You are on page 1of 9

International Conference on Renewable Energies and Power

Quality (ICREPQ’09)
European Association for the
Development of Renewable Energies,
Environment and Power Quality Valencia (Spain), 15th to 17th April, 2009

Fitting regression model and experimental validation
for a high pressure PEM electrolyzer

M. Santarelli, P. Medina, M. Calì

Dipartimento di Energetica, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy

e-mail: massimo.santarelli@polito.it, pablo.medina@polito.it

Abstract. Hydrogen production is the main obstacle today 1. Introduction
to develop a real future hydrogen economy. Research has
focused many efforts in extracting it from clean and renewable Commonly used electrolyzers (usually alkaline, which
sources. Different processes are analysed: photolysis,
thermochemical cycles, algae, etc; these processes are still far
require frequent maintenance that includes disposal and
from practical use. Electrolysis has represented the most studied replacement of the highly caustic electrolyte) generate
and experimented area for obtaining hydrogen without hydrogen at relatively low pressures. Furthermore,
employing fuel cracking. Nevertheless, for its practical storage, hydrogen has still to be compressed for storage. One
the hydrogen produced at low pressure needs to be solution to this problem is to use a compressor to increase
mechanically compressed, with a high consumption of electric the hydrogen pressure. However, the energy investment
power. Advanced materials and improved design allow to required to compress hydrogen, as well as the
obtain hydrogen from electrolysis directly at medium-high maintenance of hydrogen compressors, makes this option
pressure (70 bar) with no need of mechanical compression very expensive for large-scale application of this
stages. This single-step process is more efficient than the two-
step electrolysis + mechanical compression process. In this
technology. The dependence on a mechanical
paper the authors display the experimental results obtained with compression stage penalizes the whole efficiency, and
a prototype of high pressure PEM electrolyzer manufactured by becomes more relevant when elevated pressures are
Giner Electrochemical Systems LLC, including the description required.
of the test bench for the experimental characterization. The The alternative is to compress hydrogen
experimental design, based on Design of Experiments electrochemically, pulling protons across the electric
techniques, studied the effect of the main operation factors field to a pressure-controlled cathode camera. It has been
(temperature, pressure, water flow) at different levels of power observed that the energy that protons need to move
load, presenting a regression model of the electrolyzer voltage across a pressure gradient is much smaller than the power
as a function of the operating factors, at different values of the
electric load.
supplied to a multi-stage compressor. Some studies [1]
reveal that, at elevated pressures, a considerable fraction
of water vapour present in the produced hydrogen gas
Keywords undergoes condensation. This reduces the cost of
desiccating the gas, because at elevated pressures the
gases are virtually free of moisture. There is one more,
hydrogen production, PEM electrolyzer, high pressure implicit, advantage of electrolysis at elevated pressures,
electrolysis, design of experiments namely, the electrolysis can be run at temperatures above
100°C, which substantially reduces the cell voltage. Of

based on Design of Experiments technology.5%vol Maximum O2 in H2 content 0. prod. Görgün et al. also hydrogen production is 0.4 A/cm2) presenting their results with steady state polarization Voltage at max. The main characteristics of et al. hydrogen Table 1. operational complexity. including the description of the test drawbacks. The PEM electrolyzer uses a solid electrolyte techniques. Electrolyzer operational characteristics embrittlement. compounding a 23 experimental plot. a regression model of the electrolyzer voltage is electrolysis. prod. The results of the sustainable.course. current 0.89 bar) comparison of total energy consumption for electrolysis H2 prod. while oxygen is storage dynamics are presented. Few Authors refer to experimental research for PEM Finally. It has four fluid obtained with a prototype of high pressure PEM connections (water in. No caustic alkaline or acidic fluid levels of power load. rate 25 V measurements and electrochemical impedance Power at max. Additional advantages of PEM processed using a factorial design (Yates´ technique). displaying the trend of the stack high pressure.1 kg/h (1. they have developed [1] a mathematical model and 2. Roy et al. and its possible applications including the obtained polarization curves for different catalyst compositions 2. water flow) at different the electrolyzer. with an analytical study of the effect of enthalpy and Data Value Gibb’s free energy variation with stack temperature or Temperature 50÷60°C pressure. electro-osmotic drag pressure (max 3 bar). which allow to model the polarization curves of a PEM-based electrolyzing cell on the basis of The stack utilised for the tests is the prototype Giner GS- literature and experimental data.55 Nm3/h). pressure. and on the cell overvoltages. Schug [4] describes the operational presented. with In this paper. [3] make a general overview of PEM electrolysis different pressures. making great improvements in efficiency. It has a compact size. and cost-effective technology for generating. [8] introduce the possibilities of Anode pressure 0÷3. too: the effect of pressure on the process bench for the experimental characterization.5 l/min for different pressures and performance duration against Water resistivity at 25°C > 1 MΩ cm time. Marshall et al.4 Nm3/h at various operating pressures up to 700 atm. [5] 10 electrolyzer provided by Giner Electrochemical describe a dynamic model for PEM electrolyzer based on Systems LLC (US).1 Stack relevant software. and oxygen at low cathode. including water transport. and considering also kW.6 kinetics for electrodes polarization. [7] have made a detailed comparison between the electrolyzer are detailed in Table 1. The studies and papers [2. become more severe. even at different conditions. atmospheric and high pressure electrolysis efficiency. displaying the results of efficiency Water flow > 5. Moreover. Barbir et al. including practical considerations such as corrosion. Several polarization curves have been obtained in The trend today for future electrolysis devices. rate 5. Power developed a simple model based on Butler–Volmer loaded by the stack at maximum production rate is 5. tests have been displayed comparing the effect of the compressing. also. the Authors display the experimental results an external diameter of about 25 cm. rate 224 A (at 1. Grigoriev optimize the efficiency of the hydrogen production at et al. PEM electrolysis is potentially a simple. The thermodynamics. causes an objective of this paper is to analyse the influence of increase in the theoretical and actual voltage of water different operation factors on the voltage (and decomposition. Maximum hydrogen output and diffusion through the membrane. and the requirements pressure PEM electrolyzer at different levels of stack imposed on the electrolyzer strength and air tightness current (or hydrogen production). [10] have performed a Voltage at recommended current 22 V (at 6. as it can be seen in Figure 1. hydrogen out electrolyzer manufactured by Giner Electrochemical and nitrogen in) and two electrical poles (anode-cathode) . [6] also have produced in a half volume rate (0.1 Nm3/h). providing characteristics of high pressure electrolysis.5 bar integrating a electrolyzer with a PV array in a isolated or Cathode pressure 0÷35 bar grid connected layout.3] demonstrate the interest in this experimental design. dynamic response and cost. The electrolyzer is made of 12 planar cells.5%vol catalytic materials used for the oxygen evolving Maximum H2 production rate 1.1 Nm3/h electrode. the electrolyzer design consequently the electric power) supplied to a high becomes more complicated. [9] have developed new electro Maximum H2 in O2 content 0. water+oxygen out. The experimental data has been electrolyte is required. factors in terms of single or combined effects. many voltage against current loaded from the electrolyser. Current at max. with a the transport resistance in the polymer electrolyte. Choi et al. goes through PEM technology. rate at recom. electrolysis over alkaline electrolysis include lower using the analysis of variance (ANOVA): the factors parasitic energy losses and higher purity hydrogen have been selected in two extreme operational levels output. analyzing the an analytical expression of the behaviour of the device at responses of the main systems parameters and how the different operating conditions. studied the effect of the main operation membrane that can be expected to increase the lifetime of factors (temperature. and storing hydrogen.6 kW spectroscopy. electrolysis at elevated pressure has some Systems LLC (US). Description of the equipment and different operation temperatures and pressures. prod. The final aim is to material problems as corrosion can be solved. Onda exchange surface of 160 cm2. The stack produces hydrogen at a conservation of mole balance at the anode and the mid-high pressure (max 70 bar). as a function of the selected factors.

hydrogen production rate is directly linked with stack In the Figures 2 to 4 the main characteristics of the current.5 0.4 7 bar. gas outflow produced by electrolysis. Each tie is supplied with some tongs. As we can observe.25 Current density (A/cm2) Electrical Connection Figure 2.25 0. MEAs. capable of compensating the dilatations of the stack. the specific work is electrolyzer in stationary operation conditions are noticeable when increasing pressure. even if this effect is represented. High pressure PEM electrolyzer tested 3000 7 bar. 42 ºC 1 End Plate 0 0. Thus.75 1 1. In Figure 3 the stack power is shown.4 Anode In H2O Cathode Out 2. where the currents not have been represented because power loaded inflow and outflow channels arrive. 42 ºC between two terminal heads (end plates. Tie Bolts & Belleville Washers to Seal 2. During operation. significantly mitigated by the increase of temperature: the values are around 60 kWh/kgH2 at 70 bar.42 ºC 70 bar.8 1. different operating conditions are shown: the main factors varying system output are water temperature (considered at the entrance of the stack) and cathode pressure. around 55 kWh/kg H2. Electric power loaded from the stack in different keeping the accurate pressure in the cell sealing. 42 ºC 2000 70 bar. At low First. the stack has been tested in the extreme values allowed for these factors. which is also isolated from the base plate where the stack lays. The ties operating conditions are electrically isolated from the anodic head. because anode. installed with vertical axel. The specific work support the mechanical stress and to assure the gas rises when operating with lower temperatures or higher tightness to block any dangerous leakage from cathode to pressures: the trend is the same of the voltage. is contained 1000 70 bar. keeping sealed Current (A) the internal fluids. At pressure. 42°C) is around 10%.6 7 bar. while the cathode side is maintained at pressure maximum power loaded. Figure 3. Cell polarization curves in different operating (Cathode) conditions Electrical Connection 6000 Bipolar Cells w/ Separators. Heads are linked with ties. 58 ºC 7 bar.2 70 bar. conditions are set. As expected. which functionality is 50 100 150 200 to compress the fittings between the cells. 58 ºC The cell stack. is almost the same in any condition.with two wire connections each for optimal current increases when low temperature or high pressure conduction. Lower stack The connectors are set in the cathodic head. 58 ºC 1. the In Figure 4 the specific work consumed by the stack for electrolyzer works with a elevated pressure gradient different production rates and different conditions of between electrodes. In Figure 2 the stack cells polarization curves for pressure the values are known. that is around 500 W higher pressure: the cathode pressure rises just from the difference. the difference between the most up to 70 bar by a back pressure valve. the anode side is kept at nearly ambient maximum power input is limited at around 6 kW. which means worse results in terms of efficiency. The hydrogen is favourable condition (7 bar. or end covers). (Anode) Cell Internals 5000 Electric power (W) Tie Down Holes (For Installation) 4000 Figure 1. the cathodic (superior head) and the anodic (inferior 0 head). At high production rates. which are therefore designed to pressure and temperature are shown. the polarization . 58°C) and the worst one (70 pushed electrochemically across the membrane to a much bar.2 High Pressure H2 2 Cell voltage(V) 1. 58 ºC Anode Out H2O & O2 Out 1.

The treatments inert any flammable mix inside the ducts and to purge the have been performed for different current levels of the system before activation. The acquired data can be shared with remote workstations. measurement of many output data. The control system is 0 20 40 60 80 100 interfaced with a computer. pressure data (hydrogen. which delivers the high pressure gas to the The designed treatments have been performed with separation vessels where water is removed from completely randomized design and with two repetitions hydrogen. 42 ºC actuators in order to impose the desired values to the 70 bar. Particular emphasis has been devoted in the design in order to ensure the safe 3. stack obtaining a complete polarization curve for each Finally.3 Control panel Specific work (kWh/kg) 55 The control system is based on a PLC (Programmable 50 Logic Controller) which acquires all the information 7 bar. 2. regulated combination of the other three variables. the polymeric membrane has a permeability to hydrogen and oxygen. installed in the electrical module. Second. oxygen and Torino in collaboration with Sapio S. The eight water flow into the stack for electrochemical reaction and treatments that will be analyzed (23 factorial) can be for cooling. there is the electrical power supply. The three factors chosen have been: The test bench is made up of four main subsystems. it causes an increase of the dependent variable. from a AC/DC voltage regulator driven from the control panel. All measured parameters can be checked easily by operator. and controls the 45 7 bar. All the control devices are fluids in order to ensure the electrolyser stack operation. A operation in case of potential danger: the operation stop positive main effect of a factor means that. factorial design based on the Design of Experiment this driving force could push hydrogen from cathode to technique (Yates method). Its outlet). and the control devices. 58 ºC controlled parameters. for In this work the experiment has been designed using a 23 example due to overload. In fact. having to shut down the system in case of 35 detection of anomalous operation. The experiments have been programmed following a due to the high pressure gradient from cathode to anode. voltage). while taking away the oxygen bubbles produced in the anode. there is the hydrogen displayed geometrically as a three dimension cube. Specific work against hydrogen production rate in various operating conditions in a file for its analysis. records all the running parameters of the process Figure 4. The PLC is also responsible for 40 70 bar. 60 2. 65 characteristic curve of the electrolyzer for any current selected. Finally. The test bench must therefore has a significant effect on the output parameters detect the hydrogen-in-oxygen concentration and stop (dependent variables) or its effect is negligible. circuit. Methodology of the experiments operation of the electrolyzer. 58 ºC provided by sensors and transducers. when the has been imposed equal to 0. concentration. temperature data (water in and out the stack. Third. A nitrogen. The test bench has been designed by Politecnico di hydrogen). factorial problem.p. this concentration must always be operation by evaluating if a factor (independent variable) kept below a safety level. temperature of the stack (considered water inlet the water supply system. The regulation adjusts the source voltage to the .5% of hydrogen-in-oxygen factor is fixed at its upper level. First. all the working conditions have to be measured in order to avoid damages to the stack. which main control window Hydrogen production rate (g/h) is displayed in Figure 7. cathode pressure (or hydrogen outlet pressure) and demineralized water flow. which also carries the It allows to control several physical parameters. the nitrogen supply system is applied to for each treatment in order to analyze the simulated data with an analysis of variance (ANOVA). which sends a demineralised temperature). 42 ºC system safety. Moreover. This technique allows to perform a parametric analysis on the electrolyzer anode across the membrane and a dangerous mix with oxygen could occur.A. the data acquisition system.2 Test bench The measured quantities are: electrical data (current. differential pressure between anode inlet and schematic flow of the test bench is shown in Figure 5. levels in water separators and other data as water function is to supply the electrical power and the process conductivity and water flow. (Italy).

hydrogen production) at different current has been assumed as sufficient for a first screening of the levels. they different approach (i. Once the input factors have been chosen. Therefore. Moreover. DC models or 3k factorial analysis). The first-order regression supplied power. stack temperature. . Yates’ treatment experimental plot consistent in a regression model due to the procedure applied to obtain them. the regression models are written in coded forms. because they represent the sensitivity coefficient linking these variables when the regression models are expressed in the physical form. If current). 2k factorial analysis with Central allow to apply constrained optimization methods in order Composite Design CCD for the second-order regression to maximize one or more dependent variables (i.e.e. in the analytical regression models the coefficient linking the independent and dependent variables are not a-dimensional. Nevertheless. In the tables and figures presented in the paper. water flow and dependent variables in terms of the significant factors.Figure 5. then multiple responses of the stack operation by plotting the experimental campaign should be designed by using a contour plots and response surfaces. fundamental to determine their range of variation (the experimental domain). it is operation of the system. the unit measures are Figure 6. The models can be expressed in physical form with a simple manipulation of the equation. The regression models allow to represent higher-order regression models would be obtained. where the independent variables varies between the values −1 (corresponding to the lower bound of its experimental domain) and +1 (corresponding to the upper bound of its experimental domain). The range of the factors is shown in Figure 6. Display of the main control window of the test bench The aim of this sequential approach is to obtain analytical relations between the dependent variable (the average With this design of experiments it is possible to obtain stack voltage) and the analyzed independent variables first-order regression models for the investigated (hydrogen pressure.

As the previous case. the i=0. as transport and thermo-chemical phenomena vary significantly inside the cell membranes from case to case. The analysis is divided in two parts: the first part focus on the analysis of the experimental data recorded. i=1. almost doubling the pressure. at low temperature the cell voltage increases up to 35% more than at high temperature. This value can be observed in the left graph of Figure 8b.1 A/cm2 lines are almost parallel. Results and discussion Analysis will be displayed for two levels of current: the lowest value is related to a current density of 0. especially at high temperature and low pressure. The second part consists in the identification and selection of the best regression model. temperature has a preponderant effect. Interaction effect of two factors in the cell voltage In Figure 7 we can see that at low current. They are called interaction plots.4. in particular the significance of the effect of every factor on the dependent variable (voltage). At low current rate. This means that the interaction is weak. In Figure 8b we can see the interaction plot at high power loaded from the electrolyzer. because they detail the output parameter value for two factor combination. the pressure effect rises almost up to the temperature level. Other factors seem to be negligible. also given by the algorithm criteria. now it is noticeable the effect of the interaction of these two factors. The factor are codified to simplify the result tables and plots. looking at the interval between lowest and .1 A/cm2 (corresponding to 16 A to the stack) and the highest value is related to a current density of 1 A/cm2 (160 A). factor c the water flow. comparing the results for low and high stack current density. At high power. while the second is kept at its extreme values and the last remains fixed. First. each combination of factors is expressed as a product of their letters. The red line represents the minimum threshold of significance. Two extreme current levels have been selected to characterize the electrolyzer operation in completely different load conditions. The effect plot reveals us how every factor modifies the output variable. the effect of water flow at low pressure or high temperature is completely zero. In this case the lines are clearly not parallel: in particular. Unlike the first case.1 A/cm2 analysis.0 A/cm2 Temperature has a large effect on the voltage increase Figure 7. that means that the effect produced from the abscise factor is similar in both extreme values of the other factor displayed in the legend. Factor a represent the stack temperature. b) 1 A/cm2 Figure 8. producing a high interaction between them. and the residual a) 0. The Figure 8 shows the increment of the cell voltage varying one factor. we evaluate the significance of the effect of the three operation factors and their combinations on the stack voltage. obtained from the experimental data and used as a prevision model. Effect of each factor on the output variable caused by pressure variation: when varying cathode pressure between its limits (7-70 bar). the effect produced by the pressure is very influenced by the value of temperature. factor b the cathode pressure. We can also notice the negligible effect of water flow through the stack.

it is noticeable that there exist a curvature on the voltage plot. Once evaluated the regressor model which best fits the experimental results. Other plots reveal that.968 E-5 T·P + + 9. but the contour plots are more clear. Cell voltage dual factor contour plots In case of high current density (Figure 9b). This is obtained considering the experimental error measured between both replications of every test. at high current density the interaction of temperature and pressure becomes significant.1 E-5 T·GH2O + 7. the only significant factors (the water flow is negligible).94 E-6 P·GH2O Table 2.3. at high current densities. In the p-T contour demonstrated to be not significant in the analysis of the plot it is shown that the variation of cell voltage is similar effects. the water flow rate seems to have an effect at lower temperatures and higher pressures.9E-3 T + 4.1 A/cm2) V = 1.1 A/cm2 For high current density (1 A/cm2) V = 1. especially at high current density. A self-made program selects the best model and displays the obtained analytical expression for both current densities: Analytical regression model For low current density (0. We can observe first that much more regressors have been chosen for the high current model. they are different because. . the effect of the water flow rate has been as noticed with the first data analysis.1.88E-3 T + 9. Regression models for low and high stack current analysis In addition to the factor coefficients. that is also the different curve slope for both cases. in our statements. Pressure and temperature are positive term p*GH2O in the regression curve. This trend is In Figure 9a the contour plots obtained for low stack generated from the negative term T*GH2O and the current density are shown. which simulate the electrolyzer operational performance.94 + 2. Obviously. After the data analysis.86 E-4 GH2O – 2. Nevertheless. this effect may be associated with the when rising pressure or reducing temperature (the level experimental randomized error. This is linked to the fact of having the p*T interaction term in the regression model.highest point in both lines. to reduce the effect of the pressure in the voltage increase. the algorithm returns the confidence interval at 95% of the obtained value. 3D graphics (surface plots) are also available.8E-4 P a) 0. it is very important to have the possibility to operatre at high values of temperature. it is possible to represent the contour plots. Therefore. then it can be neglected lines displayed are almost diagonal to the axis).67 .2 E-4 p . obtaining a regression function fitting the experimental points. b) 1 A/cm2 Figure 9. as shown in Figure 8. we can proceed to build a model based on these results.

Dinamik modelling of a proton exchange conditions: 0. Hydrogen production by advanced proton represent the most significant variable on the cell voltage. 2005. Also. Technical support was provided by the Department of Energetics of Politecnico of Torino. water mass flow at the anode) on the dependent solid polymer electrolyte (SPE) water electrolysis. 31. pp. Kuleshov N. pp. lower temperature.. high-efficiency water-hidrogen-electrolysis. Tsypkin M. 661–669. Solid variable (stack and single cell voltage) and a State Ionics. Authors are also grateful to Giner Electrochemical Systems LLC. idea of the significance. Acknowledgements • water flow sent to the stack produces a variation of This work was carried out as a part of a large project 4e-4 V/l/h in cell voltage at higher pressure and financially supported by the Ministry of Instruction. Solar Energy.. Plot of the main effects of the three factors for 1 Fateev V. Hagen V. Bessarabov D. Vol. Vol. pp.. Fateev V. Italy. 1998 Two analysis have been held at extreme stack operational [5] Görgün H.. the sensitivity of the cell voltage. the tested factors (stack temperature. pp.. 1964 – 1979.. Porembsky V. 31. International Journal of Hydrogen Energy. remaining the voltage almost invariable with pp. Vol. 2006. • pressure produces a variation of 1.A. M. The aim of this paper was to display the results of an 31. 132. in order to pressure water electrolysis. Operational characteristics of high- electrolyzer. 544–550. that results for high current analysis should of each factor. N. These values confirm the negligible effect of water flow in the stack working output compared to the other two [2] Shapiro D. experimental design analysis of a high pressure PEM [4] Schug C. Hattori K. Pien M. 2006. these machines are designed to operate at high • temperature causes a variation in cell voltage of current densities for better products purity and system 5. 535–539.1 and 1 A/cm2.. [8] Barbir F. Vol.. (1 A/cm2): In fact. Universities and Research of Italy.. 64–70. We can resume this in would need higher operating temperatures to reduce the Figure 10. Pure Conclusions hydrogen production by PEM electrolysis for hydrogen energy. experimental design have been presented. Russian Journal of Electrochemistry. it has been displayed the non constant effect of [10] Roy A. Ito K. Solar Energy. Comparison of electrical energy efficiency of atmospheric and high-pressure pressure at different operational temperature. which have an almost equivalent importance in powered regenerative PEM electrolyzer/fuel cell system. where the main effects for temperature. water flow in the stack. 32.6E-3 V/bar at 60 °C and 2. 431–436. pp. producing electrolysers. The results of the Yates’ Technique pressure. temperature. 78. from the stack.. A further results has been [7] Onda K. the in both cases.. 37 (8). 79. Further studies should be done to pressure and water flow are displayed. pp. we can say that at higher current densities be especially considered because of the practical interest. The goal of the laboratory membrane (PEM) electrolyzer . pp.I. [3] Grigoriev S. For low and high power loaded [9] Marshall A. fall of performance. 2006. Kimble M. Khaliullin M. Energy. exchange membrane (PEM) water electrolysers— reducing the power load when the stack is operating at Reduced energy consumption by improved higher temperatures.N. pp. 171 – 175. Børresen B. 2001. cathode side [6] Choi P. International Journal of tests has been to make an analysis of the significance of Hydrogen Energy. and to optimize the investment.. Kyakuno T. bigger jumps in cell voltage when working at low Vol... 2004.. the temperature of the water inflow Tunold R. 175.After having a look at all the graphs. References [1] Grigor’ev S.. A simple model for pressure.... electrocatalysis. which will allow higher efficiency keeping invariable the other factors so they give us an of the hydrogen production system. for providing the S-10 electrolyzer prototype to carry out this tests... 2005. Vol. International Journal of Hydrogen Energy. pp. Datta R. 23 (12).. Vol. quantification of their effects. Solar- factors. International Journal of Hydrogen Energy.G. It would be advisable then to run the . To get an idea of the importance In particular.. Prediction of to obtain the analytical models that predict the production power for high-pressure hydrogen by high- electrolyzer operation in several conditions. PEM electrolysis for production of hydrogen Temperature and pressure have a predominant effect in from renewable energy sources. A. 29 – 38. Vol. 2004. Journal of Power Sources. we can conclude electrolyzer at higher temperatures to alleviate the that temperature and pressure have a predominant effect negative pressure effect in efficiency.. Results have been improve the membrane thermal robustness to ensure plotted for 1 A/cm2 for clarity. Watson S. remaining the voltage almost invariable pressure increasing trend of new electrolyzer models with the water flow in the stack. Vol. Vol. 2007.25·10-3 V/°C at 7 bar and a variation of 7·10-3 safety. V. 819–822. Duffy J. report the electrolyzer contour plots.... V/°C at 70 bar approximately.. both cases. Therefore. Figure 10.A. 1113-1120. This plots are obtained operation over 60ºC.1E-3 V/bar at 40 °C approximately. Infield D. Electrolysis of Water in a System with a A/cm2 Solid Polymer Electrolyte at Elevated Pressure..

Nomenclature GH2O Water flow (l/h) I Stack current (A) I Current density (A/cm2) P Cathode pressare (bar) PEM Proton Exchange Membrane T Water inlet temperature (°C) V Stack voltage (V) .