- Email: [email protected]
Measurement of geometric and radiative properties of heptane pool fires
Fire Safety Journal 96 (2018) 13–26
Contents lists available at ScienceDirect
Fire Safety Journal journal homepage: www.elsevier.com/locate/firesaf
Measurement of geometric and radiative properties of heptane pool fires Vinay C. Raj, S.V. Prabhu * Department of Mechanical Engineering, Indian Institute of Technology, Bombay, India
A R T I C L E I N F O
A B S T R A C T
Keywords: Heptane pool fires Emissivity Temperature Irradiance Thermal camera
The objective of the present work is to characterize heptane pool fires for pool diameters ranging from 0.1 m to 1.0 m, in a quiescent environment. Characterization of pool fires is done by determining various parameters such as the mass burning rate, puffing frequency, flame height, optical thickness, spatial distribution of emissivity and temperature, irradiance and fire safety distance. Measurement of the instantaneous mass burning rate indicates the presence of two steady states: an initial steady state and a bulk boiling steady state. The mass burning rate is found to decrease with increase in the free board height. Flame height is determined based on the definition of intermittency. Puffing frequency is obtained from the visible images by tracking the vortical structures in flames. The obtained results for the flame height and puffing frequency match well with the correlations presented in the literature. Flame emissivity is determined from the mass burning rate as well as a refined technique that determines the transmissivity of the electrically heated strips placed behind the flame. Radiative properties such as temperature distribution and irradiance at a distance are calculated from the thermal images. The irradiance at a distance obtained using the infrared camera is compared with Schimdt-Boelter gauge. The effect of fire size on the optical thickness, radiative fraction and fire safety distance has been examined.
1. Introduction A pool fire may be defined as a diffusion flame burning above a pool of vaporizing hydrocarbon fuel where buoyancy is the predominant transport mechanism. Air, from the surroundings, diffuses into the reacting zone resulting in combustion. Processing industries, such as the paint industry, make use of n-heptane as a solvent. They incur the risk associated with the occurrence of pool fires. Knowledge of the radiative environment of potential fire scenarios is very helpful for planning firefighting strategies. This allows for determination of whether a particular fire can be approached or not. It also allows for determination of which equipments to be used and what strategy should be employed in an emergency response plan [1]. Significant progress has been made in understanding the mass burning rate and heat feedback mechanisms to pool fires in quiescent air. The reviews by Joulian [2], Steinhaus et al. [3] and Hu [4] elaborates the state of the art research and addresses various characteristics of pool fires such as the burning rate, heat feedback mechanism, flame morphological characteristics, radiation and soot production. Despite the enormous body of work on large scale pool fires, there are still significant uncertainties in our understanding of such fires and capabilities to predict
their behavior [3]. Extensive research, conducted over the last few decades in understanding the burning characteristics of liquid pool fires, is summarized below.
1.1. Mass burning rate Mass burning rate is defined as the mass of fuel consumed per unit area per unit time. For gasoline pool fires, Babrauskas [5] observed that the mass burning rate increased for diameters up to 2 m and was nearly constant thereafter. Blinov and Khudyakov [6] observed that the burning rate for single-component fuels initially decreases at a rapid rate, reaches a minimum and finally increases with diameter before reaching their asymptotic value for large diameters. Parameters such as environmental conditions and lip height also affect the mass burning rate. Mass loss rate is routinely measured as the decrease in the height of fuel [7] or decrease in the weight of the pan containing the fuel [8]. However, Hamins [9] obtained the mass burning rate indirectly by determining the heat feedback to the fuel surface. The analysis was carried out by applying a heat balance in the liquid fuel of the pool fire for a quasi-steady state system. Fig. 1 shows the mass burning rate of n-heptane pool fire reported in various literature [7,9–16]. It is
* Corresponding author. Department of Mechanical Engineering, Indian Institute of Technology, Bombay, Powai, Mumbai, 400 076, India. E-mail address: [email protected] (S.V. Prabhu). https://doi.org/10.1016/j.firesaf.2017.12.003 Received 29 April 2017; Received in revised form 11 December 2017; Accepted 11 December 2017 0379-7112/© 2017 Elsevier Ltd. All rights reserved.
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Nomenclature a,b A A1 A2 Cp D E f F g h Δhg HPF I k L m_ '' OS Q Q* q_'' r T x,y
Empirical constants Area of pool fire (m2) Pixel Area Sensor Area Specific heat capacity, (J/kg K) Pool diameter (m) Emissive power (W/m2) Puffing frequency (Hz) View Factor acceleration due to gravity (m/s2) Heat transfer coefficient (W/m2K) Total heat of gasification (kJ/kg) Heptane Pool Fire Intensity obtained from thermal camera (W/m2.sr) Thermal conductivity (W/mK) Flame Length (m)
X/D Y/D Z Z/D
Non-dimensional radial distance Non-dimensional distance along the axis of the flame Distance between the flame and sensor plane Non-dimensional distance parallel to the flame plane
Subscripts ∞ b bf f i, j r t TC
Ambient; Infinite Body Body through flame Flame Pixel location Radiative Total Thermocouple
Greek symbols ε Emissivity κβ Absorption-extinction coefficient(m1) σ Stefan Boltzmann Constant (W/m2K4) ρ Density (kg/m3) τ Transmissivity
Mass loss rate per unit area (kg/m2s) Object signal Heat released (W) Non-dimensional heat release rate
Non-dimensional numbers
2
Heat flux (kW/m ) Radial distance (m) Temperature (K) Coordinates in the flame plane
Pr ¼
μCp
Re ¼
ρVD μ
k
Prandtl number Reynolds number
altered the mass burning rates. The effect of lip height on mass burning rate of heptane pool fires has not been studied. This is essential to explain the difference in the mass burning rates reported throughout the literature.
perplexing to observe that the mass burning rate for a particular pool diameter has been reported sometimes more than double. 1.2. Effect of lip height
1.3. Puffing frequency
Freeboard height or the lip height is the distance between the top of the pan to the top of the fuel level. Babrauskas [5] observed that the lip height significantly influences the convective and radiative heat flux when the liquid level is allowed to run down in the vessel. A fundamental study elucidating the effect of lip height on the burning rates and on flame size of small scale fires have been investigated by Dlugogorski and Wilson [17]. They measured the burning rates of ethanol fires in copper, mild steel and glass vessel for diameters less than 7 cm. They observed that the onset of fuel boiling in copper and steel cylinders dramatically
Malalasekara et al. [18] defined puffing as quasi-periodic oscillations of the diffusion flame front near the axisymmetric source of a fire with the formation of large scale flaming vortical structures. The frequency at which puffing occurs is called puffing frequency. Various researchers have determined the puffing frequency from visual as well as thermal images obtained using high speed cameras. Cetegen and Ahmed [19] recorded the pressure fluctuation across the face of the burner by means of a pressure tapping located at one quarter of the way along the burner diameter. The puffing frequency was determined by means of a frequency analyzer. They concluded that puffing is a result of buoyant flow instability which arises due to strong interaction of the toroidal vortex formed at a short distance above the burner surface. Equations (1a) and (1b) are used to determine the puffing frequency, as given by Cetegen and Ahmed [19] and Malalasekara et al. [18] respectively. 1:5 f ¼ pffiffiffiffi D
(1a)
1:68 f ¼ pffiffiffiffi D
(1b)
Experimental investigations have been carried out to determine the pulsation frequency of pool fires of different aspect ratios in subatmospheric pressures [20]. Hu et al. [21] demonstrated that the Rayleigh-Taylor instability and puffing instabilities are the main reason for the necking-in and the periodic oscillatory behavior. The dominant instability mechanism transits from the extended Rayleigh-Taylor instability to the puffing instability with increase in pool size and lip height. Fig. 1. Mass burning rate of heptane pool fires. 14
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
1.4. Flame length
1.6. Flame temperature distribution
Zukoski et al. [22] determined the mean flame height in terms of Intermittency, which is defined as the fraction of time for which the flame's length is at least higher than L. The average length of a flame (L) is defined as the length at which the intermittency reaches a value of 0.5. Munoz et al. [23] also determined the flame length by means of the visible images obtained from video recordings. Heskestad [24] developed an empirical equation for the flame length for a wide range of hydrocarbon fuels. Equation (2) shows the non-dimensional flame height as a function of the non-dimensionalized heat release rate.
Temperature distribution is another essential parameter to characterize pool fires. Over the years, various researchers have used high temperature thermocouples to measure the temperature distribution. McCaffrey [30] measured the centerline temperature of natural gas flame using K-type thermocouples without correcting for any radiation errors. Koseki [31] measured the time-averaged centerline temperature for heptane pool fire for diameters from 0.3 m to 5 m. Hayasaka et al. [32] conducted pool fire experiments on 4.85 cm diameter cylinder with heptane, kerosene and methanol as fuels and measured the centerline temperature distribution using thermocouples. Silvani and Morandini [33] have demonstrated that thermocouples located inside the flame front indicate a temperature lower than the gas temperature. This is attributed to the radiative loss from the thermocouple junction to the colder surroundings. The correction for radiative loss is necessary to obtain the true gas temperature; however, it is a very difficult process. Infrared cameras have been extensively used to measure the temperature of flame. Hayasaka et al. [32] introduced the methodology to employ high-speed thermography to measure the radiative heat flux from kerosene pool fires of 2.7 m square tank. The infrared camera captures the irradiance from the flame and transforms it into temperature values by accepting emissivity as an input parameter. The temperature distribution hence obtained is an apparent temperature distribution of the flame assuming the flame as a two-dimensional wall. Sudheer and Prabhu [34] have characterized hexane pool fires by means of infrared thermography. The temperature distribution of pool fires has been primarily obtained by using an appropriate single value of emissivity as an input to the thermal camera. The temperature values thus obtained is an approximate distribution due to the assumption of a constant emissivity value. Thus, it is of utmost importance to accurately determine the spatial distribution of temperature.
2=5 L ¼ 1:02 þ 3:7 Q* D Q* ¼
Q pffiffiffi D5=2 Cp T∞ ρ∞ g
(2a)
(2b)
where Q is the heat released; Q ¼ m_ " AΔHc (W); D is the diameter of pool fire (m); Cp is the specific heat capacity (J/kg K); T∞ is the atmospheric temperature (K) and ρ∞ is the density of surrounding air (kg/m3). Thomas [25] provided a highly simplified dimensionless analysis to relate flame height to mass burning rate. Equation (3) shows the non-dimensional flame height as a function of the mass burning rate. #0:61 " . L= ¼ 42 m_ 00 0:5 ρ∞ ðgDÞ D
(3)
where, m_ '' : mass burning rate (kg/m2s); ρ∞ : density of surrounding air (kg/m3), g: acceleration due to gravity (m/s2). Mangialavori and Rubino [26] have proposed an empirical equation for the maximum flame height of various fuels like heptane, hexane and iso-butane fires for pool diameters varying from 1 m to 6 m. The equation is of the form b' L ¼ a' m_ " D
1.7. Optical thickness
(4)
Optical thickness is defined as the ability of the pool fire to attenuate radiation along a path length. Flame emissivity in terms of optical thickness is given by Babrauskas [5] as
where, a and b are empirical constants; a ¼ 31.6 and b ¼ 0.58. 1.5. Emissivity distribution
εf ¼ 1 eκβD
Emissivity is one of the key parameters that determine the amount of radiation emitted by a pool fire. Emissivity of a pool fire is primarily due to the hot combustion products formed during the combustion process. It is thus a combination of emissivity of soot and hot gases. Limitations of earlier methodologies in the determination of emissivity [8,27,28] include assumption of axisymmetric flame and the data being spatially and temporally averaged. Raj and Prabhu [29] proposed a refined methodology to determine the spatial and temporal variation in the emissivity of diffusion flames in the optically thin regime. The instantaneous variation of the flame emissivity is obtained by determining the transmissivity of the electrically heated strips placed behind the flame using an infrared camera. The thermal camera collects infrared radiation (It) equal to the sum of the infrared radiation proceeding from the flame (If) and the infrared radiation proceeding from the body (Ib) passing through the flame. This is represented by Eq. (5). It ¼ If þ τf Ib
(6)
1.8. Irradiance measurement (a) Emissive Power Distribution By knowing the flame emissivity, the temperature distribution from thermal camera can be converted to irradiance distribution as: Ei;j ¼ σεf Ti;j4
(7)
where Ti,j is the temperature (K) of a pixel element, i and j indicate the location in the thermal image, Ei,j is the emissive power (kW/m2) at that location. Munoz et al. [23] have determined the emissive power of gasoline and diesel pool fires of diameter greater than 1.5 m. The emissivity values were set equal to 1 as in the case of optically thick pool fires (D > 1.5 m). Irradiance at a distance allows for the determination of fire safety distance. Fire safety distance is the distance at which the thermal radiation does not cause damage to humans, instruments and buildings. Sudheer et al. [35] have measured the heat flux at a distance by means of a Schmidt–Boelter gauge and also using an infrared camera. From the thermal images, the heat flux is determined as shown in Eq. (8). The view factors have been calculated for a rectangle to rectangle configuration in
(5)
It is of utmost importance to determine the complete distribution of emissivity and temperature as it allows for calculation of the true emissive power and fire safety distances. The present paper aims at providing the complete distribution of emissivity for heptane pool fires.
15
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
2. Description of experimental set-up
parallel and perpendicular planes as provided in Ref. [36]. q0012 ¼
F12 A1 E1 A2
A series of outdoor pool fire experiments are conducted with heptane as the fuel. All the experiments are conducted in an open space to allow natural air flow into the fire. The wind velocity is monitored using a vane anemometer. The velocity is predominantly between 0.5 m/s to 1 m/s. The fuel temperature and the ambient air temperature are at 33 C and the relative humidity is about 50%. The pan is made of mild steel of varying diameters from 0.1 m to 1.0 m. a) Mass burning rate: A 15 kg platform type weighing scale (Contech CT 15K1) is used for measuring the mass burning rate of pool fires of diameter 0.1 m–0.3 m. For pool fires of diameter 0.5 m, 0.7 m and 1.0 m, a 250 kg platform scale of Eldigi make is used. For determining the instantaneous mass burning rate, the lip height was initially zero and kept increasing as the fuel is consumed. b) Flame Height and Puffing frequency: The experiments are filmed using Canon EOS Digital Camera at 60 fps. Images are also obtained from thermal camera (Thermoteknix make VisIR® 640). The cameras are placed at different distances, depending upon the size of the pool fires, such that the complete flame is captured. c) Emissivity distribution: The schematic diagram for measurement of emissivity is shown in Fig. 2. A detailed technical description of the experimental setup used to determine the spatial distribution of emissivity and temperature is reported by Raj and Prabhu [29]. The experimental setup consists of 40 horizontal nichrome strips held firmly between two wooden planks and electrically insulated by means of fiber glass tape. The 40 horizontal strips are placed in such a way as to cover the entire flame length. The 40 strips are divided into 4 groups of 10 strips each. Each nichrome strip is 1.2 m in length, 0.01 m in width and 0.001 m in thickness. The ten strips are connected in series and are then connected to an AC power supply through a variac. The local emissivity distribution in heptane pool fires has been determined for pool diameters of 0.3 m, 0.5 m, 0.7 m and 1.0 m. The strips are painted with Pyromark 2500 flat black, a paint whose emissivity is a constant in the working range of the thermal camera [38,39]. The strips are placed behind the pool fire and are electrically heated to 500K. The pool fire is ignited and allowed to reach a steady mass burning rate. Fig. 3 shows the infrared images of the strips as seen through the flame. Table 1 provides the details of the thermal camera used in the present study. d) Temperature and heat flux: Temperature is measured using a
(8)
where, F1-2: View factor from pixel A1 to sensor area A2; A1: Pixel Area; A2: Sensor Area; E1: Emissive power. (b) Radiative fraction Radiative fraction is an important parameter which represents the fraction of heat released in the form of radiation. Hamins et al. [37] and Buch et al. [12] have determined the radiative fraction for heptane pool fire of diameter 0.3 m. The total energy radiated is calculated by integrating the energy flux through the cylindrical surface [37], as in Eq (9). ! R ∞ 00 00 _ Qr ¼ 2π ∫ R0 r:q ðrÞ:dr þ R0 ∫ 0 qz ðzÞ:dz
(9)
While consolidated correlations exist for estimating the flame height, radiative fraction and fire safety distances, there exists a significant degree of error in predicting various other parameters that characterize pool fires. The mass burning rate for heptane pool fires reported in the literature vary by more than two times [13]. The effect of lip height has not been captured and the uncertainty in the flame height is of the order of pool diameter [31]. There is no data on the local emissivity and temperature distribution of heptane pool fires. In an effort to improve the accuracy of the data of heptane pool fires, a series of experiments are carried out and various parameters that characterize pool fires are determined. The objective of the present study is two-fold: 1. To characterize heptane pool fires in terms of geometric and radiative parameters. 2. To address the short comings in the literature in terms of mass burning rate, effect of lip height, spatial distribution of emissivity and temperature. The characterization of heptane pool fires for diameters ranging from 0.1 m to 1.0 m would serve as a benchmark data for numerical simulation. This would aid the fire modeling community seeking to characterize fuels at bench top scale for later correlation to hazards in large fires.
Fig. 2. Schematic of the experimental set-up. 16
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
strips and the intensity of the heated body (Ib) is obtained prior to the start of the experiments. Thus, substituting the intensity values into Eq. (5), the transmissivity of the flame is obtained. A MATLAB Code performs the pixel-wise calculation of the transmissivity. Furthermore, with the assumption of zero flame reflection, emissivity is determined as:
εf ¼ 1 τf
(10)
By substituting the obtained values of emissivity at each pixel location, the temperature distribution is obtained. (d) Temperature measurement using thermocouples: The temperatures obtained from the thermocouples have been corrected for radiation losses. Silvani and Morandini [33] have provided for temperature correction for a thermocouple inside a flame front, as shown in Eq. (11). As soot gets deposited on the thermocouple bead, the emissivity of the bead has been taken as 0.8 [33]. An average value of the flame emissivity (εf) obtained using the refined technique has been used for the correction of temperature. The heat transfer coefficient in Eq. (11b) is obtained as flow over a cylinder from Nusselt number correlation provided in Bergman et al. [40]. dTC is the bead diameter (0.7 mm). Thermal conductivity (kf), Reynolds number and Prandtl number are determined for dry air at the true flame temperature by an iteration process.
Fig. 3. Instantaneous thermal image of pool fire.
Tf ¼ TTC þ
Table 1 Specifications of uncooled thermal camera used in the present study.
h¼
®
Model
VisIR 640
Detector Spectral range Pixel Resolution FOV Temperature range Accuracy Frames per second
Microbolometer Uncooled FPA 7.5–13 μm 640 480 25 19 0 C–2000 C 2% of reading 60 Hz
σεTC 1 εf Tf4 h þ 4σεTC Tf3
(11a)
kf 0:43 þ 0:53Re0:5 Pr0:31 dTC
(11b)
(e) Uncertainty Analysis: The uncertainty of various parameters obtained in experimental analysis is as shown in Table 2. The uncertainty in bead diameter is obtained by measuring it using a screw gauge. The uncertainty in the measurement of temperature by means of thermocouple is after the correction for radiation losses. 4. Results and discussion A series of pool fire experiments are conducted with heptane as fuel for diameters ranging from 0.1 m to 1.0 m. Heptane pool fires have been characterized by determining the mass burning rate, flame height, puffing frequency, optical thickness, spatial distribution of emissivity and temperature, radiative fraction and fire safety distance. Spatial variation of flame emissivity is inferred from the transmissivity of the flame measured using infrared thermography.
thermocouple tree consisting of 14 thermocouples placed at a distance of 50 mm from each other. The thermocouples are connected to a data logger (Agilent make) and the emf is recorded every 0.7 s. SchmidtBoelter gauge is placed at Z/D ¼ 1 and traversed in the vertical direction (Y/D) to measure the irradiance at various vertical locations. 3. Data reduction (a) Mass burning rate: To obtain the mass burning rate, a platform type weighing scale is used. It is interfaced with a computer through an RS232 port. Data is transferred every 0.1 s for smaller pool diameter (0.1 m–0.3 m) and every 0.2 s for larger pool diameter (0.5 m, 0.7 m and 1.0 m). To determine the effect of lip height on mass burning rate, the pan was partially filled to different levels and the mass loss rate is recorded. (b) Flame Height and puffing frequency: All the experiments are filmed using Canon EOS Digital Camera at 60 fps. Each frame is then extracted from the recorded video using the software provided by the camera manufacturer. A MATLAB program determines the maximum height of the visible portion of the flame from the instantaneous images. It also performs the Fast Fourier Transform (FFT) of the intensity signal to determine the puffing frequency. (c) Emissivity and temperature distribution: The infrared camera obtains a two-dimensional image of the three dimensional intensity distribution of the flame. Considering a 2D image of the flame with the pixels in the x-y plane, the net intensity measured by the infrared camera is a line integral along the flame thickness. The flame emissivity is obtained from the calculation of the transmissivity of the flame with reference to electrically heated strips [29]. For Eq. (5), the total infrared radiation (It) is obtained at locations where the horizontal strips are placed. The intensity of the flame (If) is obtained at locations with no
4.1. Mass burning rate Mass burning rate is one of the key parameters used to characterize pool fires, as it determines the amount of heat released. Mass burning rate is determined by measuring the weight loss of the pan containing the fuel. The pan, filled with heptane, is placed on top of the platform scale which is interfaced with a computer to record the mass loss data. The overall as well as the instantaneous mass burning rate is determined for a period of 25 min. The instantaneous mass burning rate is obtained by differentiation of the weight loss data and smoothing it out. Fig. 4 shows Table 2 Experimental uncertainties.
17
Parameter
Relative uncertainty (%)
Mass burning rate Flame length Emissivity from mass burning rate Emissivity from heated strips Bead diameter of thermocouple Thermocouple temperature Thermal camera temperature Schmidt-Boelter gauge
0.2 0.2 0.2 10 1.4 8 2 7
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 5. Effect of lip height on mass burning rate of heptane pool fire (HPF). Fig. 4. Instantaneous mass burning rate for heptane pool fire of various pool diameters.
shows the sequence of still images, for a pool fire of diameter 0.7 m, where puffing can be observed. A point as shown in Fig. 6(g) is chosen and the intensity at that point is tracked. Puffing frequency is thus determined by tracking the “flaming vortical structures” from the visible images. A MATLAB program is developed to perform a Fast Fourier Transform (FFT) of the intensity signal. The FFT provides the dominant frequency which is the puffing frequency as shown in Fig. 7. Table 3 compares the experimentally obtained puffing frequency with the empirical equation proposed by Cetegen and Ahmed [19] and Malalasekara et al. [18]. The empirical equations are obtained from the best possible fits from a large number of data points. The difference in the fits is probably due to the errors involved in retrieving the values from published reports and graphs [18]. The puffing frequency is found to match closely with that reported by Cetegen and Ahmed.
the variation of mass burning rate with normalized time which is obtained by dividing the time with the total time of the experiment. Fluctuations observed in the instantaneous mass burning rate data are caused by changes in the wind velocity. Four different stages for mass burning rate can be observed in Fig. 4: The first stage is the initial transient stage where the fire rapidly spreads and develops through the entire pool. The fuel surface temperature rises rapidly and reaches the boiling point. Small bubbles on the fuel surface can be seen. In the second stage, the flame reaches a steady state wherein the mass burning rate becomes constant. The third stage is reached after approximately 60% of the total time where the mass burning rate increases to a second plateau. This is understood to be due to the bulk boiling behavior where the fuel temperature exceeds a particular value and bubbles due to boiling are seen on the fuel surface. As more amount of fuel is at a vaporizing temperature, the mass loss rate is enhanced [13]. The last phase is the decay period where the fuel is burnt out and the fire is extinguished.
4.4. Flame length The camera is placed normal to the fire, at an average height and at a sufficient distance to capture the entire flame height. The video recordings that were used to determine the puffing frequency is used to determine the flame height. A MATLAB program is developed to identify the maximum height of the visible part of the flame. The average length of flame is obtained for an intermittency value of 0.5, as defined by Zukoski et al. [22]. Fig. 8 shows the intermittency for 0.3 m heptane pool fire. Fig. 9 shows the average flame height obtained experimentally and its comparison with the correlation proposed by Heskestad [24] and Thomas [25]. The comparison has been done for both the initial as well as the bulk boiling steady states. The flame height is observed to increase by an order of pan diameter due to the increase in the mass loss rate from initial steady state to bulk boiling steady state.
Fig. 1 shows the comparison of the experimentally obtained mass burning rate with that reported in various literature [7,9–16]. m1 denotes the mass burning rate for the initial steady state while m2 denotes the mass burning rate during the bulk boiling behavior. The mass burning rate obtained during the bulk boiling period is found to be almost twice that of the initial steady state. The large difference in the mass burning rates reported by various researchers can thus be attributed to the measurement of the burning rates at either the initial steady state or during the bulk boiling period. 4.2. Effect of lip height
4.5. Emissivity and temperature distribution
The effect of having a non-zero freeboard height is found to be significant in pool fires. Experiments are conducted on 0.2 m and 0.3 m circular pool fires to determine the lip effect. From Fig. 5 it can be observed that for heptane pool fires the mass burning rate decreases with increase in the free board height. Dlugogorcki and Wilson [17] stated that natural convection greatly influences the burning rates in vessels made from conductive materials. They observed that the onset of fuel boiling occurred at progressively increasing freeboard height which dramatically altered the mass burning rates. The uncertainty in the measurement of mass burning rate reported throughout the literature can also be attributed to this effect.
Emissivity has been determined from the mass burning rate as well as the refined technique that determines the transmissivity of the electrically heated strips placed behind the flame. The complete temperature distribution is then determined from the local emissivity distribution using it as an input to the thermal camera. 4.5.1. Using mass burning rate Babrauskas [5], by applying conservation of energy at the base of pool fire, determined emissivity as the ratio of the mass burning rate of a pool fire of particular diameter to the mass burning rate of a pool fire of infinite diameter. For heptane, the mass burning rate for a pool fire of infinite diameter is 101 g/m2s [5]. Fig. 10 shows the emissivity obtained from mass burning rate for both the initial steady state and during the bulk boiling period. The emissivity for heptane pool fires is found to increase with the increase in pool diameter as the mass burning rate also
4.3. Puffing frequency Video recordings of the experiments, captured using Canon EOS digital camera, are exported as digital images at 0.05 s interval. Fig. 6 18
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 6. Sequence of images showing puffing in pool fires of diameter 0.7 m.
Fig. 7. Power spectrum for 0.7 m pool fire indicating puffing frequency.
increases with pool diameter.
Table 3 Puffing frequency with Pool diameter. Pool Diameter
Puffing Frequency (Hz)
(m)
Measured Values
Cetegen and Ahmed [19] 1:5 ffiffiffi f ¼p D
Malalasekara et al. [18] pffiffiffi f ¼ 1:68 D
0.1 0.2 0.3 0.5 0.7 1.0
4.8 3.36 2.76 2.16 1.78 1.62
4.74 3.35 2.74 2.12 1.79 1.5
5.31 3.76 3.07 2.38 2.01 1.68
4.5.2. Using refined technique with electrically heated strips placed behind the flame The refined technique developed by Raj and Prabhu [29] is used to determine the complete spatial distribution of emissivity and temperature. For each pool diameter, 100 instantaneous infrared images are averaged to obtain the average thermal image of the strips as seen through the pool fire. The total intensity (It) is measured with the heated strips seen through the flame. The flame intensity (If) is measured two pixels to the top or bottom of the strips and intensity of strips (Ib) is measured without the flame. Fig. 11(a) shows the contour plot of emissivity distribution for the
19
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 11. (a) Contour plot of emissivity distribution (b) Contour plot of temperature distribution of heptane pool fire (HPF) of diameter 0.3 m. Fig. 8. Intermittency for heptane pool fire (HPF) of 0.3 m pool diameter.
horizontal axis (X/D) and the vertical axis (Y/D) are non-dimensionalized by dividing the axis with the diameter of the pool fire. The average emissivity at the base of the pool fire is obtained to be 0.4. While traversing from the core region of the pool fire to the plume region, the emissivity is found to decrease from 0.3 to 0.1. From the emissivity values obtained through the above methodology, temperature contours are obtained by substituting the values into the thermal images. Fig. 11(b) shows the contour plot of temperature distribution. The maximum temperature is around 1200K while the average temperature is about 750K. The thermal images are a representation of the apparent temperatures as the intensity measured by the thermal camera is line integrated over the depth of view. Fig. 12(a) and (b) show the radial variation of emissivity and temperature at various non-dimensional heights (Y/D) for heptane pool fire of diameter 0.3 m. Emissivity decreases with height from a value of 0.375 to a value of 0.25. The temperature values are also found to decrease from 1000K at the base of the pool fire to 700K in the plume region. It is observed that the emissivity and temperature profiles shift from a parabolic profile to a top hat profile along the length of the pool fire. Fig. 13(a) and (b) show the contour plot for emissivity and temperature variation for the averaged images of heptane pool fire of diameter 0.5 m. The average emissivity at the base of the pool fire is about 0.45. The emissivity values vary from 0.5 at the core region to 0.2 in the plume region. The maximum temperature is around 1050K while the average temperature is about 900K. Fig. 14(a) and (b) show the radial variation of emissivity and temperature along the length of the pool fire. It can be observed that the core region has a constant emissivity and temperature values. Fig. 15(a) shows the contour plot for the emissivity distribution for heptane pool fire of diameter 0.7 m. The average emissivity at the base of the pool fire is around 0.5 to 0.6. The emissivity at the center region is around 0.5 and decreases to 0.3 in the plume region. Fig. 15(b) shows the contour plot for the temperature distribution for heptane pool fire of diameter 0.7 m. The maximum temperature is around 1250K while the average temperature is about 900K. Fig. 16(a) and (b) show the radial variation of emissivity and temperature along the length of the pool fire. The temperatures are high at the base of the pool fire. As the combustion products move up due to buoyancy, they get cooled by mixing with the ambient air. Hence, there is a decrease in temperature values as the particles move up from the base of the pool fire. Fig. 17(a) and (b) show the contour plot of emissivity and temperature distribution for heptane pool fire of diameter 1.0 m. The average emissivity at the base of the pool fire is 0.6. The emissivity at the core region is around 0.6 and decreases to 0.5 towards the plume region. The emissivity values are high in comparison to lower pool diameters due to the presence of higher amount of soot particles. The high temperature
Fig. 9. Comparison of average flame height obtained experimentally with correlations of Heskestad [24] and Thomas [25].
Fig. 10. Comparison between the emissivity obtained by mass burning rate with the average emissivity obtained by the refined technique.
averaged images of heptane pool fire of 0.3 m diameter. The origin of contour plots is chosen at the center of the base of the pool. The
20
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 14. (a) Radial emissivity variation (b) Radial temperature variation of heptane pool fire (HPF) of diameter 0.5 m. Fig. 12. (a) Radial emissivity variation (b) Radial temperature variation of heptane pool fire (HPF) of diameter 0.3 m.
Fig. 15. (a) Contour plot of emissivity distribution (b) Contour plot of temperature distribution of heptane pool fire (HPF) of diameter 0.7 m. Fig. 13. (a) Contour plot of emissivity distribution (b) Contour plot of temperature distribution of heptane pool fire (HPF) of diameter 0.5 m.
zones at Y/D of 0.8 shows two strong vortices resulting from air entrainment. The maximum temperature is around 1150K while the average temperature is about 1000K. Fig. 18(a) and (b) show the radial variation of emissivity and temperature along the length of the pool fire. The emissivity and temperature values are uniform in the core region.
obtained from the refined technique is spatially averaged. Fig. 10 shows the comparison of emissivity values obtained from the mass burning rate with the average emissivity values obtained from the refined technique obtained during the bulk boiling period. It can be observed that the overall emissivity values are in good agreement with the average values obtained from the refined technique. It is imperative to point out that the use of an average emissivity will lead to significant errors in the spatial distribution of temperature values.
4.5.3. Comparison of emissivity obtained from mass burning rate and the refined technique Throughout the literature, various researchers have always used a single value to represent the emissivity of pool fire for a given diameter. Thus, to represent the overall emissivity and to compare it with that obtained from mass burning rate, the complete distribution of emissivity
4.5.4. Comparison of centerline temperature distribution obtained from thermal camera and thermocouple K-type thermocouples, placed at a distance of 50 mm from each other, are used to measure the centerline temperature. The data is recorded at every 0.7 s. Once the temperatures have reached steady state, an average is taken for a period of 5 min. Fig. 19 shows the comparison of the 21
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 18. (a) Radial emissivity variation (b) Radial temperature variation of heptane pool fire (HPF) of diameter 1.0 m.
Fig. 16. (a) Radial emissivity variation (b) Radial temperature variation of heptane pool fire (HPF) of diameter 0.7 m.
Fig. 17. (a) Contour plot of emissivity distribution (b) Contour plot of temperature distribution of heptane pool fire (HPF) of diameter 1.0 m.
22
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
(a) 0.3 m diameter
(b) 0.5 m diameter
(c) 0.7 m diameter Fig. 19. Comparison of centerline temperature measured by thermocouple and thermal camera.
4.6. Optical thickness
centerline temperature measured by thermocouple with that obtained from the thermal camera. It can be observed that the temperatures measured by the thermocouple are approximately 200 C higher than that obtained by the thermal camera. This can be attributed to the fact that the thermal camera captures the equivalent intensity of a three dimensional flame in a two dimensional plane. This results in a decrease in the total intensity measured as the data obtained is line integrated over the depth of view. Fig. 20 shows the comparison of the average centerline temperature obtained from the thermocouples for heptanes pool fires under study in the present paper with the centerline temperature reported by Koseki [31] for heptane pool fires and that reported by McCaffrey [30] for natural gas. The temperatures obtained match well with both Koseki and McCaffrey's data. The difference is observed for higher pool diameters of 0.7 m and 1.0 m and this is attributed to fact that McCaffrey's model is valid only for low sooting fuels (see Fig. 20).
Optical thickness in terms of flame emissivity is provided in Eq. (6). Rearranging the equation and taking natural log provides, eκβD ¼ 1 εf κβD ¼ ln 1 εf
(14)
The slope of the plot between the pool diameter (D) on the abscissa and ln(1-εf) gives the product of the absorption-extinction coefficient of the flame and the mean beam corrector length, represented by κβ, as shown in Fig. 21. The obtained value of 1.04 m-1 is found to match well with the value reported by Babrauskas [5], which is 1.1 (0.3) m1. 4.7. Irradiance measurement 4.7.1. Heat flux at a distance Emissive power from the pool fire is determined from the obtained values of emissivity and temperature using Eq. (7). Fig. 22(a) and (b) show the comparison of the heat flux obtained at a distance of 71 cm and
Fig. 20. Comparison of centerline temperatures of heptane pool fires and natural gas.
Fig. 21. Optical thickness for heptane pool fire. 23
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 22. Comparison of heat flux measurements at Z/D ¼ 1 with Schmidt–Boelter gauge and infrared camera for heptane pool fires.
133 cm with that reported by Hamins et al. [37] and Buch et al. [12] respectively. The heat flux almost doubles when the mass loss rate shifts from the initial steady state to the bulk boiling period. Fig. 22(c)–(e) show the comparison of heat flux measured using a Schmidt–Boelter gauge and that determined from the infrared camera using Eq. (8). The irradiance values obtained from the infrared camera, calculated at Z/D ¼ 1, matches well with that measured by the Schmidt–Boelter gauge. The heat flux measured at Z/D ¼ 1 is found to increase with increase in pool diameter as the total emissive power of pool fires also increases with pool diameter. The heat flux at various Z/D is calculated using Eq. (8). A heat flux of 1.4 kW/m2 is harmless for persons without any special protection and a heat flux of 4.7 kW/m2 causes pain in 15–20 s and burns after 30 s [41]. Fire safety distance is the maximum value of Z/D which matches with the above mentioned heat fluxes. Table 4 shows the fire safety distance for heptane pool fires of various pool diameters.
Table 4 Fire safety distances for heptane pool fires. Pool Diameter
Fire safety distance (Z/D)
(m)
1.4 kW/m2 (harmless for persons without any special protection)
4.7 kW/m2 (causes pain in 15–20 s and burns after 30 s)
0.3 0.5 0.7 1
3.67 4 4 4.5
1.97 2 2.25 2.5
heptane pool fires is calculated using Eq. (9). The radiative fraction for heptane pool fires is found to increase from 0.25 to 0.33 for the pool diameters considered in the present study as seen in Fig. 23. The values obtained are comparable to that reported in other literature [7,14,31, 37]. Hamins et al. [37] reported that the radiance measurement made at a single location and corrected for anisotropic effects yield an estimate of radiative fraction within 13% of the multi-location measurements.
4.7.2. Radiative fraction The radial and vertical distributions of the heat flux are calculated from the thermal image as given by Eq. (8). The total radiative power of
24
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
References [1] S. Sudheer, S.V. Prabhu, Measurement of flame emissivity of hydrocarbon pool fires, Fire Technol. 48 (2) (2012) 183–217. [2] P. Joulain, The behavior of pool fires: state of the art and new insights, in: Symposium (International) on Combustion (Vol. 27, No. 2, pp. 2691–2706), Elsevier, 1998, January. [3] T. Steinhaus, S. Welch, R.O. Carvel, J.L. Torero, Large-scale pool fires, Therm. Sci. 11 (2) (2007) 101–118. [4] L. Hu, A review of physics and correlations of pool fire behaviour in wind and future challenges, in: IAFSS 12th Symposium 2017, 91, July 2017, pp. 41–55. https://doi. org/10.1016/j.firesaf.2017.05.008. [5] V. Babrauskas, Estimating large pool fire burning rates, Fire Technol. 19 (4) (1983) 251–261. [6] V.I. Blinov, G.N. Khudyakov, Diffusion Burning of Liquids (No. AERDL-T-1490-A), Army Engineer Research and Development Labs Fort Belvoir VA, 1961. [7] K. Hiroshi, Y. Taro, Air entrainment and thermal radiation from heptane pool fires, Fire Technol. 24 (1) (1988) 33–47. [8] S. Sudheer, S.V. Prabhu, Measurement of flame emissivity of gasoline pool fires, Nucl. Eng. Des. 240 (10) (2010) 3474–3480. [9] A. Hamins, S.J. Fischer, T. Kashiwagi, M.E. Klassen, J.P. Gore, Heat feedback to the fuel surface in pool fires, Combust. Sci. Technol. 97 (1–3) (1994) 37–62. [10] H. Hayasaka, Unsteady burning rates of small pool fires, Fire Saf. Sci. 5 (1997) 499–510. [11] J. Gore, M. Klassen, A. Hamins, T. Kashiwagi, Fuel property effects on burning rate and radiative transfer from liquid pool flames, Fire Saf. Sci. 3 (1991) 395–404. [12] R. Buch, A. Hamins, K. Konishi, D. Mattingly, T. Kashiwagi, Radiative emission fraction of pool fires burning silicone fluids, Combust. Flame 108 (1) (1997) 118–126. [13] B. Chen, S. Lu, C. Li, Q. Kang, M. Yuan, Unsteady burning of thin-layer pool fires, J. Fire Sci. 30 (1) (2012) 3–15. [14] H. Koseki, H. Hayasaka, Estimation of thermal balance in heptane pool fire, J. Fire Sci. 7 (4) (1989) 237–250. [15] H. Koseki, G.W. Mulholland, The effect of diameter on the burning of crude oil pool fires, Fire Technol. 27 (1) (1991) 54–65. [16] A. Shinotake, S. Koda, K. Akita, An experimental study of radiative properties of pool fires of an intermediate scale, Combust. Sci. Technol. 43 (1–2) (1985) 85–97. [17] B.Z. Dlugogorski, M.T. Wilson, Effect of lip height on properties of small scale pool fires, Fire Saf. Sci. 2 (1995) 129–140. [18] W.M.G. Malalasekera, H.K. Versteeg, K. Gilchrist, A review of research and an experimental study on the pulsation of buoyant diffusion flames and pool fires, Fire Mater. 20 (6) (1996) 261–271. [19] B.M. Cetegen, T.A. Ahmed, Experiments on the periodic instability of buoyant plumes and pool fires, Combust. Flame 93 (1–2) (1993) 157–184. [20] F. Tang, L. Hu, Q. Wang, Z. Ding, Flame pulsation frequency of conductioncontrolled rectangular hydrocarbon pool fires of different aspect ratios in a subatmospheric pressure, Int. J. Heat Mass Tran. 76 (2014) 447–451. [21] L. Hu, J. Hu, J.L. de Ris, Flame necking-in and instability characterization in small and medium pool fires with different lip heights, Combust. Flame 162 (4) (2015) 1095–1103. [22] E.E. Zukoski, B.M. Cetegen, T. Kubota, Visible structure of buoyant diffusion flames, in: Symposium (International) on Combustion (Vol. 20, No. 1, pp. 361–366), Elsevier, 1985, January. [23] M. Mu~ noz, J. Arnaldos, J. Casal, E. Planas, Analysis of the geometric and radiative characteristics of hydrocarbon pool fires, Combust. Flame 139 (3) (2004) 263–277. [24] G. Heskestad, Luminous heights of turbulent diffusion flames, Fire Saf. J. 5 (2) (1983) 103–108. [25] P.H. Thomas, The size of flames from natural fires, in: Symposium (International) on Combustion (Vol. 9, No. 1, pp. 844–859), Elsevier, 1963, January. [26] G. Mangialavori, F. Rubino, Experimental tests on large-scale hydrocarbon pool fires, in: 7th International Symposium on Loss Prevention and Safety Promotion in Process Ind., Taormina, Italy, 1992, 83–1. [27] E. Planas-Cuchi, J.M. Chatris, C. L opez, J. Arnaldos, Determination of flame emissivity in hydrocarbon pool fires using infrared thermography, Fire Technol. 39 (3) (2003) 261–273. [28] C. Qian, K. Saito, Measurements of pool-fire temperature using IR technique, in: Proceedings Combustion Fundamentals and Applications, Combustion Institute/ Central and Western States, Mexican National Section, and American Flame Research Committee, San Antonio, Texas, 1995, pp. 81–86. [29] V.C. Raj, S.V. Prabhu, A refined methodology to determine the spatial and temporal variation in the emissivity of diffusion flames, Int. J. Therm. Sci. 115 (2017) 89–103. [30] B.J. McCaffrey, Purely Buoyant Diffusion Flames: Some Experimental Results, 1979. [31] H. Koseki, Combustion properties of large liquid pool fires, Fire Technol. 25 (3) (1989) 241–255. [32] H. Hayasaka, H. Koseki, Y. Tashiro, Radiation measurements in large-scale kerosene pool flames using high-speed thermography, Fire Technol. 28 (2) (1992) 110–122. [33] X. Silvani, F. Morandini, Fire spread experiments in the field: temperature and heat fluxes measurements, Fire Saf. J. 44 (2) (2009) 279–285. [34] S. Sudheer, S.V. Prabhu, Characterization of hexane pool fires using infrared thermography, J. Fire Sci. 31 (2) (2013) 143–165. [35] S. Sudheer, L. Kumar, B.S. Manjunath, A. Pasi, G. Meenakshi, S.V. Prabhu, Fire safety distances for open pool fires, Infrared Phys. Technol. 61 (2013) 265–273.
Fig. 23. Radiative fraction of heptane pool fire.
5. Conclusions Characterization of heptane pool fires for diameter ranging from 0.1 m to 1.0 m has been carried out by measuring key parameters like mass burning rate, effect of lip height, flame length, puffing frequency, optical thickness, radiative fraction and spatial distribution of emissivity and temperature. Mass burning rates for heptane pool fire increases with the increase in pool diameter. The instantaneous mass burning rate and total mass burning rate have been determined. From the instantaneous mass burning rate, it is observed that heptane pool fires depict two plateaus for burning rate. The first plateau is for the initial steady state while the second plateau is for the bulk boiling behavior. The effect of having a non-zero free board height is observed to significantly alter the mass burning rate for pool diameters of 0.2 m and 0.3 m. It is observed that the mass burning rate decreases with increase in free board height. Puffing frequency is determined from the visible images by tracking the intensity about a point where the vortical structures are seen. The results obtained are found to match well with Cetegen and Ahmed [19]. Maximum flame height is determined from the visible images and the average flame height is determined from the definition of intermittency of the flame. Flame emissivity has been determined as the ratio of the mass burning rate of the particular pool diameter to the mass burning rate of pool fire whose diameter is large. Spatial distribution of emissivity and temperature has also been determined by a refined technique. The emissivity and temperature are found to increase with increase in pool fire diameter and decrease along the height. The κβ value for heptane pool fires is obtained to be 1.04 m-1. The irradiance almost doubles for the second plateau of mass burning rate. The irradiance measured at a distance of Z/D ¼ 1 with infrared camera and Schmidt–Boelter gauge match reasonably well. The fire safety distances are approximately 4D. Radiative fraction for heptane pool fires varies from 0.25 to 0.33 with increase in pool diameter from 0.3 m to 1.0 m. Acknowledgements Authors thank Mr. Rahul Shirsat for his help in building the experimental setup.
25
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26 [39] J.T. Nakos, J.M. Suo-Anttila, W. Gill, Shroud Boundary Condition Characterization Experiments at the Radiant Heat Facility, SAND2004-5080, October, 2004. [40] T.L. Bergman, F.P. Incropera, D.P. DeWitt, A.S. Lavine, Fundamentals of Heat and Mass Transfer, John Wiley & Sons, 2011. [41] L. Zarate, J. Arnaldos, J. Casal, Establishing safety distances for wildland fires, Fire Saf. J. 43 (8) (2008) 565–575.
[36] J.R. Howell, A Catalog of Radiation Heat Transfer, 2016. http://www. thermalradiation.net/indexCat.html. [37] A. Hamins, M. Klassen, J. Gore, T. Kashiwagi, Estimate of flame radiance via a single location measurement in liquid pool fires, Combust. Flame 86 (3) (1991) 223–228. [38] R.S. Longenbaugh, L.C. Sanchez, A.R. Mahoney, Thermal response of a Small Scale Cask-like Test Article to Three Different High Temperature Environments, Federal Railroad Administration, 1990.
26
Contents lists available at ScienceDirect
Fire Safety Journal journal homepage: www.elsevier.com/locate/firesaf
Measurement of geometric and radiative properties of heptane pool fires Vinay C. Raj, S.V. Prabhu * Department of Mechanical Engineering, Indian Institute of Technology, Bombay, India
A R T I C L E I N F O
A B S T R A C T
Keywords: Heptane pool fires Emissivity Temperature Irradiance Thermal camera
The objective of the present work is to characterize heptane pool fires for pool diameters ranging from 0.1 m to 1.0 m, in a quiescent environment. Characterization of pool fires is done by determining various parameters such as the mass burning rate, puffing frequency, flame height, optical thickness, spatial distribution of emissivity and temperature, irradiance and fire safety distance. Measurement of the instantaneous mass burning rate indicates the presence of two steady states: an initial steady state and a bulk boiling steady state. The mass burning rate is found to decrease with increase in the free board height. Flame height is determined based on the definition of intermittency. Puffing frequency is obtained from the visible images by tracking the vortical structures in flames. The obtained results for the flame height and puffing frequency match well with the correlations presented in the literature. Flame emissivity is determined from the mass burning rate as well as a refined technique that determines the transmissivity of the electrically heated strips placed behind the flame. Radiative properties such as temperature distribution and irradiance at a distance are calculated from the thermal images. The irradiance at a distance obtained using the infrared camera is compared with Schimdt-Boelter gauge. The effect of fire size on the optical thickness, radiative fraction and fire safety distance has been examined.
1. Introduction A pool fire may be defined as a diffusion flame burning above a pool of vaporizing hydrocarbon fuel where buoyancy is the predominant transport mechanism. Air, from the surroundings, diffuses into the reacting zone resulting in combustion. Processing industries, such as the paint industry, make use of n-heptane as a solvent. They incur the risk associated with the occurrence of pool fires. Knowledge of the radiative environment of potential fire scenarios is very helpful for planning firefighting strategies. This allows for determination of whether a particular fire can be approached or not. It also allows for determination of which equipments to be used and what strategy should be employed in an emergency response plan [1]. Significant progress has been made in understanding the mass burning rate and heat feedback mechanisms to pool fires in quiescent air. The reviews by Joulian [2], Steinhaus et al. [3] and Hu [4] elaborates the state of the art research and addresses various characteristics of pool fires such as the burning rate, heat feedback mechanism, flame morphological characteristics, radiation and soot production. Despite the enormous body of work on large scale pool fires, there are still significant uncertainties in our understanding of such fires and capabilities to predict
their behavior [3]. Extensive research, conducted over the last few decades in understanding the burning characteristics of liquid pool fires, is summarized below.
1.1. Mass burning rate Mass burning rate is defined as the mass of fuel consumed per unit area per unit time. For gasoline pool fires, Babrauskas [5] observed that the mass burning rate increased for diameters up to 2 m and was nearly constant thereafter. Blinov and Khudyakov [6] observed that the burning rate for single-component fuels initially decreases at a rapid rate, reaches a minimum and finally increases with diameter before reaching their asymptotic value for large diameters. Parameters such as environmental conditions and lip height also affect the mass burning rate. Mass loss rate is routinely measured as the decrease in the height of fuel [7] or decrease in the weight of the pan containing the fuel [8]. However, Hamins [9] obtained the mass burning rate indirectly by determining the heat feedback to the fuel surface. The analysis was carried out by applying a heat balance in the liquid fuel of the pool fire for a quasi-steady state system. Fig. 1 shows the mass burning rate of n-heptane pool fire reported in various literature [7,9–16]. It is
* Corresponding author. Department of Mechanical Engineering, Indian Institute of Technology, Bombay, Powai, Mumbai, 400 076, India. E-mail address: [email protected] (S.V. Prabhu). https://doi.org/10.1016/j.firesaf.2017.12.003 Received 29 April 2017; Received in revised form 11 December 2017; Accepted 11 December 2017 0379-7112/© 2017 Elsevier Ltd. All rights reserved.
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Nomenclature a,b A A1 A2 Cp D E f F g h Δhg HPF I k L m_ '' OS Q Q* q_'' r T x,y
Empirical constants Area of pool fire (m2) Pixel Area Sensor Area Specific heat capacity, (J/kg K) Pool diameter (m) Emissive power (W/m2) Puffing frequency (Hz) View Factor acceleration due to gravity (m/s2) Heat transfer coefficient (W/m2K) Total heat of gasification (kJ/kg) Heptane Pool Fire Intensity obtained from thermal camera (W/m2.sr) Thermal conductivity (W/mK) Flame Length (m)
X/D Y/D Z Z/D
Non-dimensional radial distance Non-dimensional distance along the axis of the flame Distance between the flame and sensor plane Non-dimensional distance parallel to the flame plane
Subscripts ∞ b bf f i, j r t TC
Ambient; Infinite Body Body through flame Flame Pixel location Radiative Total Thermocouple
Greek symbols ε Emissivity κβ Absorption-extinction coefficient(m1) σ Stefan Boltzmann Constant (W/m2K4) ρ Density (kg/m3) τ Transmissivity
Mass loss rate per unit area (kg/m2s) Object signal Heat released (W) Non-dimensional heat release rate
Non-dimensional numbers
2
Heat flux (kW/m ) Radial distance (m) Temperature (K) Coordinates in the flame plane
Pr ¼
μCp
Re ¼
ρVD μ
k
Prandtl number Reynolds number
altered the mass burning rates. The effect of lip height on mass burning rate of heptane pool fires has not been studied. This is essential to explain the difference in the mass burning rates reported throughout the literature.
perplexing to observe that the mass burning rate for a particular pool diameter has been reported sometimes more than double. 1.2. Effect of lip height
1.3. Puffing frequency
Freeboard height or the lip height is the distance between the top of the pan to the top of the fuel level. Babrauskas [5] observed that the lip height significantly influences the convective and radiative heat flux when the liquid level is allowed to run down in the vessel. A fundamental study elucidating the effect of lip height on the burning rates and on flame size of small scale fires have been investigated by Dlugogorski and Wilson [17]. They measured the burning rates of ethanol fires in copper, mild steel and glass vessel for diameters less than 7 cm. They observed that the onset of fuel boiling in copper and steel cylinders dramatically
Malalasekara et al. [18] defined puffing as quasi-periodic oscillations of the diffusion flame front near the axisymmetric source of a fire with the formation of large scale flaming vortical structures. The frequency at which puffing occurs is called puffing frequency. Various researchers have determined the puffing frequency from visual as well as thermal images obtained using high speed cameras. Cetegen and Ahmed [19] recorded the pressure fluctuation across the face of the burner by means of a pressure tapping located at one quarter of the way along the burner diameter. The puffing frequency was determined by means of a frequency analyzer. They concluded that puffing is a result of buoyant flow instability which arises due to strong interaction of the toroidal vortex formed at a short distance above the burner surface. Equations (1a) and (1b) are used to determine the puffing frequency, as given by Cetegen and Ahmed [19] and Malalasekara et al. [18] respectively. 1:5 f ¼ pffiffiffiffi D
(1a)
1:68 f ¼ pffiffiffiffi D
(1b)
Experimental investigations have been carried out to determine the pulsation frequency of pool fires of different aspect ratios in subatmospheric pressures [20]. Hu et al. [21] demonstrated that the Rayleigh-Taylor instability and puffing instabilities are the main reason for the necking-in and the periodic oscillatory behavior. The dominant instability mechanism transits from the extended Rayleigh-Taylor instability to the puffing instability with increase in pool size and lip height. Fig. 1. Mass burning rate of heptane pool fires. 14
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
1.4. Flame length
1.6. Flame temperature distribution
Zukoski et al. [22] determined the mean flame height in terms of Intermittency, which is defined as the fraction of time for which the flame's length is at least higher than L. The average length of a flame (L) is defined as the length at which the intermittency reaches a value of 0.5. Munoz et al. [23] also determined the flame length by means of the visible images obtained from video recordings. Heskestad [24] developed an empirical equation for the flame length for a wide range of hydrocarbon fuels. Equation (2) shows the non-dimensional flame height as a function of the non-dimensionalized heat release rate.
Temperature distribution is another essential parameter to characterize pool fires. Over the years, various researchers have used high temperature thermocouples to measure the temperature distribution. McCaffrey [30] measured the centerline temperature of natural gas flame using K-type thermocouples without correcting for any radiation errors. Koseki [31] measured the time-averaged centerline temperature for heptane pool fire for diameters from 0.3 m to 5 m. Hayasaka et al. [32] conducted pool fire experiments on 4.85 cm diameter cylinder with heptane, kerosene and methanol as fuels and measured the centerline temperature distribution using thermocouples. Silvani and Morandini [33] have demonstrated that thermocouples located inside the flame front indicate a temperature lower than the gas temperature. This is attributed to the radiative loss from the thermocouple junction to the colder surroundings. The correction for radiative loss is necessary to obtain the true gas temperature; however, it is a very difficult process. Infrared cameras have been extensively used to measure the temperature of flame. Hayasaka et al. [32] introduced the methodology to employ high-speed thermography to measure the radiative heat flux from kerosene pool fires of 2.7 m square tank. The infrared camera captures the irradiance from the flame and transforms it into temperature values by accepting emissivity as an input parameter. The temperature distribution hence obtained is an apparent temperature distribution of the flame assuming the flame as a two-dimensional wall. Sudheer and Prabhu [34] have characterized hexane pool fires by means of infrared thermography. The temperature distribution of pool fires has been primarily obtained by using an appropriate single value of emissivity as an input to the thermal camera. The temperature values thus obtained is an approximate distribution due to the assumption of a constant emissivity value. Thus, it is of utmost importance to accurately determine the spatial distribution of temperature.
2=5 L ¼ 1:02 þ 3:7 Q* D Q* ¼
Q pffiffiffi D5=2 Cp T∞ ρ∞ g
(2a)
(2b)
where Q is the heat released; Q ¼ m_ " AΔHc (W); D is the diameter of pool fire (m); Cp is the specific heat capacity (J/kg K); T∞ is the atmospheric temperature (K) and ρ∞ is the density of surrounding air (kg/m3). Thomas [25] provided a highly simplified dimensionless analysis to relate flame height to mass burning rate. Equation (3) shows the non-dimensional flame height as a function of the mass burning rate. #0:61 " . L= ¼ 42 m_ 00 0:5 ρ∞ ðgDÞ D
(3)
where, m_ '' : mass burning rate (kg/m2s); ρ∞ : density of surrounding air (kg/m3), g: acceleration due to gravity (m/s2). Mangialavori and Rubino [26] have proposed an empirical equation for the maximum flame height of various fuels like heptane, hexane and iso-butane fires for pool diameters varying from 1 m to 6 m. The equation is of the form b' L ¼ a' m_ " D
1.7. Optical thickness
(4)
Optical thickness is defined as the ability of the pool fire to attenuate radiation along a path length. Flame emissivity in terms of optical thickness is given by Babrauskas [5] as
where, a and b are empirical constants; a ¼ 31.6 and b ¼ 0.58. 1.5. Emissivity distribution
εf ¼ 1 eκβD
Emissivity is one of the key parameters that determine the amount of radiation emitted by a pool fire. Emissivity of a pool fire is primarily due to the hot combustion products formed during the combustion process. It is thus a combination of emissivity of soot and hot gases. Limitations of earlier methodologies in the determination of emissivity [8,27,28] include assumption of axisymmetric flame and the data being spatially and temporally averaged. Raj and Prabhu [29] proposed a refined methodology to determine the spatial and temporal variation in the emissivity of diffusion flames in the optically thin regime. The instantaneous variation of the flame emissivity is obtained by determining the transmissivity of the electrically heated strips placed behind the flame using an infrared camera. The thermal camera collects infrared radiation (It) equal to the sum of the infrared radiation proceeding from the flame (If) and the infrared radiation proceeding from the body (Ib) passing through the flame. This is represented by Eq. (5). It ¼ If þ τf Ib
(6)
1.8. Irradiance measurement (a) Emissive Power Distribution By knowing the flame emissivity, the temperature distribution from thermal camera can be converted to irradiance distribution as: Ei;j ¼ σεf Ti;j4
(7)
where Ti,j is the temperature (K) of a pixel element, i and j indicate the location in the thermal image, Ei,j is the emissive power (kW/m2) at that location. Munoz et al. [23] have determined the emissive power of gasoline and diesel pool fires of diameter greater than 1.5 m. The emissivity values were set equal to 1 as in the case of optically thick pool fires (D > 1.5 m). Irradiance at a distance allows for the determination of fire safety distance. Fire safety distance is the distance at which the thermal radiation does not cause damage to humans, instruments and buildings. Sudheer et al. [35] have measured the heat flux at a distance by means of a Schmidt–Boelter gauge and also using an infrared camera. From the thermal images, the heat flux is determined as shown in Eq. (8). The view factors have been calculated for a rectangle to rectangle configuration in
(5)
It is of utmost importance to determine the complete distribution of emissivity and temperature as it allows for calculation of the true emissive power and fire safety distances. The present paper aims at providing the complete distribution of emissivity for heptane pool fires.
15
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
2. Description of experimental set-up
parallel and perpendicular planes as provided in Ref. [36]. q0012 ¼
F12 A1 E1 A2
A series of outdoor pool fire experiments are conducted with heptane as the fuel. All the experiments are conducted in an open space to allow natural air flow into the fire. The wind velocity is monitored using a vane anemometer. The velocity is predominantly between 0.5 m/s to 1 m/s. The fuel temperature and the ambient air temperature are at 33 C and the relative humidity is about 50%. The pan is made of mild steel of varying diameters from 0.1 m to 1.0 m. a) Mass burning rate: A 15 kg platform type weighing scale (Contech CT 15K1) is used for measuring the mass burning rate of pool fires of diameter 0.1 m–0.3 m. For pool fires of diameter 0.5 m, 0.7 m and 1.0 m, a 250 kg platform scale of Eldigi make is used. For determining the instantaneous mass burning rate, the lip height was initially zero and kept increasing as the fuel is consumed. b) Flame Height and Puffing frequency: The experiments are filmed using Canon EOS Digital Camera at 60 fps. Images are also obtained from thermal camera (Thermoteknix make VisIR® 640). The cameras are placed at different distances, depending upon the size of the pool fires, such that the complete flame is captured. c) Emissivity distribution: The schematic diagram for measurement of emissivity is shown in Fig. 2. A detailed technical description of the experimental setup used to determine the spatial distribution of emissivity and temperature is reported by Raj and Prabhu [29]. The experimental setup consists of 40 horizontal nichrome strips held firmly between two wooden planks and electrically insulated by means of fiber glass tape. The 40 horizontal strips are placed in such a way as to cover the entire flame length. The 40 strips are divided into 4 groups of 10 strips each. Each nichrome strip is 1.2 m in length, 0.01 m in width and 0.001 m in thickness. The ten strips are connected in series and are then connected to an AC power supply through a variac. The local emissivity distribution in heptane pool fires has been determined for pool diameters of 0.3 m, 0.5 m, 0.7 m and 1.0 m. The strips are painted with Pyromark 2500 flat black, a paint whose emissivity is a constant in the working range of the thermal camera [38,39]. The strips are placed behind the pool fire and are electrically heated to 500K. The pool fire is ignited and allowed to reach a steady mass burning rate. Fig. 3 shows the infrared images of the strips as seen through the flame. Table 1 provides the details of the thermal camera used in the present study. d) Temperature and heat flux: Temperature is measured using a
(8)
where, F1-2: View factor from pixel A1 to sensor area A2; A1: Pixel Area; A2: Sensor Area; E1: Emissive power. (b) Radiative fraction Radiative fraction is an important parameter which represents the fraction of heat released in the form of radiation. Hamins et al. [37] and Buch et al. [12] have determined the radiative fraction for heptane pool fire of diameter 0.3 m. The total energy radiated is calculated by integrating the energy flux through the cylindrical surface [37], as in Eq (9). ! R ∞ 00 00 _ Qr ¼ 2π ∫ R0 r:q ðrÞ:dr þ R0 ∫ 0 qz ðzÞ:dz
(9)
While consolidated correlations exist for estimating the flame height, radiative fraction and fire safety distances, there exists a significant degree of error in predicting various other parameters that characterize pool fires. The mass burning rate for heptane pool fires reported in the literature vary by more than two times [13]. The effect of lip height has not been captured and the uncertainty in the flame height is of the order of pool diameter [31]. There is no data on the local emissivity and temperature distribution of heptane pool fires. In an effort to improve the accuracy of the data of heptane pool fires, a series of experiments are carried out and various parameters that characterize pool fires are determined. The objective of the present study is two-fold: 1. To characterize heptane pool fires in terms of geometric and radiative parameters. 2. To address the short comings in the literature in terms of mass burning rate, effect of lip height, spatial distribution of emissivity and temperature. The characterization of heptane pool fires for diameters ranging from 0.1 m to 1.0 m would serve as a benchmark data for numerical simulation. This would aid the fire modeling community seeking to characterize fuels at bench top scale for later correlation to hazards in large fires.
Fig. 2. Schematic of the experimental set-up. 16
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
strips and the intensity of the heated body (Ib) is obtained prior to the start of the experiments. Thus, substituting the intensity values into Eq. (5), the transmissivity of the flame is obtained. A MATLAB Code performs the pixel-wise calculation of the transmissivity. Furthermore, with the assumption of zero flame reflection, emissivity is determined as:
εf ¼ 1 τf
(10)
By substituting the obtained values of emissivity at each pixel location, the temperature distribution is obtained. (d) Temperature measurement using thermocouples: The temperatures obtained from the thermocouples have been corrected for radiation losses. Silvani and Morandini [33] have provided for temperature correction for a thermocouple inside a flame front, as shown in Eq. (11). As soot gets deposited on the thermocouple bead, the emissivity of the bead has been taken as 0.8 [33]. An average value of the flame emissivity (εf) obtained using the refined technique has been used for the correction of temperature. The heat transfer coefficient in Eq. (11b) is obtained as flow over a cylinder from Nusselt number correlation provided in Bergman et al. [40]. dTC is the bead diameter (0.7 mm). Thermal conductivity (kf), Reynolds number and Prandtl number are determined for dry air at the true flame temperature by an iteration process.
Fig. 3. Instantaneous thermal image of pool fire.
Tf ¼ TTC þ
Table 1 Specifications of uncooled thermal camera used in the present study.
h¼
®
Model
VisIR 640
Detector Spectral range Pixel Resolution FOV Temperature range Accuracy Frames per second
Microbolometer Uncooled FPA 7.5–13 μm 640 480 25 19 0 C–2000 C 2% of reading 60 Hz
σεTC 1 εf Tf4 h þ 4σεTC Tf3
(11a)
kf 0:43 þ 0:53Re0:5 Pr0:31 dTC
(11b)
(e) Uncertainty Analysis: The uncertainty of various parameters obtained in experimental analysis is as shown in Table 2. The uncertainty in bead diameter is obtained by measuring it using a screw gauge. The uncertainty in the measurement of temperature by means of thermocouple is after the correction for radiation losses. 4. Results and discussion A series of pool fire experiments are conducted with heptane as fuel for diameters ranging from 0.1 m to 1.0 m. Heptane pool fires have been characterized by determining the mass burning rate, flame height, puffing frequency, optical thickness, spatial distribution of emissivity and temperature, radiative fraction and fire safety distance. Spatial variation of flame emissivity is inferred from the transmissivity of the flame measured using infrared thermography.
thermocouple tree consisting of 14 thermocouples placed at a distance of 50 mm from each other. The thermocouples are connected to a data logger (Agilent make) and the emf is recorded every 0.7 s. SchmidtBoelter gauge is placed at Z/D ¼ 1 and traversed in the vertical direction (Y/D) to measure the irradiance at various vertical locations. 3. Data reduction (a) Mass burning rate: To obtain the mass burning rate, a platform type weighing scale is used. It is interfaced with a computer through an RS232 port. Data is transferred every 0.1 s for smaller pool diameter (0.1 m–0.3 m) and every 0.2 s for larger pool diameter (0.5 m, 0.7 m and 1.0 m). To determine the effect of lip height on mass burning rate, the pan was partially filled to different levels and the mass loss rate is recorded. (b) Flame Height and puffing frequency: All the experiments are filmed using Canon EOS Digital Camera at 60 fps. Each frame is then extracted from the recorded video using the software provided by the camera manufacturer. A MATLAB program determines the maximum height of the visible portion of the flame from the instantaneous images. It also performs the Fast Fourier Transform (FFT) of the intensity signal to determine the puffing frequency. (c) Emissivity and temperature distribution: The infrared camera obtains a two-dimensional image of the three dimensional intensity distribution of the flame. Considering a 2D image of the flame with the pixels in the x-y plane, the net intensity measured by the infrared camera is a line integral along the flame thickness. The flame emissivity is obtained from the calculation of the transmissivity of the flame with reference to electrically heated strips [29]. For Eq. (5), the total infrared radiation (It) is obtained at locations where the horizontal strips are placed. The intensity of the flame (If) is obtained at locations with no
4.1. Mass burning rate Mass burning rate is one of the key parameters used to characterize pool fires, as it determines the amount of heat released. Mass burning rate is determined by measuring the weight loss of the pan containing the fuel. The pan, filled with heptane, is placed on top of the platform scale which is interfaced with a computer to record the mass loss data. The overall as well as the instantaneous mass burning rate is determined for a period of 25 min. The instantaneous mass burning rate is obtained by differentiation of the weight loss data and smoothing it out. Fig. 4 shows Table 2 Experimental uncertainties.
17
Parameter
Relative uncertainty (%)
Mass burning rate Flame length Emissivity from mass burning rate Emissivity from heated strips Bead diameter of thermocouple Thermocouple temperature Thermal camera temperature Schmidt-Boelter gauge
0.2 0.2 0.2 10 1.4 8 2 7
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 5. Effect of lip height on mass burning rate of heptane pool fire (HPF). Fig. 4. Instantaneous mass burning rate for heptane pool fire of various pool diameters.
shows the sequence of still images, for a pool fire of diameter 0.7 m, where puffing can be observed. A point as shown in Fig. 6(g) is chosen and the intensity at that point is tracked. Puffing frequency is thus determined by tracking the “flaming vortical structures” from the visible images. A MATLAB program is developed to perform a Fast Fourier Transform (FFT) of the intensity signal. The FFT provides the dominant frequency which is the puffing frequency as shown in Fig. 7. Table 3 compares the experimentally obtained puffing frequency with the empirical equation proposed by Cetegen and Ahmed [19] and Malalasekara et al. [18]. The empirical equations are obtained from the best possible fits from a large number of data points. The difference in the fits is probably due to the errors involved in retrieving the values from published reports and graphs [18]. The puffing frequency is found to match closely with that reported by Cetegen and Ahmed.
the variation of mass burning rate with normalized time which is obtained by dividing the time with the total time of the experiment. Fluctuations observed in the instantaneous mass burning rate data are caused by changes in the wind velocity. Four different stages for mass burning rate can be observed in Fig. 4: The first stage is the initial transient stage where the fire rapidly spreads and develops through the entire pool. The fuel surface temperature rises rapidly and reaches the boiling point. Small bubbles on the fuel surface can be seen. In the second stage, the flame reaches a steady state wherein the mass burning rate becomes constant. The third stage is reached after approximately 60% of the total time where the mass burning rate increases to a second plateau. This is understood to be due to the bulk boiling behavior where the fuel temperature exceeds a particular value and bubbles due to boiling are seen on the fuel surface. As more amount of fuel is at a vaporizing temperature, the mass loss rate is enhanced [13]. The last phase is the decay period where the fuel is burnt out and the fire is extinguished.
4.4. Flame length The camera is placed normal to the fire, at an average height and at a sufficient distance to capture the entire flame height. The video recordings that were used to determine the puffing frequency is used to determine the flame height. A MATLAB program is developed to identify the maximum height of the visible part of the flame. The average length of flame is obtained for an intermittency value of 0.5, as defined by Zukoski et al. [22]. Fig. 8 shows the intermittency for 0.3 m heptane pool fire. Fig. 9 shows the average flame height obtained experimentally and its comparison with the correlation proposed by Heskestad [24] and Thomas [25]. The comparison has been done for both the initial as well as the bulk boiling steady states. The flame height is observed to increase by an order of pan diameter due to the increase in the mass loss rate from initial steady state to bulk boiling steady state.
Fig. 1 shows the comparison of the experimentally obtained mass burning rate with that reported in various literature [7,9–16]. m1 denotes the mass burning rate for the initial steady state while m2 denotes the mass burning rate during the bulk boiling behavior. The mass burning rate obtained during the bulk boiling period is found to be almost twice that of the initial steady state. The large difference in the mass burning rates reported by various researchers can thus be attributed to the measurement of the burning rates at either the initial steady state or during the bulk boiling period. 4.2. Effect of lip height
4.5. Emissivity and temperature distribution
The effect of having a non-zero freeboard height is found to be significant in pool fires. Experiments are conducted on 0.2 m and 0.3 m circular pool fires to determine the lip effect. From Fig. 5 it can be observed that for heptane pool fires the mass burning rate decreases with increase in the free board height. Dlugogorcki and Wilson [17] stated that natural convection greatly influences the burning rates in vessels made from conductive materials. They observed that the onset of fuel boiling occurred at progressively increasing freeboard height which dramatically altered the mass burning rates. The uncertainty in the measurement of mass burning rate reported throughout the literature can also be attributed to this effect.
Emissivity has been determined from the mass burning rate as well as the refined technique that determines the transmissivity of the electrically heated strips placed behind the flame. The complete temperature distribution is then determined from the local emissivity distribution using it as an input to the thermal camera. 4.5.1. Using mass burning rate Babrauskas [5], by applying conservation of energy at the base of pool fire, determined emissivity as the ratio of the mass burning rate of a pool fire of particular diameter to the mass burning rate of a pool fire of infinite diameter. For heptane, the mass burning rate for a pool fire of infinite diameter is 101 g/m2s [5]. Fig. 10 shows the emissivity obtained from mass burning rate for both the initial steady state and during the bulk boiling period. The emissivity for heptane pool fires is found to increase with the increase in pool diameter as the mass burning rate also
4.3. Puffing frequency Video recordings of the experiments, captured using Canon EOS digital camera, are exported as digital images at 0.05 s interval. Fig. 6 18
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 6. Sequence of images showing puffing in pool fires of diameter 0.7 m.
Fig. 7. Power spectrum for 0.7 m pool fire indicating puffing frequency.
increases with pool diameter.
Table 3 Puffing frequency with Pool diameter. Pool Diameter
Puffing Frequency (Hz)
(m)
Measured Values
Cetegen and Ahmed [19] 1:5 ffiffiffi f ¼p D
Malalasekara et al. [18] pffiffiffi f ¼ 1:68 D
0.1 0.2 0.3 0.5 0.7 1.0
4.8 3.36 2.76 2.16 1.78 1.62
4.74 3.35 2.74 2.12 1.79 1.5
5.31 3.76 3.07 2.38 2.01 1.68
4.5.2. Using refined technique with electrically heated strips placed behind the flame The refined technique developed by Raj and Prabhu [29] is used to determine the complete spatial distribution of emissivity and temperature. For each pool diameter, 100 instantaneous infrared images are averaged to obtain the average thermal image of the strips as seen through the pool fire. The total intensity (It) is measured with the heated strips seen through the flame. The flame intensity (If) is measured two pixels to the top or bottom of the strips and intensity of strips (Ib) is measured without the flame. Fig. 11(a) shows the contour plot of emissivity distribution for the
19
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 11. (a) Contour plot of emissivity distribution (b) Contour plot of temperature distribution of heptane pool fire (HPF) of diameter 0.3 m. Fig. 8. Intermittency for heptane pool fire (HPF) of 0.3 m pool diameter.
horizontal axis (X/D) and the vertical axis (Y/D) are non-dimensionalized by dividing the axis with the diameter of the pool fire. The average emissivity at the base of the pool fire is obtained to be 0.4. While traversing from the core region of the pool fire to the plume region, the emissivity is found to decrease from 0.3 to 0.1. From the emissivity values obtained through the above methodology, temperature contours are obtained by substituting the values into the thermal images. Fig. 11(b) shows the contour plot of temperature distribution. The maximum temperature is around 1200K while the average temperature is about 750K. The thermal images are a representation of the apparent temperatures as the intensity measured by the thermal camera is line integrated over the depth of view. Fig. 12(a) and (b) show the radial variation of emissivity and temperature at various non-dimensional heights (Y/D) for heptane pool fire of diameter 0.3 m. Emissivity decreases with height from a value of 0.375 to a value of 0.25. The temperature values are also found to decrease from 1000K at the base of the pool fire to 700K in the plume region. It is observed that the emissivity and temperature profiles shift from a parabolic profile to a top hat profile along the length of the pool fire. Fig. 13(a) and (b) show the contour plot for emissivity and temperature variation for the averaged images of heptane pool fire of diameter 0.5 m. The average emissivity at the base of the pool fire is about 0.45. The emissivity values vary from 0.5 at the core region to 0.2 in the plume region. The maximum temperature is around 1050K while the average temperature is about 900K. Fig. 14(a) and (b) show the radial variation of emissivity and temperature along the length of the pool fire. It can be observed that the core region has a constant emissivity and temperature values. Fig. 15(a) shows the contour plot for the emissivity distribution for heptane pool fire of diameter 0.7 m. The average emissivity at the base of the pool fire is around 0.5 to 0.6. The emissivity at the center region is around 0.5 and decreases to 0.3 in the plume region. Fig. 15(b) shows the contour plot for the temperature distribution for heptane pool fire of diameter 0.7 m. The maximum temperature is around 1250K while the average temperature is about 900K. Fig. 16(a) and (b) show the radial variation of emissivity and temperature along the length of the pool fire. The temperatures are high at the base of the pool fire. As the combustion products move up due to buoyancy, they get cooled by mixing with the ambient air. Hence, there is a decrease in temperature values as the particles move up from the base of the pool fire. Fig. 17(a) and (b) show the contour plot of emissivity and temperature distribution for heptane pool fire of diameter 1.0 m. The average emissivity at the base of the pool fire is 0.6. The emissivity at the core region is around 0.6 and decreases to 0.5 towards the plume region. The emissivity values are high in comparison to lower pool diameters due to the presence of higher amount of soot particles. The high temperature
Fig. 9. Comparison of average flame height obtained experimentally with correlations of Heskestad [24] and Thomas [25].
Fig. 10. Comparison between the emissivity obtained by mass burning rate with the average emissivity obtained by the refined technique.
averaged images of heptane pool fire of 0.3 m diameter. The origin of contour plots is chosen at the center of the base of the pool. The
20
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 14. (a) Radial emissivity variation (b) Radial temperature variation of heptane pool fire (HPF) of diameter 0.5 m. Fig. 12. (a) Radial emissivity variation (b) Radial temperature variation of heptane pool fire (HPF) of diameter 0.3 m.
Fig. 15. (a) Contour plot of emissivity distribution (b) Contour plot of temperature distribution of heptane pool fire (HPF) of diameter 0.7 m. Fig. 13. (a) Contour plot of emissivity distribution (b) Contour plot of temperature distribution of heptane pool fire (HPF) of diameter 0.5 m.
zones at Y/D of 0.8 shows two strong vortices resulting from air entrainment. The maximum temperature is around 1150K while the average temperature is about 1000K. Fig. 18(a) and (b) show the radial variation of emissivity and temperature along the length of the pool fire. The emissivity and temperature values are uniform in the core region.
obtained from the refined technique is spatially averaged. Fig. 10 shows the comparison of emissivity values obtained from the mass burning rate with the average emissivity values obtained from the refined technique obtained during the bulk boiling period. It can be observed that the overall emissivity values are in good agreement with the average values obtained from the refined technique. It is imperative to point out that the use of an average emissivity will lead to significant errors in the spatial distribution of temperature values.
4.5.3. Comparison of emissivity obtained from mass burning rate and the refined technique Throughout the literature, various researchers have always used a single value to represent the emissivity of pool fire for a given diameter. Thus, to represent the overall emissivity and to compare it with that obtained from mass burning rate, the complete distribution of emissivity
4.5.4. Comparison of centerline temperature distribution obtained from thermal camera and thermocouple K-type thermocouples, placed at a distance of 50 mm from each other, are used to measure the centerline temperature. The data is recorded at every 0.7 s. Once the temperatures have reached steady state, an average is taken for a period of 5 min. Fig. 19 shows the comparison of the 21
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 18. (a) Radial emissivity variation (b) Radial temperature variation of heptane pool fire (HPF) of diameter 1.0 m.
Fig. 16. (a) Radial emissivity variation (b) Radial temperature variation of heptane pool fire (HPF) of diameter 0.7 m.
Fig. 17. (a) Contour plot of emissivity distribution (b) Contour plot of temperature distribution of heptane pool fire (HPF) of diameter 1.0 m.
22
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
(a) 0.3 m diameter
(b) 0.5 m diameter
(c) 0.7 m diameter Fig. 19. Comparison of centerline temperature measured by thermocouple and thermal camera.
4.6. Optical thickness
centerline temperature measured by thermocouple with that obtained from the thermal camera. It can be observed that the temperatures measured by the thermocouple are approximately 200 C higher than that obtained by the thermal camera. This can be attributed to the fact that the thermal camera captures the equivalent intensity of a three dimensional flame in a two dimensional plane. This results in a decrease in the total intensity measured as the data obtained is line integrated over the depth of view. Fig. 20 shows the comparison of the average centerline temperature obtained from the thermocouples for heptanes pool fires under study in the present paper with the centerline temperature reported by Koseki [31] for heptane pool fires and that reported by McCaffrey [30] for natural gas. The temperatures obtained match well with both Koseki and McCaffrey's data. The difference is observed for higher pool diameters of 0.7 m and 1.0 m and this is attributed to fact that McCaffrey's model is valid only for low sooting fuels (see Fig. 20).
Optical thickness in terms of flame emissivity is provided in Eq. (6). Rearranging the equation and taking natural log provides, eκβD ¼ 1 εf κβD ¼ ln 1 εf
(14)
The slope of the plot between the pool diameter (D) on the abscissa and ln(1-εf) gives the product of the absorption-extinction coefficient of the flame and the mean beam corrector length, represented by κβ, as shown in Fig. 21. The obtained value of 1.04 m-1 is found to match well with the value reported by Babrauskas [5], which is 1.1 (0.3) m1. 4.7. Irradiance measurement 4.7.1. Heat flux at a distance Emissive power from the pool fire is determined from the obtained values of emissivity and temperature using Eq. (7). Fig. 22(a) and (b) show the comparison of the heat flux obtained at a distance of 71 cm and
Fig. 20. Comparison of centerline temperatures of heptane pool fires and natural gas.
Fig. 21. Optical thickness for heptane pool fire. 23
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
Fig. 22. Comparison of heat flux measurements at Z/D ¼ 1 with Schmidt–Boelter gauge and infrared camera for heptane pool fires.
133 cm with that reported by Hamins et al. [37] and Buch et al. [12] respectively. The heat flux almost doubles when the mass loss rate shifts from the initial steady state to the bulk boiling period. Fig. 22(c)–(e) show the comparison of heat flux measured using a Schmidt–Boelter gauge and that determined from the infrared camera using Eq. (8). The irradiance values obtained from the infrared camera, calculated at Z/D ¼ 1, matches well with that measured by the Schmidt–Boelter gauge. The heat flux measured at Z/D ¼ 1 is found to increase with increase in pool diameter as the total emissive power of pool fires also increases with pool diameter. The heat flux at various Z/D is calculated using Eq. (8). A heat flux of 1.4 kW/m2 is harmless for persons without any special protection and a heat flux of 4.7 kW/m2 causes pain in 15–20 s and burns after 30 s [41]. Fire safety distance is the maximum value of Z/D which matches with the above mentioned heat fluxes. Table 4 shows the fire safety distance for heptane pool fires of various pool diameters.
Table 4 Fire safety distances for heptane pool fires. Pool Diameter
Fire safety distance (Z/D)
(m)
1.4 kW/m2 (harmless for persons without any special protection)
4.7 kW/m2 (causes pain in 15–20 s and burns after 30 s)
0.3 0.5 0.7 1
3.67 4 4 4.5
1.97 2 2.25 2.5
heptane pool fires is calculated using Eq. (9). The radiative fraction for heptane pool fires is found to increase from 0.25 to 0.33 for the pool diameters considered in the present study as seen in Fig. 23. The values obtained are comparable to that reported in other literature [7,14,31, 37]. Hamins et al. [37] reported that the radiance measurement made at a single location and corrected for anisotropic effects yield an estimate of radiative fraction within 13% of the multi-location measurements.
4.7.2. Radiative fraction The radial and vertical distributions of the heat flux are calculated from the thermal image as given by Eq. (8). The total radiative power of
24
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26
References [1] S. Sudheer, S.V. Prabhu, Measurement of flame emissivity of hydrocarbon pool fires, Fire Technol. 48 (2) (2012) 183–217. [2] P. Joulain, The behavior of pool fires: state of the art and new insights, in: Symposium (International) on Combustion (Vol. 27, No. 2, pp. 2691–2706), Elsevier, 1998, January. [3] T. Steinhaus, S. Welch, R.O. Carvel, J.L. Torero, Large-scale pool fires, Therm. Sci. 11 (2) (2007) 101–118. [4] L. Hu, A review of physics and correlations of pool fire behaviour in wind and future challenges, in: IAFSS 12th Symposium 2017, 91, July 2017, pp. 41–55. https://doi. org/10.1016/j.firesaf.2017.05.008. [5] V. Babrauskas, Estimating large pool fire burning rates, Fire Technol. 19 (4) (1983) 251–261. [6] V.I. Blinov, G.N. Khudyakov, Diffusion Burning of Liquids (No. AERDL-T-1490-A), Army Engineer Research and Development Labs Fort Belvoir VA, 1961. [7] K. Hiroshi, Y. Taro, Air entrainment and thermal radiation from heptane pool fires, Fire Technol. 24 (1) (1988) 33–47. [8] S. Sudheer, S.V. Prabhu, Measurement of flame emissivity of gasoline pool fires, Nucl. Eng. Des. 240 (10) (2010) 3474–3480. [9] A. Hamins, S.J. Fischer, T. Kashiwagi, M.E. Klassen, J.P. Gore, Heat feedback to the fuel surface in pool fires, Combust. Sci. Technol. 97 (1–3) (1994) 37–62. [10] H. Hayasaka, Unsteady burning rates of small pool fires, Fire Saf. Sci. 5 (1997) 499–510. [11] J. Gore, M. Klassen, A. Hamins, T. Kashiwagi, Fuel property effects on burning rate and radiative transfer from liquid pool flames, Fire Saf. Sci. 3 (1991) 395–404. [12] R. Buch, A. Hamins, K. Konishi, D. Mattingly, T. Kashiwagi, Radiative emission fraction of pool fires burning silicone fluids, Combust. Flame 108 (1) (1997) 118–126. [13] B. Chen, S. Lu, C. Li, Q. Kang, M. Yuan, Unsteady burning of thin-layer pool fires, J. Fire Sci. 30 (1) (2012) 3–15. [14] H. Koseki, H. Hayasaka, Estimation of thermal balance in heptane pool fire, J. Fire Sci. 7 (4) (1989) 237–250. [15] H. Koseki, G.W. Mulholland, The effect of diameter on the burning of crude oil pool fires, Fire Technol. 27 (1) (1991) 54–65. [16] A. Shinotake, S. Koda, K. Akita, An experimental study of radiative properties of pool fires of an intermediate scale, Combust. Sci. Technol. 43 (1–2) (1985) 85–97. [17] B.Z. Dlugogorski, M.T. Wilson, Effect of lip height on properties of small scale pool fires, Fire Saf. Sci. 2 (1995) 129–140. [18] W.M.G. Malalasekera, H.K. Versteeg, K. Gilchrist, A review of research and an experimental study on the pulsation of buoyant diffusion flames and pool fires, Fire Mater. 20 (6) (1996) 261–271. [19] B.M. Cetegen, T.A. Ahmed, Experiments on the periodic instability of buoyant plumes and pool fires, Combust. Flame 93 (1–2) (1993) 157–184. [20] F. Tang, L. Hu, Q. Wang, Z. Ding, Flame pulsation frequency of conductioncontrolled rectangular hydrocarbon pool fires of different aspect ratios in a subatmospheric pressure, Int. J. Heat Mass Tran. 76 (2014) 447–451. [21] L. Hu, J. Hu, J.L. de Ris, Flame necking-in and instability characterization in small and medium pool fires with different lip heights, Combust. Flame 162 (4) (2015) 1095–1103. [22] E.E. Zukoski, B.M. Cetegen, T. Kubota, Visible structure of buoyant diffusion flames, in: Symposium (International) on Combustion (Vol. 20, No. 1, pp. 361–366), Elsevier, 1985, January. [23] M. Mu~ noz, J. Arnaldos, J. Casal, E. Planas, Analysis of the geometric and radiative characteristics of hydrocarbon pool fires, Combust. Flame 139 (3) (2004) 263–277. [24] G. Heskestad, Luminous heights of turbulent diffusion flames, Fire Saf. J. 5 (2) (1983) 103–108. [25] P.H. Thomas, The size of flames from natural fires, in: Symposium (International) on Combustion (Vol. 9, No. 1, pp. 844–859), Elsevier, 1963, January. [26] G. Mangialavori, F. Rubino, Experimental tests on large-scale hydrocarbon pool fires, in: 7th International Symposium on Loss Prevention and Safety Promotion in Process Ind., Taormina, Italy, 1992, 83–1. [27] E. Planas-Cuchi, J.M. Chatris, C. L opez, J. Arnaldos, Determination of flame emissivity in hydrocarbon pool fires using infrared thermography, Fire Technol. 39 (3) (2003) 261–273. [28] C. Qian, K. Saito, Measurements of pool-fire temperature using IR technique, in: Proceedings Combustion Fundamentals and Applications, Combustion Institute/ Central and Western States, Mexican National Section, and American Flame Research Committee, San Antonio, Texas, 1995, pp. 81–86. [29] V.C. Raj, S.V. Prabhu, A refined methodology to determine the spatial and temporal variation in the emissivity of diffusion flames, Int. J. Therm. Sci. 115 (2017) 89–103. [30] B.J. McCaffrey, Purely Buoyant Diffusion Flames: Some Experimental Results, 1979. [31] H. Koseki, Combustion properties of large liquid pool fires, Fire Technol. 25 (3) (1989) 241–255. [32] H. Hayasaka, H. Koseki, Y. Tashiro, Radiation measurements in large-scale kerosene pool flames using high-speed thermography, Fire Technol. 28 (2) (1992) 110–122. [33] X. Silvani, F. Morandini, Fire spread experiments in the field: temperature and heat fluxes measurements, Fire Saf. J. 44 (2) (2009) 279–285. [34] S. Sudheer, S.V. Prabhu, Characterization of hexane pool fires using infrared thermography, J. Fire Sci. 31 (2) (2013) 143–165. [35] S. Sudheer, L. Kumar, B.S. Manjunath, A. Pasi, G. Meenakshi, S.V. Prabhu, Fire safety distances for open pool fires, Infrared Phys. Technol. 61 (2013) 265–273.
Fig. 23. Radiative fraction of heptane pool fire.
5. Conclusions Characterization of heptane pool fires for diameter ranging from 0.1 m to 1.0 m has been carried out by measuring key parameters like mass burning rate, effect of lip height, flame length, puffing frequency, optical thickness, radiative fraction and spatial distribution of emissivity and temperature. Mass burning rates for heptane pool fire increases with the increase in pool diameter. The instantaneous mass burning rate and total mass burning rate have been determined. From the instantaneous mass burning rate, it is observed that heptane pool fires depict two plateaus for burning rate. The first plateau is for the initial steady state while the second plateau is for the bulk boiling behavior. The effect of having a non-zero free board height is observed to significantly alter the mass burning rate for pool diameters of 0.2 m and 0.3 m. It is observed that the mass burning rate decreases with increase in free board height. Puffing frequency is determined from the visible images by tracking the intensity about a point where the vortical structures are seen. The results obtained are found to match well with Cetegen and Ahmed [19]. Maximum flame height is determined from the visible images and the average flame height is determined from the definition of intermittency of the flame. Flame emissivity has been determined as the ratio of the mass burning rate of the particular pool diameter to the mass burning rate of pool fire whose diameter is large. Spatial distribution of emissivity and temperature has also been determined by a refined technique. The emissivity and temperature are found to increase with increase in pool fire diameter and decrease along the height. The κβ value for heptane pool fires is obtained to be 1.04 m-1. The irradiance almost doubles for the second plateau of mass burning rate. The irradiance measured at a distance of Z/D ¼ 1 with infrared camera and Schmidt–Boelter gauge match reasonably well. The fire safety distances are approximately 4D. Radiative fraction for heptane pool fires varies from 0.25 to 0.33 with increase in pool diameter from 0.3 m to 1.0 m. Acknowledgements Authors thank Mr. Rahul Shirsat for his help in building the experimental setup.
25
V.C. Raj, S.V. Prabhu
Fire Safety Journal 96 (2018) 13–26 [39] J.T. Nakos, J.M. Suo-Anttila, W. Gill, Shroud Boundary Condition Characterization Experiments at the Radiant Heat Facility, SAND2004-5080, October, 2004. [40] T.L. Bergman, F.P. Incropera, D.P. DeWitt, A.S. Lavine, Fundamentals of Heat and Mass Transfer, John Wiley & Sons, 2011. [41] L. Zarate, J. Arnaldos, J. Casal, Establishing safety distances for wildland fires, Fire Saf. J. 43 (8) (2008) 565–575.
[36] J.R. Howell, A Catalog of Radiation Heat Transfer, 2016. http://www. thermalradiation.net/indexCat.html. [37] A. Hamins, M. Klassen, J. Gore, T. Kashiwagi, Estimate of flame radiance via a single location measurement in liquid pool fires, Combust. Flame 86 (3) (1991) 223–228. [38] R.S. Longenbaugh, L.C. Sanchez, A.R. Mahoney, Thermal response of a Small Scale Cask-like Test Article to Three Different High Temperature Environments, Federal Railroad Administration, 1990.
26