Impact of Combustible Linings in the Simulated Fluid Dynamics of a Compartment Fire

Abstract

The increasing use of engineered timber in modern architecture raises critical concerns about fire safety, particularly when combustible linings are exposed within compartments. Classical compartment fire framework, largely derived from non-combustible enclosures, may not adequately capture the dynamics introduced by materials such as cross-laminated timber (CLT). This study investigates how combustible linings influence the fluid dynamic fields of compartment fires derived from the thermal field using CFD simulations informed by experimental data. A series of configurations, from inert to fully lined compartments, were analysed to isolate the effect of burning boundaries. Results show a progressive intensification of fire conditions with additional combustible surfaces: upper-layer temperatures approach 900 °C, smoke layers thicken, and stratification becomes more pronounced. Velocity fields are similarly affected, with peak inflow and outflow velocities doubling compared to the inert case and new vortical structures emerging near burning walls. These findings highlight that exposed CLT significantly amplifies radiative and convective heat feedback, modifying both temperature distributions and flow patterns in ways not captured by the traditional framework based on the inverse opening factor. This underscores the need for performance-based fire design approaches integrating both thermal and fluid dynamic perspectives, ensuring safe implementation of timber in modern construction.

1. Introduction

The structural use of engineered timber has experienced a global increase, driven by efforts towards a more sustainable construction industry [1]. However, this trend also faces strong concerns regarding fire safety, which represent a major barrier to its broader adoption. A key challenge associated with timber is its inherent combustibility [2]. In many jurisdictions, the use of combustible enclosures is restricted or even prohibited in certain building types and heights.

At the same time, contemporary architecture has promoted open-plan layouts where structural elements are exposed within interior spaces. Exposed combustible construction materials in large open spaces challenge some of the fundamental assumptions embedded in the classical compartment fire models [3,4]. Since these models form the basis of major international codes and standards [5,6], it becomes essential to evaluate designs through Performance-Based Design (PBD) methodologies [7]. The proper application of PBD requires a deeper understanding of compartment fire behaviour and a clear quantification of how new construction technologies influence fire dynamics.

Since the 1960s [8,9,10], the single-opening compartment fire configuration has been widely used to study the interaction between fire and structural elements. Confined fire behavior has generally been classified into two fully-developed regimes: Regime I (ventilation-controlled) and Regime II (fuel-controlled) [9,11,12]. In Regime I, fire dynamics are governed by the size of the opening, with uniform compartment temperatures. In Regime II, there is sufficient oxygen available, and the fire dynamics are therefore controlled by the size of the fire (i.e., fuel availability), leading to non-uniform compartment temperatures.

Nevertheless, several authors have examined in greater depth the meaning of these regimes and their implications for different geometric configurations, window sizes, and the contribution of combustible linings. Torero et al. [13] observed that, in the case of the WTC collapse, the structural effects caused by the fire were consistent with Regime II behaviour, exhibiting temporal and spatial temperature distributions characteristic of fuel-controlled conditions in open-plan offices. Furthermore, due to the phenomenon of travelling fires [14], the fire imposed structural demands at different times and locations within the structure, challenging the assumption that Regime I necessarily represents the worst-case scenario. Along similar lines, Majdalani [15] demonstrated through scaling analysis that the regimes defined by Thomas et al. [9] should be understood as limiting cases within a broader spectrum of possible behaviours. This wider spectrum of compartment fire scenarios highlights that fire dynamics extend beyond the traditional regime classification and cannot be determined solely by the inverse opening factor ($\varphi$ as defined in Equation Equation (1)) across all compartment sizes and geometries.

$$\varphi ={\displaystyle \frac{\mathrm{Total}\mathrm{area}\mathrm{for}\mathrm{heat}\mathrm{losses}}{\mathrm{Ventilation}\mathrm{factor}}}={\displaystyle \frac{A_T}{A_w\sqrt{H}}}$$

(1)

Although this factor establishes a relationship between heat losses and heat gains, the variation of the flow field across regimes means that assumptions valid in one regime are not applicable in the other. Consequently, to characterise the regime of behaviour in a compartment fire with combustible lining (e.g., exposed timber), it is necessary to analyse how the interrelated variations of the flow and thermal fields combine.

Cross-Laminated Timber (CLT), a mass timber product composed of multiple layers of wood bonded perpendicularly, offers significant construction advantages such as lightness and rapid assembly, which reduce costs and execution times. Nevertheless, the combustible nature of wood suggests that current compartment fire models and correlations—primarily based on non-combustible enclosures—may not be applicable to CLT compartments. This is due to the considerable increase in fuel load and its redistribution within the compartment. In compartments with exposed CLT, combustible linings can ignite and drastically increase the heat release rate (HRR), which in turn can accelerate the combustion of the compartment’s contents and other combustible surfaces, directly influencing fire dynamics.

Experimental data from various campaigns at different scales support this observation. Butcher [16] conducted fire tests with combustible wall linings, finding that fibre insulation board (FIB) increased the amount of internal flames and sustained their presence. In similar conditions, Li [17] reported that, depending on the configuration of the combustible walls, secondary flashover or self-extinction can occur.

Despite these valuable contributions, it is important to note that most experimental studies primarily report thermal variables such as gas-phase temperatures, heat fluxes, or heat release rates. The velocity field, however, remains largely undiscussed in the literature. This omission is not due to a lack of relevance—indeed, velocity governs plume entrainment, smoke layer formation, and ventilation-driven flows—but rather to the intrinsic difficulty of obtaining reliable velocity measurements under fire conditions. Intrusive probes can significantly disturb the flow, while non-intrusive optical diagnostics, such as Particle Image Velocimetry (PIV), are hindered by the hostile environment of compartment fires, where high temperatures, soot, and turbulence compromise both optical access and seeding quality [18]. As a result, the term fire dynamics is often used in practice to refer to thermal responses and fire spreading, overlooking the coupled role of fluid motion [19]. This narrowed interpretation may limit the understanding of how buoyancy-driven flows, recirculation zones, and boundary layer effects influence the overall fire dynamics.

To address these limitations, Computational Fluid Dynamics (CFD) is an indispensable tool for the study of fire dynamics in its entirety. By numerically solving the governing conservation equations, CFD models such as the Fire Dynamics Simulator (FDS) [20] enable the simultaneous analysis of both the thermal and velocity fields, offering a level of detail unattainable by experimental methods alone. This capability allows for the identification of flow structures such as vortices, neutral planes, and velocity layers, that play a critical role in entrainment and smoke transport. Moreover, CFD provides the possibility of testing scenarios involving combustible linings where the interplay between pyrolysis-driven fuel release and ventilation conditions produces highly coupled thermo-fluid phenomena. Several authors have demonstrated that CFD-based studies can extend and complement experimental campaigns, bridging the gap between measurable quantities (temperatures, HRR) and the dynamic processes that control fire spread and compartment conditions [21]. Consequently, integrating CFD with experimental approaches is not only beneficial but necessary to achieve a holistic understanding of compartment fire behaviour, particularly in the context of modern construction materials such as timber linings, where the velocity field plays a decisive role.

This study contributes to the understanding of compartment fire dynamics by examining how combustible linings influence not only thermal conditions, but also the fluid dynamics. Using CFD simulations together with experimentally derived boundary conditions, the work shows that exposed CLT surfaces intensify buoyancy-driven flows, modify smoke layer development, and reshape recirculation patterns inside the compartment. These results highlight that combustible boundaries can alter the balance between fuel load, ventilation, and geometry in ways not captured by traditional opening factor frameworks, underscoring the need to refine current compartment fire models and to integrate both thermal and fluid dynamic perspectives into performance-based design strategies for timber structures.

2. Material and Methods

Computational Model

FDS version 6.7.9 [20], is used in this study. FDS is a CFD code that has been extensively validated for fire modelling. It solves a low-Mach-number form of the Navier–Stokes equations, suitable for buoyancy-driven flows. A complete description of its governing equations and submodels is provided in the technical reference guide [20].

Table 1. Simulation Parameters and their respective value.

Simulation Parameter	Value
Cell size	Far field 0.02 m × 0.02 m × 0.02 m
	Near field 0.01 m × 0.01 m × 0.01 m
Time	750 s
Fuel	Propane
Heat of Combustion	46.45 kJ g−1
Combustion Efficiency	0.9
Soot yield	0.01

, summarise the main parameters and the fixed values used in the simulations. Unless otherwise stated, default settings were used, and only explicitly mentioned parameters were modified.

The mesh size was selected following the FDS User’s Guide [20] criteria $\delta \le 0.1·D^{*}$, where $\delta$ is the edge length of the cubic mesh cell and $D^{*}$ is defined according to Equation Equation (2):

$$D^{*}={\left({\displaystyle \frac{\dot{Q}}{{\rho }_{\infty }{c}_p{T}_{\infty }\sqrt{g}}}\right)}^{2/5},$$

(2)

where $\dot{Q}$ is the heat release rate, ${\rho }_{\infty }$ is the air density, $c_p$ is the specific heat capacity of the gas, $T_{\infty }$ is the ambient temperature, and g is the acceleration due to gravity.

The influence of combustible linings on the velocity field was examined using the experimental configurations of Hadden et al. [22] (the experimental setup is shown in Figure 1). In those experiments, the lining material was cross-laminated timber (CLT), made of commercial five-layer spruce panels (100 mm thick) bonded with polyurethane adhesive. Non-exposed surfaces were protected with two layers of 12.5 mm gypsum plasterboard. From the Alpha-2 configuration onwards, a 25 mm layer of mineral wool was added behind the plasterboard to improve encapsulation. The total heat release rate (${\mathrm{HRR}}_{\mathrm{total}}$) was measured by oxygen consumption calorimetry, and crib mass loss was recorded using load cells.

From these data, the contribution of the CLT lining (${\mathrm{HRR}}_{CLT}$) to the total HRR was quantified. Figure 2 shows the average ${\mathrm{HRR}}_{CLT}$ for the Alpha configuration. For simplicity, both exposed walls were assumed to contribute equally. The heat release rate per unit area (HRRPUA) was obtained with Equation Equation (3), relating the HRR of each surface to its exposed area. On this basis, each exposed wall (or ceiling, depending on the case) contributed about 168 kW m⁻². This assumption is according to Hadden et al. [22], across all experimental configurations the contribution from the walls is of the same order of magnitude (in terms of total energy).

$$\mathrm{HRR}=\mathrm{HRRPUA}·A$$

(3)

The compartment used in these simulations has dimensions of 0.82 m × 0.96 m × 0.82 m (width × length × height) with an opening spanning the entire front, as shown in Figure 3a. This corresponds to the same scenario validated by [23]. The model includes 54 thermocouples arranged in nine vertical trees, each with six sensors at different heights, replicating the experimental distribution (Figure 3b).

The fire source model follows the analogy proposed in [23], due to mesh resolution constraints. Consequently, the original setup (120 holes with diameters of 2 mm) was replaced by 16 square burners while maintaining the same effective area. Therefore, a prescribed HRRPUA of 3769 kW m⁻² was imposed in order to obtain the target heat release rate of 24 kW reported by Majdalani [15] during their experiments.

The inert case reproduces one of the experimental setups from Majdalani [15], already validated in earlier work by the authors [23]. For this case, the compartment has 100% opening. This ensures reliable boundary conditions and consistency with experiments. Using an experimentally validated case permitted focusing on how burning walls and ceilings affect the velocity field.

Direct simulation of burning CLT surfaces was not attempted because of the high computational cost and complexity of modelling pyrolysis, charring, char combustion and delamination. Instead, the experimentally derived HRRPUA of exposed CLT surfaces was imposed through an equivalent gas burner. This simplification preserved the total heat release while reducing model complexity, thus making it possible to study the flow dynamics. Although this method does not capture detailed combustion chemistry or material degradation, it ensures the correct thermal input and reproduces the main buoyancy- and momentum-driven flow structures. Therefore, direct comparison of temperatures with experiments is not meaningful, but the method is valid for the main objective of this study: understanding the flow modifications induced by burning walls or ceilings relative to an inert compartment.

3. Results and Discussion

In this section, the main results of this study are presented, organized into two parts: thermal field and velocity field.

Figure 4 illustrates the variation in the heat release rate (HRR) as a function of the amount of combustible linings included in each configuration.

As shown in Figure 4, a clear increase in HRR is observed as the exposed timber surface area increases. This trend is both qualitatively and quantitatively consistent with previous findings reported by [2,22,24].

3.1. Thermal Field

Figure 5 presents the time-averaged temperatures during steady-state conditions for each thermocouple in each tree. In parallel, Figure 6 shows temperature slices at 500 s, highlighting the profile of the smoke layer, which is defined at locations where the average temperature exceeds 200 °C.

In Figure 5a, the inert case is shown. Based on the relationship between available fuel and ventilation–and considering the compartment’s size and shape–this case falls within a Regime II compartment fire behaviour.

This behaviour is consistent with the experiments of Majdalani et al. [12] and reflects the strong influence of ventilation: the large opening supplies significant amounts of fresh air, cooling the hot gases. Above 0.2 m, the temperature remains nearly constant (and low), indicating the absence of a well-developed smoke layer, or that the layer is too thin and attached to the ceiling. This is consistent with Figure 6a, where the red line lies above the compartment, showing that, on average, internal temperatures remain below the threshold used to define smoke layer height. Some localised exceedances are observed near the rear burners, but these occur close to the floor and are therefore attributed to burner position rather than stratification.

For alpha-(Figure 5b) and gamma-(Figure 5d) configurations, the temperature profiles are similar up to 0.4 m. Above this height, however, the gamma-configuration shows a thicker hot layer compared to the alpha case, consistent with the additional burning surface (ceiling) contributing soot and flames [2]. This effect becomes more evident in Figure 6d, where the hot layer is noticeably thicker than Figure 6b, particularly near the opening. This difference can be attributed to the ceiling flames pushing the flow downward and toward the outlet, thereby producing a thicker hot layer.

Figure 5c illustrates the vertical temperature profiles for the beta-configuration (rear wall and ceiling burning). Unlike the inert case, which exhibited moderate and nearly uniform temperatures above 0.4 m, the beta-configuration shows considerably higher temperatures across the profile, with several thermocouples exceeding 400 °C and peaks nearing 800 °C. Stratification is more evident, particularly between 0.3 m and 0.6 m, where the transition from cooler to hotter gases defines a clearer smoke layer interface. This behaviour is linked to the additional burning surfaces, which significantly increase the heat release rate (HRR) and intensify buoyancy-driven flows, thereby producing a hotter upper smoke layer.

When comparing Figure 5b,c, it becomes evident that, although both configurations present the same fuel surface, the resulting thermal fields differ markedly. This contrast highlights the influence of symmetric versus non-symmetric flow structures on the thermal environment. In the symmetric case, the flow is displaced more efficiently, leading to a reduction in compartment hot layer accumulation (thickness) and temperatures.

Finally, the delta-configuration (Figure 5e), with all four boundaries burning (rear wall, ceiling, and both lateral walls), produces the most severe conditions. Upper-layer temperatures again reach nearly 900 °C, but unlike the gamma-configuration, significant heating extends much closer to floor level as can be seen in Figure 6d. The combined radiative and convective feedback from all burning surfaces generates a nearly continuous hot gas layer throughout the compartment, leaving only a very limited cooler zone at the bottom.

In Figure 7, the maximum average thermocouple temperature over 0.4 m (hot layer) is plotted for each scenario. The results highlight the role of flow symmetry in shaping the thermal field. A symmetric flow pattern promotes a more efficient removal of the hot gas layer, resulting in lower compartment temperatures for an equivalent burning surface. Conversely, as additional surfaces are ignited, the overall temperature field increases apparently reaching a maximum as the combustion zone moves closer to the opening; however, the extent of this rise remains strongly dependent on the prevailing flow configuration. Additionally, the increase in overall HRR in the delta case tends toward an under-ventilated condition, causing part of the combustion to occur outside the compartment. This, in turn, leads to a decrease in the temperature inside the compartment [25].

In summary, the sequence from inert to alpha-, beta-, gamma-, and delta-configurations highlights the progressive influence of combustible boundaries on temperature distribution. Progressively increasing the internal surface of CLT within a constant (small) compartment and fuel configuration raises the average compartment gas temperature up to a peak value (alpha-configuration), after which it begins to decline, as schematically illustrated in Figure 7. This turning point corresponds to a shift in the compartment’s heat balance.

3.2. Flow Field

The flow field is visualised in Figure 8 and Figure 9. Figure 8 shows the time-averaged velocity streamlines over the last 100 s to ensure steady-state conditions in a vertical plane perpendicular to the opening. Figure 9 shows the time-averaged velocity streamlines over the last 100 s for a vertical plane, this time parallel to the opening.

As discussed in [12,23], the streamlines shown in Figure 8a are consistent with the classical definition of a regime II fire behaviour, given the presence of non-negligible vertical velocities. Nevertheless, due to the nearly cubic geometry of the compartment, a drastic change in flow direction can be observed. This phenomenon, referred to as the backwall effect, has been further investigated through computational simulations in [23] and theoretically in [15].

The flow field shown in Figure 9a is characterised by a central upward plume driven by buoyancy, which impinges on the ceiling and spreads laterally towards the side walls. This redistribution generates recirculating vortices in the upper corners and return currents near the floor, indicating strong mixing between hot gases and entrained cooler air. The relatively symmetric pattern reflects the influence of the nearly cubic geometry on the flow structure. Velocity magnitudes are highest in the plume core, while recirculation zones exhibit lower values. These features are consistent with regime II behavior, where vertical momentum dominates the flow dynamics, and they provide a reference case for comparison with combustible-lining compartments, where wall and ceiling burning is expected to modify both pattern and intensity of the velocity field.

The flow field characteristics displayed in Figure 8a and Figure 9a are consistent with a Regime II behaviour, and this case will be adopted as the reference behaviour for the subsequent analysis of the other compartment configurations.

Figure 8b and Figure 9b show the time-averaged velocity streamlines for the Alpha configuration. Compared to the inert case, the central vortex in the YZ plane shifts downward, indicating that the neutral plane lies closer to the floor. In addition, the burning of the back wall exerts a direct influence on the magnitude of the velocity field. Peak velocities increase from approximately 1.75 m s⁻¹ to nearly 3.5 m s⁻¹, effectively doubling the flow speed compared to inert-configuration. This intensification is attributed to the additional momentum imparted by the flames from the back and side walls, which also enhances the entrainment of fresh air into the compartment. As a result, the inflow velocities at the lower boundaries are almost twice those observed in the inert case. In the XZ plane, due to the presence of a burning lateral wall, the flow field exhibits a lateral displacement, resulting in a non-symmetric overall flow pattern. A large vortex forms near the burning wall at a height of approximately 0.5 m, which coincides with the smoke layer height observed in Figure 6b. This structural modification of the flow field highlights that combustible linings not only enhance the velocity magnitudes but, depending on their spatial location within the compartment, can also reconfigure the overall flow pattern, with direct implications for smoke-layer stratification and the distribution of thermal energy.

For the beta-configuration–Figure 8c and Figure 9c–the overall flow pattern remains symmetrical and qualitatively similar to the inert case. In the YZ plane, the main difference lies in the velocity magnitudes, which increase by nearly 120%, accompanied by a proportional rise in entrainment. This suggests that the contribution of the burning rear wall enhances the air inflow to the compartment. In the XZ plane, the fluid motion initially resembles the inert configuration, with buoyancy forces driving an upward plume through the central region of the compartment. However, a stagnation zone develops near the upper region between 0.7 m to 0.8 m, which coincides with the thickness of the smoke layer identified in Figure 6c. This indicates that, while the overall compartment flow structure is symmetrically preserved, the presence of non-lateral combustible linings modifies the stratification of the upper gas layer and intensifies local velocity magnitudes. Although the beta configuration involves a burning ceiling, the smoke layer does not reach the thickness reported in other experimental campaigns [2,26]. This discrepancy can be attributed to the presence of the large floor-to-ceiling opening, which drives the flow outward. The outflow momentum is further reinforced by the contribution of the burning rear wall, resulting in a continuous discharge of hot gases and preventing the development of a stable stagnation region. Consequently, the smoke layer remains thinner than expected despite the additional fuel contribution from the ceiling.

For the gamma-configuration–Figure 8d and Figure 9d–the overall flow pattern is similar to that observed in the alpha-configuration. The main difference between these two cases lies in the height of the neutral plane. This feature is more clearly visible in Figure 9d, where the stagnation vortex forms lower than in the alpha case. This downward shift can be attributed to the burning ceiling, which drives the flow downward, resulting in velocity neutralisation closer to the floor level.

In the delta-configuration, the overall compartment flow structure is also symmetrically preserved as in the inert and beta cases. The streamlines in Figure 8e still exhibit both vertical and horizontal components; however, the neutral velocity plane forms deeper inside the compartment. In this case, there is also a clear increase in the magnitude of both inflow of fresh air and outflow of hot gases, with the latter reaching velocities up to 5 m s⁻¹. In the XZ-plane (Figure 9e), the flow pattern reveals a general downward motion, except near the lateral walls where the flame-induced momentum drives the gases upward towards the ceiling. This interaction generates two large vortices at approximately 0.3 m, indicating that the neutral plane is located around this height. This observation is consistent with the streamline behaviour in Figure 8e and with the smoke layer height reported in Figure 6e. Among all configurations, this scenario-—with the largest number of burning surfaces-—produces the highest overall velocities.

In summary, the analysis of the flow field across the different configurations reveals that, much like the thermal field, the velocity structure is strongly influenced by the location and extent of the burning surfaces. A particularly noteworthy observation is the similarity between the alpha- and gamma-configurations: despite the additional burning ceiling in the gamma-configuration, both scenarios produce comparable flow patterns. Likewise, the inert, beta and delta scenarios produce similar flow patterns, with the latter invigorated by the burning walls, all 3 of them significantly different from the alpha- and gamma-configurations.

These observations underscore the relevance of the burning surfaces symmetry–with respect to the opening–in shaping the compartment’s overall flow pattern which subsequently drives substantial changes in the thermal field in a reciprocal manner, promoting changes in the regime of behaviour.

3.3. The Inverse Opening Factor (IOF) Breakpoint

The original framework defined the breakpoint between Regimes I and II not by the opening factor, but by the relationship between fuel load and ventilation, through the fuel surface area and the ventilation parameter, respectively. The later use of Equation Equation (1) to indicate a transition point between regimes was simply a by-product of graphically describing the correlation between average gas temperatures and the inverse opening factor under very specific conditions: a small, cubic, non-combustible compartment with a single vertical square opening, fully fuel-covered floor, and crib arrangement CIB [9]. Thus, the inverse opening factor breakpoint value reported by Thomas et al. [9] after this graph cannot be taken as a universal criterion for defining a compartment’s regime of behaviour. The transition between regimes of behaviour is not, therefore, universally defined by the compartment’s inverse opening factor (IOF), but rather by the amount of exposed fuel surface area in relation to the available ventilation—i.e., the mass flow rate approximated by the ventilation factor–and its driving mechanism. Both variables are strongly influenced by the compartment and opening characteristics, as well as by the distribution and configuration of the combustible package.

For a similar set of nearly-cubic compartments and opening configurations, but with exposed timber linings, Gorska et al. [2] proposed a modification to the inverse opening factor to account for the presence of combustible boundaries. This adjusted form is presented in Equation Equation (4).

$${\varphi }_G={\displaystyle \frac{A_T-A_{CLT}}{A_w\sqrt{H_w}}}$$

(4)

Table 2. Compartment configurations and their opening factors (traditional and adjusted [2]).

Configuration	${\mathit{A}}_{\mathit{T}}$ (m2)	$\mathit{\varphi }$ (m−1/2)	${\mathit{\varphi }}_{\mathit{G}}$ (m−1/2)	$\mathit{\varphi }/{\mathit{\varphi }}_{\mathit{G}}$
inert	3.034	4.98357	4.98357	1.000
alpha	1.5744	4.98357	2.5857	0.5188
beta	1.5744	4.98357	2.5857	0.5188
gamma	0.7872	4.98357	1.2928	0.2594
delta	0.7872	4.98357	1.2928	0.2594

summarises the IOF and the modified IOF for all configurations analysed here. Further, Figure 10a,b plot temperature vs. IOF [9], and normalised temperature vs. normalised and modified IOF proposed by [2], respectively, exhibiting their own experimental set breakpoint behaviour.

Following these graphs and the IOF and IOF-modified values from Table 2, it is evident that all configurations would fall within a Regime II compartment fire behaviour. Nevertheless, it has been clearly shown that the temperature and flow fields, as well as the consequent neutral plane location, vary widely between the different configurations, explicitly indicating different regimes of behaviour. This is to say, different ventilation modes render varying flow patterns, which combined with different burning modes consequence of a non-uniform thermal exchange, give as a result different regimes of behaviour as anticipated by [15]. It can therefore be inferred that the inclusion of combustible boundaries not only increases the HRR, enhancing the temperature and flow fields, but also drives the flow to either a symmetrical or a non-symmetrical pattern that affects the energy distribution, giving as a result different regimes of behaviour.

This last point is crucial, as Majdalani [15] also observed these hybrid behaviours when varying the opening size. Accordingly, these computational simulations, together with the conclusions drawn by Majdalani and previous studies [23], suggest that the exclusive use of the IOF to define a compartment’s regime of behaiour is misleading. Instead, new criteria must incorporate additional variables such as the ratio between fuel load and ventilation size, compartment geometry, and the amount and arrangement of combustible boundaries, among others.

Transitions between compartment fire regimes cannot be explained solely by variations in compartment configuration (i.e., size, shape, and opening characteristics) expressed through the IOF. Instead, they arise from the combined and interdependent variations of the flow and thermal fields. Consequently, the breakpoint between regimes is not universally determined by the IOF, but rather by the amount of exposed fuel, $A_f$, and the associated burning rate, in relation to the ventilation conditions or mass flow rate and its driving mechanism. Both parameters are influenced not only by the geometry of the compartment (size and shape) and the characteristics of the opening (size, shape, orientation, and relative position), but also by the distribution of the internal combustible linings and the configuration of the combustible package at floor level.

4. Conclusions

Compared with the inert configuration, which exhibited a classic Regime II compartment fire behaviour, the simulations demonstrated a progressive intensification of the HRR and a corresponding enhancement of both the thermal and flow fields as the extent of combustible linings increased. This enhancement produced a thicker hot layer and a lower neutral plane. Notably, symmetrical addition of combustible linings resulted in a symmetric flow pattern, whereas asymmetric placement generated a non-symmetric pattern, critically modifying the thermal field.

Overall, the combined influence of pressure-difference ventilation and flame-induced flows, together with the uniform or non-uniform distribution of energy exchange within the compartment, led to distinct regimes of behaviour. Consequently, the IOF cannot be regarded as a universal criterion for regime classification. While historically convenient, the IOF breakpoint represents only a subset descriptor of fire dynamics. In reality, regime transitions emerge from the relationship between exposed fuel in relation to the ventilation conditions and driving mechanisms and, therefore, regime behaviours emerge from the coupled interaction of ventilation and burning modes, both governed by compartment geometry, opening configuration, plume dynamics, and fuel distribution. This balance is further shaped by the pressure field and plume entrainment, underscoring the need to evaluate flow dynamics in conjunction with thermal exchange.