Building fire spreading simulation method, device, equipment and medium

Abstract

The invention discloses a building fire spreading simulation method, device and equipment and a medium, and the method comprises the steps: collecting the internal image data and temperature distribution data of a building, and carrying out the denoising and labeling of the data, and generating first data; inputting the first data into a recognition network, and training the recognition network; analyzing and identifying the internal structure and the temperature distribution condition of the building by using the trained identification network to obtain three-dimensional structure information of the building; based on the three-dimensional structure information of the building, building a three-dimensional model of the building by using three-dimensional modeling software; importing the building three-dimensional model into fire simulation software, setting fire simulation parameters, and performing fire simulation; the advanced neural network technology is utilized, and the three-dimensional model construction and virtual reality technology are combined, so that more accurate, real and visual fire spreading simulation is realized, and related personnel can be helped to better understand the fire development process and optimize building design and fire fighting measures.

Description

Technical Field

[0001] The present invention relates to the technical field of building fire treatment, and in particular to a building fire spread simulation method, device, equipment and medium. Background Art

[0002] With the continuous development and increase of urban buildings, fire safety issues have received increasing attention. As an important means of safety assessment, building fire spread simulation technology has important applications in the fields of building design, fire planning and emergency plan formulation. In the past few decades, building fire spread simulation technology has made great progress. Its main goal is to evaluate the impact of fire on the internal structure of the building, personnel escape and fire rescue by simulating the fire development process, and to guide building design and fire planning.

[0003] However, traditional building fire spread simulation technology has some shortcomings. First, traditional fire spread simulation usually relies on simplified building structure models, ignoring the complex structural characteristics inside the building, resulting in limited accuracy and authenticity of the simulation results. Second, existing fire spread simulations often fail to fully consider the heat distribution inside the building and lack a detailed description of the actual fire spread process. In addition, traditional fire spread simulation results are presented on a two-dimensional plane, which makes it difficult to intuitively display the fire spread process, limiting people's understanding and perception of the fire spread process. Summary of the invention

[0004] The present invention provides a method for simulating the spread of fire in a building, which makes up for the deficiencies of traditional technologies in building structure analysis, heat distribution description and simulation result presentation.

[0005] To solve the above technical problems, the present invention provides the following technical solutions, including: collecting image data and temperature distribution data inside a building, and performing denoising and labeling processing on the data to generate first data; inputting the first data into a recognition network to train the recognition network; using the trained recognition network to analyze and identify the internal structure and temperature distribution of the building to obtain the three-dimensional structural information of the building, including the spatial layout of the building, the shape and position of the building components, the size and proportion of the building, and the heat distribution inside the building; based on the three-dimensional structural information of the building, using three-dimensional modeling software to construct a three-dimensional model of the building; importing the three-dimensional model of the building into fire simulation software, setting fire simulation parameters, and performing fire simulation; the fire simulation parameters include: material properties, boundary conditions, fire sources, grids, flame temperature distribution, and smoke diffusion conditions.

[0006] As a preferred solution of the building fire spread simulation method described in the present invention, the three-dimensional model includes: the external form of the building: establishing the external outline of the building according to the spatial layout and the shape and position information of the components; the internal space layout: constructing the connection relationship between each floor and room according to the spatial layout information of the building, and presenting the layout of the internal space; structural components: creating structural components such as walls, columns, beams, and floor slabs according to the shape and position information of the building components, and ensuring their position and shape; heat distribution simulation: adding corresponding heat distribution effects to the three-dimensional model according to the heat distribution inside the building to intuitively display the heat distribution inside the building.

[0007] As a preferred solution of the building fire spread simulation method described in the present invention, the recognition network includes: a YOLO v3 network, a dynamic capsule network, a convolutional layer, and at least one fully connected layer; the first data is respectively input into the YOLO v3 network and the dynamic capsule network to obtain a first feature and a second feature; the first feature and the second feature are spliced ​​or superimposed through the convolutional layer to achieve feature fusion; the fused features are flattened into vectors and input into at least one fully connected layer for classification and regression; wherein the YOLO v3 network includes an input layer, a convolutional layer, a Darknet-53 feature extraction network, an SPP layer, a detection network and an output layer; the dynamic capsule network includes a plurality of dynamic capsule units, each dynamic capsule unit contains a feature vector represented by a vector and a capsule length represented by a scalar.

[0008] As a preferred solution of the building fire spread simulation method described in the present invention, the training includes: selecting the cross entropy loss function as the target loss function of the recognition network, and training the recognition network through the PyTorch Distributed framework, and continuously adjusting the network parameters and weight deviations until the target loss function value reaches the minimum.

[0009] As a preferred embodiment of the method for simulating the spread of building fires described in the present invention, the fire simulation includes: when performing the fire simulation, maintaining the data connection between the fire simulation software and the three-dimensional modeling software, so that the fire simulation results can be fed back to the three-dimensional modeling software in real time; observing the fire simulation results in real time through the three-dimensional modeling software, and the fire simulation results include the spread of fire and smoke diffusion.

[0010] As a preferred solution of the building fire spread simulation method described in the present invention, it also includes: transmitting the fire simulation results to a virtual reality device, entering a virtual scene through the virtual reality device, being able to freely observe the fire simulation results 360 degrees, and feeling the fire spread and smoke diffusion.

[0011] As a preferred solution of the building fire spread simulation method described in the present invention, it also includes: deploying sensors in the building to collect real-time temperature data inside the building when a fire occurs, and feeding it back to the recognition network, thereby realizing real-time updating and optimization of the simulation process.

[0012] As a preferred solution of the building fire spread simulation device described in the present invention, it includes: a data acquisition module, which is configured to execute the acquisition of internal image data and temperature distribution data of the building, and perform denoising and labeling processing on the data to generate first data; a network training module, which is configured to execute the input of the first data into the recognition network to train the recognition network; wherein the recognition network includes: a YOLO v3 network, a dynamic capsule network, a convolutional layer, and at least one fully connected layer; the first data is respectively input into the YOLO v3 network and the dynamic capsule network to obtain a first feature and a second feature; the first feature and the second feature are spliced ​​or superimposed through the convolutional layer to achieve feature fusion; the fused feature is flattened into a vector and input into at least one fully connected layer for classification and regression; the YOLO The v3 network includes an input layer, a convolutional layer, a Darknet-53 feature extraction network, an SPP layer, a detection network and an output layer; the dynamic capsule network includes multiple dynamic capsule units, each of which contains a feature vector represented by a vector and a capsule length represented by a scalar; the analysis and recognition module is configured to perform analysis and recognition of the internal structure and temperature distribution of the building through the trained recognition network, and obtain the three-dimensional structural information of the building, including the spatial layout of the building, the shape and position of the building components, the size and proportion of the building, and the heat distribution inside the building; the three-dimensional model construction module is configured to perform the construction of a three-dimensional model of the building based on the three-dimensional structural information of the building using three-dimensional modeling software; the fire simulation module is configured to perform the import of the three-dimensional model of the building into the fire simulation software, set the fire simulation parameters, and perform fire simulation; the fire simulation parameters include: material properties, boundary conditions, fire sources, grids, flame temperature distribution, and smoke diffusion conditions; the virtual reality simulation module is configured to perform the transmission of the fire simulation results to the virtual reality device, and enter the virtual scene through the virtual reality device, so that the fire simulation results can be freely observed 360 degrees, and the spread of fire and smoke can be felt.

[0013] The present invention also provides a building fire spread simulation device, comprising: one or more processors; a memory coupled to the processor and used to store one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the building fire spread simulation method as described in any of the above items.

[0014] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the building fire spread simulation method as described in any one of the above items.

[0015] Beneficial effects of the invention: The invention utilizes advanced neural network technology, and combines three-dimensional model construction and virtual reality technology to achieve a more accurate, realistic and intuitive fire spread simulation, which can help relevant personnel better understand the fire development process and optimize building design and firefighting measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them: Figure 1 A schematic diagram of the flow of a building fire spread simulation method according to a first embodiment of the present invention; Figure 2 This is a schematic diagram of the identification network structure described in the first embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0018] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0020] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0021] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0022] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0023] Example 1 Reference Figure 1~Figure 2 , which is the first embodiment of the present invention, provides a method for simulating the spread of a building fire, comprising: S1: Collect image data and temperature distribution data inside the building, and perform denoising and labeling processing on them to generate first data.

[0024] Devices such as cameras or laser scanners can be used to shoot or scan inside the building to obtain image data of the interior of the building, such as the structure and layout of different areas inside the building, and various types of equipment, facilities and equipment layout inside the building, such as air-conditioning equipment, fire-fighting equipment, lighting equipment, etc., which are not limited in the embodiments of the present invention.

[0025] Temperature distribution data can be obtained through a network of temperature sensors, which can be installed at key locations inside the building to monitor and record temperature information at different locations.

[0026] Key locations within the building include: Corridors and aisles: Corridors and aisles are the main escape routes for people. Installing sensors in these locations can timely monitor the temperature of the escape routes when a fire occurs.

[0027] Stairwell: The stairwell is also one of the escape routes. Installing a sensor in the stairwell can monitor the temperature changes in the stairwell and help guide people to escape safely.

[0028] Inside the room: Setting up sensors inside each room can monitor the temperature inside different rooms, which helps to understand the spread of fire.

[0029] Around electrical equipment: Electrical equipment is prone to become a fire source. Positions installed around electrical equipment can monitor whether there is overheating and discover potential fire hazards in time.

[0030] Ceilings and walls: Installing sensors on the ceiling and walls can monitor the temperature distribution at different heights and directions inside the building.

[0031] Ventilation vents and air conditioning vents: Ventilation vents and air conditioning vents are important channels for the spread of fire. Installing them at these locations can monitor air flow and temperature changes.

[0032] This is not limited in the embodiments of the present invention.

[0033] S2: Input the first data into the recognition network to train the recognition network.

[0034] Reference Figure 2 ,The recognition network includes: YOLO v3 network, dynamic capsule network, convolutional layer, and at least one fully connected layer.

[0035] Among them, the YOLO v3 network includes input layer, convolution layer, Darknet-53 feature extraction network, SPP layer, detection network and output layer; Convolution layer: used to extract features from input data; Darknet-53 feature extraction network: Darknet-53 is the backbone network of YOLO v3, consisting of 53 layers of convolutional neural network, which is responsible for deep feature extraction so that the subsequent target detection process can be based on these features; SPP layer (Spatial Pyramid Pooling): The SPP layer is used to process targets of different scales. By pooling the feature maps at different pooling sizes, SPP can capture information of different scales, thereby improving the accuracy of target detection; Detection network: The detection network includes various functional layers for target positioning and classification. These layers usually include convolution layers, batch normalization layers, and activation functions, etc., which are used to further process feature maps to generate target bounding boxes and category predictions; Output layer: The output layer is responsible for converting the prediction results generated by the detection network into the final target detection results. The output layer usually applies non-maximum suppression (Non-maximum Suppression, NMS) is used to eliminate overlapping bounding boxes and output the final target detection results, including target location and category.

[0036] The dynamic capsule network includes multiple dynamic capsule units, each of which contains a feature vector represented by a vector and a capsule length represented by a scalar. Multiple dynamic capsule units extract the spatial features of the input variables through a dynamic routing mechanism and learn, that is, use a dynamic routing algorithm to calculate the weights between capsules to promote better interaction and information transfer between capsules; it also includes additional layers for bounding box prediction, which are responsible for locating the target and regressing the bounding box.

[0037] The first data is input into the YOLO v3 network and the dynamic capsule network respectively to obtain the first feature and the second feature; specifically, the YOLO v3 network is responsible for detecting and extracting the objects inside the building structure, such as walls, doors, windows, etc. Through a series of convolutional layers and pooling layers, the YOLO v3 network can generate the first feature, which is used to represent the objects and structures inside the building. The dynamic capsule network will process the temperature distribution data, learn to capture the abstract features of the temperature distribution, and generate the second feature.

[0038] The first feature and the second feature are concatenated or superimposed through the convolution layer to achieve feature fusion; specifically, the first feature and the second feature obtained after being processed by their respective networks will be sent to a shared convolution layer, and the convolution layer will perform a feature fusion operation to concatenate or superimpose the first feature and the second feature to achieve feature fusion, which helps to combine visual information and temperature information and provide a more comprehensive and multi-dimensional feature representation for building structure recognition.

[0039] The fused features are flattened into vectors and input into at least one fully connected layer for classification and regression. The fully connected layer will learn how to map the fused features to specific building structure categories or attributes to achieve the final recognition and classification tasks.

[0040] Furthermore, to ensure that the recognition network can effectively learn the characteristics of the building structure and achieve good recognition results, it needs to be trained, including: The cross entropy loss function is selected as the target loss function of the recognition network, and the recognition network is trained through the PyTorchDistributed framework, and the network parameters and weight deviations are continuously adjusted until the target loss function value reaches the minimum.

[0041] Specifically, for the training of the recognition network, the cross entropy loss function is used to measure the difference between the model output and the true label. In PyTorch, the cross entropy loss function is implemented through torch.nn.CrossEntropyLoss. When using PyTorch for distributed training, follow the steps below to use the cross entropy loss function: 1. Define the loss function: In the model training code, you first need to define the cross entropy loss function. The sample code is as follows: import torch import torch.nn as nn # Define the cross entropy loss function criterion = nn.CrossEntropyLoss() 2. Use the loss function in the training loop: In the training loop, pass the output of the recognition network and the true label into the cross entropy loss function, calculate the loss value, and perform back propagation to update the recognition network parameters. The sample code is as follows: # Calculate loss and perform backpropagation at each training iteration output = model(data) # Assume that the model output is output loss = criterion(output, target) # Calculate cross entropy loss optimizer.zero_grad() loss.backward() optimizer.step() 3. In a distributed training environment, each process will calculate the loss value based on the current model parameters and data, and perform gradient aggregation to finally update the network parameters. The above is a common network training method and will not be described here.

[0042] S3: Analyze and identify the internal structure and temperature distribution of the building through the trained recognition network to obtain the three-dimensional structural information of the building.

[0043] The three-dimensional structural information of a building includes the spatial layout of the building, the shape and position of the building components, the size and proportion of the building, and the heat distribution inside the building.

[0044] Specifically, the spatial layout of the building: including the relationship between rooms and floors, the location of doors and windows, etc.; Shape and location of building components: For example, the specific location and shape of structural components such as walls, columns, beams, etc.; Size and proportion of buildings: The size and proportion of buildings can be inferred from image data; Materials and texture: Based on the image data, we can preliminarily determine the type and texture of materials used in the building. For example, we can observe the color and texture characteristics of the walls, floors, ceilings and other parts in the image. Different materials usually have different colors and textures. For example, wood, tiles, marble and other materials all have unique appearance characteristics. Or we can observe the light reflection of different parts of the image. Metal materials usually have strong reflective properties, while wood or cloth may show soft light reflection.

[0045] Temperature distribution information: Temperature distribution data can reflect the heat distribution inside the building, which helps to understand the building's insulation performance and heat loss.

[0046] In addition, the temperature data inside the building when a fire occurs can be collected in real time and fed back to the recognition network, which can re-analyze and process it to achieve real-time updating and optimization of the simulation process.

[0047] S4: Based on the three-dimensional structural information of the building, a three-dimensional model of the building is constructed using three-dimensional modeling software.

[0048] Based on the three-dimensional structural information of the building, a three-dimensional model of the building is constructed using three-dimensional modeling software (such as Blender, SketchUp, etc.).

[0049] The constructed 3D models include: External form of the building: Based on the spatial layout and the shape and location information of the components, the external outline of the building is established, including walls, windows, doors, etc. Internal space layout: Based on the spatial layout information of the building, the connection relationship between each floor and room is constructed to present the layout of the internal space; Structural components: Create walls, columns, beams, slabs and other structural components according to the shape and position information of building components, and ensure their position and shape; Heat distribution simulation: According to the heat distribution inside the building, add corresponding heat distribution effects in the 3D model, such as temperature band display, heat map, etc., to intuitively display the heat distribution inside the building.

[0050] S5: Import the 3D model into fire simulation software (such as Fire Dynamics Simulator - FDS, Pyrosim, SimScale, etc.), set fire simulation parameters, and perform fire simulation; fire simulation parameters include: material properties, boundary conditions, fire source, grid, flame temperature distribution, and smoke diffusion conditions.

[0051] Before importing the 3D model into the fire simulation software for fire simulation, some key fire simulation parameters need to be set to ensure the accuracy and effectiveness of the simulation. Specifically: 1. Material properties: Set corresponding material properties for different parts of the 3D model, including combustion characteristics, thermal conductivity, specific heat capacity, etc.

[0052] 2. Boundary condition setting: Determine the boundary conditions around the 3D model, including external airflow conditions, ambient temperature and other information to ensure the physical authenticity of the simulation process.

[0053] 3. Fire source setting: Determine parameters such as the location, size, intensity and duration of the fire, and set the properties of the fire source according to actual conditions, such as the fire source location, burning rate, burning temperature, etc.

[0054] 4. Establish a suitable grid: Establish a suitable grid in the simulation area and ensure that the grid is fine enough to accurately capture the dynamic process of the fire.

[0055] 5. Set flame temperature distribution: Set the flame temperature distribution according to the fire source, material properties, etc., including parameters such as flame height and shape.

[0056] 6. Set smoke diffusion conditions: Set smoke diffusion conditions, including smoke density, velocity, concentration and other parameters to simulate the spread of smoke in a fire.

[0057] 7. Run simulation: After completing the parameter settings, run the fire simulation software to perform simulation calculations to obtain results such as the spread of fire and smoke diffusion.

[0058] By setting these fire simulation parameters reasonably and combining them with the 3D model information of the building, accurate and reliable fire simulation can be performed, providing important reference for building fire safety design and emergency plan formulation. When setting parameters, adjustments and optimizations can be made according to actual conditions and needs to obtain simulation results that best meet actual conditions.

[0059] At the same time, when conducting fire simulation, the data connection between the fire simulation software and the 3D modeling software is maintained so that the fire simulation results can be fed back to the 3D modeling software in real time; The fire simulation results, including the spread of fire and smoke, can be observed in real time through 3D modeling software.

[0060] Furthermore, the fire simulation results are transmitted to virtual reality devices, such as head-mounted displays (VRHeadset), handheld devices (such as VR controllers), etc., and the virtual scene is entered through the virtual reality device, so that the fire simulation results can be freely observed 360 degrees, and the spread of fire and smoke can be felt. Interactive functions can also be added to allow users to interact with objects in the virtual environment, such as moving the perspective, zooming in and out of the scene, triggering specific events, etc. In addition, ensure that the virtual reality device can receive real-time updates of the fire simulation results and keep the simulation environment synchronized with the actual simulation process. Collect user feedback and behavior data in the virtual environment to evaluate the user's understanding and response to the fire scene in order to improve the simulation effect.

[0061] By combining the above methods with virtual reality technology to present fire simulation results, users can experience fire scenes in an immersive way, deepen their understanding of fire spread and evacuation, and improve their emergency response capabilities. This combination of virtual reality technology can not only improve the realism and interactivity of the simulation, but also bring better experience and results to fire safety training and education.

[0062] Example 2 This embodiment is different from the first embodiment in that it provides a building fire spread simulation device for executing the building fire spread simulation method as described in any of the above embodiments and achieving the same technical effects as the above methods, including: A data acquisition module is configured to acquire image data and temperature distribution data inside the building, and perform denoising and labeling processing on the data to generate first data; A network training module is configured to input the first data into the recognition network to train the recognition network; wherein the recognition network includes: a YOLO v3 network, a dynamic capsule network, a convolutional layer, and at least one fully connected layer; Inputting the first data into the YOLO v3 network and the dynamic capsule network respectively to obtain the first feature and the second feature; splicing or superimposing the first feature and the second feature through the convolution layer to achieve feature fusion; Flatten the fused features into a vector and input them into at least one fully connected layer for classification and regression; The YOLO v3 network includes an input layer, a convolutional layer, a Darknet-53 feature extraction network, an SPP layer, a detection network, and an output layer; the dynamic capsule network includes multiple dynamic capsule units, each of which contains a feature vector represented by a vector and a capsule length represented by a scalar; multiple dynamic capsule units extract the spatial features of the input variables through a dynamic routing mechanism, and learn, that is, use a dynamic routing algorithm (Dynamic Routing) to calculate the weights between capsules to promote better interaction and information transfer between capsules.

[0063] The analysis and recognition module is configured to perform analysis and recognition of the internal structure and temperature distribution of the building through the trained recognition network to obtain the three-dimensional structural information of the building, including the spatial layout of the building, the shape and position of the building components, the size and proportion of the building, and the heat distribution inside the building; A three-dimensional model building module is configured to build a three-dimensional model of the building using three-dimensional modeling software (such as Blender, SketchUp, etc.) based on the three-dimensional structural information of the building; The fire simulation module is configured to import the three-dimensional model into the fire simulation software (such as Fire DynamicsSimulator - FDS, Pyrosim, SimScale, etc.), set the fire simulation parameters, and perform the fire simulation; Fire simulation parameters include: material properties, boundary conditions, fire source, grid, flame temperature distribution, and smoke diffusion conditions; The virtual reality simulation module is configured to transmit the fire simulation results to virtual reality devices, such as head-mounted displays (VR Headsets), handheld devices (such as VR controllers), etc. By entering the virtual scene through the virtual reality device, the fire simulation results can be freely observed in 360 degrees, and the spread of fire and smoke can be felt.

[0064] Example 3 This embodiment provides a building fire spread simulation device, including: one or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the building fire spread simulation method as described above.

[0065] The processor is used to control the overall operation of the device to complete all or part of the steps of the above-mentioned building fire spread simulation method. The memory is used to store various types of data to support the operation of the device, and these data may include, for example, instructions for any application or method used to operate on the device, and data related to the application. The memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, referred to as EPROM), programmable read-only memory (Programmable Read-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0066] The device can be implemented by one or more application specific integrated circuits (Application Specific Integrated Circuit, referred to as ASC), digital signal processors (Digital Signal Processor, referred to as DSP), digital signal processing devices (Digital Signal Processing Device, referred to as DSPD), programmable logic devices (Programmable Logic Device, referred to as PLD), field programmable gate arrays (Field Programmable Gate Array, referred to as FPGA), controllers, microcontrollers, microprocessors or other electronic components, and is used to execute the building fire spread simulation method described in any of the above embodiments, and achieve the same technical effect as the above method.

[0067] Example 4 This embodiment provides a computer-readable storage medium, and when the program instructions are executed by a processor, the steps of the building fire spread simulation method as described in any of the above embodiments are implemented. For example, the computer-readable storage medium can be the above-mentioned memory including the program instructions, and the above-mentioned program instructions can be executed by the processor of the terminal device to complete the building fire spread simulation method as described in any of the above embodiments, and achieve the same technical effect as the above method.

[0068] It should be appreciated that embodiments of the present invention may be implemented or enforced by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable memory. The method may be implemented in a computer program using standard programming techniques - including a non-transitory computer-readable storage medium configured with a computer program, wherein the storage medium so configured causes the computer to operate in a specific and predefined manner - according to the methods and drawings described in the specific embodiments. Each program may be implemented in a high-level procedural or object-oriented programming language to communicate with a computer system. However, if desired, the program may be implemented in an assembly or machine language. In any case, the language may be a compiled or interpreted language. In addition, the program may be run on a programmed ASIC for this purpose.

[0069] Furthermore, the operations of the processes described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The processes described herein (or variations and / or combinations thereof) may be performed under the control of one or more computer systems configured with executable instructions, and may be implemented as code (e.g., executable instructions, one or more computer programs, or one or more applications) that is executed collectively on one or more processors, by hardware, or a combination thereof. The computer program includes a plurality of instructions that may be executed by one or more processors.

[0070] Further, the method can be implemented in any type of computing platform that is operably connected to a suitable computer, including but not limited to a personal computer, a minicomputer, a mainframe, a workstation, a network or distributed computing environment, a separate or integrated computer platform, or in communication with a charged particle tool or other imaging device, etc. Aspects of the present invention can be implemented in machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, an optical read and / or write storage medium, a RAM, a ROM, etc., so that it can be read by a programmable computer, and when the storage medium or device is read by the computer, it can be used to configure and operate the computer to perform the process described herein. In addition, the machine-readable code, or part thereof, can be transmitted via a wired or wireless network. When such media includes instructions or programs that implement the steps described above in conjunction with a microprocessor or other data processor, the invention described herein includes these and other different types of non-transitory computer-readable storage media. When programmed according to the methods and techniques described in the present invention, the present invention also includes the computer itself. The computer program can be applied to input data to perform the functions described herein, thereby converting the input data to generate output data stored in a non-volatile memory. The output information can also be applied to one or more output devices such as a display. In a preferred embodiment of the present invention, the converted data represents a physical and tangible object, including a specific visual depiction of the physical and tangible object produced on a display.

[0071] As used in this application, the terms "component", "module", "system", etc. are intended to refer to a computer-related entity, which can be hardware, firmware, a combination of hardware and software, software, or software in operation. For example, a component can be, but is not limited to: a process running on a processor, a processor, an object, an executable file, a thread in execution, a program, and / or a computer. As an example, an application running on a computing device and the computing device can both be components. One or more components can exist in a process and / or thread in execution, and a component can be located in a computer and / or distributed between two or more computers. In addition, these components can be executed from various computer-readable media having various data structures thereon. These components can communicate in a local and / or remote process manner, such as according to a signal having one or more data packets (e.g., data from a component that interacts with another component in a local system, a distributed system, and / or interacts with other systems in a signal manner through a network such as the Internet).

[0072] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for simulating the spread of building fire, characterized in that: include: Collecting image data and temperature distribution data inside the building, and performing denoising and labeling processing on the data to generate first data; Inputting the first data into the recognition network to train the recognition network; The trained recognition network is used to analyze and identify the internal structure and temperature distribution of the building, and obtain the three-dimensional structural information of the building, including the spatial layout of the building, the shape and position of the building components, the size and proportion of the building, and the heat distribution inside the building; Based on the three-dimensional structural information of the building, the three-dimensional model of the building is constructed using three-dimensional modeling software; Import the building 3D model into the fire simulation software, set the fire simulation parameters, and perform fire simulation; The fire simulation parameters include: material properties, boundary conditions, fire source, grid, flame temperature distribution, and smoke diffusion conditions.

2. The method for simulating the spread of building fire according to claim 1, characterized in that: The three-dimensional model includes: External form of the building: Establish the external outline of the building based on the spatial layout and the shape and location information of the components; Internal space layout: Based on the spatial layout information of the building, the connection relationship between each floor and room is constructed to present the layout of the internal space; Structural components: Create walls, columns, beams, slabs and other structural components according to the shape and position information of building components, and ensure their position and shape; Heat distribution simulation: According to the heat distribution inside the building, add the corresponding heat distribution effect in the 3D model to intuitively display the heat distribution inside the building.

3. The method for simulating the spread of building fire according to claim 2, characterized in that: The recognition network includes: YOLO v3 network, dynamic capsule network, convolutional layer, and at least one fully connected layer; Inputting the first data into the YOLO v3 network and the dynamic capsule network respectively to obtain the first feature and the second feature; The first feature and the second feature are concatenated or superimposed through a convolution layer to achieve feature fusion; Flatten the fused features into a vector and input them into at least one fully connected layer for classification and regression; Among them, the YOLO v3 network includes an input layer, a convolutional layer, a Darknet-53 feature extraction network, an SPP layer, a detection network and an output layer; the dynamic capsule network includes multiple dynamic capsule units, each of which contains a feature vector represented by a vector and a capsule length represented by a scalar.

4. The method for simulating the spread of building fire according to claim 3, characterized in that: The training includes: The cross entropy loss function is selected as the target loss function of the recognition network, and the recognition network is trained through the PyTorch Distributed framework, and the network parameters and weight deviations are continuously adjusted until the target loss function value reaches the minimum.

5. The method for simulating the spread of fire in a building according to claim 4, characterized in that: The fire simulation includes: When conducting fire simulation, the data connection between the fire simulation software and the 3D modeling software is maintained so that the fire simulation results can be fed back to the 3D modeling software in real time; The fire simulation results including the spread of fire and smoke diffusion are observed in real time through three-dimensional modeling software.

6. The method for simulating the spread of fire in a building according to claim 5, characterized in that: Also includes: The fire simulation results are transmitted to the virtual reality device. By entering the virtual scene through the virtual reality device, the fire simulation results can be freely observed 360 degrees, and the spread of fire and smoke can be felt.

7. The method for simulating the spread of fire in a building according to claim 6, characterized in that: Also includes: Sensors are deployed in buildings to collect real-time temperature data inside the building when a fire occurs, and the data is fed back to the recognition network, thereby achieving real-time updating and optimization of the simulation process.

8. A building fire spread simulation device, characterized in that: include: A data acquisition module is configured to acquire image data and temperature distribution data inside the building, and perform denoising and labeling processing on the data to generate first data; A network training module is configured to input the first data into the recognition network to train the recognition network; wherein the recognition network includes: a YOLO v3 network, a dynamic capsule network, a convolutional layer, and at least one fully connected layer; Inputting the first data into the YOLO v3 network and the dynamic capsule network respectively to obtain the first feature and the second feature; splicing or superimposing the first feature and the second feature through the convolution layer to achieve feature fusion; Flatten the fused features into a vector and input them into at least one fully connected layer for classification and regression; The YOLO v3 network includes an input layer, a convolutional layer, a Darknet-53 feature extraction network, an SPP layer, a detection network, and an output layer. The dynamic capsule network includes multiple dynamic capsule units, each of which contains a feature vector represented by a vector and a capsule length represented by a scalar. The analysis and recognition module is configured to perform analysis and recognition of the internal structure and temperature distribution of the building through the trained recognition network to obtain the three-dimensional structural information of the building, including the spatial layout of the building, the shape and position of the building components, the size and proportion of the building, and the heat distribution inside the building; The three-dimensional model building module is configured to build a three-dimensional model of the building using three-dimensional modeling software based on the three-dimensional structural information of the building; The fire simulation module is configured to import the building three-dimensional model into the fire simulation software, set fire simulation parameters, and perform fire simulation; the fire simulation parameters include: material properties, boundary conditions, fire source, grid, flame temperature distribution, and smoke diffusion conditions; The virtual reality simulation module is configured to transmit the fire simulation results to the virtual reality device. By entering the virtual scene through the virtual reality device, the fire simulation results can be freely observed 360 degrees and the spread of fire and smoke can be felt.

9. A building fire spread simulation device, characterized in that: The device includes: one or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the building fire spread simulation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the building fire spread simulation method according to any one of claims 1 to 7.

Images