Fire Experiments Research Articles

Application of Artificial Neural Networks for Mathematical Modelling of Horizontal Jet Fires

Aim: This article focuses on the use of artificial neural networks to mathematically describe the parameters that determine the size of a jet fire flame. To teach the neural network, the results of a horizontal propane jet fire, carried out experimentally and using CFD mathematical modelling, were used. Project and methods: The main part of the work consisted of developing an artificial neural network to describe the flame length and propane-air mixing path lengths with good accuracy, depending on the relevant process parameters. Two types of data series were used to meet the stated objective. The first series of data came from field tests carried out by CNBOP-PIB and from research contained in scientific articles. The second type of data was provided by numerical calculations made by the authors. The methods of computational fluid mechanics were used to develop the numerical simulations. The ANSYS Fluent package was used for this purpose. Matlab 2022a was used to develop the artificial neural network and to verify it. Results: Using the nftool function included in Matlab 2022a, an artificial neural network was developed to determine the flame length Lflame and the length of the Slift-off mixing path as a function of the diameter of the dnozzle and the mass flux of gas leaving the nozzle. Using Pearson’s correlation coefficient, a selection was made of the best number of neurons in the hidden layer to describe the process parameters. The neural network developed allows Lflame and Slift-off values to be calculated with good accuracy. Conclusions: Artificial neural networks allow a function to be developed to describe the parameters that determine flame sizes in relation to process parameters. For this purpose, the results of the CFD simulations and the results of the jet fire experiments were combined to create a single neural network. The result is a ready-made function that can be used in programmes for the rapid determination of flame sizes. Such a function can support the process of creating scenarios in the event of an emergency. A correctly developed neural network provides opportunities for the mathematical description of jet fires wherever experimental measurements are not possible. Solution proposed by the authors does not require a large investment in ongoing calculations, as the network can be implemented in any programming language. Keywords: computational fluid mechanics, artificial neural networks, jet fire

Simple model combined with data assimilation to forecast buoyancy-driven ceiling jet propagation in tunnel fires

Faster-than-real-time forecast for the ceiling jet head propagation is crucial for evacuation guidance and rescue operation in the early stages of tunnel fires. The ceiling jet head propagation is driven by buoyancy and thus significantly dominated by the heat transfer process in the head region. Simplified models have been established but proved challenging in forecasting complex dynamic processes. The cumulative deviations induced by the model limitations and the input parameter uncertainties lead to distorted forecast during the ceiling jet dynamic propagation process, thereby constraining the applicability of such models in practical scenarios. In this study, an assimilation method, the ensemble Kalman filter (EnKF), was combined with a simple theoretical model to forecast the ceiling jet propagation. Using the EnKF to assimilate the monitored local data, the key input parameters such as mass flow rate m˙ and the temperature of the ceiling jet head Ts are corrected in real time, thus alleviating model prediction deviation. Meanwhile, a heat loss rate correction coefficient λα, introduced to correct the heat loss rate during the ceiling jet propagation process, is dynamically estimated as well. This enables the model to characterize the heat transfer process more accurately, consequently ensuring faster-than-real-time forecasts of the subsequent propagation stage. Finally, the forecasting effectiveness of the proposed combined method was examined in a 1:6 reduced-size experimental tunnel fire scenario and full-scale simulated tunnel fire scenarios with various fire source locations, tunnel geometries, and suddenly changing HRR. In response to the demand of minimizing both the type and quantity of sensors in practical applications, a factor analysis was also performed to examine the robustness of this method. The results exhibited good adaptability to different fire scenarios with very limited monitored data.

Fire Resistance of Prestressed Multispan Steel Truss Composite Slab

Prestressed steel truss composite slab is composed of a precast prestressed bottom slab, a steel truss, and a postcast concrete layer. In order to investigate the fire resistance performance of prestressed multispan steel truss composite (STPC) slabs under the coupled effect of high temperature and load, four specimens of such composite slab containing two spans with various postcast concrete layer thicknesses were fabricated. Fire experiments were conducted following the ISO 834 standard temperature–time curve. Two of the specimens underwent two-span fire experiments, while the other two underwent single-span fire experiments. During the experiment, the bottom of the multispan STPC slabs exhibited serious bursting under fire and the precast concrete layer at the bursting locations fell off. The experiment result showed that the degree of bursting was much higher than that of nonprestressed composite slabs and single-span slabs. Although bond failure was observed at the composite interface where horizontal cracks developed, the STPC slabs still performed well in terms of load bearing, which could be attributed to the contribution of the steel truss. Under the coupled action of high temperature and load, concrete cracks were prone to occurring at the cut-off position of the top reinforcement near the supports; therefore, the anchorage length of these reinforcing bars should be appropriately extended. The damage to the STPC slab specimens after fire was found to be more serious for those with a less-think postcast concrete layer. In the single-span fire experiments, the structural performance of the unfired span was not prominently affected. Due to the bursting of the precast layer, the thickness of the STPC slabs was reduced, resulting in decreases in the thermal insulation and the fire resistance limit. The fire resistance limits of the STPC slabs with 60 mm and 80 mm postcast concrete layers were 90 min and 130 min, respectively. When a high fire resistance limit of the STPC slabs is required in a design, it is recommended that the thickness of the postcast concrete layer be no less than 60 mm.

Study of relationship between motion of mechanisms in gas operated weapon and its shock absorber

The article deals with the motion of the breech block carrier and the weapon casing of an automatic weapon mounted on a flexible carriage and the base of the weapon. Earlier works, which did not consider the dynamic properties of the base of the weapon, did not allow to reconcile the calculated and experimental results of the weapon casing displacement when shooting from firing rests. For the analysis of the motion of individual parts, the methods of mathematical modelling and firing experiments using a high-speed camera were chosen. Calculations show the best accord with experiment when modelling the system with 4 degrees of freedom. The oscillation of the system regarding the movement of the breech block carrier and the weapon casing was investigated under changed conditions of rate of fire, the use of a muzzle brake and different types of shock absorbers. The velocities and displacements of the weapon casing and the breech block carrier at different values of the impulse of the gases to the breech block carrier were determined.

Assessment on liquid Li fire risk under humid air condition and with heat insulators

The IFMIF-like fusion neutron sources (FNSs) such as A-FNS and IFMIF-DONES have Li fire risks and the R&D activity for the Li fire risk reduction has been conducted under the IFMIF/EVEDA project. The objectives of this study are to experimentally elucidate humidity condition in airat which no Li fire ignites and the effect of the heat insulator on the Li fire ignition. The experimental set-up was designed and fabricated in order to perform Li fire experiments safely and systematically. Around 1 g of metallic Li samples were heated with and without the heat insulators as an experimental parameter of humidity in air. In this work, no ignition was observed under the less than 0.15 vol% humidity air. The heating experiments with the heat insulators clarified that the reaction between Li and the heat insulators causing Li fire ignition didn’t occur. In conclusion, the results of the present study suggest the conditions of less than 0.15 vol% humidity air and employing the heat insulator of alkaline earth silicate wool are useful for the Li target system of FNSs.

Application of artificial intelligence methods to model the effect of grass curing level on spread rate of fires

Artificial intelligence (AI) enables new approaches to fire behaviour models of operational relevance, including prescribed burns. This is particularly important in modelling of processes that are poorly understood, such as live fuel's effect on fire propagation. The objective of this study was to apply AI algorithms to quantify the effect of the proportion of dead fuels in a senescing grassland, the curing level, on reducing the rate of fire spread relative to the fully cured condition. We applied three different machine learning (ML) models, regression trees, support vector regression (SVR) and Gene expression programming (GEP), two ensemble ML methods, Random Forest and GEP Forest, and non-linear regression analysis to an experimental fire dataset. Results show SVR and GEP as the best ML methods to model the curing level impact on fire spread. No differences in model fit were observed between the best ML methods and non-linear regression analysis.

The Effect of the Fuel Location and Ventilation Factor on the Fire Dynamics of Informal Settlement Dwellings

AbstractThree full-scale experimental compartment fires are compared to investigate the effect of the fuel location and the ventilation factor on under ventilated thermally thin bounded ISO-9705 compartments. Wood cribs were used as the fuel load and the crib placement was varied between two locations (back and middle) to study the effect of the fuel location. Furthermore, the ventilation conditions were changed from a door and window (i.e., ventilation factor of 2.58 m5/2) to only a door (i.e., ventilation factor of 2.26 m5/2) for the scenario where the cribs were placed at the back of the compartment. The novelty of this work lies in its examination of the time to flashover, gas layer temperature, heat release rate, and external radiative heat fluxes, specifically considering the impact of fuel location and ventilation factor. It was observed that placing the fuel package in the middle of the compartment led to a longer growth phase, hotter gas layer temperature, a higher Heat Release Rate (HRR) needed for flashover ($$\dot{q}_{fo}$$ q ˙ fo ), and higher external radiative heat fluxes through openings. It was also found that, decreasing the ventilation factor decreased the heat losses and therefore the $$\dot{q}_{fo}$$ q ˙ fo . Decreased ventilation also affected the height of the neutral plane, as one would expect, and the shape of the external plume, but did not have significant effect on the temperature within the compartment, the walls of the compartment, and the external radiative heat flux.

Uncertainties in the use of oxygen consumption calorimetry for heat release measurements in lithium-ion battery fires

Accurate measurement of the heat release from a battery fire is vital for risk management, product development and construction of accurate models. Oxygen consumption calorimetry is the most common method for heat release measurements in experimental fire tests. The strength of the method is that it can be applied to unknown compositions of fuel with sufficient accuracy. Despite that this method is used to estimate heat release from battery fires, the method is subject to discussion. In this work, the method is studied in-depth, and potential errors are structured and quantified. Uncertainties associated with self-generated oxygen and internal heat generation, total gas release from the battery and impact on the heat release calculations, as well as the assumed E-factor (i.e., heat release per unit mass of oxygen consumed), are thoroughly discussed. For a Li-ion battery fire, it is concluded that oxygen consumption calorimetry will exclude internal heat generation and underestimate the total heat released from the external flaming fire by up to 10 %. In addition, high rate of combustion reactions can result in that the measured peak heat release rate is underestimated much more, up to 100 %.

Experimental Analysis on the Behaviors of a Laboratory Surface Fire Spreading across a Firebreak with Different Winds

In this work, a series of laboratory surface fire experiments were performed over a pine needle fuel bed to investigate the effectiveness of a firebreak and the behaviors of a surface fire across a firebreak. Seven wind velocities of 0~3.0 m/s and six firebreak widths of 10~35 cm are varied. The behaviors of a surface fire across the firebreak, the heat flux received by fuel surface and fuel temperature before and after the firebreak are analyzed and compared simultaneously. The main conclusions are as follows: the behaviors of a surface fire spreading across a firebreak under different wind velocities are classified into three categories—no ignition, ignition by flame contact and ignition by spot fires. When the wind velocity is not more than 1.0 m/s, the surface fire cannot successfully cross the firebreak; as wind velocity changes from 1.5 m/s to 2.5 m/s, the fuel after the firebreak can be ignited by flame contact for relatively narrow firebreak conditions; when the wind velocity increases to 3.0 m/s, the burning fuel can be blown away along the fuel bed, and the fuel behind the firebreak will be ignited by spot fire. A linear relationship between the threshold of firebreak width and the fireline intensity is obtained, and the linear fitting coefficient in this paper is larger than the results reported by Wilson (0.36). For no ignition conditions, the fuel temperature and the heat flux received by the fuel after firebreak are significantly lower than those before the firebreak, whereas their variations over time are similar to those before the firebreak for ignition conditions. Moreover, for no ignition conditions, the maximum fuel temperature and the heat flux after the firebreak increase with wind velocity, but decrease with firebreak width. Additionally, when the fuel temperature (253 °C) and the heat flux received by the fuel considering the radiation and convection (43 kW/m2) after firebreak exceed a threshold value, the surface fire can successfully cross the firebreak.

Formation Mechanism of the “Green Above, Brown Below” Phenomenon in Yaozhou Kiln Celadon

Yaozhou Kiln is a famous ancient center for celadon production in China, located in present-day Shaanxi Province. While analyzing its olive-green celadon produced during the Song Dynasty, a common occurrence of brownish base (foot and bottom) was observed. This phenomenon can also be found in porcelain produced at other kilns in China and Vietnam. However, previous research has not systematically explored the coloration mechanism behind it. Through different analytical methods, coupled with reproduction firing experiments, this paper concludes that the brownish base is attributed to the diffusion of iron from the body and sand cushion into the thinly applied glaze on the base, as well as the crystallization formed by the combination of the sand cushion and the surface glaze. Factors influencing the depth of the brownish color include: (1) the iron content of the body; (2) the thickness of the base glaze; and (3) the sand cushion material.

Full-scale experimental study of the characteristics of electric vehicle fires process and response measures

Fire accidents are a source of concern for electric vehicles (EV). The spread of fire caused by thermal runaway (TR) in a scenario with a high concentration of EVs can result in a serious disaster. In this study, a full-scale EV fire experiment was conducted using thermal abuse to investigate response strategies to EV fires caused by TR. The pattern of TR fires spreading in EV was also investigated. The suppression capabilities of fire blankets, water spray, and compressed air foam on EV fires in the early stages were analyzed. During the test, the temperature distributions inside and outside the vehicle were measured. The maximum temperature recorded in the passenger compartment near the roof was 919.7 °C. The length of the jet flow was estimated from the images, with the longest being approximately 2.56 m. The fire suppression test demonstrated that the fire blankets effectively extinguished flames, and the temperature inside the car gradually decreased. A water mist device can reduce temperature and put out flames. The compressed air foam effectively extinguished the fire, with an average cooling rate of −9.8 °C/s. These results provide information on EV fire response strategies.

Study on flame height and temperature distribution of double-deck bridge fire based on large-scale fire experiments

Vehicle fire is one of the important causes of bridge fires. This study conducted large-scale actual fire experiments by simulating heat release rates (HRR) of oil tank truck fires with 50 MW, 100 MW, and 200 MW in rectangular oil pans of varying sizes. The flame shape and spatial temperature distribution at different HRRs were studied bycontrollingvariables such as wind speed,firesource location, andlayer spacing. The results indicated that the actual HRR of the large-scale fire exceeded the anticipated value. In large-scale fire experiments, flame splitting into two main peaks was observed along the long side of the rectangular oil pan due to the coupling effect of ambient wind, long side, and sidewall effect of the diagonal web member. The average flame height was determined through image processing, and subsequently, a formula for predicting flame height during bridge fires under varying heat release rates was developed through the use of dimensionless analysis and feature length analysis. The spatial temperature distribution in the vertical direction of the axis of the center of the fire source, the temperature distribution of the diagonal web member and the bottom of the dense crossbeams, which are critical components of the bridge, were also analyzed. Furthermore, the conclusion that the flame in the lower bridge space which split into two main peaks is validated by the results. The results of this study can provide important theoretical support for investigating bridge fire resistance strategies.

Experimental study of self-extinguishing behavior of hydrogen jet fires in the confined space with a ceiling vent

With the widespread utilization of hydrogen energy, the hydrogen safety issues, especially hydrogen jet fires, should be much paid attention to, since the latter present high temperature and radiation in confined space. Hydrogen jet fire experiments were conducted with various release rates and vent sizes in a confined space with ceiling vent to investigate the hydrogen jet flame height, oxygen concentration, and self-extinguishing time. The results show that, the large vent sizes with low hydrogen release rates result in well-ventilated hydrogen jet fires. As the vent size decreases or the hydrogen release rate increases, the hydrogen jet flame evolves from a well-ventilated one to one having self-extinguishing behavior with under-ventilation. The combustion process can be divided into four stages: the growth stage, the quasi-steady stage, the stretch stage and the self-extinguishing stage. The flame height increases with the hydrogen release rate at the quasi-steady stage, and the vent size has almost no effect. However, in the case of under-ventilation, the flame height at the stretch stage is determined by the joint actions of vent size and hydrogen release rate, and the flame height is positively connected with the size of the ceiling vent. The oxygen concentration and self-extinguishing time decrease dramatically as the vent size and hydrogen release rate increase. The parameter Q˙VA1/3 is proposed to reflect the coupling effects of hydrogen release rate and vent size, and the correlations of oxygen concentration and self-extinguishing time are determined with Q˙VA1/3.

The preparation and fire extinguishing mechanism research of a novel high-efficiency KHCO3 @HM dry powder

Hydromagnesite (HM) is a widely distributed primary natural carbonate mineral. To expand the application range of hydromagnesite and tap its fire-extinguishing potential, a new type of high-efficiency KHCO3 @HM dry powder was prepared by coating potassium bicarbonate on the surface of hydromagnesite with a dip coating method. KHCO3 @HM dry powder obtained a chemical inhibition effect through this method, and the fire extinguishing efficiency has been further improved. A series of tests were carried out to confirm the excellent physical and chemical inhibition effect of KHCO3 @HM dry powder. Specific surface area tests showed that the specific surface area of KHCO3 @HM dry powder (31.405 m2/g) was higher than NH4H2PO4 dry powder (12.766 m2/g). Flowability tests showed that KHCO3 @HM dry powder (0.04 g/s) achieved the same flowability as NH4H2PO4 dry powder (0.04 g/s). Local oil basin fire experiments showed that the average fire extinguishing time of KHCO3 @HM dry powder was shorter with less powder consumption. Under 0.2 MPa driving pressure, the fire extinguishing time of KHCO3 @HM dry powder was 2.2 s faster than NH4H2PO4 dry powder, and the dry powder consumption was 3.71 g less than NH4H2PO4 dry powder. The cooling effect of KHCO3 @HM dry powder (208 ℃) in 10 s was better than NH4H2PO4 dry powder (157 ℃), and KHCO3 @HM dry powder (399 ppm) showed a better smoke suppression effect than NH4H2PO4 dry powder (545 ppm). The pyrolysis of KHCO3 @HM dry powder reduced the fire temperature, diluted the oxygen concentration, and captured free radicals. The final pyrolysis products formed a thick oxide film to provide good heat insulation and asphyxiation effects. The excellent cooling, dilution, asphyxiation, and chemical inhibition mechanisms of KHCO3 @HM dry powder were demonstrated, and finally achieved a comprehensive performance comparable to NH4H2PO4 dry powder.

Flexural–torsional buckling of steel circular arches with I-section under central concentrated load in elevated-temperature environment

This article reports the fire resistance experiments conducted on I-section circular steel arches, as well as the observed out-of-plane flexural–torsional buckling (FTB) behavior and in-depth finite element numerical simulation results. The test was carried out using a large fire test furnace to provide a real fire environment with a nonuniform temperature field. A specially designed loading device allowed for unconstrained out-of-plane deformation of the specimens inside the furnace, thereby achieving effective load conditions and noncoupled response of the components. Numerical analyses were performed using a combination of the general codes Fluent and Abaqus, with the former used for heat transfer analyses, and a structural FE model including the loading device and complex geometric imperfections was established based on the measured data. Thus, accurate simulation of the out-of-plane behavior of the steel arches at elevated temperature was achieved. On this basis, an in-depth parametric study was carried out to investigate the effects of temperature field variations, load ratio (LR), and geometric imperfection. The analysis results show that FTB occurs at nonuniform temperatures for the arches under concentrated load, and the difference in temperature distribution in fire has a minimal effect on the FTB behavior. LR is negatively correlated with critical temperature; the smaller the LR, the more significant the weakening of critical loads by fire. The effect of initial geometrical imperfections on FTB at elevated temperatures is similar to that at ambient temperatures, and the effect on the critical temperature of FTB can almost be ignored when the imperfection amplitude is less than S/400.

Varying Heat Release Rates per Unit Area – The Impact in Underground Mines

Abstract The fire behaviour and smoke behaviour of a fire in an underground mine will be largely dictated by the heat release rate of the fire. The heat release rate per unit area is generally the output from small-scale fire experiments and can be applied for larger objects with varying surface area. This paper studies the impact of varying heat release rate per unit area for fires in materials and substances typically found underground. Data from earlier fire experiments were analysed and discussed, and several different heat release rate curves and flame spread velocities for various fuel components were developed and discussed. It was found that the hydraulic hose presented the highest peak heat release rate per unit area of the solid fuel items, followed by the low-voltage cable, cab interior, and tyre in descending order. The increase in the peak heat release rate was highest for the highest incident heat flux levels, representative for scenarios underneath a mining vehicle. The heat release rate per unit area of pool fires underneath a vehicle is significantly higher than a corresponding free burning case - attributed to the higher incident heat flux – but a gravel surface underneath a pool fire will reduce the heat release rate considerably. Line fires in larger bundles of electrical cables were modelled and found to attain the peak flame spread velocity more rapidly compared with a fire in hydraulic hoses. This was caused by the higher heat release rate of the initial fire and ignited segments in the electrical cable case.

Analysis of methods for assimilating fire perimeters into a coupled fire-atmosphere model

Correctly initializing the fire within coupled fire-atmosphere models is critical for producing accurate forecasts of meteorology near the fire, as well as the fire growth, and plume evolution. Improperly initializing the fire in a coupled fire-atmosphere model can introduce forecast errors that can impact wind circulations surrounding the fire and updrafts along the fire front. A well-constructed fire initialization process must be integrated within coupled fire-atmosphere models to ensure that the atmospheric component of the model does not become numerically unstable due to excessive heat fluxes released during the ignition, and that realistic fire-induced atmospheric circulations are established at the model initialization time. The primary objective of this study is to establish an effective fire initialization method in a coupled fire-atmosphere model, based on the analysis of the impact of the initialization procedure on the model’s ability to resolve fire-atmosphere circulations and fire growth. Here, we test three different fire initialization approaches leveraging the FireFlux II experimental fire, which provides a comprehensive suite of observations of the pyroconvective column, local micrometeorology, and fire characteristics. The two most effective fire initialization methods identified using the FireFlux II case study are then tested on the 380,000-acre Creek Fire, which burned across the central Sierra Nevada mountains during the 2020 Western U.S. wildfire season. For this case study, simulated pyroconvection and fire progression are evaluated using plume top height observations from MISR and airborne fire perimeter data, to assess the effectiveness of different initialization methods in the context of establishing pyroconvection and resolving the fire growth. The analyses of both the experimental fire simulation and the wildfire simulation indicate that the spin-up initialization method based on historical fire progression that masks out inactive fire regions provides the best results in terms of resolving the fire-induced vertical circulation and fire progression.

Long-term summer warming reduces post-fire carbon dioxide losses in an arctic heath tundra

The frequency and extent of wildfires in the Arctic has been increasing due to climate change. However, there is a lack of understanding about long-term impacts of climate warming on post-fire carbon dioxide (CO2) exchange in arctic tundra ecosystems. To investigate this, we conducted an experimental low-intensity fire in combination with summer warming (using open top chambers) in a dry heath tundra ecosystem in West Greenland. We report here on the impact four and five years after the fire. We also examined immediate effects of soil heating to three temperature levels (35, 55 and 80 °C), as a simulation of heat transfer during a typical tundra fire. Fire increased soil organic phosphorus concentrations up to at least four years after the burning. The burned areas remained a net CO2 source five years after the fire, mainly due to the lower aboveground vegetation biomass and reduced gross ecosystem production (GEP). However, with four to five years of summer warming, the GEP, ecosystem respiration and soil respiration significantly increased, and burned areas turned into a net CO2 sink. Ex-situ soil heating to the temperature of 55 °C, reaching the heat load comparable with in-situ burning, had minor effects on soil GHG fluxes. This suggests that soil GHG activities are not immediately affected by heat transfer and associated soil temperature increases during a typical low-intensity wildfire in arctic dry tundra. Overall, our results reveal that in a future warmer climate, vegetation is likely to recover more quickly from fires, resulting in a reduction in post-fire CO2 losses.

Experimental study of the temperature characteristics of the main cables and slings in suspension bridge fires

Fire raises key issues in the safety assessment and thermal protection of a suspension bridge, where the temperature profile of the main cables and slings (cable) is an important parameter that has not been quantified in previous studies. The present research investigated its evolution in fire scenarios on the emergency lane and inside lane. A suspension bridge fire experiment platform was built, and 30 sets of pool fire tests simulating vehicle fire were conducted considering the wind velocity, pool size, and pool location. The focus was on the mass burning rate of gasoline and the temperature profile of cables during fires in different lanes. The results showed that the mass burning rate increased with pool size and cross-wind. A new characterization of the gasoline burning rate was introduced. A vehicle fire in the emergency lane posed a greater thermal threat to cables than an inside-lane fire. An exponential model representing the temperature profile of cables was derived. A uniform prediction model for the maximum temperature of cables exposed to different lane fires on a suspension bridge was proposed for the first time. The predicted values of proposed model were compared with measured maximum temperature, and reliable results were obtained.

Corrosion behavior and failure mechanism of ball grid array packages in fire smoke atmospheres

Ball grid array (BGA) packages is a common type of electronic packaging and a crucial component in electronic devices. However, its reliability is significantly compromised when exposed to fire smoke atmospheres, thereby reducing the overall reliability of electronic equipment. Therefore, the study of failure mechanisms of BGA packages in fire smoke environments is of paramount importance for the recovery and maintenance of electronic devices post-fire. In this study, we conducted fire smoke experiments and employed electrical performance measurements along with chemical characterization techniques to delve into the failure behaviors and mechanisms of BGA packages in different fire smoke atmosphere. The results reveal that nitrile-butadiene rubber and polycarbonate smoke do not affect the electrical conductivity of BGA packages. However, the accumulation of smoke particles and water between the interconnecting solder joints temporarily reduces the insulation performance. That is induced by the deposition of smoke particles and can be restored after cleaning. On the other hand, under polyvinyl chloride smoke, BGA packages experiences insulation failure and permanent open-circuit failure. This irreversible permanent open-circuit failure is caused by the interconnection of corrosion products in solder joints. Chemical characterization analysis results indicate that solder joints dissolve due to the galvanic corrosion mechanism, transforming the solder from a spherical shape into a rose-like structure. The corrosion products primarily consist of flake-like SnO and Sn3O(OH)2Cl2. This research contributes to a deeper understanding of the failure mechanisms of electronic devices in fire smoke and provides essential guidance for post-fire equipment maintenance.