Published on

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide


  1. 1. Proton Exchange Membrane Fuel Cell Characterization for Electric Vehicle Applications D.H. Swan B.E. Dickinson M.P. Arikara Reprint UCTC No. 257The University of CaliforniaTransportationCenterUniv~sitvof CaliforniaBerkeley, CA94720
  2. 2. The University of CaliforniaTransportation CenterThe University of California Center activities. ResearchersTransportation Center (UCTC) at other universities within theis one of ten regional units region also have opportunitiesmandated by Congress and to collaborate with UCfacultyestablished in Fall 1988 to on selected studies.support research, education,and training in surface trans- UCTC’seducational andportation. The UCCenter research programs are focusedserves federal Region IX and on strategic planning foris supported by matching improving metropolitangrants from the U.S. Depart- accessibility, with emphasisment of Transportation, the on the special conditions inCalifornia Department of Region IX. Particular attentionTransportation (Caltrans), and is directed to strategies forthe University. using transportation as an instrument of economicBased on the Berkeley development, while also ac-Campus, UCTC draws upon commodatingto the region’sexisting capabilities and persistent expansion andresources of the Institutes of while maintaining and enhancoTransportation Studies at ing the quality of life there.Berkeley, Davis, Irvine, andLos Angeles; the Institute of The Center distributes reportsUrban and Regional Develop- on its research in workingment at Berkeley; and several papers, monographs, and inacademic departments at the reprints of published articles.Berkeley, Davis, Irvine, and It also publishes Access, aLos Angeles campuses. magazine presenting sum-Faculty and students on other maries of selected studies. ForUniversity of California a list of publications in print,campuses may participate in write to the address below.University of CaliforniaTransportationCenter108NavaIArchitectureBuildingBerkeley,California 94720Tel: 510/643-7378FAX:510/643-5456Thecontentsof this report reflect the viewsof the authorwho responsible isfor the facts andaccuracy the data presentedherein. Thecontentsdo not ofnecessarilyreflect the official views policiesof the State of Californiaor the orU.S.Department Transportation. of This report doesnot constitute a standard,specification,or regulation.
  3. 3. Proton Exchange MembraneFuel Cell Characterization for Electric Vehicle Applications D.H. Swan B.E. Dickinson M.P. Arikara Institute of Transportation Studies University of California at Davis Davis, CA 95616 Reprinted from Advancements Electric and HybridElectric Vehicle Technology in Society of Automotive Engineers, SP-1023, February 1994 UCTCNo. 257 TheUniversity of California Transportation Center Universityof California at Berkeley
  4. 4. 940296 Proton Exchange MembraneFuel Cell Characterization Electric for Vehicle Applications D. H. Swan, B. E. Dickinson, and M. P. Arikara University California, of DavisABSTRACT a complex group of support systems that must operate in balance for efficient performance. 3L~is paper presents experimental data and an analysis of Different types of fuel cells are conveniently dassLfieda proton exchange membranefuel cell system for electric the type of electrolyte they use. Electrolytes that are presentlyvehicle applications. The dependence of the fuel cell being considered include the proton exchange membrane (asystem’s performance on air stoichiometry, operating solid polymermaterial), phosphoric acid (a liquid), alkalinetemperature, and reactant gas pressure was assessed in terms (a liquid), molten cadxmate (a liquid) and solid oxideof the fuel cell’s polarity and powerdensity-efficiency graphs. ceramic). The choice of electrolyte directly affects a fuelAll the experimen~ were performed by loading the fuel cell cell’s operating characteristics; for example, phosphoricacidwith resistive heater coils which could be controlled to is a poor ion conductor at roomtemperature. As a result theprovi,de a constant current or constant power load. System phosphoric acid fuel cell must be heated to 150 to 200°Cparmdtic power requirements and individual cell voltage before it can be used. Todaymany researchers believe that thedia~aS[bution were also determined as a function of the proton exchange membrane (PF2eD provides the bestelectrical load. It was fotmdthat the fuel" cell’s performance characteristics for transportation applications.improved with increases in temperature, pressure and The data and analysis presented in this paper is for astoich~ometry within the range in which the fuel cell was fuel ceil system manufactured by Ballard Power Systems ofoperational. Cell voltage imbalances increased with increases Vancouwer, Canada. The fuel ceil system consists of: a 35in current otaput. The effect of such an imbalance is, cell series connected stack; gas, water and thermalhowever,not detrimental to the fuel cell system, as it is in the management subsystems; and controls and monitors allcase of a battery. assembled in a single enclosure. The area of each cell was 232 can2 and the fuet cell stack itself had a maximum grossINTRODUCTION power output greater than 3000 Watts operating on hydrogen and air. The systemwas modified by the authors to be able to An electr~-hemical fuel ceil is a device that converts independently control air stoichiometry, air/hydrogenchemical energy to direct current electrical energy. By pressure and stack exit air temperature. Previous papers thatconverting an on-board fuel to electricity it could be have presented experimental data on similar Ballard fuel cellseffetely-ely used to poweran electric vehicle. As such, a fuel systems are referencesland 2cell is an energy conversion device like an Latemal The paper is organized into sections in the followingcombustion engiue. This is in contrast to energy storage order; Fuel Cell Operating Principle, Experimentalsystems such as bat*,eries, flywheels and ultra capacitors. Apparatus, Experimental Results, Results Analysis andFurther manyof a fuel cells operating characteristics are Conclusions. The section on fuel cell operating principle iscloser to that of an engine th~n a storage battery. A fuel cell intended to give a brief overviewof howa fuel cell worksandsy~m operation involves startup, fuel and air delivery its operating characteristics. The experimental apparatuscontrol as a function of load, and removal of heat madby section briefly describes the fuel cell system used in theprochacts of the reaction. The fuel cell, in other words, is an experiments and the associated instrumentation. Theelectrochemical engine. While electrochemistry describes the experimentalresults present a series of polarity plots (voltageprincspIe of operation of a fuel cell, the engineeringchallenge - current relationship) under a variety of operating conditions.of baLancing the many variables over a wide variety of The results analysis section presents the results in terms ofoperating conditions remains. The fuel cell systemconsists of power density-efficiency plots implicitly demonstrating the 19
  5. 5. operating characteristics of the fuel cell ~stemfor electric AG= At] - T ×AS (I)vehicle applications. WhereFUEL CELL OPERATING PRINCIPLE Free energy of Formation All energy-producing oxidation reactions are T Absolute Temeperaturefundamentally the same and involve the release of chemical AS Entropyenergy through the transfer of electrons. During combustion Typically theoretical chemical to electrical conversionof hydrogen and oxygen there is art immediate transfer ofelectrons, heat is released and water is formed, efficiencies are in the range of 83% higher heating value for and 94% for the lower heating value of hydrogen. tn a fuel cell the hydrogen and oxygen do notimmediately come together but are separated by an Efficiencies of practical fuel cells using pure hydrogenand airelectrolyte. First the electrons are separated from the range from 40% to 75% based on lower heating value ofhydrogen molecule by a catalyst (oxidation) creating hydrogen.hydrogenion (no electrons). The ion then passes through the Assu~ng a near perfect cotflombic efficiency (allelectrolyte to the oxygen side. The electrons cannot pass electrons are forced to take the external circuit) a theoreticalthrough electrolyte areforced take external the and to an operational voltage can be calculated. This voltage iselectrical circuit whichleadsto the oxygen side.The calculated by considering Famday’sconstant (26.8 Amp hourselectrons provide cart useful work they as pass through the = I mole electrons) theenergy of and value ofthefuel. Tableexternal circuit. Whenthe electro~ reach the oxygen side I, provides hydrogcrJoxygen-water the reaction enthalp (heating value), freeenergy formationand the the of 3they combine with the hydrogen ion and oxygen creating restfltant theoretical cell voltages.water. By forcing the electrons to take an external path, alowtemperature direct energy is conversion achieved as Table 1 Hydrogen Thermod,namic Pro )ertiesshown Figure I. in He,tin AH Er~ai~,"GibbsFunction voyage CeU g Value of F~ AG=AH-TAS Basea on kJtMole kJ/Mae Enth~p/ Bnsed on Hydrogen At26~C, bar I Gibbs H~ -285.9 -237.2 ~A8 1.23 -241.8 -228.6 1.25 t.18 _Jil Lower Like a storage battery, when the fuel cell is under ,--~2H++ electrical load the voltage falls with maximumpower generally being producedbetween 0.5 and 0.6 volts per cell. Thevoltage asa function current due to interna drop of is resistance (electronic ionic), and electrode kinetics (particularly the air electrode), on reactant flow gas limitations product and water flooding reaction of sites. To makea useful voltage,multiple cellsare coanccted in electrical series,referredasa stack shown Figure to as in 2. !l Ii oao.oo,or Acid Electmbfte Manifolds fuel cell deliver reactant gases thereaction stack allnecessary and to sites. a~aries referred as a are to The Air (O2 & N2) H20 N2 fuel system. cell Figure 1. Fuel Cell Operation Schematic Air Air Air Air The theoretical efficiency for the conversion of chemicalenergy into electrical energy in a hydrogen-oxygen fuel cetldepends on the free energy of formation for the reaction(Gibbs Function). The free energy of formation is equal _ +the difference of the heating value of the fuel and its entropyat the temperature and pressure of conversion. This isdescribed in equation I. N2 N2 N 2 N 2 Figure 2. Fuel Cell Stack Schematic 2o
  6. 6. The fuel cell stack design must allow for heat exchange installed in the air delivery line. Pressure control wasand humidification of incomingreactant gases, thermal established by using single stage pressure regulators.management,pr~xtucl water management,exhaust gases, and Hydrogen for the fuel cell shack was provided fromelecuical management. industrial hydrogencylinders. The air system included a two FUEL CELl; SYSTEMDESCRIPTION- Given below stage compressor along with a chiller dryer and a system ofare someof the specifications for the fuel cell powersystem air filters.which among others includes a 35 cell stack with ahumidification ~ction, temperature management, reaction EXPERIMENTAL RESULTSproduct management electrical control. and The experiment was organized into exploring the effects Table 2 Fuel Cell SystemSpecifications. Electrolyte Nation-117 of three independent variables on performance: air # of AotiveCells 35 stoichiometry, air pressure and exit air temperature. Air Active Area/cell 2 cm 232 stoich/ometry is the ratio of the amountof air provided to the Total acthtearea 2 8120cm fuel cell tO that whichis necessary to react with the hydrogen Actiw; Cell TNckrP.ss 0.5 cm # of"C;oolir~Celb$ 19 fuel. Other words used to describe this ratio include Lambda Acl~e PlateThickness 0.5 cm and excess air factor. In a typical internal combustionengine # of Humidification Cells 14 the value is very close to 1. Dueto the operating nature of a System dimensions LxWxH 104x 80x 37cm System volume 307.8 ~ers fuel cell the stoichiometry of the air must be greater than one S~hekdimensions LxWxH 38x 21x 21cm" (typically 105 to 4). A lower value of stoichiometry reduces¯ Stack volume 16.76liters performancedue to the lack of oxygenat the reaction sites. A Weight thestack of 43kg high value of stoichiometry results in poor humidity control Weight hhe of system 50 kg Support SystemPower** 350Watts and excess compression energy. *lncltu~esactivestack humidirmation and with section cooling cells The following test matrix illustrates the 12 different **water pump,hydrogen recirculation fans,solenoid control pump, valve operating conditions over which the fuel cell system was and monit~ng system characterized. The stoichiometric accuracy was held between +/- 0.1, the pressure accuracy was held between +/- 0.05 barDA~[A ACQUISITION and the temperature accuracy between +/- 1.5°C. Under each operating condition the stack output current was varied (Sine of the important tasks in the experimentsconducted between a minimum and maximum amperage at 10 ampwas to monitor and control the operating conditions of the intervals. The minimumcurrent cond/tion (20 ampa) wasfuel cell. The data acquisition and control systemused signal limited by the supportsystempowerrequirements and a smallconditioning modules which were capable of either data load on the eIectric load bank . The maximum currentacqtfisition or control based on commands received from an condition was limited by a minimum stack voltage of 19 (0.54IBM compatible personal computer. The modules volts/cell). The minimum stack voltage was a Emit set bycommunicated with the computer through an RS-232 Ballard power systems and was within a safe operating rangecommunications port. Each module was isolated to I500 for continuos operation of the system.volts and includes a filter to reduce noise. The control systemim, otved both digital and analog !nput and output. Table 3 Test Matdx For Fuel Cell Testin~ ]’he load for the fuel cell wasprovidedby resistive heater Stoichi~ Hydrogenand Stack Air Exitcoils. The resistance provided to the fuel cell was controlled Air Pressure Temperature °Cby a pulse width modulator (PWM).The PWM turn was in in BarcontroUed by the analog output of the control system. The 1.5 2 50,60,70resistance could be controlled to provide a constant poweror 3 50,60,70con~uant current load_ 2.O 2 50,60,70 The variables measured by the control system included, 3 ~m.,mstack current and voltage, support system current, loadcurrent, process air exit tempematre,hydrogeninlet pressure, The results of this test matrix are organized into twoair inlet and outlet pressure. The digital signals monitored polarity plots (Stoich.=l.5 and Stoich.=2.0), andincluded the status of the load relay, hydrogenvent, oxidant individual cell voltage characterizationvent and the water drain. POLARITY PLOTS- The polarity data resulting from Onthe control side the load applied to the fuel cell was the Table 2 test matrix is presented in Figures 3 and 4. Thesecontrolled by an analog output module: The digital output plots include the measuredfull stack data (volts and amps)conlxolled the heat exchanger operation in order to maintain ~ and normalized data (volts/ceU and milli amps/cm of acth,ethe ~emperature the fuel cell at the required levels. of area) The normalized data represents an average for the In order to maintain the stoichiometry of the air provided stack; actual individual cell data is presented in the followingfor the reaction a needle valve provided on the fuel cell was sub section.controlled. The flow rate was monitored by a flow meter Figure 3 presents the results for the six cases for which the stoichiometry was equal to 1.5. Figure 4 presents the 21
  7. 7. same results with the stoichometry equal to 2.0. The current current range, a cell that shows a lower voltage at a low range over which the two plots are de~ribed is the normal current density continues to have a lower voltage at higheroperating range of the fuel cell system. (Openclrcuat voltage, current densities. The lower voltage could be the result of thealthough not shown, was found to be approximately 1 volt per difference in the electrode membrane assembly, distributioncell.). of reactant gases, internal cell resistance or a persistent Both, Figures 3 and 4, indicate a near linear relationship flooding effectively reducingthe available reactive area.between voltage and current. The higher stoichiometry Although shown, not individual performance has cellshownin Figure 4 results in a higher voltage at any given been measured several times over 6 months preparatio the ofcurrent compared to Figure 3. The difference in voltaic to conduct the expcrinacnts described thispaper. inperformance between the two stoich, rates is most pronounced Consistently cells6 and 25 havehad a lowervoltaicat the high current levels. With a Stoich of 1.5 the fuel cell performance theothercells than (0.05volts lower ~hansystem could only deliver 130 amps before the low voltage average I00amps). a result authors for As the feelthatlimit was reached. In comparison, at a stoich ram of 2.0 the difference is inlaerent to the particular cells and not afuel cell could deliver approximately 160 amps. At a transient phenomenadue to a flooding condition or otherreference performance condition of 100 amps (normal temporary phenomena. is interesting to note that cells 7, Itoperating condition) the average voltage increase with 24 and 26 perform above average, these high performanceincrease in stoich rate from 1.5 to 2 was 1.35 volts or 6.0 %. cells are adjacent to the lowest performing cell and mayAt lower current rates the percentage difference is smaller indicate that the problem one of flow distribution. isthan at the higher rates. This seemingly small increase in The following Table 4 summarizes the results of thevoltage at a given current has the twoffold effect of increasing individual cell measurements. Although Figure 5 shows apower and operating efficiency. In the following section; consistent l~ttem of ceils 6 and 25 having lower than averageq~esults Analysis’, it will be shownthat even a relatively voltage the overall stack performance is very good. Thesmall increase in voltage has a sianaLficant impact on ra~-dmum voltage spread is the resdt of ceils 6 and 25 whileperformance an electric vehicle. for the standard deviation is very small due to the generally As a function of pressure, voltaic performance also consistent performance the stack. ofincreased. Once again utilizing a reference performznce T~bie 4 Individual Cell Volta( CharacteasticsconditionI00amps typical voltage of tim increase going from StackCurren~ 20 4O 7o liE) 1302 to3 barwasbetween and1.2volts to5%).Voltaic 0.5 (2 S~¢k Vo~a~ 30.25 28.4 26.5 24.4 22.4performance also increased a function of temperature . asOnce again utilizing a reference performance condition of CurrentDens~ 86 172 3O2 431 560 2 mA/cm100 amps the typical voltage increase observed for anincrease in temperature from 50 to 60 °C was 0.5 to 1.0 volts Avera@e Votts~e 0.8e,40.812 0.758 O.698 0.640 Ceil(2 to 4%), far aa increase from 60 to 70 °C the corresponding Maximum Vo~ge 0.03 0.042 0.063 0.086 o.115voltage increase was 0.3 to 0.7 volts (1 to 3 %). The benefit spre~tof going from 60 to 70°C was less than going from 50 to CellVoltage Standard .OO6 .OO5 .0115 .017 .O2560°CThere willbe an optimum operating temperature for D~tion Vo~thefuelcellbeyond whichthevoltaic performance will SUPPORTSYSTEMSELECTRICALLOAD- The fueldecline. experimental indicates tl~s The data that value is cell system has support components that are electricallyprobably greater 70°C. than operated and controlled. Support system power is drawn For both the pressure temperature and changes the from the fuel cell stack through a voltage regulated powerpercentage change in voltaic peffomnanceis greater at high ~pply (constant 26 volt output). The support componentscurrent rates. Once again this seemingly small increase in include a combined cooling and reactant gas humi~cationvoltage at a given current has the two-fold effect of increasing system that recovers product water from the stack exit air, apower and operating efficiency. Thus a relatively small hydrogen recirculation system, and a control system thatincrease in voltage has a significant impact in performance monitors andcontrols start operation shut the up, and downfor an electric vehicle. ofthe fuel system. cell INDIVIDUAL CELL PERFORMANCE The voltage - Thecombined cooling humidification utilizes and systemperformance of individual cells was measured at a exit air a circulation pump water that provides tocooling water platestemperature of 60°C and a pressure of 3 bar. Five different followed thereactant humidification by gas section. on Basedstack current rates were used (20, 40, 70, 100, 130 amps) and stack exit air temperature the water leavingthethe respective cell voltages are presented in Figure 5. The humidification is routed section throughwater airheat a tomeasurements were manually made utilizing a digital volt exchanger. hydrogen The rccirculation system maintainsameter and pointed probes directly on to the active cell plate constant recirculation through stack a liquid rate the andedge. water knock out drum. This circulation combined with Figure 5 shows that there is a difference in voltaic periodic purges prevents flooding conditions and build up ofperformance for each cell at a given current rate. The inert gas on the hydrogen side of the tirol cell stack.differences in performance are consistent over the operating 22
  8. 8. Table 5. gives tile support system load demandduring to chemicalenergy) is less than the fuel cell stack conversionoperation of the fuel ceil. The support system power efficiencyrequirement does not include air compression which is Vinck,-pendent of the fuel cell stack output. As a result the Fuel Cell Stack ConversionEft. - x loo %(3) I. 25 ×35support systempower(a parasitic loss) varies in percentage Wherestack power from 100%(idle condition) to 10%(full powercondition. The difference in support system power (250 to V Stack Voltage400 Watts) results from the cooling of the circulating water. 1.25 Theoretical Cell VoltageWhenthe water is circulated through the water to air heat Based on Enthalt9’ of Formation ( Lowerexchangerthe additional poweris drawnby ventilation fans. 35 Number Cells in Stack of Table 5 Supporl Systems Loads Utilizing equations 2 and 3 the data presented in Figures C(~ponent 4 and 5 was recalculated and presented in Figures 6 through WaterPump 3-4 9. Figures 6 through 9 present the fuel cell stack specific HydrogenCirculation pump 1 poweras a function of fuel cell stack conversion efficiency Ventilation Fans 2 based on the lower heating value of hydrogen. In these Contro~.batteries ar~l 3-6 figures a higher electrode power demi.ty translates to a solenoids smaller electrode area to achieve a given powerlevel. A high Total Suprx~tSystem 10to 15 conversion efficiency translates to a smaller fuel storage and (250 to 4coWa~ refueling time for a given vehicle range and performance. A secondorder affect is that a higher efficiency translates toRESULTS ANALYSIS smaller heat generation within the fuel celt system thus a smaller heat exchanger system and other support components° The previous sections described fuel cell polarity plots, Figures 6 through 9 are organized as a function ofwhich is a standard methodto describefuel cell performance. stoichomettT and operating pressure. Figures 6 and 7 presentHowever,the interests of an automotive engineer in applying specific powerdata for a stoich of 1.5 and pressttres of 2 anda fuel cell to aa electric vehicle is better described in terms of 3 bar respectively. Eachfigure showsthe results for the threepower and efficiency. The specific power available different temperatures in the test matrix (50 60 and 70 °C).determinesthe size of the fuel cell system and to someextent Consider Figure 6, the specific power-efficiency curvesits cost. The conversion effidency affects the sizing of the indicate a trade-offbetween powerand efficiency. The higherfuel storage, resultant range and refueling times° the specific power density, the lower is the conversion To accoauuodate this need for a power and efficiency efficiency. The relationship between .specific power andrelationship the polarity data can be reinterpreted. The efficiency appears linear for high conversion efficiencies,spe~:~c power of the fuel cell is simply the ratio of the 700 to approximately 55%. At this point the power increaseproduct of the stack voltage and the stack current to the total for the drop in efficiency becomes less and the curve starts toavalilableactive: electrolyte area. flatten° AltI3ough the curve does not distinctly show a specific power peak the data indicates that this peak isSpec|fiePower =TotalActtveVX[ Area x 1000 mWatts ( ~ ~m ](2) 2 somewhere around 45%energy conversion efficiency, this corresponds to an average cell voltage of approximately 0.55Where (just slightly greater than one half the open circuit voltage).V Stack Voltage Increasing the current beyondthe peak powerpoint results inI Stack Current a further decline in specgfic powerand conversion efficiency. At no time wouldit be advantageousto operate the fuel cell To calculate the fuel cell slack gross conversion system beyond the peak power point.efficiency it is assumed that the system is operating at a The following analysis utilizes three different specificperfect coulombicefficiency (all ava/lable electrons from the powerconditions to illustrate the influence of stoichometry,hydrogen fuel were forced to take the external circuit pressure and temperature. The three specific powerproviding useful work). This meansthat all inefficiencies are 2, ~, conditions selected are 125 mW/cm 275 mW/cra and 325manifested in a voltaic loss. Equation 3 presents the method 2. The low powercondition (125) relates to a possible mW/cmuso~ to calculate fuel cell stack conversion efficiency. This cruise condition for the electric vehicle, the high powerapproach to calculating efficiency is generally considered condition (275) relates to acceleration or hill climbing. Thevalid for the proton exchange membrane t~ael cell, however highest specific power325 was not attainable by the systematme~lsurements will be madein the future using this samefuel low stoichometry and pressure and is used to illustrate theceil system to quantify this assumption. See the earlier enhanced capability by changing operating conditions. Tablesection, ’Fuel Cell OperatingPrinciple’, for further details. 6 relates the energy conversionefficiency to the three selectedNote that no allowance is madefor the support system power specific powerconditions.reqa/rements or air compression. As a result the netefficiency of com,ersion (ratio of delivered electrical ener~’ 23
  9. 9. First comparing the affect of stoichometry, the additional support energy required to maintain the fuelefficiencies of operation in Table 6 immediatelyshowsthat at operating temperatureat a high value.low power conditions stoichiometry has little influence on Table 7 Fuel Cell Stack Conversion Efficiency at Threeefficiency of operation resulting only in 1 to 2 percentage Specific PowerPointspoints increase. At the higher power condition the spread operating Condition Speci.Power Speci.Power Speci.Powerincreases to 4 to 5 percentage points or in other wordsa 10% 2 mW/cm 2 mW/cm 2 mWlcm 125 275 325change in stack fuel consumption for the same power level. S=2.0,T--50C, Bar P =2 64% 52% notpossibleAt the highest powercondition the increase in stoichiometry $=2.01T=70C~ Bar.~ P =2 65% 57% 51% 2 S=2.0~ F P =3Bart T=50C 66% 56% 52%is necessary for the stack to reach the 325 mWlcmspecSfie S=2.0,T=70C, =3 Bari P 67% 59% 55%power level The low stoichiometry dmply cannot reach the~.high level having l~W,~ed at approximately 290 mW/cm This analysis also indicates the clear advantage that aIncreased stoichiometry is most benefidal at high power fuel cell has at part load condition. Unlike an internalconditions. comlmstionengine the energy conversion efficiency increases substantially as the load is reduced (typically 15 percentageTable 6 FueECeg Stack Conversion Efficiency at Three points or a fuel reduction of 25%for a kWhproduced). Not specific PowerPointsoperating Condition Spe~. Power speck Power speci.Power evident in this analysis is that highly dynamicoperation does 2 mW/cm 2 mw/cm 2 mw/cm 125 275 325 not degrade the efficiency. Operating on hydrogen the fuelS=1.5,T=60C, =2 Bar P 64% 51% cell responds instantaneously to the new operating load withS=1.5,T=60C) =3 Bar P 64% 54% no loss in efficiency for the change in power condition 4. AS=2D,T=60C~=2 Bar P 65% 56% 50% difficult that does arise under dynamiccon~tions is the needS=2.0, T----ff~C,=3Bar P 66% 58% 53% to control the ~r flow at prescribed stoichiometric rates. Comparingthe influence of an air pressm’e change from This combination of high part load efficiency and no2 bar to 3 bar there is almost no difference for the low power degradation to dyrmmic due operation gives fuelcell the acase. Under high power conditions the change in pressure clear advantage theinternal over combustion engine. Thisincreases the efficiency by 2 to 3 percentage points or a advantage isparticularly important a city for drivingcycle.difference of 6% in stack fuel cor~tmption for the same AIR COMPRESSIONENERGY - The support systempower level. For the highest power condition a pressure fora fuelcell power plant include controls, cooling fans,changefrom 2 to 3 bar allows the lower stoichometry of 1.5 to recirculation (water pumps and hydrogen) compresse and 2,reach high power of 325 mW/cm however the lower process Thedifference air. between cell fuel stack power andstoichiometry has a energy conversion efficiency 7%points fuel cell systempoweris due to the parasitic losses associatedlower or a stack fuel con~unption of 15% higher. Like withthesupport system components. thesecomponent Ofstoichiometry the benefit of pressure is most pronouncedat andtheenergy required operate thecompressed to them airhigh power conditions. Both increased pressure and energy requirement farthelargest. isbystoickiometry increase the compression power needed to Pressurization improves performance fuel ofair the of theoperate under these conditions. As a result, choosing the cell whenconsideredan individual as entity. However., thestoichiometry and operating pressure is not a simple straight energy for compression an electric on vel~cle mustbeforward procedure. The affect of air compression powerwill supplied thefuel by cell system, thus power and net output isbe examinedin a following section. less thanthegross stack performance. aircompressio The Table 7 compares the influence of temperature on process can be adiabatic or isothermal and part of theperformance for a stoichiometry of 2, the two different compression energy maybe recovered from the exiting air bypressure cases are presented for comparison. Onceagain at a an expander such as a turbine. To determine the influencelow specific powerthere is little difference in efficiency. At the energy of compression would have on the fuel ceUthe high powerlevel the efficiency, spread resulting from an performance the following znzlysis assumes that theincrease temperature is between 3 and 5% points. At the compression is performed by an ideal a~abatic compressorhighest power level the low temperature low pressure case with energy no recovery upon expansion. affect theair The of 2.can not achieve 325 mW/cm By increasing the compressorberelated a sL~nple innetcell can to loss voltagetemperature to 70°C the highest power condition was as a function pressure of ratio stoichiometry. and A fullpossible at a conversion efficiency of 51%. The effect of derivation befound reference brief may in 5a explanation istemperature rise on the 3 bar pressure case was to increase provide thefollowing in paragraph°efficiency by 3 % points, or a difference in stack fuel Aircompression is a function inlet outlet power of toconsumption of 6%. There will be an optimum operating pressure ratioandthc quantity air compressed. of Thetemperature for the fuel cell beyond which the conversion relationship compressor of powerto quantity air ofefficiency, for a given specific power ~dll decline. The compressed linear. a constant is For stoichiomctric theexperimentaldata indicates that this value is probably greater quantityairneeded directly of is related thecurrent to beingthan 70°C. The high operating temperature reduces the produced. Double current, the double airis needed the tostzing of the fuel cell cooling systemby a. Further there is no maintain stoichomctry, the double compressor the power is needed maintain flow. to the Consider ideal an compressor in 24
  10. 10. series with the electrical load. When fuel cell current is the Air compression decreases the conversion efficiency ofincreased the compressor power must increase in direct fuel cell system It also impacts the peak specific powerproportion. Because the current has gone up the voltage drop available. A high stack performance from increasing theacross the compressor remains the same. The voltage operating pressure may not be enough to over come theavailable to the electrical load (net voltage) is simplythe cell additional powerrequirement of the compressor.voltage minus the effective compressor voltage that is aconstant for a Westoichiometry, and pressure ratio. Thusthe CONCLUSIONSeffect of air compressioncan be simply presented effectivelyas a reduction Lr~ voltage. Utilizing equation 3, this effective The 3 kWfuel cell system describedin this paper is ancompressorvoltage loss can be used .to determine the impact impressive performer with approximately 40 hours ofair compression has on efficiency. The following equation to operation in preparing and obtaining the presented data.ealctdate effective compressor voltage was first derived in Measured energy conversion efficiencies operating on while6reference hydrogen and air ranged between 45%and 70%. Under part load conditions (1 to 2 kWthe efficiency typically ranged 1.287 .~ #ofStoich. xcp-1-1 P2 T . I (4) between 55 and 65%, peak power output (3 kW) .typically Vc- 3600 occurring at 45%. The trade off between power and Where will efficiency influence sizing thefuel the of cell system. V Effective Adiabatic CompressorVoltage loss per Cell Fora given vehicle application a larger expensive morc fuel c Specific Heat (Air 1.004 J/(g °K)) cellwilloperate a lower at average power thushavea and~ffl CompressorInlet Air Temperature °K better economy. fuel The result of increasing the fuel cell operating pressureP1 CompressorInlet Air Pressure and stoichiometty is to increase its performance. TheP2 CbmpressorOutlet Air Pressurek :~pecific HeatRatio(Air 1.4) influence of these two parameters was found to be dependent on the fuel cell electrical load. At low load conditions Utilizing equation 4 the effective compressor voltage increasing stoichiometry or pressure had little influence. Atloss tbr the 35 cell stack data presentedin this paper is high load conditions increasing the stoichiomctry from 1.5 totabulated in Table 8. The voltage loss has been interpreted as 2 improved performance by as muchas 7 percentage points.aa efficiency reduction and is presented in brackets. This would effectively reduce stack fuel consumptionby 15% Tabte 8 Effective Adiabatic CompressionVoltage Loss for the same amount of energy conversion. Increasing the and Resultant Efficienc~ Loss operating pressure from 2 to 3 bar would improve stack $toleh Prem===2,r 1 l Pre~ -3 Bar performanceas muchas J percentage points or an effect fuel 1,5 1.23 volts (-2.8%)2.07volts (-4.7%) consumption drop of 7%. [ The benefits of increased stoichiometry and pressure Table 8 indicates that the affect of the energy of air must be balanced against the power needed to compressthecompression is to reduce operating efficiency by 2.8 to 6.3 additional air to a higher pressure. A brief analysis of thepercentagepoints. Considering that air must be forced calculated adiabatic compressionpower with no pressurethrough the fuel cell stack the 2 Bar case will be considered recovery was presented. The analysis found that based oft theas a base line,,;. Increasing the pressure from 2 to 3 bar experimental data there was no net advantage in increasingincrease the eflbctive degradation in efficiency by 1o9 to 2.6 the operational air pressure from 2 to 3 bar. This calculatedpercxmtage points. result indicates the trade offbetweenstoichometry, pressure In the previous subsection the influence of a pressure and the energy of compression is not a simple one. Otherchmagefrom 2 to 3 bar on efficiency at low SlXX~cpower support components that must be powered the fuel cell by"conditions was found to be I percentage point or less, thus the include controls, a water pumpand a hydrogenrecirc-dationnet effect of compression would be negative. At 1,hgh power pump. These componentsrequired a near constant 250 toconditions the ~fffect of pressure changefrom 2 to 3 bar is to 350 Watts (10% of maximum fuel cell stack power) and wereincrease conversion efficiency by 3 percentage points. This independent of the fuel cell system poweroutput. Bothgain is almost entirely nultified by the effective compressor pumps were operated at constant rates and sized forefficiencyloss. maximum operating conditions. In an automotive design Utilizing the data for a stoichiometry of 2 at 60°C the they will be variable and thus reduce the percentage powerperformanceof’pressure equal to 2 bar and 3 bar is compared requirementsof the punlps significantly.in Figure 10. "[’he fuel cell stack sqgecific powerfor the 3 bar The affect of increasing the operational temperaturecase is higher then the 2 bar case. However when the (stack exit air temperature) from 50 to 60 to 70 °C waseffective compressor voltage loss is considered (Table improvementin performance. There is an optimumoperatingvalues) the performance of the 2 bar case exceeds the 3 bar temperature for the fuel cell stack beyond which thecase. The 3 bar case has a higher peak powervalue but it is performance will decline. The experimental data indicatesnoT:ably lowerefficiency. that this value is probably greater than 70°C. Highoperating 25
  11. 11. temperaturereduces the sizing ofthe fuel cell cooling systemand is considered an advantage. individual cell voltaic performance was measured andfound to have slight variations between cells, Twocells inparticular had lower voltages (0.1 volts below average atcurrent of 130 amps). Adjacent to these ceils were cellperforming above average, indicating possible reactant gasflow distribution problem. Unlike a storage battery the lowervoltaic performanceis not a serious problem. The fuel cell powersystem is a strong candidate to meetthe needs of California’s ZEV mandates. Utilizing on-boardhydrogen the only byproduct of operation is water. The highconversion efficiency of a fuel cell overcomesmanyof thestorage and cost problems associated with hydrogen. Thepromise of fuel cell technology is a ZEV with theperformance, range and rapid refueling capability ofconventional vehicles.ACKNOWLEDGMENTS The authors wish to thank BalIard Power Systemsand California Department of Trarmportation (NewTechnologies Division) for the support they have provided insetting up the fuel cell laboratory at the Institute ofTransportation Studies, Davis. Wealso wish to thank theUniversity of California Transportation Center for financiallysupporting the Graduate Students involved in this project.Last but not the least we wish to th~nk Mr. Gonzalo Gomezand Mr. Manohar Prabhu for their assistance in setting up thedata acquisition and control system and helping us inobtaining the experimental dataREFERENCES1 Prater K. B., "Solid PolymerFuel Cell Developments at BaUard", Procedings of the 2nd Grove Symposium, Editors A.J. Appleby and D.G. Lovering, 1991 p 189 to 201.20tivera C.T., A. Anantaraman and W.A. Adams, "Performance Evaluation of a H2/Air PEM-FC System under Variable Load", Precedings of the 1992 Fuel Cell Seminar, Tucson, Adzons, December1992 p 451 to 4543 Masterton W.L., E.J Slo~4.nski, "ChemicalPrinciples", W.B. Saunders Co. Philadelphia PA, 1973.4 Dickinson B.E., T. Lalk, D. G. Hervey, ~Characterization of a Fuel CelL/Battery HybridSystemfor Electric Vehicle Applications", SAEPaper 931818, published in Special Publication 984, 1993.5 Swan D.H, A.J. Appleby, "Fuel Cells for Electric Vehicles, Knowledge Gaps and DevelopmentPriortities", Proceedings of The Urban Electric Vehicle, Stockholm Sweden. pp 457-468, May1992.6 SwanD.H.. O.A. Velev, I.J. Kakwan,A.C. Ferreria, S. SfinJvasan, AoJ. Appleby, "The Proton Exchange Membrane Fuel Cell - A Strong Candidate as A Power Source for Electric Vehicles", Hydrogen91 Technical Proceedings, International Association for Hydrogen Energy. 1991. 26
  12. 12. Fuel CellStack (amps) O 2O 4O 6O 80 100 120 140 160 0.9 31.5 S~oich0 Ter~0R’ess ---c---- S-~ T=F:~C, P,~r .5, F:L--2 -----o--- S--1.5,T=60 i:~ Bar C, 28 A S--1.5.T..70C, ~-2PAr O > ~.~ S~.5, T,<50C0F~Bar 24.5 -----o S--1.5, T=60C, ~ Bar ~+~ S=1.5, T=70C, ~ Bar C) u. 21 17.6 58 172 268 344 43O 616 602 688 Den~ty (matm~*’2) Figure Fuet 3. CellStack Polarity for Stoichiometry Plot of 1.5 FuelCell Stack (amps) 0 29 40 60 80 100 120 140 0.9 I 31.5 -,F I 0.8 I 28 O>o Stoich. Temp, R’ess >v J< $=2, T=50C, P=2Bar k z4.sg’j =,. S~2, T=60(3, F~2Bar / ¢Do .,- z~ S,=2,T=70 P=2 C, Bar< 0.6 ~X~ S=2, T,,~O C. P=3 Bar L 21 I J. S:=2°T=60 P,=3Bar C, ~+~ S=2, "1"=70 C, P=3 Bar 0.5 17.5 0 86 172 258 344 430 616 602 688 Current Density (ma/cm**2) Figure FuelCellStack 4. Polarity for Stoichiometry Plot of 2.0 27
  13. 13. 400 Slolch, Temp, F~ress S=1.5, "1-=50 P,,,3 Bar C, S=1.5, "1=60 C, P=,3 Mr -- ¯ ~. S-1.5,"I"-70 C~ Bar P’,3 5O 4O 46 60 55 GO ~ 70 75 FU~! Cell $1~¢k Conversion Efficiency [besecE o=t L.H.V=| (%)Figure 7. Specific Power as a Function of Temperature and Efficiency for Stoichiometry of 1.5, Pressure = 3 Bar 400 S~o~,Tecrp, R’e¢~ 360 ~J---- S=2, T=~0 P=2Bar C,"~ 300 ,, ,¢ S=2, T==50 F~ Bar C,E S=2,T=70C, P=2Bar 260E0 0 ~50 --,=_ t i l 0 40 45 50 55 60 65 70 75 Fuel Cell SlackC.,ocwe~on E~ciency on L,i-LV.] (0/~ ~Figure Specific Power a Functionof Temperature Efficiency for Stoichiometry 2.0, 8. as and of Pressure 2 Bar = 28
  14. 14. 8,50 800 550 500 O, 2 4 6 8 10 12 t4 IE 18 20 22 24 26 28 30 32 34 36 Cell # Figure individual 5. Cell Voltages DifferentCurrent at Densities 4O0 S~ich, Temp, Press 35O*--300O=4EC~,~ 150 60 ! 46 60 6$ 6O 65 70 7~ Fuel Call Stack Conversion Efficiency [basedon L. H. V.J (%)Figure6. Specific Power a Functionof Temperature Efficiency for Stoichiometry 1.5, as and of Pressure= 2 Bar 29
  15. 15. 4OO Terrp. R’ess Stoich 36O S=2, T----50 Bar C,P,=3 3O0 S=2, T=60 P=3Bar C, E 250 ¯ ’, S=2,T=70 F~3E~r C, E 200 tO0 5O 40 45 50 55 60 65 70 75 FuelCell Conversion Stack Efficiency[bas~ion LFLV.] (%) Figure SpecificPower a Function Temperature Efficiency Stoichiornetry 2°0, 9. as of and for of Pressure 3 Bar = 40O ~×~x~, x ’ "’ ~. E l Stoich, Temp, P~ess I ! --, --.~.~ S=2,1"-60 C, P~3 B~r [Gross] r t : ~,2, T-60C, P,.2 Bar[Gross] / __j I i .... > S=2. t-60 C P~2 Bzr [Net] i ! = ] -~ S-2. ]’=60 C, P=3 [Net] Bar I 36 46 45 50 56 ~3 ~ 70 76 Fue| Ce|! System Conversion Efficiency [based on LH.V.] (%)Figure O. Comparison Gross Net Fuel Cell System I of and SpecificPower Pressures 2 and3 at of Bar 30