A Multi-Physics Field Cooperative Wildfire Combustion Dynamic Simulation System

By establishing a multi-physics-field collaborative dynamic simulation system for wildfire combustion, the problem of lack of dynamic feedback in existing wildfire simulation models has been solved, enabling accurate simulation and reproduction of the dynamic evolution characteristics of wildfires and supporting forest fire prevention command and decision-making.

CN123069970APending Publication Date: 2026-11-10SICHUAN FIRE RES INST OF MEM +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611310228.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-27
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing wildfire simulation technologies neglect the dynamic correlation and bidirectional feedback between temperature field, flow field, radiation field and chemical reaction field, resulting in the inability to accurately reproduce the dynamic evolution characteristics of wildfires in complex environments.

Method used

A multi-physics collaborative dynamic simulation system for wildfire combustion was established. The system achieves bidirectional feedback between temperature field, flow field, radiation field and chemical reaction field through a physical field coupling solution module. Combined with an adaptive mesh generation module, an environmental coupling interface module, a multi-source data assimilation module and a visualization output module, a fully coupled dynamic simulation framework was constructed. Real-time calibration was performed using the large eddy simulation method and the ensemble Kalman filter algorithm.

Benefits of technology

It achieves accurate simulation of the dynamic evolution characteristics of wildfires, improves the accuracy and adaptability of simulation results, and can realistically reproduce the fire spread path and fire temperature distribution in complex environments, supporting forest fire prevention command and decision-making.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123069970A_ABST
    Figure CN123069970A_ABST
Patent Text Reader

Abstract

This invention discloses a multi-physics collaborative dynamic simulation system for wildfire combustion, belonging to the field of computer simulation technology. It includes modules for coupled physical field solution, adaptive mesh generation, environmental coupling interface, multi-source data assimilation, and visualization output. The system establishes a fully coupled dynamic simulation framework between temperature, flow, radiation, and chemical reaction fields, enabling bidirectional feedback between these physical fields. Adaptive mesh generation reduces the computational load of the fully coupled simulation. The environmental coupling interface enables bidirectional interaction between the fire field and the atmospheric boundary layer. Real-time calibration is performed using multi-source data assimilation. This invention overcomes the limitations of static superposition of multiple parameters, fully reproducing the dynamic evolution chain of wildfire combustion. While ensuring computational efficiency, it improves the accuracy and reliability of wildfire spread prediction in complex environments, providing strong support for emergency decision-making.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of computer simulation technology, specifically to a multi-physics collaborative dynamic simulation system for wildfire combustion. Background Technology

[0002] Multiphysics coupling simulation technology, as a core tool for studying complex thermodynamic systems, plays an increasingly important role in fields such as aerospace and energy. Wildfire combustion, as a typical multiphysics coupling phenomenon, involves the nonlinear interaction of temperature, flow, radiation, and chemical reaction fields. Its simulation modeling requires establishing a dynamic correlation mechanism between these physical fields.

[0003] In the field of wildfire simulation, existing technologies mostly employ a static superposition mode with multi-parameter coupling. This mode treats factors such as meteorological conditions, topography, and combustible material distribution as independent input parameters, incorporating them into the simulation model through weighted superposition. This method neglects the bidirectional feedback mechanism between various physical fields during combustion, resulting in simulation results that cannot accurately reproduce the dynamic evolution characteristics of wildfires in complex environments. Specifically, the static superposition mode treats the temperature field as a passive input to the flow field, ignoring the physical law that the heat released by the flame changes the local air density and thus affects the flow field distribution. Similarly, this mode does not consider the dynamic feedback effect of flow field changes on flame morphology and radiative heat transfer, nor the continuous driving effect of combustion chemical reactions on the temperature and component fields.

[0004] This technical solution, which ignores dynamic coordination mechanisms, reveals serious limitations when facing real wildfire scenarios. Due to the lack of a closed-loop feedback model between physical fields, existing simulation systems cannot characterize the complete dynamic evolution chain of "heat release during combustion → temperature field change → flow field adjustment → oxygen supply condition change → combustion rate change". Consequently, it is difficult to accurately predict key parameters such as fire spread path, flame morphology evolution, and fire temperature distribution. The simulation accuracy under complex terrain and meteorological conditions often fails to meet the needs of practical applications. Summary of the Invention

[0005] The present invention provides a multi-physics field collaborative wildfire combustion dynamic simulation system to solve the technical problem that the existing wildfire simulation model adopts a multi-parameter static superposition mode, which leads to the lack of closed-loop feedback and dynamic collaboration mechanism between physical fields, and thus cannot accurately predict and reproduce the dynamic evolution characteristics of wildfires in complex environments.

[0006] To solve the above problems, the technical solution adopted by the present invention is as follows:

[0007] A multi-physics collaborative wildfire combustion dynamic simulation system, the system comprising:

[0008] The physical field coupling solution module is used to establish and solve the fully coupled dynamic simulation framework between the temperature field, flow field, radiation field and chemical reaction field, and realize bidirectional feedback between the temperature field, the flow field, the radiation field and the chemical reaction field;

[0009] An adaptive mesh generation module is used to determine mesh densification regions and sparse regions based on the distribution of the temperature field and the position of the flame front. In the mesh densification regions, a high-resolution mesh is used for local fine calculations, and in the sparse regions, a low-resolution mesh is used for coarse-grained calculations. Interpolation processing is performed at the boundary between the mesh densification regions and the sparse regions to achieve data transfer between different resolutions.

[0010] The environmental coupling interface module is used to input the external environmental wind field as the upstream boundary condition for fire field calculation into the flow field, and to use the temperature anomaly and velocity disturbance caused by the heat released by the fire field as the feedback boundary condition of the external environmental wind field, so as to realize the bidirectional momentum exchange and heat exchange between the fire field and the atmospheric boundary layer.

[0011] The multi-source data assimilation module is used to obtain assimilated observations from multiple heterogeneous sources, update the state variables obtained from simulation calculations based on the assimilated observations to obtain posterior state estimates, and feed back the posterior state estimates to the physical field coupling solution module to update the initial conditions and boundary conditions of each physical field.

[0012] The visualization output module is used to present the simulation results output by the physical field coupling solution module in a graphical manner.

[0013] Furthermore, the physical field coupling solution module specifically includes: a temperature field solution submodule, a flow field solution submodule, a radiation field solution submodule, and a chemical reaction field solution submodule. The temperature field solution submodule is used to solve the energy conservation equation describing the relationship between the heat released during combustion and temperature changes, and to calculate the flow field acceleration driven by thermal buoyancy based on the temperature distribution. The flow field solution submodule is used to solve the Navier-Stokes equations describing the flow field distribution, and incorporates the thermal buoyancy source term and the component concentration gradient source term generated by the combustion chemical reaction as right-hand terms of the momentum equation into the flow field equation, thereby realizing the dynamic driving of the flow field by the temperature field and the chemical reaction field. The radiation field solution submodule is used to solve the radiation transfer equation using the discrete coordinate method to calculate the radiative heat transfer of the flame and high-temperature region, feeds back the radiative heat flux as the heat source term of the temperature field to the energy conservation equation, and calculates the shadow region using the ray tracing algorithm to introduce the shading effect of terrain and vegetation on radiation propagation; the chemical reaction field solution submodule is used to solve the chemical kinetic equation describing the pyrolysis and combustion reaction of combustibles using a multi-step reaction mechanism, calculates the combustion rate based on the oxygen concentration provided by the temperature field and the flow field, and feeds back the heat released by combustion and the reaction products generated to the energy conservation equation and the flow field equation.

[0014] Furthermore, in the calculation at each time step, the collaborative calculation process of the physical field coupled solution module is as follows: First, the chemical reaction field solution submodule calculates the combustion rate based on the oxygen concentration provided by the current temperature field and flow field, obtaining the heat released by combustion and reaction products; then, the temperature field solution submodule solves the energy conservation equation to obtain the updated temperature distribution, and simultaneously feeds back the generated reaction products as component source terms to the flow field solution submodule; next, the flow field solution submodule calculates the thermal buoyancy source term based on the updated temperature distribution, and solves the Navier-Stokes equations to obtain the updated flow field distribution; finally, the radiation field solution submodule calculates radiative heat transfer based on the updated temperature distribution and feeds back the radiative heat flux to the temperature field.

[0015] Furthermore, the environment coupling interface module uses the large eddy simulation method to model the external environment wind field. Specifically, the large-scale vortex and mean wind field are input as upstream boundary conditions to the flow field solving submodule in the following way: an inlet boundary condition is set at the upstream boundary of the computational domain, and the mean wind speed distribution is calculated using a logarithmic wind profile; a large-scale turbulent vortex is injected at the inlet boundary, and the vortex is generated using the synthetic turbulence method, simulating the large-scale vortex structure by superimposing sinusoidal perturbations of a set frequency and amplitude.

[0016] Furthermore, the multi-source data assimilation module specifically includes a satellite data processing unit, a UAV data processing unit, and a ground sensor data processing unit. The satellite data processing unit receives multispectral satellite remote sensing images, extracts the fire line location, fire line intensity, and burning area through spectral analysis algorithms, and maps the extracted parameters to the three-dimensional coordinate system of the simulation system through a coordinate transformation matrix as the assimilated observation values. The UAV data processing unit receives real-time images acquired by the infrared thermal imager and visible light camera mounted on the UAV, extracts the fire field temperature distribution from the infrared thermal image based on Planck's blackbody radiation law, and extracts flame morphology features from the visible light image as the assimilated observation values. The ground sensor data processing unit receives local environmental parameters collected by ground-deployed temperature sensors, gas sensors, and meteorological sensors as the assimilated observation values.

[0017] Furthermore, the multi-source data assimilation module utilizes an ensemble Kalman filter algorithm to fuse the assimilated observations with the state variables obtained from the simulation calculation. Specifically, the process involves: generating prior estimates of the state variables based on the simulation model to construct a state set; mapping each assimilated observation to the state space through an observation operator to obtain an observation set; calculating the prior covariance matrix and the observation covariance matrix, and combining them with the observation operator to calculate the Kalman gain matrix; and updating the prior estimates based on the Kalman gain matrix to obtain the posterior state estimates.

[0018] Furthermore, the visualization output module specifically includes a temperature field visualization submodule, a flow field visualization submodule, a radiation field visualization submodule, and a combustion process visualization submodule. The temperature field visualization submodule maps the temperature field data within the computational domain to a pseudo-color space to present a three-dimensional temperature distribution, and extracts temperature isosurfaces using an isosurface extraction algorithm to display the spatial location of the flame front. The flow field visualization submodule uses a streamline tracing algorithm to integrate along the velocity field direction to generate streamline trajectories to display the rising airflow and flow around the thermal convection, and draws velocity vector arrows on selected cross-sections to indicate the direction and magnitude of the flow velocity. The radiation field visualization submodule presents the spatial distribution of radiation intensity in the form of a heat flux density map, and uses threshold analysis to mark high-radiation hazard areas where the radiation intensity exceeds the safety threshold. The combustion process visualization submodule stores the spatial distribution data of the combustion area according to a time series, presents the advancement of the fire line over time in an animated form, and overlays the boundaries of the burned area onto the topographic map.

[0019] Furthermore, the system also includes a terrain data input module and a combustible material distribution module. The terrain data input module is used to receive digital elevation model data, set no-slip boundary conditions or logarithmic law wall functions on the terrain surface, use terrain elevation information as the terrain boundary conditions for the flow field calculation, and calculate the additional thermal buoyancy component caused by the slope terrain to reflect the influence of the slope terrain on thermal buoyancy and wind field. The combustible material distribution module is used to receive vegetation type distribution maps and combustible material moisture content distribution maps, map different vegetation types to corresponding chemical reaction parameters and input them to the chemical reaction field solving submodule, and use the combustible material moisture content distribution map as the input parameter for the moisture evaporation term to describe the inhibitory effect of the moisture evaporation term on the combustion process.

[0020] Furthermore, the system also includes a computational task scheduling module, which employs a dynamic load balancing strategy to divide the computational tasks of the physics coupling solution module into multiple sub-tasks and dynamically allocates task loads based on the available resources of each computing node in the distributed computing cluster. When the load of a node exceeds a preset threshold, some tasks are migrated to nodes with lower loads. The physics coupling solution module uses an implicit time discretization scheme and the finite volume method during the solution process. In each time step, it uses the Newton-Raphson iterative method to solve the nonlinear coupled equations and accelerates the convergence of the linear equations through a preconditional conjugate gradient algorithm or the generalized minimum residual method.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] (1) This invention establishes a fully coupled dynamic simulation framework between temperature field, flow field, radiation field and chemical reaction field, realizes bidirectional feedback between various physical fields, and fully reproduces the dynamic evolution chain of "heat release during combustion → temperature field change → flow field adjustment → oxygen supply condition change → combustion rate change", so that the simulation results can accurately reflect the dynamic evolution characteristics of real wildfires.

[0023] (2) By combining the partitioning fineness and coarseness of the adaptive mesh generation module with the dynamic load balancing and implicit time discretization of the computation task scheduling module, the problem of huge computational load faced by fully coupled computation is solved, making it possible to perform high-resolution wildfire simulation under limited computing resources.

[0024] (3) An environment-combustion bidirectional coupling mechanism was constructed, and large-scale turbulent vortices were injected through the large eddy simulation method to realistically reproduce the momentum and heat exchange process between the fire and the atmospheric boundary layer, thereby improving the adaptability and accuracy of the simulation model under complex and variable wind field conditions.

[0025] (4) Multi-source heterogeneous data, including satellite remote sensing, UAV aerial photography, and ground sensors, are introduced, and the ensemble Kalman filter algorithm is used to perform closed-loop updates on the simulation state variables. This mechanism enables the simulation system to perform real-time calibration based on measured data, avoids the accumulation of errors over time, enhances the reliability and practical value of the system, and provides support for forest fire prevention command and decision-making.

[0026] (5) The additional thermal buoyancy component brought by the digital elevation model and the influence of different vegetation types and water content on chemical kinetic processes (such as the water evaporation inhibition effect) are considered, so that the model closely matches the complex and ever-changing real forest fire terrain and combustible environment.

[0027] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, embodiments of the present invention are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a block diagram of the module composition of the multi-physics field collaborative wildfire combustion dynamic simulation system in this embodiment of the invention;

[0030] Figure 2 This is a schematic diagram illustrating the coupling and interaction of the various physics field solving submodules in an embodiment of the present invention;

[0031] Figure 3 This is a flowchart illustrating the data interaction between the environmental wind field and the fire field in an embodiment of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0033] Existing technologies in wildfire simulation mostly employ a static superposition mode with multi-parameter coupling, neglecting the dynamic correlation and bidirectional feedback between temperature, flow, radiation, and chemical reaction fields, thus failing to accurately reproduce the dynamic evolution characteristics of wildfires in complex environments. To address these technical problems, this invention provides a multi-physics collaborative wildfire combustion dynamic simulation system. This system typically runs on a distributed computing cluster server containing a multi-core central processing unit (CPU), a graphics processing unit (GPU), large-capacity memory, and high-speed network interconnection. The logical functions of each module are implemented by writing corresponding computer-readable instruction code. For example... Figure 1 As shown, the multi-physics field collaborative wildfire combustion dynamic simulation system provided in this embodiment of the invention includes: a physical field coupling solution module, an adaptive mesh generation module, an environmental coupling interface module, a multi-source data assimilation module, a visualization output module, a terrain data input module, a combustible material distribution module, and a computational task scheduling module.

[0034] The specific implementation methods of each module of the system are described in detail below:

[0035] I. Physical Field Coupling Solution Module

[0036] The physics-field coupling solution module is used to establish and solve a fully coupled dynamic simulation framework between the temperature field, flow field, radiation field, and chemical reaction field, achieving bidirectional feedback between these fields. For example... Figure 2 As shown, the physical field coupling solution module specifically includes four sub-modules: temperature field solution sub-module, flow field solution sub-module, radiation field solution sub-module, and chemical reaction field solution sub-module.

[0037] 1. Temperature Field Solving Submodule

[0038] The temperature field solution submodule is used to solve the energy conservation equation that describes the relationship between the heat released by combustion and the temperature change, and to calculate the flow field acceleration driven by thermal buoyancy based on the temperature distribution.

[0039] In its implementation, this module establishes a partial differential equation for energy conservation based on the first law of thermodynamics, and its corresponding governing equation is expressed in the following form:

[0040]

[0041] in, Air density (unit: ); This indicates the specific heat capacity of air at constant pressure (unit: );

[0042] Represents the temperature at the spatial grid node (unit: ); Indicates the simulation time; This represents the fluid velocity vector provided by the flow field solution submodule; Indicates thermal conductivity; This represents the heat source term released by combustion, passed from the chemical reaction field solution submodule. This represents the radiation heat transfer term passed from the radiation field solution submodule.

[0043] This energy conservation equation not only considers heat conduction ( ) and thermal convection ( It also receives energy feedback from the chemical reaction field and the radiation field through the right-hand source term, which reflects the coupling of multiple physics fields.

[0044] Furthermore, after obtaining the updated temperature distribution, the temperature field solution submodule passes it as a thermal buoyancy source term to the flow field solution submodule to drive the flow field to generate an upward airflow. The thermal buoyancy acceleration is calculated using the Bousinesko approximation, and the corresponding formula is:

[0045]

[0046] in, Indicates the coefficient of thermal expansion of a gas; Represents the gravitational acceleration vector; This indicates the reference temperature of the external environment.

[0047] 2. Flow field solution submodule

[0048] The flow field solution submodule is used to solve the Navier-Stokes (NS) equations that describe the flow field distribution. It incorporates the thermal buoyancy source term and the component concentration gradient source term generated by the combustion chemical reaction as right-hand terms of the momentum equation into the flow field equation, thereby realizing the dynamic driving of the flow field by the temperature field and the chemical reaction field.

[0049] In this embodiment, the fluid in the fire area is approximated as a low Mach number incompressible fluid, and the flow field solution submodule updates the flow field by solving the continuity equation and the momentum equation. The Navier-Stokes equations include:

[0050] Continuity equation:

[0051]

[0052] Momentum equation:

[0053]

[0054] in, Represents the fluid velocity vector; Indicates fluid pressure; Indicates air density; Indicates dynamic viscosity. Let represent the thermal buoyancy source term driven by the temperature field, and satisfy . ; The force source term representing the component concentration gradient generated by the combustion chemical reaction (caused by the volume expansion and density change due to the gaseous products generated by the chemical reaction) is used as the right-hand side of the momentum equation. By using these two source terms as the right-hand side of the momentum equation, the flow field solution submodule constructs a key feedback loop of "temperature field change → flow field adjustment → oxygen supply condition change (flow velocity change brings in more oxygen)".

[0055] 3. Radiation Field Solution Submodule

[0056] The radiation field solution submodule is used to solve the radiation transfer equation using the Discrete Ordinates Method (DOM) to calculate the radiative heat transfer of flames and high-temperature regions. The radiative heat flux is fed back to the energy conservation equation as the heat source term of the temperature field. The shadow region is calculated using the ray tracing algorithm to introduce the shading effect of terrain and vegetation on radiation propagation.

[0057] In wildfire scenarios, the radiative preheating of surrounding unburned combustibles by high-temperature flames is the primary mechanism for fire spread. The discrete coordinate method is used to solve the radiative transfer equation as follows:

[0058]

[0059] in, Indicates radiation intensity; Represents a spatial position vector; and The discrete direction vector represents the discretized space. Indicates the corresponding direction The direction cosine; Indicates the path length of radiation propagation; It represents the absorption coefficient of the air-fuel mixture (containing flue gas particles); Represents the scattering coefficient; This represents the Stefan-Boltzmann constant; This represents the integrator weights corresponding to different discrete directions.

[0060] During computation, this submodule acquires a 3D terrain model and vegetation canopy height, and uses a ray tracing algorithm to determine line-of-sight occlusion between mesh nodes. For occluded shadow areas, the intensity of direct radiation received is drastically attenuated; simultaneously, the calculated nodal radiative flux divergence is converted into a temperature field. item.

[0061] 4. Chemical Reaction Field Solving Submodule

[0062] The chemical reaction field solution submodule is used to solve the chemical kinetic equations describing the pyrolysis and combustion reactions of combustibles using a multi-step reaction mechanism. It calculates the combustion rate based on the oxygen concentration provided by the temperature field and the flow field, and feeds back the heat released by combustion and the reaction products generated to the energy conservation equation and the flow field equation.

[0063] The combustion of forest fuels such as timber involves both solid-phase pyrolysis and gas-phase combustion. The corresponding governing equations include:

[0064] Pyrolysis equation of combustible material:

[0065]

[0066] Combustion reaction equation:

[0067]

[0068] in, This indicates the density of solid combustible materials (such as branches and fallen leaves). This indicates the density of a gaseous combustible material (combustible gas produced by pyrolysis). This represents the oxygen density resulting from convection and diffusion in the flow field; Represents the pyrolysis rate constant; This represents the gas-phase combustion rate constant.

[0069] The core of chemical kinetic equations lies in the nonlinear characteristics of reaction rates, and the calculation of combustion rates is specifically based on the Arrhenius equation:

[0070]

[0071] in, General or ; This indicates the pre-exponential factor (frequency factor). Indicates activation energy; Represents the universal gas constant; This indicates the absolute temperature of the current mesh node, which is passed from the temperature field solving submodule.

[0072] After calculating the reaction rate within each grid cell using the above formula, this submodule calculates the generated heat based on the combustion enthalpy. And feed the mass of the gaseous products back to the flow field. Source term, complete the closed loop.

[0073] 5. Iterative collaborative workflow between modules

[0074] At each time step of the specific implementation (e.g. from) Time's up In the calculation of (time), the collaborative calculation process of the physical field coupling solution module strictly follows the following logical chain:

[0075] First, the chemical reaction field solution submodule is based on the current ( The combustion rate is calculated from the oxygen concentration provided by the temperature field and flow field at time (time). Heat released by internal combustion and reaction products;

[0076] Then, the temperature field solving submodule receives... Using the radiative heat from the previous step, we solve the energy conservation equation to obtain the updated temperature distribution. Meanwhile, the generated reaction products are fed back as component source terms to the flow field solution submodule;

[0077] Next, the flow field solution submodule calculates the flow field based on the updated temperature distribution. Calculate the latest thermal buoyancy source term And solve the Navier-Stokes equations to update the flow field velocity vector. and pressure field;

[0078] Finally, the radiation field solving submodule is based on the updated temperature distribution. Recalculate the spatial radiation field and calculate the new radiative heat transfer. The radiative heat flux is then fed back to the temperature field at the next time step. This series of processes constitutes a closed-loop feedback mechanism of "heat release from combustion → temperature field change → flow field adjustment → oxygen supply condition change → combustion rate change".

[0079] II. Adaptive Mesh Generation Module

[0080] Because wildfires involve scales ranging from several kilometers (macroscopic wind fields) to several millimeters (thin-layer chemical reactions in the flames), using a globally uniform high-resolution grid would lead to an exponential explosion in computational costs, exceeding the limits of current computing power. An adaptive grid generation module is used to determine grid-reinforced and sparse regions based on the temperature field distribution and the location of the flame front. High-resolution grids are used for fine-grained local calculations in the grid-reinforced regions, while low-resolution grids are used for coarse-grained calculations in the sparse regions. Interpolation is performed at the boundary between the grid-reinforced and sparse regions to achieve data transfer between different resolutions.

[0081] In practice, this module performs a mesh state determination before each time step iteration. First, it calculates the spatial temperature gradient of the current temperature field. ,Will Greater than a preset threshold (e.g.) The region containing the flame front position, along with the geometric envelope determined by images or calculations, is defined as the mesh-refined region; the remaining background flow field and unburned area are defined as the sparse region.

[0082] In terms of grid partitioning, the grid spacing of the encrypted region is set to... (For example The grid spacing in sparse regions is set to... (For example The two satisfy an integer multiple relationship. ,in This is the mesh density ratio, which is usually taken as 2 to 4 (e.g., octree meshing strategy).

[0083] In the fully coupled computation process, the physical field equations are solved in parallel on the two sets of meshes. At the boundary between the dense and sparse regions, physical quantities (such as velocity and temperature) need to satisfy flux conservation. The adaptive mesh generation module realizes the downsampling mapping from low resolution to high resolution data through bilinear interpolation (two-dimensional surface) or trilinear interpolation (three-dimensional volume element), and realizes the upsampling coarsening from high resolution to low resolution data through volume-weighted averaging. This effectively solves the problem of huge computational load faced by fully coupled computation and realizes high-precision simulation under limited computing power.

[0084] III. Environment Coupling Interface Module

[0085] The spread of wildfires depends not only on their own combustion but also on the strong influence of external environmental wind fields. For example... Figure 3 As shown, the environmental coupling interface module is used to input the external environmental wind field as the upstream boundary condition for fire field calculation into the flow field, and to use the temperature anomaly and velocity disturbance caused by the heat released by the fire field as the feedback boundary condition for the external environmental wind field, so as to realize the bidirectional momentum exchange and heat exchange between the fire field and the atmospheric boundary layer.

[0086] In a preferred embodiment, the environment coupling interface module uses Large Eddy Simulation (LES) to model the external wind field. LES directly analyzes the large-scale energetic turbulent eddies through filtering equations, while parametrically modeling the small-scale subgrid eddies with dissipative effects.

[0087] Specifically, this module inputs large-scale eddies and mean wind fields as upstream boundary conditions to the flow field solution submodule in the following manner:

[0088] The first step is to set inlet boundary conditions at the upstream boundary (inlet surface) of the entire fluid computational domain. To reflect the influence of surface friction on the real atmospheric boundary layer, the mean wind speed distribution is calculated using a logarithmic wind profile:

[0089]

[0090] in, Indicates vertical height Average wind speed at the location; It represents the friction velocity and characterizes the shear stress at the Earth's surface; This represents the von Kármán constant (usually taken as 0.41). Indicates zero-plane displacement (related to vegetation canopy height); It represents the length of the surface roughness.

[0091] The second step involves injecting only average wind, which is insufficient to elicit the turbulent jumping characteristics of the flame. Therefore, large-scale turbulent vortices must be introduced at the inlet boundary. These vortices are generated using synthetic turbulence methods (such as the Synthetic Eddy Method, SEM). This involves superimposing sinusoidal perturbations of a set frequency and amplitude onto the logarithmic average wind to simulate the large-scale vortex structure, generating a velocity field with time-varying fluctuations. The total inlet velocity is .

[0092] On the other hand, the environmental coupling interface module simultaneously receives velocity disturbances and temperature anomalies generated by the thermal buoyancy plume calculated from the fire scene, and feeds them back to the upper atmospheric boundary layer model. The strong updrafts in the fire scene will block or deflect the original environmental winds, forming... Figure 3 The “environment-combustion bidirectional coupling” is shown.

[0093] IV. Multi-source data assimilation module

[0094] Errors in the initial parameters of the simulation system accumulate and amplify over time. To enable real-time calibration, a multi-source data assimilation module is used to acquire assimilated observations from multiple heterogeneous sources. Based on these assimilated observations, the state variables obtained from the simulation are updated to obtain a posterior state estimate. This posterior state estimate is then fed back to the physics field coupled solution module to update the initial and boundary conditions of each physics field.

[0095] This module specifically includes three hardware / software collaborative sub-units: a satellite data processing unit, an unmanned aerial vehicle (UAV) data processing unit, and a ground sensor data processing unit.

[0096] 1. Satellite Data Processing Unit

[0097] This unit receives multispectral satellite remote sensing images (such as those from Fengyun, MODIS, and VIIRS) and extracts macroscopic dimensions of the fire line location, fire line intensity, and burn area through spectral analysis (such as the fixed brightness temperature threshold method or multi-channel context algorithm). Since the satellite images contain geographic coordinates, this unit maps the extracted parameters to the three-dimensional coordinate system of the simulation system using a coordinate transformation matrix as assimilated observation values. The mapping formula is:

[0098]

[0099] in, Represents the three-dimensional coordinates in the Cartesian coordinate system of the simulation system; This indicates the longitude, latitude, and elevation of the object extracted from the satellite imagery; This represents the rotation transformation matrix determined according to the projection system; This represents the translation vector.

[0100] 2. Unmanned Aerial Vehicle (UAV) Data Processing Unit

[0101] This unit receives real-time images from infrared thermal imagers and visible light cameras mounted on unmanned aerial vehicles (UAVs). Flame morphology features such as height and width are extracted from the visible light images. Based on Planck's blackbody radiation law, high-resolution temperature distribution of the fire scene is extracted from the infrared thermal images as assimilated observations. The temperature inversion formula is as follows:

[0102]

[0103] in, This represents the actual absolute temperature obtained from the inversion. and These are Planck's first and second radiation constants; This is the center value of the working wavelength of the infrared thermal imager; The pixel radiation intensity detected by the sensor.

[0104] 3. Ground sensor data processing unit

[0105] It is used to receive local microenvironmental parameters collected by temperature sensors, gas sensors (monitoring carbon monoxide and carbon dioxide), and meteorological sensors (wind speed and direction) deployed in forest areas through an IoT gateway, and use them as assimilated observation values.

[0106] 4. Ensemble Kalman Filter Assimilation Process

[0107] The multi-source data assimilation module utilizes the Ensemble Kalman Filter (EnKF) algorithm to fuse the assimilated observations with the state variables (model predictions) obtained from simulation calculations. Due to the extremely high dimensionality of the state variables, the specific process is as follows:

[0108] First, based on the simulation model, generate under certain perturbations. Prior estimates of the state variables of the model (e.g.) ), construct a state set ;

[0109] Each assimilated observation is passed through the observation operator. (That is, extracting operators to map the grid space to the sensor observation space) and mapping it to the state space to obtain the observation set. ;

[0110] Then the prior covariance matrix of the model state is calculated. and the sensor's own observation covariance matrix and combined with observation operators Calculate the Kalman gain matrix :

[0111]

[0112] Finally, based on the Kalman gain matrix Update the prior estimate to obtain the posterior state estimate. :

[0113]

[0114] in, This represents the prior state estimate variable before the update; Represents the observed variables in the assimilated observation set; This represents the updated posterior state estimate (i.e., the calibrated fire state). This posterior state estimate will be directly re-injected into the physics-coupled solver module to correct errors and ensure that the simulation closely matches the real fire situation.

[0115] V. Visualization Output Module

[0116] The massive amounts of multidimensional array data generated by the simulation system need to be presented intuitively to commanders through visualization. The visualization output module specifically includes four parallel sub-modules:

[0117] The temperature field visualization submodule maps temperature field data within the computational domain to a pseudo-color space (e.g., using a red-yellow-white gradient to represent dangerous temperatures from low to high) to present a three-dimensional temperature distribution. Most importantly, it uses isosurface extraction algorithms (such as the Marching Cubes algorithm) to extract isosurfaces in three-dimensional space that equal specific temperatures (e.g., [missing information]). The temperature isosurface of the ignition temperature surface is used to intuitively and clearly display the spatial location and complex topological structure of the flame front.

[0118] The flow field visualization submodule processes velocity vector data and employs a streamline tracing algorithm. Starting from a specified spatial seed point, it performs spatial integration along the velocity field direction to generate streamline trajectories. Its integral governing equations are:

[0119]

[0120] in, Represents the position vector on the streamline; Indicates the arc length parameter; This represents the velocity vector at that location; This represents the scalar magnitude of the velocity vector. Streamlines perfectly illustrate the updrafts, vortices, and wind patterns that bypass terrain obstacles in thermal convection. Simultaneously, velocity vector arrows are drawn on a selected two-dimensional cross-section (e.g., 2 meters above the ground), with the arrow pointing in the direction of the flow velocity and the length of the arrow mapping the magnitude of the flow velocity.

[0121] The radiation field visualization submodule is used to map the spatial distribution of radiation intensity in a 3D scene as a heat flux density map. In conjunction with fire safety standards, it uses threshold analysis to identify high-radiation-hazard areas where radiation intensity exceeds safety thresholds (e.g., the critical heat flux that would cause burns to firefighters), and marks them with flashing red or semi-transparent geometry.

[0122] The combustion process visualization submodule stores spatial distribution data of the combustion area according to real-time or accelerated time series. On a 3D topographic map including elevation, the red polygons representing the fire line are animated to show the progression of time, and the boundaries of the burned area (scorched earth zone) are overlaid on the topographic map to assist in the planning of emergency firefighting routes.

[0123] VI. Topographic Data Input Module and Combustible Material Distribution Module

[0124] The non-linearity of natural wildfires often stems from complex terrain and vegetation.

[0125] The terrain data input module is used to receive Digital Elevation Model (DEM) data. When using terrain elevation information as the terrain boundary condition for flow field calculation, no-slip boundary conditions or logarithmic wall functions are set on the terrain surface grid nodes (to avoid overly dense bottom-layer meshes).

[0126] Specifically, the slope of the terrain alters the relative direction of thermal buoyancy and gravity. This module calculates the terrain slope angle for each grid cell using DEM data. The additional thermal buoyancy component caused by the slope topography is calculated to reflect the influence of the slope topography on thermal buoyancy and wind field (the fire spread on the windward and leeward slopes has a significant difference of "fast uphill and slow downhill" due to the different buoyancy components). The expression for the additional thermal buoyancy component along the slope direction is:

[0127]

[0128] The combustible material distribution module receives vegetation type distribution maps and combustible material moisture content distribution maps provided by the forestry department. Different vegetation types (e.g., pine coniferous forests, shrublands, hay) are mapped to corresponding chemical reaction parameters (such as the aforementioned activation energy) through a lookup table. The module inputs the moisture content (such as the pyrolysis rate constant) into the chemical reaction field solution submodule. At the same time, the moisture content distribution map of the combustible material provides the moisture ratio. This module uses the moisture content as the input parameter of the moisture evaporation term into the energy conservation equation. Since moisture evaporation requires a large amount of heat absorption, the mathematical model can accurately describe the inhibitory effect of high moisture content (moisture evaporation term) on the combustion process.

[0129] VII. Computational Task Scheduling Module and Numerical Solution Strategy

[0130] Because this system involves solving large sparse matrices with extremely high computational complexity, a single machine cannot handle it. The system also includes a computational task scheduling module, which uses a dynamic load balancing strategy to divide the extremely large computational tasks of the physics-coupled solution module into multiple subtasks.

[0131] In practice, the computational task scheduling module monitors the available resources (CPU utilization, memory availability, network bandwidth) of each computing node (CPU core / GPU) in the distributed computing cluster in real time. The entire firefield mesh is divided into multiple computational subdomains using a region decomposition method. When the fire spread causes the load on a node (a surge in computational load due to increased mesh density) to exceed a preset threshold (e.g., CPU utilization exceeding 90%), the system automatically resets the mesh domain boundaries, migrating some computational tasks to nodes with lower loads, achieving dynamic load balancing and ensuring smooth simulation.

[0132] At the underlying numerical solution algorithm level, considering the extremely strong rigidity and nonlinearity of the fully coupled multiphysics equations, simple explicit schemes are prone to divergence. Therefore, the physics coupling solution module employs an implicit time discretization scheme to ensure absolute numerical stability (allowing for larger time steps). The time derivative term is approximately:

[0133]

[0134] Spatial discretization employs the finite volume method (FVM) to ensure strict local conservation of mass and energy. For the resulting nonlinear coupled equations, a Newton-Raphson iterative method is used for linearization at each time step, and the matrix equation form corresponding to the residual descent is as follows:

[0135]

[0136] in, This represents a large Jacobian matrix containing partial derivatives of multiple physical field variables; Represents the residual vector of a system of nonlinear equations; Indicates the number of iterations; Indicates the first The multiphysics state variable combination vector (including velocity, temperature, composition, etc.) of the next iteration. This represents the state increment to be determined.

[0137] To accelerate the convergence of the above linear equations, the system further employs Krylov subspace iterative algorithms such as the preconditional conjugate gradient algorithm (PCG) or the generalized minimum residual method (GMRES), and preprocesses the Jacobian matrix with incomplete LU decomposition (ILU) preconditioners. This reduces the computation time by tens of times while ensuring computational accuracy, thus meeting the timeliness requirements of emergency command.

[0138] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multi-physics collaborative dynamic simulation system for wildfire combustion, characterized in that, The system includes: The physical field coupling solution module is used to establish and solve the fully coupled dynamic simulation framework between the temperature field, flow field, radiation field and chemical reaction field, and realize bidirectional feedback between the temperature field, the flow field, the radiation field and the chemical reaction field; An adaptive mesh generation module is used to determine mesh densification regions and sparse regions based on the distribution of the temperature field and the position of the flame front. In the mesh densification regions, a high-resolution mesh is used for local fine calculations, and in the sparse regions, a low-resolution mesh is used for coarse-grained calculations. Interpolation processing is performed at the boundary between the mesh densification regions and the sparse regions to achieve data transfer between different resolutions. The environmental coupling interface module is used to input the external environmental wind field as the upstream boundary condition for fire field calculation into the flow field, and to use the temperature anomaly and velocity disturbance caused by the heat released by the fire field as the feedback boundary condition of the external environmental wind field, so as to realize the bidirectional momentum exchange and heat exchange between the fire field and the atmospheric boundary layer. The multi-source data assimilation module is used to obtain assimilated observations from multiple heterogeneous sources, update the state variables obtained from simulation calculations based on the assimilated observations to obtain posterior state estimates, and feed back the posterior state estimates to the physical field coupling solution module to update the initial conditions and boundary conditions of each physical field. The visualization output module is used to present the simulation results output by the physical field coupling solution module in a graphical manner.

2. The multi-physics collaborative wildfire combustion dynamic simulation system according to claim 1, characterized in that, The physical field coupling solution module specifically includes: a temperature field solution submodule, a flow field solution submodule, a radiation field solution submodule, and a chemical reaction field solution submodule. The temperature field solution submodule is used to solve the energy conservation equation describing the relationship between heat released during combustion and temperature changes, and to calculate the flow field acceleration driven by thermal buoyancy based on the temperature distribution. The flow field solution submodule is used to solve the Navier-Stokes equations describing the flow field distribution, and incorporates the thermal buoyancy source term and the component concentration gradient source term generated by the combustion chemical reaction as right-hand terms of the momentum equation into the flow field equation, thereby realizing the dynamic driving of the flow field by the temperature field and the chemical reaction field. The radiation field solution submodule is used to solve the radiation transfer equation using the discrete coordinate method to calculate the radiative heat transfer of the flame and high-temperature region. It feeds back the radiative heat flux as the heat source term of the temperature field to the energy conservation equation and calculates the shadow region using the ray tracing algorithm to introduce the shading effect of terrain and vegetation on radiation propagation. The chemical reaction field solution submodule is used to solve the chemical kinetic equations describing the pyrolysis and combustion reactions of combustibles using a multi-step reaction mechanism. It calculates the combustion rate based on the oxygen concentration provided by the temperature field and the flow field and feeds back the heat released by combustion and the reaction products generated to the energy conservation equation and the flow field equation.

3. The multi-physics collaborative wildfire combustion dynamic simulation system according to claim 2, characterized in that, In the calculation at each time step, the collaborative calculation process of the physical field coupled solution module is as follows: First, the chemical reaction field solution submodule calculates the combustion rate based on the oxygen concentration provided by the current temperature field and flow field, and obtains the heat released by combustion and reaction products; then, the temperature field solution submodule solves the energy conservation equation to obtain the updated temperature distribution, and feeds back the generated reaction products as component source terms to the flow field solution submodule; next, the flow field solution submodule calculates the thermal buoyancy source term based on the updated temperature distribution, and solves the Navier-Stokes equations to obtain the updated flow field distribution; finally, the radiation field solution submodule calculates radiative heat transfer based on the updated temperature distribution, and feeds back the radiative heat flux to the temperature field.

4. The multi-physics collaborative wildfire combustion dynamic simulation system according to claim 2, characterized in that, The environmental coupling interface module uses the large eddy simulation method to model the external environmental wind field. Specifically, large-scale vortices and the average wind field are input as upstream boundary conditions to the flow field solution submodule in the following way: inlet boundary conditions are set at the upstream boundary of the computational domain, and the average wind speed distribution is calculated using a logarithmic wind profile; large-scale turbulent vortices are injected at the inlet boundary, and the generation of the vortices is carried out using the synthetic turbulence method, which simulates the large-scale vortex structure by superimposing sinusoidal perturbations with a set frequency and amplitude.

5. The multi-physics collaborative wildfire combustion dynamic simulation system according to claim 2, characterized in that, The system also includes a terrain data input module and a combustible material distribution module. The terrain data input module receives digital elevation model data, sets no-slip boundary conditions or logarithmic law wall functions on the terrain surface, uses terrain elevation information as the terrain boundary conditions for the flow field calculation, and calculates the additional thermal buoyancy component caused by the slope terrain to reflect the influence of the slope terrain on thermal buoyancy and wind field. The combustible material distribution module receives vegetation type distribution maps and combustible material moisture content distribution maps, maps different vegetation types to corresponding chemical reaction parameters and inputs them to the chemical reaction field solving submodule, and uses the combustible material moisture content distribution map as the input parameter for the moisture evaporation term to describe the inhibitory effect of the moisture evaporation term on the combustion process.

6. The multi-physics collaborative wildfire combustion dynamic simulation system according to claim 1, characterized in that, The multi-source data assimilation module specifically includes a satellite data processing unit, a UAV data processing unit, and a ground sensor data processing unit. The satellite data processing unit receives multispectral satellite remote sensing images, extracts the fire line location, fire line intensity, and burning area using spectral analysis algorithms, and maps the extracted parameters to the three-dimensional coordinate system of the simulation system via a coordinate transformation matrix as the assimilated observation values. The UAV data processing unit receives real-time images acquired by an infrared thermal imager and a visible light camera mounted on the UAV, extracts the fire field temperature distribution from the infrared thermal image based on Planck's blackbody radiation law, and extracts flame morphology features from the visible light image as the assimilated observation values. The ground sensor data processing unit receives local environmental parameters collected by ground-deployed temperature sensors, gas sensors, and meteorological sensors as the assimilated observation values.

7. The multi-physics collaborative wildfire combustion dynamic simulation system according to claim 6, characterized in that, The multi-source data assimilation module uses an ensemble Kalman filter algorithm to fuse the assimilated observations with the state variables obtained from the simulation calculation. The specific process is as follows: generate prior estimates of the state variables based on the simulation model to construct a state set; map each assimilated observation to the state space through an observation operator to obtain an observation set. Calculate the prior covariance matrix and the observation covariance matrix, and calculate the Kalman gain matrix in conjunction with the observation operator; update the prior estimate based on the Kalman gain matrix to obtain the posterior state estimate.

8. The multi-physics collaborative wildfire combustion dynamic simulation system according to claim 1, characterized in that, The visualization output module specifically includes a temperature field visualization submodule, a flow field visualization submodule, a radiation field visualization submodule, and a combustion process visualization submodule; the temperature field visualization submodule is used to map the temperature field data in the computational domain to a pseudo-color space to present a three-dimensional temperature distribution, and to extract temperature isosurfaces through an isosurface extraction algorithm to display the spatial location of the flame front; The flow field visualization submodule is used to generate streamline trajectories by integrating along the velocity field direction using a streamline tracing algorithm to display the updraft and flow characteristics of thermal convection, and to draw velocity vector arrows on selected cross-sections to indicate the direction and magnitude of the flow velocity; the radiation field visualization submodule is used to present the spatial distribution of radiation intensity in the form of a heat flux density map, and to mark high radiation hazard areas where the radiation intensity exceeds the safety threshold through threshold analysis; the combustion process visualization submodule is used to store the spatial distribution data of the combustion area according to the time series, to present the progress of the fire line over time in the form of animation, and to overlay the boundary of the burned area on the topographic map.

9. The multi-physics collaborative wildfire combustion dynamic simulation system according to claim 1, characterized in that, The system also includes a computational task scheduling module, which uses a dynamic load balancing strategy to divide the computational tasks of the physics coupling solution module into multiple sub-tasks and dynamically allocate task loads according to the available resources of each computing node in the distributed computing cluster. When the load of a node exceeds a preset threshold, some tasks are migrated to nodes with lower loads. The physics coupling solution module uses an implicit time discrete scheme and the finite volume method to solve the nonlinear coupled equations in each time step. It uses the Newton-Raphson iterative method to solve the nonlinear coupled equations and accelerates the convergence of the linear equations by using the preconditioned conjugate gradient algorithm or the generalized minimum residual method.