Fuel-Cell Hybrid Powertrain: Toward Minimization of Hydrogen Consumption

Please download to get full document.

View again

of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Fuel-Cell Hybrid Powertrain: Toward Minimization of Hydrogen Consumption
Document Share
Document Tags
Document Transcript
  3168 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 58, NO. 7, SEPTEMBER 2009 Fuel-Cell Hybrid Powertrain: Toward Minimizationof Hydrogen Consumption Jérôme Bernard, Sebastien Delprat, Felix N. Büchi, and Thierry Marie Guerra  Abstract —In this paper, the powertrain sizing of a fuel-cellhybrid vehicle (FCHV) is investigated. The goal is to determine thefuel-cell system (FCS) size, together with the energy storage sys-tem (ESS) size, which leads to the lowest hydrogen consumption.The power source (FCS + ESS) capabilities should also respectthe vehicle driveability constraints. Batteries and supercapacitorsare considered as ESSs. The power management strategy is aglobal optimization algorithm respecting charge sustaining of theESS. The impacts of the driving cycle (urban, outer urban, andhighway), ESS technology, and vehicle driveability constraints onhydrogen consumption are analyzed in detail.  Index Terms —Battery, fuel cell, global optimization, hybridelectric vehicle, sizing, supercapacitor. I. I NTRODUCTION P OLYMER electrolyte fuel cells (PEFCs) have been identi-fied as the most appropriate technology for transportationpurposes [1], [2]. Because a fuel-cell system (FCS) is notpower reversible, an energy storage system (ESS) is added tothe powertrain. The two main electrochemical devices used asESS are either batteries (i.e., Toyota fuel-cell hybrid vehicle(FCHV)[3])orsupercapacitors(i.e.,MichelinHy-Light[4]andHonda FCX [5]). The goals of a hybrid powertrain are to reducehydrogen consumption and to satisfy driveability constraints.These goals are achieved with the second energy source (i.e.,ESS) through braking energy recovery, power assistance of the FCS, and control opportunity in the power-managementstrategy.In this paper, the influence of FCS and ESS sizing on hy-drogen consumption is investigated. These investigations takeinto account the vehicle driveability requirements. The first partdeals with the state of the art of fuel-cell hybrid powertrainsizing. In the second part, the vehicle, the FCS, and the ESS Manuscript received January 10, 2007; revised July 18, 2007,October 5, 2007, February 1, 2008, July 23, 2008, September 17, 2008, andJanuary 5, 2009. First published February 6, 2009; current version publishedAugust 14, 2009. This work was supported in part by the International Campuson Safety and Intermodality in Transportation, by the European Community,by the Délégation Régionale à la Recherche et à la Technologie, by the FrenchMinistère de l’Enseignement supérieur et de la Recherche, by the Région NordPas de Calais, by the Centre National de la Recherche Scientifique, and inpartnership with the Paul Scherrer Institute (PSI, Switzerland). The review of this paper was coordinated by Dr. M. S. Ahmed.J. Bernard and F. N. Büchi are with the Paul Scherrer Institut, 5232 Villigen,Switzerland (e-mail: Jerome.Bernard@psi.ch; felix.buechi@psi.ch).S. Delprat and T. M. Guerra are with the Laboratoire d’Automatique,de Mécanique et d’Informatique industrielles et Humaines, Université deValenciennes et du Hainaut-Cambrésis/Le Mont-Houy, 59313 ValenciennesCedex 09, France (e-mail: sebastien.delprat@univ-valenciennes.fr; guerra@univ-valenciennes.fr).Digital Object Identifier 10.1109/TVT.2009.2014684 models are specified. In the third part, the vehicle driveabilityrequirements are viewed as bounds or constraints for powersource sizing. Then, the method of analysis is presented.The core of this method is the power-management strategy,which relies on a global optimization algorithm. The globaloptimization allows fair comparisons, in terms of hydrogenconsumption, between different hybrid powertrains. Finally, thehydrogen consumption for different powertrain sizes is com-putedusinganoptimalcontrolalgorithm.Severaldrivingcycles(urban, outer urban, and highway) and two ESS technologies(supercapacitors and Ni-MH battery) are considered.II. S TATE OF THE  A RT The fuel consumption of a fuel-cell vehicle (FCV) can beimproved by using an additional energy source. Using ESS,most of the FCS operating points can be shifted to its high-efficiency power range [6]–[8]. The second role of the ESS is torecover the vehicle kinetic energy while braking. The recoveredenergy can considerably influence the hydrogen economy of theFCHV [9]. However, when hybridizing an FCV, some negativeaspects should carefully be considered. The most importantof them is the mass increase of the vehicle, which raises thepower demand and thus the hydrogen consumption [10]. Thus,the benefit of adding an ESS to the powertrain can become adrawback due to this effect [9]. The complexity and cost of such systems may also put ahead against fuel-cell powertrainhybridization [6], [11].Regarding driveability, the FCS minimal size is constrainedby the maximal sustaining power requested (during high cruisespeed) [12], [13], and the ESS is sized to power assist theFCS during brief high-power demands (basically, this occurredduring hard vehicle accelerations) [10]. Pure FCV takes the realbenefits of adding an ESS during fast transient power demandsbecause it allows relaxing the FCS dynamics constraints [8],[14]. FCS suffers from a slow transient because of the lowresponse time of the air circuit [8], which leads to the oxygen-starvation phenomena [9], [15]. FCS has some other drawbacksthat may affect vehicle driveability; for example, the cold startperiod with limited FCS power [10], [12]. A well-sized ESScan provide this lack of power and overcome this problem[10]. Finally, FCS cooling is challenging and requires particularattention for automotive applications [16].Regarding the advantages and weaknesses of a hybridizedFCS power source, the sizing procedure for FCHV has beeninvestigated during the last decade by many researchers [34].Some approaches focus on the ESS sizes with constant FCSrated power [8], [9], [15]. Considering the variation of the FCSrated power requires careful analysis of the system efficiency, 0018-9545/$26.00 © 2009 IEEE  BERNARD  et al. : FUEL-CELL HYBRID POWERTRAIN: TOWARD MINIMIZATION OF HYDROGEN CONSUMPTION 3169 TABLE IV EHICLE  C HARACTERISTICS Fig. 1. Powertrain arrangement. which is defined by the combination of the fuel-cell stack andthe auxiliary components. One may choose to consider theair compressor as a unique auxiliary (it is the main source of parasitic power) [17] or only consider a typical FCS efficiencycurve [18].Different ESS technologies can be considered and compared.Battery versus supercapacitor is the most common balancein this research field [18]–[21]. The significant differences inspecific energy and specific power explain this interest.III. V EHICLE AND  P OWERTRAIN  A SSUMPTIONS  A. Vehicle The vehicle considered in this paper is a hybrid vehicle pro-totype based on a Citroën Berlingo 1 [22]. For the simulations,the IC engine has been replaced by an FCS, the electric motor isa 75-kW ac [23], and supercapacitors or Ni-MH battery packsare used in hybrid configurations. The main vehicle parametersare given in Table I.The powertrain architecture integrates three converters (seeFig. 1): Two dc/dc converters are used to connect the FCS andthe ESS to the bus, and a dc/ac inverter is used to power theelectric motor. The topologies of the converters are supposed tomatch the voltages, currents, and power requirement in everycase. The FCS is considered as the primary power source, andthe ESS is considered as an energy buffer. The electric node(see Fig. 1) leads to power conservation.  B. Energy Storage System The ESS used in FCHV is either a battery or a su-percapacitor module, which have significant differences in 1 This Hybrid Berlingo prototype has been developed by both LAMIH UMRCNRS 8530 and PSA Peugeot-Citroën with financial support from FEDER,ADEME, and the region Nord Pas-de-Calais.TABLE IIS UPERCAPACITOR AND  B ATTERY  M ODELS their electric characteristics (see Table II). In this paper, su-percapacitor modules based on 2600-F capacitor cells andNi-MH batteries based on 28-Ah cells are considered. Theenergy density of Ni-MH is ten times higher than the en-ergy density of the supercapacitor, and the power density of the supercapacitor is three times higher than the power den-sity of the battery. The ESS models are based on a resis-tance model [24], and the parameters are given in Table II.The ESS parameters come from [25] for the supercapacitor andfrom [23] for the Ni-MH battery. The Ni-MH battery has beenchosen because it is currently the most established technologyin hybrid vehicles [26]. C. Fuel Cell System An FCS is composed of a fuel-cell stack, which convertshydrogen and air into electric energy, and a number of auxiliarycomponents. In an FCS with conventional layout, the auxiliarycomponents keep the stack operational by supplying air andhydrogen, by humidifying the reactant gases, and by coolingthe stack; they need power to operate and consume part of theenergy produced by the stack.The stack is built up from single cells electrically connectedin series. Under the assumption that all the cells are identical,the stack voltage  V  stack  is equal to the number of cell  N  cell multiplied by the cell voltage  V  cell , i.e., V  stack ( i ) =  N  cell  · V  cell ( i ) .  (1)Under the same assumption, the gross stack power  P  stack  is P  stack ( i ) =  V  cell ( i ) · i · N  cell  · A cell  (2)where  i  the current density, and  A cell  is the cell active area.To obtain the cell voltage  V  cell , a steady-state model from[27] is used, and the operating conditions retained are summa-rized in Table III.The power consumption of the auxiliary components  P  aux  isrequiredtoderivethenetpower P  FCS  availablefromtheFCSas P  FCS ( i ) =  P  stack ( i ) − P  aux ( i ) .  (3)The parasitic power  P  aux  in (5) has been estimated accordingto the air compressor power consumption, which is the main  3170 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 58, NO. 7, SEPTEMBER 2009 TABLE IIIFCS P ARAMETERS Fig. 2. Stack and system efficiency curves. source of losses [2]. According to the operating conditionsretained, the maximum net power density is 477 mW/cm 2 , andthe system peak efficiency is 50% at about 25% rated power(see Fig. 2).A fuel-cell power-rate limit is also considered. This ratelimitation mainly comes from the slow dynamic of the aircircuit [8]. Data from the literature indicate that current FCStechnologies require up to 2 s to increase their power from 10%to 90% rated power  P  maxFCS ( T  10% → 90%  = 2  s ) . The resultingmaximal power rate  ( dP  FCS /dt ) = ˙ P  FCS ( W/s )  is ˙ P  FCS  = 80% · P  maxFCS T  10% → 90% .  (4)For automotive applications, this FCS dynamic limitation mayinfluence the car’s driveability.IV. V EHICLE AND  P OWERTRAIN  A SSUMPTIONS The hybrid fuel-cell powertrain has to satisfy the prescribeddynamicperformances.Eachdynamicperformanceconstraint  j results in a domain of admissible solutions  D j . The inter-section of each domain  D j  limits the FCS and ESS sizes indomain  D  as D  =  ∩ j D j .  (5)Based on the proposal made in [7], [10], and [18], we considertwo driveability constraints  (  j  ∈ { 1 , 2 } ) , i.e., gradeability andacceleration capabilities. The hybrid Berlingo presented inSection III-A serves as an example.  A. Constraint 1: Gradeability The gradeability constraint corresponds to the capability of sustaining a constant speed on a slope. Considering the srcinalspecifications of the hybrid Berlingo [22], a sustained speedof 110 km/h on a 5% slope has to be guaranteed. Usingthe vehicle model, the corresponding constant motor powerdemand  P  sustainmot  is calculated. The problem is mass depen-dent: the heavier the vehicle, the higher the required power P  sustainmot  .With respect to our powertrain assumptions, the ESS is usedas an energy buffer (charge sustaining); thus, only the FCS isable to provide a constant power during extended time (as longas the hydrogen tank is not empty). Therefore, the gradeabilityconstraint imposes a minimal FCS size, which defines the firstdomain of admissible solutions D 1  =  P  maxFCS ,N  ESS /P  maxFCS  ≥  P  sustainmot  .  (6)  B. Constraint 2: 0–100 km/h Full Power Acceleration The0-to100-km/hacceleration time T  0 → 100km / h  isaclassicvehicle performance benchmark. It is mainly related to themotor maximal power, which has to be defined in preliminaryvehicle design specifications [17]. The motor is supposed tooperate at full torque during the 0- to 100-km/h accelerationphase (i.e., throttle at the maximal position). In our case,the motor has been fixed to 75 kW (equivalent to 100 HP);therefore, the  T  0 → 100km / h  value is somehow definitive(around12s).Onlythevehiclemassvariation,whichisinducedby the power source sizing (FCS and ESS), slightly shifts the T  0 → 100km / h  value.The FCS and ESS should satisfy the power demand  P  accelmot during this full torque acceleration. The motor torque is max-imal during the  T  0 → 100km / h  period; thus, the power demand P  accelmot  ( V  veh )  is only a function of the vehicle speed  V  veh .Considering the maximal FCS power P  maxFCS  and the FCS power-rate limitation  ˙ P  FCS  defined in (4), the FCS might not be ableto fulfill this motor power demand alone. Unless considering ahugeFCSsize,whichisnotsuitableforautomotiveapplications(problem of stack cooling, limited available space, etc.), theESS has to provide the complementary power. Within theseassumptions, the domain  D 2  is D 2  =  ( P  maxFCS ,N  ESS ) / ∀ t  ∈  [0 ,T  0 → 100 km / h ] ×  P  maxFCS  + P  maxESS  ≥  P  accelmot  ( V  veh ( t ))˙ P  FCS ( t )  ≤  ˙ P  FCS SOC  ( t )  ≥  SOC  min  .  (7)For our vehicle, the domain  D 2  is computed assuming anFCS power-rate time of   T  10% → 90%  = 2  s [7] [see (4)], aninitial ESS state of charge of 60% for the battery and 80%  BERNARD  et al. : FUEL-CELL HYBRID POWERTRAIN: TOWARD MINIMIZATION OF HYDROGEN CONSUMPTION 3171 Fig. 3.  D domain for (left) the battery case and (right) the supercapacitor case. for the supercapacitor, and a state-of-charge lower limit of 25% ( SOC  min  = 25%) . C. Area of Feasible Solutions Theintersectionofthedomains D 1  and D 2  definesaspace D of admissible FCS and ESS sizes D  =  D 1  ∩ D 2 .  (8)Notice that the domains  D 1  and  D 2  do not correspond toan exhaustive list of constraints. For example, the resultingdomain  D  is not closed (see Fig. 3); thus, infinite FCS andESS sizes are allowed. However, regarding the hydrogen con-sumption (see Section VI) as well as the acceleration time T  0 → 100km / h , huge and heavy power source configurationsmake no sense.V. H YDROGEN  C ONSUMPTION  U NDER T YPICAL  D RIVING  C ONDITION  A. MethodologyPower-Management Strategy:  Minimizing the vehicle hy-drogen consumption is the main topic of this paper. As a hybridpowertrain is under consideration, a power management strat-egy is required to define what both the FCS and ESS powersare. The power management strategy and the powertrain designstrongly influence the fuel consumption [33]. Therefore, theinfluence of the control strategy must be “eliminated” to focuson the power source sizing. A global optimization algorithmhas this powerful advantage.The global optimization algorithm computes an optimalpower splitting between the FCS and the ESS, which leads tothe lowest hydrogen consumption. This optimal power splittingcan be achieved, whatever the ESS technology (batteries orsupercapacitors), the powertrain size (FCS size, ESS size), andthe driving cycle are. The algorithm relies on the classicaloptimal control theory and was already developed for tradi-tional IC engine hybrid control strategy optimization [28]. Theauthors have extended the work of [28] to the specific FCHVapplication. Details of this algorithm are provided in [28]–[31]. Fig. 4.  ∆ SOC   influence on consumption, NEDC driving cycle, 40-kW FCS,and 300 Ni-MH battery elements SOC  (0) = 60% . For the given driving cycle and powertrain arrangement, ithasbeen proven thattheminimalhydrogen consumption isonlydependent on the final ESS state-of-charge balance (depletionor recharge of the ESS) [29]. The influence of the SOC balance ∆ SOC   on the optimal hydrogen consumption is presentedin Fig. 4 assuming the following parameters: 40-kW FCS,300 Ni-MH battery elements with an initial SOC of 60%, andthe New European Driving Cycle (NEDC). In this paper, theESS is assumed to be charge sustaining; thus, the ESS state-of-charge balance constraint is  ∆ SOC   = 0% . In this particularcase, the ESS is used as an energy buffer, and the vehiclepropelling energy is globally provided by hydrogen.  Driving Cycles:  Four driving conditions have been consid-ered to estimate the hydrogen consumption.TheNEDC(seeFig.5)isusedasabenchmarktoestimatethevehicle fuel consumption in Europe. It consists of three urbancycles followed by an outer urban cycle.The UF3, R3, and A2 driving cycles (see Fig. 5) are derivedfrom real speed measurements: UF3 is an urban run, R3 is anouter urban run, and A2 is a highway run. These three drivingcycles are derived from real driving patterns collected throughEuropean roads [32].For each driving cycle (NEDC, UF3, R3, A2), the hydrogenconsumption is obtained for different pairs of FCS and ESS  3172 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 58, NO. 7, SEPTEMBER 2009 Fig. 5. Considered driving cycles.Fig. 6. Design of experiments for hydrogen consumption estimation. sizes. Of course, the global-optimization algorithm is used tocompute the minimum consumption (for each point) under theconstraints previously depicted. The design of experiments (seeFig. 6) varies from 0 to 100 kW with a step of 10 kW for theFCS power  (∆ PAC   = 10  kW ) , from 0 to 500 elements witha step of 50 elements for the battery pack   (∆ ESS   = 50) , andfrom 0 to 500 elements with a step of 50 elements for thesupercapacitor pack   (∆ ESS   = 50) . The hydrogen consumptionis computed under the following assumptions.1) The ESS is charge sustained  (∆ SOC   = 0%) .2) The FCS 10% to 90% rated power raise time is equal to2 s  ( T  10% → 90%  = 2  s ) .Fig. 7 shows the hydrogen consumption map for the NEDCdriving cycle with some iso-consumption levels. The “Impossi-ble sizing” area represents the undersized FCS and ESS, whichcannotfulfillthemotorpowerdemand.Thelimitsofthedomain D , which are related to the driveability constraints, are alsoindicated.The maps in Fig. 7 are posttreated to make them quicklyunderstandable. For each map, a pure FCV is considered to bea reference ( N  ESS  = 0  horizontal axis). Then, the FCS poweris set to the smallest hydrogen consumption reachable (black point in Fig. 8). It does not necessarily belong to the domain of admissible solutions  D  but represents a nonhybrid powertrainreference for evaluating the fuel saving due to hybridization.Then, an optimal sizing without considering the domain  D  isset (a star in Fig. 8), and the optimal sizing regarding domain D (i.e., the driveability constraints are respected) is plotted(across in Fig. 8).  B. Results The hydrogen economy maps regarding the proposedmethodology are computed according to the driving condi-tions: urban in Fig. 9, outer urban in Fig. 10, and highway inFig. 11. For each hydrogen economy map (see Figs. 8–11),the ideal power source sizes leading to the smallest hydrogenconsumption are extracted and summarized in Table IV. Thehybridization rate HR is the ratio between the ESS maximalpower  P  maxESS  and the total power available, i.e.,HR  =  P  maxESS P  maxFCS  + P  maxESS .  (9) C. Discussion 1) The first remark concerns the driving cycle’s influenceon hydrogen consumption. The best hydrogen economiesare achieved with the urban run (see Fig. 9), with amaximum of 19% considering the supercapacitor casewithout driveability constraints. On the other hand, thelowest hydrogen economies are achieved with the high-way run (see Fig. 11), with a maximum of 1% for bothbattery and supercapacitor cases without considering thedriveability constraints. The vehicle dynamics within thedriving cycle mainly explains this result: Many accelera-tions (with peak power demand) and decelerations (withbraking energy recovery) enforce the role of an ESS,whereas thereisobviously only asmallpossiblegain withcontinuous speeds (i.e., highway).2) The second remark concerns the computed maps (seeFigs. 8–11). Within an FCHV design process, the “impos-sible sizing” boundaries should carefully be consideredsince it reveals the critical limits of FCS and ESS sizes.Above these limits, the power source will not fulfillthe motor power demand. Concerning the “inefficientsizing” area, it highlights that FCV hybridization maylead to unacceptable increase of the hydrogen consump-tion (as pointed out in [11]). Regarding the boundaryof the domain  D , the driveability constraints restrictthe location of the chosen sizing. It clearly illustrates
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks