Optimization of power and hydrogen production from glycerol by supercritical water reforming

Depósito de Investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by Elsevier in Chemical Engineering Journal, Vol. 218, on February 2013, available at: https://doi.org/10.1016/j.cej.2012.12.035 © 2012 Elsevier. En idUS Licencia Creative Commons CC BY-NC-ND 1 Optimization of Power and Hydrogen Production from Glycerol by Supercritical Water Reforming F. J. Gutiérrez Ortiz, P. Ollero, A. Serrera, S. Galera Departamento de Ingeniería Química y Ambiental, Universidad de Sevilla Camino de los Descubrimientos, s/n. 41092 Sevilla, Spain Phones: + 34 95 448 72 68 / 60 e-mail: “Francisco Javier Gutiérrez Ortiz” [email protected] Abstract A process design is proposed and simulated for reforming glycerol using supercritical water aimed to produce maximum power and hydrogen in an energy self-sufficient system. The selected route takes advantage of the huge pressure energy of product gas just at the outlet of the reformer converting that into power by a turbine. The expanded product gas is conditioned by two water gas shift reactors and a pressure swing adsorption unit, so a hydrogen-rich gas stream is sent to a proton exchange membrane fuel cell to be converted into electrical energy and the pressure swing adsorption off-gas stream is used as fuel gas to provide the thermal energy required by the reforming process. The evaluation of the global efficiency of the process is carried out by energy and exergy analysis. Required glycerol feed concentration in aqueous solution was obtained for a self-sufficient process, both for pure and pretreated crude glycerol, at reforming temperatures from 600 to 1000 ºC and 240 atm. Thus, reforming and preheating at 800ºC and 240 atm, it was obtained a power of 1592 kW per ton/h of glycerol, with exergy and energy efficiencies of 33.8% and 35.8%, respectively. Keywords Reforming, supercritical water, glycerol, hydrogen, simulation 2 1. Introduction The rising surplus of biodiesel-derived crude glycerol from the transesterification process requires a further processing. The crude glycerol stream leaving the homogeneous basecatalyzed transesterification reactor normally contains glycerol, methanol, alkalies (catalyst and soap), methyl esters and water. Normal refining of glycerol involves an acid treatment to split the soaps into free fatty acid and salts. Fatty acids are not soluble in glycerol and are separated from the top and recycled to the process; then, if methanol is recovered by vacuum distillation, some salts remain with the glycerol, which is in aqueous solution and its purity is about 85 wt%; if methanol is present, glycerol purity may be reduced to about 75 wt%. A great deal of research has been carried out following a thermochemical route of glycerol valorization to obtain H2 or synthesis gas, by steam reforming [1-4], autothermal reforming [5-6] and aqueous phase reforming [7-10]. Many of these studies are centred on the development and characterization of catalysts. Supercritical water (SCW) is an emerging and promising medium to obtain hydrogen by reforming of glycerol, due to its relevant thermophysical properties such as a high capability to solubilize gaseous organic molecules and high diffusivity, among others [11-14] with only a few papers published regarding with the glycerol [15-17]. SCW is extremely reactive and, it may be possible to perform the process in the absence of a catalyst, although this premise requires to be experimentally verified Hydrogen is very attractive as a clean fuel for proton exchange membrane fuel cells (PEMFC), which are more efficient than combustion engines and have zero carbon emissions. In order to lessen the CO content and to increase the hydrogen content in the reformate stream, two water gas shift reactors (WGS), at two temperature levels (HWGS and LWGS), are normally used. As a next stage, a pressure swing adsorption (PSA) unit or a preferential 3 oxidation reactor (PrOx) may be used to reduce the carbon monoxide concentration to less than 10 ppm to avoid the poisoning of the PEMFC catalyst. SCW reforming is a new technology, which needs a better understanding of the fundamental phenomena for a suitable reactor and process design. Since it is in its beginnings, research on the key parameters influencing the overall process efficiency is essential. Some previous works have been carried out focused on fundamental aspects of the process of glycerol reforming using supercritical water such as a thermodynamic study [15, 16, 18 , 19], but there is only one published paper that deals with a methodology based on the exergy concept to assess the process [20]. An exergy analysis quantifies the lost efficiency in a process due to the loss in energy quality by evaluating the irreversibilities of the process, once the suitable control volume is selected, providing thus where the process can be improved. This paper specifically proposes a route of achieving maximum electrical power and efficiency in a process of glycerol reforming using supercritical water from the huge pressure energy of the reformate product by means of an expander and from pure hydrogen by a PEM fuel cell. In the process, a trend of different stages is included for the conditioning of the product gas (two WGS reactors and a PSA unit), once expanded to produce electrical energy, so the final purified hydrogen-rich gas can be used in a fuel cell. Moreover, the thermal energy required by the process to achieve an energy self-sufficient system is obtained by burning the PSA off-gas, so an external fuel is not needed as a heat source for performing the process. Thus, the proposed process is assessed by an energy and exergy methodology . Therefore, in the present study, the methodology described in the previous paper [20] is applied to an energy integrated process, and an optimization procedure is developed to achieve maximum power in the process as well as to reach a maximum exergy efficiency. These are the main aspects to be claimed within the paper novelty. 4 2. Process Design and Simulation The process design and simulation have been performed by coupling the reforming process with a turbine and a PEM fuel cell, using energy integration to improve the energy use. Figure 1 shows the process flow-sheet. Appendix A gives details of the exergy analysis used to assess the process. Regarding with the energy integration, a number of exchangers were suitably located following a strategy based on minimizing the lost work or exergy losses, so small temperature-driving forces must be achieved, by using countercurrent flow and small temperature approaches at the ends of the exchangers. A 'from inside to outside' approach was used to design the heat exchanger network, in such a way that the internal streams close to each other at high thermal levels are the first to contact each other (the high-temperature hot streams heat the high-temperature cold streams) and the external more separated streams at low thermal levels are the last to contact each other (the low-temperature hot streams warm up the low-temperature cold streams), by taking the reformer as a central point, where the maximum temperature is required. Then, several possible arrangements of heat exchangers and streams pairing were performed in alternative flow sheets, and the redistribution of exchanger duties was assessed. The flow sheet that provides the best heat integration performance, i.e., minimize the use of external utilities and maximizing internal heat flows into the system, is finally selected. Besides, a series of gas-conditioning stages are added to the process for achieving a hydrogen-rich gas stream, which is fed into a PEMFC. A glycerol-water mixture is first pumped and then heated up to a desired temperature before entering the supercritical water reforming reactor. The heating is performed by means of five heat exchangers taking thus heat from hot streams leaving first the reformer-furnace, next the reforming reactor and then the two WGS reactors, which need to be cooled. In the 5 reformer, glycerol is converted at temperature and pressure previously established. The reforming reactor is simulated as a Gibbs reactor, where the product composition and the heat of overall reaction are calculated under conditions that minimize the Gibbs free energy. The furnace-combustor where the PSA off-gas stream is burnt is simulated as a stoichiometric reactor, which transfer the heat released from the combustion reaction to the reforming reactor to operate at the specified temperature. The flow-rate of the gas to be burnt is computed as that needed to match the heat flow required in the reforming reactor. A fan is used to enter the air required in the furnace-combustor, so there is an O2 content of 3%vol in the flue-gas. The product gas leaving the reformer is expanded in a turbine converting thus the pressure energy into electrical energy, and then conditioned. By two adiabatic shift reactors in series with a cooler between them to remove the heat of reaction from the exothermic water gas shift reaction, the concentration of carbon monoxide produced in the reformer is reduced and the hydrogen content is increased. The first-stage reactor (HWGS) operates at 350 ºC, and the second (LWGS) operates at 200 ºC, which proceed on exothermal reaction at equilibrium, simulated by two adiabatic equilibrium reactors. The co-existence of CO, CO2 or CH4 has been reported to have a negative influence on the PEMFC [21], so a high purity hydrogen stream is beneficial to the process efficiency. The carbon monoxide concentration can be reduced to less than 10 ppm by means of a PSA unit [22]. Against the selective oxidation of CO (PrOx), which can be also used for this purpose, the PSA is probably the most mature technology for CO removal from the reformate gas aside from it is able to remove all the other gases present to obtain a very pure hydrogen at the PSA outlet. In addition, while PrOx is mainly used for a small scale, the PSA may be used for small-to-medium scale plants in stationary applications, such as the studied process. Since water vapor is strongly adsorbed in the PSA unit, the amount of water present in the gas must be reduced as much as possible before entering this unit to minimize the loss of its adsorption 6 capacity. For this reason, the gas stream coming from the LWGS reactor is cooled down to 35 °C and the condensed water is separated from the gas stream by means of a gas-liquid separator. Furthermore, operating at low temperature allows to reduce the bed volume, and thus less of the H2 is lost during the bed regeneration steps [23]. The PSA process is based on sorbents that preferentially adsorb impurities at high partial pressures and then desorb them at a lower partial pressure (during the regeneration steps of the PSA cycle). Within the PSA unit, the adsorbent regeneration and desorption of components retained take place in the depressurization. Therefore, the removal of species adsorbed is done by total pressure reduction. This is simulated by disposing a valve at the PSA off-gas stream. On the other hand, the purified hydrogen stream exits the PSA bed at a pressure close to the PSA inlet pressure. Likewise, another valve has been disposed at the hydrogen stream to enter the PEMFC at atmospheric pressure. Increasing the pressure shifts the curves towards higher recovery although at pressures greater than 18 atm the change will be very small [23]. In addition, it is not convenient to operate the PSA unit at too high pressure in order to obtain more power in the expander. Therefore, the PSA unit has been simulated in a simplified way, by an ideal separator, although working at a realistic pressure and temperature (15 atm and 35ºC, respectively), and the pressure swing adsorber (PSA) is assumed to give a high enough purity (99.999%) with a hydrogen recovery of 80%, which are feasible values [24-25]. Thus, to yield a high purity hydrogen stream, some of the hydrogen is lost in the off-gas stream from the PSA, which is used as fuel in the furnace, and its energy is still used for heating the reforming reaction, reducing thus the amount of glycerol sent to the reformer to supply the required heat. The hydrogen-rich gas stream leaving the PSA unit is then heated up to 80 °C before it enters the PEM fuel cell to improve its performance. A simple PEM fuel cell process has been included in the scheme as a subsystem. Water is added to the air stream and the water mass 7 flow rate is calculated to reach a relative humidity of 35% at 80ºC, inside the fuel cell. Similarly, water is also added to the hydrogen stream to reach a 100% relative humidity at 80ºC. Humidification of the entering streams is necessary to maintain membrane saturation and to prolong the life of the membrane. Water diffuses through the electrodes and membrane as vapor. Hydrogen coming from the reforming process diffuses through the negatively charged electrode (anode) to the membrane and reacts with oxygen coming from the positively charged electrode (cathode) and diffusing through the cathode electrode. Thus, the electrochemical reaction occurs in the membrane electrode assembly, which is simulated as an isothermal reactor (PEMFC in the scheme). The pure hydrogen obtained in the PSA enables to operate the fuel cell without splitting a part of the hydrogen in the anode off-gas, which only consists of water vapor because practically no net hydrogen exits the cell at steady state. For the simulation of the PEM fuel cell, anode and cathode channels are represented as the streams entering the reactor separately, and leaving it in a single stream. A fraction of the chemical energy of the electrochemical reaction is converted into electrical power, another fraction is internally used to complete air and hydrogen humidification process (some integrated fuel cell and membrane humidifier systems are being developed [26]) and to heat all the reactants to 80ºC, and the rest of the energy is released as heat leaving the fuel cell body toward, respectively, the cooling water and the surroundings. The air enters the fuel cell to have a flow rate of oxygen twice the flow rate required for stoichiometric oxygen-hydrogen reaction. A fan makes it possible to feed the air into the cathode channel of the fuel cell. Likewise, the waste heat produced in the stack module is removed through the cooling loop, which usually consists of a radiator, cooling pump, radiator fan. In this study, only the pump has been considered, and the cooling water flow rate is calculated to obtain a maximum increase of 15ºC through the heat exchanger. A description of the energy and exergy analysis of the PEMFC is given in the Appendix B. 8 Specifications of the elements used in the simulation are shown in Table 1. The computation has been made with the aid of AspenPlus™ version 2006.5 (Aspen Technology, Inc., USA). The used thermodynamic method has been the predictive Soave- Redlich-Kwong (PSRK), as the most suitable [19]. Two feeds have been studied: pure and pretreated crude glycerol. The main components of crude glycerol considered were glycerol and methanol, as other researchers did [27], i.e., the starting point is taken just after separating the methyl ester fraction and performing the acid treatment, but before removing the methanol fraction. Thus, the pretreated crude glycerol feed consists of glycerol with variable methanol content (10, 20 and 30 wt%). Hydrogen, carbon monoxide, carbon dioxide, methane, ethane, propane, water, methanol, ethanol, glycerol and oxygen as well as pure carbon (in solid phase) were added manually, as they were considered the possible species from the SCW reforming of glycerol. The simulation did not predict coke formation for any of the experimental conditions in this paper, and any other compound has a mole fraction lower than 10-12. The contents of ethane, propane and ethanol were also so low that they were not included in the analyses. Likewise, for all operating conditions simulated glycerol and methanol conversion were always 100%, at equilibrium condition. Based on reforming reactions (1) and (2), two hydrogen yields are computed using equations (3) and (4): Glycerol reforming: C3H8O3 + 3H2O = 3CO2 + 7H2 [1] Methanol reforming: CH3OH + H2O = CO2 + 3H2 [2] 15 this study based on the assumption of reaching chemical equilibrium. Thus, at feed concentration of 30 wt% glycerol, Byrd et al. [15] obtained a product gas composition H2/CO/CO2/CH4 of 47.2/3.2/34.0/15.6, at a temperature of 800ºC and a pressure of 241 bars, over Ru/Al2O3 catalysts in a tubular fixed-bed flow reactor. In the present study, the product gas composition H2/CO/CO2/CH4 was 47.5/11.1/25.4/15.9, when using a glycerol feed concentration of 26.5 wt%, at 800ºC and 240 atm. At 15 wt% glycerol, Byrd et al. obtained a product gas composition H2/CO/CO2/CH4 of 57.9/0.6/30.9/10.7, and in this study, at 14 wt% glycerol, the product gas composition H2/CO/CO2/CH4 was 59.4/9.8/24.4/6.4, using the same pressure and temperature. Experimental yields of CO and CO2 are somewhat different (smaller and higher, respectively) from the predicted values at equilibrium, probably because the water-gas shift reaction takes place in presence of the used catalyst. 4. Conclusions A SCW reforming of glycerol process was designed and simulated, by including energy integration and conditioning stages (HWGS, LWGS and PSA), to obtain a hydrogen-rich gas for a PEMFC. Maximum net power production by an expander (718.6 kW @ 800ºC, 240 atm), once consumption in other machines was discounted, and by a PEMFC (873.4 kW @ 80ºC, 1 atm), once pure hydrogen production was maximized from the PSA unit (58.6 kg/h), was achieved by feeding 1000 kg/h of glycerol in an aqueous solution (26.5 wt%). The combustion of the PSA off-gas provided the heat required by the reforming process. Nomenclature B  Exergy flow (kW) H  Enthalpy flow (kW) WL  Exergy flow loss (kW) 16 m  Mass flow rate (kg/s) Q  Heat flow (kW) S  Entropy flow (kJ/(s K)) T Temperature (K) W  Power (kW) Greek  Efficiency (%) Subscripts Gly Glycerol in referred to some entering the control volume considered (unit, subsystem or system) MeOH Methanol net referred to the net power of the turbine out referred to some leaving the control volume considered (unit, subsystem or system) PEMFC Proton exchange membrane fuel cell PSA Pressure Swing Adsorption WGS Water Gas Shift Appendix A. Exergy Analysis The studied process is continuous and steady-state; it is open to interactions with the surroundings, and potential and kinetic energies are not considered. Likewise, the reactors are at equilibrium condition. The dead-state conditions (of surroundings) are 25ºC and 1atm. 17 All flows (mass, energy, exergy) entering the system are considered as positive, while flows exiting are indicated by a negative sign. The terms of energy and exergy flows refer to energy and exergy per unit of time (power). Likewise, the terms of heat flow and work flow will be used, although power will be also used for the latter. By combining energy and entropy balances, the lost work flow (exergy flow) or lost power ( WL ) for the overall process can be computed as follows:            inout j j inout inout Q T T BWWL   0 1 [A1] where: irr STWL   0 is the lost work or exergy losses due to the increase in entropy of the Universe. W  is the power relative to mechanical shaft work and electrical work as well as work resulting from the expansion or contraction of the system against the surroundings, computed as follows:    h k ink q k outk inout WWW 11 ))  [A2] Electricity and mechanical work are perfectly convertible and, for these forms, exergy contents equals to the energy content. B is the exergy flow of a stream, which is calculated via: STHB   0  [A3] where H  and S  are the total stream enthalpy and entropy flows, respectively, evaluated at the corresponding state, i.e., the liquid, vapor or supercritical state. For all the streams entering (J) and leaving (K) the system, it results 18    K s outs J s insinout BBB 11 ))  [A4] j Q  is the heat flow transferred to the system (positive) or by the system (negative) at Tj, which is the temperature of the jth heat source or sink outside the system. Only a part (given by the Carnot factor) of the heat transferred is a useful work, so:                           m j inj in j p j outj out j inout j j Q T T Q T T Q T T 1 0 1 00 )1)11  [A5] In the reforming process, the lost power is the net sum of exergy flows of inlet and outlet streams ( B ) minus the net power produced in the system ( net W ) since the overall reforming process is adiabatic. However, since the fuel cell system is not adiabatic the heat flow transferred to the surroundings by this subsystem must be taken into account (see Appendix B). Therefore:          cooling FC inout system outin Q T T WWBBWL   0 1)( [A6] As the final goal is to get maximum power through the huge pressure energy of the reformate gas, by means of an expander, plus the conversion of hydrogen energy, by means of a fuel cell, the exergy efficiency of the process to be inspected is the following: WLW W net net exergy      [A7] The term net W  includes both the net power of the turbine, once the power by other machines is discounted, and the fuel cell. The term of Lost Power includes internal and external exergy losses, both in the overall reforming process and the PEM fuel cell system. In the term of Lost Work (flow), both internal and external exergy losses are considered. The former are due to the system itself and their units and the latter is a wasted exergy no 19 longer usable by subsequent processes due to dissipations to the environment as heat losses in individual process units (not included in this study, where a perfect thermal isolation has been assumed), cooling water or smokestack effluents. Anyway, the cooling water heated in the process could be used as a local source of heating, i.e., in the plant and other surrounding plants, next to the concept of district heating. As a result, a combined heat-power exergy efficiency (h-p) may be defined as follows:         EX watercoolingnet EX watercoolingnet ph BWLW BW   [A8] where watercooling inoutwatercooling BBB   )(  [A9] Finally, for pure glycerol, the energy efficiency is defined as GlyGly net energy LHVm W     [A10] where Gly m is the mass flow rate of glycerol fed into the system and Gly LHV is the Lower Heating Value of glycerol (16 MJ/kg). In case of using pretreated crude glycerol, the energy efficiency is calculated as MeOhMeOHGlyGly net energy LHVmLHVm W      [A11] where MeOH m is the mass flow rate of methanol fed into the system and MeOH LHV is the Lower Heating Value of methanol (19 MJ/kg). More information can be found elsewhere [20]. Appendix B. Energy and exergy analysis of the PEMFC The generic energy balance in the stack is the following: 20    coolingPEMFCoutin QWHH   [B1] where in H  is the enthalpy flow of the reactants gases in, out H  is the enthalpy flow of the unused reactants gases and the product gases leaving the fuel cell (all the water is as vapour), PEMFC W  is the electricity generated, cooling Q  is the heat flow taken away from the stack by active cooling. The stack is assumed to be well thermally isolated. Assuming negligible potential and kinetic energy effects on the fuel cell electrochemical process, the total exergy transfer per unit mass of each reactant and product consists of the combination of both physical and chemical exergies. Thus, using Equation A1, the exergy loss of the PEM fuel cell unit has been computed as follows: PEMFCcooling PEMFC inoutPEMFC WQ T T BBWL             0 1 [B2] where TPEMFC is the temperature inside the PEM fuel cell. To obtain the exergy losses of this subsystem, the lost work of the fan, pump and cooler must be taken into account, and so it has been computed in the analyses, in a similar way to other similar units. Specifically, to calculate the unusable waste exergy of the PEM fuel cell through the cooler, the following equation has been used, using Equation A1: cooling PEMFC inoutCooler Q T T BBWL           0 1 [B3] where the heat source is the PEM fuel cell at TPEMFC and the heat sink is the cooling water through the cooler. The electric power generated by the fuel cell is calculated as: 22 HHFCPEMFC LHVmW   [B4] 21 where 2 H m  is the mass flow rate of hydrogen that reacts in the fuel cell, 2 H LHV is the Lower Heating Value of hydrogen and FC  is the electrochemical efficiency of the fuel cell. For the analysis, it is assumed that to obtain a power of 1 kW in the PEMFC unit, the required molar flow rate of hydrogen is 33.3 mol/h [28], corresponding to an efficiency of the fuel cell stack of about 0.45 [29]. Acknowledgment This research is supported by the Science and Technology Ministry of Spain under the research project ENE2009-13755, as a Project of Fundamental Research inside the framework of the National Plan of Scientific Research, Development and Technological Innovation 2008-2011. References [1] M. Slinn, K. Kendall, C. Mallon, J. Andrews, Steam reforming of biodiesel by-product to make renewable hydrogen, Bioresource Technol 99 (2008) 5851-5858 [2] S. Adhikari, S. Fernando, A. Haryanto, Kinetics and Reactor Modeling of Hydrogen Production from Glycerol via Steam Reforming Process over Ni/CeO2 Catalysts, Chem Eng Technol 32 (2009) 541–547 [3] B. Dou, G.L. Rickett, V. Dupont, P.T. Williams, H. Chen, Y. Ding, M. Ghadiri, Steam reforming of crude glycerol with in situ CO2 sorption, Bioresource Technol 101 (2010) 2436–2442 [4] V. Nichele, M. Signoretto, F. Menegazzo, A. Gallo, V. Dal Santo, G. Cruciani, G. Cerrato, Glycerol steam reforming for hydrogen production: Design of Ni supported catalysts, Appl Catal B-Environ 111–112 (2012) 225–232 22 [5] H. Wang, X. Wang, M. Li, S. Li, S. Wang, X. Ma, Thermodynamic analysis of hydrogen production from glycerol autothermal reforming, Int J Hydrogen Energ 34 (2009) 5683-5690 [6] S. Authayanun, A. Arpornwichanop, W. Paengjuntuek, S. Assabumrungrat, Thermodynamic study of hydrogen production from crude glycerol autothermal reforming for fuel cell applications, Int J Hydrogen Energ 35 (2010) 6617-6623 [7] R.R. Davda, J.W. Shabaker, G.W. Huber, R.D. Cortright, J.A. Dumesic, A review of catalytic issues and process conditions for renewable hydrogen and alkanes by aqueous phase reforming of oxygenated hydrocarbons over supported metal catalysts, Appl Catal B 56 (2005) 171-186 [8] G. Wen, Y. Xu, H. Ma, Z. Xu, Z. Tian, Production of hydrogen by aqueous-phase reforming of glycerol, Int J Hydrogen Energ 33 (2008) 6657-6666 [9] A.O. Menezes, M.T. Rodrigues, A. Zimmaro, L.E.P. Borges, M.A. Fraga, Production of renewable hydrogen from aqueous-phase reforming of glycerol over Pt catalysts supported on different oxides, Renew Energ 36 (2011) 595-599 [10] Y. Guo, X. Liu, M.U. Azmat, W. Xu, J. Ren, Y. Wang, G. Lu, Hydrogen production by aqueous-phase reforming of glycerol over Ni-B catalysts, Int J of Hydrogen Energ 37 (2012) 227-234 [11] P.E. Savage, Organic chemical reactions in supercritical water, Chem Rev 99 (1999) 603-622 [12] Y. Matsumura, T. Minowa, B. Potic, S.R.A. Kersten, W. Prins, WPM van Swaaij, B. van de Beld, D.C. Elliott, G.G. Neuenschwander, A. Kruse, M.J. Antal Jr., Biomass 23 gasification in near–and super-critical water: status and prospects, Biomass Bioenerg 29 (2005) 269–292 [13] A. Kruse, Hydrothermal gasification of biomass, J Supercrit Fluid 47 (2009) 391–399 [14] Y. Guo, S.Z. Wang, D.H. Xu, Y.M. Gong, H.H. Ma, X.Y. Tang, Review of catalytic supercritical water gasification for hydrogen production from biomass, Renew Sust Energ Rev 14 (2010) 334–343 [15] A.J. Byrd, R.B. Gupta, K.K. Pant, Hydrogen production from glycerol by reforming in supercritical water over Ru/Al2O3 catalyst, Fuel 87 (2008) 2956-2960 [16] F.A. Voll, C.C. Rossi, C. Silva, R. Guirardello, R.O. Souza, V.F. Cabral, L. Cardozo- Filho, Thermodynamic analysis of supercritical water gasification of methanol, ethanol, glycerol, glucose and cellulose, Int J Hydrogen Energ 34 (2009) 9737-9744 [17] A. May, J. Salvadó, C. Torras, D. Montané, Catalytic gasification of glycerol in supercritical water, Chem Eng J 160 (2010) 751–759 [18] H. Tang, K. Kitagawa. Supercritical water gasification of biomass: thermodynamic analysis with direct Gibbs free energy minimization, Chem Eng J 106 (2005) 261–267 [19] F.J. Gutiérrez Ortiz, P. Ollero, A. Serrera, A. Sanz , Thermodynamic Study of the Supercritical Water Reforming of Glycerol, Int J of Hydrogen Energ 36 (2011) 8994- 9013 [20] F.J. Gutiérrez Ortiz, P. Ollero, A. Serrera, S. Galera, An energy and exergy analysis of the supercritical water reforming of glycerol for power production, Int J of Hydrogen Energ 37 (2012) 209-226 [21] A. Unemoto, A. Kaimai, K. Sato, T. Otake, K. Yashiro, J. Mizusaki, T. Kawadab, T. Tsunekic, Y. Shirasakic, I. Yasuda, Surface reaction of hydrogen on a palladium alloy 24 membrane under co-existence of H2O, CO, CO2 or CH4, Int J Hydrogen Energ 32 (2007) 4023–4029 [22] M.J. Castaldi, Removal of trace contaminants from fuel processing reformate: preferential oxidation (Prox), in: Liu K, Song C, Subramani V (Eds.), Hydrogen and syngas production and purification technologies, Wiley-AIChE., New Jersey, 2010, pp. 329–356. [23] D.D. Papadias, S. Ahmed, R. Kumar, F. Joseck, Hydrogen quality for fuel cell vehicles – A modeling study of the sensitivity of impurity content in hydrogen to the process variables in the SMR–PSA pathway, Int J Hydrogen Energ 34 (2009) 6021-6035 [24] S. Sircar, T.C. Golden, Pressure swing adsorption for hydrogen production, in: Liu K, Song C, Subramani V (Eds.), Hydrogen and syngas production and purification technologies, Wiley-AIChE., New Jersey, 2010, pp. 414–450. [25] S. Sircar, T.C. Golden, Purification of hydrogen by pressure swing adsorption, Sep Sci Technol 35 (2000) 667–687. [26] D. Chen, H. Peng, A thermodynamic model of membrane humidifiers for PEM fuel cell humidification control, Transactions of the ASME 127 (2005) 424 -432 [27] T. Valliyappan, D. Ferdous, N. Bakhshi, A.K. Dalai. Production of hydrogen and syngas via steam gasification of glycerol in a fixed-bed reactor, Top Catal 49 (2008) 59-67 [28] Y. Choi, H.G. Stenger, Kinetics, simulation and insights for CO selective oxidation in fuel cell applications, J Power Sources 129 (2004) 246–254 [29] Y. Hou, M. Zhuang , G. Wan, The analysis for the efficiency properties of the fuel cell engine, Renew Energ 32 (2007) 1175–1186 Figure 1 WORK GLY WATER 01 02 CW 04 G 1 G 2 08 CW02 14 15 CW 03 SG SW CW 05 AIRIN HEAT CW01 AIR M IX1 P1 TURB SEP FURNACE P2 FAN1 HWGS LWGS REFORMING REACTOR Tail gas Cooling w ater Hot w ater Separated w ater HE01 HE05 HE06 HE07 M IX2 SPLIT2 PSA H 2 TG2 H2W TG3 TGIN H2FC2 R 07 09 HE03 HE04 05 06 11 HE02 03 04 13 10 12 HE08 AIRC G 3 W T UR W 3 W 4 H2IN AIR2C W 2 AIR2IN W 1 H P E M HEAT CW1PEM CW2PEM CW3PEM AIR 2 GS P3 M IX3 M IX P4 SPL-PEM P5 COOLER FAN PEMFC V1 V2 TG1 HE09 H2FC1 G4 W PEM Q PEM WORK to surroundings (a) 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 15 20 25 Glycerol feed concentration (wt%) Fraction of H2 sent to the Furnace 0 5 10 15 20 25 Mass flow-rate of H2 sent to PEMFC (kg/h) Hydrogen fraction fed to the Furnace at 700ºC Hydrogen sent to PEMFC at 700ºC (b) 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 15 20 25 30 35 Glycerol feed concentration (wt%) Fraction of H2 sent to the Furnace 0 5 10 15 20 25 30 35 Mass flow-rate of H2 sent to PEMFC (kg/h) Hydrogen fraction fed to the Furnace at 800ºC Hydrogen sent to PEMFC at 800ºC (c) 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 15 20 25 30 35 40 Glycerol feed concentration (wt%) Fraction of H2 sent to the Furnace 0 5 10 15 20 25 30 35 40 Mass flow-rate of H2 sent to PEMFC (kg/h) Hydrogen fraction fed to the Furnace at 900ºC Hydrogen sent to PEMFC at 900ºC (d) 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 15 20 25 30 35 40 45 50 Glycerol feed concentration (wt%) Fraction of H2 sent to the Furnace 0 5 10 15 20 25 30 35 40 45 50 Mass flow-rate of H2 sent to PEMFC (kg/h) Hydrogen fraction fed to the Furnace at 1000ºC Hydrogen sent to PEMFC at 1000ºC Figure 2 (a) 0 200 400 600 800 1000 1200 1400 1600 15 20 25 30 Glycerol feed concentration (wt%) Power (kW) Turbine gross power Turbine net power PEMFC power Total power Reforming temperature = 700ºC (b) 0 200 400 600 800 1000 1200 1400 1600 15 20 25 30 35 Glycerol feed concentration (wt%) Power (kW) Turbine gross power Turbine net power PEMFC power Total power Reforming temperature = 800ºC (c) 0 200 400 600 800 1000 1200 1400 1600 1800 2000 15 20 25 30 35 40 Glycerol feed concentration (wt%) Power (kW) Turbine gross power Turbine net power PEMFC power Total power Reforming temperature = 900ºC (d) 0 200 400 600 800 1000 1200 1400 1600 1800 2000 15 20 25 30 35 40 45 50 Glycerol feed concentration (wt%) Power (kW) Turbine gross power Turbine net power PEMFC power Total power Reforming temperature = 1000ºC Figure 3