Fuel Cell, Battery and Supercapacitor Hybrid System for Electric Vehicle: Modeling and Control via Energetic Macroscopic Representation

Transcription

1 Fel Cell, Battery and Spercapacitor Hybrid System for Electric Vehicle: Modeling and Control via Energetic Macroscopic Representation Lcia Gachia, Alain Boscayrol, Javier Sanz, Rochdi Trigi, Philippe Barrade To cite this version: Lcia Gachia, Alain Boscayrol, Javier Sanz, Rochdi Trigi, Philippe Barrade. Fel Cell, Battery and Spercapacitor Hybrid System for Electric Vehicle: Modeling and Control via Energetic Macroscopic Representation. Vehicle Power and Proplsion Conference, Sep 11, CHICAGO, United States. INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS - IEEE, 6 p., 11, <1.119/VPPC >. <hal > HAL Id: hal Sbmitted on 21 Oct 15 HAL is a mlti-disciplinary open access archive for the deposit and dissemination of scientific research docments, whether they are pblished or not. The docments may come from teaching and research instittions in France or abroad, or from pblic or private research centers. L archive overte plridisciplinaire HAL, est destinée a dépôt et à la diffsion de docments scientifiqes de nivea recherche, pbliés o non, émanant des établissements d enseignement et de recherche français o étrangers, des laboratoires pblics o privés.

2 Fel Cell, Battery and Spercapacitor Hybrid System for Electric Vehicle: Modeling and Control via Energetic Macroscopic Representation L. Gachia 1,5, A. Boscayrol 2,5, J. Sanz 1, R. Trigi 3,5, P. Barrade 4 1 Carlos III University. C/Btarqe 15, Leganés, Madrid (Spain) lgachia@ing.c3m.es, jsanz@ing.c3m.es 2 L2EP, Université de Lille Villeneve d Ascq cedex (France) Alain.Boscayrol@niv-lille1.fr 3 LTE, Institte Française des Sciences et Technologies des Transport, de l Aménagement et des Réseax. 25 Av. François Mitterand. F BRON cedex, France Rochdi.trigi@inrets.fr 4 LEI, École Polytechniqe de Lasanne. 115 Lasanne (Switzerland) Philippe.barrade@epfl.ch 5 MEGEVH : Modélisation Energétiqe et Gestion d Énergie des Véhicles Hybrides. (French network for HEV research) Abstract- Nowadays, no electrochemical energy system presents a competitive operation if compared with internal combstion engine, which is the reason for combining electrochemical energy systems to obtain a hybrid energy storage system. This paper presents a fel cell-battery-ltracapacitor system for vehicle application. The objective of this paper is to present the Energetic Macroscopic Representation (EMR) of the mltisorce power system. I. INTRODUCTION Crrent fossil fel vehicles present a high energy and power density operation which still cannot be matched by any electrochemical energy system on its own. However, redcing oil reserves and environmental concerns de to polltant emissions have activated the research for electrochemical energy systems. One of the crrent research lines is the hybridization of energy storage systems (ESS), in order to obtain competitive vehicle operation. Electric vehicles (EVs) allow a wide range of topologies, depending on the energy systems involved [1]. The high energy density needed to compete with ICE can be obtained from high energy density batteries or fel cells. Fel cells, especially PEM (Proton Exchange Membrane) fel cells present a particlarly high energy density if compared to the rest of electrochemical energy systems. This is de to the fact that the fel is stored in a separate tank, jst the opposite to batteries. However, they present a poor ramp p capability, de to mass transport phenomena [2] related to the protons concentration on the cathode side. Therefore, PEM fel cells can be complemented with high power energy storage systems which are capable of both generating and absorbing high power peaks. The ESS can be spercapacitors or batteries or non electrochemical systems, sch as flywheels. These electrochemical systems admit a varied combination between the fel cell, battery and spercapacitor. Depending on the topology, the power management will change. For example, for the FC-B case, the battery will be nder severe power brsts [3] which cold shorten the battery life time. The FC-SC combination cold improve this, as spercapacitors present higher power, ramp capability and life cycle than batteries [4], bt spercapacitors present mch lower energy density than batteries [5]. Moreover, it cold present problems dring vehicle start-p [6]. The B-SC combination cold solve this last aspect, bt it will redce the total energy density de to the fact that batteries still have a lower energy density than fel cells [7,8]. Therefore, the FC- B-SC combination keeps the high energy density de to the fel cell, the high power density and ramp capability de to spercapacitors and batteries, as stdied by Thonthong [9] and Schaltz [1]. The objective of this work is to present a fel cell-batteryspercapacitor electric vehicle, from modeling to control. This is done sing Energetic Macroscopic Representation, which allows representation of mltiphysical system and its strategy and control. II. EMR OF THE ELECTRIC VEHICLE A. Energetic Macroscopic Representation The FC-B-SC hybrid electric vehicle will be modelled in Energetic Macroscopic Representation (EMR). This is a graphical tool developed in [11], very helpfl when stdying complex systems and its energy management. The elements are connected considering the action-reaction principle and respecting physical casality. Physical casality is respected by considering only integral operations, discarding derivative operations. In physical systems the otpt is always delayed with respect to the inpt. In EMR each element is represented by a pictogram which internally incldes the mathematical eqations which describes the

3 relationships between action and reaction inpts and otpts. Moreover, EMR does not only consider the representation of each element, bt also complex control schemes and strategies. EMR has already proved to be sefl in, e.g. vehicle [12] and fel cell [13] applications, among others. The vehicle simlated is a 32 kw fel cell-batteryspercapacitor (Fig. 1). The power converter arrangement can be varied de to the mltiple power energy systems, and have been stdied in [5]. The complete system has three converters: one for the electric machine and two for the power sorces. The nmber of power converters for the power sorces can vary from none to three [5]. No converters imply a necessary oversizing of spercapacitors to allow its charge and discharge, as well as negative long term effects on fel cell and battery de to excessive cycling. On the other hand, three converters allow a greater control [9], bt it also increases weight, volme, costs and losses. An intermediate soltion is the se of two converters for three power sorces, leaving the third as a slack element. In this work we have chosen to interface fel cell and spercapacitors with power converters, leaving the battery directly connected to the electric machine converter. This choice allows the voltage variation needed for the fel cell and for, especially, the spercapacitors operation. B. Fel Cell System De to particlarities in vehicle applications, sch as fast start-p de to low temperatre operation, PEM fel cell (Proton Exchange Membrane) are commonly sed in road transport applications. Bat Fel cell UC / / / Fig. 1. Fel cell-battery-spercapacitor hybrid electric vehicle Vehicle mass TABLE I VEHICLE CHARACTERISTICS 1 kg Vehicle frontal srface 2 m 2 Wheel diameter.52 m Drag coefficient.35 Air density kg/m 3 Electric machine rated power Electric machine armatre voltage 32 kw 4 V v ve F res Fel cells inclde a stack, responsible for the power generation, and a nmber of ancillary systems, sch as compressor, cooling system, hmidification, etc. Therefore, it is a mltiphysical and complex system. There is a wide range of modeling approaches to fel cells. It can vary from stack centered models [14], to others which inclde axiliary systems, sch as [15]. or control oriented [16]. Eqivalent models obtained by impedance spectroscopy enable easy interaction with other electric models. The topology of the circit can present slight variations, depending on the athor [17]. The PEM fel cell model sed in this work is an eqivalent circit (see Fig. 2) obtained by impedance spectroscopy and experimentally validated [18]. The tests were carried ot on a Nexa Ballard fel cell, which is a 1.2 kw, 22-5 V system with a maximm crrent of 5 A. The model obtained has been scaled to represent a 3 cell stack with a voltage range of V. The PEM fel cell system EMR pictogram is a voltage sorce, represented by an oval block, as seen in Fig. 2. The otpt is the voltage compted by the fel cell eqations. d (1) U ( V ) = E RohmI I C1 U C1 R1 The voltage E depends on the stack temperatre, the Faraday constant and hydrogen, oxygen and water partial pressres and the fel cell standard voltage. The resistances and capacitor voltages represent the ohmic and doble layer capacitance phenomena which take places in the stack. The fel cell is connected to the rest of the system via an indctance, to smooth the crrent ripple, and a dc-dc converter. Eqations for both systems are presented. d (2) ind _ = L i + Ri ichop _ = m2 i (3) ind _ = m2 bat C. Batteries Battery modeling has been deeply stdied de to the complex phenomena which take place, both in the short and long term. A widely accepted test procedre to model batteries is the electrochemical impedance spectroscopy [19], []. Ni-Mh, have already been sccessflly been sed in Toyota Pris, and is the technology chosen for this work. The battery model sed is based on the model available in the Matlab/Simlink library. It is a 3 V 1 Ah Ni-Mh battery. The battery EMR pictogram is an oval element, with voltage as otpt and crrent as inpt, as depicted in Fig. 3. i FC + FC - FC Fig. 2. Fel cell EMR (left) and eqivalent circit (right) FC i FC

4 Batt I batt For the correct comptation of the battery model, it is necessary to estimate the battery state-of-charge (SoC). The estimation of the SoC can be a complex task. Especially in nonflooded batteries, as the acid concentration of the electrolyte, which is normally a good indication of the battery SoC, cannot be directly measred. The different methods can be conslted in [21]. The method chosen for this work is the ampere hor cont, taking into accont the ampere hor efficiency. 1 SoCt (%) = SoC (%) η t 1 i( t) (4) C D. Spercapacitors Spercapacitors enable high power delivery or braking, which is an attractive characteristic for transport applications. Spercapacitor modeling can yield a variety of eqivalent circits. Different athors present eqivalent circits with different topology, as [22] and voltage eqalization [23]. The spercapacitor model sed in this work is based on an previos experimentally validated work [24] for a 3 F 2.5 V Maxwell Boostcap element. The model has been scaled to a 1 series connected cell pack. The EMR pictogram for a spercapacitor is also a voltage sorce, represented again by an oval element (Fig. 4). 1 1 C2 c = L + ic + ic C1 C2 R2 (5) 1 C i c C3 R3 Spercapacitors are connected to the dc bs throgh an indctance and a boost dc-dc converter. d (6) sc ind _ sc = L isc + Risc i ind _ sc E = m3 i = m 3 E. Mechanical part and environment The mechanical part of a hybrid electric vehicle can present different topologies [12]. + n chop _ sc sc (7) bat Rint Fig. 3. Battery EMR (left) and eqivalent circit (right) SC U c I c Fig. 4. Spercapacitor EMR (left) and eqivalent circit (right) + _ Ubatt BAT FC SC U i Energy generation U sc i batt i U ind_ U ind_sc i chop_ i chop_sc m 2 m 3 Parallel copling i load Fig. 5. EMR for the fel cell-battery-spercapacitor vehicle ES The machine sed for this simlation is a permanentmagnet machine. This implies that no external excitation system needs to be controlled. The electric machine is not directly copled to the wheels, bt needs to be interfaced by a differential in order to redce the shaft speed, increase the torqe and distribte it to both driven wheels. The gear mltiplies the torqe to be transmitted to the differential by a factor of 5. The differential will transmit the torqe to left and right wheels, allowing a different speed for each if necessary. All the mechanical part has been inclded in an electric sorce at the dc bs, as seen in Fig. 5. F. Parallel copling At the dc bs the voltage will be imposed by the battery, whilst the crrents will be distribted in a parallel node. To facilitate later control schemes, the parallel connection is separated in two. iload = ibat + i1 (8) i1 = ichop _ + ichop _ sc III. Once the complete system is represented in EMR, it is necessary to be able to control it. The ability to do so is one of the interesting characteristics of EMR. The control is done by firstly defining the control loops in the system. For this prpose it is important to define which are the system tning and objective variables. In this case there are two control loops. One is related to the mechanical part of the system and the other one to the power generation system. The control loop for the mechanical part has the speed as the objective variable and the power converter of the electric machine as the tning variable. This loop is inclded in the electric sorce which simlates the mechanical part. i 1 Mechanical part INVERSION BASED CONTROL THROUGH EMR

5 Fig. 6. EMR of the fel cell-battery-spercapacitor system with inversion based control The power generation control loop has as objective the control of the crrent distribtion in the dc bs. The crrent distribtion is related to the first Kirchhoff law (8), where the crrent on the dc bs (i total ) is imposed by the dc machine. Therefore, only two crrents need to be controlled to obtain the three crrents. The choice of which of the two crrents mst be controlled will depend on the strategy, as seen in Fig. 6. This work will present the example of controlling the fel cell crrent and battery crrent. Fel cell crrent control eqations: m BAT FC SC ind _ 2 _ ref Energy generation i total = i bat + i chop _ + i chop _ sc (9) U i mes U sc U mes i batt isc ref i ref _ ref = ind _ = i U sc mes _ mes _ ref bat _ mes U ind_ Uind_sc C i chop_ i chop_sc m 2 U ind_sc ref m i 2 mes U ind_ ref m 3 PI i 1 i chop_sc ref Parallel copling Ubatt mes i load [ i i ] _ ref i 1 ref Strategy Mechanical part i bat ref MS I chop_ mes _ mes i total mes (9) Battery crrent control loop related to the inversion based control eqations are presented. i1 _ ref = iload _ mes ibat _ ref i chop _ sc _ ref = i1 _ ref ichop mes i ( ) chop _ sc _ ref t i _ = (1) sc ref m3 _ ref ( ) = ( ) ( )[ ( ) ( )] ind _ c _ ref t sc _ mes t C PI t isc _ ref t isc _ mes t ( ) ind _ sc _ ref t m3 _ ref = bat _ mes The advantage is that fel cell and battery are the elements that sffer more severe aging processes, so it is important to avoid excessive cycling. Known the tning and objective variables, it is possible to define the control system sing the inversion principle. This principle is based on the idea of inverting each block ntil the objective variable is obtained. This objective shold follow the reference generated by the strategy. IV. SIMULATION RESULTS The simlation reslts present the hybridization of a fel cell-battery-spercapacitor system nder a NE cycle (New Eropean Driving Cycle), as shown in Fig. 7. A first strategy based on freqency selection is presented. To avoid oxygen starvation on fel cell systems, it is sal to avoid sbjecting the fel cell system to power brsts. This can be achieved by assigning as reference load crrents with a freqency lower than 1 Hz. On the other hand, spercapacitors are specially designed to endre sdden power brsts, both dring spply or regenerative braking. Therefore, high freqency loads, above 1 Hz will be spplied or absorbed by the spercapacitor. This implies that intermediate freqencies will be the reference for battery crrent. Reslts are presented in Fig. 8 and V ev (km/h) - Fig. 7. NE driving cycle (pper) and load crrent (bottom)

6 1 1 I load (A) 81 8 SoC (%) I bat (A) U sc (s) I chop (A) I chop sc (A) -1 Fig. 8. Crrent distribtion on the parallel copling The strategy mst inclde limitations in each of the electrochemical energy sorces. Limitations for maximm fel cell crrent (5 A) are inclded. This limitation is de to the stack cross section. Any limitations for stack environment to garantee correct operation is already carried ot by the fel cell system control. This control will make sre temperatre, gas pressre and water content are the expected vale for each operation point. The battery state-of-charge (SoC), is limited to 6-8%, to avoid deep discharges. Maximm voltage for spercapacitors mst also be limited to 2.5 V per element. Spercapacitor voltage range is taken from to 2.5 V per element, which is a wider range than the normally sed V range. This will allow a deeper discharge of the spercapacitors, which will enable a higher power recharge dring regenerative braking. V. CONCLUSIONS In this paper we have obtained the energetic macroscopic representation of a fel cell-battery-spercapacitor hybrid electric vehicle. The electrochemical energy systems models sed were previosly experimentally validated and adapted to the case power levels. An initial energy storage strategy was simlated, demonstrating EMR is a powerfl tool for this isse. The next step will be to stdy how other energy management strategies affect each of the electrochemical energy systems involved. ACKNOWLEDGMENT Fig. 9. Battery state-of-charge (pper) and spercapacitor voltage (bottom) Athors acknowledge financial spport from the LEESS research grop of the University Carlos III of Madrid (UC3M), from the scholarships for mobility of researchers of UC3M 11 and the L2EP of the University of Lille.

7 REFERENCES [1] C. C. Chan, The state-of-the-art of electric, hybrid, and fel cell vehicles, Proceedings of the IEEE., vol. 95, no. 4, pp , April 7. [2] M. Ceraolo, C. Milli, and A. Pozio, Modelling static and dynamic behavior of PEM fel cells on the basis of electrochemical description, Jornal of Power Sorces, vol. 113, pp , Janary 3. [3] M.J. Kim, and H. Peng, Power management and design optimization of fel cell/battery hybrid vehicles, Jornal of Power Sorces, vol. 165, pp , 7. [4] A. Borke, Ultracapacitors: why, how and where is the technology, Jornal of Power Sorces, vol. 91, pp. 37-5, November. [5] J. Baman, and M. Kazerani, A comparative stdy of fel cell-battery, fel cell-ltracapacitor and fel cell-battery-ltracapacitor vehicles, IEEE Trans. Vehiclar Techn, vol 57, no.2, pp , March 8. [6] W. Gao, Performance comparison of a fel cell-battery hybrid powertrain and a fel cell-ltracapacitor hybrid powertrain, IEEE Trans. On Veh. Techn., vol. 54, no. 3, pp , May 5. [7] R. M. Dell, and D.A.J. Rand, Energy storage technology for global energy sstainability, Jornal of Power Sorces, vol. 1, pp. 2-17, November 1. [8] J.M. Miller, Energy storage system technology challenges facing strong hybrid, plg-in and battery electric vehicles, IEEE Vehicle Power Proplsion Conference, September 7-1 9, Dearborne U.S.A. [9] P. Thonthong, S. Raël, and B. Davat, Energy management of fel cell/battery/spercapacitor hybrid power sorce for vehicle applications, Jornal of Power Sorces, 9. [1] E. Shaltz, A. Khaligh, and P.O. Rasmssen, Investigation of battery/ltracapacitor energy storage rating for a fel cell hybrid electric vehicle, IEEE Vehicle Power and Proplsion Conference, September 3-5 8, Harbin, China. [11] A. Boscayrol, B. Davat, B. de Fornel, B. François, J.P. Hatier, F. Meibody-Tabar and M. Pietrzak-David, Mltimachine mlticonverter system: application for electromechanical drives, Er. Phys. J, Appl. Phys., vol. 1, no. 2, pp , May. [12] K. Chen, A. Boscayrol, A. Berthon, P. Delare, D. Hissel, and R. Trigi, Global modeling of different vehicles, IEEE Vehiclar Technology Magazine, vol. 4, no. 2, pp. 8-89, Jne 9. [13] L. Bolon, D. Hissel, A. Boscayrol, and M.-C. Pera, From modelling to control of a PEM fel cell sing energetic macroscopic representation, IEEE. Trans. Ind. Electron., vol 57, no.6, pp , Jne 1. [14] P. Pathapati, X. Xe, and J. Tang, A new dynamical model for predicting transient phenomena in a PEM fel cell, Renewable Energy, vol. 3, pp. 1-22, 5. [15] L. Bolon, D. Hissel, A. Boscayrol, and M.C. Péra, From modelling to control of a PEM fel cell sing energetic macroscopic representation, IEEE Trans. Ind. Electron., vol. 57, no. 6, pp , Jne 1. [16] C. Bordons, A. Arce, and A.J. del Real, Constrained predictive control strategies for PEM fel cells, Proceedings American Control Conference, Minneapolis, USA, Jne 14-16, 6. [17] X. Yan, J. Sn, H. Wang, and J. Zhang, AC impedance diagnosis of a 5 W PEM fel cell stack. Part I: stack impedance, Jornal of Power Sorces, vol. 161, pp , 6. [18] L Gachia, and J. Sanz, A per-nit hardware-in-the-loop simlation of a fel cell/battery hybrid energy system, IEEE Trans. Ind. Electron.., vol. 57, no.4, pp , April 1. [19] E. Karden, S. Bller, and R.D. Doncker, A method for measrement and interpretation of impedance spectra for indstrial batteries, Jornal of Power Sorces, vol. 85, pp ,. [] M. Thele, O. Bohlen, D. Saer, and E. Karden, Development of a voltage behavior model for NiMh batteries sing an impedance based modeling concept, Jornal of Power Sorces, vol. 175, pp , 8. [21] S. Piller, M. Perrin and A. Jossen, Methods for state-of-charge determination and their applications, Jornal of Power Sorces, vol. 95, pp , Jne 1. [22] S. Bller, E. Karden, D. Kok, and R.D. Doncker, Modeling the dynamic behavior of spercapacitors sing impedance spectroscopy, 36th Indstry Application Conference, vol. 4, pp , Chicago, USA, September 1. [23] P. Barrade, Series Connection of Spercapacitors: Comparative Stdy of Soltions for the Active eqalization of the Voltages, 7th International Conference on Modeling and Simlation of Electric Machines, Converters and Systems, Agst, Ecole de Technologie Spériere, Montreal, Canada. [24] L. Gachia, S. Castaño, and J. Sanz, New approach to spercapacitor testing and dynamic modelling, IEEE Vehicle Power and Proplsion Conference, September 1-3 1, Lille, France.