Downward and Upward Spread of Smoldering Peat Fire

Transcription

1 10 th U. S. National Combustion Meeting Organized by the Eastern States Section of the Combustion Institute April 23-26, 2017 College Park, Maryland Downward and Upward Spread of Smoldering Peat Fire Xinyan Huang 1,* and Guillermo Rein 2 1 Department of Mechanical Engineering, University of California at Berkeley, USA 2 Department of Mechanical Engineering, Imperial College London, UK * Corresponding Author xinyan.huang@berkeley.edu Abstract: Smoldering fire in peatland is different from flaming wildland fire. The spread of smoldering peat fire is slow, low-temperature and more persistent, releasing a large amount of carbon and smoke into the atmosphere. In this work, we investigated the downward and upward smoldering spread of a peat column up to 30 cm height under variable moisture content (MC) and packing density. In the experiment, the downward fire spread rate was found to decrease with the increasing depth. As MC increased from 10% to 70%, the spread rate increased, which was the opposite to what was previously observed in the horizontal spread of smoldering. For the case of in-depth ignition, upward smoldering spread was not observed if the igniter was more than 15 cm below the free surface. This is thought to be because of the limited oxygen diffusion. If ignited, the smoldering fire first spread upward to the free surface, and then a secondary-stage downward spread was activated because of the multi-step oxidation process. These results suggested that (1) the oxygen diffusion controls the downward spread of peat fire, and (2) the incomplete peat oxidation sustains the upward fire spread. This study provides the novel physical understanding of the role of oxygen and moisture in the downward and upward spread of smoldering peat fires and may provide guidelines to prevent and suppress fires in peatlands. Keywords: Wildland Fire; 2-stage Spread; Critical Ignition Depth; Spread Rate 1. Introduction Smoldering wildfires in peatlands are the largest and longest combustion phenomena on Earth and contribute considerably to greenhouse gas emissions [1]. Annually, peat fires release huge amounts of ancient carbon roughly equivalent to 15% of man-made emissions [2], and result in the widespread destruction of ecosystems and regional haze events. Moreover, recent global warming dries the peatlands and increases the depth of belowground soil combustion, creating a positive feedback to the climate system [3]. Smoldering combustion is the slow, low-temperature, flameless burning of porous fuels and the most persistent type of combustion phenomena [4]. Smoldering involves heterogeneous reactions and is especially common in solid fuels like coal and organic soils with a charring tendency. Peat, as a typical organic soil, is a porous and charring natural fuel [5], thus prone to smoldering [1,4]. Two mechanisms control the spread of smoldering combustion: oxygen supply and heat losses [4]. The probability of ignition depends on the ignition source, environmental conditions, peat moisture content (MC, wt.%), inorganic content (IC). After ignition, the smoldering fire can spread laterally along the free surface and vertically downward into the deep peat layers, both of which are dominated by the forward smoldering spread [1,6], as illustrated in Fig. 1(a). The forward spread of smoldering peat fire has been mainly studied with small sample size (~ 5 cm thick) and under a sufficient oxygen supply. Frandsen [7] experimentally studied the ignition 1

2 threshold for various bench-scale peat and other soil samples and found a correlation between critical MC and IC, which were recently verified by a numerical model in [8]. For the large-scale peat fires in field, the downward spread of smoldering fire was found to consume peat up to depths of 50 cm [9,10]. The depth of burn (DOB) and critical MC for extinction at the in-depth spread of peat fires have been investigated by various experiments [11 14] and numerical simulation [15 17]. Recently, experiments on the horizontal peat fire spread showed that the spread rate decreased with MC while increased with the forward wind speed [18]. Figure 1. Diagrams of multi-dimensional smoldering peat fire spreading (a) downward and (b) upward. The upward spread of peat fire can also exist (see Fig. 1(b)). For example, because of the strong persistence of peat fire, if it is not completely suppressed in the deep layer, the hidden in-depth fire could spread upward in the drought season. Such upward fire spread may play an important role in managing the fire suppression strategy and preventing the fire hazard of re-ignition. Previous bench-scale experiments [18] also showed that if the unburned peat was placed above the burning peat, it could be ignited and consumed in the form of the upward spread. Torero and Fernandez-Pello [19] studied upward smoldering spread controlled by the natural convection and observed the transition to flame when smoldering front reaches the upper free surface. Palmer [20] found the upward smoldering spread could be achieved in a 5.2-cm cork dust pile and faster than the downward spread. So far, upward smoldering spread is still not well understood. In this work, the downward and upward spread of smoldering peat fires is studied using the bench-scale moss peat samples in the laboratory under different peat moisture and density conditions. 2. Experimental Setup Figure 2 shows the schematic diagram of the experimental setup for (a) downward fire spread, and (b) upward fire spread. A fire reactor was built using 1.27 cm thick insulation ceraboard to contain the peat sample and had an inner cross-section area of cm2 and a height of 30 cm. The fire reactor was further covered by several layers of aluminum foil to prevent the gas leakage and reduce the radiative heat loss. Such tall peat sample was used to ensure a 1-D vertical spread mode. The peat used in the experiment was a commercial moss peat from Ireland and was also used in [18]. After the oven-drying process, this moss peat has a bulk density of 135 ± 5 kg/m3 in its natural state and a low mineral content (IC~2%). To obtain the desired MC, the peat was first dried at 90 C for 48 h, and then well mixed with the corresponding amount of water, and left to rest inside a sealed basket for homogenization for another 48 h. Once the dry peat came in contact with air, it would quickly absorb the ambient moisture and reach an equilibrium with an MC of about 10%. Other targeted values for peat were 30%, 70%, and after mixing the uncertainty of the actual MC is ±5%. Moreover, during the water absorbing process, the volume of peat sample tended to expand naturally. Thus, although 2

3 the measured bulk density of wet peat increases with MC, the bulk density of dry peat or the fuel density decreases with MC, as shown in Fig. 2(d). Figure 2. Diagram of the experimental setup and the arrangement of thermocouples array: (a) downward fire spread, and (b) upward fire spread. To initiate a downward fire spread (Fig. 2(a)), a 10-cm long coil heater was placed to 5 cm below the top free surface. The ignition protocol was fixed to be 100 W for 30 min, which was strong enough to initiate a uniform smoldering front in a peat sample of MC < 150%. For initiating an upward fire spread (Fig. 2(b)), the same coil heater was placed right above the bottom insulation board. The thickness of peat layer and the heating protocol were adjusted to ensure a successful ignition and fire spread. A visual camera was hung above the sample to record the fire spread process. Also, thermocouples (TC) probes were inserted into peat and placed from 5 cm (coil heater) to 29 cm (near the bottom) away from the free surface with 2 cm interval (see Fig. 2(a)). TCs aimed to monitor both the temperature and the location of the smoldering front. At least 2-3 experiments were conducted at each condition. 3. Experimental Results and Discussions 3.1 Downward fire spread In the experiment, a typical downward fire spread over a 30-cm deep sample could last from 20 to 50 hours, depending on the peat MC and density. Figure 3 shows the thermocouple measurements in the downward fire spread under different peat MCs, where the negative sign of the thermocouple position means the distance below the initial free surface (z = 0). During the forced ignition by coil heater (t < 30 min), the peat temperature near the ignition zone (i.e. -7 and -9 cm) rapidly increased above 600 C. After the ignition, their temperature first decreased, and then increased again, which indicates a self-sustained smoldering front. During the downward fire spread, the peak temperature was 550~600 C which was not varied under different peat MCs or the packing densities. As the fire spread deeper, the peak temperature slowly decreased with the depth, and such decrease became less prominent as MC increased. At the same time, the peat sample shrank, and the top free surface regressed. 3

4 Figure 3: Thermocouple measurements in the downward fire spread over a 30-cm deep peat sample: (a) MC = 10% (dried), ρ wp = 135 kg/m3; and (b) MC = 70%, ρ wp = 160 kg/m3. During the fire spread, a thin black char layer was observed on the free surface and was not converted into white probably because of a large heat loss to the environment. Below this thin char layer, a white ash layer was observed. Near the end of spread, the peat temperature was found to be below 200 C, where the smoldering front (or the char-oxidation zone) could not be sustained. Therefore, another black char layer of 1-4 cm thick was left right above the bottom. Figure 4: (a) Downward spread rate and (b) 2-stage (upward + downward) spread rate. Figure 4(a) shows the downward smoldering spread rate vs. MC and depth. Clearly, as the peat MC increases, the downward spread rate continuously increases. This result is opposite to the previous finding in the horizontal fire spread near the free surface for the identical peat, where the spread rate decreased with increasing MC [18]. There are two possible reasons: (1) the water in the porous peat matrix increases the overall conductivity in the preheat zone and facilitates the heat transfer from the upstream oxidation zone; and (2) the reduction of fuel density due to the natural expansion of peat sample after the water absorption (see Fig. 2(d)). 3.2 Upward fire spread To initiate a peat fire with an upward spread, the coil heater was placed right above the bottom insulation. However, it was found that for the same 30-cm deep dried sample, the ignition was not successful when applying 100 W from 30 min to even 120 min. Then, the sample depth was reduced to find the maximum (critical) depth for ignition. Interestingly, such maximum depth 4

5 was found to be about 15 cm for all MCs and packing densities. Smaller than that depth, ignition was successful using 100 W for 30 min (MC = 10%) and 90 min (MC = 70%). Figure 5(a) shows a typical thermocouple measurement of a 15-cm deep dried and compressed peat sample ignited from the bottom, where the negative sign of the thermocouple position means the distance below the free surface (z = 0). After ignition, the smoldering front spread upward towards the free surface, as illustrated in Fig. 5(b)I-III. Specifically, the upward spread lasted for 4~5 h, and the peak temperature was relatively low (about 300 C) compared to the downward spread in Fig. 3. This low temperature suggested the existence of a weaker oxidation. During the upward spread, no smoke was observed before the color of free surface became black (charring process). Although no regression was observed on the free surface, in some experiments internal collapses might take place, as indicated by the discontinuities found in thermocouple measurement. A sampling of the sample also indicated that all the original peat was converted into char during the upward spread (see Fig. 5(b)III). Figure 5: (a) Thermocouple measurements in the upward fire spread over a 15-cm deep dry peat sample ρ wp = 190 kg/m3; and (b) schematic diagram of the upward spread process of smoldering peat fire. Afterward, the smoldering front started to spread downward, as illustrated in Fig. 5(b)IV and such a process was also clearly indicated by the thermocouple measurements from 5 to 22 h. During this second-stage downward spread, the peak temperature is similar to the direct downward spread in Fig. 3, as dominated by the char oxidation. Due to the heat loss from the bottom, similar low-temperature profile near extinction and char residues were observed. After extinction, a sandwich structure of fire residue (char + ash + char) was also observed, similar to the downward spread. Figure 4(b) shows the measured upward and downward spread rate from Fig. 5(a), and the error bar indicates the uncertainty of repeating the experiment. Clearly, for both the first-stage upward spread and the second-stage downward spread, their spread rates decrease with the depth, same as the direct downward spread. Moreover, the upward spread was found to be faster than downward spread despite a lower peak temperature (see Fig. 5(a)). 4. Conclusion In this experimental and numerical study, we investigate the downward and upward spread of peat fire under variable moisture content (MC). In the experiment, the downward fire spread rate is found to decrease with increasing depth. Moreover, the spread rate increases with MC, opposing to the previous experiment on the horizontal spread. For the upward fire spread, there is a maximum depth for ignition (~15 cm), deeper than when no ignition occurs regardless of the 5

6 ignition power and duration. After ignition, a first-stage upward spread and a second-stage downward spread are observed because of the multi-step oxidations. Results suggest that (1) the oxygen diffusion controls the downward spread of peat fire, and (2) the incomplete peat oxidation sustains the upward fire spread. Acknowledgements The authors thank Dr Rory Hadden (University of Edinburgh) for his assistance in experiments. References [1] Rein G. Smouldering Fires and Natural Fuels. In: Claire M. Belcher, editor. Fire Phenomena in the Earth System, New York: John Wiley & Sons, Ltd.; 2013, p doi: / ch2. [2] Page SE, Siegert F, Rieley JO, Boehm H V, Jayak A, Limink S. The amount of carbon released from peat and forest fires in Indonesia during ;1999:61 5. doi: /nature [3] Turetsky MR, Benscoter B, Page S, Rein G, van der Werf GR, Watts A. Global vulnerability of peatlands to fire and carbon loss. Nature Geoscience 2015;8:11 4. doi: /ngeo2325. [4] Rein G. Smoldering Combustion. SFPE Handbook of Fire Protection Engineering, New York, NY: Springer New York; 2016, p doi: / _19. [5] Chambers FM, Beilman DW, Yu Z. Methods for determining peat humification and for quantifying peat bulk density, organic matter and carbon content for palaeostudies of climate and peatland carbon dynamics. Mires and Peat 2011;7:1 10. [6] Huang X, Rein G. Smouldering combustion of peat in wildfires: Inverse modelling of the drying and the thermal and oxidative decomposition kinetics. Combustion and Flame 2014;161: doi: /j.combustflame [7] Frandsen WH. The influence of moisture and mineral soil on the combustion limits of smouldering forest duff. Canadian Journal of Forest Research 1987;16: doi: /x [8] Huang X, Rein G, Chen H. Computational smoldering combustion: Predicting the roles of moisture and inert contents in peat wildfires. Proceedings of the Combustion Institute 2015;35: doi: /j.proci [9] Ballhorn U, Siegert F, Mason M, Limin S., Limin S. Derivation of burn scar depths and estimation of carbon emissions with LIDAR in Indonesian peatlands. Proceedings of the National Academy of Sciences of the United States of America 2009;106: doi: /pnas [10] Rein G, Cleaver N, Ashton C, Pironi P, Torero JL. The severity of smouldering peat fires and damage to the forest soil. Catena 2008;74: doi: /j.catena [11] Benscoter BW, Thompson DK, Waddington JM, Flannigan MD, Wotton BM, de Groot WJ, et al. Interactive effects of vegetation, soil moisture and bulk density on depth of burning of thick organic soils. International Journal of Wildland Fire 2011;20:418. doi: /wf [12] Davies GM, Gray A, Rein G, Legg CJ. Peat consumption and carbon loss due to smouldering wildfire in a temperate peatland. Forest Ecology and Management 2013;308: doi: /j.foreco [13] Watts AC. Organic soil combustion in cypress swamps: Moisture effects and landscape implications for carbon release. Forest Ecology and Management doi: /j.foreco [14] Zaccone C, Rein G, D Orazio V, Hadden RM, Belcher CM, Miano TM. Smouldering fire signatures in peat and their implications for palaeoenvironmental reconstructions. Geochimica et Cosmochimica Acta 2014;137: doi: /j.gca [15] Huang X, Rein G. Computational study of critical moisture and depth of burn in peat fires. International Journal of Wildland Fire 2015: doi:dx.doi.org/ /wf [16] Huang X, Rein G. Interactions of Earth s atmospheric oxygen and fuel moisture in smouldering wildfires. Science of the Total Environment 2016;572: doi: /j.scitotenv [17] Yang J, Chen H, Liu N. Modeling of Two-Dimensional Natural Downward Smoldering of Peat. Energy & Fuels 2016:acs.energyfuels.6b doi: /acs.energyfuels.6b [18] Huang X, Restuccia F, Gramola M, Rein G. Experimental study of the formation and collapse of an overhang in the lateral spread of smouldering peat fires. Combustion and Flame 2016;168: doi: /j.combustflame [19] Torero JL, Fernandez-Pello a. C. Natural convection smolder of polyurethane foam, upward propagation. Fire Safety Journal 1995;24: doi: / (94)00030-j. [20] Palmer KN. Smouldering combustion in dusts and fibrous materials. Combustion and Flame 1957;1: doi: / (57)90041-x. 6