Method and system for evaluating spontaneous combustion risk of fire in limited space

Abstract

The invention provides a method and system for evaluating the spontaneous combustion risk of a fire in a confined space, and relates to the technical field of fire evaluation, and the method comprises the steps: constructing a ceiling impact fire model based on the confined space, the optimized discretization parameters of the ceiling impact fire model are determined by simulating and verifying free fire plume based on an open space in advance; based on the ceiling impact fire model, performing numerical simulation under various ventilation limitation degrees and various heat release rates to obtain a simulation result; based on the simulation result, extracting concentration distribution data and flue gas temperature distribution data of unburned fuel at the ceiling position; and comparing the concentration distribution data of the unburned fuel and the flue gas temperature distribution data with corresponding combustion thresholds, and evaluating the spontaneous combustion risk of the limited space according to a comparison result. According to the invention, the assessment of the fire spontaneous combustion risk can be accurately quantified, and the serious dependence on artificial experience is avoided.

Description

Technical Field

[0001] This invention relates to the field of fire assessment technology, and more specifically, to a method and system for assessing the risk of spontaneous combustion in confined spaces. Background Technology

[0002] The rapid pace of urbanization and the increasing complexity of building structures and industrial systems have greatly increased the challenges associated with fire safety, making it a critical and evolving research area. Fully identifying flame behavior, particularly the spread of flames impacting ceilings in various constrained configurations such as buildings, tunnels, pipes, and vehicles, is of paramount importance.

[0003] In related technologies, the assessment of such risks mainly relies on two approaches: one is to make judgments based on empirical formulas of simplified theories, and the other is to simulate based on regional models. However, both approaches have significant limitations, such as verification depending on the level of experience, poor accuracy, and inability to quantify risks. Summary of the Invention

[0004] The problem that this invention aims to solve is that related technologies for risk identification of fires in confined spaces suffer from at least one of the following issues: verification relies on experience, accuracy is poor, and risk cannot be quantified.

[0005] To address the aforementioned problems, in a first aspect, the present invention provides a method for assessing the spontaneous combustion risk of fires in confined spaces, comprising: A ceiling impact fire model based on confined space is constructed, wherein the optimized discretization parameters of the ceiling impact fire model are determined in advance by simulating and verifying free fire plumes based on open space; Based on the aforementioned ceiling impact fire model, numerical simulations were performed under various degrees of ventilation restriction and various heat release rates to obtain simulation results. Based on the simulation results, the concentration distribution data of unburned fuel and the flue gas temperature distribution data at the ceiling location were extracted. The concentration distribution data of the unburned fuel and the temperature distribution data of the flue gas are compared with the corresponding combustion thresholds, and the spontaneous combustion risk of the confined space is assessed based on the comparison results.

[0006] The method for assessing the spontaneous combustion risk of confined space fires provided by this invention completes simulation and verification in a free fire plume benchmark scenario using determined optimized discretization parameters. When constructing a more complex confined space ceiling impact fire model, the optimized discretization parameters are directly inherited and applied, fundamentally solving the problems of parameter setting relying on experience and wasting computational resources in numerical simulations while ensuring computational accuracy. Secondly, based on the confined space ceiling impact fire model, full-condition simulation is conducted by changing ventilation conditions and fire source power, revealing the coupling mechanism between the two on fire dynamics and risk formation. This overcomes the shortcomings of traditional methods that may miss key risk scenarios due to single operating conditions or reliance solely on experience. Then, data on unburned fuel concentration and flue gas temperature near the ceiling are specifically extracted, and the extracted data are compared with the corresponding combustion thresholds to quantitatively compare and assess the spontaneous combustion risk of confined spaces. This completely changes the fundamental limitation of traditional methods that rely on subjective experience and cannot quantify risk.

[0007] Secondly, embodiments of the present invention provide a system for assessing the risk of spontaneous combustion in confined spaces, employing the method for assessing the risk of spontaneous combustion in confined spaces as described in any of the preceding claims, including: An optimized construction module is used to: construct a ceiling impact fire model based on a confined space, wherein the optimized discretization parameters of the ceiling impact fire model are determined in advance by simulating and verifying a free fire plume based on an open space; The numerical simulation module is used to: perform numerical simulations based on the ceiling impact fire model under various degrees of ventilation restriction and various heat release rates to obtain simulation results; The data extraction module is used to: extract the concentration distribution data of unburned fuel and the flue gas temperature distribution data at the ceiling location based on the simulation results; The spontaneous combustion risk assessment module is used to: compare the concentration distribution data of the unburned fuel and the flue gas temperature distribution data with the corresponding combustion thresholds, and assess the spontaneous combustion risk of the confined space based on the comparison results.

[0008] Thirdly, the present invention provides an electronic device, including a memory and a processor; The memory is used to store computer programs; The processor is configured to, when executing the computer program, implement the method for assessing the risk of spontaneous combustion in confined space fires as described in the first aspect.

[0009] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for assessing the risk of spontaneous combustion in confined spaces as described in the first aspect.

[0010] The confined space fire spontaneous combustion risk assessment system, electronic device, and computer-readable storage medium provided by this invention have the same beneficial effects as the confined space fire spontaneous combustion risk assessment method compared to the prior art, and will not be repeated here. Attached Figure Description

[0011] Figure 1 A flowchart illustrating a method for assessing the risk of spontaneous combustion in a confined space fire, as described in an embodiment of the present invention, is shown. Figure 2 This diagram illustrates the effect of mesh size on temperature and velocity along the centerline in an embodiment of the present invention. Figure 3 This diagram illustrates the influence of the computational domain on temperature and velocity along the centerline in an embodiment of the present invention. Figure 4 This diagram illustrates a comparison of dimensionless flame heights obtained by FDS, experiments, and Heskestad correlation in an embodiment of the present invention. Figure 5 A schematic diagram illustrating the calculation settings for open environments and confined cabins in an embodiment of the present invention is shown; Figure 6 The diagram shows the centerline temperature distribution of the constrained configuration under different HRRs in an embodiment of the present invention. Figure 7 A schematic diagram of the instantaneous temperature field under different configurations when the HRR is 4.6 kW in an embodiment of the present invention is shown; Figure 8 This diagram illustrates the change of actual HRR over time in a constrained configuration according to an embodiment of the present invention. Figure 9 The figure shows the centerline oxygen concentration variation curves of four confined configurations under different HRRs in embodiments of the present invention; Figure 10 A schematic diagram of the instantaneous oxygen concentration field when the HRR is 4.6 kW is shown in an embodiment of the present invention; Figure 11 The figure shows the CO2 concentration variation curves of the centerline under different HRR for four confined configurations in embodiments of the present invention; Figure 12 The instantaneous CO2 concentration field at an HRR of 4.6 kW in an embodiment of the present invention is shown; Figure 13 The figure shows the propane concentration variation curves at the centerline for four restricted configurations under different HRRs in embodiments of the present invention. Figure 14 A schematic diagram of the structure of a confined space fire spontaneous combustion risk assessment system is shown in an embodiment of the present invention. Detailed Implementation

[0012] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0013] It should be noted that relational terms such as "first" and "second" in this invention are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0014] In the description of this specification, references to terms such as "embodiment," "one embodiment," and "one implementation" indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or implementation is included in at least one embodiment or illustrative implementation of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or implementation. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or implementations.

[0015] Reference Figure 1 As shown in the figure, this embodiment of the invention proposes a method for assessing the risk of spontaneous combustion in confined spaces; The methods for assessing the spontaneous combustion risk of confined space fires include: A ceiling impact fire model based on confined space is constructed, wherein the optimized discretization parameters of the ceiling impact fire model are determined in advance by simulating and verifying free fire plumes based on open space.

[0016] Specifically, a computational domain that accurately reflects the geometric characteristics of a confined space (such as a 1:10 scale student cabin, including walls, ceilings, and doors and windows) can be established based on the open-source computational framework Fire Dynamics Simulator (FDS). Then, according to the five ventilation-constrained configurations designed in the research, corresponding open or closed boundary conditions are assigned to the doors and windows, and actual material properties are assigned to the walls. Finally, a propane burner ignition source was set at a predetermined location (e.g., directly below the ceiling at the center of the floor), and different levels of heat release rates were input according to the experimental scheme. This completed a ceiling impact fire calculation model that both inherited the verified numerical reliability and fully contained all key elements of the target physical scenario. Among these, optimizing the discretization parameters is an important setting parameter before simulation. For example, the mesh size is not arbitrarily set, but determined based on a systematic mesh sensitivity analysis of the benchmark scenario of a free fire plume in an open space. In this analysis, by simulating a series of meshes with different coarsenesses (e.g., from 8 mm to 2.5 mm), it was observed that when the mesh was fined to about 3 mm, the changes in key physical quantities (such as flame temperature and velocity) tended to stabilize. Further refining the mesh did not significantly improve the accuracy. This indicates that 3 mm reached the best balance point between accuracy and computational cost, i.e., "optimized mesh size". Subsequently, when constructing a more complex confined space ceiling impact fire model, this verified 3 mm mesh size was directly inherited and applied. This approach ensures that complex models can fully resolve the same basic physical processes (such as turbulence, combustion, and heat transfer) as the baseline scenario, while avoiding result distortion or waste of computational resources due to improper mesh settings.

[0017] Based on the aforementioned ceiling impact fire model, numerical simulations were conducted under various degrees of ventilation restriction and various heat release rates to obtain simulation results.

[0018] Based on the established, parameter-reliable ceiling impact fire model, two key variables were mainly changed: the degree of ventilation restriction (e.g., setting five configurations with increasing ventilation from open to completely closed, Conf. 1-5) and the fire source power (i.e., heat release rate HRR, representing the heat released by the fire per unit time, such as 0.5 kW, 4.6 kW, and 18.6 kW). By conducting numerical simulations under various combinations of these two variables, a comprehensive dataset of simulation results can be obtained. This dataset records in detail the dynamic evolution of physical fields such as temperature, flow velocity, and concentration of key gas components within the confined space under each condition. These data results can systematically reveal how ventilation conditions and fire source intensity jointly affect flame behavior, flow field structure, gas distribution, and temperature field within the confined space.

[0019] Based on the simulation results, the concentration distribution data of unburned fuel and the flue gas temperature distribution data at the ceiling location were extracted.

[0020] Specifically, in the absence of a ceiling, if a fire occurs, the flames are free-rising buoyant plumes. The physical obstruction of the ceiling forces the vertically rising flames to collide, changing the flame shape from vertical to horizontal spread, forming a complex impact zone. After the impact, the high-temperature flames and smoke spread horizontally along the ceiling, forming a rapidly moving layer of high-temperature smoke (i.e., the ceiling jet). In reality, building fires (such as room, corridor, and tunnel fires) almost always occur in confined spaces with ceilings. Therefore, studying impact flames with ceilings has more direct engineering guidance significance. From the aforementioned simulation results, we focused on and extracted two core parameters near the ceiling: first, the concentration distribution data of unburned fuel (such as propane); and second, the temperature distribution data of flue gas. In the FDS simulation model, detectors or slices for recording, such as "propane mass fraction" and "gas temperature", were set up and located directly near the ceiling. Thus, from the multiple physical field variables output by the simulation (such as velocity, pressure, concentration of each component, radiation flux, etc.), data directly related to these two essential conditions were selected for analysis.

[0021] The concentration distribution data of the unburned fuel and the temperature distribution data of the flue gas are compared with the corresponding combustion thresholds, and the spontaneous combustion risk of the confined space is assessed based on the comparison results.

[0022] Specifically, for each "ventilation configuration-heat release rate" combination, simulation results are used to extract data on unburned fuel concentration and flue gas temperature distribution in key areas near the ceiling. For example, the concentration data is compared with the lower flammability limit (LFL) of the fuel, and the temperature data is compared with the auto-ignition temperature (AIT) to determine whether each operating condition simultaneously meets the dual conditions of "concentration exceeding limits" and "temperature exceeding limits." Finally, by comprehensively analyzing the comparison results of all operating conditions, confined spaces (corresponding to the corresponding configurations) with auto-ignition risk are identified.

[0023] In practical application, this embodiment completes simulation and verification in a free fire plume benchmark scenario using determined optimized discretization parameters. When constructing a more complex confined space ceiling impact fire model, the optimized discretization parameters are directly inherited and applied. This fundamentally solves the problems of parameter setting relying on experience and wasting computational resources in numerical simulation, while ensuring computational accuracy. Secondly, based on the confined space ceiling impact fire model, full-condition simulation is conducted by changing ventilation conditions and fire source power, revealing the coupling mechanism between the two on fire dynamics and risk formation. This overcomes the shortcomings of traditional methods that may miss key risk scenarios due to single operating conditions or reliance solely on experience. Then, data on unburned fuel concentration and flue gas temperature near the ceiling are specifically extracted, and the extracted data are compared with the corresponding combustion thresholds to quantitatively compare and assess the spontaneous combustion risk of confined spaces. This completely changes the fundamental limitation of traditional methods that rely on subjective experience and cannot quantify risk.

[0024] Studies have shown that the highest risk does not occur in configurations with excellent or poor ventilation, but rather is often concentrated in "semi-restricted" configurations with doors and windows partially open. In such configurations, limited ventilation is insufficient to achieve complete combustion (leading to the accumulation of combustible gases under the ceiling) and does not completely prevent heat loss (maintaining a high-temperature environment). This invention provides a direct and profound scientific basis for prioritizing the identification and control of such high-risk "semi-open and semi-closed" space configurations in building fire protection design.

[0025] As an optional embodiment of the present invention, the optimized discretization parameters include the grid size for spatial discretization and / or the computational domain size for defining the computational boundary, and the prior simulation and verification of free fire plumes based on open space includes: A baseline model of free fire plumes based on the open space is constructed.

[0026] Specifically, in the open-space fire plume simulation, open boundary conditions were applied to all domain boundaries except the floor; the ambient temperature was set to 20°C; for the simulation experiment, the fire source was modeled as propane being injected at a uniform velocity through a circular vent with a diameter of 5 mm placed at the center of a solid wall; a Cartesian grid was used in the FDS, therefore the circular burner was approximated as a "stepped" structure; the heat release rate (HRR) varied from 0.5 kW to 18.6 kW; the required HRR was generated by specifying the total mass flow rate per unit area through the burner's MASS_FLUX parameter; therefore, the HRR in the simulation was determined by the product of the mass flow rate per unit area, the burner area, and the heat of combustion of the fuel; the heat of combustion of propane was equal to 46.45 x 10⁻⁶. 3 kJ / kg; all simulation cases run for 10 seconds, with the last 5 seconds used for steady-state averaging.

[0027] Mesh sensitivity analysis and computational domain independence verification are performed on the baseline model to determine the optimized discretization parameters that enable the simulation results of the baseline model to achieve mesh independence and eliminate boundary effects.

[0028] Specifically, during the simulation, the continuous space is divided into grids (pixels). The size of the grid directly affects the calculation accuracy. If the grid is too coarse, the results will be inaccurate; if the grid is infinitely fine, the computational cost will be infinitely high. By selecting an appropriate grid size (e.g., from coarse to fine), the changes in the calculation results (such as flame temperature and height) are observed to be small enough to be acceptable. Secondly, the interference of the simulation "boundary" on the results is eliminated. The computational domain is an artificially defined virtual "area" in which the fire is simulated. If the box is too small, the flame or smoke will quickly hit the boundary, producing unrealistic reflections or limitations. This is the "boundary effect." The goal of this verification is to find a sufficiently large computational domain size so that the simulation results of the core area we are concerned with (such as the flame body) no longer change due to the size of the domain (i.e., "eliminating the boundary effect"). Finally, a set of verified optimal numerical parameters is determined, namely, an optimal grid size that ensures accuracy while controlling computational cost, and an optimal computational domain size that eliminates boundary interference without excessively wasting computational resources.

[0029] In practical applications, this embodiment uses mesh sensitivity analysis to determine an optimal mesh size, ensuring that the calculation results do not change significantly with further mesh refinement, thus achieving "mesh independence" and guaranteeing that the accuracy of the results is not affected by the degree of artificial discretization. Simultaneously, computational domain independence verification determines a sufficiently large simulation region to eliminate the interference of artificially set computational boundaries on core physical phenomena (such as flames and flow), i.e., eliminating "boundary effects." This ultimately ensures the accuracy of subsequent simulations and avoids the problem of excessively large computational domains and insufficient boundary effects leading to high computational costs.

[0030] As an optional embodiment of the present invention, the step of performing grid sensitivity analysis and computational domain independence verification on the benchmark model to determine the optimized discretization parameters that enable the simulation results of the benchmark model to achieve grid independence and eliminate boundary effects includes: The characteristic fire diameter is determined based on the minimum heat release rate corresponding to the benchmark model, and the theoretically feasible range of the grid size is determined based on the characteristic fire diameter and the preset grid resolution ratio range.

[0031] Specifically, FDS provides a widely used characteristic fire diameter that takes into account the effects of fire source size, burner diameter, etc. The reference range for the mesh size in the simulation can be obtained using the following formula:

[0032] Where D* represents the characteristic fire diameter. It is the heat release rate (HRR) of the heat source, c p It is specific heat capacity. and The density and temperature are under environmental conditions, and g is the acceleration due to gravity. Reference range for grid size.

[0033] Within the theoretically feasible range, by implementing comparative simulations of multi-level grid resolution and multi-size computational domains, the grid size that achieves grid independence and the computational domain size that eliminates boundary effects are determined based on the convergence characteristics of the temperature and velocity distribution along the flame centerline.

[0034] Specifically, eight different heat release rates (HRRs) were considered in each series of experiments: HRR = {0.5, 1.2, 2.3, 4.6, 9.3, 14.0, 16.3, 18.6} kW, corresponding to fuel mass flow rates of {0.01, 0.025, 0.05, 0.1, 0.2, 0.3, 0.35, 0.4} g / s. To maintain consistency with the experimental dataset, these HRRs were also implemented in the numerical simulations. Considering a minimum HRR of 0.5 kW, the estimated grid sizes calculated using Equations (1) and (2) ranged from 0.7 mm to 1 cm.

[0035] Therefore, six different grid resolutions—8 mm, 6 mm, 5 mm, 4 mm, 3 mm, and 2.5 mm—were systematically employed to evaluate the impact of grid sensitivity on the simulation results.

[0036] To balance computational efficiency and numerical accuracy, a uniform and non-uniform mesh distribution was used. The computational domain size was 0.3 × 0.3 × 0.25 m, defined in a rectangular Cartesian coordinate system with the origin at the center of the fire source. Considering the burner radius of 2.5 mm, the mesh size near the burner was limited to ≤2.5 mm. Therefore, a non-uniform mesh was applied, and a refined region with a 0.25 mm mesh spacing was achieved in a 0.06 m × 0.06 m subdomain centered on the burner using the TRNX and TRNY mesh stretching functions in FDS.

[0037] Table 1 summarizes the grid configurations.

[0038] Table 1

[0039] The criteria for determining the mesh size to achieve mesh independence and the computational domain size to eliminate boundary effects are the "convergence characteristics of the temperature and velocity distribution along the flame centerline." This directly corresponds to the disclosure. Figure 1 (Grid effect) and Figure 3The analysis of (computational domain influence) shows that when the curves tend to coincide and the changes are negligible as the parameters change (the mesh becomes finer or the computational domain becomes larger), it is considered that "mesh independence" and "elimination of boundary effects" have been achieved.

[0040] Figure 2 The effects of mesh resolution on centerline temperature and velocity were explained. It was observed that the peak centerline temperature increases with increasing mesh size, and at a given height on the centerline, the temperature also increases with increasing mesh size. The temperature distribution between M5 and M6 meshes tends to converge, indicating that further mesh refinement does not significantly alter the results. Similarly, the centerline velocity exhibits a similar trend with mesh refinement. Based on these sensitivity analyses, the M5 mesh (3 mm mesh size) was chosen for subsequent simulations because it provides mesh-independent results while optimizing computational cost.

[0041] In numerical simulation, the computational domain plays a crucial role in the accuracy of the simulation results. To select a suitable computational domain, three numerical simulations were conducted in three computational domains of different sizes, using the maximum HRR of 18.6 kW as the input firepower, as shown in Table 2. The sizes of the three computational domains are 0.1 m × 0.1 m × 1.2 m, 0.3 m × 0.3 m × 1.2 m, and 0.5 m × 0.1 m × 1.2 m, respectively.

[0042] Table 2

[0043] Considering fire plumes in small domains, the fire plumes are significantly constrained by the computational domain boundaries, thus necessitating an increase in the computational domain size. Considering fire plumes in medium and large domains, the size of the computational domain has little impact on the flame shape.

[0044] Figure 3 The effects of the computational domain on centerline (a) temperature and (b) velocity are described. Results show that the fire plume centerline velocities predicted by the large-domain and medium-domain simulations converge, indicating that simulations using the medium-domain computational domain can provide similar results compared to simulations using the large-domain domain. For centerline temperature, the large-domain simulation predicts slightly different results from the medium-domain simulation. This can be explained by the time-averaging effect; although the fire reaches an approximate steady state within 2 seconds, it fluctuates over time during the 10-second simulation. Therefore, the centerline temperature may differ depending on the time-averaging method. Furthermore, the large-domain simulation runs three times longer than the medium-domain simulation. Since the simulation results were similar, the medium-domain computational domain was used in subsequent work.

[0045] The determined grid size and / or computational domain size are applied to the baseline model, and the final verification is completed by comparing the degree of agreement between the simulated flame height and the theoretical prediction data, thus obtaining the optimized discretization parameters.

[0046] Specifically, in this invention, a temperature threshold of 520°C is used to define the flame height; that is, the flame height corresponds to the region where the temperature exceeds this value. Figure 4 The paper presents a comparison between the dimensionless flame height obtained by FDS and experimental data and Heskestad (empirical correlation).

[0047] FDS predictions fit the Heskestad correlation very well at lower heat release rates (HRR). However, as HRR increases, FDS slightly overestimates the flame height, indicating potential limitations of the model under more intense fire conditions. Despite this difference at higher HRRs, the overall consistency is strong, with the predicted flame height fitting the correlation very well and the relative gap (RG) being only 3.84%.

[0048] Contrary to the numerical results, the experimental data showed a large overall discrepancy with the Heskestad correlation, with a relative difference (RG) of 13.45%. Detailed examination revealed that this discrepancy was primarily driven by low HRR conditions, under which the measured flame height was consistently higher than theoretically predicted. However, at higher HRRs, the consistency was better. The systematic overestimation at low HRRs may be attributed to a slight underestimation of the fuel mass flow rate during the experiment, possibly due to the limited accuracy of the flowmeter used.

[0049] In practical application, this embodiment first uses the characteristic fire diameter formula to accurately define the parameter range, then selects the optimal value of the optimized discretization parameter based on data convergence through multi-level comparative simulation, and finally verifies the accuracy of the result using classical empirical formulas. This ensures that the numerical foundation of all subsequent complex fire simulations (especially ceiling-impact flame simulations) based on the optimized discretization parameter is solid and reliable, thus guaranteeing the reliability of the risk assessment results from the source.

[0050] As an optional embodiment of the present invention, the construction of the ceiling impact fire model based on confined space includes: Establish a geometry comprising a confined compartment and its extended domain, wherein the confined compartment is configured with material parameters, boundary conditions, and a burner.

[0051] Specifically, to study ceiling impact flames in open and confined compartments, numerical simulations were performed in a cuboid computational domain measuring 0.6 m × 0.8 m × 0.3 m. Due to the presence of the ceiling, non-uniform meshes could not be used; therefore, a uniform mesh of 2.5 mm (i.e., δx = δy = δz) was used in all dimensions, totaling 9.216 million computational cells. Previous mesh sensitivity analysis showed that this resolution provided sufficiently fine discretization to accurately capture key flow and thermal features. An additional computational area is added outside the ceiling or compartment model to allow the flue gas flowing out of the opening to have room to develop, avoid boundary reflections, and improve the simulation realism and numerical stability.

[0052] To maintain consistency with experimental work, five configurations were also examined, such as Figure 5 As shown, the first configuration, Conf.1, refers to the situation where the flame impacts the ceiling in an open environment, with a height of 0.24m between the burner and the ceiling. The last four configurations, from Conf.2 to Conf.5, refer to the situation where the flame impacts the ceiling in confined compartments with different degrees of confinement. More precisely, the degree of confinement for each configuration is determined by the closing / opening conditions of the windows and doors in the compartment, so the degree of confinement increases from Conf.2 to Conf.5.

[0053] In one example, the cabin geometry is based on a 1:10 scale model of a typical French student cabin. The detailed dimensions of the small-scale cabin are 0.6 m (length) × 0.4 m (width) × 0.24 m (height); the window dimensions are 0.15 m (height) × 0.15 m (width), while the door dimensions are 0.2 m (height) × 0.09 m (width); cabin wall material properties: side walls are concrete with a density of 2200. Thermal conductivity 1.2 Specific heat capacity 0.88 The roof is made of steel with a density of 7850. Thermal conductivity 46 Specific heat capacity 0.5 Emission rate 0.9.

[0054] Burner: refers to the ignition source, which is configured as a circular propane injection port with a diameter of 5 mm, located at the center of the floor and 0.24 m directly below the ceiling. The heat release rate (HRR) is controlled by setting MASS_FLUX (mass flow rate).

[0055] Based on the geometric configuration, a dynamic solver for turbulence-combustion-radiation coupling is configured, and the optimized discretization parameters are applied to complete the model discretization, thereby obtaining the ceiling impact fire model based on confined space.

[0056] Specifically, the dynamics solver as a whole adopts the Navier-Stokes solver based on the low Mach number approximation, uses large eddy simulation (LES), and employs the wall model of Nicoud and Ducros. It adopts the concept of eddy dissipation (EDC) and enables a simple flameout model based on a critical flame temperature of 1427℃ for propane burners. The radiation model adopts the gray gas radiation model, combined with the finite volume method (FVM) to solve the radiation transfer equation, and sets the lower limit of radiation fraction to 0.35 to control uncertainty.

[0057] The mesh and computational domain parameters determined by the open space plume verification are applied to the ceiling impact fire model in a confined space. That is, the geometric model is discretized directly using the optimized mesh configuration and computational domain settings to generate a corresponding number of computational units, thereby completing the transformation from a continuous physical system to a discrete numerical model.

[0058] In practical application, this embodiment firstly ensures the objectivity and reliability of the numerical basis by applying the optimized discretization parameters, which have been verified reliably in open space, to the complex constrained model, and effectively saves costs. Secondly, by introducing an extended domain and configuring real materials and boundary conditions, a highly realistic and numerically stable refined geometric-physical model is constructed. Finally, by programmatically controlling ventilation and fire source parameters for system simulation, a targeted dataset covering a wide range of hazardous scenarios is efficiently generated, laying the foundation for subsequent system revelation of the coupling law between configuration and heat release rate, as well as for locating the highest risk scenario (configuration).

[0059] As an optional embodiment of the present invention, the numerical simulation based on the ceiling impact fire model under various degrees of ventilation restriction and various heat release rates to obtain simulation results includes: The opening and closing states of ventilation openings in the confined compartment of the ceiling impact fire model are controlled to determine multiple configurations that correspond one-to-one with the degree of ventilation confinement. For each configuration, a heat release rate with multiple different levels is set, and numerical simulations are performed on all defined parameterized scenarios until a quasi-steady state is reached.

[0060] Specifically, the degree of geometric confinement is altered by controlling the "opening" and "closing" of the boundary conditions of doors and windows (ventilation openings) in the model. A series of configurations from Conf. 1 (open) to Conf. 5 (completely closed) are defined by the opening and closing of doors and windows. The first variable is the degree of ventilation confinement (configuration). This is the core control variable of this study, used to investigate the "confinement effect." The second variable is the heat release rate (HRR). For each fixed ventilation configuration, the power of the fire source is systematically changed. This corresponds to the setting of using six HRRs in the open configuration (Conf. 1) and three representative HRRs (0.5, 4.6, 18.6 kW) in the confined compartment configurations (Conf. 2-5) as described in the briefing. The simulation time for the open environment is 10 seconds (the fire stabilizes after about 2 seconds), and the simulation time for the confined compartment is at least 30 seconds to ensure that the flow field and combustion in the compartment reach dynamic equilibrium (quasi-steady state).

[0061] Configure and output the simulation results, wherein the simulation results include the flow state parameters at the ceiling location and its neighborhood location, the concentration distribution data, and the relevant heat flux data.

[0062] Specifically, this involves setting the parameters in an input file within software like FDS before running the simulation. This isn't post-processing; rather, it's a pre-defined process that requires recording certain data. Flow state parameters, including velocity and pressure, are fundamental to understanding flow field structure, entrainment, and mixing, and help explain the causes of temperature and concentration distribution.

[0063] Concentration distribution data: Specifically refers to the concentration fields of key gaseous components such as O2, CO, and propane (C3H8). This is the direct basis for assessing the accumulation of unburned fuel (propane) and the combustion state (O2, CO).

[0064] Relevant heat flux data include convective heat flux, radiative heat flux, and net heat flux. These data directly reflect the intensity of the thermal shock to the ceiling and surrounding structure caused by a fire, and are important parameters for assessing structural response and fire spread risk.

[0065] In practical applications, this embodiment generates a series of "virtual experimental" conditions covering different ventilation conditions and fire intensities by programmatically controlling the model opening and fire source power. These conditions are then fully calculated until they stabilize. The generated data provides a complete raw database for the subsequent extraction of "temperature distribution data" and "concentration distribution data".

[0066] As an optional embodiment of the present invention, the extraction of unburned fuel concentration distribution data and flue gas temperature distribution data at the ceiling location based on the simulation results includes: From the simulation results, read the original time-series monitoring data at the monitoring locations pre-configured for monitoring the ceiling location.

[0067] Specifically, according to the original result file output after the CFD simulation, the data in the original result file is monitoring data that was pre-configured before the simulation and is specifically used to record the physical state of the ceiling area. It is a continuous record that changes over time and contains all the instantaneous pulse information during the fire development process. For the monitoring locations in the simulation, in addition to the horizontal and vertical devices placed above and below the ceiling, temperature, velocity, pressure, concentration and mixing fraction of O2, CO, and propane are measured by setting up slice fields in the middle of the compartment and below the ceiling. In order to facilitate detailed visualization of flow and thermal patterns, boundary file (BNDF) groups are configured to record convective heat flux, radiative heat flux, net heat flux, combustion rate and wall temperature distribution.

[0068] The original time-series monitoring data is averaged over time to obtain the flue gas temperature parameters and unburned fuel concentration parameters corresponding to each monitoring location.

[0069] Specifically, by performing time averaging on the above-mentioned raw monitoring data, random high-frequency fluctuations caused by turbulence are effectively filtered out, thereby extracting representative physical quantities of each monitoring location under quasi-steady-state combustion conditions, namely a statistically significant steady-state value of flue gas temperature and a steady-state value of unburned fuel concentration.

[0070] Based on the spatial coordinates of each monitoring location, the obtained flue gas temperature parameters are reconstructed into flue gas temperature distribution data at the ceiling location, and the obtained unburned fuel concentration parameters are reconstructed into unburned fuel concentration distribution data at the ceiling location.

[0071] Specifically, based on the spatial coordinate information of each monitoring point, the discrete temperature and concentration steady-state values ​​obtained in the previous step are spatially reconstructed (e.g., interpolated) to generate continuous and visualized flue gas temperature distribution data and unburned fuel concentration distribution data covering the entire target area. The distribution data has clear spatial coordinate attributes and can accurately locate a small area near the ceiling.

[0072] In practical applications, this embodiment effectively filters out random fluctuations caused by turbulence by performing time averaging on the original monitoring data, and extracts characteristic temperature and concentration values ​​representing quasi-steady-state combustion. The generated continuous distribution data field (discrete sampling and digital representation of continuous physical states such as temperature field and concentration field in the actual area near the ceiling, with each data point corresponding to a specific location and a specific physical quantity value in space) is the input for subsequent dual-criteria spatial coupling analysis.

[0073] As an optional embodiment of the present invention, the step of comparing the concentration distribution data of the unburned fuel and the flue gas temperature distribution data with the corresponding combustion thresholds, and assessing the spontaneous combustion risk of the confined space based on the comparison results, includes: Based on the unburned fuel, its auto-ignition temperature threshold and ultimate combustible concentration threshold are determined and used as a comparison benchmark.

[0074] Specifically, based on the inherent properties of fuel, two key physical quantity thresholds are determined: the auto-ignition temperature threshold (AIT), which is the standard for determining whether the local environment has sufficient ignition energy; and the limiting flammability concentration threshold (LFL), which is the standard for determining whether the unburned gas mixture reaches a flammable concentration. These two thresholds are the unified benchmarks for comparison of all subsequent grid cells.

[0075] The flue gas temperature distribution data corresponding to each confined space is compared with the auto-ignition temperature threshold one by one to determine the first grid cell whose flue gas temperature distribution data exceeds the limit.

[0076] Specifically, this step performs a full-domain scan of the temperature field, traversing every grid cell in the computational domain and checking whether its stored time-averaged flue gas temperature value is greater than or equal to the auto-ignition temperature threshold (AIT). All grid cells that satisfy the condition "temperature ≥ AIT" are marked and set together, referred to as the "first grid cell set".

[0077] The concentration distribution data of the unburned fuel corresponding to each confined space is compared one by one with the lower limit of combustible concentration to determine the second grid cell in which the concentration distribution data of the unburned fuel exceeds the limit.

[0078] Specifically, this step involves a synchronous full-domain scan of the concentration field. The algorithm traverses each grid cell, checking whether its stored time-averaged unburned fuel concentration value is greater than or equal to the lower flammability limit (LFL) threshold (possibly adjusted for temperature or multi-component factors). All grid cells satisfying the "concentration ≥ LFL" condition are marked and grouped together, referred to as the "second grid cell" set.

[0079] The concentration of total unburned gas in confined compartments under three different HRRs was calculated using the following formula:

[0080] in, , and These represent the concentrations of carbon monoxide, alkane, and hydrogen, respectively.

[0081] See Figure 6 , Figure 6The centerline temperature distribution of the confined compartment configuration is shown at three representative heat release rates (HRR): (a) 0.5 kW, (b) 4.6 kW, and (c) 18.6 kW. At the lowest HRR of 0.5 kW, the semi-confined configuration exhibits a similar vertical temperature distribution, consistent with adequately ventilated combustion conditions. In contrast, the fully confined configuration (Conf. 5) shows a significant temperature decay towards the ceiling, indicating insufficient ventilation or oxygen-limited combustion.

[0082] As the HRR increased from 0.5 kW to 18.6 kW, the peak gas temperature generally decreased across all configurations. This trend reflects the gradual limitation of available oxygen within the compartment: despite the higher thermal power, the restricted ventilation reduced combustion efficiency, leading to a decrease in local gas temperature. This behavior differs from typical experimental observations and may be attributed to the lack of leakage or infiltration flow in the numerical model.

[0083] Figure 7 The instantaneous temperature fields (t=1.5 s) for configurations 2 through 5 at HRR=4.6 kW are shown: (a) Conf. 2; (b) Conf. 3; (c) Conf. 4; (d) Conf. 5. In Conf. 2, with its open doors and windows, a coherent flame plume with well-defined flue gas stratification is observed, confined to the upper part of the compartment. As the degree of confinement increases from Conf. 3 to Conf. 5, the flue gas layer gradually descends, eventually forming a complete volumetric flue gas fill in the fully enclosed configuration. This evolution demonstrates a transition from a ventilated controlled to a ventilated confined combustion state.

[0084] These thermal and flue gas dynamics highlight how increasing confinement transforms the internal compartment environment from one exhibiting pronounced stratification to one characterized by uniform flue gas filling and suppressed peak temperatures. These insights provide a crucial foundation for understanding subsequent gas buildup, flammability hazards, and spontaneous combustion potential within confined compartments.

[0085] Through spatial correlation analysis, the overlapping grid cells in the first grid cell and the second grid cell are identified as grid cells with spontaneous combustion risk. When the confined space corresponding to the combination of ventilation restriction degree and heat release rate contains grid cells with spontaneous combustion risk, it is determined that the confined space has spontaneous combustion risk.

[0086] Specifically, spatial logic operations are performed on the "first set of grid cells" (high-temperature zone) and the "second set of grid cells" (fuel-rich zone) to identify grid cells that belong to both sets. These cells are located in locations that simultaneously meet the dual conditions of "sufficient heat" and "sufficient fuel," and are therefore identified as "grid cells with a risk of spontaneous combustion." In a simulation scenario defined by a specific combination of "ventilation restriction degree (configuration) + heat release rate (HRR)," if there is at least one grid cell marked as risky, then from an engineering safety perspective, it can be determined that the confined space has a risk of spontaneous combustion under this fire condition.

[0087] To evaluate flame behavior under different confinement levels, the numerical heat release rate (HRR) obtained for four confinement configurations in the range of specified fire source power from 0.5 kW to 18.6 kW was analyzed. Figure 8 (a) illustrates the variation of the actual HRR over time when the theoretical HRR is 0.5 kW. For Conf. 2, 3, and 4, the actual HRR stabilizes at 0.5 kW, indicating adequate ventilation combustion. In contrast, the fully confined configuration Conf. 5 exhibits significant HRR fluctuations due to insufficient ventilation conditions.

[0088] At a theoretical HRR of 4.6 kW, the actual HRR in the semi-confined configurations (Conf. 2 and Conf. 3) remained close to the specified value, despite increased instability. In Conf. 4 (a highly confined and poorly ventilated case), combustion became significantly unstable. In Conf. 5, the flame extinguished completely in approximately 1.5 seconds. As the theoretical HRR increased to 18.6 kW, the semi-confined configurations (Conf. 2 and Conf. 3) showed a gradual increase in actual HRR, eventually stabilizing near the specified value, a typical pattern for poorly ventilated combustion. In Conf. 4, stronger fluctuations were observed, while in Conf. 5, the flame extinguished in approximately 0.5 seconds.

[0089] Figure 8 The theoretical HRRs are: (a) 0.5 kW; (b) 4.6 kW; (c) 18.6 kW. Figure 8 The actual HRR evolution depicted is key evidence for understanding the transition of combustion states: from adequate ventilation to controlled ventilation, and finally flame extinction as the degree of confinement increases. Configurations can be categorized as follows: Conf. 2 and Conf. 3 (slightly insufficient ventilation due to the opening but maintaining combustion), Conf. 4 (height confinement and insufficient ventilation, unstable combustion), and Conf. 5 (extreme confinement, leading to rapid flame extinction). It should be noted that simulation accuracy is limited under conditions of height confinement and insufficient ventilation (e.g., Conf. 4 and Conf. 5), and FDS reliability is known to be low under these conditions.

[0090] It is also possible to focus on improving prediction accuracy and addressing this inherent limitation of FDS.

[0091] However, these limitations do not diminish the validity of the analysis of the semi-restricted configurations (Conf. 2 and Conf. 3), in which continued combustion leads to significant unburned gas accumulation and high-temperature risk.

[0092] In the above configuration scenarios, the highest risk does not occur in completely open or completely sealed environments, but rather in certain semi-confined configurations. In these cases, limited ventilation leads to the accumulation of unburned fuel under the ceiling, while the presence of openings creates the possibility of sudden air mixing, such as when a door is opened, triggering spontaneous combustion or a flashback explosion. Such semi-confined scenarios are common in real-world environments, such as rooms with slightly ajar doors or partially open windows, or in equipment compartments, where the associated fire risks are often overlooked. It should be emphasized that the risk assessment methodology used in this study provides a reference approach for assessing such hazards.

[0093] Therefore, through precise spatial correlation analysis, overlapping areas that simultaneously meet the conditions of high temperature and rich fuel are located, which are the potential spontaneous combustion risk points. Finally, based on whether the risk point exists in a fire scenario defined by a specific ventilation configuration and ignition source power, a clear conclusion is output regarding whether there is a spontaneous combustion risk in that scenario. This method transforms complex fire dynamics phenomena into automatically executable quantitative criteria, realizing the transition from simulation data to engineering decisions, and can reveal the highest risks inherent under "partially restricted" ventilation conditions.

[0094] As an optional embodiment of the present invention, determining the auto-ignition temperature threshold and the limiting combustible concentration threshold of the unburned fuel and using them as a comparison benchmark includes: The auto-ignition temperature and standard lower flammability limit concentration of the unburned fuel under standard conditions were obtained. Specifically, the key safety parameters of each fuel (such as propane) under normal temperature and pressure (standard conditions) were obtained from standard data sheets or experiments, as shown in Table 3 below: Table 3

[0095] Among these, the auto-ignition temperature (AIT, e.g., 455°C for propane) and the lower flammability limit (LFL, e.g., 2.1% for propane) are the baseline values ​​for all subsequent comparisons and corrections.

[0096] When the local temperature of a grid cell in the flue gas temperature distribution data deviates from the standard state, the standard lower limit concentration of combustible gas is dynamically corrected based on the local temperature to obtain a first corrected concentration threshold that matches the grid cell.

[0097] Specifically, the LFL, UFL, and AIT values ​​for carbon monoxide, hydrogen, and propane at ambient temperature and pressure can be found in the table above. The flammability limits can be extended to a wider temperature range using the following formula:

[0098] in and These represent the lower flammability limits at temperatures T1 and T2, respectively. UFL(T1) and These represent the upper limits of flammability at temperatures T1 and T2, respectively.

[0099] Since the LFL of fuel decreases linearly with increasing temperature, using the LFL at room temperature (2.1%) as a criterion in a high-temperature fire environment would overestimate the safety margin and miss risks. Therefore, this method uses the local flue gas temperature of each high-temperature grid cell in the simulation to dynamically correct the standard LFL, resulting in a more stringent and realistic "first correction concentration threshold". For example, if the flue gas temperature is 500°C, the LFL of propane may be corrected to about 1.5%.

[0100] When the unburned fuel is a mixture of multiple combustible gas components, the equivalent lower flammability limit concentration of the mixture is determined based on the component concentration of each combustible gas component and the corresponding standard lower flammability limit concentration, and is used as the second corrected concentration threshold.

[0101] Specifically, in actual fires, "unburned fuel" is often a mixture of gases (such as propane, carbon monoxide, and hydrogen). Different gases have different flammability, and the overall flammability of the mixture is not simply added together. When multiple flammable gases are present, the "equivalent lower flammability limit concentration" (i.e., the second corrected concentration threshold) of the entire gas mixture is calculated using the following formula based on the concentration of each component and its respective LFL. Use the following formula to predict the lower and upper limits of flammability:

[0102] in, These represent the volume fraction percentage, lower flammability limit, and upper flammability limit of the i-th component, respectively; n represents the total number of components. This indicates the lower limit of total flammability of the gas mixture; It indicates the total flammability limit of the gas mixture; considering only flammable species, it is more accurate than LFL using a single fuel.

[0103] The auto-ignition temperature is used as a temperature comparison benchmark, and the first or second corrected concentration threshold is established as a concentration comparison benchmark according to applicable conditions.

[0104] Specifically, in actual use, if the main fuel is a single fuel but the temperature distribution is uneven, the "first corrected concentration threshold" (the value after temperature correction) is selected; if the main fuel is a mixture of multiple fuels, the "second corrected concentration threshold" (the equivalent value of multiple components) is selected. In the actual algorithm, both may be calculated and the lower (more stringent) value may be selected as the final concentration comparison benchmark to ensure the conservatism of the risk assessment.

[0105] Following the specific numerical simulation, the following analysis examines the concentration distributions of O2, CO2, and propane. Figure 9 Centerline oxygen concentrations were plotted for four confined configurations at (a) HRR = 0.5 kW; (b) HRR = 4.6 kW; and (c) HRR = 18.6 kW. At HRR = 0.5 kW, the oxygen concentration increases from floor to ceiling, indicating adequate ventilation. At HRR = 4.6 kW and HRR = 18.6 kW, for Conf. 2, Conf. 3, and Conf. 4, the oxygen concentration decreases to approximately 0.5% along the height from floor to ceiling, indicating insufficient ventilation. Considering the fully confined configuration, Conf. 5, the oxygen concentration is consistently higher than in the other confined configurations, even exceeding 0.1 (mol / mol) at HRR = 18.6 kW. However, in Conf. 5, the flame extinguishes rapidly because the temperature is below the temperature specified in the combustion model. Figure 10 The instantaneous (t=1.5s) oxygen concentration field at HRR=4.6 kW is shown: (a) Conf. 2; (b) Conf. 3; (c) Conf. 4; (d) Conf. 5. It can be observed that for the less confined configurations, i.e., Conf. 2, Conf. 3, and Conf. 4, the oxygen layer in the upper region is thinner as combustion consumes oxygen. In the fully confined configuration, i.e., Conf. 5, only the oxygen near the burner is consumed, while the oxygen level is sufficient throughout the compartment. It can be further inferred that oxygen concentration is not a decisive parameter for combustion in the current numerical simulation. Figure 11 Centerline CO2 concentrations were plotted for four confined configurations at (a) HRR = 0.5 kW; (b) HRR = 4.6 kW; and (c) HRR = 18.6 kW. At HRR = 0.5 kW, the CO2 concentration in the less confined configurations (i.e., Conf. 2, Conf. 3, and Conf. 4) initially increased from floor to ceiling and then decreased, due to the fire plume being below the ceiling. At HRR = 4.6 kW and HRR = 18.6 kW, the CO2 concentration initially increased and then remained constant near the ceiling, due to a stable smoke layer near the ceiling. The CO2 concentration in Conf. 5 was consistently lower than in the other confined configurations, likely due to insufficient ventilation.

[0106] Combination Figure 12 The instantaneous (t=1.5s) CO2 concentration field at HRR=4.6 kW is shown, where (a) Conf. 2; (b) Conf. 3; (c) Conf. 4; (d) Conf. 5. As can be seen from the figure analysis, the flue gas stratification is clearly shown in the configuration with a lower degree of confinement, while in the fully confined configuration (i.e. Conf. 5), the entire compartment is filled with flue gas, which is consistent with the previous discussion.

[0107] Figure 13 The study describes the centerline propane concentrations in four confined configurations at (a) HRR = 0.5 kW; (b) HRR = 4.6 kW; and (c) HRR = 18.6 kW. At HRR = 0.5 kW, the propane concentration in all configurations rapidly decreases to zero near the ceiling along the centerline. A similar trend is observed at 4.6 kW, although the concentration stabilizes at a low but non-zero value near the ceiling. A significant change occurs at 18.6 kW, with significant unburned propane accumulation observed near the ceiling, particularly in the semi-confined configurations (Conf. 2 and Conf. 3), where the concentration reaches approximately 0.2%. While this value is 2.1% (volume fraction) below the lower flammability limit (LFL) in absolute terms, it indicates a significant and hazardous accumulation trend under high HRR conditions. This propane accumulation is not present in the more open Conf. 1 and the fully confined Conf. 5.

[0108] This phenomenon arises from the combined effects of confined geometry, poorly ventilated yet partially open conditions (allowing for limited gas exchange), and sufficiently high ignition source power. These factors collectively promote incomplete combustion and lead to localized enrichment of unburned fuel in the upper compartment, highlighting the unique risk of gas accumulation in partially confined fire scenarios.

[0109] It should be noted that, within the scope of the risk analysis, minimum values ​​of LFL and AIT were considered for a more reliable risk assessment. Therefore, the concentration of total unburned gas in confined compartments under three different HRRs was compared with the LFL value of propane. Figure 6 The roof temperature for each configuration at the three HRRs was also compared with the ignition temperature (AIT) of propane.

[0110] Therefore, the spontaneous combustion risk of the four restricted configurations under the three HRRs can be comprehensively assessed by considering the unburned gas concentration and temperature, as summarized in Tables 4, 5 and 6 below.

[0111] Table 4 - Spontaneous combustion risk of four restricted configurations at HRR=0.5 kW

[0112] As shown in Table 4, there is no risk of spontaneous combustion in the four confined configurations with an HRR of 0.5 kW, provided that the conditions are low temperature and no unburned gas is generated. This lack of risk is due to the very well-ventilated environment with an equivalence ratio of less than 1.

[0113] Table 5 - Spontaneous combustion risk of four restricted configurations at HRR=4.6 kW

[0114] For HRR=4.6 kW, Table 5 shows that the temperature is one of the criteria for achieving spontaneous combustion, but the concentration of total unburned gas remains below the flammable range.

[0115] Therefore, it can be assumed that there is no risk of spontaneous combustion at HRR=4.6 kW. Furthermore, this situation represents an under-ventilated environment with an equivalence ratio of approximately 1.

[0116] Table 6 - Spontaneous combustion risk of four restricted configurations at HRR=18.6 kW

[0117] As shown in Table 6, at an HRR of 18.6 kW, the risk of spontaneous combustion is most significant in Conf. 2 (compartments with open doors and windows) and Conf. 3 (compartments with open windows). This increased risk stems from the accumulation of large amounts of unburned gas near the ceiling and the high temperature. In contrast, more constrained configurations such as Conf. 4 exhibit reduced spontaneous combustion potential, primarily due to the ceiling temperature being below the auto-ignition temperature (AIT).

[0118] Therefore, by introducing temperature correction and multi-component equivalent calculation algorithms in the above steps, the risk assessment can adapt to the actual situation of uneven temperature field and complex gas composition in real fires, thereby significantly improving the accuracy and reliability of the risk assessment.

[0119] As an optional embodiment of the present invention, the method for assessing the spontaneous combustion risk of confined space fires further includes: Based on the frequency with which the configuration corresponding to the degree of ventilation restriction is determined to have a risk of spontaneous combustion under different heat release rates, the configuration is divided into different risk levels; wherein, the higher the frequency with which it is determined to have a risk of spontaneous combustion, the higher the corresponding risk level.

[0120] Specifically, the spontaneous combustion risk of different ventilation configurations under different fire source powers is determined. Based on this, a configuration classification method based on the frequency of risk occurrence is introduced: by statistically analyzing the number of times each ventilation configuration is judged as "risky" in all test conditions, the configuration can be divided into different risk levels (such as high and low), according to the aforementioned risk assessment table, for example: Configuration Conf. 2 (doors and windows open): No risk at HRR=0.5 kW, no risk at HRR=4.6 kW, risk at HRR=18.6 kW → risk frequency=1.

[0121] Configuration Conf. 3 (window only): The judgment result is the same as Conf. 2 → risk frequency = 1.

[0122] Configuration Conf. 4 (door open only): No risk under all HRRs → risk frequency = 0.

[0123] Configuration Conf. 5 (Completely Closed): No risk under all HRRs → Risk frequency = 0. High-risk level: Conf. 2, Conf. 3 (frequency = 1).

[0124] Low risk level: Conf. 4, Conf. 5 (frequency = 0).

[0125] This clearly reveals the key pattern that "partially restricted" ventilation configurations (such as slightly open doors and windows) pose the highest risk under high fire source power. This classification result can be directly applied to building fire protection design and safety assessment, guiding designers to avoid high-risk ventilation modes or to take targeted prevention and control measures for existing high-risk spaces.

[0126] like Figure 14 As shown, this embodiment of the invention provides a confined space fire spontaneous combustion risk assessment system 200, which applies the confined space fire spontaneous combustion risk assessment method described in the above embodiment, including: The optimization construction module 210 is used to: construct a ceiling impact fire model based on a confined space, wherein the optimized discretization parameters of the ceiling impact fire model are determined in advance by simulating and verifying a free fire plume based on an open space; The numerical simulation module 220 is used to: perform numerical simulations based on the ceiling impact fire model under various degrees of ventilation restriction and various heat release rates to obtain simulation results; Data extraction module 230 is used to: extract the concentration distribution data of unburned fuel and the flue gas temperature distribution data at the ceiling location based on the simulation results; The spontaneous combustion risk assessment module 240 is used to: compare the concentration distribution data of the unburned fuel and the flue gas temperature distribution data with the corresponding combustion threshold, and assess the spontaneous combustion risk of the confined space based on the comparison results.

[0127] The specific implementation method of this embodiment can be referred to the corresponding implementation method described above, and will not be described again here.

[0128] An electronic device provided by an embodiment of the present invention includes a memory and a processor; the memory is used to store a computer program; the processor is used to implement the method for assessing the risk of spontaneous combustion in confined space fires as described above when the computer program is executed.

[0129] Alternatively, an electronic device includes a memory and a processor coupled to the memory; the memory is configured to store a computer program; the processor is configured to perform the following operations when the computer program is executed: A ceiling impact fire model based on confined space is constructed, wherein the optimized discretization parameters of the ceiling impact fire model are determined in advance by simulating and verifying free fire plumes based on open space; Based on the aforementioned ceiling impact fire model, numerical simulations were performed under various degrees of ventilation restriction and various heat release rates to obtain simulation results. Based on the simulation results, the concentration distribution data of unburned fuel and the flue gas temperature distribution data at the ceiling location were extracted. The concentration distribution data of the unburned fuel and the temperature distribution data of the flue gas are compared with the corresponding combustion thresholds, and the spontaneous combustion risk of the confined space is assessed based on the comparison results.

[0130] This invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the method for assessing the risk of spontaneous combustion in confined spaces as described above.

[0131] Alternatively, a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the following operations: A ceiling impact fire model based on confined space is constructed, wherein the optimized discretization parameters of the ceiling impact fire model are determined in advance by simulating and verifying free fire plumes based on open spaces. Based on the ceiling impact fire model, numerical simulations are performed under various degrees of ventilation restriction and various heat release rates to obtain simulation results. Based on the simulation results, the concentration distribution data of unburned fuel and the flue gas temperature distribution data at the ceiling location are extracted. The concentration distribution data of unburned fuel and the flue gas temperature distribution data are compared with the corresponding combustion thresholds, and the spontaneous combustion risk of the confined space is assessed based on the comparison results.

[0132] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Although the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A method for assessing the spontaneous combustion risk of fires in confined spaces, characterized in that, include: A ceiling impact fire model based on confined space is constructed, wherein the optimized discretization parameters of the ceiling impact fire model are determined in advance by simulating and verifying free fire plumes based on open space; Based on the aforementioned ceiling impact fire model, numerical simulations were performed under various degrees of ventilation restriction and various heat release rates to obtain simulation results. Based on the simulation results, the concentration distribution data of unburned fuel and the flue gas temperature distribution data at the ceiling location were extracted. The concentration distribution data of the unburned fuel and the temperature distribution data of the flue gas are compared with the corresponding combustion thresholds, and the spontaneous combustion risk of the confined space is assessed based on the comparison results.

2. The method for assessing the spontaneous combustion risk of confined space fires according to claim 1, characterized in that, The optimized discretization parameters include the grid size for spatial discretization and / or the computational domain size for defining the computational boundary. The prior simulation and verification of free fire plumes based on open space includes: Construct a baseline model of free fire plumes based on the aforementioned open space; Mesh sensitivity analysis and computational domain independence verification are performed on the baseline model to determine the optimized discretization parameters that enable the simulation results of the baseline model to achieve mesh independence and eliminate boundary effects.

3. The method for assessing the spontaneous combustion risk of confined space fires according to claim 2, characterized in that, The step of performing grid sensitivity analysis and computational domain independence verification on the benchmark model to determine the optimized discretization parameters that enable the simulation results of the benchmark model to achieve grid independence and eliminate boundary effects includes: The characteristic fire diameter is determined based on the minimum heat release rate corresponding to the benchmark model, and the theoretically feasible range of the grid size is determined based on the characteristic fire diameter and the preset grid resolution ratio range. Within the theoretically feasible range, by implementing comparative simulations of multi-level grid resolution and multi-size computational domains, the grid size that achieves grid independence and the computational domain size that eliminates boundary effects are determined based on the convergence characteristics of temperature and velocity distributions along the flame centerline, respectively. The determined grid size and / or computational domain size are applied to the baseline model, and the final verification is completed by comparing the degree of agreement between the simulated flame height and the theoretical prediction data, thus obtaining the optimized discretization parameters.

4. The method for assessing the spontaneous combustion risk of confined space fires according to claim 1, characterized in that, The construction of the ceiling impact fire model based on confined space includes: Establish a geometry comprising a confined compartment and its extended domain, wherein the confined compartment is configured with material parameters, boundary conditions, and a burner; Based on the geometric configuration, a dynamic solver for turbulence-combustion-radiation coupling is configured, and the optimized discretization parameters are applied to complete the model discretization, thereby obtaining the ceiling impact fire model based on confined space.

5. The method for assessing the spontaneous combustion risk of confined space fires according to any one of claims 1-4, characterized in that, The numerical simulation based on the ceiling impact fire model was conducted under various degrees of ventilation restriction and various heat release rates to obtain simulation results including: The opening and closing state of the ventilation openings in the confined compartment in the ceiling impact fire model is controlled to determine multiple configurations that correspond one-to-one with the degree of ventilation confinement. For each configuration, multiple heat release rates with different levels are set, and numerical simulations are performed on all defined parameterized scenarios until a quasi-steady state is reached. Configure and output the simulation results, wherein the simulation results include the flow state parameters at the ceiling location and its neighborhood location, the concentration distribution data, and the relevant heat flux data.

6. The method for assessing the spontaneous combustion risk of confined space fires according to claim 5, characterized in that, Based on the simulation results, the extraction of unburned fuel concentration distribution data and flue gas temperature distribution data at the ceiling location includes: From the simulation results, read the original time-series monitoring data at the monitoring locations pre-configured for monitoring the ceiling location; The original time-series monitoring data are averaged over time to obtain the flue gas temperature parameters and unburned fuel concentration parameters corresponding to each monitoring location. Based on the spatial coordinates of each monitoring location, the obtained flue gas temperature parameters are reconstructed into flue gas temperature distribution data at the ceiling location, and the obtained unburned fuel concentration parameters are reconstructed into unburned fuel concentration distribution data at the ceiling location.

7. The method for assessing the spontaneous combustion risk of confined space fires according to any one of claims 1-4, characterized in that, The step of comparing the concentration distribution data of the unburned fuel and the flue gas temperature distribution data with the corresponding combustion thresholds, and assessing the spontaneous combustion risk of the confined space based on the comparison results, includes: Based on the unburned fuel, its auto-ignition temperature threshold and the ultimate combustible concentration threshold are determined and used as a comparison benchmark; The flue gas temperature distribution data corresponding to each confined space is compared with the auto-ignition temperature threshold one by one to determine the first grid cell whose flue gas temperature distribution data exceeds the limit. The concentration distribution data of the unburned fuel corresponding to each confined space is compared with the lower limit of combustible concentration threshold one by one to determine the second grid cell in which the concentration distribution data of the unburned fuel exceeds the limit; Through spatial correlation analysis, the overlapping grid cells in the first grid cell and the second grid cell are identified as grid cells with spontaneous combustion risk. When the confined space corresponding to the combination of ventilation restriction degree and heat release rate contains grid cells with spontaneous combustion risk, it is determined that the confined space has spontaneous combustion risk.

8. The method for assessing the spontaneous combustion risk of confined space fires according to claim 7, characterized in that, The step of determining the auto-ignition temperature threshold and the limiting combustible concentration threshold based on the unburned fuel and using them as a comparison benchmark includes: Obtain the auto-ignition temperature and standard lower flammability limit concentration of the unburned fuel under standard conditions; When the local temperature of the grid cell in the flue gas temperature distribution data deviates from the standard state, the standard lower limit of combustible concentration is dynamically corrected according to the local temperature to obtain a first corrected concentration threshold that matches the grid cell. When the unburned fuel is a mixture of multiple combustible gas components, the equivalent lower flammability limit concentration of the mixture is determined based on the component concentration of each combustible gas component and the corresponding standard lower flammability limit concentration, and is used as the second corrected concentration threshold. The auto-ignition temperature is used as a temperature comparison benchmark, and the first or second corrected concentration threshold is established as a concentration comparison benchmark according to applicable conditions.

9. The method for assessing the spontaneous combustion risk of confined space fires according to claim 5, characterized in that, Also includes: Based on the frequency with which the configuration corresponding to the degree of ventilation restriction is determined to have a risk of spontaneous combustion under different heat release rates, the configuration is divided into different risk levels; wherein, the higher the frequency with which it is determined to have a risk of spontaneous combustion, the higher the corresponding risk level.

10. A system for assessing the risk of spontaneous combustion in confined spaces, characterized in that, The method for assessing the spontaneous combustion risk of confined space fires as described in any one of claims 1-9 includes: An optimized construction module is used to: construct a ceiling impact fire model based on a confined space, wherein the optimized discretization parameters of the ceiling impact fire model are determined in advance by simulating and verifying a free fire plume based on an open space; The numerical simulation module is used to: perform numerical simulations based on the ceiling impact fire model under various degrees of ventilation restriction and various heat release rates to obtain simulation results; The data extraction module is used to: extract the concentration distribution data of unburned fuel and the flue gas temperature distribution data at the ceiling location based on the simulation results; The spontaneous combustion risk assessment module is used to: compare the concentration distribution data of the unburned fuel and the flue gas temperature distribution data with the corresponding combustion thresholds, and assess the spontaneous combustion risk of the confined space based on the comparison results.

Images