph attention network (GAT), wherein the graph-theory algorithm is used for defining nodes and edges among fire-fighting facilities, personnel and fire sources, constructing the fire-fighting relation graph and representing basic relations among entities; GAT analyzes dynamic relation among entities by distributing different attention weights to edges of multiple nodes, and optimizes judgment on entity importance; the dynamic graph updating technology adjusts the states or attributes of the nodes and the edges according to the real-time environmental change or event triggering, and maintains the timeliness of the relationship graph; the generated dynamic relation diagram can reflect real-time change of the relation among entities in the fire scene, provides a basis for subsequent nonlinear identification and optimization, and enhances adaptability and efficiency of a fire response strategy.
In step S6, nonlinear identification and optimization are carried out on the fire dynamic model through a Gaussian Process Regression (GPR) and a kernel method, and the GPR is used as a Bayesian regression technology to provide a flexible nonlinear model for predicting the development trend of the fire, so that the model can process complex nonlinear relations; performing deep analysis on the data generated by the dynamic relation diagram by using GPR, and adjusting model parameters to adapt to different environmental conditions, so that the accuracy of fire prediction is improved; the optimized dynamic model can more accurately predict the behavior of the fire disaster under various conditions, and provides a solid scientific basis for the fire-fighting resource allocation and the rescue action decision.
In step S7, optimizing and distributing the fire fighting resources and planning rescue paths by using a resource evaluation algorithm and a path planning algorithm, wherein the resource evaluation algorithm determines an optimal allocation scheme of the resources according to real-time resource conditions and predicted requirements; the path planning algorithm takes factors such as distance, speed, environmental obstacle and the like into consideration, and calculates the most effective rescue path; the fire-fighting resources can be effectively utilized according to actual needs, and rescue time is shortened to the greatest extent through scientific path planning, so that the rescue success rate is improved; the generated resource allocation scheme and rescue path planning provide important execution guidelines for fire control command, and are beneficial to improving the overall efficiency and effect of fire control rescue operation.
In step S8, a data visualization technology and a three-dimensional modeling technology are adopted, a fire scene is simulated based on a resource allocation scheme, real-time updating and dynamic displaying of data are carried out through a real-time dynamic data processing method, complex fire data and a resource allocation scheme are converted into an intuitive three-dimensional fire scene digital twin model, and dynamic tracking and displaying of key indexes are realized through the data visualization technology; not only can the fire command intuitively understand the real-time distribution of fire resources and rescue path planning, but also provides real-time data feedback and supports the dynamic adjustment of the decision process; the generated comprehensive visual interface provides an efficient and visual decision support tool for fire control command and rescue workers, and enhances the response speed and accuracy of fire control rescue actions.
The present invention is not limited to the above embodiments, and any equivalent embodiments which can be changed or modified by the technical disclosure described above can be applied to other fields, but any simple modification, equivalent changes and modification made to the above embodiments according to the technical matter of the present invention will still fall within the scope of the technical disclosure.
Claims (5)
1. A digital twin fire control room graphic display system, characterized in that: the system comprises a data compression mapping module, a flame trend prediction module, a scene recognition module, a correlation analysis module, a dynamic model optimization module, a resource allocation decision module and a visual presentation module;
The data compression mapping module performs linear transformation on the data based on the multidimensional fire fighting data, analyzes and reflects key components of data variability, reduces data dimension, performs nonlinear mapping on the multidimensional data by adopting a t-distribution random neighborhood embedding method, and generates a compression mapping data set;
The flame trend prediction module predicts the speed and direction of fire development by combining historical fire data with real-time environmental parameters including temperature, humidity and wind speed based on a compressed mapping data set by adopting a random forest algorithm, and generates a fire development prediction result;
The scene recognition module analyzes a data structure in a fire scene by adopting a topology data analysis method based on a fire development prediction result, extracts key topology features from a data set, recognizes potential fire risks and risk structures in the fire scene, and generates a scene risk assessment result;
The association analysis module analyzes dynamic relations among fire facilities, personnel and fire sources by adopting a graph attention network based on scene risk assessment results, codes and analyzes entities and relations in a fire scene, optimizes resource allocation and analyzes rescue paths, and generates a dynamic relation graph;
The dynamic model optimization module carries out nonlinear identification and optimization on the fire dynamic model by adopting a kernel method and Gaussian process regression based on the dynamic relation diagram, describes the dynamic development of the fire under the differential environmental conditions, and generates an optimized dynamic model;
The resource allocation decision module is used for optimizing allocation of fire resources and planning rescue paths based on the optimized dynamic model by adopting a decision support algorithm and combining real-time environment data and the fire dynamic model, and comprises fire engines and firefighters to generate a resource allocation scheme;
the visual presentation module is based on a resource allocation scheme, adopts a data visualization technology, constructs a three-dimensional model of a fire scene through a three-dimensional graph display method, captures the spatial distribution and dynamic change of fire resources, and utilizes a dynamic graph display technology to update key data indexes in real time to generate a comprehensive visual interface;
The data compression mapping module comprises a dimension reduction sub-module, a nonlinear mapping sub-module and a data integration sub-module;
The dimension reduction submodule is used for carrying out linear conversion on data characteristics based on multidimensional fire fighting data by adopting a principal component analysis method, reducing the data dimension, and generating limited dimension data, wherein the limited dimension data comprises key component scores, variable contribution rates and accumulated contribution rates;
the nonlinear mapping submodule performs nonlinear mapping on multidimensional data by adopting a t-distribution random neighborhood embedding method based on limited dimensional data, shows the distribution condition of the data in a limited dimensional space and generates nonlinear mapping data;
The data integration submodule performs normalization processing on the data after nonlinear mapping by adopting a data formatting processing technology based on nonlinear mapping data, wherein the normalization processing comprises data standardization processing, data type conversion and missing value processing, and a compressed mapping data set is generated;
the flame trend prediction module comprises a historical data analysis sub-module, a real-time data analysis sub-module and a prediction model training sub-module;
The historical data analysis submodule analyzes the historical fire data based on the compressed mapping data set, extracts key characteristics and modes including the development speed, the direction and the influenced environmental parameters of the historical fire, provides a training data set for constructing a prediction model, and generates the historical fire characteristic data;
The real-time data analysis sub-module is used for acquiring real-time environmental parameters including temperature, humidity and wind speed in real time based on the historical fire characteristic data, and predicting the development trend of fire by combining the historical fire characteristic data to generate comprehensive prediction input data;
the prediction model training submodule is used for training a fire behavior prediction model by adopting a random forest algorithm based on comprehensive prediction input data, predicting the speed and direction of fire development, and carrying out model training and optimization by combining historical data and real-time environmental parameters to generate a fire behavior development prediction result;
the association analysis module comprises a graph construction sub-module, an attention mechanism analysis sub-module and a dynamic relationship updating sub-module;
The map construction submodule adopts a graph theory algorithm based on a scene risk assessment result, utilizes an adjacency matrix and an edge list construction algorithm to encode and structure the relation among fire facilities, personnel and fire sources, captures the dynamic relation among entities in a fire scene and generates a fire relation map;
the attention mechanism analysis submodule is used for carrying out dynamic relation analysis among entities in a fire scene by adopting a graph attention network based on the fire-fighting relation graph, dynamically distributing differentiated weights to nodes in the graph by introducing an attention mechanism, and generating a key characteristic relation graph;
the dynamic relation updating sub-module updates a time attenuation model and an event-triggered algorithm by adopting a dynamic diagram updating technology based on the key characteristic relation diagram, adjusts and updates nodes and edges in the diagram in real time, reflects the real-time state of a fire scene and generates a dynamic relation diagram;
the dynamic model optimization module comprises a nonlinear identification sub-module, an environmental factor analysis sub-module and a model optimization sub-module;
The nonlinear identification submodule carries out nonlinear identification on the fire dynamic model by adopting a kernel method and Gaussian process regression based on the dynamic relation diagram, solves the nonlinear problem by converting data into a multidimensional space, provides probability distribution for prediction by utilizing Gaussian process regression, and generates a nonlinear identification model result;
the environmental factor analysis submodule analyzes the influence of various environmental factors on fire dynamics based on a nonlinear identification model result, wherein the influence of various environmental factors on fire dynamics comprises temperature, humidity and wind speed, quantitatively evaluates the influence of environmental variables on fire development by using a statistical analysis method and an environmental influence model, and generates an environmental influence analysis result;
The model optimization submodule adjusts and optimizes the dynamic model by adopting a data-driven optimization method based on an environmental impact analysis result, identifies key environmental factors of dynamic development influence of fire by analyzing the environmental impact analysis result, adjusts parameters of the dynamic model and generates an optimized dynamic model;
The visual presentation module comprises a data display sub-module, an interactive design sub-module and an interface optimization sub-module;
The data display submodule adopts a data visualization technology based on a resource allocation scheme, simulates a fire scene through a three-dimensional graph, reflects data change in real time by utilizing a dynamic graph, tracks the dynamic state of key indexes and generates data visualization display;
The interactive design submodule adopts a user interface interactive design principle based on data visual display, and comprises responsive design and user-oriented interaction strategy, so that user experience is optimized, required information is acquired and analyzed, and an interactive design interface is generated;
The interface optimization submodule is based on an interactive design interface, adopts an interface performance optimization technology, reduces interface loading time through performance optimization, improves the fluency of user interaction, and utilizes the design of a visual effect strengthening method and an animation effect to improve interface attraction so as to generate a comprehensive visual interface.
2. The digital twin fire control room graphic display system of claim 1, wherein: the compressed mapping data set comprises a feature set after dimension reduction and data point coordinates after nonlinear mapping, the fire development prediction result comprises a predicted fire expansion speed and a predicted fire expansion direction, the scene risk assessment result comprises a data structure of an identified fire mode and a potential risk, the dynamic relation graph comprises relations among facilities, personnel and fire sources and a criticality score, the optimized dynamic model comprises adjusted model parameters, the resource allocation scheme comprises firefighting resources of a target place and firefighting rescue paths, and the comprehensive visual interface comprises three-dimensional display of real-time data, a dynamic diagram of fire expansion and a visual path of resource allocation.
3. The digital twin fire control room graphic display system of claim 1, wherein: the scene recognition module comprises a data structure analysis sub-module, a pattern recognition sub-module and a risk assessment sub-module;
The data structure analysis submodule adopts continuous coherent group analysis based on a fire development prediction result, performs fire data structure analysis by constructing and analyzing a multi-scale topology abstract of data, including life cycle and cave characteristics, extracts key topology characteristics in the data set and generates a topology characteristic set;
the pattern recognition submodule carries out multidimensional feature classification by using a vector machine based on a topological feature set and a machine learning classification algorithm, and establishes a classification rule by using a decision tree to recognize potential fire risks in a fire scene and generate a potential fire pattern recognition result;
the risk assessment submodule adopts a risk assessment model to construct a risk assessment framework based on fire probability based on a potential fire pattern recognition result, calculates and predicts occurrence probability of multiple types of risk events by combining a statistical analysis method and a Bayesian network, assesses risk levels in a fire scene, and generates a scene risk assessment result.
4. The digital twin fire control room graphic display system of claim 1, wherein: the resource allocation decision module comprises a resource evaluation sub-module, a path planning sub-module and a decision generation sub-module;
The resource evaluation submodule evaluates the fire protection resources based on the optimized dynamic model by using a resource evaluation algorithm, comprises linear programming and a resource optimization model, evaluates the real-time resource condition, predicts the resource requirement and generates a resource state evaluation result;
The path planning submodule adopts a path planning algorithm to plan a rescue path according to real-time fire resources based on the resource state evaluation result, comprehensively refers to various factors including path length, traffic conditions and barriers, calculates a path from a real-time position of the resources to a fire scene, and generates a rescue path planning scheme;
The decision generation submodule adopts a decision support method to carry out resource allocation and path decision based on a rescue path planning scheme, analyzes the influence of differentiated rescue resource allocation and path selection on rescue efficiency, evaluates the differentiated resource allocation scheme and the path selection through multi-criterion decision analysis, and generates a resource allocation scheme.
5. A digital twin fire control room graphic display method, characterized in that the digital twin fire control room graphic display system according to any one of claims 1-4 is executed, comprising the steps of:
based on multidimensional fire fighting data, adopting a principal component analysis method to perform linear conversion on data characteristics, reducing the dimension of a data set, and reserving key component scores, variable contribution rates and accumulated contribution rates to generate limited dimension data;
based on the limited dimension data, adopting a t-distribution random neighborhood embedding method to carry out nonlinear mapping on the dimension-reduced data, and carrying out data integration and formatting processing to generate a compressed mapping data set;
Based on the compression mapping data set, a random forest algorithm is adopted, and a fire development prediction model is trained by combining historical fire data and real-time environmental parameters, so that the development speed and direction of a fire are predicted, and a fire development prediction result is generated;
Based on the fire development prediction result, adopting continuous coherent group analysis, analyzing a fire data structure, extracting key topological features, and identifying potential fire risks and risk grades in a fire scene by using a support vector machine through a machine learning classification algorithm to generate a scene risk assessment result;
Based on the scene risk assessment result, a firefighting relation graph between firefighting facilities, personnel and fire sources is constructed by adopting a graph theory algorithm, dynamic relations among entities are analyzed through a graph attention network, a dynamic graph updating technology is applied, and nodes and edges in the graph are adjusted in real time to generate a dynamic relation graph;
Based on the dynamic relation diagram, adopting a Gaussian process regression and kernel method to carry out nonlinear identification and optimization on a fire dynamic model, adjusting model parameters, analyzing fire dynamic development under differential environmental conditions, and generating an optimized dynamic model;
Based on the optimized dynamic model, optimizing and distributing fire resources and planning rescue paths by adopting a resource evaluation algorithm and a path planning algorithm, and generating a resource allocation scheme by referring to various factors including journey, speed and environmental obstacle;
Based on the resource allocation scheme, a data visualization technology is adopted, a fire scene is simulated, real-time data change is analyzed, a digital twin model of the fire scene is constructed by utilizing a three-dimensional modeling technology, key indexes in the fire resource allocation scheme are extracted through a real-time dynamic data processing method, and data are updated and dynamically displayed in real time, so that a comprehensive visual interface is generated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410309990.0A CN117910673B (en) | 2024-03-19 | 2024-03-19 | Digital twin firefighting control room graphic display system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410309990.0A CN117910673B (en) | 2024-03-19 | 2024-03-19 | Digital twin firefighting control room graphic display system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117910673A CN117910673A (en) | 2024-04-19 |
| CN117910673B true CN117910673B (en) | 2024-06-18 |
Family
ID=90695060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410309990.0A Active CN117910673B (en) | 2024-03-19 | 2024-03-19 | Digital twin firefighting control room graphic display system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117910673B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118627936A (en) * | 2024-08-09 | 2024-09-10 | 成都森拓沐然科技有限公司 | Forest fire prevention prediction method based on digital twinning |
Family Cites Families (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101437417B1 (en) * | 2013-03-08 | 2014-09-16 | 한국건설기술연구원 | Volcanic disaster response system |
| KR102087000B1 (en) * | 2019-08-13 | 2020-05-29 | 주식회사 지에스아이엘 | Method And System for Monitoring Fire |
| KR102518678B1 (en) * | 2021-01-27 | 2023-04-06 | 안상직 | Digital twin evacuation platform that provides safe evacuation route information in case of fire or disaster through digital twin |
| KR102521990B1 (en) * | 2021-04-13 | 2023-04-14 | 한기성 | Evacuation route simulation device using machine learning and learning method |
| KR102693387B1 (en) * | 2021-12-17 | 2024-08-12 | 국립안동대학교 산학협력단 | Disaster Monitoring System based on Digital Twin using Indoor Positioning and Real-time Remote Control and Method for Controlling the same |
| US20230342526A1 (en) * | 2022-03-28 | 2023-10-26 | Willow Labs, Inc. | System and method for wildfire spread behavior forecasting and on-parcel wildfire risk evaluation |
| CN115051926B (en) * | 2022-06-22 | 2023-06-09 | 烽火通信科技股份有限公司 | Digital twin device, model evaluation system and model operation method |
| CN115345371A (en) * | 2022-08-19 | 2022-11-15 | 中用科技有限公司 | Intelligent fire safety assistant decision-making method |
| CN115392708A (en) * | 2022-08-25 | 2022-11-25 | 爱瑞克(大连)安全技术集团有限公司 | Fire risk assessment and early warning method and system for building fire protection |
| CN115619609B (en) * | 2022-09-27 | 2024-02-27 | 无锡布塔信息科技有限公司 | Fire-fighting warning method and system based on three-dimensional visual data |
| CN116362113B (en) * | 2023-02-21 | 2024-03-12 | 北京图安世纪科技股份有限公司 | Accident disaster evolution simulation method and system based on digital twinning |
| CN116362549B (en) * | 2023-05-22 | 2023-08-04 | 北京航天常兴科技发展股份有限公司 | Fire disaster prevention and control method based on data information mining technology |
| CN116957484A (en) * | 2023-06-08 | 2023-10-27 | 中国消防救援学院 | Fire rescue plan method and system based on digital twinning |
| CN117408447A (en) * | 2023-09-18 | 2024-01-16 | 中国消防救援学院 | Fire-extinguishing rescue auxiliary decision-making method and system |
| CN117649130A (en) * | 2024-01-29 | 2024-03-05 | 深圳天益建设工程有限公司 | Intelligent fire safety monitoring system |
-
2024
- 2024-03-19 CN CN202410309990.0A patent/CN117910673B/en active Active
Non-Patent Citations (2)
| Title |
|---|
| 储罐火灾消防资源需求预测及辅助决策系统开发;周宁;印峥;崔汝卿;刘晅亚;黄维秋;毕海普;;油气田地面工程;20180820(08);全文 * |
| 火灾模拟分析与消防指挥决策的融合研究;冯凯, 徐志胜, 周庆;消防科学与技术;20050625(03);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117910673A (en) | 2024-04-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117910673B (en) | Digital twin firefighting control room graphic display system and method | |
| Izakian et al. | Anomaly detection and characterization in spatial time series data: A cluster-centric approach | |
| CN109508360B (en) | Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton | |
| CN116681250A (en) | Building engineering progress supervisory systems based on artificial intelligence | |
| Szabo et al. | Formalization of weak emergence in multiagent systems | |
| Vasudevan et al. | Concurrent consideration of evacuation safety and productivity in manufacturing facility planning using multi-paradigm simulations | |
| CN117725537A (en) | Real-time metering data processing platform | |
| CN117829534A (en) | BIM-based intelligent safety control method and system for road construction project | |
| CN117434912B (en) | Method and system for monitoring production environment of non-woven fabric product | |
| CN117787569B (en) | Intelligent auxiliary bid evaluation method and system | |
| CN117669384B (en) | Intelligent monitoring method and system for temperature sensor production based on Internet of things | |
| CN117787711A (en) | Arch bridge construction single-end management system based on big data | |
| CN117892887A (en) | Land use optimizing system based on big data | |
| CN112348285A (en) | Crowd evacuation simulation method in dynamic environment based on deep reinforcement learning | |
| CN117314006A (en) | Intelligent data analysis method and system | |
| CN117912717A (en) | Public health early warning platform | |
| Michel et al. | Weak interaction and strong interaction in agent based simulations | |
| CN117711131A (en) | Constructor warning protection system based on gesture danger analysis | |
| CN118133695A (en) | Subway station fire disaster temperature field simulation system and method based on big data processing and neural network | |
| CN113922502A (en) | Intelligent video operation and maintenance management system and management method | |
| KR20070043126A (en) | Artificial intelligence engine system for computer game | |
| Derix et al. | Near futures: associative archetypes | |
| CN113592311A (en) | Method for selecting human factors for complex man-machine system | |
| Li et al. | Construction Technology Safety Management under the Background of BIM and Information System Modeling | |
| CN118505021B (en) | Intelligent demonstration method and system for ecological protection red line project |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |