- Email: [email protected]
Experimental and analytical study of a porous media reformer with passive air entrainment
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Experimental and analytical study of a porous media reformer with passive air entrainment S.R. Addamane, M. Hajilou, E.L. Belmont* University of Wyoming, Mechanical Engineering Dept, 1000 E University Ave, Laramie WY 82071, USA
article info
abstract
Article history:
A porous media fuel reformer featuring passive air entrainment and a surface-stabilized
Received 8 February 2016
flame is demonstrated and utilized for the reformation of methane. Passive air entrain-
Received in revised form
ment negates the need for auxiliary resources beyond a source of compressed fuel gas for
25 April 2016
operation, making a reformer of this design promising for applications such as replace-
Accepted 5 May 2016
ment of hydrocarbon flaring with syngas generation at isolated oil and gas production sites
Available online 4 June 2016
and generation of syngas in other remote locations. The porous media reformer design presented in this study incorporates an eductor to entrain ambient air and a swirl mixing
Keywords:
chamber and porous media bed to premix fuel and oxidizer before combustion. A range of
Porous media
methane flow rates was tested in this study, and the effect on air entrainment, and
Combustion
therefore equivalence ratio, was examined. A wide range of stable operating conditions
Passive entrainment
shows a large turndown ratio of the reformer and burning rate ratios greater than unity.
Reforming
Fuel conversion efficiency and syngas production were evaluated and concentrations of
Syngas
methane, hydrogen, carbon monoxide and carbon dioxide are presented for a range of
Flare
methane flow rates. Product compositions are compared to equilibrium and a high extent of fuel conversion efficiency is shown. An analytical model that accounts for reformer geometry and operating conditions to predict air entrainment and equivalence ratio is presented and compared with the experimental results. The analytical model compares favorably with the experimental results and can guide future reformer development and scaling. © 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Introduction Extensive research has examined catalytic and non-catalytic routes for reforming fuels to hydrogen-rich synthesis gas, or syngas. Thermal partial oxidation for syngas production has been shown to be a robust technique that is tolerant to wide variations in fuel composition [1e6], and heat recirculation is frequently used to extend flammability limits and accelerate conversion rates [7e12]. While highly effective in reforming
fuels to hydrogen-rich syngas, filtration combustors and other heat-recirculating reactors often need to be actively controlled to overcome operational challenges such as propagating flame fronts, flashback, and soot formation [5,13e15]. There are scenarios in which fuel reformation could be implemented but the costs and complexities associated with catalytic or excess enthalpy techniques are currently prohibitive. Oil and gas production, for example, produce significant amounts of fuel waste that could be converted to syngas, and the likelihood of industries doing so is significantly improved if a
* Corresponding author. E-mail address: [email protected] (E.L. Belmont). http://dx.doi.org/10.1016/j.ijhydene.2016.05.035 0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
Nomenclature A BRR dp keff L n_ P Pe Q SL T U V
cross-sectional area, [m2] burning rate ratio bead diameter, [m] effective loss coefficient length of the bed, [m] molar flow rate pressure, [Pa] peclet number flow rate at room temperature, [m3/s] flame speed, [m/s] temperature, [K] superficial velocity at the cross section of the bed, [m/s] velocity, [m/s]
Greek symbols f equivalence ratio a thermal diffusivity, [m2/s] m dynamic viscosity, [Ns/m2] ∞ atmospheric condition ε porosity h conversion efficiency of methane, [%] r density, [kg/m3] Subscript a FS g h i l m m,m p t
air at ambient temperature flame speed methane at ambient temperature exit of the burner fuel injector orifice in eductor inlet of the burner air-fuel mixture mixture at average porous media bed temperature porous media bed eductor throat
robust and low cost method is available. Significant quantities of natural gas and larger hydrocarbons are currently flared at oil production sites that do not have sufficient auxiliary resources, such as electricity, to provide active control of flare gas combustion. Associated gas is sometimes used at these sites to provide pneumatic motive force for early oil production and equipment, after which the gas is flared. Additionally, natural gas wells that are drilled but not used for production will have a period of flaring prior to being temporarily shut [16]. Since hydrocarbons have at least ten times higher global warming potential by weight than carbon dioxide [17], and direct release of large amounts of combustible gas can produce an explosion hazard near the production site, unused gas is flared instead of released directly into the atmosphere. Flaring of hydrocarbon-rich gases has increased significantly in the past decade due to advanced oil recovery efforts in formations that produce natural gas in addition to oil [18], and flaring has contributed significantly to an increase
12739
in greenhouse gas emissions. Incomplete combustion of flare gas can produce significant amounts of pollutants that damage the environment and human health, leading to cancer and neurological, reproductive and developmental effects [19]. One of these pollutants is soot, which is hazardous to inhale due to small particle sizes that can travel readily into lungs and is known to contribute to climate change through high absorption of solar radiation, called radiative forcing [20,21]. This study focuses on development and use of a porous media reformer to convert hydrocarbon fuels, which are currently flared or otherwise wasted, to clean burning and utilizable synthesis gas. Such synthesis gas could be upgraded, utilizing only the hydrogen content or the hydrogen and carbon monoxide, or burned directly for power production. The reactor presented in this study incorporates design features to promote air entrainment and premixing in order to produce combustible reactants for maximal combustion efficiency. One of these design features is an eductor, which utilizes the momentum of a pressurized fuel gas jet to entrain ambient air into the fuel gas stream. Another feature is a swirl chamber, which mixes the fuel and gas by inducing swirl in the flow as it enters the burner body [22]. A third significant feature is the porous media bed; porous media has been widely used in the construction of experimental combustion devices to promote premixing and preheating of reactants prior to combustion [23e25]. Heat transfer through porous media has been extensively studied, and found to improve with decreasing media porosity [26]. In addition to heat recirculation, porous media also promotes premixing of gases through its pore networks [27]. The burner design presented in this study utilizes porous media to premix fuel and air and recirculate heat from the flame in order to preheat incoming reactants and promote flame stability over a wide range of fuel flow rates and equivalence ratios. Porous media has also been observed to slightly dampen burner acoustics during operation, which could assist in mitigating noise from industrial flares [28]. The present study experimentally and analytically examines operation of the porous media reformer with a surfacestabilized flame because the surface flame permits safe operation without risk of flashback over a wide range of operating conditions, and the surface flame reduces the risk of carbon build-up within the porous media bed from pyrolysis of large hydrocarbons at elevated temperatures [5]. This mode of operation is achieved by the use of fine diameter porous media and a range of high fuel flow rates over which the reformer is operated. A surface-stabilized flame distinguishes the current approach from many other burner methods and designs which utilize a submerged flame, and is intended to permit operation over long periods of time without the need for active control or additional methods of flame stabilization [29]. Equivalence ratios and syngas production are quantified by experimental measurements. An analytical model of the reformer is presented that predicts the rate of air entrainment into the fuel stream, and therefore reactant equivalence ratio, as a function of geometric and operating parameters. The model also provides guidelines for geometric scaling, permitting the future construction of a reformer that is capable of accommodating a wide range of reactant flow rates for a given application.
12740
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
Materials and methods Reactor apparatus Fig. 1 is a schematic of the porous media reformer used in this study. The apparatus includes an eductor, into which fuel is metered using a mass flow controller and air is passively entrained. Fig. 2 is a schematic of the eductor showing orifice and throat diameters, as well as the area of the air intake at the eductor. The fuel and air reactant mixture is fed from the eductor to a swirl chamber, where the flow is introduced tangentially to induce swirl and promote mixing. The swirl chamber has an internal diameter of 0.2 m and a height of 0.08 m. The geometric swirl number, Sg, of the chamber is estimated to be on the order of 1, which has been found to promote mixing [30]. The reactant mixture enters the bottom of the porous media bed upon exiting the swirl chamber. The porous media bed is retained by a steel cylinder, 8 cm in height, which is positioned atop the swirl chamber. An alumina sleeve inside the steel cylinder acts as an insulator, and is the same height as the steel cylinder. The alumina sleeve is filled with 1.6 mm diameter alumina beads. Five Ktype thermocouples are inserted into the porous media to the center of the reactor, and temperature variation along the height of the bed is continuously monitored during experiments. An Agilent CP 490 gas chromatograph (GC) is used to measure reactant and product species concentrations at the reactor outlet. Two columns are used to measure species in this study: a molecular sieve measures diatomic hydrogen (H2), oxygen (O2), nitrogen (N2), carbon monoxide (CO) and methane (CH4), and a porous polymer unit measures carbon dioxide (CO2). GC samples are drawn from the sampling chimney using a quartz probe with a 2 mm inner diameter and a 0.1 mm orifice diameter. Samples are transported through inert Silco stainless steel tubing that connects the probe to the GC, and pass through a membrane filter prior to entering the GC.
Fig. 1 e Schematic of the porous media reformer is shown, including components that promote passive air entrainment and fuel-air premixing.
Fig. 2 e Schematic of eductor with dimensions of fuel orifice and air intake shown.
Experimental method Methane fuel is utilized in this study because, while one potential envisioned application is replacement of flaring, typical flares vent CH4, larger alkanes, olefins and traces of hydrogen sulfide, and the volume fraction of CH4 can exceed 75% [16]; hence, CP grade CH4 (99% purity) was used as a single component representative fuel in this study. Fuel was introduced to the eductor over a range of flow rates from 2 to 22 SLPM. Ambient air, not dried or otherwise altered, was entrained for combustion. The equivalence ratio of unignited CH4 and air exiting the reactor was measured by GC at the top of the porous media. Equivalence ratio was determined from measured concentrations of O2 and CH4 at the porous media burner surface. The reactant mixture was ignited at the reactor outlet using a secondary flame from a butane lighter. Ignition was rapid and performed at a CH4 flow rate of 12 SLPM. Upon ignition, bed temperatures were continuously monitored and observed to rise until steady state was reached. Steady state was defined as less than ±0.5 K variation over 15 min of continuous burner operation. Once steady state was reached, exhaust gas combustion products were measured by GC and extractive probe sampling. In order to prevent secondary air entrainment at the reactor outlet, and permit the study of combustion resulting from primary air entrainment at the eductor, a quartz chimney of 0.6 m height and 0.2 m internal diameter was placed on top of the steel porous media bed housing during sampling. The chimney had no noticeable effect on burner performance based upon reactor bed temperatures and observed stability. Following the acquisition of burned gas samples, the reactor was extinguished by briefly interrupting CH4 flow to the burner. Once CH4 flow was resumed, unburned gas samples were taken again in order to determine the high temperature and steady state equivalence ratio of the operating point. The surface temperature of the porous bed was measured using a K-type thermocouple after steady state was achieved and the flame was extinguished. Measurement of CH4, O2, H2, CO, and CO2 was performed for a range of CH4 flow rates from 12 to 22 SLPM. This flow rate range subset was chosen for product speciation because of the nearly steady equivalence ratio observed over this range. Species concentrations are reported on a dry basis. Experimentally measured concentrations of CH4, H2, CO, and CO2 in combustion products were compared with chemical
12741
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
equilibrium, which was calculated for the measured equivalence ratios of the reactant mixtures over CH4 flow rates of 12e22 SLPM, to determine the extent of fuel conversion and the progression of combustion towards equilibrium. Each operating point was tested twice. Error analysis was conducted to evaluate the uncertainty of reactant flow rates, equivalence ratios, and species concentrations as measured by GC. Flow rate uncertainty is due to uncertainty of the methane mass flow controller and is ±0.5 SLPM. Uncertainty in equivalence ratio is due to repeatability of measurements and uncertainty in GC calibration gas compositions, and is calculated according to a student-t distribution to have an average value of ±0.04 across the tested flow rates. Uncertainties in measured gas species are attributed to repeatability, which is assessed according to a student-t distribution, and uncertainty in GC calibration gases. The average uncertainties of CH4, O2, H2, CO and CO2 concentrations are calculated to be ±0.21, 0.20, 1.40, 1.14 and 0.75%, respectively.
The rate of air entrainment is a critical performance parameter because it determines the equivalence ratio of the reactant mixture. One-dimensional analysis is utilized in this model, in which parameters are assumed to vary in the bulk flow path through the reactor. Assuming Ai ≪ At and velocity of air significantly less than that of CH4, a momentum balance between the eductor inlet and throat is given by Ref. [31]. rg Qg2 Ai
rm Qm2 At
(1)
Assuming heat transfer to the bed is approximately equal to change in enthalpy of the reactants, and accounting for change in kinetic energy of the reactants by heating as well as the pressure loss in the porous bed and minor losses in the remainder of the reactor, an energy balance between the eductor throat and the burner outlet is given by 2 rm;h Qm;h r Q2 P∞ Pt ¼ m 2m 1 keff DPp : 2 2Ap 2At
(2)
Combining Eqs. (1) and (2), using the ideal gas approximation to relate volumetric flow rate to temperature, and rearranging, Eq. (3) is obtained 2 rg Qg2 rm;h Qm;h r Q2 þ DPp : 0 ¼ m 2m 1 þ keff 2 2Ap At Ai 2At
(3)
Using relations obtained by conservation of mass Qm ¼ Qa þ Qg and rm ¼
ra Qa þ rg Qg Qa þ Qg
(4)
Eq. (3) can be rearranged to: ra Qa2
þ
rg Qg2
1 þ keff 2 r Q 2 þ r Q 2 þ r þ r Q Q a g rg Qg2 a a g g a g Th þ DPp : Tl 2A2p At Ai
0¼
! 2 ð1 εÞ2 Qm Tm 150mm L Tm ¼ DPp d2p Tl Tl ε3 Ap 2 2 1:75Lrm Tm ð1 εÞ Qm þ ε3 dp Tl A2p
! 2 2 150mm L Tm ð1 εÞ2 1:75Lrm Tm ð1 εÞ and N ¼ : M¼ 2 2 dp Ap ε3 Tl ε3 Tl dp Abed (8) Combining Eqs. (7) and (5) and dividing by
gives
)
(
(9) Using the definition of equivalence ratio n_fuel n_air actual f¼ n_fuel n_air stoichiometric
(10)
which can be rewritten for methane-air combustion in terms of the volumetric flow rates of fuel and air as Qg f ¼ 9:52 Qa
(11)
the ratio of volumetric flow rates in Eq. (9) can be rewritten and Eq. (9) can be simplified as ) !# 2 1 1 Th 1 M N 1 þ k eff At Ai 2A2t Qg Tl 2A2p ( ) " 2 i 1 h Th 1 þ N 1 þ k þ 9:52*f ra þ rg eff Tl 2A2p 2A2t !# ( (12) M 1 1 þ k þ 9:522 * ra þ eff Qg 2A2t )! 2 Th 1 N ¼ 0: Tl 2A2p 2 Taking a ¼ rg ; b ¼ ra , c ¼ At1A ; d ¼ 2A1 2 ð1 þ keff Þ; e ¼ TThl 2A1 2 , i p t Eq. (12) can be represented as "
(
rg
f2
aðc d e NÞ
2A2t
þ
The pressure drop through the porous media bed, DPp, can be found by applying the Ergun equation [32].
Qg2
! 2 1 1 Th 1 M N 1 þ keff 0 ¼ 2 rg At Ai 2A2t Qg Qg Tl 2A2p )! ( 2 2 Q 1 Th 1 N 1 þ keff þ a2 ra Qg Tl 2A2p 2A2t ! ( ) 2 i 1 Qa Qg h Th 1 M r : þ r þ N þ 1 þ k þ eff a g Qg Qg2 Tl 2A2p 2A2t
f2
(5)
(7)
where the coefficients of Eq. (7) can be represented by
"
þ ra þ rg Qa Qg
(6)
Rearrangement of Eq. (6) to reference all flow rates at ambient temperature gives
Qg2
Analytical model
ðPt P∞ ÞAt ¼
! 150mL ð1 εÞ2 1:75Lr ð1 εÞ 2 DPp ¼ U : U þ d2p dp ε3 ε3
M Qg
!#
M Qg
!#
" 9:52*f
ða þ bÞðd þ e þ NÞ (13)
9:522 *bðd þ e þ NÞ ¼ 0:
Using the quadratic equation shown in Eq. (13), f can be
12742
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
predicted for a given burner geometry, operating condition and bed temperature.
Results and discussion Equivalence ratio The reactant equivalence ratio achieved within the porous burner is determined by the flow rate of CH4 that is introduced into the system by mass flow controller and the amount of air that is passively entrained by the eductor. The amount of air entrainment will be determined by the momentum of CH4 introduced to the system and the pressure drop through the burner; therefore, the equivalence ratio is indirectly controlled through CH4 volumetric flow rate and burner geometry. Fig. 3 shows experimental measurements and analytical model predictions of equivalence ratio as a function of CH4 flow rate. Experimental measurements at ambient temperature and elevated steady-state temperature are included to assess the effect of bed heating and resultant additional pressure drop on air entrainment. Experimental results show a decrease in equivalence ratio from 2.32 to 1.49 with increase in CH4 flow rate from 2 to 22 SLPM. Flammable gas mixtures are produced at CH4 flow rates of 6 SLPM and greater, where sufficient air is entrained to produce reactant mixtures with equivalence ratios below the upper flammability limit of CH4, f ¼ 1.67. An increase in CH4 flow rate above 6 SLPM leads to increased air entrainment and a lower equivalence ratio up to a CH4 flow rate of 12 SLPM. Further increase in methane flow rate above 12 SLPM produces a slight increase in equivalence ratio from f ¼ 1.49 to f ¼ 1.56. Average uncertainty in equivalence ratio is calculated to be ±0.04, therefore f is constant within uncertainty at the highest tested CH4 flow rates. The calibrated analytical model (keff ¼ 0.57) shows quantitative agreement with the experimental results within experimental uncertainty. Slight deviation is seen at the highest tested CH4 flow rate but is within equivalence ratio uncertainty. The otherwise excellent agreement between experimental and analytical results gives insight into the physical effects that underlie the observed equivalence ratio results. The eductor functions by creating a low pressure
Fig. 3 e Experimentally measured equivalence ratios (f) at ambient conditions (Ta) and steady state temperature (Tss) over a range of CH4 flow rates are compared to analytical model results calculated at Ta of 300 K.
region near the air inlet due to the high velocity of the CH4 jet exiting the eductor orifice, thereby entraining air into the CH4 fuel stream. An increase in CH4 flow rate increases the pressure difference between the air opening in the eductor and the atmosphere, leading to higher entrainment of air. Simultaneously, the pressure drop through the porous media bed also increases with an increase in CH4 and air flow rates. This increased pressure drop produces a competing effect that lowers air entrainment. The results of experiments and modeling indicate that air entrainment enhancement by low pressure at the air intake is dominant at low flow rates, leading to a decrease in f with increase in CH4 flow rate. At higher flow rates, the effects of low pressure and increased pressure drop through the bed balance each other and a nearly constant f is produced with change in CH4 flow rate. The rate of experimental air flow entrainment is calculated for a given fuel flow rate using the measured oxygen and CH4 concentrations at the reactor outlet prior to ignition. Entrained air flow rates increase from 8 SLPM to 140 SLPM with increase in CH4 flow rates from 2 SLPM to 22 SLPM; thus, equivalence ratio and total flow rate results highlight the large turndown ratio of the reformer and operating range over which a rich reactant mixture can be produced for combustion and syngas generation. The rate of air flow entrainment into the burner is dependent on a number of factors, including burner geometry and methane flow rate. Control of air flow rate independent of methane flow rate would therefore require variation of burner geometry, such as eductor orifice and throat diameters, and porous media bed height and diameters. The reactor geometry tested in this study was chosen to be similar to that used in Yoksenul and Jugjai [23] in order to validate the concept of passive air entrainment for a porous media burner with surface-stabilized flame, and the analytical model derived in this study gives an indication of how equivalence ratio will vary with changes to reformer geometry. Fig. 4 shows analytical model results for the variation of equivalence ratio calculated for the geometry of the reformer used in this study, and for variation in bed heights and particle diameters. The equivalence ratio is predicted by the model to increase with increase in bed height and with decrease in particle diameter due to increase in pressure drop across the bed, and the resultant reactant mixture remains in the rich and
Fig. 4 e Variation of f with change in geometry and porosity at 20 SLPM CH4 flow rate and ambient temperature of 300 K, calculated using the analytical model, is shown. The geometry of the burner used in this study is marked (*).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
flammable regime over the wide range of particle diameters and bed heights of 0.8e2.0 mm and 0.04e0.16 m, respectively, modeled and shown in Fig. 4.
Bed temperature The extent to which the porous media bed was heated by the surface flame at steady state burner operation was evaluated through monitoring of temperatures during experiments, and the impact of this heating on air entrainment and equivalence ratio was assessed by comparison of experimental measurements and analytical model results. Fig. 5 shows experimentally measured temperatures within the porous media bed. Bed temperatures are slightly elevated at less than 20 K above ambient temperature (300 K) within the majority of the bed. The surface thermocouple shows the most significant temperature elevation, with surface temperature varying from 408 K at 12 SLPM to 342 K at 22 SLPM. These temperatures, which are measured using thermocouples embedded within the porous bed of the reactor, support the observed position of the flame, which is stabilized at the outlet of the burner and not submerged within the porous bed. Experimental temperature results show that heat recirculation is confined to porous media bed depths within 25 mm of the surface of the burner, and temperatures at the bed surface are observed to decrease with increasing CH4 flow rate. A bed Peclet number which represents the ratio of convective to conductive heat transfer is given by Pep ¼
Vm dp : ap
(14)
Pep increases monotonically from 11 to 18 over the range of tested reactant flow rates, indicating an increased impact of convection over conduction within the bed; thus, the depth of temperature penetration within the bed is reduced with increasing reactant flow rate and Pep. As bed temperature is increased through heat recirculation from the flame and conductive and radiative heat transfer through the media, reactant gas temperature also increases as it is preheated by the porous media. Increase in gas temperature corresponds to a decrease in gas density, which leads to an increase in volumetric flow rate of gas through the porous media
Fig. 5 e Experimentally measured temperatures along the depth of the porous media bed at steady state for a range of methane flow rates.
12743
bed. The Ergun equation (Eq. (6)) shows a proportional effect of volumetric flow rate, or velocity U, on pressure drop through the bed, DPp. Fig. 6 shows the effect of elevated bed temperatures on equivalence ratio as calculated using the calibrated analytical model. Equivalence ratio is observed to increase with porous media bed temperature; this affect is attributed to the increase in pressure drop with increased bed temperature and volumetric flow rate, which leads to lower air entrainment and therefore higher equivalence ratios. The pressure drop across the bed, calculated for the operating conditions tested in this study, is between 5 and 40 Pa over the tested conditions, suggesting the utility of the reformer for industrial applications that require minimal backpressure [33]. The analytically observed effect of bed heating on air entrainment and equivalence ratio can be utilized to understand the experimentally measured ambient and steady state equivalence ratios, shown in Fig. 3, for CH4 flow rates from 12 to 22 SLPM. The experimentally measured equivalence ratios show no significant dependence on temperature when measured at ambient or steady state conditions. The average temperature measured throughout the porous bed over the range of tested methane flow rates was 313e321 K, with a substantial portion of the bed depth showing little temperature rise of less than 20 K above ambient. Model predictions, shown in Fig. 6, indicate that a change in average bed temperature from 300 K to 321 K induces a change in equivalence ratio of less than 0.03, which is below the calculated uncertainty of the equivalence ratio. Hence, the lack of significant change in f at steady state temperature observed in experiments is attributed to low average temperature rise throughout the porous media bed at experimentally measured conditions, but would be expected to be more substantial if a higher bed temperature rise occurred.
Emissions and conversion efficiency The equivalence ratio produced by passive entrainment was observed in Fig. 3 to decrease with increase in CH4 flow rate, and nearly stabilize at the highest tested flow rates. In order to achieve high extents of conversion of CH4 to syngas, a rich
Fig. 6 e Analytical model results show the variation of f for a range of mean bed temperatures at the CH4 flow rates tested in this study.
12744
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
equivalence ratio is desired. In the present study, a nearly constant equivalence ratio was achieved over a CH4 flow rate range of 12e22 SLPM and combustion product measurements were made at these operating conditions. This range of flow rates is highlighted in Fig. 3. Combustion product gases were analyzed in the reactor exhaust by GC after the burner reached steady state. Fig. 7 shows H2 and CO2 concentrations measured in the reactor exhaust, as well as equilibrium concentrations calculated for the equivalence ratios measured under steady state conditions. H2 increases and CO2 decreases slightly with increase in CH4 flow rate from 12 SLPM to 22 SLPM, following the trends exhibited by H2 and CO2 equilibrium concentrations for the slightly increasing equivalence ratios measured across this range of operating conditions. Fig. 8 shows CO and CH4 concentrations measured in the reactor exhaust, as well as equilibrium concentrations calculated for the equivalence ratios measured under steady state conditions. CO equilibrium concentrations increase slightly from 10.0% to 10.7% over the tested CH4 flow rate range of 12e22 SLPM due to the slight increase in equivalence ratio over this range. Experimentally measured CO concentrations also increase with CH4 flow rate, and are slightly lower than equilibrium at CH4 flow rates of 12 SLPM and 14 SLPM and equal to equilibrium within calculated uncertainty at CH4 flow rates of 16 SLPM and higher. The product concentrations of CO, as well as those of H2 and CO2 in Fig. 7, are further analyzed through examination of CH4 concentrations in combustion products. As seen in Fig. 8, product concentrations of CH4 are less than 0.4% at all tested flow rates, indicating a high extent of CH4 conversion. Equilibrium concentrations of CH4 are below the detectable limit, with mole fractions on the order of 1 ppt or less, thus the experimentally measured values of CH4 are above equilibrium within the calculated uncertainty of 0.2% at methane flow rates of 20 SLPM and 22 SLPM. These results suggest that low residence times at high reactant flow rates may limit complete conversion of CH4, and the discrepancies at low flow rates may likewise be due to low amounts of unreacted CH4. The efficiency of the passive entrainment reformer examined in this study can be further evaluated by the extent of CH4 conversion, or conversion efficiency, defined as
Fig. 8 e CO and CH4 concentrations in reactor exhaust over methane flow rates of 12e22 SLPM.
hCH4 ¼
½CH4 reac tan ts ½CH4 equilibrium
! $100:
(15)
Methane conversion efficiency calculated for the results shown in Fig. 8 is greater than 98% at all tested flow rates. Taken together with the measured product concentrations of H2, CO and CO2, the results suggest that the conversion of CH4 is nearly complete at all tested firing rates and a significant amount of syngas was generated at the tested operating conditions.
Suitability for industrial use Experimental results indicate the utility of the passive entrainment reformer for the production of syngas as a clean burning fuel [34]. The results of experiments and analytical modeling suggest a broad operating range that produces high conversion efficiency to syngas with insensitivity to variation in CH4 flow rates. Stable operation over a wide range of operating conditions is advantageous when utilizing a syngas reformer, whether in field applications with little or no utilities available for active control or in a highly controlled industrial setting, as instabilities can result in release of unburned fuel and inconsistent syngas production. This is important for reformer operation with minimal supervision requirements. A flame is stabilized at the porous media bed outlet when the effective burning speed equals the unburned gas velocity exiting the burner. A burning rate ratio (BRR) is defined as BRR ¼
Fig. 7 e H2 and CO2 concentrations in reactor exhaust over CH4 flow rates of 12e22 SLPM.
½CH4 reac tan ts ½CH4 products
Vm SL
(16)
where the velocity of gas exiting the bed, Vm, is calculated by dividing the reactant mixture flow rate, Qm, by the crosssectional area of the bed. Fig. 9 shows the burning rate ratios calculated for flow rates from 12 to 22 SLPM, which are above unity for all tested flow rates. The ability of the burner to support stable flames at reactant velocities above the adiabatic flame speed is attributed to two major effects:
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
Fig. 9 e Burning rate ratio and quenching Peclet number over a range of methane flow rates from 12 to 22 SLPM.
preheating and flame wrinkling. Preheating of a reactant mixture leads to an increase in flame speed above its adiabatic value, and hence an increase in the burning rate ratio. The bed temperatures shown in Fig. 5 indicate that there is heat recirculation and preheating occurring in the porous media bed, but that effect alone is insufficient to explain the burning rate ratios significantly above unity achieved in this study. The remainder of the effective burning rate increase over adiabatic flame speed is attributed to flame wrinkling at the reformer outlet, which increases the flame area and permits increased reactant throughput. The turndown ratio and flow rates achieved in the reformer suggest that it may be suitable for use in place of industrial flares, with average fuel flow rates around 350 SLPM or higher, if the reformer in this study were scaled up in diameter [16]. Another critical consideration for the use of reformers in industrial applications is safety of operation. Eq. (15) defines the quenching Peclet number,PeFS, which characterizes the surface stability or flashback potential of the flame. The potential for a flame to propagate into the porous bed or, conversely, the ability of the porous media to quench the flame and prevent propagation of the flame into the bed can be quantified by the quenching Peclet number PeFS ¼
SL dp : a
(17)
Fig. 9 shows the variation of quenching Peclet number across the CH4 flow rate range of 12e22 SLPM. PeFS ranges from 12 to 14 over the tested range, which is well below the critical quenching condition of PeFS < 65 [35,36]; thus, the reformer can be operated safely without the possibility of flame backpropagation. The values of PeFS also suggest that the porosity of the medium could be safely increased in order to reduce pressure drop in the porous media if needed for a particular application.
Conclusions A porous media reformer was demonstrated and utilized to study fuel reformation to syngas using an approach that is suited towards field or remote operation. Two distinguishing
12745
features of the reformer are passive entrainment of air, which reduces the requirements for utilities and active control of the reactant mixture, and surface stabilization of the flame, which also reduces active control requirements and flashback potential. An analytical model of the reformer was developed which predicts air entrainment and resultant equivalence ratio for user-defined reformer geometric parameters and methane flow rate operating conditions. Experimental results are presented which compare favorably with model predictions. Equivalence ratios of 2.32e1.49 were attained through passive air entrainment over the tested range of methane flow rates from 2 to 22 SLPM, and flammable mixtures were achieved for methane flow rates of 6e22 SLPM. Methane flow rates from 12 to 22 SLPM were selected for product composition measurement as high throughput operating points are expected to be of significant interest in industrial applications. Combustion product concentrations of hydrogen, carbon monoxide, carbon dioxide and unburned methane are reported and compared with chemical equilibrium. A high extent of methane conversion to hydrogen, carbon monoxide, and carbon dioxide was observed, suggesting that sufficient premixing of reactants was achieved in the burner design and significant amounts of syngas can be produced via passive air entrainment and surface flame stabilization. Fuel conversion efficiencies of greater than 98% were achieved at all operating conditions, suggesting the suitability of the passive entrainment burner for meeting emissions requirements in envisioned industrial applications.
references
[1] Pedersen-Mjaanes H, Chan L, Mastorakos E. Hydrogen production from rich combustion in porous media. Int J Hydrogen Energy 2005;30(6):579e92. [2] Toledo M, Vergara E, Saveliev AV. Syngas production in hybrid filtration combustion. Int J Hydrogen Energy 2011;36(6):3907e12. [3] Belmont EL, Solomon SM, Ellzey JL. Syngas production from heptane in a non-catalytic counter-flow reactor. Combust Flame 2012;159(12):3624e31. [4] Gentillon P, Toledo M. Hydrogen and syngas production from propane and polyethylene. Int J Hydrogen Energy 2013;38(22):9223e8. [5] Smith CH, Pineda DI, Zak CD, Ellzey JL. Conversion of jet fuel and butanol to syngas by filtration combustion. Int J Hydrogen Energy 2013;38(2):879e89. [6] Lai M-P, Horng R-F, Chen C-Y, Lai W-H. The infrared thermograph observation of a porous medium assisted catalyst packed-bed under excess enthalpy reforming. Int J Hydrogen Energy 2014;39(32):18612e7. [7] Mendes MAA, Pereira JMC, Pereira JCF. On the stability of ultra-lean H2/CO combustion in inert porous burners. Int J Hydrogen Energy 2008;33(13):3416e25. [8] Schoegl I, Newcomb SE, Ellzey JL. Ultra-rich combustion in parallel channels to produce hydrogen-rich syngas from propane. Int J Hydrogen Energy 2009;34(12):5152e63. [9] Toledo M, Bubnovich V, Saveliev A, Kennedy L. Hydrogen production in ultrarich combustion of hydrocarbon fuels in porous media. Int J Hydrogen Energy 2009;34(4):1818e27. [10] Belmont EL, Schoegl I, Ellzey JL. Experimental and analytical investigation of lean premixed methane/air combustion in a
12746
[11]
[12]
[13]
[14]
[15]
[16] [17]
[18]
[19] [20]
[21]
[22]
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
mesoscale counter-flow reactor. Proc Combust Inst 2013;34(2):3361e7. Belmont EL, Ellzey JL. Lean heptane and propane combustion in a non-catalytic parallel-plate counter-flow reactor. Combust Flame 2014;161(4):1055e62. Yang SI, Wu MS. Properties of premixed hydrogen/propane/ air flame in ceramic granular beds. Int J Hydrogen Energy 2014;39(30):17347e57. Bingue JP, Saveliev AV, Fridman AA, Kennedy LA. Hydrogen production in ultra-rich filtration combustion of methane and hydrogen sulfide. Int J Hydrogen Energy 2002;27(6):643e9. Loukou A, Frenzel I, Klein J, Trimis D. Experimental study of hydrogen production and soot particulate matter emissions from methane rich-combustion in inert porous media. Int J Hydrogen Energy 2012;37(21):16686e96. Su S-S, Hwang S-J, Lai W-H. On a porous medium combustor for hydrogen flame stabilization and operation. Int J Hydrogen Energy 2014;39(36):21307e16. U.S. EPA Office of Air Quality Planning and Standards. Parameters for properly designed and operated flares. 2012. Lashof DA, Ahuja DR. Relative contributions of greenhouse gas emissions to global warming. Nature 1990;344(6266):529e31. Caulton DR, Shepson PB, Cambaliza MO, McCabe D, Baum E, Stirm BH. Methane destruction efficiency of natural gas flares associated with shale formation wells. Environ Sci Technol 2014;48(16):9548e54. Ajugwo AO. Negative effects of gas flaring: the Nigerian experience. J Environ Pollut Hum Heal 2013;1(1):6e8. McEwen JDN, Johnson MR. Black carbon particulate matter emission factors for buoyancy driven associated gas flares. J Air Waste Manag 2012;62(3):307e21. Stohl A, Klimont Z, Eckhardt S, Kupiainen K, Shevchenko VP, Kopeikin VM, et al. Black carbon in the Arctic: the underestimated role of gas flaring and residential combustion emissions. Atmos Chem Phys 2013;13(17):8833e55. de Lemos MJS. Turbulence in porous media: modeling and applications. Elsevier; 2012.
[23] Yoksenakul W, Jugjai S. Design and development of a SPMB (self-aspirating, porous medium burner) with a submerged flame. Energy 2011;36(5):3092e100. [24] Barra AJ, Ellzey JL. Heat recirculation and heat transfer in porous burners. Combust Flame 2004;137(1e2):230e41. [25] Makmool U, Jugjai S, Tia S, Vallikul P, Fungtammasan B. Performance and analysis by particle image velocimetry (PIV) of cooker-top burners in Thailand. Energy 2007;32(10):1986e95. s-Parada FJ, Goyeau B, Ochoa[26] Aguilar-Madera CG, Valde Tapia JA. Convective heat transfer in a channel partially filled with a porous medium. Int J Therm Sci 2011;50(8):1355e68. [27] Tamir A, Elperin I, Yotzer S. Performance characteristics of a gas with a swirling central flame. Energy 1989;14(7):373e82. [28] Meadows J, Agrawal AK. Time-resolved PIV of lean premixed combustion without and with porous inert media for acoustic control. Combust Flame 2015;162(4):1063e77. € cker K, Pickena € cker O, Trimis D, [29] Brenner G, Pickena Wawrzinek K, Weber T. Numerical and experimental investigation of matrix-stabilized methane/air combustion in porous inert media. Combust Flame 2000;123(1e2):201e13. [30] Claypole TC, Syred N. The effect of swirl burner aerodynamics on NOx formation. 18th Int Symp Comb 1981:81e9. [31] Namkhat A, Jugjai S. Primary air entrainment characteristics for a self aspirating burner: model and experiments. Energy 2010;35:1701e8. [32] Ergun S. Fluid flow through packed columns. Chem Eng Prog 1952;48(2):89e94. [33] American Petroleum Institute. API standard 520 Part I: sizing, selection, and installation of pressure-relieving devices in refineries. 2005. [34] Whitty KJ, Zhang HR, Eddings EG. Emissions from syngas combustion. Combust Sci Technol 2008;180(6):1117e36. [35] Babkin VS, Korzhavin AA, Bunev VA. Propagation of premixed gaseous explosion flames in porous media. Combust Flame 1991;87(2):182e90. [36] Al-Hamamre Z, Al-Zoubi A. The use of inert porous media based reactors for hydrogen production. Int J Hydrogen Energy 2010;35(5):1971e86.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Experimental and analytical study of a porous media reformer with passive air entrainment S.R. Addamane, M. Hajilou, E.L. Belmont* University of Wyoming, Mechanical Engineering Dept, 1000 E University Ave, Laramie WY 82071, USA
article info
abstract
Article history:
A porous media fuel reformer featuring passive air entrainment and a surface-stabilized
Received 8 February 2016
flame is demonstrated and utilized for the reformation of methane. Passive air entrain-
Received in revised form
ment negates the need for auxiliary resources beyond a source of compressed fuel gas for
25 April 2016
operation, making a reformer of this design promising for applications such as replace-
Accepted 5 May 2016
ment of hydrocarbon flaring with syngas generation at isolated oil and gas production sites
Available online 4 June 2016
and generation of syngas in other remote locations. The porous media reformer design presented in this study incorporates an eductor to entrain ambient air and a swirl mixing
Keywords:
chamber and porous media bed to premix fuel and oxidizer before combustion. A range of
Porous media
methane flow rates was tested in this study, and the effect on air entrainment, and
Combustion
therefore equivalence ratio, was examined. A wide range of stable operating conditions
Passive entrainment
shows a large turndown ratio of the reformer and burning rate ratios greater than unity.
Reforming
Fuel conversion efficiency and syngas production were evaluated and concentrations of
Syngas
methane, hydrogen, carbon monoxide and carbon dioxide are presented for a range of
Flare
methane flow rates. Product compositions are compared to equilibrium and a high extent of fuel conversion efficiency is shown. An analytical model that accounts for reformer geometry and operating conditions to predict air entrainment and equivalence ratio is presented and compared with the experimental results. The analytical model compares favorably with the experimental results and can guide future reformer development and scaling. © 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Introduction Extensive research has examined catalytic and non-catalytic routes for reforming fuels to hydrogen-rich synthesis gas, or syngas. Thermal partial oxidation for syngas production has been shown to be a robust technique that is tolerant to wide variations in fuel composition [1e6], and heat recirculation is frequently used to extend flammability limits and accelerate conversion rates [7e12]. While highly effective in reforming
fuels to hydrogen-rich syngas, filtration combustors and other heat-recirculating reactors often need to be actively controlled to overcome operational challenges such as propagating flame fronts, flashback, and soot formation [5,13e15]. There are scenarios in which fuel reformation could be implemented but the costs and complexities associated with catalytic or excess enthalpy techniques are currently prohibitive. Oil and gas production, for example, produce significant amounts of fuel waste that could be converted to syngas, and the likelihood of industries doing so is significantly improved if a
* Corresponding author. E-mail address: [email protected] (E.L. Belmont). http://dx.doi.org/10.1016/j.ijhydene.2016.05.035 0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
Nomenclature A BRR dp keff L n_ P Pe Q SL T U V
cross-sectional area, [m2] burning rate ratio bead diameter, [m] effective loss coefficient length of the bed, [m] molar flow rate pressure, [Pa] peclet number flow rate at room temperature, [m3/s] flame speed, [m/s] temperature, [K] superficial velocity at the cross section of the bed, [m/s] velocity, [m/s]
Greek symbols f equivalence ratio a thermal diffusivity, [m2/s] m dynamic viscosity, [Ns/m2] ∞ atmospheric condition ε porosity h conversion efficiency of methane, [%] r density, [kg/m3] Subscript a FS g h i l m m,m p t
air at ambient temperature flame speed methane at ambient temperature exit of the burner fuel injector orifice in eductor inlet of the burner air-fuel mixture mixture at average porous media bed temperature porous media bed eductor throat
robust and low cost method is available. Significant quantities of natural gas and larger hydrocarbons are currently flared at oil production sites that do not have sufficient auxiliary resources, such as electricity, to provide active control of flare gas combustion. Associated gas is sometimes used at these sites to provide pneumatic motive force for early oil production and equipment, after which the gas is flared. Additionally, natural gas wells that are drilled but not used for production will have a period of flaring prior to being temporarily shut [16]. Since hydrocarbons have at least ten times higher global warming potential by weight than carbon dioxide [17], and direct release of large amounts of combustible gas can produce an explosion hazard near the production site, unused gas is flared instead of released directly into the atmosphere. Flaring of hydrocarbon-rich gases has increased significantly in the past decade due to advanced oil recovery efforts in formations that produce natural gas in addition to oil [18], and flaring has contributed significantly to an increase
12739
in greenhouse gas emissions. Incomplete combustion of flare gas can produce significant amounts of pollutants that damage the environment and human health, leading to cancer and neurological, reproductive and developmental effects [19]. One of these pollutants is soot, which is hazardous to inhale due to small particle sizes that can travel readily into lungs and is known to contribute to climate change through high absorption of solar radiation, called radiative forcing [20,21]. This study focuses on development and use of a porous media reformer to convert hydrocarbon fuels, which are currently flared or otherwise wasted, to clean burning and utilizable synthesis gas. Such synthesis gas could be upgraded, utilizing only the hydrogen content or the hydrogen and carbon monoxide, or burned directly for power production. The reactor presented in this study incorporates design features to promote air entrainment and premixing in order to produce combustible reactants for maximal combustion efficiency. One of these design features is an eductor, which utilizes the momentum of a pressurized fuel gas jet to entrain ambient air into the fuel gas stream. Another feature is a swirl chamber, which mixes the fuel and gas by inducing swirl in the flow as it enters the burner body [22]. A third significant feature is the porous media bed; porous media has been widely used in the construction of experimental combustion devices to promote premixing and preheating of reactants prior to combustion [23e25]. Heat transfer through porous media has been extensively studied, and found to improve with decreasing media porosity [26]. In addition to heat recirculation, porous media also promotes premixing of gases through its pore networks [27]. The burner design presented in this study utilizes porous media to premix fuel and air and recirculate heat from the flame in order to preheat incoming reactants and promote flame stability over a wide range of fuel flow rates and equivalence ratios. Porous media has also been observed to slightly dampen burner acoustics during operation, which could assist in mitigating noise from industrial flares [28]. The present study experimentally and analytically examines operation of the porous media reformer with a surfacestabilized flame because the surface flame permits safe operation without risk of flashback over a wide range of operating conditions, and the surface flame reduces the risk of carbon build-up within the porous media bed from pyrolysis of large hydrocarbons at elevated temperatures [5]. This mode of operation is achieved by the use of fine diameter porous media and a range of high fuel flow rates over which the reformer is operated. A surface-stabilized flame distinguishes the current approach from many other burner methods and designs which utilize a submerged flame, and is intended to permit operation over long periods of time without the need for active control or additional methods of flame stabilization [29]. Equivalence ratios and syngas production are quantified by experimental measurements. An analytical model of the reformer is presented that predicts the rate of air entrainment into the fuel stream, and therefore reactant equivalence ratio, as a function of geometric and operating parameters. The model also provides guidelines for geometric scaling, permitting the future construction of a reformer that is capable of accommodating a wide range of reactant flow rates for a given application.
12740
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
Materials and methods Reactor apparatus Fig. 1 is a schematic of the porous media reformer used in this study. The apparatus includes an eductor, into which fuel is metered using a mass flow controller and air is passively entrained. Fig. 2 is a schematic of the eductor showing orifice and throat diameters, as well as the area of the air intake at the eductor. The fuel and air reactant mixture is fed from the eductor to a swirl chamber, where the flow is introduced tangentially to induce swirl and promote mixing. The swirl chamber has an internal diameter of 0.2 m and a height of 0.08 m. The geometric swirl number, Sg, of the chamber is estimated to be on the order of 1, which has been found to promote mixing [30]. The reactant mixture enters the bottom of the porous media bed upon exiting the swirl chamber. The porous media bed is retained by a steel cylinder, 8 cm in height, which is positioned atop the swirl chamber. An alumina sleeve inside the steel cylinder acts as an insulator, and is the same height as the steel cylinder. The alumina sleeve is filled with 1.6 mm diameter alumina beads. Five Ktype thermocouples are inserted into the porous media to the center of the reactor, and temperature variation along the height of the bed is continuously monitored during experiments. An Agilent CP 490 gas chromatograph (GC) is used to measure reactant and product species concentrations at the reactor outlet. Two columns are used to measure species in this study: a molecular sieve measures diatomic hydrogen (H2), oxygen (O2), nitrogen (N2), carbon monoxide (CO) and methane (CH4), and a porous polymer unit measures carbon dioxide (CO2). GC samples are drawn from the sampling chimney using a quartz probe with a 2 mm inner diameter and a 0.1 mm orifice diameter. Samples are transported through inert Silco stainless steel tubing that connects the probe to the GC, and pass through a membrane filter prior to entering the GC.
Fig. 1 e Schematic of the porous media reformer is shown, including components that promote passive air entrainment and fuel-air premixing.
Fig. 2 e Schematic of eductor with dimensions of fuel orifice and air intake shown.
Experimental method Methane fuel is utilized in this study because, while one potential envisioned application is replacement of flaring, typical flares vent CH4, larger alkanes, olefins and traces of hydrogen sulfide, and the volume fraction of CH4 can exceed 75% [16]; hence, CP grade CH4 (99% purity) was used as a single component representative fuel in this study. Fuel was introduced to the eductor over a range of flow rates from 2 to 22 SLPM. Ambient air, not dried or otherwise altered, was entrained for combustion. The equivalence ratio of unignited CH4 and air exiting the reactor was measured by GC at the top of the porous media. Equivalence ratio was determined from measured concentrations of O2 and CH4 at the porous media burner surface. The reactant mixture was ignited at the reactor outlet using a secondary flame from a butane lighter. Ignition was rapid and performed at a CH4 flow rate of 12 SLPM. Upon ignition, bed temperatures were continuously monitored and observed to rise until steady state was reached. Steady state was defined as less than ±0.5 K variation over 15 min of continuous burner operation. Once steady state was reached, exhaust gas combustion products were measured by GC and extractive probe sampling. In order to prevent secondary air entrainment at the reactor outlet, and permit the study of combustion resulting from primary air entrainment at the eductor, a quartz chimney of 0.6 m height and 0.2 m internal diameter was placed on top of the steel porous media bed housing during sampling. The chimney had no noticeable effect on burner performance based upon reactor bed temperatures and observed stability. Following the acquisition of burned gas samples, the reactor was extinguished by briefly interrupting CH4 flow to the burner. Once CH4 flow was resumed, unburned gas samples were taken again in order to determine the high temperature and steady state equivalence ratio of the operating point. The surface temperature of the porous bed was measured using a K-type thermocouple after steady state was achieved and the flame was extinguished. Measurement of CH4, O2, H2, CO, and CO2 was performed for a range of CH4 flow rates from 12 to 22 SLPM. This flow rate range subset was chosen for product speciation because of the nearly steady equivalence ratio observed over this range. Species concentrations are reported on a dry basis. Experimentally measured concentrations of CH4, H2, CO, and CO2 in combustion products were compared with chemical
12741
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
equilibrium, which was calculated for the measured equivalence ratios of the reactant mixtures over CH4 flow rates of 12e22 SLPM, to determine the extent of fuel conversion and the progression of combustion towards equilibrium. Each operating point was tested twice. Error analysis was conducted to evaluate the uncertainty of reactant flow rates, equivalence ratios, and species concentrations as measured by GC. Flow rate uncertainty is due to uncertainty of the methane mass flow controller and is ±0.5 SLPM. Uncertainty in equivalence ratio is due to repeatability of measurements and uncertainty in GC calibration gas compositions, and is calculated according to a student-t distribution to have an average value of ±0.04 across the tested flow rates. Uncertainties in measured gas species are attributed to repeatability, which is assessed according to a student-t distribution, and uncertainty in GC calibration gases. The average uncertainties of CH4, O2, H2, CO and CO2 concentrations are calculated to be ±0.21, 0.20, 1.40, 1.14 and 0.75%, respectively.
The rate of air entrainment is a critical performance parameter because it determines the equivalence ratio of the reactant mixture. One-dimensional analysis is utilized in this model, in which parameters are assumed to vary in the bulk flow path through the reactor. Assuming Ai ≪ At and velocity of air significantly less than that of CH4, a momentum balance between the eductor inlet and throat is given by Ref. [31]. rg Qg2 Ai
rm Qm2 At
(1)
Assuming heat transfer to the bed is approximately equal to change in enthalpy of the reactants, and accounting for change in kinetic energy of the reactants by heating as well as the pressure loss in the porous bed and minor losses in the remainder of the reactor, an energy balance between the eductor throat and the burner outlet is given by 2 rm;h Qm;h r Q2 P∞ Pt ¼ m 2m 1 keff DPp : 2 2Ap 2At
(2)
Combining Eqs. (1) and (2), using the ideal gas approximation to relate volumetric flow rate to temperature, and rearranging, Eq. (3) is obtained 2 rg Qg2 rm;h Qm;h r Q2 þ DPp : 0 ¼ m 2m 1 þ keff 2 2Ap At Ai 2At
(3)
Using relations obtained by conservation of mass Qm ¼ Qa þ Qg and rm ¼
ra Qa þ rg Qg Qa þ Qg
(4)
Eq. (3) can be rearranged to: ra Qa2
þ
rg Qg2
1 þ keff 2 r Q 2 þ r Q 2 þ r þ r Q Q a g rg Qg2 a a g g a g Th þ DPp : Tl 2A2p At Ai
0¼
! 2 ð1 εÞ2 Qm Tm 150mm L Tm ¼ DPp d2p Tl Tl ε3 Ap 2 2 1:75Lrm Tm ð1 εÞ Qm þ ε3 dp Tl A2p
! 2 2 150mm L Tm ð1 εÞ2 1:75Lrm Tm ð1 εÞ and N ¼ : M¼ 2 2 dp Ap ε3 Tl ε3 Tl dp Abed (8) Combining Eqs. (7) and (5) and dividing by
gives
)
(
(9) Using the definition of equivalence ratio n_fuel n_air actual f¼ n_fuel n_air stoichiometric
(10)
which can be rewritten for methane-air combustion in terms of the volumetric flow rates of fuel and air as Qg f ¼ 9:52 Qa
(11)
the ratio of volumetric flow rates in Eq. (9) can be rewritten and Eq. (9) can be simplified as ) !# 2 1 1 Th 1 M N 1 þ k eff At Ai 2A2t Qg Tl 2A2p ( ) " 2 i 1 h Th 1 þ N 1 þ k þ 9:52*f ra þ rg eff Tl 2A2p 2A2t !# ( (12) M 1 1 þ k þ 9:522 * ra þ eff Qg 2A2t )! 2 Th 1 N ¼ 0: Tl 2A2p 2 Taking a ¼ rg ; b ¼ ra , c ¼ At1A ; d ¼ 2A1 2 ð1 þ keff Þ; e ¼ TThl 2A1 2 , i p t Eq. (12) can be represented as "
(
rg
f2
aðc d e NÞ
2A2t
þ
The pressure drop through the porous media bed, DPp, can be found by applying the Ergun equation [32].
Qg2
! 2 1 1 Th 1 M N 1 þ keff 0 ¼ 2 rg At Ai 2A2t Qg Qg Tl 2A2p )! ( 2 2 Q 1 Th 1 N 1 þ keff þ a2 ra Qg Tl 2A2p 2A2t ! ( ) 2 i 1 Qa Qg h Th 1 M r : þ r þ N þ 1 þ k þ eff a g Qg Qg2 Tl 2A2p 2A2t
f2
(5)
(7)
where the coefficients of Eq. (7) can be represented by
"
þ ra þ rg Qa Qg
(6)
Rearrangement of Eq. (6) to reference all flow rates at ambient temperature gives
Qg2
Analytical model
ðPt P∞ ÞAt ¼
! 150mL ð1 εÞ2 1:75Lr ð1 εÞ 2 DPp ¼ U : U þ d2p dp ε3 ε3
M Qg
!#
M Qg
!#
" 9:52*f
ða þ bÞðd þ e þ NÞ (13)
9:522 *bðd þ e þ NÞ ¼ 0:
Using the quadratic equation shown in Eq. (13), f can be
12742
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
predicted for a given burner geometry, operating condition and bed temperature.
Results and discussion Equivalence ratio The reactant equivalence ratio achieved within the porous burner is determined by the flow rate of CH4 that is introduced into the system by mass flow controller and the amount of air that is passively entrained by the eductor. The amount of air entrainment will be determined by the momentum of CH4 introduced to the system and the pressure drop through the burner; therefore, the equivalence ratio is indirectly controlled through CH4 volumetric flow rate and burner geometry. Fig. 3 shows experimental measurements and analytical model predictions of equivalence ratio as a function of CH4 flow rate. Experimental measurements at ambient temperature and elevated steady-state temperature are included to assess the effect of bed heating and resultant additional pressure drop on air entrainment. Experimental results show a decrease in equivalence ratio from 2.32 to 1.49 with increase in CH4 flow rate from 2 to 22 SLPM. Flammable gas mixtures are produced at CH4 flow rates of 6 SLPM and greater, where sufficient air is entrained to produce reactant mixtures with equivalence ratios below the upper flammability limit of CH4, f ¼ 1.67. An increase in CH4 flow rate above 6 SLPM leads to increased air entrainment and a lower equivalence ratio up to a CH4 flow rate of 12 SLPM. Further increase in methane flow rate above 12 SLPM produces a slight increase in equivalence ratio from f ¼ 1.49 to f ¼ 1.56. Average uncertainty in equivalence ratio is calculated to be ±0.04, therefore f is constant within uncertainty at the highest tested CH4 flow rates. The calibrated analytical model (keff ¼ 0.57) shows quantitative agreement with the experimental results within experimental uncertainty. Slight deviation is seen at the highest tested CH4 flow rate but is within equivalence ratio uncertainty. The otherwise excellent agreement between experimental and analytical results gives insight into the physical effects that underlie the observed equivalence ratio results. The eductor functions by creating a low pressure
Fig. 3 e Experimentally measured equivalence ratios (f) at ambient conditions (Ta) and steady state temperature (Tss) over a range of CH4 flow rates are compared to analytical model results calculated at Ta of 300 K.
region near the air inlet due to the high velocity of the CH4 jet exiting the eductor orifice, thereby entraining air into the CH4 fuel stream. An increase in CH4 flow rate increases the pressure difference between the air opening in the eductor and the atmosphere, leading to higher entrainment of air. Simultaneously, the pressure drop through the porous media bed also increases with an increase in CH4 and air flow rates. This increased pressure drop produces a competing effect that lowers air entrainment. The results of experiments and modeling indicate that air entrainment enhancement by low pressure at the air intake is dominant at low flow rates, leading to a decrease in f with increase in CH4 flow rate. At higher flow rates, the effects of low pressure and increased pressure drop through the bed balance each other and a nearly constant f is produced with change in CH4 flow rate. The rate of experimental air flow entrainment is calculated for a given fuel flow rate using the measured oxygen and CH4 concentrations at the reactor outlet prior to ignition. Entrained air flow rates increase from 8 SLPM to 140 SLPM with increase in CH4 flow rates from 2 SLPM to 22 SLPM; thus, equivalence ratio and total flow rate results highlight the large turndown ratio of the reformer and operating range over which a rich reactant mixture can be produced for combustion and syngas generation. The rate of air flow entrainment into the burner is dependent on a number of factors, including burner geometry and methane flow rate. Control of air flow rate independent of methane flow rate would therefore require variation of burner geometry, such as eductor orifice and throat diameters, and porous media bed height and diameters. The reactor geometry tested in this study was chosen to be similar to that used in Yoksenul and Jugjai [23] in order to validate the concept of passive air entrainment for a porous media burner with surface-stabilized flame, and the analytical model derived in this study gives an indication of how equivalence ratio will vary with changes to reformer geometry. Fig. 4 shows analytical model results for the variation of equivalence ratio calculated for the geometry of the reformer used in this study, and for variation in bed heights and particle diameters. The equivalence ratio is predicted by the model to increase with increase in bed height and with decrease in particle diameter due to increase in pressure drop across the bed, and the resultant reactant mixture remains in the rich and
Fig. 4 e Variation of f with change in geometry and porosity at 20 SLPM CH4 flow rate and ambient temperature of 300 K, calculated using the analytical model, is shown. The geometry of the burner used in this study is marked (*).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
flammable regime over the wide range of particle diameters and bed heights of 0.8e2.0 mm and 0.04e0.16 m, respectively, modeled and shown in Fig. 4.
Bed temperature The extent to which the porous media bed was heated by the surface flame at steady state burner operation was evaluated through monitoring of temperatures during experiments, and the impact of this heating on air entrainment and equivalence ratio was assessed by comparison of experimental measurements and analytical model results. Fig. 5 shows experimentally measured temperatures within the porous media bed. Bed temperatures are slightly elevated at less than 20 K above ambient temperature (300 K) within the majority of the bed. The surface thermocouple shows the most significant temperature elevation, with surface temperature varying from 408 K at 12 SLPM to 342 K at 22 SLPM. These temperatures, which are measured using thermocouples embedded within the porous bed of the reactor, support the observed position of the flame, which is stabilized at the outlet of the burner and not submerged within the porous bed. Experimental temperature results show that heat recirculation is confined to porous media bed depths within 25 mm of the surface of the burner, and temperatures at the bed surface are observed to decrease with increasing CH4 flow rate. A bed Peclet number which represents the ratio of convective to conductive heat transfer is given by Pep ¼
Vm dp : ap
(14)
Pep increases monotonically from 11 to 18 over the range of tested reactant flow rates, indicating an increased impact of convection over conduction within the bed; thus, the depth of temperature penetration within the bed is reduced with increasing reactant flow rate and Pep. As bed temperature is increased through heat recirculation from the flame and conductive and radiative heat transfer through the media, reactant gas temperature also increases as it is preheated by the porous media. Increase in gas temperature corresponds to a decrease in gas density, which leads to an increase in volumetric flow rate of gas through the porous media
Fig. 5 e Experimentally measured temperatures along the depth of the porous media bed at steady state for a range of methane flow rates.
12743
bed. The Ergun equation (Eq. (6)) shows a proportional effect of volumetric flow rate, or velocity U, on pressure drop through the bed, DPp. Fig. 6 shows the effect of elevated bed temperatures on equivalence ratio as calculated using the calibrated analytical model. Equivalence ratio is observed to increase with porous media bed temperature; this affect is attributed to the increase in pressure drop with increased bed temperature and volumetric flow rate, which leads to lower air entrainment and therefore higher equivalence ratios. The pressure drop across the bed, calculated for the operating conditions tested in this study, is between 5 and 40 Pa over the tested conditions, suggesting the utility of the reformer for industrial applications that require minimal backpressure [33]. The analytically observed effect of bed heating on air entrainment and equivalence ratio can be utilized to understand the experimentally measured ambient and steady state equivalence ratios, shown in Fig. 3, for CH4 flow rates from 12 to 22 SLPM. The experimentally measured equivalence ratios show no significant dependence on temperature when measured at ambient or steady state conditions. The average temperature measured throughout the porous bed over the range of tested methane flow rates was 313e321 K, with a substantial portion of the bed depth showing little temperature rise of less than 20 K above ambient. Model predictions, shown in Fig. 6, indicate that a change in average bed temperature from 300 K to 321 K induces a change in equivalence ratio of less than 0.03, which is below the calculated uncertainty of the equivalence ratio. Hence, the lack of significant change in f at steady state temperature observed in experiments is attributed to low average temperature rise throughout the porous media bed at experimentally measured conditions, but would be expected to be more substantial if a higher bed temperature rise occurred.
Emissions and conversion efficiency The equivalence ratio produced by passive entrainment was observed in Fig. 3 to decrease with increase in CH4 flow rate, and nearly stabilize at the highest tested flow rates. In order to achieve high extents of conversion of CH4 to syngas, a rich
Fig. 6 e Analytical model results show the variation of f for a range of mean bed temperatures at the CH4 flow rates tested in this study.
12744
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
equivalence ratio is desired. In the present study, a nearly constant equivalence ratio was achieved over a CH4 flow rate range of 12e22 SLPM and combustion product measurements were made at these operating conditions. This range of flow rates is highlighted in Fig. 3. Combustion product gases were analyzed in the reactor exhaust by GC after the burner reached steady state. Fig. 7 shows H2 and CO2 concentrations measured in the reactor exhaust, as well as equilibrium concentrations calculated for the equivalence ratios measured under steady state conditions. H2 increases and CO2 decreases slightly with increase in CH4 flow rate from 12 SLPM to 22 SLPM, following the trends exhibited by H2 and CO2 equilibrium concentrations for the slightly increasing equivalence ratios measured across this range of operating conditions. Fig. 8 shows CO and CH4 concentrations measured in the reactor exhaust, as well as equilibrium concentrations calculated for the equivalence ratios measured under steady state conditions. CO equilibrium concentrations increase slightly from 10.0% to 10.7% over the tested CH4 flow rate range of 12e22 SLPM due to the slight increase in equivalence ratio over this range. Experimentally measured CO concentrations also increase with CH4 flow rate, and are slightly lower than equilibrium at CH4 flow rates of 12 SLPM and 14 SLPM and equal to equilibrium within calculated uncertainty at CH4 flow rates of 16 SLPM and higher. The product concentrations of CO, as well as those of H2 and CO2 in Fig. 7, are further analyzed through examination of CH4 concentrations in combustion products. As seen in Fig. 8, product concentrations of CH4 are less than 0.4% at all tested flow rates, indicating a high extent of CH4 conversion. Equilibrium concentrations of CH4 are below the detectable limit, with mole fractions on the order of 1 ppt or less, thus the experimentally measured values of CH4 are above equilibrium within the calculated uncertainty of 0.2% at methane flow rates of 20 SLPM and 22 SLPM. These results suggest that low residence times at high reactant flow rates may limit complete conversion of CH4, and the discrepancies at low flow rates may likewise be due to low amounts of unreacted CH4. The efficiency of the passive entrainment reformer examined in this study can be further evaluated by the extent of CH4 conversion, or conversion efficiency, defined as
Fig. 8 e CO and CH4 concentrations in reactor exhaust over methane flow rates of 12e22 SLPM.
hCH4 ¼
½CH4 reac tan ts ½CH4 equilibrium
! $100:
(15)
Methane conversion efficiency calculated for the results shown in Fig. 8 is greater than 98% at all tested flow rates. Taken together with the measured product concentrations of H2, CO and CO2, the results suggest that the conversion of CH4 is nearly complete at all tested firing rates and a significant amount of syngas was generated at the tested operating conditions.
Suitability for industrial use Experimental results indicate the utility of the passive entrainment reformer for the production of syngas as a clean burning fuel [34]. The results of experiments and analytical modeling suggest a broad operating range that produces high conversion efficiency to syngas with insensitivity to variation in CH4 flow rates. Stable operation over a wide range of operating conditions is advantageous when utilizing a syngas reformer, whether in field applications with little or no utilities available for active control or in a highly controlled industrial setting, as instabilities can result in release of unburned fuel and inconsistent syngas production. This is important for reformer operation with minimal supervision requirements. A flame is stabilized at the porous media bed outlet when the effective burning speed equals the unburned gas velocity exiting the burner. A burning rate ratio (BRR) is defined as BRR ¼
Fig. 7 e H2 and CO2 concentrations in reactor exhaust over CH4 flow rates of 12e22 SLPM.
½CH4 reac tan ts ½CH4 products
Vm SL
(16)
where the velocity of gas exiting the bed, Vm, is calculated by dividing the reactant mixture flow rate, Qm, by the crosssectional area of the bed. Fig. 9 shows the burning rate ratios calculated for flow rates from 12 to 22 SLPM, which are above unity for all tested flow rates. The ability of the burner to support stable flames at reactant velocities above the adiabatic flame speed is attributed to two major effects:
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
Fig. 9 e Burning rate ratio and quenching Peclet number over a range of methane flow rates from 12 to 22 SLPM.
preheating and flame wrinkling. Preheating of a reactant mixture leads to an increase in flame speed above its adiabatic value, and hence an increase in the burning rate ratio. The bed temperatures shown in Fig. 5 indicate that there is heat recirculation and preheating occurring in the porous media bed, but that effect alone is insufficient to explain the burning rate ratios significantly above unity achieved in this study. The remainder of the effective burning rate increase over adiabatic flame speed is attributed to flame wrinkling at the reformer outlet, which increases the flame area and permits increased reactant throughput. The turndown ratio and flow rates achieved in the reformer suggest that it may be suitable for use in place of industrial flares, with average fuel flow rates around 350 SLPM or higher, if the reformer in this study were scaled up in diameter [16]. Another critical consideration for the use of reformers in industrial applications is safety of operation. Eq. (15) defines the quenching Peclet number,PeFS, which characterizes the surface stability or flashback potential of the flame. The potential for a flame to propagate into the porous bed or, conversely, the ability of the porous media to quench the flame and prevent propagation of the flame into the bed can be quantified by the quenching Peclet number PeFS ¼
SL dp : a
(17)
Fig. 9 shows the variation of quenching Peclet number across the CH4 flow rate range of 12e22 SLPM. PeFS ranges from 12 to 14 over the tested range, which is well below the critical quenching condition of PeFS < 65 [35,36]; thus, the reformer can be operated safely without the possibility of flame backpropagation. The values of PeFS also suggest that the porosity of the medium could be safely increased in order to reduce pressure drop in the porous media if needed for a particular application.
Conclusions A porous media reformer was demonstrated and utilized to study fuel reformation to syngas using an approach that is suited towards field or remote operation. Two distinguishing
12745
features of the reformer are passive entrainment of air, which reduces the requirements for utilities and active control of the reactant mixture, and surface stabilization of the flame, which also reduces active control requirements and flashback potential. An analytical model of the reformer was developed which predicts air entrainment and resultant equivalence ratio for user-defined reformer geometric parameters and methane flow rate operating conditions. Experimental results are presented which compare favorably with model predictions. Equivalence ratios of 2.32e1.49 were attained through passive air entrainment over the tested range of methane flow rates from 2 to 22 SLPM, and flammable mixtures were achieved for methane flow rates of 6e22 SLPM. Methane flow rates from 12 to 22 SLPM were selected for product composition measurement as high throughput operating points are expected to be of significant interest in industrial applications. Combustion product concentrations of hydrogen, carbon monoxide, carbon dioxide and unburned methane are reported and compared with chemical equilibrium. A high extent of methane conversion to hydrogen, carbon monoxide, and carbon dioxide was observed, suggesting that sufficient premixing of reactants was achieved in the burner design and significant amounts of syngas can be produced via passive air entrainment and surface flame stabilization. Fuel conversion efficiencies of greater than 98% were achieved at all operating conditions, suggesting the suitability of the passive entrainment burner for meeting emissions requirements in envisioned industrial applications.
references
[1] Pedersen-Mjaanes H, Chan L, Mastorakos E. Hydrogen production from rich combustion in porous media. Int J Hydrogen Energy 2005;30(6):579e92. [2] Toledo M, Vergara E, Saveliev AV. Syngas production in hybrid filtration combustion. Int J Hydrogen Energy 2011;36(6):3907e12. [3] Belmont EL, Solomon SM, Ellzey JL. Syngas production from heptane in a non-catalytic counter-flow reactor. Combust Flame 2012;159(12):3624e31. [4] Gentillon P, Toledo M. Hydrogen and syngas production from propane and polyethylene. Int J Hydrogen Energy 2013;38(22):9223e8. [5] Smith CH, Pineda DI, Zak CD, Ellzey JL. Conversion of jet fuel and butanol to syngas by filtration combustion. Int J Hydrogen Energy 2013;38(2):879e89. [6] Lai M-P, Horng R-F, Chen C-Y, Lai W-H. The infrared thermograph observation of a porous medium assisted catalyst packed-bed under excess enthalpy reforming. Int J Hydrogen Energy 2014;39(32):18612e7. [7] Mendes MAA, Pereira JMC, Pereira JCF. On the stability of ultra-lean H2/CO combustion in inert porous burners. Int J Hydrogen Energy 2008;33(13):3416e25. [8] Schoegl I, Newcomb SE, Ellzey JL. Ultra-rich combustion in parallel channels to produce hydrogen-rich syngas from propane. Int J Hydrogen Energy 2009;34(12):5152e63. [9] Toledo M, Bubnovich V, Saveliev A, Kennedy L. Hydrogen production in ultrarich combustion of hydrocarbon fuels in porous media. Int J Hydrogen Energy 2009;34(4):1818e27. [10] Belmont EL, Schoegl I, Ellzey JL. Experimental and analytical investigation of lean premixed methane/air combustion in a
12746
[11]
[12]
[13]
[14]
[15]
[16] [17]
[18]
[19] [20]
[21]
[22]
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 1 2 7 3 8 e1 2 7 4 6
mesoscale counter-flow reactor. Proc Combust Inst 2013;34(2):3361e7. Belmont EL, Ellzey JL. Lean heptane and propane combustion in a non-catalytic parallel-plate counter-flow reactor. Combust Flame 2014;161(4):1055e62. Yang SI, Wu MS. Properties of premixed hydrogen/propane/ air flame in ceramic granular beds. Int J Hydrogen Energy 2014;39(30):17347e57. Bingue JP, Saveliev AV, Fridman AA, Kennedy LA. Hydrogen production in ultra-rich filtration combustion of methane and hydrogen sulfide. Int J Hydrogen Energy 2002;27(6):643e9. Loukou A, Frenzel I, Klein J, Trimis D. Experimental study of hydrogen production and soot particulate matter emissions from methane rich-combustion in inert porous media. Int J Hydrogen Energy 2012;37(21):16686e96. Su S-S, Hwang S-J, Lai W-H. On a porous medium combustor for hydrogen flame stabilization and operation. Int J Hydrogen Energy 2014;39(36):21307e16. U.S. EPA Office of Air Quality Planning and Standards. Parameters for properly designed and operated flares. 2012. Lashof DA, Ahuja DR. Relative contributions of greenhouse gas emissions to global warming. Nature 1990;344(6266):529e31. Caulton DR, Shepson PB, Cambaliza MO, McCabe D, Baum E, Stirm BH. Methane destruction efficiency of natural gas flares associated with shale formation wells. Environ Sci Technol 2014;48(16):9548e54. Ajugwo AO. Negative effects of gas flaring: the Nigerian experience. J Environ Pollut Hum Heal 2013;1(1):6e8. McEwen JDN, Johnson MR. Black carbon particulate matter emission factors for buoyancy driven associated gas flares. J Air Waste Manag 2012;62(3):307e21. Stohl A, Klimont Z, Eckhardt S, Kupiainen K, Shevchenko VP, Kopeikin VM, et al. Black carbon in the Arctic: the underestimated role of gas flaring and residential combustion emissions. Atmos Chem Phys 2013;13(17):8833e55. de Lemos MJS. Turbulence in porous media: modeling and applications. Elsevier; 2012.
[23] Yoksenakul W, Jugjai S. Design and development of a SPMB (self-aspirating, porous medium burner) with a submerged flame. Energy 2011;36(5):3092e100. [24] Barra AJ, Ellzey JL. Heat recirculation and heat transfer in porous burners. Combust Flame 2004;137(1e2):230e41. [25] Makmool U, Jugjai S, Tia S, Vallikul P, Fungtammasan B. Performance and analysis by particle image velocimetry (PIV) of cooker-top burners in Thailand. Energy 2007;32(10):1986e95. s-Parada FJ, Goyeau B, Ochoa[26] Aguilar-Madera CG, Valde Tapia JA. Convective heat transfer in a channel partially filled with a porous medium. Int J Therm Sci 2011;50(8):1355e68. [27] Tamir A, Elperin I, Yotzer S. Performance characteristics of a gas with a swirling central flame. Energy 1989;14(7):373e82. [28] Meadows J, Agrawal AK. Time-resolved PIV of lean premixed combustion without and with porous inert media for acoustic control. Combust Flame 2015;162(4):1063e77. € cker K, Pickena € cker O, Trimis D, [29] Brenner G, Pickena Wawrzinek K, Weber T. Numerical and experimental investigation of matrix-stabilized methane/air combustion in porous inert media. Combust Flame 2000;123(1e2):201e13. [30] Claypole TC, Syred N. The effect of swirl burner aerodynamics on NOx formation. 18th Int Symp Comb 1981:81e9. [31] Namkhat A, Jugjai S. Primary air entrainment characteristics for a self aspirating burner: model and experiments. Energy 2010;35:1701e8. [32] Ergun S. Fluid flow through packed columns. Chem Eng Prog 1952;48(2):89e94. [33] American Petroleum Institute. API standard 520 Part I: sizing, selection, and installation of pressure-relieving devices in refineries. 2005. [34] Whitty KJ, Zhang HR, Eddings EG. Emissions from syngas combustion. Combust Sci Technol 2008;180(6):1117e36. [35] Babkin VS, Korzhavin AA, Bunev VA. Propagation of premixed gaseous explosion flames in porous media. Combust Flame 1991;87(2):182e90. [36] Al-Hamamre Z, Al-Zoubi A. The use of inert porous media based reactors for hydrogen production. Int J Hydrogen Energy 2010;35(5):1971e86.