- Email: [email protected]
Numerical study of methane TPOX within a small scale Inert Porous Media based reformer
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Numerical study of methane TPOX within a small scale Inert Porous Media based reformer Miguel A.A. Mendes*, Jose´ M.C. Pereira, Jose´ C.F. Pereira Universidade de Lisboa, Instituto Superior Te´cnico, Mech. Eng. Dept., LASEF, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal
article info
abstract
Article history:
Methane Thermal Partial Oxidation (TPOX) within a small scale Inert Porous Media (IPM)
Received 11 September 2013
based reactor was investigated numerically in order to explore the operating conditions
Received in revised form
and possible procedures for maximizing the reforming efficiency and minimizing the soot
17 December 2013
formation. A quasi-1D model of the TPOX reactor was validated and further used to study
Accepted 28 December 2013
the process. The model considers detailed chemistry and solves the energy balances for
Available online 31 January 2014
both gas and solid phases, including radiative heat transfer in the solid phase. The parametric results of the reactor operation show that the optimal airefuel ratio is a compro-
Keywords:
mise between soot formation and reforming efficiency. Moreover, a high preheating
Hydrogen production
temperature of the reactants is found to be always beneficial for the process, and the effect
Methane reforming
of power input is negligible for the reforming efficiency. The numerical investigations also
Thermal partial oxidation
suggest that shorting the IPM length, as well as mixing small amounts of water vapor with
Inert Porous Media
the reactants, appear to be effective procedures for improving the operation performance
Combustion modeling
of the TPOX reactor. Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Power generation from fossil fuels has been widely considered as one of the main contributors for global warming, while at the same time the reserves of these non-renewable energy sources are rapidly depleting worldwide. Therefore, there is an increasing interest in developing power generation systems based on clean and sustainable energy sources. However, around 80% of the total energy consumption in the world is based yet on fossil fuel reserves. Consequently, in order to gradually convert the actual power generation systems from conventional to renewable, a smooth transition is required to meet the actual energy demand [1]. Emerging power generation technologies, such as SolidOxide Fuel Cells (SOFCs), are particularly attractive, when
compared to conventional technologies, due to their potential for efficient use of hydrocarbon fuels and reduced emissions. SOFCs present very high efficiencies at elevated working temperatures (500e1000 C) being therefore attractive for use in commercial and industrial Combined Heat and Power (CHP) applications [2e4]. Although other fuel cell types, such as, Molten-Carbonate Fuel Cells (MCFCs) or High-Temperature Polymer Electrolyte Membrane Fuel Cells (HT-PEMFC), also present a great potential for CHP [5,6]. In addition, SOFCs are fuel flexible and are able to operate either on pure hydrogen or hydrocarbon fuels. Therefore, systems based on SOFC technology have the potential to provide a transition solution from fossil fuels to more clean and sustainable energy carriers, such as bio-fuels and hydrogen. However, SOFC operation directly on hydrocarbon fuels is very limited and fuel reforming processes are generally required in order to convert hydrocarbon
* Corresponding author. E-mail addresses: [email protected], [email protected] (M.A.A. Mendes). 0360-3199/$ e see front matter Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.12.192
4312
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Nomenclature _ 00 m A av Cp dp hk Hv k MWk Nu Qr Qw Re T t u vk x Yk
axial mass flux, kg/m2 s reactor cross-sectional area, m2 surface area to volume ratio, m1 specific heat capacity at constant pressure, J/ kg K characteristic pore dimension, m enthalpy of k species, J/kg volumetric convective heat transfer coefficient, W/m3 K thermal conductivity, W/m K molar weight of k species, kg/kmol Nusselt number radiative heat flux, W/m2 wall heat loss per unit of length, W/m Reynolds number temperature, K time coordinate, s axial mean velocity of gas phase, m/s diffusion velocity of k species, m/s axial coordinate, m mass fraction of k species
Greek symbols b solid extinction coefficient, m1 u_ k production rate of k species, kmol/m3 s ˛ solid emissivity l mass ratio of air to fuel m dynamic viscosity, kg/m s u solid scattering albedo f porosity r mass density, kg/m3 s residence time, s Subscript g gas phase s solid phase
fuels into synthesis-gas (consisting mainly of H2 and CO) [7]. Owing to the carbon deposition and anode deactivation problems, external reforming is generally preferred over internal reforming, specially on practical small scale systems. There are three main reforming technologies used for synthesis-gas production from liquid and gaseous hydrocarbon fuels: steam reforming, catalytic or thermal partial oxidation and autothermal reforming [8]. Among these, Thermal Partial Oxidation (TPOX) offers several advantages [9,10], such as: no need for external heat sources and additional feeds; absence of catalysts which eliminates the catalyst deactivation problems; good dynamic response; and hydrocarbon flexibility. However, at the same time it presents comparatively low hydrogen yield and tendency to produce soot. Performing the TPOX within Inert Porous Media (IPM) based reformers is a practical solution to prevent flame stability problems, which are promoted by the slow reaction rates at low adiabatic flame temperatures existing in the TPOX process [11,12]. The high heat recirculation from the hot products to the reactants, provided by the solid matrix, originates local super-adiabatic combustion temperatures, which
increase the reaction rates and the stability of the TPOX process improving its operational characteristics [12]. Comprehensive reviews on IPM combustion can be found in Refs. [13e15]. Several recent studies have investigated TPOX of hydrocarbons within IPM, and different techniques have been successfully applied in order to stabilize the reforming process, see, e.g., Refs. [11,12,16e18]. Studies regarding transient TPOX within IPM can be found in Refs. [16,17]. These works applied the filtration combustion principle, which involves a traveling combustion wave freely propagating within a IPM, where the pore size of the solid matrix is typically sub-critical. Therefore, the IPM must be preheated to high temperatures in order to initialize the combustion wave propagation. On the other hand, stationary techniques for TPOX within IPM can be found in Refs. [11,12,18]. In these techniques, the combustion zone is steadily stabilized inside a porous matrix of supercritical pore size, and the nature of stabilization phenomenon is completely different from the one existing in transient approaches. In Ref. [18], experimental investigations of methaneeair TPOX process within a small scale IPM based reformer were performed for different ranges of inlet temperature, power input and airefuel ratio. The TPOX process was steadily stabilized by employing a reactor formed by a diffuser-like upstream section followed by a downstream cylindrical section, similar to the one presented in Ref. [12], and the IPM consisted of a packed bed of Al2O3 rings. Temperature profiles were measured along the reactor central axis, and concentrations of synthesis-gas species (H2, CO, CO2, CH4, C2H2) were measured at the reactor outlet. To the authors knowledge, there are no studies in the literature regarding the optimization of the TPOX process within IPM based reactors similar to the ones used in Refs. [12,18]. The objective of the present study is to investigate possible solutions in order to maximize the reforming efficiency and minimize the soot formation on the methane TPOX process within a small scale IPM based reactor. For this purpose a quasi-1D combustion model have been used, which was validated against the experimental results from Ref. [18]. The numerical model considers detailed chemical kinetics and solves the gas and solid phase energy balances, including radiative heat transport in the solid phase. With the help of this numerical model, a wide range of operating conditions as well as different procedures were explored in order to evaluate which conditions and procedures maximize the reforming efficiency and minimize the soot formation on this particular TPOX reactor. The present paper is organized in the following form: First, the numerical model of the methane TPOX reactor is described. Further, simulation results are presented, including model validation, parametric study, and different procedures for improving the process. Finally, the main conclusions are summarized.
2.
Numerical model
This section describes the numerical model used in order to simulate the methane TPOX process within a small scale IPM based reactor, experimentally tested in Ref. [18]. The TPOX
4313
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
reactor model was numerically implemented within an inhouse Fortran code, which is a modified version of the PREMIX code [19] that additionally incorporates the solid phase energy balance and the convective heat transfer between gas and solid phases. This numerical model was previously validated with excellent agreement against experimental results [20,21]. The geometrical configuration of the reactor is presented in Fig. 1, where the IPM consists of a packed bed of small Al2O3 rings. For modeling purposes, the IPM is assumed to be an homogeneous continuum with negligible catalytic effects. The ideal gas assumption is considered and the pressure loss along the reactor is neglected. A quasi-1D modeling approach is considered, which solves the 1D governing equations taking into account the cross-sectional area variation of the reactor by correcting the axial mass flux. With these assumptions, the averaged forms of the conservation equations for mass, gaseous species, and gas and solid phase energy balances are given by: Mass Conservation v Afrg _ 00 Þ vðAm ¼0 þ vx vt
(1)
where m00 ¼ frg u is the axial mass flux, Species Mass Fraction Balance v Afr v k Yk g vYk vY _ 00 k þ þ Am Afu_ k MWk ¼ 0 Afrg vt vx vx
(2)
Gas Phase Energy Balance
X vTg vT vTg vTg v _ 00 Cp;g g Afkg þ Am þ Af rg Cp;k vk vt vx vx vx vx k X u_ k hk MWk þ AHv Tg Ts ¼ 0 þAf (3)
Afrg Cp;g
k
Solid Phase Energy Balance vAQr v rs Cp;s Ts v vTs Að1fÞks AHv Tg Ts þ ¼0 Að1fÞ vx vt vx vx (4) Both gas and solid phase energy equations (3) and (4) are coupled by the convective heat transfer term AHv(Tg Ts), in which the volumetric convective heat transfer coefficient Hv is obtained from the following correlation [22]: dp Nu ¼ 2 1:3 þ 4:43 pffiffiffiffi þ 0:17Re0:75 A
INSULATION
POROUS MEDIA
Fig. 1 e Small scale IPM based reactor configuration.
(5)
where the Nusselt number Nu ¼ (Hvdp)/(kgav) and the Reynolds 00 number Re ¼ (m_ dp)/(mg) are based on the characteristic pore dimension dp of the IPM. For radiation purposes, the porous medium is assumed as a diffuse, gray medium together with a non-participative gas mixture. The term v(AQr)/vx appearing in equation (4) represents the radiative heat transport in the solid phase and was numerically solved using the Discrete-Ordinates method (S8 approximation) [23]. The boundary conditions applied to equations (1)e(4) and to the radiative heat transport equations are similar to the ones presented in Ref. [24]. The methaneeair TPOX reactions were modeled on the basis of the chemical kinetic mechanism from Refs. [25], where only the C1C2 chemistry was considered, consisting of 32 species and 150 elementary reactions. The thermo-physical properties of the gas mixture (rg, kg, mg, vk and Cp,g), required by equations (1)e(3), were obtained based on [26,27]. Moreover, the thermo-physical properties (rs, ks and Cp,s) of the Al2O3 ceramic material, required by equation (4), were obtained from Ref. [28] as functions of temperature. The values of the IMP properties used for the simulations are resumed in table 1.
3.
Results
This section presents the numerical predictions of methane TPOX process within IPM obtained with the reactor model described in the previous section. First, the model validation results are resumed, followed by the results of the parametric study of the reactor operation. Finally, the results regarding different procedures for improving the TPOX process are presented.
3.1.
Model validation
In order to validate the model, numerical predictions of the methane TPOX process within a small scale IPM based reactor were compared against the respective experimental results, obtained from Ref. [18]. This comparison was performed for ranges of airefuel ratio (l), inlet gas mixture temperature (Tin), and power input (P). Moreover, reactor heat losses were accounted for by the model for this comparison, i.e., radiative heat losses from the reactor outlet to the downstream environment at 1250 K, as well as heat losses from the reactor side walls to the surroundings at 300 K. The wall heat losses were modeled by including the term Qw in the left-hand side of the solid energy balance, given by equation (4). This term was assumed to be uniform and equal to 0.9 kW/m (based on the experiments), in a similar way to [20].
Table 1 e IPM properties used for the simulations. Porosity, f Pore dimension, dp [m] Specific surface, av [m1] Extinction coefficient, b [m1] Scattering albedo, u Emissivity, ˛
0.63 0.015 320 100 0.8 0.4
4314
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Fig. 2(a) presents the simulated gas and solid temperature profiles together with the experimental temperatures, measured along the reactor center line, for several l and fixing Tin ¼ 823 K and P ¼ 5 kW. Overall temperature measurements are well predicted by the model, particularly the temperatures at the post-flame region. The figure also shows that the position of the flame front is reasonably well captured by the model, however the predicted temperature peak of the gas phase is not clearly observed in the measurements. This can be explained by the fact that the thermocouples inserted in the reactor measured a temperature value that is neither the one of the gas nor solid phase but something in between, see, e.g., Ref. [29]. Moreover, it is expected that the thermocouples strongly felt the presence of the solid due to its high radiative heat transfer and also due to possible contact between the solid matrix and the thermocouples. Nevertheless, the temperature of the porous matrix is lower at the upstream region of the reactor and the higher flow velocities, induced by the smaller cross-sectional area of the reactor, increase the convective heat transfer between the gas phase and thermocouples. Therefore, in this region, the measured temperatures are expected to be closer to the ones of the gas mixture. The comparison between numerical predictions and measurements of the molar fraction of main synthesis-gas species (H2, CO, H2O and CO2) at the reactor outlet are shown in Fig. 2(b) as function of l for Tin ¼ 823 K and P ¼ 5 kW. Similar results are presented in Fig. 2(c) for the minor
synthesis-gas species (CH4 and C2H2). One can observe from Fig. 2(b) that the numerical predictions of the main species are in good agreement with the measurements. Fig. 2(c) shows a good qualitative trend of the minor species prediction. Nevertheless, the quantitative values are far from perfect, and this may be attributed to the soot formation, which is not taken into account by the current model, see, e.g., Ref. [30]. In Ref. [31], the same TPOX reactor considered for the present study was tested for different operation conditions (l, Tin and P) and soot measurements were obtained at the exit of the reactor. They found that for higher values of l, e.g., l ¼ 0.5, no soot was detected. However, Fig. 2(a) shows that the agreement between predicted and measured temperature profiles is not significantly influenced by l. Therefore, it can be concluded that the soot formation, which according to Ref. [31] occurs for lower values of l, has a negligible effect on the temperatures inside the TPOX reactor and hence, the same is expected for the temperatures predicted by the present model. Nevertheless, the soot formation influences the minor species, as pointed above. The over-prediction of the measured C2H2 molar fractions, observed in Fig. 2(c) for lower values of l, can be explained by the fact that C2H2 participates in the chemical pathways that lead to molecular grow of aromatic compounds and formation of soot particles. From the validation results, one can conclude that this simple quasi-1D model is able to capture the main features of the methane TPOX process within the small scale IPM based
Fig. 2 e Comparison between simulations and experimental results for the TPOX reactor operating at several l, fixing P [ 5 kW and Tin [ 823 K: (a) temperature profiles; (b) outlet main species; (c) outlet minor species.
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
4315
reactor tested in Ref. [18]. Therefore, further investigations with this model are expected to be realistic. However, for the sake of simplicity and objectivity, the following results are obtained assuming that the TPOX reactor is adiabatic, i.e., wall heat losses and radiative heat losses at the reactor outlet are neglected.
3.2.
Parametric study
Before presenting the parametric study results, the methane TPOX process within IPM was compared against free flame results in order to understand the main differences between both processes. This comparison was carried out assuming Tin ¼ 823 K. Fig. 3(a) and (b) compares the model predictions for the evolution of the reforming efficiency (h) and the gas temperature profiles, respectively, of the TPOX process for IPM and free flame modes as function of the residence time. The reforming efficiency is defined as the ratio between the low heating value of the gas mixture produced from the reforming process to the low heating value of the original fuel (CH4), and given by: h¼
YH2 LHVH2 þ YCO LHVCO YCH4 LHVCH4
(6)
Fig. 3(a) shows that the presence of IPM accelerates the initial stage of the TPOX process due to the super-adiabatic temperatures induced by the heat recirculation from the hot products to the reactants by both conduction and radiation heat transfer. The gas temperature profiles presented in Fig. 3(b) confirm this efficiency improvement. However, the heat recirculation excessively decreases the products temperature in the post-flame region slowing down the TPOX process reactions. Therefore, after a certain residence time, the free flame mode shows a higher h than the IPM mode. This phenomenon is more evident for lower l values, since at these ultra-rich conditions the TPOX process requires higher residence times to reach equilibrium. A parametric study was conducted using the original TPOX reactor configuration, shown in Fig. 1, operating at different l, Tin and P, keeping as reference the nominal operating conditions: Tin ¼ 823 K and P ¼ 5 kW. The results presented in this section consider the influence of the residence time in an implicit manner through l, Tin and P, which directly influence the value of the mass flow rate of the gas mixture. However, an explicit influence of the residence on reforming efficiency or soot formation is not considered, because the residence time is not an independent input parameter to the TPOX reactor operation, but l, Tin and P are. Fig. 4(a) shows h as function of l for different values of Tin (fixing P ¼ 5 kW), along with the equilibrium reforming efficiency and the maximum theoretical reforming efficiency. The maximum theoretical reforming efficiency is calculated assuming that all CH4 is converted to H2 and CO. One can observe from the figure that, for high values of l, the h of the process is very close to the equilibrium efficiency and is little sensitive to Tin. However, for l lower than a critical value, h departs from the equilibrium efficiency, and the departure starts at higher l values if Tin is lower. This is mainly related to the fact that the residence time required to achieve
Fig. 3 e Comparison of the methane TPOX process, predicted by the model, performed within IPM and in the free flame mode as function of the residence time for two different l: (a) h evolution; (b) gas temperature evolution.
equilibrium increases with the decrease in l and Tin due to the extremely slow reaction rates at low adiabatic temperatures existing at these conditions [12]. Therefore, there exists an optimal l that maximizes h. In order to shift the optimal l to lower values and to increase h to equilibrium values, one would require to increase the residence time of the TPOX process. Fig. 4(b) presents the reforming efficiency as function of P for several l values, fixing Tin ¼ 823 K. It can be observed from the figure that h is nearly independent of P. However, this will not be the case if the reactor heat losses are not negligible, because the resulting lower temperatures inside the reactor would decrease the reforming efficiency, especially at lower P.
4316
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Fig. 4 e Evolution of h: (a) h as function of l for different Tin fixing P [ 5 kW; (b) h as function of P for different l fixing Tin [ 823 K.
Fig. 5 e Evolution of C2H2 molar fraction: (a) C2H2 as function of l for different Tin fixing P [ 5 kW; (b) C2H2 as function of P for different l fixing Tin [ 823 K.
Although the main interest in the TPOX process is to maximize h, the formation of soot particles also plays an important role and it must be minimized. The lack of oxygen and the lower reaction temperatures existing at lower l values, promote soot formation due to the increase of intermediate hydrocarbon species that are the precursors of soot particles. The C2H2 species is the main indicator for the presence of soot due to its deep involvement in the critical steps of soot formation, i.e., inception and surface growth [30]. Therefore, it is reasonable to use the C2H2 concentration in the synthesis-gas as a rough quantifier of soot produced. Fig. 5(a) shows the C2H2 molar fraction as function of l for different values of Tin, fixing P ¼ 5 kW. It can be observed from the figure that the C2H2 increases with the decreasing on both l and Tin, and this increase is rather more drastic due to l than
due to Tin. Therefore, a similar behavior should be expected for the soot produced. Fig. 5(b) presents the C2H2 molar fraction as function of P for different values of l, fixing Tin ¼ 823 K. One can observe from the figure that the C2H2 fraction slightly increases with P, mainly due to the decrease of the residence time, which does not allow a complete conversion of C2H2 into more stable gaseous products. Consequently, this should not be totally interpreted as an increase of soot production, because its formation process also requires some time.
3.3.
Process improvement procedures
Taking into account the previous results for the parametric study, three different procedures were investigated in order to
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
improve the performance of the TPOX process within the small scale reactor. The first procedure is based on modifying the original reactor geometry by excluding the IPM from the downstream cylindrical section (see Fig. 1). The second procedure is focussed on changing the inlet composition by mixing water vapor to the reactants. A third procedure explored the effect of partially recirculating the TPOX products from the outlet to inlet of the reactor. In first place, it was investigated the effect of excluding the IPM from the downstream cylindrical section of the TPOX reactor presented in Fig. 1. The IPM improves the heat recirculation from the hot products to the reactants and, therefore, increases the temperature in the flame front region promoting flame stabilization. However, the high heat recirculation reduces the temperature of the products in the post-combustion zone, strongly decelerating the rates of the slower reforming reactions taking place there. Consequently, by excluding the IPM from the downstream cylindrical section (where it is not relevant for the flame stabilization), one will reduce the heat recirculation from the post-flame hot products to the reactants, keeping these products at a higher temperature. This will also enhance the free space volume and increase the residence time of the gas mixture on the cylindrical region of the reformer. Fig. 6(a) and (b) compares the h calculated using the TPOX reactor with and without the IPM in the cylindrical section (see Fig. 1). Fig. 6(a) presents h as function of l for different Tin values, fixing P ¼ 5 kW, and Fig. 6(b) shows the same information as function of P for several l values, fixing Tin ¼ 823 K. This simple modification in the reactor geometry improves h, especially for lower l and Tin, see Fig. 6(a). Fig. 6(b) shows that h was improved for higher P. However, this efficiency improvement is smaller at lower power operation, because for these cases the reaction front is located deep inside the conical section, where the heat recirculation provided by the IPM is still able to strongly decrease the temperature of the post-flame products. Fig. 7 presents the gas phase temperature profiles for both reformer versions operating at 2 kW and 7 kW, where it is notorious the effect of excluding the IPM from the cylindrical section of the TPOX reactor. Moreover, Fig. 8 presents the axial velocity u and residence time s of the gas mixture inside the TPOX reactor with and without IPM in the cylindrical section, for a particular operating condition: Tin ¼ 823 K, P ¼ 5 kW and l ¼ 0.36. One can observe from the figure that the exclusion of the IPM from the cylindrical section of the reactor increases the residence time around 1.5 times, which is a consequence of the lower axial velocity. Fig. 9(a) and (b) shows the C2H2 molar fraction calculated as function of l and P, respectively, for the TPOX reactor with and without IPM in the cylindrical section. By removing the IPM from the cylindrical section, the C2H2 concentration and consequently the soot formation decrease for high P and lower l. This is justified by the extended residence time and higher temperatures within the cylindrical section of the TPOX reactor, which allow a more complete conversion of C2H2 into more stable gaseous species. The second procedure, considered in order to reduce soot formation during the TPOX process, was to mix water vapor with the reactants. Previous studies have shown that the addition of small amounts of water vapor can influence, by
4317
Fig. 6 e Comparison of h evolution with and without the IPM in the cylindrical section of the TPOX reactor: (a) h as function of l for different Tin fixing P [ 5 kW; (b) h as function of P for different l fixing Tin [ 823 K.
thermal and chemical effects, the formation of soot particles in the ultra-rich premix regime, see, e.g., Ref. [32]. Fig. 10(a) and (b) presents the C2H2 molar fraction and Tin, respectively, as functions of the water vapor molar fraction present in the reactant mixture for two different inlet conditions. Fig. 10(a) shows that even small amounts of water can drastically reduce the C2H2 molar fraction in the TPOX products, and this is mainly justified by an increase of the OH radical concentration in the reaction front, which affects the process chemistry [32]. This finding can also be extrapolated to the soot formation due to reasons explained above. In Fig. 10(b), the Tin value was calculated in order to have the same outlet temperature obtained without water addition.
4318
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Fig. 7 e Gas phase temperature profiles for the TPOX reactor with and without the IPM in the cylindrical section operating at 2 kW and 7 kW (for Tin [ 823 K and l [ 0.45).
One can conclude that Tin must be increased in order to attain a constant outlet temperature. Fig. 10(c) shows that h is not negatively affected by the addition of water vapor to the reactants. However, one must note that equation (6), used to calculate h, does not account for the H2 introduced through the water and neglects the energy required for water vaporization. The third procedure consists in partially recirculating synthesis-gas from the outlet to the inlet of the TPOX reactor. This procedure influences the performance of the TPOX process mainly due to two effects: 1) modification/dilution of the reactant mixture composition; 2) heat recirculation from the hot products in order to preheat the reactant mixture.
Fig. 9 e Comparison of C2H2 molar fraction evolution with and without the IPM in the cylindrical section of the TPOX reactor: (a) C2H2 as function of l for different Tin fixing P to 5 kW; (b) C2H2 as function of P for different l fixing Tin to 823 K.
Fig. 8 e Axial velocity and residence time of the gas mixture inside the TPOX reactor with and without the IPM in the cylindrical section operating at 5 kW, Tin [ 823 K and l [ 0.36.
Simulations were carried out for a test case with the following conditions: P ¼ 5 kW, l ¼ 0.45 and Tin ¼ 823 K. The recirculation of TPOX products was quantified by introducing a recirculation factor (r) which is defined as the mass fraction of products recirculated to the reactor inlet. Fig. 11(a) presents the gas and solid temperature profiles within the TPOX reactor for several r values. One can observe from the figure that the increase of the recirculation factor, r, increases the mixture temperature at the inlet (Tmix) while slightly decreases both maximum and outlet temperatures. The increase on Tmix is mainly due to the heat recirculation from the hot products to the reactants. The dilution of the
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
4319
Fig. 10 e Effect of mixing water vapor with the reactants on the TPOX process performance: (a) C2H2 molar fraction evolution; (b) Tin evolution; (c) h evolution.
reactants by the recirculated products is responsible for the decrease on the maximum and outlet temperatures. The effect of the recirculation factor r on the concentration of main and minor gaseous species at the reactor outlet are shown in Fig. 11(b) and (c), respectively. It can be observed from Fig. 11(b) that the main species, particularly H2 and CO, are negligibly affected by the products recirculation. This is mainly related with the small influence of r on the postreaction temperature profiles (see Fig. 11(a)). However, Fig. 11(c) shows that the increase of the recirculation factor largely increase the CH4 and C2H2 concentrations at the reactor outlet. Therefore, it is expected that the soot formation will also increase with the recirculation factor. As observed in the previous section, the increase in Tin is always beneficial for the performance of the TPOX process. Therefore, taking also into account the results of TPOX products recirculation, one should note that the only direct effect of the products recirculation which benefits the TPOX process is the heat recirculation from the hot products to the reactants. This can be achieved by using a recuperative heat exchanger, avoiding the negative effects of reactants dilution. From the analysis of the three improving procedures presented above, one can conclude that the first two appear to be potentially effective. In the first procedure, the exclusion of the IPM from the downstream cylindrical section of the
reactor allows a higher radiative heat transfer from the reaction zone to the outlet, and therefore, radiative heat losses at the outlet must be prevented for this procedure to be effective. In the case of mixing water vapor with the reactants (second procedure), the energy for water vaporization may play an important role on the process efficiency, depending on its origin. Experimental investigations are required in order to explore the practical applicability of these improvement procedures.
4.
Conclusions
A quasi-1D combustion model was used in order to investigate possible procedures for maximizing the reforming efficiency and minimize the soot formation on the methane TPOX process within a small scale IPM based reactor. The reforming reactions were modeled by a detailed chemical kinetic mechanism, and the gas and solid phase energy balances were solved, including radiative heat transfer in the solid phase. From the comparison of the numerical predictions with experimental results, it was possible to validate the model and to conclude that this is able to capture the main features of the process. The numerical model was further used in order to perform a parametric study of the TPOX
4320
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Fig. 11 e Effect of synthesis-gas recirculation (quantified by r) from the outlet to the inlet of the TPOX reactor: (a) gas phase temperature profiles; (b) main species at the reactor outlet; (c) minor species at the reactor outlet.
reactor operation. From the parametric study the following conclusions can be drawn: 1. There is an optimal airefuel ratio which maximizes the reforming efficiency of the TPOX reactor, and it is dependent on the inlet mixture temperature. 2. High airefuel ratios minimize soot formation tendency during the TPOX process. 3. High inlet mixture temperatures are always beneficial in order to achieve both high reforming efficiency and low soot formation. 4. Power input has negligible direct effect on the reforming efficiency of the TPOX reactor. Taking into account the parametric study, the numerical model was used in order to explore possible procedures for improving the performance of the TPOX process within the small scale reactor. Three different procedures were investigated: removal of the IPM from the downstream cylindrical section of the TPOX reactor; mixing water vapor with the inlet reactants; and partially recirculating the TPOX products from the outlet to the inlet of the reactor. From the results of applying these procedures to the TPOX process, one can conclude that:
1. Removing the IPM from the downstream cylindrical section of the TPOX reactor improves the reforming efficiency and decreases the soot formation. 2. Mixing water vapor with the inlet reactants appears to be an effective solution for reducing the soot formation, without negatively affecting the reforming efficiency. 3. Recirculating TPOX products from the outlet to the inlet of the reactor increases soot formation and has negligible influence on the reforming efficiency. Therefore, the first two procedures appear to be potentially effective on improving the performance of the TPOX reactor operation, however, experimental investigations are required in order to conclude about the practical applicability of these procedures.
Acknowledgments The authors acknowledge the European Commission for the financial support of this work within the project FlameSOFC, contract no. 019875 (SES6). The first author also would like to acknowledge the financial support through fellowship from Fundac¸a˜o para a Cieˆncia e a Tecnologia e FCT.
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
references
[1] Kothari R, Buddhi D, Sawhney RL. Comparison of environmental and economic aspects of various hydrogen production methods. Renew Sustain Energy Rev 2008;12:553e63. [2] Fontell E, Kivisaari T, Christiansen N, Hansen J-B, Palsson J. Conceptual study of a 250 kW planar SOFC system for CHP application. J Power Sources 2004;131:49e56. [3] Santangelo PE, Tartarini P. Fuel cell systems and traditional technologies. Part I: experimental CHP approach. Appl Therm Eng 2007;27:1278e84. [4] Hawkes A, Staffell I, Brett D, Brandon N. Fuel cells for microcombined heat and power generation. Energy Environ Sci 2009;2:729e44. [5] Varbanov P, Klemes J, Shah RK, Shihn H. Power cycle integration and efficiency increase of molten carbonate fuel cell systems. Trans ASME J Fuel Cell Sci Technol 2006;3:375e83. paper No. FC-05-1079. [6] Korsgaard AR, Nielsen MP, Koer SK. Part one: a novel model of HTPEM-based micro-combined heat and power fuel cell system. Int J Hydrogen Energy 2008;33:1909e20. [7] Barnett SA. In: Vielstich W, Lamm A, Gasteiger H, editors. Handbook of fuel cell technology: IV. Fundamentals of technology and applications. San Francisco: Wiley; 2003. p. 1098. [8] Song X, Guo Z. Technologies for direct production of flexible H2/CO synthesis gas. Energy Convers Manag 2006;47:560e9. [9] Kumar R, Ahmed S, Krumpelt M. The low-temperature partial-oxidation reforming of fuels for transportation fuel cell systems. In: 1996 Fuel cell seminar abstracts, Orlando, FL 1996. pp. 750e3. [10] Tindall B, Crews M. Alternative technologies to steammethane reforming. Hydrocarb Process 1995;74:75e81. [11] Pedersen-Mjaanes H, Mastorakos E. Hydrogen production from rich combustion in porous media. Int J Hydrogen Energy 2005;30:579e92. [12] Al-Hamamre Z, Voss S, Trimis D. Hydrogen production by thermal partial oxidation of hydrocarbon fuels in porous media based reformer. Int J Hydrogen Energy 2009;34:827e32. [13] Howell JR, Hall MJ, Ellzey JL. Combustion of hydrocarbon fuels within porous inert media. Prog Energy Combust Sci 1996;22:121e45. [14] Wood S, Harris AT. Porous burners for lean-burn applications. Prog Energy Combust Sci 2008;34:667e84. [15] Abdul Mujeebu M, Abdullaha MZ, Abu Bakarb MZ, Mohamadc AA, Abdullaha MK. Combustion in porous media and its applications e a review. J Environ Manag 2009;86:1365e75. [16] Toledo M, Bubnovich V, Saveliev A, Kennedy L. Hydrogen production in ultrarich combustion of hydrocarbon fuels in porous media. Int J Hydrogen Energy 2009;34:1818e27.
4321
[17] Dhamrat RS, Ellzey JL. Numerical and experimental study of the conversion of methane to hydrogen in a porous media reactor. Combust Flame 2006;144:698e709. [18] Loukou A, Voss S, Mendes M, Raimondi A, Trimis D. Parametric experimental investigation of a small scale packed bed reactor for thermal partial oxidation. In: 4th European combustion meeting e proceedings CD 2009. Vienna, Austria, paper no. 810340. [19] Kee RJ, Grcar JF, Smooke MD, Miller JA. A Fortran program for modelling steady laminar one-dimensional premixed flames. Report SAND85-8240. Sandia National Laboratories; 1996. [20] Zhou XY, Pereira JCF. Numerical study of combustion and pollutants formation in inert nonhomogeneous porous media. Combust Sci Technol 1997;130:335e64. [21] Voss S, Mendes M, Pereira JC, Trimis D. Comparison of experimental and numerical results of ultra-lean H2/CO combustion within inert porous media. In: 4th European combustion meeting e proceedings CD 2009. Vienna, Austria, paper no. 810339. [22] VDI-Wa¨rmeatlas. 10th ed. Berlin: Springer-Verlag; 2006. p. Mh10. [23] Modest MF. Radiative heat transfer. New York: McGraw-Hill; 1993. [24] Mendes MAA, Pereira JMC, Pereira JCF. A numerical study of the stability of one-dimensional laminar premixed flames in inert porous media. Combust Flame 2008;153:525e39. [25] Lindstedt P, Maurice L, Meyer M. Thermodynamic and kinetic issues in the formation and oxidation of aromatic species. Faraday Discuss 2001;119:409e32. [26] Kee RJ, Rubley FM, Meeks E. Chemkin-II: a Fortran chemical kinetic package for the analysis of gas-phase chemical kinetics. Sandia National Laboratories; 1989. Report SAND89-8009. [27] Kee RJ, Dixon-Lewis G, Warnatz J, Coltrin ME, Miller JA. A Fortran computer code package for the evaluation of the gasphase transport properties. Sandia National Laboratories; 1986. Report SAND86-8246. [28] Munro RG. Evaluated material properties for a sintered alpha-Al2O3. J Am Ceram Soc 1997;80:1919e28. [29] Zheng C, Cheng L, Saveliev A, Luo Z, Cen K. Gas and solid phase temperature measurements of porous media combustion. Proc Combust Inst 2011;33:3301e8. [30] Xu F, Lin K-C, Faeth GM. Soot formation in laminar premixed methane/oxygen flames at atmospheric pressure. Combust Flame 1998;115:195e209. [31] 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;37:16686e96. [32] Skjoth-Rasmussen MS. Modelling of soot formation in autothermal reforming [PhD thesis]. Technical University of Denmark, Department of Chemical Engineering, CHEC; 2003, ISBN 87-90142-91-8; 2003.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Numerical study of methane TPOX within a small scale Inert Porous Media based reformer Miguel A.A. Mendes*, Jose´ M.C. Pereira, Jose´ C.F. Pereira Universidade de Lisboa, Instituto Superior Te´cnico, Mech. Eng. Dept., LASEF, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal
article info
abstract
Article history:
Methane Thermal Partial Oxidation (TPOX) within a small scale Inert Porous Media (IPM)
Received 11 September 2013
based reactor was investigated numerically in order to explore the operating conditions
Received in revised form
and possible procedures for maximizing the reforming efficiency and minimizing the soot
17 December 2013
formation. A quasi-1D model of the TPOX reactor was validated and further used to study
Accepted 28 December 2013
the process. The model considers detailed chemistry and solves the energy balances for
Available online 31 January 2014
both gas and solid phases, including radiative heat transfer in the solid phase. The parametric results of the reactor operation show that the optimal airefuel ratio is a compro-
Keywords:
mise between soot formation and reforming efficiency. Moreover, a high preheating
Hydrogen production
temperature of the reactants is found to be always beneficial for the process, and the effect
Methane reforming
of power input is negligible for the reforming efficiency. The numerical investigations also
Thermal partial oxidation
suggest that shorting the IPM length, as well as mixing small amounts of water vapor with
Inert Porous Media
the reactants, appear to be effective procedures for improving the operation performance
Combustion modeling
of the TPOX reactor. Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Power generation from fossil fuels has been widely considered as one of the main contributors for global warming, while at the same time the reserves of these non-renewable energy sources are rapidly depleting worldwide. Therefore, there is an increasing interest in developing power generation systems based on clean and sustainable energy sources. However, around 80% of the total energy consumption in the world is based yet on fossil fuel reserves. Consequently, in order to gradually convert the actual power generation systems from conventional to renewable, a smooth transition is required to meet the actual energy demand [1]. Emerging power generation technologies, such as SolidOxide Fuel Cells (SOFCs), are particularly attractive, when
compared to conventional technologies, due to their potential for efficient use of hydrocarbon fuels and reduced emissions. SOFCs present very high efficiencies at elevated working temperatures (500e1000 C) being therefore attractive for use in commercial and industrial Combined Heat and Power (CHP) applications [2e4]. Although other fuel cell types, such as, Molten-Carbonate Fuel Cells (MCFCs) or High-Temperature Polymer Electrolyte Membrane Fuel Cells (HT-PEMFC), also present a great potential for CHP [5,6]. In addition, SOFCs are fuel flexible and are able to operate either on pure hydrogen or hydrocarbon fuels. Therefore, systems based on SOFC technology have the potential to provide a transition solution from fossil fuels to more clean and sustainable energy carriers, such as bio-fuels and hydrogen. However, SOFC operation directly on hydrocarbon fuels is very limited and fuel reforming processes are generally required in order to convert hydrocarbon
* Corresponding author. E-mail addresses: [email protected], [email protected] (M.A.A. Mendes). 0360-3199/$ e see front matter Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.12.192
4312
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Nomenclature _ 00 m A av Cp dp hk Hv k MWk Nu Qr Qw Re T t u vk x Yk
axial mass flux, kg/m2 s reactor cross-sectional area, m2 surface area to volume ratio, m1 specific heat capacity at constant pressure, J/ kg K characteristic pore dimension, m enthalpy of k species, J/kg volumetric convective heat transfer coefficient, W/m3 K thermal conductivity, W/m K molar weight of k species, kg/kmol Nusselt number radiative heat flux, W/m2 wall heat loss per unit of length, W/m Reynolds number temperature, K time coordinate, s axial mean velocity of gas phase, m/s diffusion velocity of k species, m/s axial coordinate, m mass fraction of k species
Greek symbols b solid extinction coefficient, m1 u_ k production rate of k species, kmol/m3 s ˛ solid emissivity l mass ratio of air to fuel m dynamic viscosity, kg/m s u solid scattering albedo f porosity r mass density, kg/m3 s residence time, s Subscript g gas phase s solid phase
fuels into synthesis-gas (consisting mainly of H2 and CO) [7]. Owing to the carbon deposition and anode deactivation problems, external reforming is generally preferred over internal reforming, specially on practical small scale systems. There are three main reforming technologies used for synthesis-gas production from liquid and gaseous hydrocarbon fuels: steam reforming, catalytic or thermal partial oxidation and autothermal reforming [8]. Among these, Thermal Partial Oxidation (TPOX) offers several advantages [9,10], such as: no need for external heat sources and additional feeds; absence of catalysts which eliminates the catalyst deactivation problems; good dynamic response; and hydrocarbon flexibility. However, at the same time it presents comparatively low hydrogen yield and tendency to produce soot. Performing the TPOX within Inert Porous Media (IPM) based reformers is a practical solution to prevent flame stability problems, which are promoted by the slow reaction rates at low adiabatic flame temperatures existing in the TPOX process [11,12]. The high heat recirculation from the hot products to the reactants, provided by the solid matrix, originates local super-adiabatic combustion temperatures, which
increase the reaction rates and the stability of the TPOX process improving its operational characteristics [12]. Comprehensive reviews on IPM combustion can be found in Refs. [13e15]. Several recent studies have investigated TPOX of hydrocarbons within IPM, and different techniques have been successfully applied in order to stabilize the reforming process, see, e.g., Refs. [11,12,16e18]. Studies regarding transient TPOX within IPM can be found in Refs. [16,17]. These works applied the filtration combustion principle, which involves a traveling combustion wave freely propagating within a IPM, where the pore size of the solid matrix is typically sub-critical. Therefore, the IPM must be preheated to high temperatures in order to initialize the combustion wave propagation. On the other hand, stationary techniques for TPOX within IPM can be found in Refs. [11,12,18]. In these techniques, the combustion zone is steadily stabilized inside a porous matrix of supercritical pore size, and the nature of stabilization phenomenon is completely different from the one existing in transient approaches. In Ref. [18], experimental investigations of methaneeair TPOX process within a small scale IPM based reformer were performed for different ranges of inlet temperature, power input and airefuel ratio. The TPOX process was steadily stabilized by employing a reactor formed by a diffuser-like upstream section followed by a downstream cylindrical section, similar to the one presented in Ref. [12], and the IPM consisted of a packed bed of Al2O3 rings. Temperature profiles were measured along the reactor central axis, and concentrations of synthesis-gas species (H2, CO, CO2, CH4, C2H2) were measured at the reactor outlet. To the authors knowledge, there are no studies in the literature regarding the optimization of the TPOX process within IPM based reactors similar to the ones used in Refs. [12,18]. The objective of the present study is to investigate possible solutions in order to maximize the reforming efficiency and minimize the soot formation on the methane TPOX process within a small scale IPM based reactor. For this purpose a quasi-1D combustion model have been used, which was validated against the experimental results from Ref. [18]. The numerical model considers detailed chemical kinetics and solves the gas and solid phase energy balances, including radiative heat transport in the solid phase. With the help of this numerical model, a wide range of operating conditions as well as different procedures were explored in order to evaluate which conditions and procedures maximize the reforming efficiency and minimize the soot formation on this particular TPOX reactor. The present paper is organized in the following form: First, the numerical model of the methane TPOX reactor is described. Further, simulation results are presented, including model validation, parametric study, and different procedures for improving the process. Finally, the main conclusions are summarized.
2.
Numerical model
This section describes the numerical model used in order to simulate the methane TPOX process within a small scale IPM based reactor, experimentally tested in Ref. [18]. The TPOX
4313
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
reactor model was numerically implemented within an inhouse Fortran code, which is a modified version of the PREMIX code [19] that additionally incorporates the solid phase energy balance and the convective heat transfer between gas and solid phases. This numerical model was previously validated with excellent agreement against experimental results [20,21]. The geometrical configuration of the reactor is presented in Fig. 1, where the IPM consists of a packed bed of small Al2O3 rings. For modeling purposes, the IPM is assumed to be an homogeneous continuum with negligible catalytic effects. The ideal gas assumption is considered and the pressure loss along the reactor is neglected. A quasi-1D modeling approach is considered, which solves the 1D governing equations taking into account the cross-sectional area variation of the reactor by correcting the axial mass flux. With these assumptions, the averaged forms of the conservation equations for mass, gaseous species, and gas and solid phase energy balances are given by: Mass Conservation v Afrg _ 00 Þ vðAm ¼0 þ vx vt
(1)
where m00 ¼ frg u is the axial mass flux, Species Mass Fraction Balance v Afr v k Yk g vYk vY _ 00 k þ þ Am Afu_ k MWk ¼ 0 Afrg vt vx vx
(2)
Gas Phase Energy Balance
X vTg vT vTg vTg v _ 00 Cp;g g Afkg þ Am þ Af rg Cp;k vk vt vx vx vx vx k X u_ k hk MWk þ AHv Tg Ts ¼ 0 þAf (3)
Afrg Cp;g
k
Solid Phase Energy Balance vAQr v rs Cp;s Ts v vTs Að1fÞks AHv Tg Ts þ ¼0 Að1fÞ vx vt vx vx (4) Both gas and solid phase energy equations (3) and (4) are coupled by the convective heat transfer term AHv(Tg Ts), in which the volumetric convective heat transfer coefficient Hv is obtained from the following correlation [22]: dp Nu ¼ 2 1:3 þ 4:43 pffiffiffiffi þ 0:17Re0:75 A
INSULATION
POROUS MEDIA
Fig. 1 e Small scale IPM based reactor configuration.
(5)
where the Nusselt number Nu ¼ (Hvdp)/(kgav) and the Reynolds 00 number Re ¼ (m_ dp)/(mg) are based on the characteristic pore dimension dp of the IPM. For radiation purposes, the porous medium is assumed as a diffuse, gray medium together with a non-participative gas mixture. The term v(AQr)/vx appearing in equation (4) represents the radiative heat transport in the solid phase and was numerically solved using the Discrete-Ordinates method (S8 approximation) [23]. The boundary conditions applied to equations (1)e(4) and to the radiative heat transport equations are similar to the ones presented in Ref. [24]. The methaneeair TPOX reactions were modeled on the basis of the chemical kinetic mechanism from Refs. [25], where only the C1C2 chemistry was considered, consisting of 32 species and 150 elementary reactions. The thermo-physical properties of the gas mixture (rg, kg, mg, vk and Cp,g), required by equations (1)e(3), were obtained based on [26,27]. Moreover, the thermo-physical properties (rs, ks and Cp,s) of the Al2O3 ceramic material, required by equation (4), were obtained from Ref. [28] as functions of temperature. The values of the IMP properties used for the simulations are resumed in table 1.
3.
Results
This section presents the numerical predictions of methane TPOX process within IPM obtained with the reactor model described in the previous section. First, the model validation results are resumed, followed by the results of the parametric study of the reactor operation. Finally, the results regarding different procedures for improving the TPOX process are presented.
3.1.
Model validation
In order to validate the model, numerical predictions of the methane TPOX process within a small scale IPM based reactor were compared against the respective experimental results, obtained from Ref. [18]. This comparison was performed for ranges of airefuel ratio (l), inlet gas mixture temperature (Tin), and power input (P). Moreover, reactor heat losses were accounted for by the model for this comparison, i.e., radiative heat losses from the reactor outlet to the downstream environment at 1250 K, as well as heat losses from the reactor side walls to the surroundings at 300 K. The wall heat losses were modeled by including the term Qw in the left-hand side of the solid energy balance, given by equation (4). This term was assumed to be uniform and equal to 0.9 kW/m (based on the experiments), in a similar way to [20].
Table 1 e IPM properties used for the simulations. Porosity, f Pore dimension, dp [m] Specific surface, av [m1] Extinction coefficient, b [m1] Scattering albedo, u Emissivity, ˛
0.63 0.015 320 100 0.8 0.4
4314
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Fig. 2(a) presents the simulated gas and solid temperature profiles together with the experimental temperatures, measured along the reactor center line, for several l and fixing Tin ¼ 823 K and P ¼ 5 kW. Overall temperature measurements are well predicted by the model, particularly the temperatures at the post-flame region. The figure also shows that the position of the flame front is reasonably well captured by the model, however the predicted temperature peak of the gas phase is not clearly observed in the measurements. This can be explained by the fact that the thermocouples inserted in the reactor measured a temperature value that is neither the one of the gas nor solid phase but something in between, see, e.g., Ref. [29]. Moreover, it is expected that the thermocouples strongly felt the presence of the solid due to its high radiative heat transfer and also due to possible contact between the solid matrix and the thermocouples. Nevertheless, the temperature of the porous matrix is lower at the upstream region of the reactor and the higher flow velocities, induced by the smaller cross-sectional area of the reactor, increase the convective heat transfer between the gas phase and thermocouples. Therefore, in this region, the measured temperatures are expected to be closer to the ones of the gas mixture. The comparison between numerical predictions and measurements of the molar fraction of main synthesis-gas species (H2, CO, H2O and CO2) at the reactor outlet are shown in Fig. 2(b) as function of l for Tin ¼ 823 K and P ¼ 5 kW. Similar results are presented in Fig. 2(c) for the minor
synthesis-gas species (CH4 and C2H2). One can observe from Fig. 2(b) that the numerical predictions of the main species are in good agreement with the measurements. Fig. 2(c) shows a good qualitative trend of the minor species prediction. Nevertheless, the quantitative values are far from perfect, and this may be attributed to the soot formation, which is not taken into account by the current model, see, e.g., Ref. [30]. In Ref. [31], the same TPOX reactor considered for the present study was tested for different operation conditions (l, Tin and P) and soot measurements were obtained at the exit of the reactor. They found that for higher values of l, e.g., l ¼ 0.5, no soot was detected. However, Fig. 2(a) shows that the agreement between predicted and measured temperature profiles is not significantly influenced by l. Therefore, it can be concluded that the soot formation, which according to Ref. [31] occurs for lower values of l, has a negligible effect on the temperatures inside the TPOX reactor and hence, the same is expected for the temperatures predicted by the present model. Nevertheless, the soot formation influences the minor species, as pointed above. The over-prediction of the measured C2H2 molar fractions, observed in Fig. 2(c) for lower values of l, can be explained by the fact that C2H2 participates in the chemical pathways that lead to molecular grow of aromatic compounds and formation of soot particles. From the validation results, one can conclude that this simple quasi-1D model is able to capture the main features of the methane TPOX process within the small scale IPM based
Fig. 2 e Comparison between simulations and experimental results for the TPOX reactor operating at several l, fixing P [ 5 kW and Tin [ 823 K: (a) temperature profiles; (b) outlet main species; (c) outlet minor species.
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
4315
reactor tested in Ref. [18]. Therefore, further investigations with this model are expected to be realistic. However, for the sake of simplicity and objectivity, the following results are obtained assuming that the TPOX reactor is adiabatic, i.e., wall heat losses and radiative heat losses at the reactor outlet are neglected.
3.2.
Parametric study
Before presenting the parametric study results, the methane TPOX process within IPM was compared against free flame results in order to understand the main differences between both processes. This comparison was carried out assuming Tin ¼ 823 K. Fig. 3(a) and (b) compares the model predictions for the evolution of the reforming efficiency (h) and the gas temperature profiles, respectively, of the TPOX process for IPM and free flame modes as function of the residence time. The reforming efficiency is defined as the ratio between the low heating value of the gas mixture produced from the reforming process to the low heating value of the original fuel (CH4), and given by: h¼
YH2 LHVH2 þ YCO LHVCO YCH4 LHVCH4
(6)
Fig. 3(a) shows that the presence of IPM accelerates the initial stage of the TPOX process due to the super-adiabatic temperatures induced by the heat recirculation from the hot products to the reactants by both conduction and radiation heat transfer. The gas temperature profiles presented in Fig. 3(b) confirm this efficiency improvement. However, the heat recirculation excessively decreases the products temperature in the post-flame region slowing down the TPOX process reactions. Therefore, after a certain residence time, the free flame mode shows a higher h than the IPM mode. This phenomenon is more evident for lower l values, since at these ultra-rich conditions the TPOX process requires higher residence times to reach equilibrium. A parametric study was conducted using the original TPOX reactor configuration, shown in Fig. 1, operating at different l, Tin and P, keeping as reference the nominal operating conditions: Tin ¼ 823 K and P ¼ 5 kW. The results presented in this section consider the influence of the residence time in an implicit manner through l, Tin and P, which directly influence the value of the mass flow rate of the gas mixture. However, an explicit influence of the residence on reforming efficiency or soot formation is not considered, because the residence time is not an independent input parameter to the TPOX reactor operation, but l, Tin and P are. Fig. 4(a) shows h as function of l for different values of Tin (fixing P ¼ 5 kW), along with the equilibrium reforming efficiency and the maximum theoretical reforming efficiency. The maximum theoretical reforming efficiency is calculated assuming that all CH4 is converted to H2 and CO. One can observe from the figure that, for high values of l, the h of the process is very close to the equilibrium efficiency and is little sensitive to Tin. However, for l lower than a critical value, h departs from the equilibrium efficiency, and the departure starts at higher l values if Tin is lower. This is mainly related to the fact that the residence time required to achieve
Fig. 3 e Comparison of the methane TPOX process, predicted by the model, performed within IPM and in the free flame mode as function of the residence time for two different l: (a) h evolution; (b) gas temperature evolution.
equilibrium increases with the decrease in l and Tin due to the extremely slow reaction rates at low adiabatic temperatures existing at these conditions [12]. Therefore, there exists an optimal l that maximizes h. In order to shift the optimal l to lower values and to increase h to equilibrium values, one would require to increase the residence time of the TPOX process. Fig. 4(b) presents the reforming efficiency as function of P for several l values, fixing Tin ¼ 823 K. It can be observed from the figure that h is nearly independent of P. However, this will not be the case if the reactor heat losses are not negligible, because the resulting lower temperatures inside the reactor would decrease the reforming efficiency, especially at lower P.
4316
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Fig. 4 e Evolution of h: (a) h as function of l for different Tin fixing P [ 5 kW; (b) h as function of P for different l fixing Tin [ 823 K.
Fig. 5 e Evolution of C2H2 molar fraction: (a) C2H2 as function of l for different Tin fixing P [ 5 kW; (b) C2H2 as function of P for different l fixing Tin [ 823 K.
Although the main interest in the TPOX process is to maximize h, the formation of soot particles also plays an important role and it must be minimized. The lack of oxygen and the lower reaction temperatures existing at lower l values, promote soot formation due to the increase of intermediate hydrocarbon species that are the precursors of soot particles. The C2H2 species is the main indicator for the presence of soot due to its deep involvement in the critical steps of soot formation, i.e., inception and surface growth [30]. Therefore, it is reasonable to use the C2H2 concentration in the synthesis-gas as a rough quantifier of soot produced. Fig. 5(a) shows the C2H2 molar fraction as function of l for different values of Tin, fixing P ¼ 5 kW. It can be observed from the figure that the C2H2 increases with the decreasing on both l and Tin, and this increase is rather more drastic due to l than
due to Tin. Therefore, a similar behavior should be expected for the soot produced. Fig. 5(b) presents the C2H2 molar fraction as function of P for different values of l, fixing Tin ¼ 823 K. One can observe from the figure that the C2H2 fraction slightly increases with P, mainly due to the decrease of the residence time, which does not allow a complete conversion of C2H2 into more stable gaseous products. Consequently, this should not be totally interpreted as an increase of soot production, because its formation process also requires some time.
3.3.
Process improvement procedures
Taking into account the previous results for the parametric study, three different procedures were investigated in order to
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
improve the performance of the TPOX process within the small scale reactor. The first procedure is based on modifying the original reactor geometry by excluding the IPM from the downstream cylindrical section (see Fig. 1). The second procedure is focussed on changing the inlet composition by mixing water vapor to the reactants. A third procedure explored the effect of partially recirculating the TPOX products from the outlet to inlet of the reactor. In first place, it was investigated the effect of excluding the IPM from the downstream cylindrical section of the TPOX reactor presented in Fig. 1. The IPM improves the heat recirculation from the hot products to the reactants and, therefore, increases the temperature in the flame front region promoting flame stabilization. However, the high heat recirculation reduces the temperature of the products in the post-combustion zone, strongly decelerating the rates of the slower reforming reactions taking place there. Consequently, by excluding the IPM from the downstream cylindrical section (where it is not relevant for the flame stabilization), one will reduce the heat recirculation from the post-flame hot products to the reactants, keeping these products at a higher temperature. This will also enhance the free space volume and increase the residence time of the gas mixture on the cylindrical region of the reformer. Fig. 6(a) and (b) compares the h calculated using the TPOX reactor with and without the IPM in the cylindrical section (see Fig. 1). Fig. 6(a) presents h as function of l for different Tin values, fixing P ¼ 5 kW, and Fig. 6(b) shows the same information as function of P for several l values, fixing Tin ¼ 823 K. This simple modification in the reactor geometry improves h, especially for lower l and Tin, see Fig. 6(a). Fig. 6(b) shows that h was improved for higher P. However, this efficiency improvement is smaller at lower power operation, because for these cases the reaction front is located deep inside the conical section, where the heat recirculation provided by the IPM is still able to strongly decrease the temperature of the post-flame products. Fig. 7 presents the gas phase temperature profiles for both reformer versions operating at 2 kW and 7 kW, where it is notorious the effect of excluding the IPM from the cylindrical section of the TPOX reactor. Moreover, Fig. 8 presents the axial velocity u and residence time s of the gas mixture inside the TPOX reactor with and without IPM in the cylindrical section, for a particular operating condition: Tin ¼ 823 K, P ¼ 5 kW and l ¼ 0.36. One can observe from the figure that the exclusion of the IPM from the cylindrical section of the reactor increases the residence time around 1.5 times, which is a consequence of the lower axial velocity. Fig. 9(a) and (b) shows the C2H2 molar fraction calculated as function of l and P, respectively, for the TPOX reactor with and without IPM in the cylindrical section. By removing the IPM from the cylindrical section, the C2H2 concentration and consequently the soot formation decrease for high P and lower l. This is justified by the extended residence time and higher temperatures within the cylindrical section of the TPOX reactor, which allow a more complete conversion of C2H2 into more stable gaseous species. The second procedure, considered in order to reduce soot formation during the TPOX process, was to mix water vapor with the reactants. Previous studies have shown that the addition of small amounts of water vapor can influence, by
4317
Fig. 6 e Comparison of h evolution with and without the IPM in the cylindrical section of the TPOX reactor: (a) h as function of l for different Tin fixing P [ 5 kW; (b) h as function of P for different l fixing Tin [ 823 K.
thermal and chemical effects, the formation of soot particles in the ultra-rich premix regime, see, e.g., Ref. [32]. Fig. 10(a) and (b) presents the C2H2 molar fraction and Tin, respectively, as functions of the water vapor molar fraction present in the reactant mixture for two different inlet conditions. Fig. 10(a) shows that even small amounts of water can drastically reduce the C2H2 molar fraction in the TPOX products, and this is mainly justified by an increase of the OH radical concentration in the reaction front, which affects the process chemistry [32]. This finding can also be extrapolated to the soot formation due to reasons explained above. In Fig. 10(b), the Tin value was calculated in order to have the same outlet temperature obtained without water addition.
4318
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Fig. 7 e Gas phase temperature profiles for the TPOX reactor with and without the IPM in the cylindrical section operating at 2 kW and 7 kW (for Tin [ 823 K and l [ 0.45).
One can conclude that Tin must be increased in order to attain a constant outlet temperature. Fig. 10(c) shows that h is not negatively affected by the addition of water vapor to the reactants. However, one must note that equation (6), used to calculate h, does not account for the H2 introduced through the water and neglects the energy required for water vaporization. The third procedure consists in partially recirculating synthesis-gas from the outlet to the inlet of the TPOX reactor. This procedure influences the performance of the TPOX process mainly due to two effects: 1) modification/dilution of the reactant mixture composition; 2) heat recirculation from the hot products in order to preheat the reactant mixture.
Fig. 9 e Comparison of C2H2 molar fraction evolution with and without the IPM in the cylindrical section of the TPOX reactor: (a) C2H2 as function of l for different Tin fixing P to 5 kW; (b) C2H2 as function of P for different l fixing Tin to 823 K.
Fig. 8 e Axial velocity and residence time of the gas mixture inside the TPOX reactor with and without the IPM in the cylindrical section operating at 5 kW, Tin [ 823 K and l [ 0.36.
Simulations were carried out for a test case with the following conditions: P ¼ 5 kW, l ¼ 0.45 and Tin ¼ 823 K. The recirculation of TPOX products was quantified by introducing a recirculation factor (r) which is defined as the mass fraction of products recirculated to the reactor inlet. Fig. 11(a) presents the gas and solid temperature profiles within the TPOX reactor for several r values. One can observe from the figure that the increase of the recirculation factor, r, increases the mixture temperature at the inlet (Tmix) while slightly decreases both maximum and outlet temperatures. The increase on Tmix is mainly due to the heat recirculation from the hot products to the reactants. The dilution of the
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
4319
Fig. 10 e Effect of mixing water vapor with the reactants on the TPOX process performance: (a) C2H2 molar fraction evolution; (b) Tin evolution; (c) h evolution.
reactants by the recirculated products is responsible for the decrease on the maximum and outlet temperatures. The effect of the recirculation factor r on the concentration of main and minor gaseous species at the reactor outlet are shown in Fig. 11(b) and (c), respectively. It can be observed from Fig. 11(b) that the main species, particularly H2 and CO, are negligibly affected by the products recirculation. This is mainly related with the small influence of r on the postreaction temperature profiles (see Fig. 11(a)). However, Fig. 11(c) shows that the increase of the recirculation factor largely increase the CH4 and C2H2 concentrations at the reactor outlet. Therefore, it is expected that the soot formation will also increase with the recirculation factor. As observed in the previous section, the increase in Tin is always beneficial for the performance of the TPOX process. Therefore, taking also into account the results of TPOX products recirculation, one should note that the only direct effect of the products recirculation which benefits the TPOX process is the heat recirculation from the hot products to the reactants. This can be achieved by using a recuperative heat exchanger, avoiding the negative effects of reactants dilution. From the analysis of the three improving procedures presented above, one can conclude that the first two appear to be potentially effective. In the first procedure, the exclusion of the IPM from the downstream cylindrical section of the
reactor allows a higher radiative heat transfer from the reaction zone to the outlet, and therefore, radiative heat losses at the outlet must be prevented for this procedure to be effective. In the case of mixing water vapor with the reactants (second procedure), the energy for water vaporization may play an important role on the process efficiency, depending on its origin. Experimental investigations are required in order to explore the practical applicability of these improvement procedures.
4.
Conclusions
A quasi-1D combustion model was used in order to investigate possible procedures for maximizing the reforming efficiency and minimize the soot formation on the methane TPOX process within a small scale IPM based reactor. The reforming reactions were modeled by a detailed chemical kinetic mechanism, and the gas and solid phase energy balances were solved, including radiative heat transfer in the solid phase. From the comparison of the numerical predictions with experimental results, it was possible to validate the model and to conclude that this is able to capture the main features of the process. The numerical model was further used in order to perform a parametric study of the TPOX
4320
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
Fig. 11 e Effect of synthesis-gas recirculation (quantified by r) from the outlet to the inlet of the TPOX reactor: (a) gas phase temperature profiles; (b) main species at the reactor outlet; (c) minor species at the reactor outlet.
reactor operation. From the parametric study the following conclusions can be drawn: 1. There is an optimal airefuel ratio which maximizes the reforming efficiency of the TPOX reactor, and it is dependent on the inlet mixture temperature. 2. High airefuel ratios minimize soot formation tendency during the TPOX process. 3. High inlet mixture temperatures are always beneficial in order to achieve both high reforming efficiency and low soot formation. 4. Power input has negligible direct effect on the reforming efficiency of the TPOX reactor. Taking into account the parametric study, the numerical model was used in order to explore possible procedures for improving the performance of the TPOX process within the small scale reactor. Three different procedures were investigated: removal of the IPM from the downstream cylindrical section of the TPOX reactor; mixing water vapor with the inlet reactants; and partially recirculating the TPOX products from the outlet to the inlet of the reactor. From the results of applying these procedures to the TPOX process, one can conclude that:
1. Removing the IPM from the downstream cylindrical section of the TPOX reactor improves the reforming efficiency and decreases the soot formation. 2. Mixing water vapor with the inlet reactants appears to be an effective solution for reducing the soot formation, without negatively affecting the reforming efficiency. 3. Recirculating TPOX products from the outlet to the inlet of the reactor increases soot formation and has negligible influence on the reforming efficiency. Therefore, the first two procedures appear to be potentially effective on improving the performance of the TPOX reactor operation, however, experimental investigations are required in order to conclude about the practical applicability of these procedures.
Acknowledgments The authors acknowledge the European Commission for the financial support of this work within the project FlameSOFC, contract no. 019875 (SES6). The first author also would like to acknowledge the financial support through fellowship from Fundac¸a˜o para a Cieˆncia e a Tecnologia e FCT.
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 3 9 ( 2 0 1 4 ) 4 3 1 1 e4 3 2 1
references
[1] Kothari R, Buddhi D, Sawhney RL. Comparison of environmental and economic aspects of various hydrogen production methods. Renew Sustain Energy Rev 2008;12:553e63. [2] Fontell E, Kivisaari T, Christiansen N, Hansen J-B, Palsson J. Conceptual study of a 250 kW planar SOFC system for CHP application. J Power Sources 2004;131:49e56. [3] Santangelo PE, Tartarini P. Fuel cell systems and traditional technologies. Part I: experimental CHP approach. Appl Therm Eng 2007;27:1278e84. [4] Hawkes A, Staffell I, Brett D, Brandon N. Fuel cells for microcombined heat and power generation. Energy Environ Sci 2009;2:729e44. [5] Varbanov P, Klemes J, Shah RK, Shihn H. Power cycle integration and efficiency increase of molten carbonate fuel cell systems. Trans ASME J Fuel Cell Sci Technol 2006;3:375e83. paper No. FC-05-1079. [6] Korsgaard AR, Nielsen MP, Koer SK. Part one: a novel model of HTPEM-based micro-combined heat and power fuel cell system. Int J Hydrogen Energy 2008;33:1909e20. [7] Barnett SA. In: Vielstich W, Lamm A, Gasteiger H, editors. Handbook of fuel cell technology: IV. Fundamentals of technology and applications. San Francisco: Wiley; 2003. p. 1098. [8] Song X, Guo Z. Technologies for direct production of flexible H2/CO synthesis gas. Energy Convers Manag 2006;47:560e9. [9] Kumar R, Ahmed S, Krumpelt M. The low-temperature partial-oxidation reforming of fuels for transportation fuel cell systems. In: 1996 Fuel cell seminar abstracts, Orlando, FL 1996. pp. 750e3. [10] Tindall B, Crews M. Alternative technologies to steammethane reforming. Hydrocarb Process 1995;74:75e81. [11] Pedersen-Mjaanes H, Mastorakos E. Hydrogen production from rich combustion in porous media. Int J Hydrogen Energy 2005;30:579e92. [12] Al-Hamamre Z, Voss S, Trimis D. Hydrogen production by thermal partial oxidation of hydrocarbon fuels in porous media based reformer. Int J Hydrogen Energy 2009;34:827e32. [13] Howell JR, Hall MJ, Ellzey JL. Combustion of hydrocarbon fuels within porous inert media. Prog Energy Combust Sci 1996;22:121e45. [14] Wood S, Harris AT. Porous burners for lean-burn applications. Prog Energy Combust Sci 2008;34:667e84. [15] Abdul Mujeebu M, Abdullaha MZ, Abu Bakarb MZ, Mohamadc AA, Abdullaha MK. Combustion in porous media and its applications e a review. J Environ Manag 2009;86:1365e75. [16] Toledo M, Bubnovich V, Saveliev A, Kennedy L. Hydrogen production in ultrarich combustion of hydrocarbon fuels in porous media. Int J Hydrogen Energy 2009;34:1818e27.
4321
[17] Dhamrat RS, Ellzey JL. Numerical and experimental study of the conversion of methane to hydrogen in a porous media reactor. Combust Flame 2006;144:698e709. [18] Loukou A, Voss S, Mendes M, Raimondi A, Trimis D. Parametric experimental investigation of a small scale packed bed reactor for thermal partial oxidation. In: 4th European combustion meeting e proceedings CD 2009. Vienna, Austria, paper no. 810340. [19] Kee RJ, Grcar JF, Smooke MD, Miller JA. A Fortran program for modelling steady laminar one-dimensional premixed flames. Report SAND85-8240. Sandia National Laboratories; 1996. [20] Zhou XY, Pereira JCF. Numerical study of combustion and pollutants formation in inert nonhomogeneous porous media. Combust Sci Technol 1997;130:335e64. [21] Voss S, Mendes M, Pereira JC, Trimis D. Comparison of experimental and numerical results of ultra-lean H2/CO combustion within inert porous media. In: 4th European combustion meeting e proceedings CD 2009. Vienna, Austria, paper no. 810339. [22] VDI-Wa¨rmeatlas. 10th ed. Berlin: Springer-Verlag; 2006. p. Mh10. [23] Modest MF. Radiative heat transfer. New York: McGraw-Hill; 1993. [24] Mendes MAA, Pereira JMC, Pereira JCF. A numerical study of the stability of one-dimensional laminar premixed flames in inert porous media. Combust Flame 2008;153:525e39. [25] Lindstedt P, Maurice L, Meyer M. Thermodynamic and kinetic issues in the formation and oxidation of aromatic species. Faraday Discuss 2001;119:409e32. [26] Kee RJ, Rubley FM, Meeks E. Chemkin-II: a Fortran chemical kinetic package for the analysis of gas-phase chemical kinetics. Sandia National Laboratories; 1989. Report SAND89-8009. [27] Kee RJ, Dixon-Lewis G, Warnatz J, Coltrin ME, Miller JA. A Fortran computer code package for the evaluation of the gasphase transport properties. Sandia National Laboratories; 1986. Report SAND86-8246. [28] Munro RG. Evaluated material properties for a sintered alpha-Al2O3. J Am Ceram Soc 1997;80:1919e28. [29] Zheng C, Cheng L, Saveliev A, Luo Z, Cen K. Gas and solid phase temperature measurements of porous media combustion. Proc Combust Inst 2011;33:3301e8. [30] Xu F, Lin K-C, Faeth GM. Soot formation in laminar premixed methane/oxygen flames at atmospheric pressure. Combust Flame 1998;115:195e209. [31] 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;37:16686e96. [32] Skjoth-Rasmussen MS. Modelling of soot formation in autothermal reforming [PhD thesis]. Technical University of Denmark, Department of Chemical Engineering, CHEC; 2003, ISBN 87-90142-91-8; 2003.