In this study, an ISO 5659-2-based smoke density chamber [5] and an ISO/TS 5660-4-based cone calorimeter [7] were used to compare smoke generation behavior and resulting optical properties under different measurement environments. Figure 1 presents schematic diagrams of the two apparatuses: Figure 1(a) represents the combustion environment under static conditions and Figure 1(b) represents that under dynamic conditions. Both apparatuses use optical measurement systems to measure the light intensity attenuated by smoke and determine light transmittance from the ratio of the attenuated light intensity to the reference intensity before smoke generation. The relationship between light transmittance and the extinction coefficient, k, is described by Bouguer’s law [23], as shown in Eq. (1).

where k is the extinction coefficient (m^-1), L is the optical path length through which smoke passes (m), I is the light intensity reduced by smoke, and I₀ is the light intensity in the absence of smoke. Therefore, the extinction coefficient considered in this study was calculated by accounting for the optical path length.

The smoke density chamber consists of a cone heater that provides radiant heat, a specimen holder positioned beneath the cone heater, a light source, and an optical measurement system. The radiant heat flux applied to the combustible surface of the material was controlled by adjusting the cone heater temperature, and the resulting smoke accumulated inside the chamber. During the experiment, a 6.5 V monochromatic light source at the bottom of the chamber transmitted light upward, while an optical detector at the top measured the reduction in light intensity caused by smoke to calculate the time-dependent light transmittance, T(t). The extinction coefficient, k(t), was then determined using Eq. (1), with an optical path length of 0.915 m.

The cone calorimeter consists of a cone heater, a specimen holder, a smoke collection duct, and a blower for exhaust flow rate control. Radiant heat flux control was the same as in the smoke density chamber. The generated smoke was collected in the duct and discharged under a constant exhaust flow condition of 18 L/s. Smoke concentration in the duct was measured using a laser-based optical attenuation measurement system. Light transmittance, T(t), was calculated from the light intensity received by the detector after laser light passed through the optical path. The extinction coefficient, k(t), was then derived using Eq. (1) based on the same principle as in the smoke density chamber, with an optical path length of 0.118 m.

In this study, wood, a typical combustible material in indoor fires, was selected as the experimental material. Wood forms a char layer during combustion, allowing a smoldering stage to occur [24], which enables analysis of smoke generation and resulting changes in optical properties. Accordingly, experiments were conducted using Douglas fir wood [25], which is widely used in furniture, interior finishing materials, and various outdoor construction materials, making it highly representative of real fire scenarios. Figure 2 shows the geometry of the Douglas fir specimens used in this study.

Typically, specimens for ISO 5660-4 experiments are prepared with dimensions of 100 mm × 100 mm. However, in this study, specimens with dimensions identical to those used in ISO 5659-2 experiments were used to ensure that differences in measurement environments between the two apparatuses were the only variable. The specimen dimensions were 75 mm × 75 mm × 24 mm, and the specimens were wrapped with aluminum foil so that the area directly exposed to the cone heater was 65 mm × 65 mm. A standard radiant heat flux of 25 kW/m² was applied, and no pilot ignition source was used during the experiments. Each experiment was conducted for 30 min. Prior to testing, all specimens were conditioned for at least 24 h at a temperature of 23 °C and a relative humidity of 50% to ensure consistent specimen conditions. To ensure the reliability of the results; repeated experiments were conducted at least three times under each condition. The results are presented as the average values of repeated measurements, and the standard deviations are indicated as shaded regions in the graphs.

This section presents smoke characteristic indices derived from light transmittance and the extinction coefficient to compare smoke generation behavior between the apparatuses. Subsequently, the temporal behavior and representative values of the extinction coefficient are comparatively analyzed in the following section. For the smoke density chamber, light transmittance is measured as smoke accumulates, allowing smoke concentration in an enclosed space to be represented over time in terms of optical density. The specific optical density, D_s, reflecting the geometry of the smoke density chamber, is defined by Eq. (2).

where, V is the volume of the smoke density chamber (m³), is the exposed area of the combustible material (m²), and L is the optical path length (m). T is the light transmittance (%), and F is an optical filter correction term. The optical filter is used when the light transmittance falls below 1%; F is set to 0.3 when the filter is used and 0 when it is not. In the cone calorimeter, the extinction coefficient, k(t), is determined from the light transmittance measured in the duct section, and smoke generation characteristics are quantified in terms of smoke production using the exhaust volumetric flow rate, V˙_e(t), which reflects smoke flow within the duct. The smoke production rate (SPR) is defined by Eq. (3).

In addition, the cumulative amount of smoke generated during the experiment is defined as the total smoke production (TSP), which is calculated by time integration of SPR, as expressed in Eq. (4).

where SPR(t) is the smoke production rate (m²/s), k(t) is the extinction coefficient (m^-1), V˙_e(t) is the duct exhaust volumetric flow rate (m³/s), and TSP is the total smoke production (m²).

Figure 3 shows the temporal variation in specific optical density under static conditions. Figure 4 presents the smoke generation behavior over time under dynamic conditions, with Figure 4(a) showing the SPR and Figure 4(b) showing the TSP. As shown in Figure 3, under static conditions, the specific optical density increased rapidly during 0-400 s and then remained nearly constant from 400-1,800 s. Because smoke generated under static conditions accumulates inside the chamber without being discharged, directly interpreting combustion behavior solely based on changes in optical density is limited. Therefore, additional measurements, such as mass loss rate or combustion gas concentration, are required. In contrast, under dynamic conditions, as shown in Figure 4(a), the smoke production rate increased and then decreased during 0-400 s, followed by a second increase and decrease from 800-1,400 s, exhibiting two peak values. To interpret the combustion behavior of Douglas fir wood, thermal thickness was evaluated by calculating the thermal penetration depth (δ_p=A at)) using thermal diffusivity (k/ac_p) and exposure time (t). The thermal diffusivity [25] and coefficient [26] were obtained from previous studies, and because ignition was not observed, the evaluation was based on the value at the end of the experiment (1,800 s). The calculated thermal penetration depth was much smaller than the specimen thickness, indicating that Douglas fir exhibited thermally thick behavior under these conditions. When thermally thick charring materials are exposed to radiant heat, combustion reactions begin through pyrolysis, and heat transfer into the interior occurs actively. However, pyrolysis becomes suppressed due to char layer formation and thermal thickness effects. As combustible material consumption continues, the reacting surface approaches the rear surface of the specimen, resulting in a rapid increase in pyrolysis and combustion activity [25]. Therefore, the two peaks in smoke production rate observed under dynamic conditions in this study can be interpreted in relation to the staged pyrolysis and combustion behavior of thermally thick charring materials. In addition, the smoke production rate from 800-1,400 s was greater than during 0-400 s. This is considered to result from smoke initially being generated primarily by combustion of combustible gases, followed by increased smoke generation due to smoldering reactions of the char layer combined with combustible gas combustion as char formation progressed. Furthermore, as shown in Figure 4(b), the total smoke production increased over time, and its growth rate increased during 800-1,400 s, confirming that the second peak in smoke production rate contributed dominantly to cumulative smoke generation.

Figure 5 shows the variation in extinction coefficient calculated from light transmittance measured under static and dynamic conditions. As with the specific optical density and smoke production characteristics discussed in Section 3.1, clear differences were observed depending on the combustion environment. Under static conditions, the extinction coefficient increased rapidly during the initial stage of the experiment and then remained nearly constant. This indicates that smoke generated within the sealed chamber was not discharged, resulting in a rapid increase in smoke particle concentration and sustained high levels of light extinction due to smoke accumulation. In contrast, under dynamic conditions, the extinction coefficient increased and then decreased as smoke concentration varied over time due to exhaust flow. In addition, two distinct peaks were observed depending on combustion behavior. The maximum extinction coefficient under static conditions was approximately 5 to 15 times greater than those under dynamic conditions (first and second peaks), suggesting that smoke accumulation can have a dominant influence on the intensity of light extinction.

The extinction coefficient is an indicator representing the degree of light attenuation as light passes through a smoke layer and serves as a key parameter for quantifying visibility reduction. In fire simulations, the extinction coefficient is calculated as the product of the mass specific extinction coefficient, k_m, and the smoke mass concentration, ρY_s=ρ_s [21]. Therefore, mass-based smoke generation data, such as soot yield, are required as input parameters for fire simulations. Although the extinction coefficient can be experimentally determined under both static and dynamic conditions, the smoke density chamber under static conditions is an accumulation-based test method, making it difficult to obtain exhaust flow information and corresponding mass-based generation data, which limits soot yield estimation. In contrast, the cone calorimeter under dynamic conditions includes exhaust flow, allowing soot yield to be determined by combining the extinction coefficient divided by the mass specific extinction coefficient with exhaust flow rate and fuel consumption data. Accordingly, these values have been used as input parameters in fire simulations. However, as confirmed in this study, when the same combustible material was exposed to the same radiant heat flux, the extinction coefficient measured under static conditions was higher than that measured under dynamic conditions. This suggests that using only input parameters established under dynamic conditions to represent visibility reduction during fires may introduce uncertainty. Therefore, further investigation is needed to generalize extinction coefficient values derived under both static and dynamic conditions for use as simulation input parameters.