CC BY-NC-ND
0CrossMark
Available online at www.sciencedirect.com
ScienceDirect
Energy Procedia 69 (2015) 947 - 956
International Conference on Concentrating Solar Power and Chemical Energy Systems,
SolarPACES 2014
Numerical modeling and optimization of an entrained particle-flow thermochemical solar reactor for metal oxide reduction
J. P. Muthusamya, S. Abanadesb, T. Shamima, N. Calvet a*
aInstitute Center for Energy (iEnergy), Department of Mechanical & Materials Engineering, Masdar Institute of Science & Technology, Masdar City, P.O. Box 54224, Abu Dhabi, United Arab Emirates bProcesses, Materials and Solar Energy Laboratory (PROMES - CNRS), 7 rue du Four Solaire, 66120 Font Romeu Odeillo, France.
Abstract
The endothermic thermochemical process of metal oxide reduction in an indirectly-irradiated particle-laden flow solar reactor was modeled and analyzed using computational fluid dynamics (CFD) tool Ansys-Fluent. CFD modeling includes chemically reactive multiphase flow including solid-gas interactions, radiation heat transfer among particles, inner reactor walls and gas phase, and particle surface reaction chemical kinetics. A novel indirect heating cavity-type tubular solar reactor designed for continuous metal oxide reduction was simulated for predicting the temperature distribution profiles and benchmarked with on-sun testing results under similar conditions. Further, design optimization on cavity size was performed for the targeted reaction temperature with enhanced handling capacity. A 50 mm cavity height was found to be suited for required temperature of above 1900 K for zinc oxide thermal reduction. Prior to reaction kinetics implementation, the study of inert particle case was carried out to understand the influence of particle heating on thermal profile. Finally, reactive particle-laden flow was simulated using Eulerian-Lagrangian combined approach. The chemical conversion efficiency of the ZnO reduction process and the solar-to-chemical energy conversion efficiency were also calculated for varied inlet particle mass flow rates.
© 2015TheAuthors.Publishedby Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).
Peer review by the scientific conference committee of SolarPACES 2014 under responsibility of PSE AG Keywords: thermochemical; solar reactor; reactive particle laden flow; thermal dissociation; CFD; metal oxide; redox
* Corresponding author. ncalvet@masdar.ac.ae
1876-6102 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer review by the scientific conference committee of SolarPACES 2014 under responsibility of PSE AG doi:10.1016/j.egypro.2015.03.178
1. Introduction
Developments in concentrated solar power (CSP) technologies may provide high temperatures with zero-carbon emission. However, CSP technology demands for highly efficient energy storage, conversion and recovery methods. The effective conversion of solar energy into storable and transportable chemical energy is a long term research objective and offers an efficient path for long-term storage and long-range transport of solar energy. Latest evolutions in commercially advanced CSP technologies make high-temperature thermochemical (TC) energy conversion a sustainable, clean, and efficient path for solar fuel production and thermochemical energy storage (TCES), which offers a promising alternative to the conventional fossil-based fuel technologies.
TC systems combined with CSP can be used to produce chemical fuels such as hydrogen (H2) from water splitting thermochemical cycles that is a sustainable alternative to fossil fuels-based processes [1]. TCES is also potentially interesting for CSP applications because of pros such as long term stable storage, energy loss avoidance, and high thermal storage density, compared to conventional sensible or even latent heat thermal energy storage methods. Among various TC energy conversions, metal oxide cycles using reduction-oxidation (redox) reactions have a significant potential for both solar fuel production and energy storage as they meet future operating temperatures of solar thermal power plants and electricity cost requirements by energy storage methods [1, 2, 3]. In the first step, an endothermic reduction reaction must be carried out in a high temperature solar reactor driven by CSP to produce the pure metal or the lower valence oxide, followed by an oxidation or hydrolysis step targeting TCES or H2 production, respectively. The oxidation step is used to regenerate the metal oxide for subsequent recycling [4]. As the energy conversion and heat recovery efficiency greatly depend on the efficiency of the reduction step, a significant effort must be dedicated to solar reactor design and optimization. Some of earlier studies dealt with solar reactor modelling for high-temperature thermochemical applications [4, 5, 6].
Two-step redox cycle based on zinc oxide (ZnO) is a promising candidate for thermochemical fuel production and energy conversion due to its high energy storage density and fuel productivity potential (15 mmolH2/gZn) compared to other metal oxides such as tin oxide, iron oxide, or manganese oxide [1, 7, 8, 11, 12]. A specific feature of ZnO reduction step is that the Zn species formed during the reaction is in gaseous state (boiling point of 907 °C) and it can be recovered in condensed state by quenching the flue gases to limit the reverse recombination process with the evolved oxygen. The required decomposition temperature for ZnO is around 2000 K and this endothermal reaction can be driven by CSP as the external process heat source. The solid metal zinc can be used in different industrial and automobile applications such as a fuel cell or battery, or in a water-splitting reaction producing hydrogen and regenerating ZnO which can be further recycled to complete the redox cycle [2, 8, 13, 16]. Although ZnO/Zn thermochemical cycle is known for its high process energy efficiency due to high reaction temperature and reduced number of steps, the operating temperature makes challenging the choice of thermally and chemically stable reactor materials and the appropriate reactor design [10]. In order to obtain and withstand high reaction temperatures in the reactor with acceptable efficiency, it is generally necessary to use the cavity-type receiver concept where the effective cavity absorption effect can be increased by focusing concentrated solar energy through an aperture [9, 14, 17]. Cavity- type receiver systems subjected to incoming concentrated sunlight can further be classified as directly or indirectly heated depending on the mode of heat transfer to the reaction site during the endothermic process (direct irradiation of an absorbing particle flow or indirect irradiation via an opaque heat transfer wall) [10, 18-20].
This study aims at simulating and validating a small-scale indirect heating solar chemical reactor applied to ZnO dissociation. The simulation also aims to optimize the cavity design for reaching maximum energy conversion efficiency and reaction temperature requirements. The continuously injected reactive particle flow consists of ZnO reacting particles that also act as volumetric radiant absorber inside the tube. The indirect irradiation through an intermediate opaque tubular absorber results in a more uniform heating of the particles and a better temperature control, while the particle deposition on the optical window is avoided. The separation of the solar cavity receiver from the chemical reactor part is also beneficial for controlling the reacting volume, and avoiding any pollution evolving from insulation degassing. The solar reduction of zinc oxide (ZnO(s) ^ Zn(g) + 0.5O2) in the current continuously-operated reactor configuration has been simulated with Ansys-Fluent using chemical kinetic data for ZnO dissociation available from literature [6, 7]. CFD is applied as a useful tool to analyze the fluid flow, heat transfers, reaction kinetics with multiphase approach at quick turnaround time with much lower costs than expensive experimental techniques [15]. CFD results can be benchmarked with experimental results at defined key conditions.
Thereafter, the problem set up can be repeated for the similar flow physics and sensitivity studies can be performed to understand the behavior of the system under various operating conditions. In this study, numerical analysis using CFD helped to optimize the reactor design and to understand the reaction mechanisms involving ZnO particles at different flow rates.
Nomenclature
k kinetic rate of particle surface reaction per unit area
A pre-exponential factor, kg m~2 s_1
E activation energy, J mol-1
T temperature of the particle, K
^ solar to chemical energy conversion efficiency, %
AHr reaction enthalpy of ZnO dissociation, 456 kJ mol-1 at 2000 K
cp,ZnO specific heat of ZnO, J mol-1 K-1
T0 ambient temperature, K
FZn Zn flow rate, mol s-1
Psolar solar power input, W
2. Description of the solar reactor setup
The device consists of two distinct parts, i.e., solar receiver cavity and chemical reactor that are isolated from atmosphere and to each other, as shown in Fig. 1. (a). The tubular reactor is inserted vertically into the square cavity solar receiver that is surrounded by outer graphite and ceramic fiber insulators and closed at the front by the glass window. The insulation materials are used to minimize the heat losses from the cavity and receiver. A 10 mm-thick insulating carbon felt is wrapped over the graphite cavity walls to avoid the direct contact between ceramic wool and hot cavity walls. A 20 mm-diameter aperture at the front cavity wall is used to absorb the concentrated solar irradiation through a hemispherical quartz window. Then, the tube placed inside the cavity is heated mainly by radiation from the surrounding cavity walls. Two different pyrometers placed on either side of the reactor serve as optical devices for direct temperature measurements on the tubular reactor external walls. The reactive medium is heated mainly by the conduction through the tubular walls and subsequent indirect radiation offered by inner tubular walls. Particles of solid reactant and inert carrier gas are supplied to the system through the tubular path in the downward direction for redox reactions and the necessary energy for driving the endothermic reaction (for both reactant heating and reaction enthalpy) is provided using the indirect solar heating as explained. The experimental set up during the on-sun testing is shown in Fig. 1. (b).
Fig. 1. (a) Schematic of the solar thermochemical tubular cavity-reactor
Fig. 1. (b) On-sun experimental testing of the reactor
3. Numerical simulation of solar thermochemical reactor
3.1. Description of the computational model
The existing reactor system was modeled including the flow domain, cavity, cylindrical tubular reactor, and insulators (felt and ceramic fibers) for the CFD analysis as shown in Fig. 2. The temperature measurement locations in the experiment for comparison with numerical model are also highlighted in Fig. 2. For the domain boundaries, the equilibrium convection conditions on the outer wall and equivalent radiation flux on the aperture opening was defined. The existing cavity height is 30 mm and it was used for the baseline study to establish the thermal profiles and compare with the numerical results. A 3D computational domain as shown in Fig. 2 was modelled and discretized (meshed) using Gambit. The mesh consists of both hexagonal and tetragonal elements (1938441 cells and 386818 nodes) with quality parameters well within the reference values in order to obtain better results and the cut section of mesh on x-y plane is shown in Fig. 3. Physical properties of the various components and chemical species used in this study are referred from the previous studies [4, 6].
The discretized domain was then imported to the CFD solver Ansys-Fluent for further solving the appropriate governing equations for fluid flow, thermal and chemical kinetics along with particle-flow interactions and radiative heat exchange between particles and walls.
3.2. CFD model setup and boundary conditions
The system was first simulated for predicting thermal profiles without considering the particle flow and reaction kinetics. Once the residuals stabilized and satisfactory convergence achieved, particles were introduced into the analysis to establish the particle-flow interaction with heat and momentum exchange. Before activating the reaction kinetics, the particles were assumed to be inert and subjected to only force and heat balance. The reason of introducing the inert particles was to better understand their influence on the thermal performance inside the reactor due to particle heating. Once the effect of particle heating on thermal behavior was predicted, chemical kinetics were activated to simulate the complete flow physics. The model includes heat and mass transfer, chemical reaction, multiphase flow (particles and inert carrier gas flows), and the coupling of these phenomena. The details of the governing equations can be found elsewhere [4, 6, 10].
The reactive particle-laden flow was simulated by a Lagrangian discrete phase model combined with the Eulerian continuum approach. The dispersed particle phase was solved by tracking a number of particles across the established flow field, while the carrier phase was treated as a continuum by solving the time-averaged Navier-
Fig. 2. Computational domain and temperature monitored positions
Fig. 3. Discretized domain (X-Y section plane)
Stokes equations. The dispersed phase exchanges momentum, mass, and energy with the carrier gaseous phase. The coupling between the carrier and dispersed phases and its impact on each other was included in the CFD model.
The radiative heat transfer was included using discrete ordinates (DO) grey radiation model. The radiation effect between particles and particle-walls was also included in the radiation model. The participation of the particulate discrete phase in the radiation calculation consists of absorption, emission and scattering. The heterogeneous surface decomposition reaction consumes the ZnO particles (particle shrinkage) and this is a source of Zn species in the gas phase during the computation of the transport equations. The surface reaction also consumes energy, in an amount determined by the heat of reaction. The reaction rate is determined by inherent kinetics of ZnO dissociation on the basis of the Arrhenius law [6, 7].
The gas and particle mass flow rates with constant homogeneous particle diameter was the boundary condition at the tubular reactor inlet and the outlet condition was considered to be pressure outlet with zero gauge (atmospheric pressure). The aperture face was defined as a radiation inlet with specified heat flux of 1 kW in line with a typical solar radiation flux during experiments. An annular face around the aperture was also created as a gas inlet to input the nitrogen flow rate for ventilation of the cavity and mass flow rate was defined. The outer walls of the whole system were defined with equilibrium convective heat transfer coefficient and free stream temperature at lab test conditions. Appropriate solver settings including steady, 3D, double precision with higher order schemes for discretization of variables were defined. The governing transport equations were solved until the residuals and monitored variables are stabilized and post-processing was done from the converged solution.
3.3. Base line simulation results and validation
The base line study with 30 mm height cavity receiver was first simulated and validated with the experimental results for the similar operating conditions (no particle injection). For the input heat flux of 1 kW, the temperatures at different locations were compared between experimental and CFD simulation, and good agreement of results was found as shown in Fig. 4. The temperature measurements at the front side of the tube and at the center of the tube show very good agreement (difference lower than 5 %) with respect to numerical predictions. It shows that the radiation heat exchange between aperture, cavity walls, tube walls and particles are captured appropriately. However, the over prediction of temperatures at back side of the tube and external cavity by CFD can be justified such a way that the contact resistance between the solids are not modeled in the numerical study. Properties of the different materials used in the computational study may also be somewhat underestimated (e.g., actual thermal conductivity of insulators may become higher than the theoretical one given by the manufacturer), which is also the reasons for the deviation as the physical properties may change over time, temperature and test cycles due to materials ageing. Overall, the temperatures predicted by CFD analysis are higher than the experimental measurements because the CFD simulations are solved until the complete thermal equilibrium is achieved, whereas the experiments had not reached the steady state while the measurements were being acquired. However, the temperature measurements at the front side of the tube and inside the tube are more crucial than the other locations, and the values are not influenced by contact resistance, properties of materials and thermocouple intrusions. Hence, it can be inferred that the problem setup and applied boundary conditions are appropriate enough to predict the temperature profiles reasonably. Experimentally measured time-dependent temperatures at different locations along with the numerically predicted steady state temperatures embedded are shown in Fig. 5. It can be observed that the experiment is yet to reach the equilibrium state and hence temperature values are lower than numerical predictions.
Fig. 4. Results comparison (CFD vs. experiments) Fig. 5. Transient temperature by experiment vs. CFD
3.4. Characterization of cavity size vs. input solar power
After the validation of the base case, a number of simulations were performed with different cavity sizes in order to optimize it for enhanced reactor performance. The cavity sizes of 30 mm, 40 mm, 50 mm and 60 mm were simulated with power input levels ranging from 600 W to 1000 W with the interval of 100 W. The peak temperature inside the tube axis is shown in the Fig. 6 and it is evident that the smallest cavity (30 mm height) reaches the maximum temperature. The peak temperature inside the tubular reactor for 30 mm cavity is around 2100 K whereas the 60 mm cavity has the peak temperature of 1750 K for the same input heat flux. The temperature distribution at the centerline of the tube axis for 1 kW solar power as a function of the cavity height is shown in Fig. 7. It can also be observed that larger sized cavity reactor results in a more homogenous temperature distribution along the tube axis across the cavity length. This is due to uniform radiation heating with increased surface area. However, this homogeneity is obtained at the expense of maximum temperature limitation for larger sized cavities, and the tradeoff between uniform heating and maximum temperature needs to be considered for the specific case of study. This necessitates the design optimization of cavity size for the improved reactor performance in terms of uniform heating as well as temperature requirements.
—•—Cavity height: 30mm —Cavity height: 40mm Position along the center of the tube
—♦—Cavity height: 50mm—4 Cavity height: 60mm .30mm i 40mm «50mm .60mm
Fig. 6. Peak temperature in the tube Fig. 7. Axial temperature distribution in the tube
Size of cavity can be chosen from Figs. 6 and 7 based on the metal oxide to be studied and the temperature required for the reaction. As the interest of this study is to analyze ZnO thermal dissociation, the cavity size of 50 mm was selected for further analysis due to the reaction temperature requirement of greater than 1900 K.
3.5. Thermal study for the case of inert particle flow
In order to study the effect of particles heating on the temperature distribution inside the reactor, the case of inert particles injection was considered (for the 50 mm cavity height reactor without chemical kinetics). This approach provides understanding of the response of the system due to particles.
The inert particles were introduced into the computational domain as discrete particles using Lagrangian approach and the nominal mass flow rate was 1x10-6 kg/s. The particles were coupled with the flow field calculated by Eulerian approach for the force and energy balance. The input heat flux was 1000 W and inert particles were also participating to radiative heat transfer with particle-particle, particle-gas, wall-particle and wall-wall heat exchanges. The emissivity of ZnO particles was considered to be 0.8 [6].
The axial temperature distribution along the reactor tube is compared considering a flow with and without particles as shown in Fig. 8. The particle heating results in a more homogeneous temperature distribution along the reactor tube length corresponding to the cavity height. The temperature also rises more rapidly from the inlet to the cavity region. It can also be observed that the maximum temperature along the reactor tube was increased in the cavity region. This is caused by the particle radiation absorption effects predicted by the model.
The radial temperature distribution along the tubular reactor at cavity entry, cavity center and cavity exit is shown in Fig. 9. It shows that the radial temperature distribution is more uniform at the cavity exit and less uniform at the entry of the cavity. The temperature at the cavity exit is above 1900 K and it is homogenous, which is beneficial to uniform conversion. It can also be observed that the temperature at the reactor walls are higher than at the tube center as the reactor walls are exposed to radiation and they emit back into the reactor tube.
1500 Gas outlet
900 / / • • / Cavity a » WithoutZnO
• • / 1 *
600 • With ZnO (Inert)
300 Gas inlet $ ,
S. 1900
3 1S50 n
S 1800 a.
5 1750
1650 0.025
t * «Í!
0.05 0 -0.05
Position along the center of the tube
O.OS 0.035 0.04 0.M5
Radial position across the tube
* Cavity entry * Cavity center »Cavity exit
Fig. 8. Axial temperature distribution in the tube
Fig. 9. Radial temperature distribution along the tube
3.6. Sensitivity study in the case of reactive particles
In the continuing section, an injection of ZnO particles with reaction kinetics is considered in addition to the argon carrier gas. The two-phase reactive flow was simulated, enabling the species transport equations and reaction, in order to predict the flow patterns of both gas and solid phases, and reaction conversion. The Lagrangian approach (Discrete Phase Modeling) was considered for the particles and it is coupled with Eulerian phase flow field and species transport equations. The reaction kinetics was computed on the basis of the Arrhenius law. The kinetic rate of particle surface reaction, k, per unit area is defined as:
k = Ae( kt)
Kinetic data were taken from the literature [6] giving A = 1.4 x 106 kg m s and E = 328.5 kJ mol . The mass flow rate of the particles considered for the first simulation was 1x10-6 kg/s with the constant particle diameter of 1 ^m. The temperature contours at the y-z section plane is shown in Fig. 10. The heat transfer by conduction across
the solid components and radiation effects can be seen. At these conditions, the entire particle flow rate was converted into gaseous zinc with 100 % chemical conversion efficiency as shown in Fig. 11. This shows that the reactor performance was favorable at the specified flow and thermal conditions. The reactor could reach the required reaction temperature for an efficient reduction of zinc oxide.
Static temperature (K)
Particle temperature (K)
Fig. 10. Temperature contours in a vertical plane
Fig. 11. Particles streams colored by temperature
In order to study the effect of particle flow rate on conversion efficiency, the mass flow rate of ZnO particle was varied from 1x10-6 kg/s to 100x10-6 kg/s. The chemical conversion is plotted as a function of the particle flow rate in Fig. 12. As the particle mass flow rate increases, the chemical conversion decreases. This trend is not significant when the flow rate is changed from 1x10" kg/s to 5x10" kg/s. However, as it increases from 5x10 kg/s to 50x10"6 kg/s, the chemical conversion drops by more than 50%. At the flow rate of 100x10"6 kg/s, almost no particles were dissociated. This could be due to the large optical thickness of the particle flow that prevents radiation absorption in the bulk of the gas/solid flow. High particle loading implies higher radiation attenuation from the wall, as the optically"thicker particle cloud serves as a radiation shield, which limits particle heating.
Using the chemical conversion data and ZnO particle flow rate with corresponding solar power input, the solar-to-chemical energy conversion efficiency is calculated as:
,zno-dT)/Psoiar (2)
The solar-to-chemical energy conversion efficiency for different mass flow rates of ZnO particles is plotted in Fig. 13. It can be inferred that the system has a maximum efficiency when the flow rate is
50x10-6 kg/s for the
defined conditions of initial particle diameter (1 ^m) and solar power input. Even though the chemical conversion decreases with the increase in particle flow rate, the solar-to-chemical energy conversion efficiency shows an optimum value of about 11% for a given particle flow rate.
4. Conclusion
An indirectly irradiated tubular particle-laden flow solar reactor was successfully modeled using CFD analysis, and the temperature profiles were validated with experimental test results under similar operating conditions. A number of thermal simulations were performed to optimize the cavity dimensions and to match the existing solar power with the required high-temperature thermochemical reaction consisting of ZnO thermal dissociation. Based on the simulation results, the optimum cavity size of 50 mm was selected for further study based on targeted reaction temperature and capacity of the reactor. The magnitude of temperatures reached by the system in the
reaction zone is in agreement with simulation predictions, thus allowing further study of thermochemical reactions at more than 2000 K. The modeling of inert particle flow was then conducted to study the effect of particles on thermal response of the reactor. It was observed that particles heating and combined radiation heat exchange result in a more uniform temperature distribution in the axial and radial directions of the tube, especially in the zone corresponding to the cavity height.
Fig. 12. Conversion rate vs. particle flow rate
Fig. 13. Solar-to-chemical efficiency vs. particle flow rate
Finally, a reactive particle-laden flow was simulated by including chemical kinetics of ZnO dissociation and by using combined Lagrangian-Eulerian approach. The chemical conversions were determined as a function of the mass flow rate of ZnO particles and it was found that the increase in particle flow rate caused the decrease in chemical conversion. Solar-to-chemical energy conversion efficiency was also estimated and an optimum of about 11 % was obtained for a particle flow rate of 50x10-6 kg/s.
The performed simulations contributed to better understanding of the reactor thermal behavior and performance predictions including the reaction extent for varied inlet ZnO particle mass flow rates. The results will be used in the optimization of the solar chemical reactor design and operation, with further studies on performance assessment. The simulation procedure could also be used to predict the solar reactor performances when upscaling the concept to a higher solar power level.
Acknowledgements
Authors gratefully thank PROMES-CNRS for providing advices in reactor design and concepts, technical support and experimental reactor, equipment and facilities to conduct the on-sun testing.
This research was supported by the Government of Abu Dhabi to help fulfil the vision of the late President Sheikh Zayed Bin Sultan Al Nayhan for sustainable development and empowerment of the UAE and humankind.
References
[1] Steinfeld, A. Solar hydrogen production via a two-step water-splitting thermochemical cycle based on Zn/ZnO redox reactions. International Journal of Hydrogen Energy 2002; 27 (6), p.611-619.
[2] Palumbo, R., Keunecke, M., Moller, S., Steinfeld, A. Reflections on the design of solar thermal chemical reactors: thoughts in transformation. Energy 2004; 29, p.727-744.
[3] Xiao, L., Wu, S.-Y., Li, Y.-R., Advances in solar hydrogen production via two-step water-splitting thermochemical cycles based on metal redox reactions. Renew. Energy 2012; 41, p1-12.
[4] Gael Leveque, Stephane Abanades, Design and operation of a solar-driven thermogravimeter for high temperature kinetic analysis of solidgas thermochemical reactions in controlled atmosphere, Solar Energy 105, 2014; 225-235
[5] Meier, A., Ganz, J., Steinfeld, A. Modeling of a novel high-temperature solar chemical reactor. Chemical Engineering Science 1996; 51 (11), p.3181-3186.
[6] Abanades, S., Charvin, P., Flamant, G. Design and Simulation of a Solar Chemical Reactor for the Thermal Reduction of Metal Oxides: Case Study of ZnO Dissociation, Chem. Eng. Sci., 2007; 62(22), p. 6323-6333.
[7] Moller, Palumbo, R. Solar thermal decomposition kinetics of ZnO in the temperature range 1950-2400 K. Chemical engineering science 2001;56, p.4505-4515.
[8] Weidenkaff, A., Reller, A.W., Wokaum, A., Steinfeld, A. Thermogravimetric analysis of the ZnO/Zn water splitting cycle. Thermochemica Acta 2000; 359, p.69-75.
[9] Diver R. Receiver/reactor concepts for thermochemical transport of solar energy. Journal of Solar Energy Engineering 1987;109(3):p199-207.
[10] Erik Koepf, Suresh G. Advani, Aldo Steinfeld, Ajay K. Prasad. A novel beam-down, gravity-fed, solar thermochemical receiver/reactor for direct solid particle decomposition: Design, modeling, and experimentation, international journal of hydrogen energy 37, 2012;p16871-16887
[11] Steinfeld A. Solar thermochemical production of hydrogeneea review. Solar Energy 2005;78(5):603-15.
[12] Perkins C, Weimer AW. Solar-thermal production of renewable hydrogen. AIChE Journal 2009;55(2):286-93.
[13] Abanades S, Charvin P, Flamant G, Neveu P. Screening of water-splitting thermochemical cycles potentially attractive for hydrogen production by concentrated solar energy. Energy 2006;31(14):2805-22.
[14] Moller S, Palumbo R. The development of a solar chemical reactor for the direct thermal dissociation of zinc oxide. Journal of Solar Energy Engineering 2001;123(2):83-90.
[15] ANSYS. Fluent user documentation. www.fluentusers.com.
[16] Loutzenhiser P, Meier A, Steinfeld A. Review of the two-step H2O/CO2-splitting solar thermochemical cycle based on Zn/ZnO redox reactions. Materials 2010;3:4922-38.
[17] Steinfeld A, Schubnell M. Optimum aperture size and operating temperature of a solar cavity-receiver. Solar Energy 1993;50(1):19-25
[18] Schunk LO, Haeberling P, Wepf S, Wuillemin D, Meier A, Steinfeld A. A receiver-reactor for the solar thermal dissociation of zinc oxide. Journal of Solar Energy Engineering 2008;130(2):021009.
[19] Melchior T, Perkins C, Weimer A, Steinfeld A. A cavity-receiver containing a tubular absorber for high-temperature thermochemical processing using concentrated solar energy. International Journal of Thermal Sciences 2008;47(11): 1496-503.
[20] Muller R, Haeberling P, Palumbo RD. Further advances toward the development of a direct heating solar thermal chemical reactor for the thermal dissociation of zno(s). Solar Energy 2006;80(5):500-11.
