Resident indoor explosion multi-load decoupling test method and system

Abstract

The invention relates to a resident indoor explosion multi-load decoupling test method and system, and belongs to the field of explosion effect test. The method comprises the following steps: acquiring original sensor data and original image data; based on the original sensor data and the original image data, analyzing different loads by adopting a plurality of single-load algorithms to obtain a plurality of single-load analysis results; integrating all single-load analysis results into an explosion effect data set, and processing the explosion effect data set by adopting a multi-load decoupling model to obtain a multi-load decoupling analysis result; the multi-load decoupling model is constructed by adopting a multi-load correlation analysis algorithm, a load contribution degree evaluation algorithm and a multi-load decoupling algorithm; and comparing an actual explosion test result with a multi-load decoupling analysis result, and determining an explosion multi-load decoupling effect according to a comparison result. According to the invention, the multi-load decoupling test on the indoor explosion of residents can be realized, and the test accuracy and reliability are improved.

Description

Technical Field

[0001] This application relates to the field of explosion effect testing, and specifically relates to a method and system for decoupling multi-loads in indoor explosion of residents, which is used to evaluate and study the influence of loads on building structures and internal facilities. Background Art

[0002] With the acceleration of the urbanization process, the safety of indoor spaces for residents has received increasing attention. In some special cases, such as explosion accidents caused by gas leakage, it will pose a serious threat to the lives and property safety of residents. At present, the testing methods for indoor explosion of residents mainly focus on small-scale experiments or measurements of specific parameters, lacking a method for decoupling multi-loads in the full-scale large indoor space of residents. The existing testing methods often cannot comprehensively and accurately reflect the changes of various physical parameters during the explosion process, and it is also difficult to effectively decouple and analyze different loads, such as overpressure, thermal radiation, shock waves, etc. The coupling effects of these loads are crucial for structural safety and personnel safety.

[0003] In order to improve the accuracy of the investigation and aftermath handling of indoor explosion accidents of residents, there is an urgent need to develop a new testing method to simulate the explosion process in a real indoor environment of residents and be able to accurately measure and analyze various loads generated by the explosion. Therefore, developing a method for decoupling multi-loads in the full-scale large indoor space of residents is of great significance for improving the prevention and response capabilities of indoor explosion accidents of residents and ensuring the safety of people's lives and property. Summary of the Invention

[0004] The embodiments of this application provide a method and system for decoupling multi-loads in indoor explosion of residents to solve the problems of low testing accuracy and poor reliability existing in the prior art.

[0005] In the first aspect, the embodiments of this application provide a method for decoupling multi-loads in indoor explosion of residents, including: Obtaining the original sensor data collected by various types of sensors and the original image data collected by various types of image acquisition devices in the system for decoupling multi-loads in indoor explosion of residents; Based on the original sensor data and the original image data, using multiple single-load algorithms to analyze different loads, and obtaining multiple single-load analysis results; Integrating all single-load analysis results into an explosion effect data set, and using a multi-load decoupling model to process the explosion effect data set to obtain a multi-load decoupling analysis result; the multi-load decoupling model is constructed by using a multi-load correlation analysis algorithm, a load contribution degree evaluation algorithm, and a multi-load decoupling algorithm; Compare the actual explosion test results with the multi-load decoupling analysis results, and determine the explosion multi-load decoupling effect according to the comparison results.

[0006] Preferably, a multi-load decoupling model is used to process the explosion effect data set to obtain multi-load decoupling analysis results; the multi-load decoupling model is constructed by using a multi-load correlation analysis algorithm, a load contribution degree evaluation algorithm, and a multi-load decoupling algorithm, and includes: Use the multi-load correlation analysis algorithm to perform correlation analysis on the explosion effect data set to obtain a load correlation matrix; Use the load contribution degree evaluation algorithm to evaluate and process the load correlation matrix to obtain a load contribution degree matrix; Use the multi-load decoupling algorithm to process the load contribution degree matrix to obtain a decoupling effect matrix; among them, the multi-load decoupling algorithm performs different decoupling analyses on different sub-regions based on the position of the explosion source and the explosion influence range; Use the time series analysis algorithm to analyze the decoupling effect matrix to obtain multi-load decoupling analysis results.

[0007] Preferably, the step of using the multi-load decoupling algorithm to process the load contribution degree matrix to obtain a decoupling effect matrix includes: Determine the position of the explosion source and the explosion influence range according to the original sensor data and the original image data, and divide the indoor area of the residents into multiple sub-regions according to the position of the explosion source and the explosion influence range; the multiple sub-regions include sub-regions of the first level and sub-regions of the second level; the second level is lower than the first level, and the explosion source is in the sub-region of the first level; Split the load contribution degree matrix into the load contribution degree matrix of each sub-region according to the sub-region; Perform preprocessing on the load contribution degree matrix of each sub-region to obtain the preprocessed load contribution degree matrix of each sub-region; For each sub-region, use a feature extraction method to extract features from the preprocessed load contribution degree matrix to obtain a feature matrix; Use different decoupling analysis methods to perform decoupling analysis on the feature matrices of sub-regions of different levels to obtain a decoupling intermediate matrix for each sub-region; Construct an explosion effect knowledge graph, and use the domain knowledge in the explosion effect knowledge graph to adjust the decoupling intermediate matrix of each sub-region respectively to obtain the adjusted decoupling intermediate matrix of each sub-region; Perform effect separation on the adjusted decoupling intermediate matrix of each sub-region to obtain a decoupling separation matrix for each region, and splice the decoupling separation matrices of all regions to obtain a decoupling effect matrix.

[0008] Preferably, the decoupling analysis of the feature matrices of sub-regions at different levels by using different decoupling analysis methods to obtain the decoupled intermediate matrix of each sub-region includes: Performing regression analysis on the feature matrix of the sub-region at the first level by using the principal component regression algorithm to obtain the first principal component regression result, performing decoupling analysis on the feature matrix of the sub-region at the first level by using the neural network model to obtain the first neural network decoupling result, performing ensemble learning processing on the feature matrix of the sub-region at the first level by using the ensemble learning method to obtain the first ensemble learning decoupling result; performing regression analysis on the feature matrix of the sub-region at the first level by using the partial least squares regression algorithm to obtain the first partial least squares regression result; weighting and averaging the first principal component regression result, the first neural network decoupling result, the first ensemble learning decoupling result, and the first partial least squares regression result to obtain the decoupled intermediate matrix of the sub-region at the first level; Performing regression analysis on the feature matrix of the sub-region at the second level by using the principal component regression algorithm to obtain the second principal component regression result, performing regression analysis on the feature matrix of the sub-region at the second level by using the partial least squares regression algorithm to obtain the second partial least squares regression result; weighting and averaging the first principal component regression result and the second partial least squares regression result to obtain the decoupled intermediate matrix of the sub-region at the second level.

[0009] Preferably, the determination of the position of the explosion source and the explosion influence range according to the original sensor data and the original image data includes: Performing time difference positioning processing on the original sensor data by using the sensor data positioning algorithm to obtain the first explosion source position; Performing video analysis, image segmentation, feature extraction, and target tracking processing on the original image data by using the image analysis algorithm to obtain the second explosion source position; Performing simulation analysis processing on the three-dimensional model of the residential indoor by using the numerical simulation method to obtain the third explosion source position and the initial influence range to be adjusted; Based on the first explosion source position and the second explosion source position, adjusting the third explosion source position and the initial influence range to obtain the position of the explosion source and the explosion influence range.

[0010] Preferably, the preprocessing of the load contribution matrix of each sub-region to obtain the preprocessed load contribution matrix of each sub-region includes: Performing noise and outlier removal processing on the load contribution matrix of each sub-region by using the data cleaning algorithm to obtain the cleaned data; Performing standardization processing on the cleaned data by using the standardization algorithm to make the scales of different features consistent to obtain the standardized data; The missing value filling algorithm is used to fill the missing values in the standardized data, and the load contribution matrix after preprocessing for each sub-region is obtained.

[0011] Preferably, the feature extraction method is used to extract features from the preprocessed load contribution matrix to obtain a feature matrix, including: The feature selection algorithm is used to perform feature selection processing on the preprocessed load contribution matrix to obtain a list of feature parameters; The feature extraction algorithm is used to extract the selected feature parameters from the list of feature parameters to obtain a parameter matrix; The feature normalization algorithm is used to normalize all the feature parameters in the parameter matrix to generate a feature matrix.

[0012] Preferably, the explosion effect knowledge graph is constructed, and the domain knowledge in the explosion effect knowledge graph is used to adjust the decoupling intermediate matrix of each sub-region respectively to obtain the adjusted decoupling intermediate matrix of each sub-region, including: The knowledge graph construction algorithm is used to construct an explosion effect knowledge graph; The knowledge fusion algorithm is used to fuse the domain knowledge in the explosion effect knowledge graph with the decoupling intermediate results of each sub-region to obtain the fused decoupling intermediate matrix of each sub-region; The result correction algorithm is used to correct the fused decoupling intermediate matrix of each sub-region by combining the historical data and prior knowledge in the explosion effect knowledge graph to obtain the adjusted decoupling intermediate matrix of each sub-region.

[0013] Preferably, based on the original sensor data and the original image data, multiple single-load algorithms are used to analyze different loads to obtain multiple single-load analysis results, including: The free-field pressure sensor data, temperature sensor data, fragment image data collected by a high-speed camera, velocity measurement target paper data, wall pressure sensor data, concentration sensor data, flame image data recorded by a high-speed camera, building image data collected by an ordinary camera, overhead video data taken by a global photography UAV, and local video data taken by a moving camera are respectively preprocessed; The shock wave propagation model algorithm is used to process the preprocessed free-field pressure sensor data to construct a shock wave propagation model, and the propagation path and intensity of the shock wave are obtained; The thermal radiation transfer algorithm is used to process the preprocessed temperature sensor data to establish a thermal radiation transfer model, and the flame temperature change, flame temperature distribution, and thermal radiation influence range during the explosion process are obtained; The fragment dynamics algorithm is used to process the pre - processed fragment image data and the pre - processed velocity - measuring target paper data to calculate the velocities and trajectories of the fragments of the glass on the interior walls of residential buildings and the exterior glass fragments, and obtain the motion laws of the fragments. The structural dynamic response algorithm is used to process the pre - processed wall pressure sensor data to evaluate the dynamic response of the building structure under blast loads and obtain the damage degree of the building structure. The gas concentration distribution algorithm is used to process the pre - processed concentration sensor data to analyze the changes in the concentration distributions of natural gas and liquefied petroleum gas indoors and obtain the gas leakage situation before the explosion. The flame propagation speed algorithm is used to process the pre - processed flame image data to analyze the propagation path and speed of the flame and obtain the influence of the flame on different regions. The building damage mode recognition algorithm is used to process the pre - processed building image data to identify the building damage mode. The global perspective analysis algorithm is used to process the pre - processed overhead video data to analyze the blast influence range and blast mode. The local detail analysis algorithm is used to process the pre - processed local video data to analyze the blast influence and damage details in the key areas.

[0014] Through the test method of this application, the blast effects in the large - space environment of residential interiors can be simulated and evaluated more accurately, especially the multi - load coupling effect. This embodiment can use the multi - load decoupling model constructed based on the multi - load correlation analysis algorithm, load contribution degree evaluation algorithm, and multi - load decoupling algorithm to obtain the multi - load decoupling analysis results, improving the test accuracy and reliability. Furthermore, it provides a scientific basis for building structure design and safety protection measures, effectively improving the safety of residential interiors.

[0015] In a second aspect, an embodiment of this application provides a multi - load decoupling test system for indoor explosions in residential buildings, including: a test platform serving as a residential room, various types of sensors and various types of image acquisition devices deployed around the test platform, and a processing device for executing the method described in any item of the first aspect. The processing device can be a data acquisition and synchronization control system.

[0016] Exemplarily, the multi - load decoupling test system for indoor explosions in residential buildings includes a test platform, a gas distribution system, an ignition system, a concentration sensor, a temperature sensor, multiple pressure sensors, an image acquisition system, a building fragment velocity acquisition system, and a data acquisition and synchronization control system; the multiple pressure sensors include wall pressure sensors and free - field pressure sensors. The test platform is a two-bedroom, two-living-room and one-bathroom structure with a regular rectangular shape, equipped with bedroom, kitchen, bathroom, and living room scenarios. The east-west length of the test platform is 14.15 m, and the north-south width is 8.3 m. The gas distribution system includes combustible gas cylinders, air cylinders, a circulating gas distributor, one-way valves, and gas pipelines. The combustible gas cylinders and air cylinders are connected to one end of the circulating gas distributor through gas pipelines, and the other end of the circulating gas distributor is connected to the one-way valve and injects combustible gas into the test platform through gas pipelines. The ignition system includes an igniter and an ignition rod. The igniter is an adjustable multi-functional igniter with adjustable ignition energy and ignition mode. The ignition rod is installed at the bottom of the kitchen switch, electrical appliances, the top of the kitchen, and the living room switch to restore the ignition positions in gas explosion accidents in residential buildings. The concentration sensor is installed on the mounting rod, and the mounting rod is located at the center of each indoor space to monitor the concentration distribution changes of natural gas and liquefied petroleum gas in the room. The temperature sensor is installed at the center of each indoor space and outside the doors and windows to collect the flame temperature changes during the explosion process. The pressure sensors include free-field pressure sensors and wall pressure sensors. The free-field pressure sensors are installed outside each door and window to collect the outdoor overpressure changes to obtain the propagation path and intensity of the shock wave. The wall pressure sensors are installed on the inner side of the roof and walls to collect the indoor overpressure changes to determine the damage degree of the building structure. The image acquisition system includes the following image acquisition devices: high-speed cameras, global imaging drones, and motion cameras. The high-speed cameras are installed outside the rooms. The global imaging drones are installed at the top of the rooms, hovering at a safe position directly above the rooms. The motion cameras are installed at the top corners of the living room and bedrooms. The building fragment velocity acquisition system includes velocity measurement target papers and high-speed cameras. The velocity measurement target papers are installed on the side where the indoor brick-concrete partition walls may collapse to test the fragment velocity of the indoor walls. The high-speed cameras are used to capture and calculate the fragment velocity of the outdoor glass. The data acquisition and synchronization control system includes a data acquisition instrument and a synchronization controller. The data acquisition instrument is used to store, record, and analyze the test data. The synchronization controller is used to coordinately control the multi-load decoupling test system for indoor explosions in residential buildings. The purpose of the embodiments of this application is to provide a multi-load decoupling test system for indoor explosions in residential buildings to solve the problems of low test accuracy and poor reliability in existing methods. The specific purposes include: realizing explosion tests on full-scale large indoor spaces in residential buildings, comprehensively obtaining various physical parameters during the explosion process; and performing decoupling analysis on the multi-loads during the explosion process to provide a scientific basis for the prevention and control of explosion accidents, thereby improving the accuracy and reliability of the test method, and reducing the test cost and risk.

[0017] Preferably, the total floor area of the test platform is 117.445 m 2 , and the usable area is 100.125 m 2 , and each room is provided with a window.

[0018] Preferably, 15 concentration sensors are provided, with 3 concentration sensors installed on each mounting rod, for a total of 5 mounting rods; the concentration sensors are respectively arranged at positions 0.5 m from the ground, in the middle position, and 0.1 m from the top, for real-time monitoring of the change in gas concentration in the vertical direction of the room (i.e., the change in the concentration of natural gas and liquefied petroleum gas), providing a basis for the control of explosion conditions; the accuracy of the concentration sensors is ≤±1% FS, and they have explosion-proof functions.

[0019] Preferably, the temperature sensors can be of a certain model, with a measuring range of 0~2300 °C and a response time of ≤2 ms; the temperature sensors are 1.4 m from the ground, and a total of 11 are provided to achieve temperature monitoring of different areas.

[0020] Preferably, the free-field pressure sensors are pressure sensors of a certain model, with a measuring range of 0.001 - 200 MPa and a response frequency of 50 - 500 kHz; a total of 22 free-field pressure sensors are provided, with 3 arranged outside the entrance door, in the kitchen, in the bathroom, and outside the secondary bedroom window each, and 5 arranged outside the living room and master bedroom window each, with a sensor spacing of 1 m.

[0021] Preferably, the wall pressure sensors are pressure sensors of a specified type, with a measuring range of 0.001 - 20 MPa; a total of 19 wall pressure sensors are provided, among which, 1 is arranged at the center position of the space top, 1 is set at a height of 1.4 m beside the door and window each, and 1 is set at a height of 2.1 m on the wall without a window each.

[0022] Preferably, 2 high-speed cameras are arranged outside the room for recording the explosion flame propagation speed and the building fragment speed.

[0023] Preferably, an ordinary camera is also included for recording the building damage condition before and after the explosion process.

[0024] Through the collaborative work of the above system, the embodiments of the present application can comprehensively and accurately obtain various physical parameters during the explosion process, and perform decoupling analysis on multiple loads during the explosion process, providing a scientific basis for the prevention and control of explosion accidents. At the same time, the embodiments of the present application adopt reasonable fixed installation and protection measures, reducing the risk and equipment damage rate during the test process, and improving the safety and sustainability of the test. Description of the Drawings

[0025] Figure 1 Structural schematic diagram of a full-size large indoor space test platform provided in this embodiment; Figure 2 Schematic diagram of the arrangement of concentration sensors provided in this embodiment; Figure 3 Schematic diagram of the arrangement of temperature sensors provided in this embodiment; Figure 4 Schematic diagram of the arrangement of free-field pressure sensors provided in this embodiment; Figure 5 Schematic diagram of the arrangement of wall pressure sensors provided in this embodiment; Figure 6 Schematic diagram of the arrangement of image acquisition devices provided in this embodiment; Figure 7 Schematic diagram of the arrangement of building fragment velocity acquisition devices provided in this embodiment; Figure 8 Structural schematic diagram of a multi-load decoupling test system for indoor explosion of residents provided in an embodiment of this application; Figure 9 Flow schematic diagram of a multi-load decoupling test method for indoor explosion of residents provided in an embodiment of this application.

[0026] Among them, 1 - test platform; 2 - velocity measurement target paper; 3 - wall pressure sensor; 4 - igniter; 5 - ignition rod; 6 - synchronization controller; 7 - data acquisition instrument; 8 - concentration sensor; 9 - temperature sensor; 10 - free-field pressure sensor; 11 - infrared thermometer; 12 - high-speed camera; 13 - ordinary camera; 14 - gas cylinder; 15 - circulating gas distributor; 16 - one-way valve; 17 - global imaging drone; 18 - moving camera. Specific implementation manners

[0027] Next, the method solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0028] A multi-load decoupling test system for indoor explosion of residents in an embodiment of this application includes: a test platform serving as a resident room, various types of sensors and various types of image acquisition devices deployed around the test platform, and a processing device for executing the multi-load decoupling test method for indoor explosion of residents.

[0029] Exemplarily, such as Figure 8As shown in the figure, an indoor explosion multi-load decoupling test system for residents provided by an embodiment of the present application includes: a test platform 1, a gas distribution system, an ignition system, a concentration sensor 8, a temperature sensor 9, a wall pressure sensor 3, a free field pressure sensor 10, an image acquisition system, a building fragment velocity acquisition system, and a data acquisition and synchronization control system.

[0030] The test platform 1 is a regular rectangular structure of a typical full-size residential house. Exemplarily, it can be a two-bedroom, two-living-room, one-bathroom house, with scenes such as bedrooms, kitchens, bathrooms, and living rooms, and is furnished with furniture and electrical appliances such as tables, chairs, sofas, and beds inside, truly restoring the real indoor environment, as Figure 1 shown. Exemplarily, the east-west length of the test platform can be 14.15 m, the north-south width is 8.3 m, the total building area is 117.445 m², and the floor area within the unit is 100.125 m². It is the first full-size experimental test platform for gas explosion of a typical house type structure not less than 100 m² in China. In this embodiment, each room is provided with a window for realizing and recording the gas explosion pressure relief process.

[0031] The gas distribution system is responsible for mixing combustible gas and air in a certain proportion and then transporting them into the test platform 1 to simulate the situation of gas leakage. The gas distribution system includes components such as gas cylinders 14, a circulating gas distributor 15, a check valve 16, and gas pipelines. Among them, the gas cylinders 14 can include combustible gas cylinders, air cylinders, etc.

[0032] The ignition system includes an igniter 4 and an ignition rod 5. The igniter 4 can adjust the ignition energy and ignition method. The ignition rod 5 is arranged at typical positions where gas explosion is triggered, such as at the kitchen switch, the bottom of electrical appliances, the top of the kitchen, and the living room switch, for restoring the ignition position (or the position of the explosion source) of a gas explosion accident in a residential house.

[0033] Exemplarily, the concentration sensor 8 is arranged on an installation rod, and the installation rod is located at the center position of each indoor space, as Figure 2 shown. 15 concentration sensors are set. 3 concentration sensors are installed on each installation rod, and there are a total of 5 installation rods. The concentration sensors are respectively arranged at positions 0.5 m from the ground, the middle position, and 0.1 m from the top, for real-time monitoring of the gas concentration change characteristics in the vertical direction of the room, providing a basis for the control of explosion conditions. The accuracy of the concentration sensor is ≤±1% FS, and it has an explosion-proof function.

[0034] The temperature sensor 9 is arranged at the center position of each indoor space and outside the doors and windows, as Figure 3 shown. The temperature sensor is a certain model sensor with a measurement range of 0~2300 °C and a response time of ≤2 ms; the temperature sensor is 1.4 m from the ground, and a total of 11 are set to realize temperature monitoring of different areas.

[0035] The pressure sensors include free-field pressure sensors 10 and wall pressure sensors 3. The free-field pressure sensors are installed outside each door and window. They are pressure sensors of a certain model, with a measuring range of 0.001 - 200 MPa and a response frequency of 50 - 500 kHz. As Figure 4 shown, a total of 22 are set, with 3 arranged outside the entrance door, in the kitchen, in the bathroom, and outside the secondary bedroom window respectively, and 5 arranged outside the living room and master bedroom window respectively. The distance between sensors is 1 m. The wall pressure sensors are pressure sensors of a specified type, with a measuring range of 0.001 - 20 MPa. As Figure 5 shown, a total of 19 are set, with 1 arranged at the center position of the space ceiling, 1 set at a height of 1.4 m beside the door and window respectively, and 1 set at a height of 2.1 m on the wall without a window respectively.

[0036] The image acquisition system includes a high-speed camera 12, a global imaging drone 17, a motion camera 18, and a normal camera 13, as Figure 6 shown. The high-speed camera is installed outside the room; the global imaging drone is installed at the top of the room, hovering at a safe position directly above the room; the motion camera is installed at the top corner of the living room and bedroom; the normal camera is used to record the damage condition of the building before and after the explosion process.

[0037] The building fragment velocity acquisition system includes a velocity measurement target paper 2 and a high-speed camera 12, as Figure 7 shown. The velocity measurement target paper is installed on the side where the indoor brick-concrete partition wall may collapse, for testing the velocity of indoor wall fragments; the high-speed camera is used to capture and calculate the velocity of outdoor glass fragments.

[0038] Optionally, the building fragment velocity acquisition system further includes an infrared thermometer 11, which can be used to measure and monitor the temperature change during and after the explosion process.

[0039] The data acquisition and synchronization control system includes a synchronization controller 6 and a data acquisition instrument 7. The data acquisition instrument is used to store, record, and analyze test data; the synchronization controller is used to cooperate in controlling the multi-load decoupling test system for indoor explosions of residents.

[0040] Through the collaborative work of the above system, this application can comprehensively and accurately obtain various physical parameters during the explosion process, decouple and analyze multiple loads during the explosion process, and provide a scientific basis for the prevention and control of explosion accidents. At the same time, this application adopts reasonable fixed installation and protection measures, reduces the risks and equipment damage rate during the testing process, and improves the safety and sustainability of the testing. Through the testing method of this application, the explosion effects in the large indoor space of residents can be more accurately simulated and evaluated, providing a scientific basis for building structure design, explosion accident prevention and control, and effectively improving the safety of the indoor environment of residents.

[0041] Figure 9 FIG. is a schematic flow chart of a method for decoupling multiple loads in an indoor explosion of residents provided by an embodiment of this application. As Figure 9 shown, the method includes: S91. Obtain the original sensor data collected by various types of sensors in the indoor explosion multi-load decoupling test system for residents and the original image data collected by various types of image acquisition devices.

[0042] Among them, the original sensor data is the unprocessed data collected by various types of sensors (such as pressure sensors, temperature sensors, etc.) in the indoor explosion multi-load decoupling test system for residents. The original image data is the unprocessed image or video data collected by various types of image acquisition devices (such as high-speed cameras, ordinary cameras, etc.) in the indoor explosion multi-load decoupling test system for residents.

[0043] S92. Based on the original sensor data and the original image data, use multiple single-load algorithms to analyze different loads to obtain multiple single-load analysis results.

[0044] Among them, the single-load algorithm is an algorithm for analyzing a single type of load (such as pressure, temperature, etc.). By analyzing a single type of load, the characteristics and effects of the load are extracted.

[0045] S93. Integrate all the single-load analysis results into an explosion effect data set, and use a multi-load decoupling model to process the explosion effect data set to obtain a multi-load decoupling analysis result; the multi-load decoupling model is constructed by using a multi-load correlation analysis algorithm, a load contribution degree evaluation algorithm, and a multi-load decoupling algorithm.

[0046] Among them, the explosion effect dataset is a comprehensive dataset formed by integrating all single-load analysis results. The multi-load decoupling model is a model used to process and analyze multi-load data, and is used to decompose complex multi-load effects into multiple independent load effects. The multi-load correlation analysis algorithm is used to construct a load correlation matrix to reveal the correlation between different loads. The load contribution degree evaluation algorithm is used to construct a load contribution degree matrix to quantify the influence of each load. The multi-load decoupling analysis results include the effects of each independent load, time-varying curves, etc.

[0047] S94. Compare the actual explosion test results with the multi-load decoupling analysis results, and determine the explosion multi-load decoupling effect according to the comparison results.

[0048] Among them, the actual explosion test results are the data and observation results obtained through actual explosion tests, and are used to compare with the multi-load decoupling analysis results to verify the accuracy of the multi-load decoupling model.

[0049] In this embodiment, by using a variety of sensors and image acquisition devices, different types of data are obtained, which helps to comprehensively understand the explosion process. By using the single-load algorithm to analyze each load separately, the characteristics and effects of each load can be extracted more precisely. Integrating all single-load analysis results into a comprehensive dataset facilitates comprehensive and integrated analysis. Through the multi-load decoupling model, complex multi-load effects are decomposed into multiple independent load effects, improving the clarity and accuracy of the analysis. Therefore, in this embodiment, by decomposing complex multi-load effects into multiple independent load effects, the role of each load can be identified and analyzed more clearly, the accuracy of the decoupling analysis can be improved, and further the accuracy and reliability of the test can be improved.

[0050] In some embodiments, S92. Based on the original sensor data and original image data, use multiple single-load algorithms to analyze different loads, and obtain multiple single-load analysis results, including: Step 921: Preprocess the free-field pressure sensor data, temperature sensor data, fragment image data collected by the high-speed camera, velocity measurement target paper data, wall pressure sensor data, concentration sensor data, flame image data recorded by the high-speed camera, building image data collected by the ordinary camera, overhead video data captured by the global camera drone, and local video data captured by the moving camera respectively. Step 922: Process the preprocessed free-field pressure sensor data using the shock wave propagation model algorithm to construct a shock wave propagation model and obtain the propagation path and intensity of the shock wave. Step 923: Process the preprocessed temperature sensor data using the thermal radiation transfer algorithm to establish a thermal radiation transfer model and obtain the flame temperature change, flame temperature distribution, and thermal radiation influence range during the explosion. Step 924: Process the preprocessed fragment image data and the preprocessed velocity measurement target paper data using the fragment dynamics algorithm to calculate the velocities and trajectories of the fragments of the indoor wall glass and outdoor glass of the residents and obtain the motion law of the fragments. Step 925: Process the preprocessed wall pressure sensor data using the structural dynamic response algorithm to evaluate the dynamic response of the building structure under the explosion load and obtain the damage degree of the building structure. Step 926: Process the preprocessed concentration sensor data using the gas concentration distribution algorithm to analyze the change in the indoor natural gas and liquefied petroleum gas concentration distribution and obtain the gas leakage situation before the explosion. Step 927: Process the preprocessed flame image data using the flame propagation speed algorithm to analyze the propagation path and speed of the flame and obtain the influence of the flame on different regions. Step 928: Process the preprocessed building image data using the building damage mode recognition algorithm to identify the building damage mode. Step 929: Process the preprocessed overhead video data using the global perspective analysis algorithm to analyze the explosion influence range and explosion mode. Step 930: Process the preprocessed local video data using the local detail analysis algorithm to analyze the explosion influence and damage details of the key area.

[0051] In this embodiment, by preprocessing and performing special analysis on the data of the free-field pressure sensor, temperature sensor, fragment image data collected by the high-speed camera, velocity measurement target paper data, wall pressure sensor data, concentration sensor data, flame image data recorded by the high-speed camera, building image data collected by the ordinary camera, overhead video data captured by the global camera drone, and local video data captured by the moving camera respectively, the characteristics and effects of each load can be comprehensively and meticulously extracted. This not only improves the accuracy and reliability of the data, but also provides a solid foundation for the subsequent multi-load decoupling analysis. Specifically, these steps can construct a shock wave propagation model, establish a heat radiation transfer model, calculate the motion law of fragments, evaluate the dynamic response of the building structure, analyze the gas concentration distribution, study the flame propagation speed, identify the building damage mode, analyze the explosion influence range and mode, and analyze the explosion influence and damage details in the key area, thereby comprehensively improving the accuracy and reliability of the explosion multi-load decoupling test.

[0052] In some embodiments, in S93, a multi-load decoupling model is used to process the explosion effect data set to obtain a multi-load decoupling analysis result; the multi-load decoupling model is constructed by using a multi-load correlation analysis algorithm, a load contribution degree evaluation algorithm, and a multi-load decoupling algorithm, and includes: Step 931: Use the multi-load correlation analysis algorithm to perform correlation analysis on the explosion effect data set to obtain a load correlation matrix.

[0053] Step 932: Use the load contribution degree evaluation algorithm to evaluate the load correlation matrix to obtain a load contribution degree matrix.

[0054] Step 933: Use the multi-load decoupling algorithm to process the load contribution degree matrix to obtain a decoupling effect matrix; wherein the multi-load decoupling algorithm performs different decoupling analyses on different sub-regions based on the position of the explosion source and the explosion influence range.

[0055] Step 934: Use the time series analysis algorithm to analyze the decoupling effect matrix to obtain a multi-load decoupling analysis result.

[0056] From the above process, it can be seen that in this embodiment, by using a multi-load correlation analysis algorithm, a load contribution degree evaluation algorithm, and a multi-load decoupling algorithm, the explosion effect data set is analyzed and processed layer by layer, which can comprehensively reveal the correlation between different loads, quantify the contribution degree of each load, and decompose the complex multi-load effect into multiple independent load effects. In particular, the multi-load decoupling algorithm performs differential decoupling analysis on different sub-regions based on the position and influence range of the explosion source, further improving the accuracy and reliability of the decoupling result. Finally, by using the time series analysis algorithm to analyze the decoupling effect matrix, a detailed multi-load decoupling analysis result is obtained.

[0057] In the above embodiments, the design of the multi-load decoupling algorithm has a great impact on improving the test accuracy. Therefore, in this embodiment, the design of the multi-load decoupling algorithm will be specifically described. Exemplarily, in step 933, the load contribution matrix is processed using the multi-load decoupling algorithm to obtain a decoupling effect matrix, including: Step a1: Determine the location and explosion influence range of the explosion source based on the original sensor data and original image data, and divide the indoor area of the residents into multiple sub-areas according to the location and explosion influence range of the explosion source; the multiple sub-areas include sub-areas of the first level and sub-areas of the second level; the second level is lower than the first level, and the explosion source is in the sub-area of the first level.

[0058] Step a2: Split the load contribution matrix into the load contribution matrix of each sub-area according to the sub-areas.

[0059] Step a3: Preprocess the load contribution matrix of each sub-area to obtain the preprocessed load contribution matrix of each sub-area.

[0060] Step a4: For each sub-area, use a feature extraction method to extract features from the preprocessed load contribution matrix to obtain a feature matrix.

[0061] Step a5: Use different decoupling analysis methods to perform decoupling analysis on the feature matrices of sub-areas of different levels to obtain the decoupling intermediate matrix of each sub-area.

[0062] Step a6: Construct an explosion effect knowledge graph, and use the domain knowledge in the explosion effect knowledge graph to adjust the decoupling intermediate matrix of each sub-area respectively to obtain the adjusted decoupling intermediate matrix of each sub-area.

[0063] Step a7: Perform effect separation on the adjusted decoupling intermediate matrix of each sub-area to obtain the decoupling separation matrix of each area, and splice the decoupling separation matrices of all areas to obtain a decoupling effect matrix.

[0064] As can be seen from the above description, in this embodiment, by determining the location and influence range of the explosion source based on the original sensor data and image data, and dividing the indoor area of the residents into sub-areas of different levels, it is possible to perform detailed decoupling analysis on each sub-area specifically. Specifically, through preprocessing, feature extraction, and processing with different decoupling analysis methods on the load contribution matrix of each sub-area, and combining with the domain knowledge in the explosion effect knowledge graph for adjustment, the precise decoupling of each sub-area is finally achieved. This method not only improves the accuracy and reliability of the decoupling result, but also can analyze the explosion effect of each sub-area in detail, improving the accuracy and reliability of the explosion multi-load decoupling test.

[0065] In the above embodiment, step a1, determining the location of the explosion source and the explosion influence range according to the original sensor data and the original image data, includes: step a11, performing time difference positioning processing on the original sensor data by using a sensor data positioning algorithm to obtain the first explosion source location. Step a12, performing video analysis, image segmentation, feature extraction, and target tracking processing on the original image data by using an image analysis algorithm to obtain the second explosion source location. Step a13, performing simulation analysis processing on the three-dimensional model of the residential indoor by using a numerical simulation method to obtain the third explosion source location and the initial influence range to be adjusted. Step a14, adjusting the third explosion source location and the initial influence range based on the first explosion source location and the second explosion source location to obtain the location of the explosion source and the explosion influence range.

[0066] In this embodiment, by using a sensor data positioning algorithm, an image analysis algorithm, and a numerical simulation method, the location of the explosion source and the initial influence range are determined from multiple perspectives such as time difference positioning, video analysis, and simulation analysis. Based on the preliminary results obtained by multiple methods, comprehensive adjustment is carried out to finally obtain the accurate location of the explosion source and the influence range. This method not only improves the accuracy and reliability of positioning, but also can provide more comprehensive and detailed information in a complex environment, providing support for subsequent multi-load decoupling analysis and explosion effect evaluation.

[0067] In the above embodiment, step a3, preprocessing the load contribution degree matrix of each sub-region to obtain the preprocessed load contribution degree matrix of each sub-region, includes: step a31, performing noise and outlier removal processing on the load contribution degree matrix of each sub-region by using a data cleaning algorithm to obtain the cleaned data; Step a32, performing standardization processing on the cleaned data by using a standardization algorithm to make the scales of different features consistent to obtain the standardized data; step a33, performing missing value filling processing on the standardized data by using a missing value filling algorithm to obtain the preprocessed load contribution degree matrix of each sub-region.

[0068] Therefore, in this embodiment, by using a data cleaning algorithm to remove noise and outliers, using a standardization algorithm to make the scales of different features consistent, and using a missing value filling algorithm to process missing data, the quality and consistency of the load contribution degree matrix of each sub-region can be significantly improved. This not only ensures the accuracy and reliability of subsequent analysis, but also provides a clean, complete, and standardized data basis for multi-load decoupling analysis, thus improving the effect and credibility of the entire decoupling process.

[0069] In the above embodiment, in step a4, a feature extraction method is used to extract features from the preprocessed load contribution degree matrix to obtain a feature matrix, including: Step a41: Use a feature selection algorithm to perform feature selection on the preprocessed load contribution matrix to obtain a list of feature parameters; Step a42: Use a feature extraction algorithm to extract the selected feature parameters from the list of feature parameters to obtain a parameter matrix; Step a43: Use a feature normalization algorithm to normalize all the feature parameters in the parameter matrix to generate a feature matrix.

[0070] In this embodiment, by using a feature selection algorithm to screen out important feature parameters, a feature extraction algorithm to extract the selected feature parameters, and a feature normalization algorithm to normalize the feature parameters, the quality and representativeness of the feature matrix can be significantly improved. It not only reduces data redundancy and noise but also ensures the comparability and consistency between different features, providing efficient, accurate, and reliable feature data for subsequent multi-load decoupling analysis, thereby enhancing the effect and credibility of the entire analysis process.

[0071] In the above embodiment, in Step a5: Use different decoupling analysis methods to perform decoupling analysis on the feature matrices of sub-regions at different levels to obtain a decoupling intermediate matrix for each sub-region, including: Step a51: Use the principal component regression algorithm to perform regression analysis on the feature matrix of the sub-region at the first level to obtain the first principal component regression result, use a neural network model to perform decoupling analysis on the feature matrix of the sub-region at the first level to obtain the first neural network decoupling result, use an ensemble learning method to perform ensemble learning on the feature matrix of the sub-region at the first level to obtain the first ensemble learning decoupling result; use the partial least squares regression algorithm to perform regression analysis on the feature matrix of the sub-region at the first level to obtain the first partial least squares regression result; weight and average the first principal component regression result, the first neural network decoupling result, the first ensemble learning decoupling result, and the first partial least squares regression result to obtain the decoupling intermediate matrix of the sub-region at the first level.

[0072] Optionally, the principal component regression algorithm can be sparse principal component analysis. The neural network model includes an attention mechanism and a basic model, where: ; ; ; where A is the attention weight matrix, representing the attention degree of each query vector to each key vector, is the query vector, which can refer to the features extracted from the feature matrix of the sub-region at the first level, is the key vector, also the features extracted from the feature matrix of the sub-region at the first level; represents the transpose of K; is the dimension of the key vector, is the square root of the dimension of the key vector, which is used to scale the dot product result to prevent gradient vanishing or explosion; is the value vector, which is usually the feature representation in the feature matrix of the first-level sub-region. is the decoupling result of the first neural network, is the attention output result; is the output result of the base model; is the weight coefficient of the attention output result, which is used to balance the contributions of the attention mechanism and the base model and can be adjusted according to the actual situation. The attention mechanism allows the neural network model to more flexibly capture the important parts of the input features when processing complex data, thereby improving the quality of feature representation.

[0073] Step a52: Use the principal component regression algorithm to perform regression analysis on the feature matrix of the second-level sub-region to obtain the second principal component regression result, and use the partial least squares regression algorithm to perform regression analysis on the feature matrix of the second-level sub-region to obtain the second partial least squares regression result; Weight and average the first principal component regression result and the second partial least squares regression result to obtain the decoupling intermediate matrix of the second-level sub-region.

[0074] In this embodiment, by using multiple algorithms such as principal component regression, neural network, ensemble learning, and partial least squares regression to perform detailed decoupling analysis on sub-regions of different levels and weighting and averaging multiple results, the accuracy and reliability of the decoupling result can be significantly improved. This method not only makes full use of the advantages of different algorithms, reduces the limitations of a single algorithm, but also can adopt differentiated analysis strategies for sub-regions of different levels, thereby providing more comprehensive and refined data support for multi-load decoupling analysis and enhancing the effect and credibility of the entire decoupling process.

[0075] In the above embodiment, step a6: Construct an explosion effect knowledge graph and use the domain knowledge in the explosion effect knowledge graph to adjust the decoupling intermediate matrix of each sub-region respectively to obtain the adjusted decoupling intermediate matrix of each sub-region, including: Step a61: Use the knowledge graph construction algorithm to construct an explosion effect knowledge graph. Step a62: Use the knowledge fusion algorithm to fuse the domain knowledge in the explosion effect knowledge graph with the decoupling intermediate results of each sub-region to obtain the fused decoupling intermediate matrix of each sub-region. Step a63: Use the result correction algorithm to correct the fused decoupling intermediate matrix of each sub-region in combination with the historical data and prior knowledge in the explosion effect knowledge graph to obtain the adjusted decoupling intermediate matrix of each sub-region.

[0076] In this embodiment, by constructing an explosion effect knowledge graph, using a knowledge fusion algorithm to fuse domain knowledge with the decoupled intermediate results, and then using a result correction algorithm to correct the results in combination with historical data and prior knowledge, the accuracy and reliability of the decoupling results can be significantly improved. This method not only makes full use of domain knowledge and historical data, enhances the interpretability and credibility of the model, but also can effectively correct potential errors, thus providing more scientific and accurate data support for multi-load decoupling analysis and improving the effect of the entire decoupling process.

[0077] The embodiment of the present application also provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it can implement the multi-load decoupling test method for indoor explosion of residents in the above Figure 9 illustrated embodiment.

[0078] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0079] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0080] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as a magnetic disk, an optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present application.

Claims

1. A method for decoupling indoor explosion loads in residents, characterized in that: include: Obtaining the original sensor data collected by various types of sensors in the residential indoor explosion multi-load decoupling test system and the original image data collected by various types of image acquisition devices; Based on the original sensor data and the original image data, multiple single load algorithms are used to analyze different loads to obtain multiple single load analysis results; Integrate all single load analysis results into an explosion effect data set, and use a multi-load decoupling model to process the explosion effect data set to obtain a multi-load decoupling analysis result; The multi-load decoupling model is constructed by using a multi-load correlation analysis algorithm, a load contribution evaluation algorithm and a multi-load decoupling algorithm; The actual explosion test results are compared with the multi-load decoupling analysis results, and the explosion multi-load decoupling effect is determined based on the comparison results.

2. A method for testing indoor explosion multi-load decoupling of residents according to claim 1, characterized in that: The explosion effect data set is processed using a multi-load decoupling model to obtain a multi-load decoupling analysis result; The multi-load decoupling model is constructed by using a multi-load correlation analysis algorithm, a load contribution evaluation algorithm and a multi-load decoupling algorithm, and includes: A multi-load correlation analysis algorithm is used to perform correlation analysis on the explosion effect data set to obtain a load correlation matrix; The load contribution evaluation algorithm is used to evaluate the load correlation matrix to obtain a load contribution matrix; The load contribution matrix is ​​processed by a multi-load decoupling algorithm to obtain a decoupling effect matrix; wherein the multi-load decoupling algorithm performs different decoupling analyses on different sub-areas based on the location of the explosion source and the explosion impact range; The time series analysis algorithm is used to analyze the decoupling effect matrix and obtain the multi-load decoupling analysis results.

3. A method for testing indoor explosion multi-load decoupling of residents according to claim 2, characterized in that: The multi-load decoupling algorithm is used to process the load contribution matrix to obtain a decoupling effect matrix, including: Determine the location of the explosion source and the explosion influence range according to the original sensor data and the original image data, and divide the indoor area of ​​the residents into a plurality of sub-areas according to the location of the explosion source and the explosion influence range; the plurality of sub-areas include a first-level sub-area and a second-level sub-area; the second level is lower than the first level, and the explosion source is in the first-level sub-area; Splitting the load contribution matrix into a load contribution matrix of each sub-region according to the sub-region; Preprocessing the load contribution matrix of each sub-region to obtain a preprocessed load contribution matrix of each sub-region; For each sub-region, a feature extraction method is used to extract features from the preprocessed load contribution matrix to obtain a feature matrix; Different decoupling analysis methods are used to perform decoupling analysis on the characteristic matrices of sub-regions of different levels, and the decoupling intermediate matrix of each sub-region is obtained; Constructing an explosion effect knowledge graph, and using the domain knowledge in the explosion effect knowledge graph to adjust the decoupling intermediate matrix of each sub-region respectively, to obtain an adjusted decoupling intermediate matrix of each sub-region; The adjusted decoupling intermediate matrix of each sub-region is subjected to effect separation to obtain a decoupling separation matrix of each region, and the decoupling separation matrices of all regions are spliced ​​to obtain a decoupling effect matrix.

4. A method for testing indoor explosion multi-load decoupling of residents according to claim 3, characterized in that: The decoupling analysis is performed on the feature matrices of sub-regions of different levels by using different decoupling analysis methods to obtain a decoupling intermediate matrix of each sub-region, including: Using a principal component regression algorithm to perform regression analysis on the feature matrix of the first-level sub-region to obtain a first principal component regression result, using a neural network model to perform decoupling analysis on the feature matrix of the first-level sub-region to obtain a first neural network decoupling result, using an ensemble learning method to perform ensemble learning processing on the feature matrix of the first-level sub-region to obtain a first ensemble learning decoupling result; using a partial least squares regression algorithm to perform regression analysis on the feature matrix of the first-level sub-region to obtain a first partial least squares regression result; taking a weighted average of the first principal component regression result, the first neural network decoupling result, the first ensemble learning decoupling result, and the first partial least squares regression result to obtain a decoupled intermediate matrix of the first-level sub-region; A principal component regression algorithm is used to perform regression analysis on the characteristic matrix of the second-level sub-region to obtain a second principal component regression result, and a partial least squares regression algorithm is used to perform regression analysis on the characteristic matrix of the second-level sub-region to obtain a second partial least squares regression result; the first principal component regression result and the second partial least squares regression result are weighted and averaged to obtain a decoupled intermediate matrix of the second-level sub-region.

5. A method for testing indoor explosion multi-load decoupling of residents according to claim 3, characterized in that: The step of determining the location of the explosion source and the explosion impact range according to the original sensor data and the original image data includes: Using a sensor data positioning algorithm to perform time difference positioning processing on the original sensor data to obtain the first explosion source position; Using an image analysis algorithm to perform video analysis, image segmentation, feature extraction and target tracking processing on the original image data to obtain the second explosion source position; The numerical simulation method is used to simulate and analyze the three-dimensional model of the residents' room to obtain the location of the third explosion source and the initial impact range to be adjusted; Based on the first explosion source position and the second explosion source position, the third explosion source position and the initial impact range are adjusted to obtain the explosion source position and the explosion impact range.

6. A method for testing indoor explosion multi-load decoupling of residents according to claim 3, characterized in that: The preprocessing of the load contribution matrix of each sub-region to obtain the preprocessed load contribution matrix of each sub-region includes: The data cleaning algorithm is used to remove noise and outliers from the load contribution matrix of each sub-area to obtain cleaned data; Using a standardization algorithm to standardize the cleaned data to make the scales of different features consistent, thereby obtaining standardized data; A missing value filling algorithm is used to fill in the missing values ​​of the standardized data to obtain a preprocessed load contribution matrix for each sub-area.

7. A method for testing indoor explosion multi-load decoupling of residents according to claim 3, characterized in that: The feature extraction method is used to extract features from the preprocessed load contribution matrix to obtain a feature matrix, including: Using a feature selection algorithm to perform feature selection processing on the preprocessed load contribution matrix to obtain a feature parameter list; extracting selected feature parameters from the feature parameter list using a feature extraction algorithm to obtain a parameter matrix; A feature normalization algorithm is used to normalize all feature parameters in the parameter matrix to generate a feature matrix.

8. A method for testing indoor explosion multi-load decoupling of residents according to claim 3, characterized in that: The step of constructing the explosion effect knowledge graph and adjusting the decoupling intermediate matrix of each sub-region respectively by using the domain knowledge in the explosion effect knowledge graph to obtain the adjusted decoupling intermediate matrix of each sub-region includes: Use the knowledge graph construction algorithm to build the explosion effect knowledge graph; The knowledge fusion algorithm is used to fuse the domain knowledge in the explosion effect knowledge graph with the decoupled intermediate results of each sub-region to obtain the fused decoupled intermediate matrix of each sub-region. The result correction algorithm is used to correct the fused decoupled intermediate matrix of each sub-region in combination with the historical data and prior knowledge in the explosion effect knowledge graph to obtain the adjusted decoupled intermediate matrix of each sub-region.

9. The method for testing indoor explosion multi-load decoupling of residents according to claim 1, characterized in that: Based on the original sensor data and the original image data, multiple single load algorithms are used to analyze different loads to obtain multiple single load analysis results, including: Preprocess the free-field pressure sensor data, temperature sensor data, fragment image data collected by high-speed cameras, speed target paper data, wall pressure sensor data, concentration sensor data, flame image data recorded by high-speed cameras, building image data collected by ordinary cameras, bird's-eye view video data taken by global camera drones, and local video data taken by motion cameras respectively; The pre-processed free-field pressure sensor data is processed using a shock wave propagation model algorithm to construct a shock wave propagation model and obtain the propagation path and intensity of the shock wave; The pre-processed temperature sensor data is processed using a thermal radiation transfer algorithm to establish a thermal radiation transfer model to obtain the flame temperature change, flame temperature distribution and thermal radiation influence range during the explosion process; The fragment dynamics algorithm is used to process the preprocessed fragment image data and the preprocessed velocity measurement target paper data to calculate the speed and trajectory of indoor wall glass fragments and outdoor glass fragments of residents, and obtain the movement law of the fragments; The pre-processed wall pressure sensor data is processed using a structural dynamic response algorithm to evaluate the dynamic response of the building structure under blast loads and obtain the degree of damage to the building structure. The pre-processed concentration sensor data is processed using a gas concentration distribution algorithm to analyze the changes in indoor natural gas and liquefied petroleum gas concentration distribution and obtain the gas leakage situation before the explosion. The flame propagation speed algorithm is used to process the preprocessed flame image data to analyze the flame propagation path and speed and obtain the impact of the flame on different areas; The pre-processed building image data is processed using a building damage pattern recognition algorithm to identify the building damage pattern; The pre-processed bird's-eye view video data is processed using a global perspective analysis algorithm to analyze the explosion impact range and explosion mode; The preprocessed local video data are processed using a local detail analysis algorithm to analyze the explosion impact and damage details in key areas.

10. A residential indoor explosion multi-load decoupling test system, characterized in that: include: A test platform for a residential room, various types of sensors and various types of image acquisition equipment deployed around the test platform, and a processing device for executing a residential room explosion multi-load decoupling test method as described in any one of claims 1 to 9.

Images