Advances in Fuel Cells and Hydrogen

P.J.R. Pinto, C.M. Rangel

Solar-hydrogen stand-alone systems

Materials for Energy Workshop “Advances in Fuel Cells and Hydrogen” April 2010, Torres Vedras, Portugal Performance of a solar-hydrogen stand-alone system for residential applications P.J.R. PINTO*, C.M. RANGEL* Laboratório Nacional de Energia e Geologia (LNEG), Fuel Cells and Hydrogen Unit, Estrada do Paço do Lumiar 22, 1649-038 Lisboa, Portugal *[email protected]; *[email protected]

ABSTRACT: Hydrogen, as an energy storage medium, is considered a promising solution to overcome the limitation of intermittent renewable energy sources. In this paper, a residential scale solar-hydrogen based stand-alone energy system is designed, modelled and the simulated system performance under real end-use load, representative of standard European domestic electrical energy consumption, and meteorological conditions is analyzed. The sun is the primary energy source of the system and a fuel cell-electrolyzer combination is used as a backup and a long-term storage system. A battery bank is also used as energy buffer and for short time storage. Matlab/Simulink® is used for the overall system modelling and simulation. The results show that the designed solar-hydrogen system is in principle capable of operating autonomously and in a sustainable manner. The designed system is able to convert 7.6% of the total energy irradiated in one year. Keywords: Solar hydrogen, stand-alone system, modelling, energy management, system sizing. RESUMO: O hidrogénio, como um meio de armazenamento de energia, é considerado uma solução promissora para superar a limitação da intermitencia das fontes de energia renováveis. Neste trabalho é dimensionado e modelado um sistema de energia solar-hidrogénio isolado da rede de distribuição de energia eléctrica e à escala residencial. São analisados os resultados da simulação do seu desempenho sob condições climatéricas e de carga, esta última representativa do padrão de consumo de energia eléctrica no sector residencial na Europa. O sol é a fonte primária de energia do sistema e uma combinação pilha de combustível-electrolisador é usada como um sistema de apoio e como um sistema de armazenamento de longo prazo. É também usado um banco de baterias como reserva de energia e de armazenamento de curto prazo. O Matlab / Simulink® é utilizado para a modelação e simulação de todo o sistema. Os resultados mostram que o sistema solar-hidrogénio projetado é, em princípio, capaz de operar autonomamente e de forma sustentável. O sistema projectado é capaz de converter 7,6% do total de energia irradiada num ano. Palavras chave: Hidrogénio solar, sistema autónomo, modelação, gestão de energia, dimensionamento do sistema.

NOMENCLATURE Acronyms AC DC MPPT PEMFC PHOEBUS PV RE

30

alternating current direct current maximum power point tracker proton exchange membrane fuel cell PHOtovoltaik-Elektrolyse-Brennstoffzelle Und Systemtecknik photovoltaic renewable energy

Symbols A B C D DoA DoD E f1 f2 I, i

area (m2) slope of Tafel line (V) equivalent capacitance (F) constant in the mass transfer term (V) days of autonomy depth of discharge energy (kWh) parameter related to Faraday efficiency (mA2cm-4) parameter related to Faraday efficiency current (A)

Ciência & Tecnologia dos Materiais, Vol. 23, n.º 1/2, 2011

Solar-hydrogen stand-alone systems K M MH n N P r R s t T U, u V .. n n α η

thermal correction factor (ΩºC-1) number of moles (mol) metal hydride diode quality factor number of cells or modules power (W) parameter related to ohmic resistance of electrolyte (Ωm2) resistance (Ω) coefficient for overvoltage on electrodes (V) coefficient for overvoltage on electrodes (A-1m2) temperature (ºC) voltage (V) volume (m3) molar flow rate (mols-1) ratio between the current irradiance and the irradiance at standard rating conditions efficiency

Subscripts BAT battery bank BUS DC-bus BUS-BAT from DC-bus to battery bank BUS-EZ from DC-bus to electrolyzer BUS-EZ_MAX electrolyzer maximum input seen from the DC-bus BUS-EZ_MIN electrolyzer minimum input seen from the DC-bus BUS-LOAD from DC-bus to end-use load CELL cell DAI daily average solar irradiation DAL daily average end-use load EZ electrolyzer EZR electrolyzer rated EZ_EL electrolyzer electrode EZ_H2 hydrogen produced by the electrolyzer EZ_OFF limit for electrolyzer operation EZ_ON level to activate electrolyzer EZ_REF electrolyzer reference EZ_R electrolyzer reversible F Faraday FC fuel cell FCR fuel cell rated FC-BUS from fuel cell to DC-bus FC-BUS_MAX fuel cell maximum output seen from the DC-bus FC_H2 hydrogen consumed by the fuel cell FC_L fuel cell limiting FC_N fuel cell internal FC_O fuel cell exchange FC_OFF limit for fuel cell operation FC_ON level to activate fuel cell FC_REF fuel cell reference FC_R fuel cell reversible M membrane and contact MAI month average solar irradiation MAL month average end-use load MAX maximum acceptable


P.J.R. Pinto, C.M. Rangel MH MH_I MH_T MIN NET PV PV-BUS PV_L PV_MIN0 PV_MIN1 PV_O PV_STC REF S SH SOC SOC_MAX SOC_MIN YAI

metal hydride initial hydrogen content in the metal hydride total hydrogen capacity in the metal hydride minimum acceptable production minus consumption photovoltaic from photovoltaic array to DC-bus photocurrent minimum initial photovoltaic array minimum photovoltaic array photovoltaic module reverse saturation photovoltaic module electrical values under standard test conditions at standard rating conditions series shunt state of charge maximum state of charge minimum state of charge yearly average solar irradiation

Constants F Faraday constant (F=96485Cmol-1) LHV low heating value of hydrogen (LHV=3kWhm-3) z number of electrons transferred per reaction (z=2 1. INTRODUCTION The combined effect of the rising prices of fossil fuels and the growing awareness of the impact of environmental pollution has stimulated great interest in alternative and clean energy sources, such as solar and wind. They produce little or no environmental pollution, are unlimited and available almost everywhere on the earth. However, their inherent intermittency and variability renders storage of energy indispensable for fitting time-varying load demand. A hydrogen system comprising a hydrogen producing unit (electrolyzer), a hydrogen storage unit and a hydrogen utilizing unit (fuel cell) is considered a promising solution for renewable energy storage. Compared to commonly used battery storage, hydrogen storage has higher storage density and lesser environmental effect. During the past decade, the interest on the concept of integrating renewable energy sources with hydrogen storage systems for stand-alone applications has increased. Experimental results of prototype or actual systems have been reported by several researchers [1-6] focusing on system performance and viability. Many simulation studies focusing on system energy management and control strategies [7-10], sizing [11-15] and modeling [16-19] are also available in the literature. The integrated system unit-sizing and energy flow control are two major research challenges because of their interdependence, the temporal mismatch between energy supply and demand, the non-linear characteristics of the individual units of the system and the high number of variables and param-

31

P.J.R. Pinto, C.M. Rangel eters. To properly size the integrated system units, researchers [11,12] used the “yearly average monthly” method in which the unit-sizing is conducted by considering the yearly averaged monthly values of meteorological and load demand data. When the goal is to find the optimum unit-size and/or the control strategy for the integrated system, researchers [13-15] used time-series simulation coupled with optimization methods. In most studies, the energy flow is controlled through the state of charge of the battery bank, which is commonly used for short-term energy storage. Due to their inherent characteristics (poor dynamic response of fuel cells and intermittent nature of RE sources), stand-alone RE hydrogen systems must include a short-term response energy storage device. The focus of this paper is to examine the performance of a stand-alone solar-hydrogen system for residential applications in terms of the final amount of hydrogen in the storage unit, the level of utilization and operating pattern of battery bank, electrolyzer and fuel cell. Starting with a description of the system to be investigated, the dynamic model of each unit is briefly described. The major equations are provided and the key model parameters are given. Simulation results obtained using Matlab/Simulink® are then presented and analyzed.

Solar-hydrogen stand-alone systems 2.2. Energy management and control strategy The main requirements of the designed energy management strategy for the stand-alone solar-hydrogen system are to satisfy the end-use load under variable meteorological conditions and to manage the energy flow while ensuring efficient operation of the different system units. The PV-generated energy is primarily used to meet the end-use load. Any excess of energy is used to charge the battery bank or to produce hydrogen through water electrolysis. Any shortage of energy is met by the PEMFC and/or the battery bank. The inherent variability in the solar generation produces variability in the operation of the PEMFC and the water electrolyzer. Since stable operation of these energy systems is vital for efficiency, lifetime and cost, the management strategy uses the battery bank to mitigate the effects of energy fluctuations on their operating pattern. In this implementation, a double hysteresis loop control strategy is used for this purpose, see Fig. 2. The on/off switching actions of the PEMFC and the water electrolyzer are determined by DC-bus voltage, uBUS, where UEZ_ON, UEZ_OFF, UFC_ON and UFC_OFF are the key decision parameters.

2. SYSTEM DESCRIPTION 2.1. Configuration Fig. 1 shows the schematic diagram of the considered system, which consists of a photovoltaic (PV) array for solar energy conversion, an alkaline water electrolyzer for hydrogen production, a metal-hydride for hydrogen storage, a proton exchange membrane fuel cell (PEMFC) for hydrogen energy conversion and a battery bank for energy buffering. All the units are linked via a common DC-bus, which allows energy to be managed between sources, storages and end-use load. The battery bank is directly connected to the DC-bus, while the PV generator, the electrolyzer and the PEMFC are coupled to the DC-bus via boost-based maximum power point tracker (MPPT), buck and boost converters, respectively. The end-use load is powered by the DC-bus through a DC-AC converter (inverter).

Fig. 2. Representation of the operation of the considered solar-hydrogen stand-alone system.

To apply the energy management strategy easily, all quantities of the energy conversion units (PV array, PEMFC and water electrolyzer) and the end-use load are referred into the DC-bus by taking the losses associated with the power conversion units into consideration. Control flow is shown in Fig. 3 and is based on the following instantaneous current balance equation at the DC-bus: iNET = iBUS-EZ – iFC-BUS ± iBUS-BAT (A)

(1)

where iNET is the PV-generated current, iPV-BUS, minus the enduse load current, iBUS-LOAD.

Fig. 1. Layout of the considered solar-hydrogen stand-alone system.

32

Basically, at any given time, any excess PV-generated energy (iNET > 0) is supplied to the battery bank to charge it or to the water electrolyzer to produce hydrogen. The electrolyzer is activated when MHSOC < MHSOC_MAX, uBUS >= UEZ_ON and iNET >= IBUS-EZ_MIN (a minimum electrolyzer threshold input current has been assigned for safety reasons). If 0 < iNET < IBUS-EZ_MIN and in the previous time step the electrolyzer was operating, then the electrolyzer is not disconnected and the deficit current, iNET – IBUS-EZ_MIN, is provided by the battery bank. On the other hand, if iNET > IBUS-EZ_MAX, then the electrolyzer utilizes current equal to IBUS-EZ_MAX and the surplus current, iNET – IBUS-EZ_MAX, is used to charge the battery bank without exceeding UMAX




(a)

Fig. 3. Control logic flow chart.

to avoid overcharging. The electrolyzer on/off state from the previous time step is maintained unaltered for UEZ_OFF MHSOC_MIN and uBUS IFC-BUS_MAX, then the remaining current deficit is covered by the battery bank. The PEMFC on/off state from the previous time step is maintained unaltered for UFC_ON < uBUS 0) or discharged (iNET < 0) depending on the net current level. 2.3. Unit-Sizing The solar-hydrogen stand-alone system is assumed to be installed in a house located near Lisbon, Portugal (Lat.: 38.84; Long.: -9.12). Local hourly averaged solar irradiation intensity on a surface tilted at 38.8 degrees and ambient temperature profiles are used for modeling the performance of the PV generator. These data are derived from the measured data available in Ref. [20]. The end-use load is taken from Ref. [21] and it represents standard European domestic electrical energy consumption. The data is from a mother and two children dwelling, averaged over a 5 minute time interval. The monthly averaged solar irradiation on the tilted plane and end-use load energies for the time period spanned by the simulation run (one year) are shown in Fig. 4. The key data are listed in Table 1. !

(b)

Fig. 4. Monthly averaged (a) solar irradiation on the tilted plane and (b) enduse load energies.

Table 1. End-use load and solar irradiation. End-use load Daily average energy EDAL (kWhday-1) Yearly average energy (kWhm-2 per year) Total energy consumption (kWh) Solar irradiation Daily average energy EDAI (kWhday-1) Yearly average energy EYAI (kWhm-2 per year) Total energy usable (kWhm-2)

4 121 1453

Based on the designed energy management strategy and on the yearly solar irradiation and end-use load energies profiles just characterized, the following unit-sizing procedure is used to determine the size of the battery bank, the PEMFC, the PV array, the alkaline water electrolyzer and the metal-hydride. In this study, the assumed efficiencies of the main system units are as follows: battery bank (ηBAT), 85%; PEMFC (ηFC), 35%; PV array (ηPV), 14%; alkaline water electrolyzer (ηEZ), 75%; DC-DC converters (ηPV-BUS, ηBUS-EZ and ηFC-BUS), 95%; DC-AC converter (ηBUS-LOAD), 90%. The battery bank energy capacity EBAT is determined by:

E BAT =

E DAL DoA (kWh) " BUS#LOAD" BAT DoD

(2)

daily average end-use load energy (kWhwhere EDAL is Ethe DAL PFCR-1=), DoA is the (W ) day " " days 24hof autonomy (one day of autonomy BUS#LOAD

APV _ MIND =


8 252 3029

! A

FC#BUS

E DAL E DAI " PV " PV #BUS" BUS#LOAD

=A

+

(m ) 2

33

PFCR =


!

is considered here) and DoD is the depth of discharge of the battery bank (UEZ_ON-UFC_ON in this case). E DAL DoA ! (kWh) E BAT = The PEMFC is"also sized for the daily average end-use load " BUS#LOAD DoD BAT energy and its rated power is determined by:

PFCR =

!

!

! ! !

E DAL (W ) " BUS#LOAD" FC#BUS 24h

!

(3) ! A safety margin of 25% calculated values of the E DALis added to the = energy capacity m 2 PEMFC APV _ MINDbank ( ) battery and the rated power. E E" PV " PV #BUSDoA " BUS#LOAD (kWh) E BAT = DAI DAL DoD " BUS#LOAD The PV array size " isBAT determined through the following steps: APV MIN 1 = APV MIN 0 + • In order to assure the autonomy of the system, the!daily E average = end-useDALload energy P (W )must come directly from the FCR # # # #the ) (1" " BUS#LOAD 24h BUS"EZ EZ " FCFC#BUS FC"BUS PV array. Hence, minimum $ initial area of the PV array #BUS"EZ #EZ #FC #FC"BUS 12 months is determined by: September E 2 ) (E MAI = # PV # PV "BUS "DALE MAL / # BUS"LOAD m 2)(m APV _ MIND ( ) March E DAI " PV " PV #BUS" BUS#LOAD ! (4)

%

! !

! !

!

! ! ! ! !

!

! !! ! ! ! ! ! ! ! !

! ! !!

! !! ! !

!

! ! ! !!

!

! ! !

" BUS#EZ " EZ due to the mismatch between energy genera• However, VMHA $ PV MIN 1 = APV MIN 0 + tionLHV and consumption, there are additional energy losses October 3 in the hydrogen loop. on the)(m initial E DAL (E #DoA EBased /)" BUS#LOAD ) area, APV_MIN0, MAI " PV " PV #BUS MAL (kWh) # # # =# E BAT(1" February BUS"EZ EZ FC FC"BUS E DoA of the PV array, the monthly average mismatch between $ DAL " BAT DoD (kWh) E BAT =#" BUS#LOAD # #FC #and 12DoD months BUS"EZenergy FC"BUS solar load energy is shown in! Fig. "EZBUS#LOAD " BATend-user September $ we (T #Tref ) ]+iPV Rsaverages (" [ uPV +KiPVmonthly ) / anT ' to =5."In iPV this #1) 2 simplify the iPV % (Eicase, # PVuse _L # PV 0& EPVeDAL MAI_# "BUS " E MAL / # BUS"LOAD )(m ) March % ( the system proW PFCR = calculations. From March to September, ( ) E"DAL 24h " = P W BUS#LOAD FC#BUS ( Rload ) demand. The excess of FCR duces kiPVenergy Trefthan + ithe a u PVmore (T"#FC#BUS )24h PV s "+BUS#LOAD ! " BUS#EZ "be EZ used to produce (A) # energy can hydrogen. From October to VMH $R sh E DAL of energy 2and the system needs LHVthere is a deficit February, = APVOctober (m 2 ) the3 load demand. _ MIND E DAL E " " to consume hydrogen in to (cover APV _ MIND(E=MAI "DAIPV "PVPV #BUS m ) )(m ) PV #BUS BUS#LOAD #" Eorder MAL / " BUS#LOAD February # PV "BUSEiPVDAIuPV "PVPV " " BUS#LOAD ! the The required to accommodate = minimum APV) #BUSarea iPV "BUS (array u energy losses in the hydrogen loop can be calculated by: BUS APV MIN 1 = APV MIN 0 + $+(" [ uPV +KiPV (T #Tref ) ]+iPV Rs ) / anT ' A = A "i1PV _ L #PViPVMIN_ 00& e #1) iPV PV= MIN % ( i u FC"REF BUS (1" iFC#=BUS"EZ #EZ #FC #(W ) ) $ PFC = uFC FC"BUS +#ki# a u#PVBUS"EZ #T ##Tref ) +) iPV Rs (1" PV FC"BUS EZ( FC FC"BUS #BUS"EZ # #EZ #FC #FC"BUS 12 months $ (A) ! #BUS"EZ #EZ #FC #FC"BUS September Rsh12 months 2 (E MAI # PV # PV "BUS " E MAL # BUS"LOAD )(m 2) ) %September # t /+t & (5) 2 r2T (E+MAIr1# # PVi "BUS " E MAL )(m 1 / # BUS"LOAD 2 /TEZ +t 3TEZ) %=March PVEZ +uCELL u + slog i "1 % ( MarchEZ _ R EZ EZ # PV "BUS * i AEZiPV_ ELuPV ( A) $ AEZ_EL '(V ) = PV "BUS" + V BUS#EZ "u EZ BUS uCELL"(V )$ ,uEZ =MHN"EZBUS#EZ EZ VMH LHV $ OctoberLHV " u BUS # E MAL / " BUS#LOAD )(m 3 ) (E MAI "iFC"REF = uFC(E iFC =" PV " PV #BUS ) /" PFC October February #(W E MAL )(m 3 ) MAI PV PV #BUS BUS#LOAD # FC"BUS February ! #i + i & FC FC _ N "B uFC = u FC _ R " (iFC + iFC $_ N )("R[ uMPV +Ki % ( T #T +i R / anT ' ( ref ) ] PV s ) PV ln '#1)'( 2 )iPV = "iPV _ L # iPVr_r0 &$%Te (" [ uPV +KiPV (T$#Tref# t) ]i+iFC+t PV_RO s ) / anT & +t) 3TEZ = " i # i i 1_ 02& eEZ 1 2 /TEZ#1 PV PV _ L PV +uCELL =# u EZ _ R + %& iEZ + slog% iEZ "1( ( ! '(V ) *+ D lna%1" + ikiFCPVA_ NEZ u PViFC (T(_(VEL# T) ref ) + iPV$Rs AEZ_EL + # a$ u PV +iFCki_PVL (T' # Tref ) + iPV Rs (A) ,uEZ#= N EZ uCELL (V )Rsh (A) Rsh • ! N EZ # PV n EZ _iH 2 = "= iEZ iPV uPV A "BUS F ( ) PV "BUS #zF u iPV uPV #i + i & iPV "BUS = PV "BUS BUS ( A) FC FC _ N = u + i _ N ) RMthe" B u ( uiBUS FC _ R " ( FC ln % Fig. 5.FCPV-generated energyFCminus end-use load energy iFC _ O ' (net energy), • $ N referred i iFC"REF u BUS = theFCDC-bus. n FC _ H 2 into ! i#FC FC =ii i u& (W ) PFC ="uFFCZF FC FC _ N BUS FC"REF # + D ln 1" ) )to handle the maximal power (W ((Vable FC"BUS The Pelectrolyzer be FC = u FC i%FC =should iFC # FC"BUS $ 2The ' _L of the PV array. maximum possible rated power of the iEZ / AEZis_ ELgiven ( ) electrolyzer by: ) # t +t /T +t T 2 ! & "F • 0...1) rf r(T + 12 2 EZ iEZ + slog%# t 1 +t 2 /TEZ +t 3TEZ2 iEZ "1(& f+)1u+CELL iEZ=/ uAEZEZN ( __REL EZ) r r T EZ 3 EZ iAEZ1PVEZ2 _–EZ EZ _N H2 = PEZRn+*= P"uFSTC –+ h uCELL iEZh+ slog(W) iEZ "1(6) $% 1 2 AEZ_EL '((V ) ELBUS PV= PV BUS–EZ EZ _zF R AEZ _ EL A +* $ '(V ) EZ_EL (V ) dt # J" ,+uEZ =MN EZ uCELL + J " dt _I EZ _ H 2 FC _ H 2 ) uEZ== NMH ! MH•,50C (0...1) EZ NuFCCELL (VM n i FC _ H 2 = MH _T FC 34 " F ZF # & # " i 2 + i 1 R " B %#& iFC + iFC _ N (& = u u ( ) FC FC _ R FC FC _ N M ln i + i i dt )FC (VFC)_ N u =U + i R +

%

[

]

%

[

]

% %

[ [

] ]

E DAL (W ) Solar-hydrogen stand-alone systems " BUS#LOAD" FC#BUS 24h

However, due to theEinherent intermittency and variability of DAL = APV _ MIND (m 2to) the electrolyzer would solar irradiation, the power supplied E DAI " PV " PV #BUS" BUS#LOAD be below the PV maximum power output most of the time. In addition, an electrolyzer with a higher power rating means APV MIN = APV MIN 0 + also a 1higher minimum operation threshold and thus lower operation time. Hence, a more economical option may be to ) size(1" the#electrolyzer atFC"BUS a power BUS"EZ # EZ # FC # $lower than the PV maximum #BUS"EZ #output # # 12 months power (typically between 60 and 80% [15]). In this EZ FC FC"BUS September case, a percentage value of 70% is applied. 2 %March (E MAI #PV #PV "BUS " E MAL /#BUS"LOAD )(m ) The hydrogen capacity of the metal hydride is determined by: " " VMH BUS#EZ EZ $ LHV October

%

February

(E MAI " PV " PV #BUS # E MAL / " BUS#LOAD )(m 3 )

(7)

where LHV is the lower (3kWhm-3). +Ki PV (T #T refvalue / anThydrogen $ (" [ uPVheating ' ) ]+iPV Rs )of #1) iPV = "iPV _ L # iPV _ 0 & e The bounds of summation in the above expression represent % ( the months for which there is an excess of energy. These new a u PV + kiPV (T # Tref ) + iPV Rs bounds PV array area. In order to # are a result of increased (A) accommodate the Rmetal hydride state of charge initial level sh and seasonal fluctuation, a factor of two is applied to the cal# i u culated value. iPV "BUS = PV "BUS PV PV ( A) u BUS Following the unit-sizing procedure described in this section, the sizes of the battery bank, the PEMFC, the PV array, the aliFC"REF u BUS kaline and) the metal-hydride are estimated i =electrolyzer (W PFC = uwater FC FC # FC"BUS and listed in Table 2.

[

]

Table 2. System unit-sizing. # ) & rrT t +t /T +t T 2 +uCELL = u EZ _ R + 1 2 EZ iEZ + slog% 1 2 EZ 3 EZ iEZ "1( Battery bank * AEZ _ EL AEZ_EL $ '(V ) + Energy capacity EBAT (kWh) 33.9 ,uEZ = N EZ uCELL (V ) PEMFC Rated power PFCR (W) PV array #i + i & Area A (m2) FC FC _ N uFC = u FC _PV_MIN1 ( R " i FC + iFC _ N RM " Bln % Electrolyzer $ iFC _ O ' Rated power (W)& # iPEZR i Metal + hydride D ln%1" FC FC _ N ((V ) iFC _ LV ' (m3) $ capacity Hydrogen MH

(

)

500 27.4 2432 1113

• The PV array, N battery bank and electrolyzer unit sizes used in n EZ _ H 2 = " F EZ iEZ the simulation zFare 28.1m2, 38.4kWh and 2500W, respectively. The sizes of the PV array and battery bank were chosen based • commercially on available units, while the size of the electroN FC = chosen n FC _ H 2was iFC to be a scaled-down proportion of an eleclyzer " F ZF trolyzer whose empirical model is reported in the literature and used in this 2study. The sizes of all the other units used (iEZ / AEZ _ EL ) in "F the simulation aref 2 (those 0...1) listed in the above table. Further f1 + (iEZ / A information isEZcontained in the next section. _ EL )

M + J" EZ _ H 2 dt # J" FC _ H 2 dt MH 50C = MH _ I MODEL (0...1) 3. SYSTEM’S M MH _T 3.1. PV Array’s and Boost-based MPPT Converter’s & Model # 1 uBAT = U BAT + %iBAT RBAT + iBAT dt () BAT (V ) " c BAT The model of$ the PV array refers 'to the electrical model with one diode. The current-voltage characteristic curve is


%

MAI

March

! ! !

PV "BUS

PV

MAL

!

E DAL PFCR = (W ) ! " " " " stand-alone 24h BUS#EZ EZ Solar-hydrogen systems BUS#LOAD$ FC#BUS ! VMH LHV October 3 E DAL# E MALfive-parameter ! (E / " BUS#LOAD ) described the equation % MAI " PV "following PV #BUS = by m 2 ) )(m implicit APVFebruary ( _ MIND ! E " " " [22]: DAI PV PV #BUS BUS#LOAD !! $ (" [ u +Ki (T #T ) ]+i R ) / anT ' ! # iPV _ 0 &+e #1) iAPV = "iPV=_ A L PV MIN 1 PV MIN 0 % ( ! ! ! a u PV + kiPV (T # Tref ) + iPV Rs #EZ #FC #FC"BUS ) (1" #BUS"EZ (A) # ! $ R12 ! (8) sh #BUS"EZ #EZ #FC #FC"BUS months PV

! ! !

BUS"LOAD

[

PV

ref

PV s

]

2 !! (E MAIthe # PVvalues # PV "BUS of " Ethe / # BUS"LOAD )(m ) Table 3 lists parameters obtained % MAL empirical # March PV "BUS iPV u PV ! iPV "BUS ( A) for the=PV module u BUS Kyocera KC175GHT-2 [23] with a!peak power of 175W in standard conditions, using the procedure " " EZ VMH BUS#EZ $ reported in [22]. ! LHV iFC"REF u BUS P = u i = (W ) October FC maximum FC FC ! !temThe power output of the PV array3 varies with! (E MAI " PV#"FC"BUS PV #BUS # E MAL / " BUS#LOAD )(m ) February and solar irradiation. Therefore, designing efficient perature ! PV systems heavily emphasizes to track the maximum power 2 ) & r$Tthis +t /T perturb +t +Ki PV (T #T ref +i PV Rs ) / anT ' T and observe operating point. rIn a#) ]tclassic (" [ uPV study, +iPVuCELL ( = "=iPVu_EZL _#R i+PV _10 &2e EZ iEZ + slog% 1 2 EZ#1) 3 EZ iEZ "1! (P&O) tracking A algorithm is used ) * AEZ_EL (the model of % _ EL $ involving '(Vthe EZ + nonlinear current-voltage characteristic of the PV module. ! u+ ki (V T ) # Tref ) + iPV Rs ,uEZ =a NuEZ PV CELL PV ( (A) # ! By taking into proper Rsh account the boost-based MPPT conver!! by: sion efficiency, ηPV-BUS, the current to the DC-bus is given September

! ! !

! !

%

[

!

!

!

]

#i + i & # i u iuPVFC"BUS = u=FC _ RPV""BUS (uiFCPV+ PViFC(_ AN )RM " Bln % FC i FC _ N ( BUS $ FC _ O ' # i i & FC FC _parameters. N Table 3.ln PV E model DoA + D 1" ) % ((V (kWh) DALiFC"REF u BUS EPBAT = = iFC _ L ' (W ) FC = u FC iFC $ " BUS#LOAD "#BATFC"BUS DoD α=1; T=25ºC I

! ! ! ! !

(A)

8.0923

! !

! ! ! ! !

! !

!

!

!

! !

!

! !! !

• PV_L )n EZ _ H 2 = " F NEEZDAL &! iIEZr T (A) (W ) # t +t9.6024x10 r /T +t -12 T2 = P 1 PV_O +FCR uCELL " = u EZ _ RzF +" 2 EZ24h iEZ + slog% 1 2 EZ 3-3 EZ iEZ "1! (! BUS#LOAD FC#BUS * AnEZ(V/K) AEZ_EL 3.5897x10 $ '(V ) _ EL +• ! R (Ω) 99.158 N EZNuCELL ,nuFCEZ_ H=2 = FC (V )SH ! i E DAL ! APV _ MIND =" F ZF FC RS (Ω) m 2 ) 0.282 ( ! E DAI " PV " PV #BUS" BUS#LOAD

α=1; 25-75ºC

-3 K (ΩºC-1) # i +1.1218x10 iFC _ N & (iu1EZ=/ AAEZPV"_ MIN EL ) + FC A MIN 0 + i f ( 0...1 " = i R " B uPV % ( ) ( ) F FC FC2_ N M ln f1 + FC AEZFC_ EL )T=25ºC (iEZ_ R / α=0.2; $ iFC _ O ' 2

! ! !

!! ! (9)

IPV_O (A)

1.4356x10-11

! ! !

(1" #BUS"EZ # #iEZ #i FC #FC"BUS & ) $ FC FC _ N +D ) %1" ((V + J " #BUS"EZ #EZln#M # 12 MH _ I EZ _months H 2 dt # J" FC _ H 2 dt ! arFC FC"BUS iFC 22 AMH total number of are needed to cover (0...1 ) the PV ' _ L modules 50C = $ ! ! September Mthe 2(see Table 2). MH _T ray area determined in previous section %March (E MAI #PV #PV "BUS " E MAL /#BUS"LOAD )(m ) • N n EZ _ H 2 = " F #EZ iLoop 3.2. Hydrogen Model & ! EZ 1 ! ! U BAT"+EZzF iBAT RBAT + iBAT dt () BAT (V ) uBAT "=BUS#EZ % " ! V $is the set cof MH $ ' Hydrogen loop electrical energy BATunits converting LHV • intoOctober chemical storing it and N FC energy via water electrolysis, 3 n FC _ H 2 =(E MAI iFCPVto "back # E MAL / "energy )(musing ) a fuel cell. % PV " #BUS BUS#LOADby ! converting electrical February " it F ZF !

Based on the working conditions, the energy flow controller 2 $ (" [ u PV +Ki PV (T #T ref ) ]+i PV Rs ) / anT ' produces current reference for the electrolyzer or the PEMiiEZPV/_aA iPV = " #1) ( EZ i _ PV EL _)0 & e L # % ( " f 0...1 ( ) F Current controlled 2 power converters could then use! FC. !these f1 + (iEZ / AEZ _ EL ) ! current to control the electrolyzer/PEMFC current + ki T # T + i R a ureferences ( ) PV PV ref PV s (A) converter is represented # input/output. In this study, each power RJsh" EZ _ Hwhere Mpower by an ideal the MH _ I +source 2 dt # J" FC _ratio H 2 dt from its output power MH (0...1[8]. ) Hence, !input/ 50C = to its input power is dictated by its efficiency M MH _T ! # PV "BUSreferences iPV uPV output can thus be determined by: iPV "BUS =power A ( ) u#BUS & i = h iEZ REF1hBUS i(W) (10) PuEZ ==uU EZ EZ + %i BUS–EZ ! " BAT BAT BAT RBAT + BAT dt () BAT (V ) ! c $i ' BAT ! u PFC = uFC iFC = FC"REF BUS (W ) # FC"BUS (11) ! ) # t +t /T +t T 2 & rrT +uCELL = u EZ _ R + 1 2 EZ iEZ + slog% 1 2 EZ 3 EZ iEZ "1( Materiais,$ Vol. 23,An.º 1/2, 2011 *Ciência & Tecnologia AEZ _dos '(V ) EL EZ_EL + ,uEZ = N EZ uCELL (V )

[

]

iPV =October "iPV _ L # iPV _ 0 & e #1) %" E # E / " ( 3) )(m (E MAI "EPVDAL 2 DAL PV #BUS MAL BUS#LOAD A = m = PFCR E DAL (W ) ( 2 ) PVFebruary _ MIND T" #PV Tref#BUS iPV Rs ( m ) a "uBUS#LOAD APV _ MIND = +Eki "(PV ) "+BUS#LOAD 24h PV PV DAI (A) # E DAI "FC#BUS PV " PV #BUS " BUS#LOAD P.J.R. Pinto, C.M. Rangel R$sh (" [ uPV +KiPV (T #Tref ) ]+iPV Rs ) / anT '

%

[

]

iPV = "iPV= _AL # iPV _ 0 + #1) & eE A 2 DAL ( PV MIN 1 PV MIN 0 % A = m ( ) = A + PV _ MIND The model of PV MIN 1# PV ithe MINu 0 electrolyzer is developed for a scaled-down PV E "BUS PV PV " " " DAI PV PV #BUS BUS#LOAD iPV "BUS a = u + ki (T (#AT) ) + i R PV u PV of the ref PHOEBUS version (2.5kW) BUS #EZ #FC #FC"BUS ) PV s (A)electrolyzer (26kW, 21 (1" # #BUS"EZ $ unit is an advanced type # # # (1"7#bar) cells, reported in [24]. This R EZ FCsh FC"BUS ) #PV #=BUS"EZ #APV#MIN + 12 months $ A BUS"EZ FC"BUS MIN 1 EZ FC 0 of alkaline electrolyzer having low voltage-high current re#BUS"EZ # # # 12 months iFC"REF u BUS FC"BUS PFCSeptember = uFCEZ iFC FC = (W ) lationship. The current-voltage characteristic September # # " E / # )(m 22 ) of the electro(E % # PVMAI"BUS#PViPV uPVPV"BUS MAL BUS"LOAD FC"BUS March (E # " E / # # # # ) (1" = A ilyzer %PV March ( ) MAL $ BUS"LOAD )(m ) MAI PV PV "BUS BUS"EZ EZ FC FC"BUS "BUSis expressed u BUS as: # #EZ #FC #FC"BUS 12 months ) BUS"EZ # t +t /T +t T 2 & " BUS#EZ " EZ r1r2TEZ September 1 2 EZ 3 EZ V 2 +uMH " BUS#EZ " +#$ # iEZ +" slog u(E i "1 % ( CELL = EZ _ R EZ EZ E MAL / # BUS"LOAD % u LHVMAI i$ PV PV "BUS *V A AEZ_EL)(m ) FC"REF MH $ '(V ) EZ _ EL BUS March = u i = (W ) P LHV FC FC FC October + 3 #) FC"BUS = N(E uEZOctober "(V ,% EZ uMAI CELL PV " PV #BUS # E MAL / " BUS#LOAD )(m ) (12) February (E " PV " PV #BUS # E MAL / " BUS#LOAD )(m 3 ) % MAI " " BUS#EZ EZ February VMH $ ) values LHV of the # t +t & 4 The r1r$2empirical T(EZ /T are+tlisted T 2 in Table " [ u +Ki (T parameters #T ) ]+i 1 R ) /2anT EZ ' 3 EZ +PVuOctober = u + i + slog iEZ "1( % = " i # i e #1 iand & ) CELL PV _ EZ _ R PV _ 0 $ (" [ u EZ +Ki (T #T# ) ]+i R ) / anT & 3' L were obtained from measurements performed at an i + i # E MAL / "$FC )(m (V ) A0 %&PV FC _A N EZ_EL % BUS#LOAD == "uiPV(E_ LMAI #"iPVPV_" #1()) iu*PV February 'operEZe#BUS _ EL FC _ R " (i FC + i%FC _ N ) RM " Bln % +FC temperature ating of 80ºC. $ iFC _ O (' ( u kiPV(V(T) # Tref ) + iPV Rs u PV ,uEZ =a N EZ +CELL + ki +(Ti#TPV R) ]s+i(A) a u PV ) developed # of PV The#model based R is / anT $TN #(&" [Tu ref+Ki ' on the empirical ithe i (PEMFC FC FC _R (A)) #1) iPV =#+"DiPV ln % _ 1" L # iPV _ 0 & esh((V ) current-voltagei equation reported in [25] and %Rsh ( is expressed as:

[

]

[[

PV

PV

ref

PV s

PV

PV

ref

PV s

ref

PV s

PV

$

FC _ L

[

]]

PV

'

]

kiiPVPV(uTPV# Tref ) + iPV Rs# iFC + iFC _ N & a u#PV +"BUS i•uPVFC"BUS " ( #= u=FC#_PV (iiiFCPV +uRPViFC((_AAN )) RM " Bln %(A)i N R PV "BUS inPVEZ"BUS uEZBUS _ H 2 = "F $ FC _ O ' EZ sh u zFBUS # i i & FC _ N + D# lnNPV%1" i FC u ((V ) • "BUS iFC"REF PV u PV BUS = u=FC iFC$FC= iiFC"REF ) inPPVFC iFC _ Lu(BUS FC _ H 2 = 'A)(W FC# (13) = uFC"iFCZF =u BUS (W ) PFC"BUS FC"BUS F # FC"BUS • Table 5 lists Nthe values of the empirical parameters obtained 2 2 )n EZ _ Hi2 =/ A iFC"REF irEZTustack " F EZ # +t /Ttemperature & BUS r ( ) for a 500W under a EZ i EZPEMFC _= EL = u (W ) %#at 1stack P 1 2 EZ 2 EZ +t 3TEZ 2 ) FC zF +uFFCCELL =FCu EZ + i + slog i of"155ºC, (& " f 0...1 r t r T +t /T +t T ( _R EZ ) # 1 FC"BUS 2 2 EZ 1 2 EZ 3 EZ EZ humidifier temperature of 45ºC, and a use of hydrogen and air, (V ) *+uCELL + i + slog i "1 A A f1 +=(iu / A % ( ) $ ' EZ EZEZ _ REZ _ EL EZ _ EL EZ EZ_EL AEZand AEZ_ELpressure [26]. $ '(V ) +*• _ EL 25% at atmospheric respectively, of(V 50 N u ) FC +,nuFCEZ_ H=2 N EZ CELL i = ),uEZ = N EZ #_tH 2+t & (V +FCJ)" EZ_ H(10) dt # J" FC(12) dt 2 /TEZ an MuMHCELL +t TEZ2 _ I r1 r2TEZ 2 ZF Combining F equations = u" +MH (0...1) 3expression uCELL iEZ + and slog% 1 yields iEZ "1( link50C = EZ _ R + * the electrolyzer’s AEZ _M AEZ_ELconverter’s output _T ing current to$ the buck '(V ) EL MH + 2 power which is(Vsolved using the# Newton-Raphson method. ) / AuEZCELL ,uEZ =(iEZN iFC + iFC _ N & EZ _ EL ) & # & f 2_(N0...1 The voltage is Bthen obtained by substituting u"FCF =electrolyzer’s u FC _ R "#(iFC + iFC % ( 1RM)" i + i ) ln FC _ N uBAT +_ N ) RM ""iBAT ) BAT ) '( +FC(BAT i_EZR+" / %A(iBAT =f=1 uU iEZFC_REL+BAT Blndt$% (FC iFC (V )iFCof FC estimated _ O current the electrolyzer’s in equation cthe $ value BAT $ ' iFC _ O ' # i from & (12). Similarly, equations (11) and (13) we obtain the FC iFC _ N + D lnoperating 1" ) # J".#FCThe iFC +rates iFC _ N &of hydrogen pro%#M "N EZ(&(V iFCJ_point PEMFC MHi_FC I + _ H 2idt -u _ H 2 dt FC % uMH =50CuDFC=ln_ R$%1" " (iFCiFC+_iLFC _'(N(V "B () (0...1 )R)MFCconsumption FC + ln iFC _ L M duction (electrolyzer) can then $ iFC _ O (PEMFC) ' $ ' and MH _T be computed using the following expressions: # Ni i & • EZFC FC _ N n• EZ _+H 2D=ln "%F1" & N#EZ iEZ ((V )1 iiFCEZR nuEZBAT_ H=2 = $F +zF U" iBAT iBAT dt () BAT (V ) _ LBAT '+ % " BAT -1 zF (molsc BAT ) (Hydrogen $ ' production rate) (14) • NN • EZi = "N FC FC _ H 2 = FC nn• EZ iEZ FC n FC __ HH 22 = "FF ZF zF iFC (mols-1) (Hydrogen consumption rate)(15) " F ZF 2 • The Faraday of the PEMFC was assumed constant i = / ANEZFC_efficiency ( ELi) 2 n FC _ H 2 EZ FC "F equal f 2 (0...1 iEZ /"AtoEZZF ( and the)Faraday efficiency of the electro_0.9, EL ) while "F f1 + (iEZ F/ AEZ _ EL ) f 2 (0...1) lyzerf1 is + (described iEZ / AEZ _ EL by ) the following expression [24]: 2

(iEZ / AMEZMH_ EL_ I)+ J"f EZ(0...1 _ H 2 dt # J" FC _ H 2 dt " ) MH F 50C = M MH _ I + J"2EZ _ H 2 dt # J" FC _ H 2 dt ( 0...1) + i / A MHf50C = ( ) M (0...1) 1 EZ EZ _ EL MH _T M MH _T J" FC _ H 2&dt [24]. M MH# _ I + J" EZmodel _ H 2 dt 1 #parameters Table 4. Electrolyzer (0...1 # 50C = iBAT dt (&) BAT (V )) uMH 1 " BAT = U BAT + % i BAT RBAT + _T uBAT = U2BAT + $%iBAT RBATM+MHc BAT iBAT dt '() BAT (V ) " AEZ_EL (m ) 0.25c BAT s (V) ' $

f1 (mA2cm-4) 250 # 1 -4 fu2 (mA=2cm ) 0.96 U BAT + %iBAT RBAT + BAT $ NEZ 2 c BAT 2 r1 (Ωm ) 8.05x10-5 2 -1 r2 (Ωm ºC ) -2.5x10-7

"

t1 (m2A-1) & -1 2 ti2 (mdt ºCA ) (V ) () BAT BAT t3 (m2ºC'2A-1) uEZ_R (V)

(16)

0.185 -1.002 8.424 247.3 1.184

35

!

!

!

(ViFC ) _N & ,uEZ = N EZ#uCELLiFC + D ln%1" ((V ) iFC _ L ' $ P.J.R. Pinto, C.M. Rangel • #i + i & N EZ FC FC _ N EZ _= H 2u= " F " (i iEZ unFC ( FC _ R FC + iFC _ N ) RM " Bln % Table 5. FuelzFcell model parameters $ i[26]. ' FC _ O # i i & A• (V) 1.35 iO (A) FC FC _ N N1" FC + D ln (V ) % ( = n i FC _ H 2 B (V) 1.19 RM (Ω) FC "$F ZF iFC _ L ' iL (A) i•N (A)

uFC_R (V)

100 0.23


6.54x10-3 42x10-3 27.1

! !

N EZ 2 n EZ _ H(2iEZ= /"AF EZ _ EL )iEZ f (0...1) "F zF 2 The f1metal + (iEZ hydride / AEZ _ EL ) was modeled as a simple hydrogen summation unit as described by the following expression: • N n FC _ H 2 = M FC iFC+ J" MH _ I EZ _ H 2 dt # J" FC _ H 2 dt MH 50C = " F ZF (0...1) M MH _T (17) 2 (iEZ / AEZ _#EL ) f (0...1) " & F 2 1 3.3. Battery / A iBank ) BATModel uBATf1=+U(iBAT + iBAT dt () BAT (V ) EZ + %EZ _ ELR BAT c $ ' A simplified model of theBATlead-acid battery bank can be constructed byMan ideal constant voltage source, UBAT, in series with MH _ I + J" EZ _ H 2 dt # J" FC _ H 2 dt MHequivalent (0...1an) equivalent ca50C = an internalMresistance, RBAT, and MH _T pacitance, CBAT [27]. Knowing the current, voltage is given by:

!

# 1 uBAT = U BAT + %iBAT RBAT + c $ BAT

! !

! !

! !

!

"

"i

BAT

& dt () BAT (V ) '

(18)

The battery bank efficiency, ηBAT, is considered to be equal to 0.85 (as defined in section 2.3) during charging and equal to 1 during discharging. An Autosil EE 2-800, 2V 800Ah [28], was chosen in this study as base battery for building the battery bank. Table 6 lists the values of the parameters used in the simulation for the battery bank. Table 6. Battery bank model parameters. Voltage source UBAT (V) Capacitance CBAT (F) Internal resistance RBAT (Ω)

42 218182 1.68x10-3

4. SIMULATION RESULTS A simulation in Matlab/Simulink® environment was conducted to evaluate the performance of the solar-hydrogen standalone system over a one-year period under real end-use load and meteorological data. Table 6 lists the values of the control parameters used.

Table 7. System control parameters. Electrolyzer Minimum input current IBUS-EZ_MIN (A) Maximum input current IBUS-EZ_MAX (A) Metal hydride Initial state of charge (0…1) Minimum limit of state of charge MHSOC_MIN (0…1) Maximum limit of state of charge MHSOC_MAX (0…1) Battery bank Rated voltage (V) Initial voltage (V) Maximum charge current (A) Maximum acceptable voltage level UMAX (V) Voltage level to activate the electrolyzer UEZ_ON (V) Voltage limit for electrolyzer UEZ_OFF and PEMFC UFC_OFF operation (V) Voltage level to activate the PEMFC UFC_ON (V) Minimum acceptable voltage level UMIN (V)

Figs. 6, 7 and 8 show the metal hydride state of charge correlation with the battery bank voltage, the power consumed by the electrolyzer and the power supplied by the fuel cell, respectively. We notice that, in accordance to the energy management strategy, the fluctuation in the metal hydride state of charge and the operating pattern of the battery bank, electrolyzer and PEMFC reflect the fluctuation in the system’s net energy. In fact, from February to October, the system produces more energy than the load demand and thus: • the metal hydride state of charge increases; • the battery bank voltage fluctuates mainly between UEZ_OFF and UEZ_ON;

36

9.5 55.7 0.5 0.3 0.9 48 48 80 55.2 52.5 49.9 47.3 42

• the electrolyzer operates whenever the DC-bus voltage constraint and device current limitation are satisfied; • the PEMFC is always disconnected (except at the beginning of February, when the PEMFC is operating and the DC-bus voltage is below UFC_OFF). In contrast, from November to January, there is a deficit of energy and thus: • the metal hydride state of charge decreases; • the battery bank voltage fluctuates mainly between UFC_ON and UFC_OFF; • the electrolyzer is always disconnected; • the PEMFC operates whenever the DC-bus voltage constraints are satisfied.



P.J.R. Pinto, C.M. Rangel is beneficial in terms of electrolyzer performance and lifetime [7, 29]. The control of the PEMFC operation is done to achieve the same objective. Table 8 shows the performance parameters of the electrolyzer and PEMFC. In Fig 8, it can be seen that the PEMFC operates always at rated output power. This is the result from setting the PEMFC to provide the deficit current and to charge the battery bank. Table 8. Electrolyzer and PEMFC performance parameters.

Fig. 6. Metal hydride state of charge and battery bank voltage.

Fig. 7. Metal hydride state of charge and power consumed by the electrolyzer.

Fig. 8. Metal hydride state of charge and power supplied by the PEMFC.

Fig. 6 shows that the battery bank is subject to many chargedischarge partial cycles and to a long period of operation (the period of deficit of energy) at voltages below UEZ_ON (i.e., the battery bank never reached a high recharging level), which shorten its lifetime. The frequency of these charge-discharge partial cycles is higher during the period of excess of energy than during the period of deficit of energy. This is due to the use of the battery bank to maintain electrolyzer operation when the net current becomes lower than the minimum electrolyzer threshold input current. Such an operating policy prolongs the duration of the electrolyzer operation after each activation, reducing the number of start and stops, which


Number of starts Minimum operating time (h) Minimum idling time (h) Total operating time (h)

Electrolyzer 170 5.1 6.5 2352.3

PEMFC 28 7.7 11.7 860.0

On an energy basis, the designed system is able to power the end-use-load continuously given the input energy available from the renewable resource. The battery bank plays an important role in maintaining the energy flows and in extending the electrolyzer operation, supplying approximately 49% of the end-use load demand and 50% of the energy consumed by the electrolyzer. But over the one year period, the system is able to keep the battery bank voltage well above the minimum acceptable limit (i.e., the system operated in a sustainable manner). The energy produced by the PV array is equal to 5.22MWh. About 29% of this energy is supplied to the end-use load, 17% is supplied to the electrolyzer and 54% is used to charge the battery bank. The relatively low percentage of energy supplied to the end-use load is due to the temporal mismatch between energy generation and consumption, which emphasizes the need for energy storage. The difference between the percentages of PV-generated energy given to the electrolyzer and to the battery bank is mainly a result of the implemented energy management strategy. In fact, only about 6% of the energy given to the battery bank is due to the constraint placed by the minimum electrolyzer current input requirement. The PEMFC provides 0.4MWh yearly energy. About 45% of this energy is given to the end-use load and 55% is given to the battery bank. The final state of charge of the metal hydride is 0.53, meaning that the system is in principle capable of operating as a stand-alone unit. 7.4% and 0.2% of the total solar irradiation are found to be supplied to the end-user and stored in the system, respectively, at the end of the testing year. 5. CONCLUSIONS The performance of a residential scale solar-hydrogen based stand-alone energy system over a one-year period under real end-use load and meteorological conditions was analyzed by numerical simulation. The results show that the designed system is able to service the end-use load in a sustainable manner given the input energy available from the renewable resource. Moreover, the end of period surplus hydrogen (+3%) shows that the designed system is capable of stand-alone operation. However, these

37

P.J.R. Pinto, C.M. Rangel results are at the cost of an intense usage of the battery bank, which affects its life span. There are many trade-offs between battery bank and hydrogen loop operation, depending on the energy management strategy employed. By using the current energy management strategy, the total number and magnitude of battery bank charge/discharge cycles could be reduced by, for example, narrowing the hysteresis band size. The drawbacks, however, would be an increased PEMFC operation time, an increased number of the electrolyzer start-up/stop cycles and a lower hydrogen inventory. REFERENCES [1] P. Hollmuller, J. M. Joubert, B. Lachal, K. Yvon, International Journal of Hydrogen Energy 25 (2000) 97_109. [2] T. K. Bose, K. Agbossou, M. Kolhe, J. Hamelin, Available online < http://www.ieahia.org/pdfs/res_uquebec. pdf>, 2004. [3] G. Hoffheinz, N. Kelly, A. Ete, Evaluation of hydrogen demonstration systems & United Kingdom hydrogen infrastructure (Years 2–3 of Task 18 of The IEA Hydrogen Implementing Agreement) Contract number: F/04/00287/00/Rep, URN number: 07/770, Contractor: Sgurr Energy Ltd., 2007. [4] J. F. Martins, A. Joyce, C. M. Rangel, J. Sotomayor, R. Castro, A. Pires, J. Carvalheiro, R. A. Silva, S. Viana, International Conference on Power Engineering, Energy and Electrical Drives, Setúbal, Portugal, 12-14 April, 2007. [5] H. Miland, O. Ulleberg, Solar Energy (2008), doi:10.1016/j.solener.2008.04.013. [6] A. Bergen, L. Pitt, A. Rowe, P. Wild, N. Djilali, Journal of Power Sources 186 (2009) 158-166.

Solar-hydrogen stand-alone systems [13] C. Darras, S. Sailler, C. Thibault, M. Muselli, P. Poggi, J. C. Hoguet, S. Melscoet, E. Pinton, S. Grehant, F. Gailly, C. Turpin, S. Astier, G. Fontès, International Journal of Hydrogen Energy 35 (2010) 3322-3332. [14] S. Jalilzadeh, H. Kord, A. Rohani, ECTI Transactions on Electrical Eng., Electronics, and Communications 8 (1) (2010) 118-125. [15] P. Poggi, C. Cristofari, J. I. Canaletti, C. Darras, M. Muselli, Global Journal on Technology and Optimization (1) (2010). Available online: . [16] S. S. Deshmukh, R. F. Boehm, Renewable and Sustainable Energy Reviews 12 (2007) 2301-2330. [17] O. C. Onar, M. Uzunoglu, M. S. Alam, Renewable Energy 34 (2009) 509-520. [18] M. J. Khan, M. T. Iqbal, Applied Energy 86 (2009) 2429-2442. [19] S. Pedrazzi, G. Sini, P. Tartarini, Energy Conversion and Management 51 (2010) 122-129. [20] Sistema Nacional de Informação de Recursos Hídricos, Available from: . [21] European Electrical Specific Profiles – Annex 42, Available online: . [22] V. Lo Brano, A. Orioli, G. Ciulla, A. Di Gangi, Solar Energy Materials and Solar Cells 94 (8) (2010) 1358-1370. [23] Kyocera KC175GHT-2 Datasheet. Available online: .

[7] O. Ulleberg, Solar Energy 76 (2004) 323-329.

[24] O. Ulleberg, International Journal of Hydrogen Energy 28 (2003) 21-33.

[8] A. Bilodeau, K. Agbossou, Journal of Power Sources 162 (2006) 757-764.

[25] J. Larminie, A. Dicks, “Fuel cell systems explained”, Second ed., John Wiley & Sons, Chichester, 2003, UK.

[9] D. Ipsakis, S. Voutetakis, P. Seferlis, F. Stergiopoulos, S. Papadopoulou, C. Elmasides, Energy 33 (2008) 1537-1550.

[26] I. Sadli, P. Thounthong, J.-P. Martin, S. Raël, B. Davat, Journal of Power Sources 156 (1) (2006) 119-125.

[10] D. Ipsakis, S. Voutetakis, P. Seferlis, F. Stergiopoulos, C. Elmasides, International Journal of Hydrogen Energy 34 (2009) 7081–7095. [11] C. Wang, M. H. Nehrir, IEEE Transactions on Energy Conversion 23 (3) (2008) 957-967. [12] K. Zhou, J. A. Ferreira, S. W.. H. de Haan, International Journal of Hydrogen Energy 33 (2008) 477-489.

38

[27] B. S. Borowy, Z. M. Salameh, IEEE Transactions on Energy Conversion 12 (1) (1997) 73-78. [28] Autosil EE 2-800 Datasheet. Available online: . [29] A. Bergen, L. Pitt, A. Rowe, P. Wild, N. Djilali, International Journal of Hydrogen Energy 34 (2009) 64-70.
