Effects of Time-Dependent Heat Fluxes on Pyrolysis and Spontaneous Ignition of Wet Wood. Anhui, China

Transcription

1 Effects of Time-Dependent Heat Fluxes on Pyrolysis and Spontaneous Ignition of Wet Wood Zhai C. J. 1,2, Yang Z. 1, *, Zhou X. D. 1, Peng F. 1, Gong J. H. 3 1 University of Science and Technology of China, State Key Laboratory of Fire Science, Hefei, Anhui, China 2 Nanjing Forest Police College, Department of Forest Fire Protection, Nanjing, Jiangsu, China 3 Nanjing Tech University, College of Safety Science and Engineering, Nanjing, Jiangsu, China *Corresponding author yanglz@ustc.edu.cn ABSTRACT This work investates experimentally and theoretically the effects of time-dependent incident heat flux (HF), which is more reasonable in fire-like environment, on thermal degradation process of wet pine wood. A feedback method was utilized to generate a time-dependent HF by controlling the output power of radiative heater, namely qin=αt β, and both quadratic and linear heat fluxes are focused in this study. Comparison between the measured varying heat fluxes and desned values indicates that this method provides hh accuracy necessary. Measurements of mass loss rate, temperature distribution at different depths of material and nition time were implemented in the tests to examine the influence of time-dependent heat fluxes. The results showed that the mass loss rate is affected snificantly by the changed heat flux compared with constant scenario. The critical mass flux, which keeps almost unchanged, can be employed as the nition criterion due to the fact that the nition temperature increases with increasing HF, which also certifies the conclusions of other researchers. The heat penetration layer is restricted to a thinner depth with larger α and β. A simplified theoretical model is used to predict the surface temperature before nition and good agreement exists between the experimental and theoretical results. Furthermore, a linear relationship was found between nition time and α 2/(1+2β), which is also validated by the experimental data and is reexamined by the constant circumstance. KEYWORDS: Time-dependent heat flux, pyrolysis, wood, nition time. NOMENCLATURE A pre-exponential factor (s 1 ) C thermal capacity (J kg 1 K 1 ) E reaction active energy (J mol 1 ) h conv Convective coefficient (W m 2 ) H enthalpy per unit mass (kj m 3 ) k thermal conductivity (W m 1 K 1 ) k D Darcyʼs coefficient (m 3 s kg 1 ) L thickness of sample (mm) m mass (kg m 2 ) m'' mass flux (kg m 2 s 1 ) M molar mass of gas (g mol 1 ) p pressure (Pa) P power (kw) q heat flux (kw m 2 ) Q reaction heat (J kg 1 ) R ideal gas constant (J mol 1 K 1 ) t time (s) T temperature (K) u velocity of gas (m s 1 ) x spatial variable (mm) Y moisture content Proceedings of the Ehth International Seminar on Fire and Explosion Hazards (ISFEH8), pp Edited by Chao J., Liu N. A., Molkov V., Sunderland P., Tamanini F. and Torero J. Published by USTC Press ISBN: DOI: /c.sklfs.8thISFEH Greek α, β coefficient in heat flux ε porosity ζ emissivity μ gas viscosity coefficient (N s m 2 ) ρ density (kg m 3 ) σ Stefan-Boltzmann constant Subscripts 0 initial condition ambient a active material g gas in incident l water s solid v pyrolyzed volatile w wood

2 Proceedings of the Ehth International Seminar on Fire and Explosion Hazards (ISFEH8) INTRODUCTION Burning of solid fuels has been an important subject in the field of fire safety engineering [1]. When the fuel is heated, the surface and inner temperature rises. Fuel starts to pyrolyze, release combustible gas from the surface and finally lead to flame while the critical conditions for nition are met. Research on the characteristics of the early stage of fire will help to prevent the fire disaster and reduce the corresponding danger. One of the most interesting materials to be examined is wood which plays an important role in building fires [2]. Various experiments and models have been conducted [3-9] so far to study the pyrolysis and burning behavior of wood considering different factors, such as external energy source, materials and environmental conditions. In this paper, we focused on the effect of heat flux (HF) that is very important for the process of energy transfer. Many studies have been done against this problem under the condition of standard time-temperature curve prescribed by ASTME 119 [10] or ISO 834 [11] for purpose of interchangeable data. But the HF used in the literature was often set to be constant while it always varies with the growth of a fire in a practical fire environment. Therefore, it will be more rational to study the pyrolysis and nition of wood under variable HFs. Some researchers did the experiments of different materials under simple variable HFs, trying to get more practical results. Bilbao et al. [12] carried out an experimental study on the fire behavior of wood exposed to decreasing HFs corresponding to the thermal radiation of a flame produced in an accidental release of materials. Results show that the decreasing rate of HF had little influence on nition time. As heat-release rate is generally time-dependent in fire, Yang et al. [13] did the experiments under linearly- dependent HFs, demonstrating an agreement of power function between the nition time and the increasing rate of the HFs. Ji et al. [14] gave the nition criteria of wood under linearly increasing heat fluxes according to experiment and simulation, showing that wood will catch fire when the surface temperature reaches 500 o C and the increasing rate of heat flux is larger than 0.06 kw/(m 2 s). However, in their work, the heat flux can only be linearly-increasing due to the simple controlling method. More real fire environment is not presented and logical desn of HFs is not found in previous literature, so we need to desn a more complex developing HF, which will certainly be helpful to get further understanding of the influence of HF on the pyrolysis behavior of wood. In this paper, we extend the work in [15-18] by desning two kinds of time-dependent HFs, linearlyincreasing HF and quadratically-increasing HF that are more reasonable in a practical fire environment and more robust than that in [15] with improved controlling algorithm. Linearlyincreasing HFs are desned for the fundamental research, and quadratically-increasing HFs are desned according to Heskestadʼs work [19], where it is found that heat release rate increases quadratically in the early stages of fire development. Under the desned HFs, wet white pine that is very common in our surroundings is chosen as the sample to conduct the experiments. Temperature distribution, mass loss and nition time are studied. With a numerical model considering basic chemical and physical processes, such as pyrolysis, water evaporation and gas transportation, mass loss rate of pyrolyzed volatile is calculated. The aim of this paper is to show basic effects of timedependent HFs on the pyrolysis of wet white pine and give some suggestions for further research on this subject. MATERIALS AND METHODS Materials 50 mm 50 mm square white pine wood samples with measured moisture of 10.3% were employed in the experiments, and the thickness is 20 mm. The samples were first wrapped by a double-layer of fiberglass fabric and aluminum foil on sides and bottom surface to guarantee the insulated boundary condition, and then were mounted in a 100 mm 100 mm 50 mm refractory holder before tests. A volume of 53 mm 53 mm 23 mm was desned on the holder to fix the specimen. Thus, one 286

3 Part II Fire dimensional scenario was achieved. The top surface of the samples and the holder are in exactly the same position below heater during the tests, namely 10 cm. Detailed information of the wet pine wood, including thermodynamics and kinetics, are listed in Table 1. Table 1. Thermodynamics and kinetics parameters of wood used. Parameter Value Source Parameter Value Source Ea 150 (kj mol 1 ) [20] Ql 2440 (kj kg 1 ) [22] Aa 1.4e10 (s 1 ) [20] ζ 0.78 [22] Qa 300 (kj kg 1 ) [21] hconv 10 (W m 2 K 1 ) [22] El 88 (kj mol 1 ) [22] Cw 1950 (J kg 1 K 1 ) [23] Al 5.13e10 (s 1 ) [22] Cc 1390 (J kg 1 K 1 ) [23] Calibration of heat flux Fure 1. Schematic of the experimental setup. The experimental apparatus utilized in pyrolysis and nition tests are illustrated in F. 1. Six silicon carbide bars were employed as the radiation source to generate the HF. To validate the uniformity of the heat flux of heater on top surface of samples during tests, a Schmidt-Boelter heat flux gauge was used in the experiments to obtain the measured HF in 5 locations on the horizontal position equivalent to top surface of sample. The HF gauge, with measurement range of 0 to 100 kw/m 2, was calibrated before the experiment by the first metrology and measurement institute of CASC. Constant HF was used here for simplification, and same conclusions can be derived for time-dependent HF. Four locations were selected at the corner of the holder and the left one is in the center. The uncertainty of the measured HF was found to be lower than 10%, which indicates that the uniformity of HF is acceptable for the tests. Also for the center one, additional measurement was conducted 5 mm, maximum value of the regression depth, below the other tests to examine the influences of regression of wood. No more than 5% of the HF decay was found, and thus the effects can be neglected. For time-dependent HF, the output power of heater can be controlled to increase continuously from 0 to 55 kw by computer using a heat-flux module. To obtain controllable heat flux, it is important to precisely measure the exact heat flux value during the heating to obtain the feedback. The real-time measurement of the HF gauge can be recorded by computer through the heat-flux interface to construct a feedback loop. In [15], linearly time-increasing heat fluxes were generated by adjusting the output power that keeps constant in the experiments. This simple controlling method fails when more complex HF is required, and a feedback method was adopted in this work. A series of experiments under different constant output power were performed, and subsequently the relationship among the power, HF and the increasing rate of HF was recorded. For a desned HF curve, the HF and the rate of HF at different time were calculated. The output power was then adjusted by the deviation between the measured and desned value, as given in Eq. (1): 287

4 Proceedings of the Ehth International Seminar on Fire and Explosion Hazards (ISFEH8) P 2=P 1+χ(q 1 q 2), (1) where P 2 is the output power to be set, P 1 is the output power recorded, q 1 is the measured HF, q 2 is the desned HF, χ is a parameter set manually for adjusting the effects of feedback. Only one parameter needs to be adjusted in this method, simplifying the problem compared to classic proportional-integral-derivative (PID) method [24]. Two groups of HF were desned based on this method that can be expressed as q in=αt β. In this study, the scenarios with β=2, and β=1 were defined as group A and B for convenience, respectively. Results are shown in Table 2 and F. 2. The good agreement between desned and measured HF verifies the validity of this method. Table 2. Expressions of incident heat flux versus time. Group A Expressions Group B Expressions a1 qin(t) = 4.84e 4t 2 b1 qin(t) = 0.04t a2 qin(t) = 9.32e 4t 2 b2 qin(t) = 0.064t a3 qin(t) = 1.70e 3t 2 b3 qin(t) = 0.11t a4 qin(t) = 2.40e 3t 2 b4 qin(t) = 0.16t a5 qin(t) = 3.55e 3t 2 b5 qin(t) = 0.27t Fure 2. Comparison between desned and measured HF: (a) group A, (b) group B. Experimental methods In pyrolysis tests, an electronic balance, with 0.01 g resolution, was located below the sample to obtain the mass data and no temperature measurement was conducted. While in the temperature measurement tests, only the temperature distribution is measured by thermocouples (K-type, NiCr- NiSi) with 0.5 mm diameter, which is embedded in the sample at depth of 0, 5, 10 and 15 mm below the exposed surface. In both circumstances, nition time measurement was implemented by a stopwatch once a visible flame was observed on the sample. Spontaneous nition of the sample is the focus in this paper and no external spark niter is used. A gypsum shutter was employed below the heater before the exposure of sample. Both mass and temperature data were recorded at a frequency of 1 Hz. Each scenario was repeated at least 5 times to guarantee the repeatability and uncertainty of data. The data after nition was not recorded, and no nition occurs under HF b 1 and b 2. The experiments were conducted at room temperature in quiescent air. 288

5 Part II Fire RESULTS AND DISCUSSION Mass loss rate Mass loss rate (MLR) is an important parameter in pyrolysis process, which can be regarded as an nition criterion of wood nition. F. 3 illustrates the measured MLR in different circumstances. A one-step global model based on [3, 25] is adopted to simulate the MLR of pyrolyzed volatiles, accounting for kinetics and thermodynamics of degradation processes, such as water evaporation, gas transportation and wood pyrolysis. Fure 3. Measured MLR of the two groups of HF: (a) group A. (b) group B. The employed model is briefly described below, and the detailed information can be found in [3]. The parameters used in the simulation are listed in Table 1. Mass conservation equations for solid, liquid and vapors are expressed as: ρ E s a = Aρ exp a t RT, (2) ρ A exp E l l l l = ρ t RT, (3) ερ ρ u ρ g g s + = t x t, (4) ερgl ρglu ρ l + =. (5) t x t The velocity of gas in solid can be described by Darcyʼs law: kd p u = µ x. (6) Inside the sample, ideal gas law is utilized for the pressure of gas: p = ρ RT / M + ρ RT / M. (7) gv gv gl gl Energy conservation equation: ρ sh s T m gvh g m glh gl a l ks Q ρ a Q ρ = + l t x x x x t t. (8) 289

6 Proceedings of the Ehth International Seminar on Fire and Explosion Hazards (ISFEH8) Boundary condition: T 4 4 x = 0: ks x = 0 = qin σζ ( T T ) hconv ( T T ), x (9) T x = L: x= L= 0. x (10) The thermal conductivity of wet wood was obtained from Ref. [24]: k w = Y (W m 1 K 1 ). (11) In previous literatures, a constant nition temperature, T, was most comely used as the nition criterion, and it was regarded as an inherent thermal parameter of material. Recently, some other researchers found [26-29] that the nition temperature actually increases with increasing incident HF, which also can be validated by the measured data in this paper. Lautenberger [29] also predicted an increase in the nition temperature with incident radiant fluxes when studying the nition process under various incident radiant heat fluxes from 0 to 200 kw/m 2. In that theoretical model, the time to piloted nition used the critical mass flux as the nition criterion as the incident radiant flux increases. The interpretation given was that the pyrolysis reaction is increasingly confined to a thin layer near the surface and this layer must be raised to a hher temperature to achieve the critical mass flux for nition than at lower radiant flux. Table 3 lists the measured nition temperature, and apparently the T increases with increasing HF. Extreme conditions are b 1 and b 2, and no nition occurred. Simulation results of critical MLR are also given in Table 3, showing that it is not affected by the varying incident HF. This indicates that the critical MLR is a more reliable nition criterion for time-dependent HF rather than T. T under HF with larger α has to be larger to keep the critical MLR constant due to the thinner thermal penetration depth. Char layer oxidation affects the pyrolysis process snificantly in the experiments, especially under HF b 1 and b 2 where the surface was covered by ash, indicating that char oxidation should be taken into consideration when developing a more accurate numerical model. Table 3. Simulated critical MLR, measured T and surface temperature on T. T (s) MLRcir (g m 2 s 1 ) T (ºC) T (s) MLRcir (g m 2 s 1 ) T (ºC) a b1 N/A N/A N/A a b2 N/A N/A N/A a b a b a b Temperature distribution in solid Temperature distribution before nition in condensed phase of material can be predicted by a simplified theoretical model in which the sample is assumed to be thermal inert, namely the thermal parameters are constant. The energy conservation equation in solid can be expressed as: ρ C T t = k w p,w w 2 2 T. (12) x 290

7 Part II Fire The effects of thermal degradation and transportation of volatiles are neglected. Combining with boundary and initial conditions, the surface temperature can be derived as: Ts = T + 1 kw rw Cp,w q&in - q&l dt, 0 t -t ò π t (13) where q&l denotes the heat loss on surface by convection and radiation. With q&in = a t b, Eq. (13) can be integrated as: Ts = T + 2 (a t b - q&l ) t 0.5 k w r w Cp,w π. (14) When the surface temperature is relative low before the nition, q&l can be nored, and thus Eq. (14) is simplified as: Ts = T + 2a t b kw rw Cp,w π. (15) F. 4 illustrates the measured and predicted surface temperatures under the two groups of HFs. The surface temperature was first estimated by Eq. (15), and subsequently was employed to obtain q& l. Finally the surface temperature was recalculated by Eq. (14). Fure 4. Measured and predicted surface temperature under group (a) A and (b) B of HF. In both cases, the surface temperature increases with time and the increasing profiles are similar to the HF curves. As the energy loss from the exposed surface by re-radiation and convection is related to its temperature, the energy received by sample decreases with the increase of surface temperature, leading to a gradual decrease of the increasing rate of temperature over time under constant HF [22]. Being different from that constant HF condition, the increasing rate of temperature increases over time under time-dependent HFs since the received energy keeps increasing. Therefore, the increasing process maintains almost the entire period before nition under time- dependent HFs, while it mainly occurs at the beginning under constant HF. As the temperature rises much faster on the surface than that at depth of 5 mm under time- dependent HFs, there is a large temperature gradient around the surface. This difference leads to an important phenomenon that the thermal penetration depth under time-dependent HF is much thinner than that 291

8 Proceedings of the Ehth International Seminar on Fire and Explosion Hazards (ISFEH8) under constant HF [22] as shown in F. 5 where the temperature distributions under HF a1, a5 and b1, b5 are given. Similar to the difference between time-dependent HF and constant HF with β = 0, the thermal penetration depth is thinner under HF with hher β as shown in the comparison between HF a5 and b5. Furthermore, larger a leads to thinner thermal penetration depth for a given β due to the fact that the surface temperature increases in a shorter time and less energy is transported into the interior of sample. As the sample is wet, there is an obvious drying plateau at depth of 5 mm as shown in Fs. 4(b) and 4(c). No plateau was found in other graphs of F. 4 as the temperature at depth of 5 mm does not reach 100 ºC before nition. Fure 5. Temperature distributions at different depths: (a) a1, (b) a5, (c) b1, (d) b5. Ignition time Table 3 lists the measured nition time and calculated critical MLR, where N/A denotes that no nition occurred. The relationship between t and a and b can be obtained by Eq. (16): 1 b +0.5 t = (T 2a - T ) kw rwcp,w π. (16) 292

9 Part II Fire In Ref. [17], when considering a ramped heat flux with no surface losses, the nition time was expressed as: 1/2 w w p,w q net ( t ) ( ) = t π 3 k ρ C T T. (17) Integrating Eq. (17), the nition time can be related to the squared integral of the incident energy: 2 1 t t = q 0 netdt 3( T T ), (18) πkwρwcp,w t 2 t is proportional to ( q 0 netdt), which indicates that the nition delay time is expressed as a function of the total energy delivered to the surface. Substituting for α t β net be obtained as Eq. (16). With β = 0, 1 and 2, Eq. (16) can be expressed as: 1 2α = t T T k C ( ) 0.5 wρw p,w 1 2α = t T T k C ( ) 1.5 wρw p,w 1 2α = t T T k C ( ) 2.5 wρw p,w q, similar correlation can, (19) π, (20) π. (21) π Eq. (19) is the classical nition correlation in literatures for constant HF when neglecting the surface heat loss. Also, another correlation can be derived by Eq. (16): 2 2β 1 2 wρw p,wπ + 2β+ 1 ( ) T T k C 2 t = α. When β = 0, 1 and 2, t α 2/3, t α 2/5 and t α, 2 which agree with the conclusion of constant HF, [17] and [24] where it was found t is proportional to α by experimental data. F. 6 shows the relationship between t and α 2/(1+2β) In F. 6(b), the measured nition time of toon, paulownia, elm and acacia under linear external heat flux are also plotted, and good linearity is found. The slops and intercepts of the fitted lines are determined by the thermal parameters of the materials. CONCLUSION In summary, experimental investations on pyrolysis and auto-nition of wet white pines exposed to two groups of time-dependent heat fluxes that are desned to be more reasonable in practical applications were performed, including three aspects: temperature at different depths, mass loss and nition time. A one-step global model is used to calculate the MLR cri at nition time. Main results of the study are: (1) The dependent curve of the surface temperature is similar to that of the time-dependent HFs, indicating that the increasing rate of temperature raises over time which results in thinner thermal penetration depth for hher β. Besides, there is obvious drying behavior inside the sample but none on the surface during the heating. 293

10 Proceedings of the Ehth International Seminar on Fire and Explosion Hazards (ISFEH8) (2) MLR cri is a more reliable nition criterion than surface temperature under time-dependent HF. The surface temperature at nition time is hher under HF with larger increasing rate because of the thinner thermal penetration depth. Char oxidation is observed in the experiment, especially under HF with slow increasing rate. The surface of the sample turns into ash and no flame occurs under HF b 1 and b 2 no matter how large the HF and surface temperature can be. (3) Power function is found to be able to describe the relationship between nition time and the increasing rate of HF, demonstrating that t is proportional to α 2/(1+2β). Fure 6. Ignition time versus α -2/(1+2β) under group (a) A and (b) B of HF. ACKNOWLEDGMENT This research was supported by National Natural Science Foundation of China (No: ), Natural Science Foundation of Jiangsu Province (No: BK ) and The Open Fund of the State Key Laboratory of Fire Science (SKLFS) Program (HZ2015-KF09). The authors deeply appreciate the support. REFERENCE 1. Babrauskas, V. Ignition Handbook, Fire Science Publishers, Di Blasi, C. Modeling and Simulation of Combustion Process of Charring and Non-Charring Solid Fuels, Progress in Energy and Combustion Science, 19(1): , Kung, H. C. A Mathematical Model of Wood Pyrolysis, Combustion and Flame, 18(2): , Chen, Y., Delichatsios, M. A., and Motvalli, V. Material Pyrolysis Properties, Part I: An Integral Model for One-Dimensional Transient Pyrolysis of Charring and Non-Charring Materials, Combustion Science and Technology, 88(5-6): , Delichatsios, M. A. Ignition Times for Thermally Thick and Intermediate Conditions in Flat and Cylindrical Geometries, In: Curtat, M. (Ed.), Fire Safety Science Proceedings of the Sixth International Symposium, pp , Spearpoint, M. J., and Quintiere, J. G. Predicting the Piloted Ignition of Wood in the Cone Calorimeter Using an Integral Model: Effect of Species, Grain Orientation and Heat Flux, Fire Safety Journal, 36(4): , Yang, L. Z., Wang, Y. F., Zhou, X. D., Dai, J. K., and Deng, Z. H. Experimental and Numerical Study of the Effect of Sample Orientation on the Pyrolysis and Ignition of Wood Slabs Exposed to Radiation, Journal of Fire Sciences, 30(3): ,

11 Part II Fire 8. Fereres, S., Lautenberger, C., Fernandez-Pello, A. C., Urban, D. L., and Ruff, G. A. Understanding Ambient Pressure Effects on Piloted Ignition through Numerical Modeling, Combustion and Flame, 159(12): , Shi, L., and Michael, Y. L. C. Experimental Study of Woods under External Heat Flux by Autonition, Journal of Thermal Analysis and Calorimetry, 111(2): , ASTM E119. Standard Test Methods for Fire Tests of Building Construction and Materials, West Conshohocken, PA, ASTM International, ISO 834, Fire-Resistance Tests Elements of Building Construction, International Organization for Standardization, Geneva, Bilbao, R., Mastral, J. F., Lana, J. A., Ceamanos, J., Aldea, M. E., and Betran, M. A Model for the Prediction of the Thermal Degradation and Ignition of Wood under Constant and Variable Heat Flux, Journal of Analytical and Applied Pyrolysis, 62(1): 63-82, Yang, L. Z., Guo, Z. F., Ji, J. W., and Fan, W. C. Experimental Study on Spontaneous Ignition of Wood Exposed to Variable Heat Flux, Journal of Fire Sciences, 23(5): , Ji, J. W., Hao, Y. H., Wang, Y. E., Zhang, J. Y., and Li, J. G. On the Comprehensive Auto-Ignition Mechanism for Wooden Material under Linearly Increasing Heat-Flux, Journal of Safety and Environment, 9: , Yang, L. Z., Guo, Z. F., Zhou, Y. P., and Fan, W. C. The Influence of Different External Heating Ways on Pyrolysis and Spontaneous Ignition of Some Woods, Journal of Analytical and Applied Pyrolysis, 78(1): 40-45, Reszka, P., Borowiec, P., Steinhaus, T., and Torero, J. L. A Methodology for the Estimation of Ignition Delay Times in Forest Fire Modelling, Combustion and Flame, 159(12): , Vermesi, I., Roenner, N., Pironi, P., Hadden, R. M., and Rein, G. Pyrolysis and Ignition of a Polymer by Transient Irradiation, Combustion and Flame, 163: 31-41, Lamorlette, A. Analytical Modeling of Solid Material Ignition under a Radiant Heat Flux Coming from a Spreading Fire Front, Journal of Thermal Science and Engineering Applications, 6(4): 1-4, Heskestad, G. Society of Fire Protection Engineers, Technology Report, pp , Wagenaar, B. M., Prins, W., and Van Swaajj, W. P. M. Flash Pyrolysis Kinetics of Pine Wood, Fuel Processing Technology, 36(1-3): , Yang, L. Z., Chen, X. J., Zhou, X. D., and Fan, W. C. A Modified Model of Pyrolysis for Charring Materials in Fire, International Journal of Engineering Science, 40(9): , Shen, D. K., Fang, M. X., Luo, Z. Y., and Cen, K. F. Modeling Pyrolysis of Wet Wood under External Heat Flux, Fire Safety Journal, 42(3): , Alves, S. S., and Fueiredo, J. L. A Model for Pyrolysis of Wet Wood, Chemical Engineering Science, 44(12): , Astrom, K. J., and Hagglund, T. PID Controllers-Theory Desn and Tuning, Instrument Society of America, USA, Wang, Y. F., Yang, L. Z., Zhou, X. D., Dai, J. K., Zhou, Y. P., and Deng, Z. H. Experiment Study of the Altitude Effects on Spontaneous Ignition Characteristics of Wood, Fuel, 89(5): , Beaulieu, P. A., and Dembsey, N. A. Flammability Characteristics at Applied Heat Flux Levels up to 200 kw/m 2, Fire and Materials, 32(2): 61-86, Bal, N., and Rein, G. Numerical Investation of the Ignition Delay Time of a Translucent Solid at Hh Radiant Heat Fluxes, Combustion and Flame, 158(6): , Jiang, F., De Ris, J. L., and Khan, M. M. Absorption of Thermal Energy in PMMA by In-Depth Radiation, Fire Safety Journal, 44(1): , Lautenberger, C., and Fernandez-Pello, C. A. Approximate Analytical Solutions for the Transient Mass Loss Rate and Piloted Ignition Time of a Radiatively Heated Solid in the Hh Heat Flux Limit, In: Gottuk, D. T. and Lattimer, B. Y. (Eds.), Fire Safety Science Proceedings of the Ehth International Symposium, pp ,