Patent Application
20150010874
The present invention relates generally to a miniaturized power generation device, and more particularly, but not exclusively, to a microscale combustor and heat exchanger (μCHX) that may include several repeating unit cells each of which performs combustion, recuperation, and heat exchange.
Advances in fabrication technology have helped spur research in miniaturized thermal devices for distributed power generation and heat and mass transport process intensification. These devices have the advantages of being lightweight and compact, thereby yielding high energy and power densities. With a decrease in channel dimensions, the surface area per unit volume of fluid flowing in a microchannel increases thereby increasing the heat transfer rate. The increased surface area per unit volume also enhances completion of heterogeneous surface chemical reactions within microchannels. Therefore, catalytic combustion and heat transfer in microchannels can result in increased efficiency and reduced size of thermal power generation and exchange devices. In addition, the rate of heat generation from combustion and device temperatures can be controlled to some extent by tailoring heterogeneous reactions to occur at specific locations within the channel walls. Thus, high temperatures that cause the formation of NOX can potentially be mitigated. Because of these advantages, several microscale combustor geometries for different fuels have been presented in literature over the past decade. As the combustion channel size decreases, homogeneous (gas-phase) reactions face thermal and radical quenching. Depending on the heat loss and the gas mixture velocity, flame extinction of gas phase hydrogen combustion can occur in channels as large as 1000 μm. Combustion in channels smaller than this limit can occur via heterogeneous (surface) reactions that are promoted by catalysts. Heterogeneous combustion can further trigger homogeneous combustion in the bulk of the fluid.
Combustion within microchannels has been documented in several numerical and experimental studies. For example, Boyarko et al. tested hydrogen-oxygen mixture combustion in a platinum microtube and found that the 400 μm and 800 μm tubes used in their experiments were below quenching size under most atmospheric pressure test conditions. (Boyarko, G. A., Sung, C. J., and Schneider, S. J., 2005, “Catalyzed combustion of hydrogen-oxygen in platinum tubes for micro-propulsion applications,” Proc. Combust. Inst., 30, pp. 2481-2488.) In both numerical simulations as well as experiments, Boyarko et al. observed that there was a minimum threshold heat flux necessary for ignition. When the ignition heat flux was increased further, the gas temperature got so high that a choked flow resulted inside the tube. Zhou et al. modeled conjugate heat transfer within a 500 μm channel made of different wall materials (quartz glass, alumina ceramic, copper) to investigate the effect of wall thermal conductivity on homogeneous/heterogeneous combustion of hydrogen-air mixture. (Zhou, J., Wang, Y., Yang, W., Liu, J., Wang, Z., and Cen, K., 2009, “Combustion of hydrogen-air in catalytic micro-combustors made of different material,” Int. J. Hydrog. Energy, 34, pp. 3535-3545.) They observed that heterogeneous reaction became dominant as thermal conductivity of the material increased. Chen et al., who simulated combustion in a 1-mm channel with different wall materials, reached a similar conclusion. (Chen, G. B., Chen, C. P., Wu, C. Y., and Chao, Y. C., 2007, “Effects of catalytic walls on hydrogen/air combustion inside a micro-tube,” Appl. Catal. A-Gen., 332, pp. 89-97.) They found that heterogeneous reaction was dominant in the beginning of the tube, followed by homogeneous reaction downstream of the tube. Lower wall conductivity was observed to lead to a larger temperature gradient on the surface causing homogeneous combustion to shift upstream. At the highest studied velocity (20 m/s), no homogeneous reaction was observed in channel heights lower than 200 μm in diameter. In a follow-on study, Chen et al. investigated the effect of differences in catalyst configurations in the same geometry as that used in their previous study. (Chen, G. B., Chao, Y. C., and Chen, C. P., 2008, “Enhancement of hydrogen reaction in a micro-channel by catalyst segmentation,” Int. J. Hydrog. Energy, 33, pp. 2586-2595.) A multi-segment catalyst was compared with a single segment catalyst of the same total length. The multi-segment catalyst configuration showed better conversion due to the occurrence of homogeneous reaction in the regions between segments. Karagiannidis and Mantzaras used a 2-D model to simulate transient hetero-homogeneous combustion of methane over platinum catalyst within a 1000 micrometer channel. (Karagiannidis, S., and Mantzaras, J., 2010, “Numerical investigation on the start-up of methane-fueled catalytic microreactors,” Combust. Flame, 157, pp. 1400-1413.) For the pressures in the range of 1 bar-5 bar, they found that ignition and steady state microreactor residence times decreased with an increase in pressure. Combustors with lower thermal conductivity walls had smaller ignition times.
Recuperation has been used alongside combustion in order to preheat the gas mixture by several groups. Lloyd and Weinberg fabricated a spiral counterflow combustor, often referred to as a “Swiss roll” type combustor to improve the efficiency of combustion processes. (Lloyd, S. A., and Weinberg, F. J., 1974, “A burner for mixtures of very low heat content,” Nature, 251 (5470), pp. 47-49.) Peterson et al. developed a microscale hydrogen combustor with counterflow heat recuperator. (Peterson, R. B., and Vanderhoff, J. A., 2000, “A catalytic combustor for microscale applications,” Combust. Sci. Technol. Comm., 1, pp. 10-13.) They observed that preheating helped to keep a sustained homogeneous reaction. In addition to a microscale combustor, an efficient heat exchanger is required in order to transfer heat produced by the reaction to a working fluid. The heat transfer to the working fluid will alter the wall temperature distribution, which will in turn affect the combustion process. Janicke et al. used hydrogen combustion over a platinum covered surface to heat a gas stream cross-flow to the combustion gas flow in a microscale heat exchanger. (Janicke, M. T., Kestenbaum, H., Hagendorf, U., Schüth, F., Fichtner, M., and Schubert, K., 2000, “The controlled oxidation of hydrogen from an explosive mixture of gases using a microstructured reactor/heat exchanger and Pt/Al2O3 catalyst,” J. Catal., 191 (2), pp. 282-293.)
There have been several studies involving stacked microchannel arrays for various applications. The review articles by Fan and Luo and Khan and Fartaj provide some of the recent examples of stacked microchannel devices including heat exchangers and chemical reactors. (Fan, Y., and Luo L., 2008, “Recent applications of advances in microchannel heat exchangers and multi-scale design optimization,” Heat Transf. Eng., 29(5), pp. 461-474. Khan, M. G., and Fartaj, A., 2011, “A review on microchannel heat exchangers and potential applications”, Int. J. Energy Res., 35, pp. 553-582.) In stacked microscale reactors, one layer could have several parallel microchannels wherein a reaction occurs while exchanging heat with a working fluid that flows in an adjacent layer. Such an arrangement has been used for methane steam reforming where a fuel combusts in the combustor layers and transfers the produced heat to the reformer sheets. Ryi et al. tested methane steam reforming with hydrogen catalytic combustion in an integrated microchannel reactor. (Ryi, S. K., Park, J. S., Choi, S. H., Cho, S. H., and Kim S. H., 2005, “Novel micro fuel processor for PEMFCs with heat generation by catalytic combustion,” Chem. Eng. J., 113, pp. 47-53.) The designed device consisted of cover plate, a base plate and 50 plates (25 alternating combustor and reformer plates) with microchannels. Inconel plates were used to fabricate the microchannel sheets and stainless steel sheets were used for the cover and base plates. Each sheet had 22 microchannels in parallel with 500 μm in diameter, 250 μm in depth and 17 mm in length. Pt—Sn/Al2O3 and Rh—Mg/Al2O3 were impregnated by wash-coating in the combustor and reformer for catalytic reactions respectively. Hwang et al. developed a similar combined combustor and methane reformer device and were able to achieve 95% conversions and hydrogen production rate of 0.78 mol/h in the reformer. (Hwang, K. R., Lee, C. B., Lee, S. W., Ryi, S. K., and Park, J. S., 2011, “Novel micro-channel methane reformer assisted combustion reaction for hydrogen production,” Int. J. Hydrog. Energy, 36, pp. 473-481.) Their device consisted of a variety of chemically etched metal plates, such as half-etched straight channel plates (10 sheets), fully etched 3D mixing channel plates (2 sheets), and cover/holder/separator plates (5 sheets). Hydrogen and/or methane were used as the fuel in the combustor sheets to provide heat for methane reformation. A Pt-coated mesh catalyst was used as an igniter at the inlet of the combustor until a flame was generated.
Mettler et al. used CFD simulations to model stacks of different sizes and characterize the effects of scaling up of microchemical systems. (Mettler, M. S., Stefanidis, G. D., and Vlachos, D. G., 2011, “Enhancing stability in parallel plate microreactor stacks for syngas production,” Chem. Eng. Sci., 66, pp. 1051-1059.) They studied syngas production from methane using a parallel-plate reactor with alternating combustion and steam reforming channels. The author compared stacks of 3 units to 15 units, each comprised a combustion channel and reformer channel. They found that heat losses caused extinction of combustion in the outer channels and consequently reduced the efficiency of the smaller stack. Whereas extinction of combustion occurred in the outer channels even for the larger stack, the interior channels sustained combustion, resulting in a higher efficiency. They also recommended stack materials with thermal conductivities higher than 100 W/m-K for a more stable device.
Very recently Zhang et al. synthesized a Pt-based catalyst, and investigated the behavior of hydrogen catalytic combustion at low temperatures of the hydrogen/dry air mixture. (Zhang, C., Zhang, J., and Ma, J., 2012, “Hydrogen catalytic combustion over a Pt/Ce0.6Zr0.4O2/MgAl2O4 mesoporous coating monolithic catalyst,” Int. J. Hydrog. Energy, 37, pp. 12941-12946.) They found that for low temperature catalytic combustion of hydrogen, the initial reaction temperature, H2 concentration, and flow rates were very important parameters. They tried hydrogen combustion at mixture temperatures of 298 K and 263 K and their results show that higher H2 concentration was helpful in initiating and sustaining catalytic combustion. For the 263 K combustion, the authors could not achieve conversions higher that 40% for low hydrogen concentrations and although they could start the catalytic combustion, they described the largest challenge to be avoiding product water from freezing.
Additionally, several reaction mechanisms are available in literature on hydrogen oxidation. Although the rates were determined for macroscale channels, these reaction rate coefficients can be used for microscale simulations since they are surface reactions. Warnatz et al. studied stagnation flow of hydrogen-oxygen mixture over a platinum surface and developed a reaction mechanism for H2/O2 combustion. (Warnatz, J., Allendorf, M. D., Kee, R. J., and Coltrin, M. E., 1994, “A model of elementary chemistry and fluid mechanics in the combustion of hydrogen on platinum surfaces,” Combust. Flame, 96, pp. 393-406.) Warnatz et al.'s as well as three other homogeneous reaction mechanisms as well as three heterogeneous reaction schemes were tested by Appel et al. (Appel, C., Mantzaras, J., Schaeren, R., Bombach, R., Inauen, A., Kaepperli, B., Hemmerling, B., and Stampanoni, A., 2002, “An experimental and numerical investigation of homogeneous ignition in catalytically stabilized combustion of hydrogen/air mixtures over platinum,” Combust. Flame, 128, pp. 340-368.) When combustion in a 7 mm high channel with platinum covered walls was considered, they found differences from 8 to 66 percent between the modeling results using these schemes and their own experimental results for ignition characteristics. Their study showed that Warnatz's homogeneous reaction mechanism and Deutschmann's heterogeneous reaction mechanism give the best predictions, within 8% of the experimental results. (Deutschmann, O., Schmidt, R., Behrendt, F., and Warnatz, J., 1996, “Numerical modeling of catalytic ignition,” Proc. 26th Symposium (International) on Combustion/The Combustion Institute, Pittsburgh, Pa., pp. 1747-1754.)
However, a need remains in the art for a compact and efficient microchannel heat exchangers for low temperature applications, in particular, for example, ones which provides combustion of fuels and heat exchange from the combustion gases to a working fluid.
In one of its aspects the present invention relates to a miniaturized power generation device, such as a microscale combustor and heat exchanger (μCHX). The μCHX may include several repeating unit cells each of which performs three unit operations: combustion, recuperation, and heat exchange. Heterogeneous catalytic combustion may occur on the walls of microchannels in the presence of a platinum catalyst. In one particular configuration, the present invention may include a distributed catalyst arrangement which deters extinction of the reaction due to a cold gas stream and which provides a high hydrogen conversion (greater than 95 percent) for a range of operating conditions.
For instance, in one of its aspects, the present invention may provide several microchannels that are connected in parallel in order to meet the thermal power requirements of the desired application. The parallel microchannels may be linked together by inlet and outlet headers that distribute the flow uniformly amongst the microchannels. When only one working fluid is involved, a single layer of parallel microchannels could be sufficient, as in the case of a heat sink or simple chemical reactors. However, for heat or mass exchangers and more complex chemical reactors, a stacked up, multi-layer parallel microchannel architecture may be needed. In the design of such microchannel devices, it may be sufficient to optimize the performance of a single microchannel “unit cell” (for example, two microchannels separated by a non-permeable wall if the device is a heat exchanger) and to ensure that the flow distribution between the microchannel unit cells and between the stacked layers is uniform. Typically, the unit cells and headers may also be designed with pressure drop constraints in mind.
In one exemplary application, the devices and methods of the present invention may find use in an automotive cryo-adsorbant storage system for hydrogen. In such an exemplary application, hydrogen gas that exits a cryo-adsorbant storage tank needs to be heated to a minimum temperature of 233K (−40° C.) prior to entering the fuel cell. During cold start conditions, heat exchange with ambient air or with the fuel cell coolant is insufficient to provide this minimum temperature, thereby requiring an additional source of thermal energy. This thermal energy can be provided by combusting a small portion of a cold hydrogen stream in a device capable of transferring heat of reaction back to the cold stream. To maintain a high on-board efficiency and low storage system weight and volume, it is desirable for the device to be small, lightweight and operate at a high efficiency. Another exemplary application is building or distributed heating. Exemplary configurations of compact devices based on parallel microchannel architecture are presented herein.
For example, the present invention may provide a microscale combustor and heat exchanger, comprising a plurality of layers each having one or more respective channels extending therethrough. The layers may be joined to one another to permit gaseous communication between selected respective channels of the layers. The plurality of layers may include a combustor layer comprising at least one combustion channel having a catalyst disposed therein and a recuperator layer comprising at least one recuperation channel having a catalyst disposed therein. The recuperation channel may be disposed in gaseous communication with a respective combustion channel to receive a combustion gas therefrom. The combustor layer and recuperator layer may be disposed in stacked arrangement so that the at least one recuperation channel and the at least one combustion channel are disposed over one another with a common channel wall therebetween. The plurality of layers may also include a heat exchange layer having at least one heat exchange channel disposed therein, the recuperation layer may be disposed between the heat exchange layer and the combustor layer.
In addition, the present invention may provide a microscale combustor and heat exchanger comprising a plurality of layers each having one or more respective channels extending therethrough. The plurality of layers may include a combustor layer having at least one combustion channel disposed therein, and a heat exchange layer having at least one heat exchange channel disposed therein. In addition, a casing may be provided around and enclosing the plurality of layers and may include an inner wall defining a cavity disposed therein; the cavity may be dimensioned to provide a gap between at least a portion of the inner wall and the plurality of layers, with at least one heat exchange channel disposed in gaseous communication with the gap. A recuperator layer may be provided having at least one recuperation channel disposed therein, and the recuperation layer may be disposed between the heat exchange layer and the combustor layer. The at least one heat exchange channel may include an inlet in gaseous communication with the gap and the heat exchange and combustor layers may each include respective working fluid outlet ports in gaseous communication with the gap. Further, one or more of the plurality of layers may include a groove or two concentric grooves disposed therein between the respective channels and an edge of the respective layer. The at least one combustion channel, at least one recuperation channel, and/or at least one heat exchange channel may include pin fins disposed therein, and the at least one combustion channel and/or at least one recuperation channel may include a catalyst disposed therein.
In yet an additional aspect, the present invention may provide a microscale heat exchanger, comprising a plurality layers each comprising one or more respective channels extending therethrough. The plurality of layers may include one or more heat exchange layers having at least one heat exchange channel disposed therein, and may be disposed in stacked arrangement so that at least one channel from each of two or more layers is disposed adjacent one another with a common wall therebetween through which heat may be exchanged. The plurality of layers may include a shroud disposed between the one or more respective channels and associated respective edges of the layers. One or more of the plurality of layers may include a groove disposed therein between the shroud and an edge of the layer.
The foregoing summary and the following detailed description of exemplary embodiments of the present invention may be further understood when read in conjunction with the appended drawings, in which:
In one of its aspects the present invention provides a general device design which includes several repeating unit cells 100 each of which may perform three unit operations: combustion, recuperation, and heat exchange,
Two levels of numerical simulations are performed to realize the design. The first level represents a single unit cell 100 comprising a combustion channel 110, two recuperator channels 140, and two heat exchange channels 120,
In one exemplary application, the multiple unit cell microscale combustor and heat exchanger (μCHX) 600 can operate at temperatures as low as 200 K. One particularly useful application of the μCHX 600 is cryo-adsorbent hydrogen storage systems for fuel cell cars under cold start conditions. In this exemplary application, desired operating conditions for the μCHX 600 are shown in Table 1. A fraction of the incoming cold hydrogen gas would be premixed with air to provide the thermal energy rate needed to increase the temperature of the rest of the hydrogen flow from 200 K to 233 K. The hydrogen flow rate to the fuel cell could vary between 0.5 g/s to 2 g/s and the working pressure could vary between 5 to 20 bars. For the cold start condition, the environment, from which air is drawn for the combustion process, is assumed to be at 233 K (−40° C.).
Each unit cell 100 of the multi-unit μCHX 600 was designed to perform, at the minimum, the unit operations of combustion and heat exchange to the working fluid (e.g., a cold hydrogen stream), by having a combustion channel 110 surrounded on both sides by heat exchange channels 120. The combustion channel 110 may include inner catalyst beds 130 disposed on respective inner surfaces 112 of the combustion channel 110, and may include outer catalyst beds 132 disposed on respective outer surfaces 114 of the combustion channel 110, which catalysts beds 130, 132 may include a noble metal, such as platinum, palladium, rhodium, and/or other suitable material, for example. The catalyst can be in the form of a thin layer (coating) deposited on the channels walls, or the catalyst can be in the form of insets put inside the channels (or attached to the walls). The catalyst (coating or insert) may comprise a porous material.
Results from simulations (described below) for such a unit cell design showed that the flow of a very cold working fluid (e.g., 200 K hydrogen gas) in the heat exchange channels 120 reduced the temperature of the catalyst beds 130, 132 and prevented catalytic combustion. Therefore, recuperator channels 140 were provided between the combustion and heat exchange channels 110, 120 to provide a thermal buffer therebetween,
Geometrical arrangements and thermofluidic parameter values that ensure high efficiency and conversion were determined Efficiency is used to identify the overall performance of the unit cell 100, and is defined as the ratio of the amount of heat transferred to the working fluid (e.g., cold hydrogen stream) to the chemical energy of input hydrogen in the combustible gas mixture,
where MH
As part of the design considerations, to produce a high specific power within the unit cell 100, an equivalence ratio of unity was considered. The equivalence ratio, φ, is the ratio of the molar fuel-to-air ratio at the desired test conditions to that at stoichiometric conditions. The large equivalence ratio also results in reduced pressure drop for the same thermal power generated when compared with lower equivalence ratio mixtures.
In the exemplary design, the height of the combustion and heat exchange channels 110, 120 were 300 μm each, and height of the recuperator channel 140 was 150 μm. The width of all channels 110, 120, 140 in the unit cell was 2 mm while the length, L, of the unit cell 100 was kept at 15 mm. Based on simulations, this length L was found to provide sufficient area for heat exchange between the recuperator and heat exchange channels 140, 120 while keeping the pressure drop low. All outer walls of the recuperator and heat exchange channels 140, 120 were considered insulated (as indicated by the cross-hatching in
Initially the catalyst beds 130, 132 were located entirely on the inner walls 112 of the combustion channel 110 alone; however, simulations showed that almost half of the catalyst length was not being efficiently utilized because of the low reactant mixture temperature. Despite the thermal buffer provided by the recuperator channels 140, the cold hydrogen gas flow in the heat exchange channels 120 tended to decrease the mixture temperature rapidly past the initial part of the catalyst bed 130. As a result, hydrogen conversion of only around 80 percent was typically achieved.
In order to obtain higher conversions while managing the amount of catalyst used, a novel catalyst bed arrangement was used. The basic premise of the new arrangement was that higher gas mixture and catalyst bed temperatures resulted in higher reaction rates and hence more complete hydrogen conversion. To achieve this larger catalyst temperature, the catalyst bed was redistributed such that ⅖ of the catalyst bed was shifted from the end of the inner wall 112 of the combustion channel 110 and placed in the recuperator channels 140 on the outer surface 114, (i.e., outer catalyst bed 132). The location of the outer catalyst beds 132 within the recuperator channels 140 coincided with the location of the inner catalyst beds 130 in the combustion channel 110,
The catalyst used in a microscale combustor and heat exchanger 600 in accordance with the present invention may have any site density. Catalyst site density is defined as the amount of catalytically active site per unit area (cm2). Catalyst site density may be adjusted based on the required reactivity of the catalyst, e.g., the lower the site density, the lower the reaction rate (and therefore produced power). In addition, the catalyst in the combustor and recuperator channels 110, 140 may have different site densities. In the case where the catalyst surface (catalyst bed) in the combustor and recuperator channels 110, 140 have different site densities, the total amount of catalyst can be calculated by multiplying the site density value by the area. The ratio of the amount of catalyst in each channel 110, 140 over the summation of the catalyst sites (total catalyst amount in the combustion channel 110+total catalyst amount in the recuperator channel(s) 140) can be in the range of 1/10 to 9/10.
Two-dimensional, steady-state simulations were carried out on the mass, momentum, energy and species (both gas-phase and surface species) balance equations for the unit cell geometry indicated in
OH (s) + Pt(s)
H2O(s) + Pt(s)
OH(s) + OH(s)
aThe hydrogen adsorption (first reaction) is first order with respect to platinum.
bSticking coefficient.
The mass fraction of species at the inlet of the unit cell 100 was defined by the equivalence ratio. For brevity, the governing equations and boundary conditions were set according to previous work by two of the presently named inventors. (Ghazvini, M., and Narayanan, V., 2011, “Performance characterization of a microscale integrated combustor recuperator oil heat exchanger,” Proc. AJTEC2011: ASME/JSME 2011 8th Thermal Engineering Joint Conference, Honolulu, Hi., 2011). The numerical model was validated against the combined experimental and numerical study on hetero-/homogeneous combustion of hydrogen/air mixtures over platinum in a single channel by Appel et al. (Appel, C., Mantzaras, J., Schaeren, R., Bombach, R., Inauen, A., Kaepperli, B., Hemmerling, B., and Stampanoni, A., 2002, “An experimental and numerical investigation of homogeneous ignition in catalytically stabilized combustion of hydrogen/air mixtures over platinum,” Combust. Flame, 128, pp. 340-368.) The current simulations were seen to predict the experimental data in Appel et al. at least as well as the parity between their own numerical simulations and experiments.
The governing equations and boundary conditions were solved in FLUENT® V14 (ANSYS, Inc., Cannonsburg, Pa., USA) in conjunction with Chemkin-CFD™ (Reaction Design, San Diego, Calif., USA) for the chemical reactions. A non-uniform mesh was used to refine the near-wall regions. The total number of grids was 34,300 for the whole model domain. It was initially verified that an orthogonal staggered grid of 606×20 grid points (in x and y, combustion channel, or 12,120 grids) was sufficient to produce a grid independent solution. Additionally, adaptive mesh refinement was also applied in several locations based on the gradients of mass imbalance for better convergence. The simulation convergence was decided when the residuals approached steady values asymptotically and when the relative residuals were smaller than 10−7 for continuity, momentum, energy and species, with the exception of residuals for O and H species which were less than 10−3.
Based on the required mass flow rate of cold hydrogen in Table 1, the first step in a unit cell simulation was to determine the flow rate of cold hydrogen and flow rate of hydrogen in the combustion channels 110 within each unit cell 100. Based on previous work with oil as the heat exchange fluid, the initial inlet mixture velocity to the combustion channel was picked to be 4 m/s. In that study, this flow rate provided sufficient residence time for conversions in excess of 90 percent. The inlet velocity of the cold hydrogen stream was set arbitrarily to a value. The exit temperature of the cold hydrogen stream was checked against the requirement in Table 1 (>233 K). Simultaneously, the hydrogen conversion was verified to be in excess of 90 percent. If the exit temperature was found to be lower, the flow rate of cold hydrogen was lowered. If the hydrogen conversion was found to be lower, the inlet mixture velocity of the combustion gases was lowered. Once conditions that met both the cold hydrogen exit temperature as well as conversion were achieved, the fraction of hydrogen stream required for combustion to that being heated in the heat exchange channels 120 was determined. Based on the ratio of the maximum mass flow rate of the cold gas in the μCHX to the mass flow rate in each unit cell, a total number of 168 unit cells with the channel width of 2 mm was found to be sufficient to provide heat to 2 g/s of cold hydrogen. For the lower mass flow rate limit of 0.5 g/s in Table 1, the same number of unit cells was retained, with a corresponding decrease in velocity of the cold hydrogen and combustion mixture gases. For example, at a pressure of 5 bar the flow rate of cold hydrogen gas through each of the 168 unit cells were 6.1 mg/s and 1.5 mg/s for a total flow rate of 2 g/s and 0.5 g/s, respectively.
A representative temperature contour plot within the unit cell 100 for the conditions indicated in Table 1 and heat exchange fluid (cold hydrogen) inlet flow rate of 2 g/s, inlet temperature of 200 K, and pressure of 5 bars is shown in
In the unit cell simulations, based on the requirements stated in Table 1, two parameters were varied—the inlet pressure and mass flow rate of the cold hydrogen gas. Table 3 summarizes the achieved conversions and unit cell efficiencies for the extremities in the range of the desired pressures and hydrogen mass flow rates presented in Table 1. It can be seen that the exit temperature in all cases is in excess of 233K and that the conversions and efficiencies are in excess of 99% and 92% respectively. Pressure drop values within the combustor and recuperator channels 110, 140 are also shown in Table 3. A larger cold hydrogen mass flow rate requires a proportionally larger hydrogen/air mixture flow rate, thereby increasing the pressure drop in the combustion and recuperator channels 110, 140 with increase in cold hydrogen mass flow rate. However, pressure drop is almost independent of the working fluid pressure, because the mass flow rate of hydrogen is kept the same for both 5 bar and 20 bar. The slight increase in pressure drop with higher pressure is a result of the larger specific heat value at higher pressures. With the temperature difference between the inlet and exit fixed at 33 K, a larger Cp resulted in an increase in the amount of heat rate needed and hence a slightly larger flow of hydrogen-air mixture within the combustion channel 110.
One goal was to fabricate and characterize the performance of a multi-unit cell μCHX and validate the numerical simulations with laboratory measurements. However, due to safety considerations in the laboratory, cold nitrogen gas was used in place of cold hydrogen gas as the heat transfer fluid. Since density and thermal properties of nitrogen are considerably different from those of hydrogen, additional simulations were performed using cold nitrogen gas for the unit cell 100 of
There are two thermal resistances in the path for the requisite amount of heat to be transferred within the unit cell 100 from the heat exchange wall 116 (separating the recuperator and heat exchange channels 140, 120) to the cold gas. The first one pertains to the convection resistance, Rconv=1/(hcold gasA), while the second one is the resistance due to heating of the cold gas stream, Rheat=1/({dot over (m)}Cp). For the range of hydrogen flow rates considered, the flow is laminar and hence the heat transfer coefficient, assuming fully developed flow, is about 1120 W/m2-K. The Rconv and Rheat for cold hydrogen flow within each unit cell 100 are 35.7 K/W and 5,926 K/W respectively. The Rheat estimate is based on a working pressure of 5 bar and for a flow rate of 2 g/s. Since Rheat is the dominant thermal resistance in transferring heat to the cold gas, in order to preserve the same representative thermal conditions as cold hydrogen flow, it is clear that the Rheat between hydrogen and nitrogen flows has to be matched. Hence, the heat capacity rates (1/Rheat) between the cold hydrogen and cold nitrogen flows, as well as the temperature rise (see Table 1) were kept identical. This meant that, because of the high specific heat of hydrogen compared to nitrogen (CpH
Table 4 shows the result of the simulations with nitrogen as the working fluid. The working pressure was fixed at 5 bar for these simulations. Nitrogen mass flow rate of 26.1 g/s and 6.57 g/s have the same heat capacity of 2 g/s and 0.5 g/s of hydrogen, respectively. By a comparison of results in Tables 3 and 4, it can be seen that the hydrogen conversion remains largely unaffected by changing the heat exchange fluid. This result is to be expected since changing the heat exchange fluid only changes the boundary condition on the combustion process. When the heat capacity rates are matched, the temperature drop along the heat exchange channels 140 would remain similar for both cold hydrogen and cold nitrogen cases, thereby causing little variation in the hydrogen conversion. It can also be seen that the unit cell efficiency remains unchanged between the two cases which is an indication that the convective resistances on the recuperator and heat exchange channels 140, 120 are smaller than that of the thermal resistance along the heat exchange channel 120 (1/{dot over (m)}Cp). Table 4 shows that, similar to cold hydrogen flow, pressure drop in the combustor and recuperator channels 110, 140 are higher when the working fluid flow rate is higher.
As described above, a total of 168 unit cells 100 are needed to increase the temperature of hydrogen flow of 2 g/s from 200 K to 233K. The same amount of heat (911 W) can be removed using 26.1 g/s of nitrogen flow with the same inlet and outlet temperatures. In the present device level design, cold nitrogen gas flow is used for the working fluid. Because the flow rate of cold nitrogen is 13 times larger than that of cold hydrogen, the former presents a limiting case in the design of the headers for uniform flow distribution.
Only fluid flow is simulated at this level due to computational requirements. Properties of nitrogen were estimated at an average temperature of 216.5 K while the properties of hydrogen-air fuel mixture were kept fixed at 400 K. Another important index for performance is pressure drop. The performance measure at the device level pertains to uniform flow distribution amongst the unit cells and an overall low pressure drop within the device.
The microscale combustor and heat exchanger 600 comprising multiple unit cell stacks 400 was designed so it could be fabricated using chemical etching and diffusion bonding,
Since the width of the microchannels 415, 425, 315, 435, 445 (hereafter “415-445”) within each unit cell 100 of the design of
In the combustor layer 300, four combustion channels 315 may be located on each side of the combustor layer 300. Locations of different inlets 312 for the combustion gas mixture are also shown in
In particular, combustion gas inlets 312 may be provided in the combustor layer 300 to introduce a combustion gas into the combustion channels 315,
As to the working fluid, the upper and lower heat exchange layers 410, 440 may each include respective working fluid inlets 418, 448 through which a working fluid may be introduced into the respective heat exchange channels 415, 445,
The thickness of the five layer stack 400 of eight unit cells may be about 3 mm. Twenty-one such stacks 400 of layers 410-440 may be stacked up with a top and a bottom header caps 602, 604,
At the layer level, the design ensured that two criteria were satisfied: (1) uniform flow distribution between the microchannels 415-445 within each layer 410-440 and (2) manufacturability. Three-dimensional simulations of fluid flow were used to verify the uniform flow among the recuperator, combustion, and heat exchange channels 415-445.
In addition to uniform flow distribution, the design needed to accommodate the diffusion bonding manufacturing requirements. One of these requirements was that the wall thicknesses had to be sufficiently thick to provide a leak free seal between fluids. To meet this requirement, walls that separated different fluids (combustion mixture, exhaust, and cold nitrogen) were thickened to 1 mm. In addition, a conservative 5 mm of solid material was added around the periphery of each layer 410-440 to ensure that fluids did not leak out of the stack 400.
Turning to the header, the location of the inlet and exit manifolds 603, 605, 608, 616 of the microscale combustor and heat exchanger 600 were designed to promote uniform flow distribution amongst the different layers 410, 420, 300430, 440,
There are two different fluids flowing in the microscale combustor and heat exchanger 600, therefore there needs to be at least two inlets 605, 608 and two outlets 603, 616. In the current design, for better flow distribution and lower pressure drops in both streams, two inlets 605, 608 are included for each stream. In the present design, hexagonal inlets 416-446 are used to provide better flow distribution and to reduce the size of the headering section,
Several three-dimensional simulations of fluid flow were performed in FLUENT® V14 (ANSYS, Inc., Cannonsburg, Pa., USA) in order to determine the proper headering design. While not optimized, the designs presented here represent iterative efforts at obtaining uniform flow distribution amongst layers 410-440. In these simulations, 3 sets of 8-unit cell stacks 400 were placed on top of each other to make a 24 unit cell μCHX. The flow rate through each unit cell was kept identical to the largest flow rate 5.75×10−5 for heat exchange layers 410, 440 and 6.70×10−6 for the combustor layer 300, and hence the flow distribution for larger stacks should be similar to the one presented herein.
Four additional layers 702, 704, 706, 708 on top and three layers 705, 707, 709 at the bottom of the stacks 400 were necessary to transition from circular tubing of the header caps 602, 604 to the plena that distribute flow amongst different layers 410-440,
Pressure contours for flow within the headers 602, 700, combustion gas inlet ports 412-442, and combustion channels 315 is shown in
Pressure contours of the nitrogen gas flow within the headers 602, 701, nitrogen inlet ports 418-448, and the heat exchange channels 415, 445 are shown in
The design presented for the nitrogen headering is an extreme case since the flow rates for nitrogen are 13 times that of hydrogen, which will be the actual working fluid for the device. The thermo-fluidic design presented herein does not consider conjugate heat transfer effects, which could be significant especially when the thermal conductivity of the material from which the layers 410-440 is fabricated is large.
A prototype of the μCHX 600 of
In the experiments, due to safety considerations, cold nitrogen gas was used as the heat transfer fluid instead of cold hydrogen gas to be used in the actual application.
An experimental facility 230 was built to characterize the performance of the device, labeled “μCHX” in
Inlet and outlet temperatures, pressures and flow rates of the streams were measured using digital data acquisition. The gathered data was used to calculate the overall performance of the μCHX and was defined as the ratio of the amount of heat transferred to nitrogen to the chemical energy of input hydrogen as given by Eqn. (1) above, with the work fluid, wf, being nitrogen in the present experiment.
Heat losses were calculated by subtracting the heat transferred to nitrogen from the heat produced in the combustor by the combusted hydrogen.
The temperature of the body of the device was measured at four different locations. To determine the body temperature, readings from two of the thermocouples that were located closest to the catalyst section of the μCHX were averaged.
Varied parameters include the inlet temperature of cold nitrogen (150-273 K), mass flow rate of cold nitrogen, and mass flow rate of the combustible gas mixture. The performance of the device was characterized using three parameters: pressure drop across the combustor and recuperator channels, hydrogen conversion, and efficiency index. The equivalence ratio, which is defined as the ratio of the molar fuel-to-air ratio at the test conditions to that at stoichiometric conditions, was kept constant at unity.
One of the important factors in hydrogen conversion was the residence time over the catalyst for reactions. Residence time is defined as the time a hydrogen molecule has to react before leaving the catalyst region.
t
r
=L
Cat
/
where LCat is the catalyst bed length and
It can be seen that H2 conversion increases by about 10% with increasing residence time from 9-56 ms. However, the efficiency of the device drops from 82% to 66% due to heat losses. When the power generated in a device is small, the effect of heat losses on the overall efficiency of the system can be significant. The device was covered with Pyrogel® and Cryogel® (Aspen Aerogels, Inc., Northborough, Mass., USA) insulations, and the body temperature was kept approximately the same between all experiments at about 186° C. and therefore there was not much difference between heat losses (in watts, see Table 8). However, the ratio of the heat loss to total power ratio which is defined as,
increased with an increase in residence time. Since the total length of the catalyst and the equivalence ratio were fixed, in order to achieve different residence times the flow rate of the hydrogen/air mixture had to be varied (see Eq. 4). Different hydrogen/air flow rates resulted in different input power to the system (Table 8). Therefore, that heat loss ratio was different for different cases and is shown in
The body temperature of the combustor was a second important factor in conversion and device efficiency since higher reaction temperatures result in higher reaction rate and hence conversion.
The heat loss values varied from 1.1 W for the body temperature of 124° C. to 4.6 W for the body temperature of 196° C. Since the power input was high in these cases (120 W), although heat loss increased, the heat loss ratios were small numbers and had insignificant effect on the overall efficiency of the system. Therefore the efficiency also increased with increasing the body temperature. The experimental results showed that hydrogen residence time and body temperature had significant effects on the overall efficiency of the device. Conversions as high as 94.1% and efficiencies as high as 88.3% were achieved,
In another of its aspects, the present invention provides an alternative exemplary configuration of a microscale combustor and heat exchanger (μCHX) 900, which retains the basic features of the unit cell design 100 of
Turning to the layer structure in more detail, consistent with the schematic in
Combustion gas inlets 852 may be provided in the combustor layer 850 to introduce a combustion gas into the combustion channels 855,
As to the working fluid, the upper and lower heat exchange layers 810, 890 may each include respective working fluid inlets 816, 896 through which a working fluid may be introduced into the respective heat exchange channels 815, 895,
The multi-unit cell stack 800 may be designed such that it may be fabricated using chemical etching and diffusion bonding/laser welding. (The device could readily also be manufactured using additive manufacturing technologies such as 3D printing.) The manufacturing methods impose some important constraints on the design of the layers 810-890 which are addressed in this design. Etching does not provide sharp corners at the bottom of the walls and the cross section of the channels will be U-shaped. That reduces the cross section area and can increase pressure drop and cause maldistribution. In order to reduce the effect of curved corners, the width of the walls should be at least twice the height of the channels 815-895. On the other hand, bonding methods have other limitations. For example for diffusion bonding there should not be an unsupported span, and force should be transmitted from top to bottom in the regions that needs to be diffusion bonded. In addition, diffusion bonding as well as laser welding needs a minimum width for the bonding surface.
Each layer 810-890 may have a thickness of 600 μm, and thus the thickness of the five layer stack 800 may be about 3 mm. Unlike the design of
There are two reasons for having 150 μm deep features on both sides instead of 300 μm deep features on one side. The first reason pertains to the constraints of the fabrication process. Isotropic chemical etching limits the distance between features to be at least twice the etching depth. For example, if the etch depth is 300 μm, the distance between features has to be at least 600 μm. However, in the designs in
In addition to uniform flow distribution, the design needed to accommodate the bonding manufacturing requirements. One of these requirements was that the layer wall thicknesses had to be sufficient to provide a leak free seal between fluids. To meet this requirement, walls that separated different fluids (combustion mixture, recuperator, and cold nitrogen) were thickened to at least 2 mm. In addition, a conservative 3 mm of solid material was added around the periphery of each layer 810-890 to ensure that fluids did not leak out of the stack 800.
At the layer level, the design again ensured that two criteria were satisfied: (1) uniform flow distribution between the microchannels 815-895 within each layer 810-890, and (2) manufacturability. Three-dimensional simulations of fluid flow were used to verify the uniform flow among the channels 815-895.
Heat loss from a combustor by convection is directly proportional to the body temperature. By designing a shroud 500 around the cell stack 800, heat loss is minimized. Specifically, the cell stacks 800 may be stacked together to provide four unit cells which may then be placed in an enclosing casing 901 comprising lower and upper casing portions 902, 904 which are larger than the stacks 800,
The heat exchange fluid may flow in the shroud 500 around all the layers 810, 890 that have openings 816, 819 for the intake of the working fluid. In this regard there may also be a space between a top distribution layer 920 and the layer stacks 800 and also between the lower casing portion 902 and the layer stacks 800 which lets the working fluid flow around the layer stacks 800. Therefore in addition to the sides, there is working fluid on top and bottom of the layer stacks 800. It can be seen from
The flow distribution arrangement between the shroud 500 and the heat exchange channels 815, 895 causes only a fraction (typically ⅕th) of the fluid, to enter the heat exchange channels 815, 895. Fluid going through these channels 815, 895 gets heated by heat transfer from the recuperator channels 835, 875 and exits at temperatures typically in excess of 100° C. The hot fluid then mixes with the bypassed fluid in the casing, and the desired air temperature is attained at the exit 903 of the μCHX 900. The distance between the lower casing portion 902 and the bottom of the cell stacks 800 may be adjusted, such as by raised spacers which may be part of the recuperator gas outlets 926, such that the desired mixture ratio is achieved (e.g., 600 μm in this case). This flow distribution arrangement is important for two reasons: (i) by allowing only a fraction of the inlet working fluid to enter the heat exchange channels 815, 895, the overall body temperature within the combustion, recuperation and heat exchange channels 815-895 is maintained relatively high, thereby resulting in high reaction rates and high fuel conversion; and, (ii) the pressure drop through the μCHX 900 is reduced when compared with the entire fluid flowing through the heat exchange channels 815, 895.
To further supplement the insulative effect of the shroud 500, one or two sets of concentric grooves 818, 838, 858, 878, 898 (hereafter “818-898”) may be positioned around the periphery of each of the layers 810-890 exterior to the respective channels 815-895 to further reduce the heat transfer out of the cell stack 800,
Several three-dimensional simulations of fluid flow were performed in FLUENT® V14 (ANSYS, Inc., Cannonsburg, Pa., USA) in order to determine the proper headering design. Pressure contours for flow within the heat exchange layers is shown in
A comparison between the pressure drop between the designs of
An important achievement here is the reduction of pressure drop in the heat exchange stream. The pressure drop in the
In yet a further possible design 200 in accordance with the present invention is incorporation of the casing 901 and shroud 500 of
In another aspect of it aspects, the present invention may provide a microscale combustor 2600, 2601 which includes an auxiliary heat source to heat a selected portion of the combustion channel 2610, 2620 to assist in initiating a catalytic reaction within the combustion channel 2610, 2620. For example, the combustion channel 2610, 2620 may be included a part of a device having the configuration shown in
These and other advantages of the present invention will be apparent to those skilled in the art from the foregoing specification. For example, while the exemplary configurations illustrate parallel flow among the channels of the layers, cross flow channel configurations may also be utilized in accordance with the present invention. Accordingly, it will be recognized by those skilled in the art that changes or modifications may be made to the above-described embodiments without departing from the broad inventive concepts of the invention. It should therefore be understood that this invention is not limited to the particular embodiments described herein, but is intended to include all changes and modifications that are within the scope and spirit of the invention as set forth in the claims.
The contents of all publications cited throughout the text of this disclosure are hereby incorporated herein by reference. In addition, the following symbols used throughout this disclosure have the following meanings:
A Pre-exponential factor (Arrhenius equation) (mol-cm-K-s)
Cp Specific heat at constant pressure (J/kg·K)
Ea Activation energy (Arrhenius equation) (kJ/mol)
h Enthalpy (J/kg)
Hc Combustion channel height (m)
Ho Working fluid channel height (m)
HR Recuperator channel height (m)
M Molar mass (kg/kmol)
{dot over (m)} Mass flow rate (kg/s)
P Pressure (Pa)
T Temperature (K)
Yg Mass fraction of gth gaseous species
Greek Symbols
β Temperature exponent (Arrhenius equation)
φ Species equivalence ratio in a reaction
Γ Surface site density (mol/cm2)
η Efficiency of the unit cell
Subscripts
0 Value at the inlet
in Inlet
g Index for species
out Outlet
This application claims the benefit of priority of U.S. Provisional Application No. 61/842,547, filed on Jul. 3, 2013, the entire contents of which application(s) are incorporated herein by reference.
This invention was made with government support under DE-FC36-09GO19005 awarded by United States Department of Energy. The government has certain rights in this invention.
| Number | Date | Country | |
|---|---|---|---|
| 61842547 | Jul 2013 | US |
