- Email: [email protected]
A study on the fire extinguishing characteristics of deep-seated fires using the scale model experiment
Fire Safety Journal 80 (2016) 38–45
Contents lists available at ScienceDirect
Fire Safety Journal journal homepage: www.elsevier.com/locate/firesaf
A study on the fire extinguishing characteristics of deep-seated fires using the scale model experiment Nam Kyun Kim a, Dong Ho Rie b,n a b
Department of Safety Engineering, Incheon National University, Incheon 406-772, Democratic People's Republic of Korea Fire Disaster Prevention Research Center, Incheon National University, Incheon 406-772, Democratic People's Republic of Korea
art ic l e i nf o
a b s t r a c t
Article history: Received 17 November 2014 Received in revised form 28 November 2015 Accepted 9 January 2016 Available online 1 February 2016
NFPA 18 is used as a standard for the performance evaluation of wetting agents used in fire fighting. The performance evaluation of the wetting agent involves three tests: a wood crib fire test, a deep-seated fire test, and a wood fiberboard penetration test. In this study, a scale model experimental device, which can evaluate both extinguishing performance and penetration performance at once, was prepared to evaluate the performance of a wetting agent in various types of porous materials. Then, this result was compared with the results of an NFPA 18 test to validate the result of the novel test. Three types of commercial firefighting solutions and water were used as extinguishing agents in the experiment. The experimental materials used in this performance evaluation were wood crib which is used in the existing class A fire test and wood flour from New Zealand pine trees which has 75% of the domestic market share in wood cribs. The discrimination of the study result used in this study to determine penetration and extinguishing performances was found to be equivalent to the results of the NFPA18 test. Moreover, the study established the validity of the scale model that could check on both penetration and extinguishing performance, permitting a more economical and efficient performance evaluation of wetting agents. & 2016 Elsevier Ltd. All rights reserved.
Keywords: Scale model experiment Penetration performance Extinguishing performance Deep-seated fire Wetting agent
1. Introduction In the case of a deep-seated fire in a wooden building or in a space loaded with porous materials, high temperature could be transmitted into the depths of the structure, in turn increasing the amount of combustion. A deep-seated fire refers to the smoldering status where contact with air is limited so no ignition is possible [1,2]. The incomplete combustion rate of a heat source is much higher in a deep seated fire as compare to a surface fire, which causes a high percentage of fuel to be converted into toxic compounds. In addition, a deep-seated fire cannot be easily detected [3,4]. Because liquid extinguishing agents cannot penetrate far below the surface, deep-seated fires have a risk of re-ignition if they contact air later. Moreover, because a heat source for a deepseated fire cools down slowly, the penetration performance and internal residence period of the extinguishing agent must be outstanding [5]. Due to these aspects of a deep-seated fire, there is a growing demand for a wetting agent that is effective on a deepseated fire in a porous material as well as a method for evaluation of the penetration and extinguishing performances of a wetting agent. n
Corresponding author. E-mail address: [email protected] (D.H. Rie).
http://dx.doi.org/10.1016/j.firesaf.2016.01.003 0379-7112/& 2016 Elsevier Ltd. All rights reserved.
For this, Popovicha et al. investigated the effect of physical and chemical properties of the wet liquid on compressed carbon black and silica powders. They found that a change in the fine structure, caused by wetting, is a significant factor for the diffusion of the liquid and that the diffusion is also related to the viscosity of the wet liquid [6]. Lazghab et al. investigated various methods reported in the literature to evaluate the degree of wetting of solid powder and reported their strength and weakness. Their findings have been used to develop a proper method to collect study data as well as the specifications of a system used in this study [7]. To study a scale fire, Chattaway et al. developed an experimental device, which scaled down to a full-scale fire experiment using a wood crib and conducted a fire extinguishing performance evaluation. They succeeded in validating this setup using this experiment conducted with the aforementioned device [8]. In addition to these studies, there are continuous efforts on the study of a deep-seated fire and its transmission speed and extinguishing of a deep-seated fire [9–11]. The previous study was conducted by separating the penetration and extinguishing performance of the wetting agent in a porous material. However, there is no ongoing study on the complex fire extinguishing property; a correlation between penetration and extinguishing performance. NFPA 18 evaluates the performance of the wetting agent through three independent tests [12]. The wood crib fire test is an actual scale fire test using a wood crib in accordance with UL 711
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
and it incurs a high experiment cost. Moreover, because both the deep-seated fire test and the wood fiberboard penetration test require specific experimental specimens, their applicability is limited [13]. In this study, a scale model experimental device was planned and constructed. This scale model experimental device was designed to evaluate both the penetration and extinguishing performance of the wetting agent in a comprehensive and quantitative fashion, and it can be used for various porous materials including a powder-phase porous material. Also, the experimental device of this study can measure the penetration amounts of liquid extinguishing agents and the internal temperature of holder simultaneously. Thus, it is possible to determine a results comprised correlation between the penetration and extinguishing performance of the wetting agent, which previous studies were not able to ascertain.
39
extinguishing agent running off the outside of the holder [12], a scale for measuring the level of penetration by detecting a change in the weight of the holder, and [13] a scale for measuring of the amount of watering. The specimen holder is designed as a cylindrical stainless steel holder with a diameter and height of 20 cm and 40 cm, respectively, and the bottom of the holder is attached with a mesh net to prevent the runoff of liquid extinguishing agent. On the side of the holder, a 1.5 cm flange is installed every 3 cm for thermocouple insertion. The spray nozzle has a gauge of 1 mm and a spray angle of 45°. The thermocouple can measure temperature up to 900 °C and the scale is a digital strain gauge type which has a minimum measurement unit of 0.1 g. The weight and temperature measured in real-time are saved in the data logger in 1 s increments, and program it to be displayed on the test instrument in real time as to allow real-time observations of movement and extinguishing phenomenon of a deep-seated fire.
2. Experimental methods 2.2. Experimental materials 2.1. Experimental apparatus Fig. 1 shows a scale model experimental device, devised for the study. This experimental device is designed to evaluate both penetration and extinguishing performances. This scale model experimental device is composed of [1] a manometer that displays the internal pressure and watering pressure of a pressure tank [2], a spray nozzle [3], a pressure tank that stores and pressurizes the liquid extinguishing agent [4], a holder that is used as a combustion space by charging specimens [5], a flow regulation valve that controls the flux of the liquid extinguishing agent [6], supports to hold the weights of the specimen holder and pressure tank [7–9], K-type thermocouple to measure a temperature below the surface [10], a data logger and display to show experimental data in real time [11], a scale equipped with a pan to collect and measure the amount of liquid
In this scale model experiment, wood cribs and wood flour were used as experimental materials. For the wood crib, a pine wood crib generally used in a class A fire test was chosen for the study. In the case of the wood crib, a surface carbonization process initializes during combustion and this prevents any further increase in the internal temperature. Consequently, the surface carbonization process stops at a certain stage and the wood crib can maintain its shape [14]. This phenomenon can be observed in wood cribs above a certain level of width. Therefore, if the wood crib is scaled down to a scale model experiment, there would be an increase in the surface area in contact with heat source as well as a decrease in the wood thickness and consequently, accelerate the carbonization process and collapse of the wood crib. To prevent such a collapse of the wood crib in a scale model experiment, the wood cribs were scaled down to a size of 1.5 1.5 1.5 cm and six wood cribs were loaded on top of each other. The volume of the scale model was determined to be 2.205 cm2, or 1/185th of the volume of a wood crib used in a typical fire experiment as described in Chattaway et al. [8]. Fig. 2 shows the formation of wood cribs loaded for the experiment. Radiata pine wood flour, which occupies more than 75% of the domestic wood market, was used and the grain size was maintained between 500 and 1000 μm with a sieve shaker. To maintain a controlled level of moisture content, the wood cribs and wood flour were dried in a dryer at a temperature of 80 °C for 62 h, and their mass was measured every 4 h. The drying process was continued until the change in specimen mass was (70.1) g. In accordance with Eq. (1), the resulting moisture contents for the specimens were 15% for wood cribs and 10.6% for wood flour [8,15]. Eq. (1) states:
MC =
Fig. 1. Schematic diagram of experimental apparatus.
Wm − Wd × 100 Wd
(1)
where MC is the moisture content, Wm is the weight of sample before drying, and Wd is the weight of sample after drying. Fig. 3 shows wood cribs and wood flour used in the experiment. To validate the performance of the scale model experimental device, the experiment was conducted with wetting agents typically used in fire safety evaluation performance. Three types of wetting agents, commercialized in South Korea, were selected, and the concentration of these wetting agents was limited to the median of the recommended concentration for use (if any), to minimize experimental errors. The average concentration of these wetting agents diluted in liquid extinguishing agent was 0.77
40
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
Fig. 2. Loading type of wood crib in the scale model experiment.
(70.24)%, and the average surface tension and viscosity were 30.3 (70.4) mN/m and 1.027( 70.01) cP, respectively. Table 1 shows the physical properties of the three types of wetting agents used in the experiment. 2.3. Experimental conditions and methods Table 2 shows the experimental conditions of scale model experiment. The initial temperature, humidity, and pressure of the lab were 278( 7 3) K, 50( 75)% and 1 atm pressure, and the experiment was repeated three times under the same experimental conditions. Test specimens were heated with a methane burner from a distance of 10 cm and the methane spray pressure was 2.5 bar. The injection angle of the spray nozzle for liquid extinguishing agent was 45°. To evenly spray liquid extinguishing agent onto the surface of the specimen, the height of watering was calculated with Eq. (2):
⎛ θ⎞ h tan ⎜ 90 − ⎟ = ⎝ 2⎠ r
(2)
where θ is the injection angle of nozzle, h is the height of watering, and r represents the radius of specimen holder. Liquid extinguishing agent was sprayed at the height of 24 cm (as calculated by Eq. (2)) and at the speed of 0.5 L/min. The amount of penetration and runoff were measured in real time by determining the change in the mass of liquid extinguishing agent that ran off and the change in the mass of specimen holder. At the same time, the cooling temperature was measured in real time at three points; 2 cm, 5 cm, and 8 cm apart from the lower part of the specimen. Fig. 4 represents a flow chart of the experiment. Wood cribs were loaded in accordance with the instructions of
Table 2 The experimental conditions of the scale model experiment. Fig. 3. Pictures of wood crib and wood flour.
Table 1 Physical properties of extinguishing agents that are used in the experiment. Name of wetting agent
Experimental concentration [%]
Recommended concentration [%]
Surface tension [mN/m]
Viscosity after dilution [cP]
A B C
0.75 0.55 1
0.5–1 0.1–1 1
31 34 26
1.036 1.017 1.027
Division
Experimental Condition
Temperature Pressure
278( 7 3) K 1 atm
Height of injection Position of thermocouples
Division
Humidity Flow rate of liquid extinguishing agent 25 cm Angle of injection pattern (nozzle) 2/5/8 cm from the Distance of methane burner bottom of the sample Gas pressure of methane burner
Experimental Condition 50(7 5)% 0.5 L/min
45° 10 cm 2.5 bar
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
41
Fig. 4. The flow chart of the scale model experiment.
the class A fire test stated in the UL 711, and 3 pieces, 4 pieces and 3 pieces of wood crib were loaded in 2 layers, using 20 pieces of wood crib in total. For extinguishing, 0.1 L of liquid extinguishing agent was sprayed after being heated the wood crib for 90 s. The temperature of thermocouple in the middle and mass loss were 968.0(7 16.20) K and 73.5( 73.02) g just before extinguishing. In a specimen holder, wood flour was stacked to the height of 13 cm and its charging density was maintained at 0.076 g/cm3. To induce a deep-seated fire, heat was applied for 30 min. In addition, liquid extinguishing agent was sprayed after a 30 min period of heat and smoldering transmission. This process was done to minimize the deviation caused by the internal arrangement of the porous material and openings as well as the loss of wood flour due to combustion of the lower part of a test specimen. The temperature of thermocouple in the middle and mass loss were 467.2 (7 3.42) K and 30.1( 72.14) g just before extinguishing. The quantity dispensed was 0.5 L liquid extinguishing agent stored in a liquid extinguishing agent tank. NFPA 18 testing device was prepared according to the experiment standard stated in Sections 6.3.3 and 6.4.3. The holder, used in deep-seated fire test, was 7 in. in height and 4.5 in. in diameter. The weight of used cotton was 100 g. Moreover, Steel rod was 33 mm in height and 35 mm in diameter, and it was used after being heated to 593 °C. The size of wood fiberboard used in wood fiberboard penetration was 305 305 13 mm, and its heat source was metal burner which is same as the one used in Scale model experiment [12].
3. Results 3.1. Extinguishing performance Experimental errors were calculated and it was found that the experimental error rate of thermocouple (T/C) 1, the farthest from a heat source, was 9.8%. Errors of 3.1% were found for T/C 2, located at the center, and for 3.7% T/C 3, closest to the heat source. As the temperature reproducibility of T/C data in the central area is the most precise, the temperature of T/C 2 was chosen as the representative temperature. In the case of extinguishing, the water evaporation temperature of 373 K was chosen as the standard extinguishing temperature and the extinguishing period was calculated with it. To validate the scale model experimental device, a wetting agent experiment that met NFPA 18 standards was also conducted. Fig. 5 represents the real-time temperature results of Wood crib fire test; one of NFPA 18 experiments. In this wood crib fire test, the wood crib collapsed due to carbonization prior to extinguishment so it was not possible to measure an extinguishing period. In a wetting agent extinguishing test, conducted under the same test conditions, there was a 240.3 s extinguishing period on average. Fig. 6 shows the results of a scale model experiment made with a wood crib. In the case of water, after about 200 s from the watering, temperature was rising and re-ignition occurred. After about 400 s from the watering, during a wood crib fire test of NFPA 18, the wood crib collapsed due to carbonization prior to
42
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
Table 3 Extinguishing time result according to the experimental method and liquid extinguishing agent. Liquid extinguishing agent
NFPA 18
Scale model experiment
Wood crib fire test Using wood crib Using wood flour Wetting agent A [s] Wetting agent B [s] Wetting agent C [s] Average of wetting agents [s] Water [s]
Fig. 5. Real-time temperature result of Wood crib fire test in NFPA 18.
234 ( 7 4.24) 229 ( 7 13.59) 258 ( 7 17.72) 240.3
192 ( 7 8.64) 210 ( 7 9.27) 229 (7 6.48) 210.3
68 ( 7 4.55) 67 ( 73.56) 72 (7 5.89) 69.7
Inextinguishable
Inextinguishable Inextinguishable
temperature, induced by watering, and only a consistent increase in temperature was detected at the T/C. In the case of a wetting agent, there was a dramatic decrease in temperature after 55 s, and the extinguishing period was approximately 55 s on average. Table 3 shows the extinguishing period by tests. The average extinguishing period for a wetting agent was found to be 240.3 s in a wood crib fire test, 210.3 s for a scale model experiment using wood crib, and 69.7 s for a scale model experiment using wood flour. In addition because re-ignition was detected in every test with water, it would be proper to assume that these tests have the power to discriminate between extinguishing performances. Also, the extinguishing performance can be confirmed by wood fiberboard penetration test performed as per NFPA 18. It can be confirmed by measuring the weight change before and after experiment. The weight after experiment was measured after drying a sample. The weight change represents the combustion loss of the sample. After going through a drying process in a 105 °C dryer for 12 h, there was a 65 g decrease in water and a 9 g decrease in a wetting agent in comparison with the weight of wood crib prior to the test. 3.2. Penetration performance
Fig. 6. Real-time temperature result of scale model experiment using wood crib.
Fig. 8 shows the results of a deep-seated fire test done by NFPA 18. In the case of water, the amount of runoff was 71.2 g, and a wetting agent had an average runoff amount of 5.9 g. Fig. 9 shows the results of a wood fiberboard penetration test performed as per NFPA 18. In the case of water, there was a runoff amount of 144.4 g, and a wetting agent had an average runoff amount of 93.5 g. Fig. 10 shows the data on runoff amount of a scale model experiment using a wood crib. In the case of water, there was a runoff amount of 17.6 g, and a wetting agent had an average runoff
Fig. 7. Real-time temperature result of scale model experiment using wood flour.
extinguishing. In a wetting agent extinguishing test, conducted under the same test conditions, there was a 210.3 s extinguishing period on average. Fig. 7 shows the results of a scale model experiment with wood flour. In the case of water, the T/C showed no change in
Fig. 8. Real-time amount of runoff result of deep-seated fire test in NFPA 18.
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
43
Table 4 Amount of runoff result according to the experimental method and liquid extinguishing agent. Liquid extinguishing agent
NFPA 18 Wood fiberboard penetration
Using wood crib
Using wood flour
Wetting agent A [g] 0 (7 0)
80.2 (7 3.63)
0 ( 7 0)
Wetting agent B [g] 7.8 ( 7 0.45) Wetting agent C [g] 10.0 ( 7 0.82) Average of wetting 5.9 agents [g] Water [g] 71.2
99.1 ( 7 5.73)
93.5
11.7 ( 7 2.58) 15.6 ( 7 1.77) 15.1 ( 7 2.49) 14.1
144.4
17.6
335.6
Deep-seated fire test
Fig. 9. Real-time amount of runoff result of wood fiberboard penetration in NFPA 18.
Fig. 10. Real-time amount of runoff result of scale model experiment using wood crib.
Scale model experiment
101.0 ( 7 5.40)
0 ( 7 0) 27.5 ( 7 12.75) 12.2
amount of 14.1 g. In addition, there was a decrease in mass, which was attributed to the evaporation of runoff after approximately 300 s caused by re-ignition in the wood crib. Fig. 11 shows the data on runoff amount of a scale model experiment using wood flour. In the case of water, there was a runoff amount of 335.6 g, and a wetting agent had an average runoff amount of 12.2 g. In addition, there was a decrease in a mass, which was attributed to the evaporation of runoff after approximately 300 s caused by re-ignition in the wood crib. Table 4 shows the runoff amount of liquid extinguishing agent to the outside of the holder by test and/or experiment. The average runoff amounts of the wetting agent and water were 5.9 g and 71.2 g respectively in a deep-seated fire test, 93.5 g and 14.4 g in wood fiberboard penetration, 14.1 g and 17.6 g in a scale model experiment using wood flour, and 12.2 g and 335.6 g in a scale model experiment using wood flour. These data verify that the results have the ability to discriminate penetration performances except for the scale model experiment using wood crib.
4. Consideration Q, the quantity of heat of an object, was calculated by Eq. (3):
Q = c × m × ΔT
Fig. 11. Real-time amount of runoff result of scale model experiment using wood flour.
(3)
where c is the specific heat of an object, m is the mass of an object, and ΔT is the change in temperature. In this study, the ratio of the quantity of heat over time was measured based on the quantity of heat of a material at the commencement of liquid extinguishing agent addition to evaluate the extinguishing efficiency. When the specific heat of a material and the change in mass are assumed to be the same in every test, the integral value of the quantity of heat upon time can be defined as a function with temperature as its variable. Because the materials, heat source, and heating period remain identical in the same test. It was assumed that the T/C temperature represents the overall temperature of an object, and the Wood crib fire test data were calculated under the supposition that the temperature was maintained due to collapse in the test with water. In other words, EE, the extinguishing efficiency, was calculated with Eq. (4):
44
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
⎛ ∫ Q E dt ⎞⎟ EE = ⎜⎜ 1 − × 100 ∫ Q T dt ⎟⎠ ⎝
Table 5 Extinguishing efficiency difference between water and average of wetting agents. NFPA 18
⎛ c × m × ∫ ΔTE dt ⎞ ⎟ × 100 = ⎜⎜ 1 − c × m × ∫ ΔTT dt ⎟⎠ ⎝ ⎛ c × m × ∫ (TE − T0 ) dt ⎞ ⎟ × 100 = ⎜⎜ 1 − c × m × ∫ (TT − T0 ) dt ⎟⎠ ⎝
Scale model experiment
Wood crib fire Using wood test crib
(4)
where QE is the quantity of heat for an experiment result, QT is the quantity of heat upon commencement of watering. ΔTE is an increase in temperature (TE) after the experiment in comparison with the initial temperature(T0, 25 °C) of a specimen, and ΔTT is an increase in temperature (TT) upon commencement of watering in comparison with the initial temperature (T0, 25 °C) of a specimen. The extinguishing efficiency is considered as an index showing a degree of decrease in the quantity of heat in comparison with the quantity of heat upon commencement of watering. The ratio of heat quantity is equivalent to the ratio of temperature due to the deletion of specific heat and mass on the denominator and numerator. Thus, the extinguishing efficiency represents a ratio of temperature decrease based on an integral value of the temperature upon commencement of watering. In other words, as the extinguishing efficiency is closer to 100%, it could be interpreted that the extinguishing proceeds at faster speed and in return creating a more dramatic temperature decrease. Additionally, in the case that an integral value of temperature data of the experiments exceeds the integral value of temperature upon commencement of watering, indicating that extinguishing failed and that there was a continuous increase in temperature, the extinguishing efficiency becomes negative. The method of application like this about the integral of the temperature data can be identified even where XFED calculation of ISO 13571 [16]. Fig. 12 shows the extinguishing efficiency calculated with Eq. (4) by test. In the case of a wood crib fire test under NFPA 18, a wetting agent showed an extinguishing efficiency of 73.9% on average and water showed an efficiency of 13.2% on average. In addition, in a scale model experiment using a wood crib, the average extinguishing efficiencies of the wetting agent and water were 78.9% and 19.5%, respectively. In a scale model experiment using wood flour, the average extinguishing efficiencies of the wetting agent and water were 69.7% and 24.2%, respectively, showing that a heat source was not eliminated and there was a continuous increase in the quantity of heat even after watering. To validate the discriminating power of a test method on an extinguishing performance, the difference between the wetting
Fig. 12. Extinguishing efficiency result according to the experimental method and liquid extinguishing agent.
Extinguishing efficiency differ- 60.7 ence [%]
59.4
Using wood flour 91.9
agent and water in extinguishing efficiency was compared and analyzed. Table 5 shows the difference between the wetting agent and water in extinguishing efficiency by test. In the case of a wood crib fire test under NFPA 18, the difference between the wetting agent and water in extinguishing efficiency was 60.7%, and for the scale model experiment using wood crib and scale model experiment using wood flour it was 59.4% and 91.9%, respectively. Therefore, the discriminating power of the scale model experiment was verified to be equivalent or better on an extinguishing performance than NFPA 18. Penetration performance is confirmed through runoff amounts. However, quantitative comparison of the experimental results, which are the Scale model experiment, and deep-seated fire test, and wood fiberboard penetration, may lead to inaccurate results. Because the liquid extinguishing agent amounts and the properties of the targeted porous material differ depending on the test method. Thus, PR, penetration ratio, was calculated with Eq. (5) and then used for a comparative analysis on the penetration performance. Eq. (5) states:
⎛ W ⎞ PR = ⎜ 1 − r ⎟ × 100 ⎝ Ww ⎠
(5)
where Wr is the weight of runoff amounts, and Ww is the weight of watering amounts. Fig. 13 shows penetration ratio calculated with Eq. (5) by test. In the case of a scale model experiment using wood crib, water had a penetration ratio of 82.4% on average and a wetting agent had a ratio of 85.8% on average. In addition, in a deep-seated fire test, the average penetration ratios of the water and wetting agent were 71.9% and 96.2%, respectively. In wood fiberboard penetration, the average penetration ratio of the water and wetting agents were 45.4% and 63.8%, respectively. In a scale model experiment using wood flour, the average extinguishing efficiencies of the water and wetting agents were 67.0% and 98.8%, respectively. To validate the discriminating power of penetration performance in test method as suggested in the study, the difference
Fig. 13. Penetration ratio result according to the experimental method and liquid extinguishing agent.
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
Table 6 Penetration ratio difference between water and average of wetting agents. NFPA 18
Penetration ratio difference [%]
Scale model experiment
Deep-seated fire test
Wood fiberboard penetration
Using Using wood crib wood flour
24.3
18.4
3.4
31.8
between the wetting agent and water in penetration ratio was compared and analyzed. Table 6 shows the difference between the wetting agent and water in a penetration ratio by each test. The difference between the wetting agents and water in a penetration ratio was 3.4%, 24.3%, 18.4%, and 31.8% for the scale model experiment using the wood crib method, the deep-seated fire test of NFPA 18, the wood fiberboard penetration of NFPA 18, and the scale model experiment using wood flour, respectively. Therefore, it could be assumed that the scale model experiment using a wood crib is not sufficient to evaluate penetration performance. In addition, as scale model experiment using wood flour shows the highest difference in penetration ratio, the study confirmed that the discriminating power of a scale model experiment using wood flour on a penetration performance is equivalent to or better than that of NFPA 18.
5. Conclusions This study aims to investigate the reliability of data acquired with a scale model experiment device. For the study, performance evaluation methods of wetting agent listed in NFPA18 and three agents were used to perform a comparative analysis. Major findings of the study are as follows: (1) The difference in the quantity of heat between wetting agent and water, calculated with real-time temperature data, was 26.3% in the scale model experiment using wood crib, 23.7% in the scale model experiment using wood flour, and 31% in the wood crib fire test of NFPA 18. Thus, it is possible to assume that this experiment method suggested in the study induce test/experiment data similar or equivalent with the existing methods. (2) In the case of a wood crib fire test under NFPA 18, the difference between the wetting agent and water in extinguishing efficiency was 60.7%, and for the scale model experiment using wood crib and scale model experiment using wood flour it was 59.4% and 91.9%, respectively. Therefore, the discriminating power of the scale model experiment was verified to be equivalent or better on an extinguishing performance than NFPA 18. (3) For the difference in penetration ratio between wetting agent and water, calculated with real-time data on the amount of penetration, a scale model experiment using a wood crib showed only 3.4%. Thus, it could not verify discrimination on
45
the penetration performance of wood crib. However, because a scale model experiment using a wood flour showed only 31.8%, this experiment method suggested in the study is revealed to be effective on analyzing the extinguishing mechanism for a deep-seated fire with powder-type combustibles. (4) Since the difference in penetration ratio between the wetting agent and water were 24.3% and 18.4% in the deep-seated fire test and wood fiberboard penetration of NFPA 18, respectively, it could be confirmed that the experiment method suggested in the study can perform tests similar or equivalent to the existing methods. (5) These findings show that the experiment method suggested in the study can evaluate both extinguishing and penetration performances of a wetting agent in a deep-seated fire.
Acknowledgment This work was supported by the Incheon National University (International Cooperative). Research Grant in 2012 (No.20120281).
References [1] I.C. Kong, I.G. Park, K.B. Lim, D.H. Rie, A study for characteristics of water that penetrates wood flour due to changes for concentration of BDG, J. Korean Soc. Saf. 28 (2013) 74–79. [2] Fire Prevention Division of National Emergency Management Agency, Carbon dioxide extinguishing of fire safety standards(NFSC 106), NEMA Notice No. 2012-122 (2012). [3] J.S. Kim, D.H. Rie, Downward smoldering fire characteristics of wood chips and wood flour, J. KOSHAM 13 (2013) (2013) 269–274. [4] S.W. Hwang, W.H. Lee, Hardness and dimensional stability of Radiata Pine (Pinus radiata D.Don) heat-compressed wood, Mokchae Konghak 39 (2011) 206–212. [5] National Fire Protection Association, NFPA 12: Standard on Carbon Dioxide Extinguishing Systems, Quincy, USA, 2005. [6] Laura L. Popovich, Donald L. Feke, Ica Manas-Zlocxower, Influence of physical and interfacial characteristics on the wetting and spreading of fluids on powders, Powder Technol. 104 (1999) 68–74. [7] M. Lazghab, K. Saleh, I. Pezron, P. Guigon, L. Komunjer, Wettability assessment of finely divided solids, Powder Technol. 157 (2005) 79–91. [8] A. Chattaway, G.G. Cox, S.R. Preece, D.J. Spring, The development of a smallscale class a fire test, Halon Options Technical Working Conference, vol. 5, 1997, pp. 498–508. [9] B.C. Hagen, V. Frette, G. Kleppe, B.J. Arntzen, Onset of smoldering in cotton: effects of density, Fire Saf. J. 46 (2011) 73–80. [10] F. He, F. Behrendt, Experimental investigation of natural smoldering of char granules in a packed bed, Fire Saf. J. 46 (2011) 406–413. [11] M.T. Gratkowski, N.A. Dembsey, C.L. Beyler, Radiant smoldering ignition of plywood, Fire Saf. J. 41 (2006) 427–443. [12] National Fire Protection Association, NFPA 18: Standard on Wetting Agents, Quincy, USA, 2011. [13] Underwriters Laboratories Inc., Standard for Safety, UL711: Rating and Fire Testing of Fire Extinguishers, Northbrook, USA, 2009. [14] I.H. Yeo, M.O. Yoon, J.B. Yoon, Burning behavior of heavy timber subjected to standard fire, J. Arch. Inst. Korea Struct. Constr. 25 (2009) 123–132. [15] M.H. Kim, Effects of Additives on the Low Pressure Water Mist for Extinguishing Wood Cribs Fire (Master's Thesis), Chungbuk National University, South Korea, 2013. [16] The International Organization for Standardization, ISO 13571:2012 : Lifethreatening components of fire - Guidelines for the estimation of time to compromised tenability in fires, Geneva, Switzerland (2012).
Contents lists available at ScienceDirect
Fire Safety Journal journal homepage: www.elsevier.com/locate/firesaf
A study on the fire extinguishing characteristics of deep-seated fires using the scale model experiment Nam Kyun Kim a, Dong Ho Rie b,n a b
Department of Safety Engineering, Incheon National University, Incheon 406-772, Democratic People's Republic of Korea Fire Disaster Prevention Research Center, Incheon National University, Incheon 406-772, Democratic People's Republic of Korea
art ic l e i nf o
a b s t r a c t
Article history: Received 17 November 2014 Received in revised form 28 November 2015 Accepted 9 January 2016 Available online 1 February 2016
NFPA 18 is used as a standard for the performance evaluation of wetting agents used in fire fighting. The performance evaluation of the wetting agent involves three tests: a wood crib fire test, a deep-seated fire test, and a wood fiberboard penetration test. In this study, a scale model experimental device, which can evaluate both extinguishing performance and penetration performance at once, was prepared to evaluate the performance of a wetting agent in various types of porous materials. Then, this result was compared with the results of an NFPA 18 test to validate the result of the novel test. Three types of commercial firefighting solutions and water were used as extinguishing agents in the experiment. The experimental materials used in this performance evaluation were wood crib which is used in the existing class A fire test and wood flour from New Zealand pine trees which has 75% of the domestic market share in wood cribs. The discrimination of the study result used in this study to determine penetration and extinguishing performances was found to be equivalent to the results of the NFPA18 test. Moreover, the study established the validity of the scale model that could check on both penetration and extinguishing performance, permitting a more economical and efficient performance evaluation of wetting agents. & 2016 Elsevier Ltd. All rights reserved.
Keywords: Scale model experiment Penetration performance Extinguishing performance Deep-seated fire Wetting agent
1. Introduction In the case of a deep-seated fire in a wooden building or in a space loaded with porous materials, high temperature could be transmitted into the depths of the structure, in turn increasing the amount of combustion. A deep-seated fire refers to the smoldering status where contact with air is limited so no ignition is possible [1,2]. The incomplete combustion rate of a heat source is much higher in a deep seated fire as compare to a surface fire, which causes a high percentage of fuel to be converted into toxic compounds. In addition, a deep-seated fire cannot be easily detected [3,4]. Because liquid extinguishing agents cannot penetrate far below the surface, deep-seated fires have a risk of re-ignition if they contact air later. Moreover, because a heat source for a deepseated fire cools down slowly, the penetration performance and internal residence period of the extinguishing agent must be outstanding [5]. Due to these aspects of a deep-seated fire, there is a growing demand for a wetting agent that is effective on a deepseated fire in a porous material as well as a method for evaluation of the penetration and extinguishing performances of a wetting agent. n
Corresponding author. E-mail address: [email protected] (D.H. Rie).
http://dx.doi.org/10.1016/j.firesaf.2016.01.003 0379-7112/& 2016 Elsevier Ltd. All rights reserved.
For this, Popovicha et al. investigated the effect of physical and chemical properties of the wet liquid on compressed carbon black and silica powders. They found that a change in the fine structure, caused by wetting, is a significant factor for the diffusion of the liquid and that the diffusion is also related to the viscosity of the wet liquid [6]. Lazghab et al. investigated various methods reported in the literature to evaluate the degree of wetting of solid powder and reported their strength and weakness. Their findings have been used to develop a proper method to collect study data as well as the specifications of a system used in this study [7]. To study a scale fire, Chattaway et al. developed an experimental device, which scaled down to a full-scale fire experiment using a wood crib and conducted a fire extinguishing performance evaluation. They succeeded in validating this setup using this experiment conducted with the aforementioned device [8]. In addition to these studies, there are continuous efforts on the study of a deep-seated fire and its transmission speed and extinguishing of a deep-seated fire [9–11]. The previous study was conducted by separating the penetration and extinguishing performance of the wetting agent in a porous material. However, there is no ongoing study on the complex fire extinguishing property; a correlation between penetration and extinguishing performance. NFPA 18 evaluates the performance of the wetting agent through three independent tests [12]. The wood crib fire test is an actual scale fire test using a wood crib in accordance with UL 711
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
and it incurs a high experiment cost. Moreover, because both the deep-seated fire test and the wood fiberboard penetration test require specific experimental specimens, their applicability is limited [13]. In this study, a scale model experimental device was planned and constructed. This scale model experimental device was designed to evaluate both the penetration and extinguishing performance of the wetting agent in a comprehensive and quantitative fashion, and it can be used for various porous materials including a powder-phase porous material. Also, the experimental device of this study can measure the penetration amounts of liquid extinguishing agents and the internal temperature of holder simultaneously. Thus, it is possible to determine a results comprised correlation between the penetration and extinguishing performance of the wetting agent, which previous studies were not able to ascertain.
39
extinguishing agent running off the outside of the holder [12], a scale for measuring the level of penetration by detecting a change in the weight of the holder, and [13] a scale for measuring of the amount of watering. The specimen holder is designed as a cylindrical stainless steel holder with a diameter and height of 20 cm and 40 cm, respectively, and the bottom of the holder is attached with a mesh net to prevent the runoff of liquid extinguishing agent. On the side of the holder, a 1.5 cm flange is installed every 3 cm for thermocouple insertion. The spray nozzle has a gauge of 1 mm and a spray angle of 45°. The thermocouple can measure temperature up to 900 °C and the scale is a digital strain gauge type which has a minimum measurement unit of 0.1 g. The weight and temperature measured in real-time are saved in the data logger in 1 s increments, and program it to be displayed on the test instrument in real time as to allow real-time observations of movement and extinguishing phenomenon of a deep-seated fire.
2. Experimental methods 2.2. Experimental materials 2.1. Experimental apparatus Fig. 1 shows a scale model experimental device, devised for the study. This experimental device is designed to evaluate both penetration and extinguishing performances. This scale model experimental device is composed of [1] a manometer that displays the internal pressure and watering pressure of a pressure tank [2], a spray nozzle [3], a pressure tank that stores and pressurizes the liquid extinguishing agent [4], a holder that is used as a combustion space by charging specimens [5], a flow regulation valve that controls the flux of the liquid extinguishing agent [6], supports to hold the weights of the specimen holder and pressure tank [7–9], K-type thermocouple to measure a temperature below the surface [10], a data logger and display to show experimental data in real time [11], a scale equipped with a pan to collect and measure the amount of liquid
In this scale model experiment, wood cribs and wood flour were used as experimental materials. For the wood crib, a pine wood crib generally used in a class A fire test was chosen for the study. In the case of the wood crib, a surface carbonization process initializes during combustion and this prevents any further increase in the internal temperature. Consequently, the surface carbonization process stops at a certain stage and the wood crib can maintain its shape [14]. This phenomenon can be observed in wood cribs above a certain level of width. Therefore, if the wood crib is scaled down to a scale model experiment, there would be an increase in the surface area in contact with heat source as well as a decrease in the wood thickness and consequently, accelerate the carbonization process and collapse of the wood crib. To prevent such a collapse of the wood crib in a scale model experiment, the wood cribs were scaled down to a size of 1.5 1.5 1.5 cm and six wood cribs were loaded on top of each other. The volume of the scale model was determined to be 2.205 cm2, or 1/185th of the volume of a wood crib used in a typical fire experiment as described in Chattaway et al. [8]. Fig. 2 shows the formation of wood cribs loaded for the experiment. Radiata pine wood flour, which occupies more than 75% of the domestic wood market, was used and the grain size was maintained between 500 and 1000 μm with a sieve shaker. To maintain a controlled level of moisture content, the wood cribs and wood flour were dried in a dryer at a temperature of 80 °C for 62 h, and their mass was measured every 4 h. The drying process was continued until the change in specimen mass was (70.1) g. In accordance with Eq. (1), the resulting moisture contents for the specimens were 15% for wood cribs and 10.6% for wood flour [8,15]. Eq. (1) states:
MC =
Fig. 1. Schematic diagram of experimental apparatus.
Wm − Wd × 100 Wd
(1)
where MC is the moisture content, Wm is the weight of sample before drying, and Wd is the weight of sample after drying. Fig. 3 shows wood cribs and wood flour used in the experiment. To validate the performance of the scale model experimental device, the experiment was conducted with wetting agents typically used in fire safety evaluation performance. Three types of wetting agents, commercialized in South Korea, were selected, and the concentration of these wetting agents was limited to the median of the recommended concentration for use (if any), to minimize experimental errors. The average concentration of these wetting agents diluted in liquid extinguishing agent was 0.77
40
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
Fig. 2. Loading type of wood crib in the scale model experiment.
(70.24)%, and the average surface tension and viscosity were 30.3 (70.4) mN/m and 1.027( 70.01) cP, respectively. Table 1 shows the physical properties of the three types of wetting agents used in the experiment. 2.3. Experimental conditions and methods Table 2 shows the experimental conditions of scale model experiment. The initial temperature, humidity, and pressure of the lab were 278( 7 3) K, 50( 75)% and 1 atm pressure, and the experiment was repeated three times under the same experimental conditions. Test specimens were heated with a methane burner from a distance of 10 cm and the methane spray pressure was 2.5 bar. The injection angle of the spray nozzle for liquid extinguishing agent was 45°. To evenly spray liquid extinguishing agent onto the surface of the specimen, the height of watering was calculated with Eq. (2):
⎛ θ⎞ h tan ⎜ 90 − ⎟ = ⎝ 2⎠ r
(2)
where θ is the injection angle of nozzle, h is the height of watering, and r represents the radius of specimen holder. Liquid extinguishing agent was sprayed at the height of 24 cm (as calculated by Eq. (2)) and at the speed of 0.5 L/min. The amount of penetration and runoff were measured in real time by determining the change in the mass of liquid extinguishing agent that ran off and the change in the mass of specimen holder. At the same time, the cooling temperature was measured in real time at three points; 2 cm, 5 cm, and 8 cm apart from the lower part of the specimen. Fig. 4 represents a flow chart of the experiment. Wood cribs were loaded in accordance with the instructions of
Table 2 The experimental conditions of the scale model experiment. Fig. 3. Pictures of wood crib and wood flour.
Table 1 Physical properties of extinguishing agents that are used in the experiment. Name of wetting agent
Experimental concentration [%]
Recommended concentration [%]
Surface tension [mN/m]
Viscosity after dilution [cP]
A B C
0.75 0.55 1
0.5–1 0.1–1 1
31 34 26
1.036 1.017 1.027
Division
Experimental Condition
Temperature Pressure
278( 7 3) K 1 atm
Height of injection Position of thermocouples
Division
Humidity Flow rate of liquid extinguishing agent 25 cm Angle of injection pattern (nozzle) 2/5/8 cm from the Distance of methane burner bottom of the sample Gas pressure of methane burner
Experimental Condition 50(7 5)% 0.5 L/min
45° 10 cm 2.5 bar
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
41
Fig. 4. The flow chart of the scale model experiment.
the class A fire test stated in the UL 711, and 3 pieces, 4 pieces and 3 pieces of wood crib were loaded in 2 layers, using 20 pieces of wood crib in total. For extinguishing, 0.1 L of liquid extinguishing agent was sprayed after being heated the wood crib for 90 s. The temperature of thermocouple in the middle and mass loss were 968.0(7 16.20) K and 73.5( 73.02) g just before extinguishing. In a specimen holder, wood flour was stacked to the height of 13 cm and its charging density was maintained at 0.076 g/cm3. To induce a deep-seated fire, heat was applied for 30 min. In addition, liquid extinguishing agent was sprayed after a 30 min period of heat and smoldering transmission. This process was done to minimize the deviation caused by the internal arrangement of the porous material and openings as well as the loss of wood flour due to combustion of the lower part of a test specimen. The temperature of thermocouple in the middle and mass loss were 467.2 (7 3.42) K and 30.1( 72.14) g just before extinguishing. The quantity dispensed was 0.5 L liquid extinguishing agent stored in a liquid extinguishing agent tank. NFPA 18 testing device was prepared according to the experiment standard stated in Sections 6.3.3 and 6.4.3. The holder, used in deep-seated fire test, was 7 in. in height and 4.5 in. in diameter. The weight of used cotton was 100 g. Moreover, Steel rod was 33 mm in height and 35 mm in diameter, and it was used after being heated to 593 °C. The size of wood fiberboard used in wood fiberboard penetration was 305 305 13 mm, and its heat source was metal burner which is same as the one used in Scale model experiment [12].
3. Results 3.1. Extinguishing performance Experimental errors were calculated and it was found that the experimental error rate of thermocouple (T/C) 1, the farthest from a heat source, was 9.8%. Errors of 3.1% were found for T/C 2, located at the center, and for 3.7% T/C 3, closest to the heat source. As the temperature reproducibility of T/C data in the central area is the most precise, the temperature of T/C 2 was chosen as the representative temperature. In the case of extinguishing, the water evaporation temperature of 373 K was chosen as the standard extinguishing temperature and the extinguishing period was calculated with it. To validate the scale model experimental device, a wetting agent experiment that met NFPA 18 standards was also conducted. Fig. 5 represents the real-time temperature results of Wood crib fire test; one of NFPA 18 experiments. In this wood crib fire test, the wood crib collapsed due to carbonization prior to extinguishment so it was not possible to measure an extinguishing period. In a wetting agent extinguishing test, conducted under the same test conditions, there was a 240.3 s extinguishing period on average. Fig. 6 shows the results of a scale model experiment made with a wood crib. In the case of water, after about 200 s from the watering, temperature was rising and re-ignition occurred. After about 400 s from the watering, during a wood crib fire test of NFPA 18, the wood crib collapsed due to carbonization prior to
42
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
Table 3 Extinguishing time result according to the experimental method and liquid extinguishing agent. Liquid extinguishing agent
NFPA 18
Scale model experiment
Wood crib fire test Using wood crib Using wood flour Wetting agent A [s] Wetting agent B [s] Wetting agent C [s] Average of wetting agents [s] Water [s]
Fig. 5. Real-time temperature result of Wood crib fire test in NFPA 18.
234 ( 7 4.24) 229 ( 7 13.59) 258 ( 7 17.72) 240.3
192 ( 7 8.64) 210 ( 7 9.27) 229 (7 6.48) 210.3
68 ( 7 4.55) 67 ( 73.56) 72 (7 5.89) 69.7
Inextinguishable
Inextinguishable Inextinguishable
temperature, induced by watering, and only a consistent increase in temperature was detected at the T/C. In the case of a wetting agent, there was a dramatic decrease in temperature after 55 s, and the extinguishing period was approximately 55 s on average. Table 3 shows the extinguishing period by tests. The average extinguishing period for a wetting agent was found to be 240.3 s in a wood crib fire test, 210.3 s for a scale model experiment using wood crib, and 69.7 s for a scale model experiment using wood flour. In addition because re-ignition was detected in every test with water, it would be proper to assume that these tests have the power to discriminate between extinguishing performances. Also, the extinguishing performance can be confirmed by wood fiberboard penetration test performed as per NFPA 18. It can be confirmed by measuring the weight change before and after experiment. The weight after experiment was measured after drying a sample. The weight change represents the combustion loss of the sample. After going through a drying process in a 105 °C dryer for 12 h, there was a 65 g decrease in water and a 9 g decrease in a wetting agent in comparison with the weight of wood crib prior to the test. 3.2. Penetration performance
Fig. 6. Real-time temperature result of scale model experiment using wood crib.
Fig. 8 shows the results of a deep-seated fire test done by NFPA 18. In the case of water, the amount of runoff was 71.2 g, and a wetting agent had an average runoff amount of 5.9 g. Fig. 9 shows the results of a wood fiberboard penetration test performed as per NFPA 18. In the case of water, there was a runoff amount of 144.4 g, and a wetting agent had an average runoff amount of 93.5 g. Fig. 10 shows the data on runoff amount of a scale model experiment using a wood crib. In the case of water, there was a runoff amount of 17.6 g, and a wetting agent had an average runoff
Fig. 7. Real-time temperature result of scale model experiment using wood flour.
extinguishing. In a wetting agent extinguishing test, conducted under the same test conditions, there was a 210.3 s extinguishing period on average. Fig. 7 shows the results of a scale model experiment with wood flour. In the case of water, the T/C showed no change in
Fig. 8. Real-time amount of runoff result of deep-seated fire test in NFPA 18.
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
43
Table 4 Amount of runoff result according to the experimental method and liquid extinguishing agent. Liquid extinguishing agent
NFPA 18 Wood fiberboard penetration
Using wood crib
Using wood flour
Wetting agent A [g] 0 (7 0)
80.2 (7 3.63)
0 ( 7 0)
Wetting agent B [g] 7.8 ( 7 0.45) Wetting agent C [g] 10.0 ( 7 0.82) Average of wetting 5.9 agents [g] Water [g] 71.2
99.1 ( 7 5.73)
93.5
11.7 ( 7 2.58) 15.6 ( 7 1.77) 15.1 ( 7 2.49) 14.1
144.4
17.6
335.6
Deep-seated fire test
Fig. 9. Real-time amount of runoff result of wood fiberboard penetration in NFPA 18.
Fig. 10. Real-time amount of runoff result of scale model experiment using wood crib.
Scale model experiment
101.0 ( 7 5.40)
0 ( 7 0) 27.5 ( 7 12.75) 12.2
amount of 14.1 g. In addition, there was a decrease in mass, which was attributed to the evaporation of runoff after approximately 300 s caused by re-ignition in the wood crib. Fig. 11 shows the data on runoff amount of a scale model experiment using wood flour. In the case of water, there was a runoff amount of 335.6 g, and a wetting agent had an average runoff amount of 12.2 g. In addition, there was a decrease in a mass, which was attributed to the evaporation of runoff after approximately 300 s caused by re-ignition in the wood crib. Table 4 shows the runoff amount of liquid extinguishing agent to the outside of the holder by test and/or experiment. The average runoff amounts of the wetting agent and water were 5.9 g and 71.2 g respectively in a deep-seated fire test, 93.5 g and 14.4 g in wood fiberboard penetration, 14.1 g and 17.6 g in a scale model experiment using wood flour, and 12.2 g and 335.6 g in a scale model experiment using wood flour. These data verify that the results have the ability to discriminate penetration performances except for the scale model experiment using wood crib.
4. Consideration Q, the quantity of heat of an object, was calculated by Eq. (3):
Q = c × m × ΔT
Fig. 11. Real-time amount of runoff result of scale model experiment using wood flour.
(3)
where c is the specific heat of an object, m is the mass of an object, and ΔT is the change in temperature. In this study, the ratio of the quantity of heat over time was measured based on the quantity of heat of a material at the commencement of liquid extinguishing agent addition to evaluate the extinguishing efficiency. When the specific heat of a material and the change in mass are assumed to be the same in every test, the integral value of the quantity of heat upon time can be defined as a function with temperature as its variable. Because the materials, heat source, and heating period remain identical in the same test. It was assumed that the T/C temperature represents the overall temperature of an object, and the Wood crib fire test data were calculated under the supposition that the temperature was maintained due to collapse in the test with water. In other words, EE, the extinguishing efficiency, was calculated with Eq. (4):
44
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
⎛ ∫ Q E dt ⎞⎟ EE = ⎜⎜ 1 − × 100 ∫ Q T dt ⎟⎠ ⎝
Table 5 Extinguishing efficiency difference between water and average of wetting agents. NFPA 18
⎛ c × m × ∫ ΔTE dt ⎞ ⎟ × 100 = ⎜⎜ 1 − c × m × ∫ ΔTT dt ⎟⎠ ⎝ ⎛ c × m × ∫ (TE − T0 ) dt ⎞ ⎟ × 100 = ⎜⎜ 1 − c × m × ∫ (TT − T0 ) dt ⎟⎠ ⎝
Scale model experiment
Wood crib fire Using wood test crib
(4)
where QE is the quantity of heat for an experiment result, QT is the quantity of heat upon commencement of watering. ΔTE is an increase in temperature (TE) after the experiment in comparison with the initial temperature(T0, 25 °C) of a specimen, and ΔTT is an increase in temperature (TT) upon commencement of watering in comparison with the initial temperature (T0, 25 °C) of a specimen. The extinguishing efficiency is considered as an index showing a degree of decrease in the quantity of heat in comparison with the quantity of heat upon commencement of watering. The ratio of heat quantity is equivalent to the ratio of temperature due to the deletion of specific heat and mass on the denominator and numerator. Thus, the extinguishing efficiency represents a ratio of temperature decrease based on an integral value of the temperature upon commencement of watering. In other words, as the extinguishing efficiency is closer to 100%, it could be interpreted that the extinguishing proceeds at faster speed and in return creating a more dramatic temperature decrease. Additionally, in the case that an integral value of temperature data of the experiments exceeds the integral value of temperature upon commencement of watering, indicating that extinguishing failed and that there was a continuous increase in temperature, the extinguishing efficiency becomes negative. The method of application like this about the integral of the temperature data can be identified even where XFED calculation of ISO 13571 [16]. Fig. 12 shows the extinguishing efficiency calculated with Eq. (4) by test. In the case of a wood crib fire test under NFPA 18, a wetting agent showed an extinguishing efficiency of 73.9% on average and water showed an efficiency of 13.2% on average. In addition, in a scale model experiment using a wood crib, the average extinguishing efficiencies of the wetting agent and water were 78.9% and 19.5%, respectively. In a scale model experiment using wood flour, the average extinguishing efficiencies of the wetting agent and water were 69.7% and 24.2%, respectively, showing that a heat source was not eliminated and there was a continuous increase in the quantity of heat even after watering. To validate the discriminating power of a test method on an extinguishing performance, the difference between the wetting
Fig. 12. Extinguishing efficiency result according to the experimental method and liquid extinguishing agent.
Extinguishing efficiency differ- 60.7 ence [%]
59.4
Using wood flour 91.9
agent and water in extinguishing efficiency was compared and analyzed. Table 5 shows the difference between the wetting agent and water in extinguishing efficiency by test. In the case of a wood crib fire test under NFPA 18, the difference between the wetting agent and water in extinguishing efficiency was 60.7%, and for the scale model experiment using wood crib and scale model experiment using wood flour it was 59.4% and 91.9%, respectively. Therefore, the discriminating power of the scale model experiment was verified to be equivalent or better on an extinguishing performance than NFPA 18. Penetration performance is confirmed through runoff amounts. However, quantitative comparison of the experimental results, which are the Scale model experiment, and deep-seated fire test, and wood fiberboard penetration, may lead to inaccurate results. Because the liquid extinguishing agent amounts and the properties of the targeted porous material differ depending on the test method. Thus, PR, penetration ratio, was calculated with Eq. (5) and then used for a comparative analysis on the penetration performance. Eq. (5) states:
⎛ W ⎞ PR = ⎜ 1 − r ⎟ × 100 ⎝ Ww ⎠
(5)
where Wr is the weight of runoff amounts, and Ww is the weight of watering amounts. Fig. 13 shows penetration ratio calculated with Eq. (5) by test. In the case of a scale model experiment using wood crib, water had a penetration ratio of 82.4% on average and a wetting agent had a ratio of 85.8% on average. In addition, in a deep-seated fire test, the average penetration ratios of the water and wetting agent were 71.9% and 96.2%, respectively. In wood fiberboard penetration, the average penetration ratio of the water and wetting agents were 45.4% and 63.8%, respectively. In a scale model experiment using wood flour, the average extinguishing efficiencies of the water and wetting agents were 67.0% and 98.8%, respectively. To validate the discriminating power of penetration performance in test method as suggested in the study, the difference
Fig. 13. Penetration ratio result according to the experimental method and liquid extinguishing agent.
N.K. Kim, D.H. Rie / Fire Safety Journal 80 (2016) 38–45
Table 6 Penetration ratio difference between water and average of wetting agents. NFPA 18
Penetration ratio difference [%]
Scale model experiment
Deep-seated fire test
Wood fiberboard penetration
Using Using wood crib wood flour
24.3
18.4
3.4
31.8
between the wetting agent and water in penetration ratio was compared and analyzed. Table 6 shows the difference between the wetting agent and water in a penetration ratio by each test. The difference between the wetting agents and water in a penetration ratio was 3.4%, 24.3%, 18.4%, and 31.8% for the scale model experiment using the wood crib method, the deep-seated fire test of NFPA 18, the wood fiberboard penetration of NFPA 18, and the scale model experiment using wood flour, respectively. Therefore, it could be assumed that the scale model experiment using a wood crib is not sufficient to evaluate penetration performance. In addition, as scale model experiment using wood flour shows the highest difference in penetration ratio, the study confirmed that the discriminating power of a scale model experiment using wood flour on a penetration performance is equivalent to or better than that of NFPA 18.
5. Conclusions This study aims to investigate the reliability of data acquired with a scale model experiment device. For the study, performance evaluation methods of wetting agent listed in NFPA18 and three agents were used to perform a comparative analysis. Major findings of the study are as follows: (1) The difference in the quantity of heat between wetting agent and water, calculated with real-time temperature data, was 26.3% in the scale model experiment using wood crib, 23.7% in the scale model experiment using wood flour, and 31% in the wood crib fire test of NFPA 18. Thus, it is possible to assume that this experiment method suggested in the study induce test/experiment data similar or equivalent with the existing methods. (2) In the case of a wood crib fire test under NFPA 18, the difference between the wetting agent and water in extinguishing efficiency was 60.7%, and for the scale model experiment using wood crib and scale model experiment using wood flour it was 59.4% and 91.9%, respectively. Therefore, the discriminating power of the scale model experiment was verified to be equivalent or better on an extinguishing performance than NFPA 18. (3) For the difference in penetration ratio between wetting agent and water, calculated with real-time data on the amount of penetration, a scale model experiment using a wood crib showed only 3.4%. Thus, it could not verify discrimination on
45
the penetration performance of wood crib. However, because a scale model experiment using a wood flour showed only 31.8%, this experiment method suggested in the study is revealed to be effective on analyzing the extinguishing mechanism for a deep-seated fire with powder-type combustibles. (4) Since the difference in penetration ratio between the wetting agent and water were 24.3% and 18.4% in the deep-seated fire test and wood fiberboard penetration of NFPA 18, respectively, it could be confirmed that the experiment method suggested in the study can perform tests similar or equivalent to the existing methods. (5) These findings show that the experiment method suggested in the study can evaluate both extinguishing and penetration performances of a wetting agent in a deep-seated fire.
Acknowledgment This work was supported by the Incheon National University (International Cooperative). Research Grant in 2012 (No.20120281).
References [1] I.C. Kong, I.G. Park, K.B. Lim, D.H. Rie, A study for characteristics of water that penetrates wood flour due to changes for concentration of BDG, J. Korean Soc. Saf. 28 (2013) 74–79. [2] Fire Prevention Division of National Emergency Management Agency, Carbon dioxide extinguishing of fire safety standards(NFSC 106), NEMA Notice No. 2012-122 (2012). [3] J.S. Kim, D.H. Rie, Downward smoldering fire characteristics of wood chips and wood flour, J. KOSHAM 13 (2013) (2013) 269–274. [4] S.W. Hwang, W.H. Lee, Hardness and dimensional stability of Radiata Pine (Pinus radiata D.Don) heat-compressed wood, Mokchae Konghak 39 (2011) 206–212. [5] National Fire Protection Association, NFPA 12: Standard on Carbon Dioxide Extinguishing Systems, Quincy, USA, 2005. [6] Laura L. Popovich, Donald L. Feke, Ica Manas-Zlocxower, Influence of physical and interfacial characteristics on the wetting and spreading of fluids on powders, Powder Technol. 104 (1999) 68–74. [7] M. Lazghab, K. Saleh, I. Pezron, P. Guigon, L. Komunjer, Wettability assessment of finely divided solids, Powder Technol. 157 (2005) 79–91. [8] A. Chattaway, G.G. Cox, S.R. Preece, D.J. Spring, The development of a smallscale class a fire test, Halon Options Technical Working Conference, vol. 5, 1997, pp. 498–508. [9] B.C. Hagen, V. Frette, G. Kleppe, B.J. Arntzen, Onset of smoldering in cotton: effects of density, Fire Saf. J. 46 (2011) 73–80. [10] F. He, F. Behrendt, Experimental investigation of natural smoldering of char granules in a packed bed, Fire Saf. J. 46 (2011) 406–413. [11] M.T. Gratkowski, N.A. Dembsey, C.L. Beyler, Radiant smoldering ignition of plywood, Fire Saf. J. 41 (2006) 427–443. [12] National Fire Protection Association, NFPA 18: Standard on Wetting Agents, Quincy, USA, 2011. [13] Underwriters Laboratories Inc., Standard for Safety, UL711: Rating and Fire Testing of Fire Extinguishers, Northbrook, USA, 2009. [14] I.H. Yeo, M.O. Yoon, J.B. Yoon, Burning behavior of heavy timber subjected to standard fire, J. Arch. Inst. Korea Struct. Constr. 25 (2009) 123–132. [15] M.H. Kim, Effects of Additives on the Low Pressure Water Mist for Extinguishing Wood Cribs Fire (Master's Thesis), Chungbuk National University, South Korea, 2013. [16] The International Organization for Standardization, ISO 13571:2012 : Lifethreatening components of fire - Guidelines for the estimation of time to compromised tenability in fires, Geneva, Switzerland (2012).