4.1 Thermal Environment and Qualitative Discussion of Fire Dynamics
The instructor sampling locations were exposed to less severe conditions when using similar training fuel materials in the multi-compartment concrete training structure compared to the temperatures reported by Stakes et al. in a single compartment, metal container-based Fire Behavior Lab [
44]. First floor temperatures measured in the multi-compartment structure at the instructor location with
Pallet,
OSB, and
Fiberboard fuel packages (medians 76°C to 83°C) were lower than those from the Fire Behavior Lab front instructor location (medians 87°C to 102°C). Likewise, the second-floor instructor location temperatures (medians 45°C to 49°C) were lower than those from the Fire Behavior Lab rear instructor location (medians 60°C to 72°C). Similarly, heat fluxes at the first-floor instructor location (medians 1.7 kW/m
2 to 2.5 kW/m
2) were lower than those from the Fire Behavior Lab front instructor location (medians 5.1 kW/m
2 to 8.4 kW/m
2), and second-floor instructor location heat fluxes (medians 0.4 kW/m
2 to 0.6 kW/m
2) were lower than those from the Fire Behavior Lab rear instructor location (medians 1.4 kW/m
2 to 2.1 kW/m
2). Median and peak air temperatures and heat flux values were classified as
Ordinary operating conditions (72°C to 200°C and 2 kW/m
2 to 12 kW/m
2) on the first floor and
Routine operating conditions (20°C to 72°C and < 2 kW/m
2) on the second floor according to Utech’s classification scheme as modified by Madrzykowski [
59,
60]. In contrast, heat fluxes at the front instructor location in the Fire Behavior Lab regularly reached
Emergency conditions (> 12 kW/m
2) in the timeframe necessary to complete training objectives. The average time instructors would need to work in these conditions in order to complete six ventilation cycles of fire dynamics training ranged from 25 s (Palletspallets) to 153 s (OSB) due to the higher heat fluxes produced by the OSB fuel packages in this structure [
44].
Mean air temperatures were much lower at the instructor locations for the OSB fuel package with extra ventilation (due in part to the open door on the second floor), Pallet fuel package with excelsior (due to minimal involvement of pallets), and smoke barrel experiments (approximately ambient temperature). Mean heat fluxes at the second-floor instructor location for the smoke barrels and Pallet fuel package with excelsior experiments were similar to the background measurements collected prior to ignition. On the other hand, second-floor instructor location heat fluxes for the OSB fuel package with extra ventilation experiment were the highest of any scenario tested, nearly double the next highest measurement recorded. This result was likely indicative of higher convective heat flux due to higher temperature gases and higher gas flow velocity in the flow path created by the additional ventilation openings. This experiment was the only one conducted where heat flux values at the second-floor instructor location would be classified as the higher risk (Ordinary; 2 kW/m2 to 12 kW/m2) operating condition for any amount of time (peak values of 3.1 kW/m2 to 3.4 kW/m2).
Many factors contribute to the differences in thermal conditions between these two training structures (single compartment Fire Behavior Lab vs. multi-compartment) even though the same materials were used in the fuel packages for replicate experiments. The total weight of fuels was different, and the ratio of pallets to sheet goods (OSB, fiberboard) in the multicompartment experiments described here was different than in the Fire Behavior Lab. The orientation of the fuel can impact combustion efficiency and evolved thermal conditions. The total volume (Fire Behavior Lab: 51.3 m3 vs. multicompartment: 347.5m3) of the structure can impact initial dilution and subsequent mixing of the combustion gases with ambient air. The materials used in construction (Fire Behavior Lab: steel vs. multicompartment: concrete) can impact the heat loss to the environment. Finally, structure ventilation can impact thermal conditions as shown in the OSB fuel package with extra ventilation scenario. However, ventilation is often decided based on training objectives, and in the case of the Fire Behavior Lab, the cycling of ventilation is required to achieve training objectives.
4.2 Airborne Chemical Compounds
Concentrations of airborne chemical compounds were approximately an order of magnitude lower than those reported using identical sampling methods in the Fire Behavior Lab [
41], with the notable exception of HCl. Median total PAH concentrations ranged from 0.6 mg/m
3 to 3.0 mg/m
3 in this study, compared to 6.0 mg/m
3 to 33.7 mg/m
3 in [
41] (Fig. S5 in Supplemental Materials). The live-fire experiments conducted in this study were similar in duration to the 3-cycle experiments with the Fire Behavior Lab, but 10 min to 20 min shorter than the median times of the 6-cycle Fire Behavior Lab experiments, particularly for the pallet fuel packages [
41]. In the Fire Behavior Lab experiments, the pallet scenarios resulted in median total PAH concentrations that were much higher than fiberboard or OSB scenarios (14.4 mg/m
3 vs. 7.8 mg/m
3 and 9.1 mg/m
3, respectively). However, when the fuel load was placed on a metal burn rack slightly elevated above the ground with increased access to oxygen, the
Pallet fuel package resulted in significantly lower median total PAH concentrations compared to the
Fiberboard and
OSB fuel packages in this study (0.6 mg/m
3 to 0.7 mg/m
3 vs. 1.6 mg/m
3 to 1.8 mg/m
3 and 1.9 mg/m
3 to 3.0 mg/m
3, respectively). In related studies, Fent et al. measured total PAH personal gas concentrations (sampled by personal samplers located at chest height) with medians ranging from 2.78 mg/m
3 for fire instructors conducting pallet and straw fuels scenarios in a concrete training structure to 34.0 mg/m
3 for firefighters in an OSB, pallet, and straw-fueled scenario in a metal container-based structure [
37]. Personal gas samples collected using similar methods from firefighters responding to controlled residential fires measured a median of 23.8 mg/m
3 total PAHs (range 7.46 mg/m
3 to 78.2 mg/m
3) and 17.8 mg/m
3 total PAHs (range 9.77 mg/m
3 to 43.8 mg/m
3) for firefighters assigned to attack and search job assignments, respectively [
16]. Other studies have reported total PAH concentrations of 0.43 mg/m
3 to 2.70 mg/m
3 for particle board-fueled training fires in Australia [
31], 75 mg/m
3 to 180 mg/m
3 for particle board-fueled training fires also in Australia [
35], and 19 mg/m
3 to 41 mg/m
3 (sum of 22 PAHs) for chipboard-fueled training fires in Sweden [
34].
Similar to previous reports from training fire research, benzene was the most abundant BTEXS compound measured. Area gas concentrations of benzene measured at the 0.9 m working height inside the structure (median range of 2.5 mg/m
3 to 28 mg/m
3) were 10 to 100 times lower than that measured in the Fire Behavior Lab (median range of 19 mg/m
3 to 270 mg/m
3, Fig. S6 in Supplemental Materials) [
41], but similar to the personal gas concentrations measured by Fent et al. where the median range was 9.6 mg/m
3 to 29 mg/m
3 for instructors during live-fire exercises involving different fuels and structures [
37]. Kirk and Logan reported comparable gas concentrations of benzene during compartment fire behavior training sessions using particle board (4.5 mg/m
3 to > 7.8 mg/m
3) [
39]. On the other hand, Laitinen et al. reported area gas concentrations of benzene ranging from 0.62 mg/m
3 for pure spruce and pine wood-fueled fires to 1.0 mg/m
3 for chipboard-fueled fires (which also included polyurethane foam and kerosene) to 2.5 mg/m
3 for conifer plywood-fueled fires [
29]. In comparison, Fent et al. measured much higher personal gas concentrations of benzene from firefighters who were assigned to search and attack firefighting tasks inside controlled residential fires (median 121 mg/m
3 and 129 mg/m
3, respectively with peak concentrations near 1000 mg/m
3 for both groups) [
16].
In the Fire Behavior Lab experiments, the impact of fuel package on benzene concentrations was dependent on the different sampling locations inside the structure (front or rear instructor location) [
41]. The OSB fuel package experiments resulted in median benzene concentrations that were higher than the fiberboard or pallet scenarios at the rear instructor location (220 mg/m
3 vs. 72 mg/m
3 and 100 mg/m
3, respectively). On the other hand, OSB and pallet fuel packages resulted in similar median concentrations at the front instructor location (89 mg/m
3 and 98 mg/m
3, respectively) [
41]. In that study, it was noted that the fuel packages had differing impacts on the fire dynamics environment, which can impact the distribution of smoke within the structure. When fiberboard and OSB fuel packages were compared in shorter duration experiments where similar fire dynamic environments were created, there were no notable differences in airborne benzene concentrations. In the current study, when the fuel package was on a metal rack slightly elevated above the ground with increased access to oxygen, the
Pallet fuel package consistently resulted in the lowest median benzene concentrations compared to
Fiberboard and
OSB fuel packages (2.5 mg/m
3 to 3.2 mg/m
3 vs. 13 mg/m
3 to 17 mg/m
3 and 28 mg/m
3, respectively).
Median 1, 3 butadiene concentrations in the multicompartment structure were typically two orders of magnitude lower than in the Fire Behavior Lab for the same training fuels (multicompartment: 0.27 mg/m3 to 0.44 mg/m3, Fire Behavior Lab: 15 mg/m3 to 65 mg/m3 at the rear instructor location). However, the three highest overall airborne concentrations of 1, 3 butadiene were measured in the smoke barrel experiments with straw (1.9 mg/m3 and 12 mg/m3) and excelsior (3.9 mg/m3). The Pallet fuel package with excelsior experiment also stands out in its relatively high concentrations of 1, 3-butadiene (1.4 mg/m3 and 1.7 mg/m3 at the first and second floor respectively) relative to the distributions from with Pallet experiments. The common factor in each of these experiments is that smoldering combustion was a relatively larger contributor to smoke production compared to flaming combustion in the others.
Formaldehyde was present in high concentrations relative to its occupational exposure limits as has been reported in several other studies. Also, similar to previous studies, the relative differences between median concentrations produced by the different fuel packages (
Pallet: 3.7 mg/m
3 to 4.7 mg/m
3,
Fiberboard: 4.9 mg/m
3,
OSB: 4.6 mg/m
3 to 5.7 mg/m
3, with considerable overlap in overall ranges) was not as large as with total PAHs or benzene. In the Fire Behavior Lab, the same three fuels studied here produced levels of formaldehyde at the rear instructor location at the 0.9 m height that were similar to each other (median range of 44 mg/m
3 to 48 mg/m
3), though nearly an order of magnitude larger than measured in this study (Fig. S7 in Supplemental Materials) [
41]. Fent et al. also found comparable levels of formaldehyde between a pallet and straw scenario and one type (labeled as ‘Alpha’) of OSB (4.6 mg/m
3 vs. 4.5 mg/m
3), with similar magnitudes as reported here (also in similar sized structures) [
37]. Laitinen et al. reported mean formaldehyde concentrations ranging from 0.3 mg/m
3 to 1.5 mg/m
3 for training fires involving wood-based fuels in a ‘fire house’, and 11 mg/m
3 for training fires in a ‘gas simulator’ [
29]. Kirk and Logan reported similar gas concentrations of formaldehyde during compartment fire behavior training sessions (0.53 mg/m
3 to 5.0 mg/m
3) [
39]. The median acrolein concentrations reported here were also similar across the different fuels (
Pallet: 2.2 mg/m
3 to 3.0 mg/m
3,
Fiberboard: 1.7 mg/m
3 to 2.0 mg/m
3,
OSB: 2.7 mg/m
3 to 2.8 mg/m
3, with considerable overlap in overall ranges). These results are on the low end of those measured inside the Fire Behavior Lab structure (median range of 3.4 mg/m
3 to 32 mg/m
3) [
41] or the levels measured during live-fire exercises in Fent et al. where the median range was 4.9 mg/m
3 to 60.6 mg/m
3 [
37].
Airborne HCl in these training fire environments were similar across the different fuel packages (median concentrations for
Pallet: 9.3 mg/m
3 to 9.8 mg/m
3,
Fiberboard: 10 mg/m
3 to 11 mg/m
3,
OSB: 12 mg/m
3 to 13 mg/m
3, with considerable overlap in overall ranges). These concentrations are notably higher than median values reported in the Fire Behavior Lab with the same fuel packages (non-detect to 2.8 mg/m
3) [
41], contrary to the other airborne compounds reported here. In Horn et al. and Fent et al., the pallet-fueled scenario resulted in the highest concentrations of HCl [
37,
41]. The fuel packages used here and in Fent et al. utilized a larger proportion of pallets and straw in the OSB and fiberboard scenarios than in Horn et al., which may account for the higher relative concentrations in those studies. It is interesting to note that HCl was detected in all 16 experiments where pallets and straw were used (with or without fiberboard/OSB), but HCl was not detected at either sampling location in any of the three exploratory experiments where excelsior was substituted in place of the straw. Chlorine and other halogens occur in nature and may be absorbed by trees and/or straw. While these three independent studies have now shown higher concentrations with pallet and straw fuel packages, it is unknown whether this fuel package would contain more chlorine than timber used in the other wood-based products.
Considering all the fuels studied here, the highest concentrations of benzene and PAHs were measured during the
OSB fuel package experiments. However, the peak concentrations of most of the other compounds were from a smoldering smoke barrel with straw. Compared to flaming combustion, smoldering fires have been shown to result in higher yields of many chemical compounds for the same mass of material consumed due to less efficient combustion, though often at lower mass consumption rates [
61]. Although the fire service has been made aware of the concern in using engineered wood products in training fuel packages, anecdotally, they are often less concerned with the smoke produced by smoldering straw or excelsior in smoke barrels. The air temperature increase when using a smoke barrel is much lower than with a flaming wood-based fuel package, so some instructors have been traditionally less vigilant about using airway protection while inside or immediately outside training structures using these fuel packages. HCl is a respiratory irritant, which may encourage more stringent use of respiratory protection, but relatively low concentrations were detected in the straw smoke barrel compared to the flaming wood-based fuel packages [below OSHA PEL Ceiling limit (7 mg/m
3)], and HCl was not detected for the excelsior smoke barrel experiment. These data should further reinforce the need for use of respirator protection when firefighters are working in and around smoke, and more research should be conducted with replicate experiments using smoldering fuel packages.
On the other hand, the lowest concentration of nearly all airborne compounds measured here were from the single experiment using OSB, pallets, and straw with two important changes in ventilation. The open window on the first floor and doorway on the second floor increased ventilation to the fuel package and resulted in the highest ceiling temperatures in the fire room and heat flux at the second-floor instructor location of all 20 tests. However, this flow path also allowed much of the smoke to escape the second floor, resulting in the least impact on visibility and overall lowest concentration of contaminants at each instructor location.
Even though air temperatures were much lower on the second floor of the multi-compartment training structure, the airborne concentrations of nearly all compounds measured in this study were similar between the first-floor and second-floor instructor locations and in some cases higher on the second floor. Adding smoke barrels on the second floors may further increase airborne concentrations of these compounds at higher elevation, though with minimal impact on air temperature. The low air temperatures and distance from the radiant flame (or lack of flames in the smoke barrel sources) may lure some instructors into less consistent use of respiratory protection. Once again, these measurements reinforce the need to maintain respiratory protection even at these remote instructor locations.
By combining this data set with that from a Fire Behavior Lab using identical fuels, measurement techniques, and measurement height (simulating a crouching instructor’s head location) [
41], the ability for a fire instructor to impact airborne air concentrations through fuel package and structure ventilation practices can be further understood. It is apparent that the material selected for fuel packages in training fires can have an important impact on compounds in the air, particularly VOCs and PAHs. However, the orientation of those fuels and how they are incorporated into the fuel package also impact the relative contribution of contaminants in the air, particularly for the PAHs. Likewise, changing ventilation configurations had the most dramatic impact on concentrations of the measured airborne contaminants but also impacted the training fire environment created and thus, may not always be appropriate. Finally, the layout and volume of the training structure play an important role in the concentration of contaminants in the structure.
The smoke created with the intention of impacting visibility and/or fire dynamics in the training fire environment must be understood along with its contribution to airborne toxicants. Of the replicate fuel packages tested in this study,
OSB had the largest impact on visibility followed by
Fiberboard, then
Pallet. The overall ranking of airborne chemical concentrations followed the same trend. The straw smoke barrel on the second floor likewise resulted in a dramatic reduction in visibility, much more so than on the first floor. The excelsior smoke barrels resulted in slightly less impact on visibility than the straw for both floors. Once again, decreasing visibility was associated with increasing concentration of airborne contaminants. Finally, the
OSB fuel package with extra ventilation had the least impact on visibility and resulted in the lowest overall concentrations of airborne contaminants. The Fire Behavior Lab experiments (with noted differences in fuel package load, structure geometry, and ventilation) resulted in darker smoke with sufficiently high concentrations of unburned hydrocarbons to ignite and create rollover and/or flashover. Those experiments resulted in concentrations of contaminants that were often 10 to 100 times higher than those measured in this study. The Fire Behavior Lab study also showed that when less smoke was created (resulting in only surface burning in the fire area), the concentrations of contaminants in the smoke were much lower than when sufficient smoke was created to support rollover/flashover. These results further our understanding of the relationship between visibility in fires and concentrations of airborne chemicals in the smoke. Similarly, Purser has shown carbon monoxide and hydrogen cyanide concentrations increase with decreasing visibility and the concentrations of CO are higher for under-ventilated fires than well-ventilated fires for the same visibility [
62].
4.4 Conclusions
When training fire environments were created with three different fuel packages on the first floor of a multi-compartment training structure, first-floor temperatures and heat fluxes at the instructor locations were considerably higher than those from the second floor due to the proximity to the fire room. While the Pallet fuel package consistently resulted in the lower median time-averaged temperatures and heat fluxes compared to the OSB and Fiberboard fuel packages, there was no significant differences outside of the fire room at either simulated instructor location.
Fuel packages that included OSB and Fiberboard produced the highest median concentrations of total PAHs and VOCs (e.g. benzene) measured in this study, while the Pallet fuel package produced the lowest median concentrations of these compounds. These trends generally followed the qualitative visual obscuration created by each fuel. Airborne concentrations of compounds measured in this study at both the first and second floor instructor locations were similar, which reinforces the need to maintain respiratory protection throughout the entire training structure.
Exploratory tests were conducted to investigate the impact of increased ventilation and the use of smoldering smoke barrels. The training fire environment created by the OSB fuel package with increased ventilation resulted in the highest fire room temperatures but the lowest impact on visibility throughout the structure. Further, this test resulted in the lowest overall concentrations of contaminants in this study. In contrast, the smoldering straw-filled smoke barrel had minimal impact on the thermal environment but created a highly obscured environment and some of the highest concentrations of the targeted contaminants of any test.
It should be highlighted that decreasing visibility was associated with increasing concentration of airborne contaminants in this study. The data clearly demonstrates the need for consistent use of respiratory protection when firefighters are working in and around smoke and the need for post fire PPE cleaning and skin hygiene regardless of the fuel package. These data may be useful in balancing obscuration for training with potential exposure to thermal stressors and contaminants.
The fire service should understand the risk–benefit tradeoff when creating an elevated thermal environment for training. Trainees benefit from working in these hot environments for acclimatization, familiarization, sensing, and reacting to the environment as well as understanding PPE and firefighting tool operations, capabilities, and limitations. However, this benefit is balanced with increasing the risk for skin burns and PPE damage. The fire service and fire training organizations should take a similar approach to balancing the benefit of creating an environment where fire instructors and trainees work in reduced visibility and/or ignitable smoke with the increased risk for exposure to contaminants through dermal absorption, inhalation (when airway protection is not properly worn), and contaminated PPE.