3D Thermal Analysis of Li-ion Battery Cells with Various Geometries and Cooling Conditions Using Abaqus

Transcription

1 3D Thermal Analyss of L-on Battery Cells wth Varous Geometres and Coolng Condtons Usng Abaqus Km Yeow, Ho Teng, Marna Thellez, Eugene Tan AVL Powertran Engneerng Abstract: Modelng thermal behavor of L-on cells for vehcle electrfcaton applcatons s a challengng task. AVL has developed 3D-FEA models to smulate the electro-thermal behavor of L-on battery cells wth varous geometres usng Abaqus. In these models, the L-on battery system s smplfed and modeled as heterogeneous sold medum consstng of a sngle or multple equvalent battery layers wth composte electrcal and thermal conductvtes for the equvalent anode, cathode and separator. Thermal behavors of cylndrcal, prsmatc and pouch L-on battery cells and modules were analyzed under dfferent electrcal loads and coolng condtons. Smulaton results were compared wth avalable battery temperature measurements (coverng cylndrcal-cell and pouch-cell modules) and good agreements were observed. Ths ndcates that the 3D electro-thermal model employed n ths study characterzes the electro-thermal behavor of the L-on battery cells reasonably well. Keywords: Lthum-on battery, cells, modules, electro-thermal modelng, coolng 1. Introducton Lthum-on (L-on) batteres are wdely selected as the energy storage devces for Hybrd Electrcal Vehcles (HEV), Plug-n Hybrd Electrcal Vehcles (PHEV), and Electrcal Vehcles (EV) due largely to ther hgh energy densty, hgh power densty, good stablty and low charge loss when not n use. In a battery pack, the cells are assembled n groups or modules to obtan the requred pack capacty, and modules are connected n seres to provde the requred pack voltage. For hgh-power battery packs such as those for HEV and PHEV applcatons, consderable amount of heat can be generated n the cells as a result of hgh dscharge/regen pulse currents durng duty cycles, causng rapd rse n cell temperature. For optmal performance of a battery pack, workng temperatures of the cells n the pack should be controlled to wthn a proper range (deally between 20 C to 40 C) and the temperature dstrbuton n the cells should be as unform as possble. The pack power capablty s affected sgnfcantly by temperatures of the cells wthn the pack: n low temperature operatons, the pack power capablty s lmted by the coolest cell; when operated at elevated temperature, the pack safety and thus the maxmum allowed pack power are determned by the hottest cell. The maxmum cell temperature and the maxmum dfferental cell temperature are crucal factors to the cell safety and durablty. Fgures 1 to 3 show the structures of cylndrcal, prsmatc and pouch L-on cells used n the battery packs for HEV, PHEV or EV applcatons. These three types of cells have advantages and dsadvantages. Cylndrcal cells (Fgure 1) are easer to manufacture and have good mechancal stablty and hgh energy densty. However they have a low packng effcency, resultng n a 2012 SIMULIA Communty Conference 1

2 relatvely low energy densty for the pack. Prsmatc cells (Fgure 2) typcally have jelly roll or stacked electrodes, and they are mechancally robust and have a hgh packng effcency. They have slghtly lower energy densty and are more expensve to manufacture compared to cylndrcal cells. Pouch cells (Fgure 3) have hgher energy densty than the other two desgns. They are relatvely nexpensve and provde desgn freedom on dmensons, whch often makes the pouch cells the frst choce n the cell selecton for hgh capacty PHEV or EV packs. However, ther dsadvantages are: (1) mechancally vulnerable and thus requrng cell cartrdges to hold them, (2) prone to swellng durng operatons especally when the cells age, and (3) has no mechansm for gas ventng (as opposed to cylndrcal and prsmatc cells). Gas ventng for the pouch cell nvolves swellng/breakng of the pouch and hence causng cell falure. Fgure 1. (a) Spral wound structure for a cylndrcal cell; (b) Sectonal vew of spral wound core of a cylndrcal cell and detals of battery layers n the cell [1]. Fgure 2. (a) Mult-folded-layer structure of a prsmatc cell (b) Sectonal vew of X- ray pctures showng battery layers n the cell [2] SIMULIA Customer Conference

3 Fgure 3. (a) Mult-stacked-layer structure of a pouch cell; (b) Sectonal vew (electronc magnfcaton scannng) of the core of a pouch cell [3]. All the three types of L-on cells contan dozens of parallel thn battery layers whose dmensons are on the order of 10 2 mcrons [4,5]. Each of the battery layers n the cells conssts of two electrodes (cathode and anode), a separator and two current collectors (copper for anode and alumnum for cathode). The electrodes and the separator are porous meda flled wth electrolyte as llustrated n Fgure 4. Durng cell usage, the current flow (from one electrode to the other n each of the battery layers n the cell) nvolves electronc charge transfer through an external electrcal crcut and onc charge transfer through the nternal path,.e., the electrolyte [6]. Modelng the thermal behavor of the L-on cells for vehcle electrfcaton applcatons s a challengng task. AVL has developed 3D electro-thermal models usng the Fnte Element Analyss (FEA) tool Abaqus [7] for smulatng the electro-thermal behavor of L-on battery cells wth varous geometres. The 3D-FEA model and the smulaton results for the varous battery cells/modules wll be dscussed n the followng sectons of ths paper. Fgure 4. Illustraton of the structure of a sngle battery layer n a L-on cell SIMULIA Communty Conference 3

4 2. Modelng Approach As aforementoned, a L-on battery cell has many parallel thn battery layers across ts thckness. In an deal desgn, the current flows n all of these battery layers are very smlar. In a L-on battery cell, the onc charge (the lthum ons) transfer from one electrode to the other takes place only through the electrolyte (Fgure 4),.e., the electro-chemcal processes n the cell are confned n the space between the two electrodes n each of the thn battery layers wth a dmenson n the order of 10 2 mcrons. In contrast, the electronc charge (the electrons) transfer n each of the battery layers n the cell takes place along the entre current-collector surfaces whose dmensons are several orders of magntude greater than that of the battery-layer thckness. Under a gven cell termnal current, the current densty dstrbutons are smlar n the current collectors as well as n the sources for the currents the electrodes for all the battery layers. As ndcated n Fgure 4, the electrolyte n each battery layer s dstrbuted n the pores of the electrodes and the separator. Dmensons for these pores are several orders of magntude smaller than that of the thckness of a sngle battery layer. For a cell under a dscharge process wth a current I, the dsspaton of chemcal energy nto heat n the cell can be characterzed by the dfference between the open crcut voltage of the cell E 0 (the best voltage that the cell can provde at a gven state of charge and temperature for a gven cell chemstry) and the termnal voltage V as V = E ( V + V + V ) (1) where V 1, V 2 and V 3 represent the three major voltage losses due respectvely to: (1) the ohmc resstance of the electrodes and current collectors, (2) the actvaton polarzaton at the electrode-electrolyte nterfaces, and (3) the concentraton polarzaton as a result of the unbalanced transent electronc current n the electrodes and onc current n the electrolyte. V 1 s related to the electronc current, and V 2 and V 3 are due largely to the onc current resstances n the electrolyte. For a gven State of Charge (SOC) and temperature (T), Fgure 5 llustrates the E 0 -V relatonshp at dfferent cell currents. All three voltage losses ncrease wth ncreasng cell current wth the polarzaton resstances domnate the overall resstance at hgh cell currents as llustrated n Fgure 5. Fgure 5. Voltage losses due to varous resstances under dfferent cell currents SIMULIA Customer Conference

5 The cell heat generaton Q resultng from cell chemcal energy dsspaton can be descrbed by the product of the cell current and the voltage drop due to the chemcal energy dsspaton as Q= I ( E0 V ) (2) Alternatvely Equaton 2 may be expressed as Q = I 2 (3) R where R = (E 0 V)/I s the cell nternal resstance. For battery cells n HEV and PHEV applcatons, R s commonly determned wth the Hybrd Pulse Power Characterzaton (HPPC) current I HPPC, whch can be defned based on the target pack load [8]. USDOE [8,9] and USCAR / USABC [10] recommend that R be evaluated wth a 10-second pulse HPPC current, for whch contrbutons are ncluded of the resstances to electrcal and onc charge transfers. The nternal resstance so-determned s only a functon of State of Charge (SOC) and temperature (T). Fgure 6 shows a R -DOD-T map for a reference L-on cell, where DOD (= 1 SOC) s the Depth of Dscharge. As llustrated n Fgure 6, the cell nternal resstance ncreases wth decreasng cell temperature and ncreasng depth of dscharge. Fgure 6. Internal resstance of a L-on cell wth temperature and DOD. The nternal resstance determned from the HPPC tests s a bulk property for a battery cell. In the thermal analyss of a cell, t s not practcal or necessary to model the detals of the dozens of thn porous battery layers n the cell. If the cell heat generaton can be estmated from the cell performance data, then based on the characterstcs of the current densty dstrbutons n the current collectors and electrodes n the battery layers, the cell may be smplfed to contan just one or several equvalent battery layers and modeled as a contnuous heterogeneous sold medum, 2012 SIMULIA Communty Conference 5

6 n whch heat conducton takes place [11,12]. Changes n the cell temperature due to the cell usage can thus be characterzed wth energy balance on a unt cell volume as T ρ C p = ( k T ) + q (4) t where ρ, C p and k are the local densty, heat capacty and thermal conductvty of the cell medum, T s the temperature, t s the tme and q s the heat generated. Because of the layered structure, thermal behavor of the cell should be characterzed wth the effectve thermal propertes. Each of the battery layers wth the same thermal propertes may be treated as a component of the heterogeneous sold medum. The composte local volumetrc heat capacty may be expressed as ρ C p, V ρ C p = (5) V where the subscrpt ndcates the propertes for the component and V s volume. Thermal conductvtes at the component nterfaces should be determned based on connectons of the components. For seres connectons, the composte thermal conductvty s gven as L k = ( L / k ) For parallel connectons, the expresson for the composte thermal conductvty becomes Lk k = (7) L In Equatons 6 and 7, L and k are the thckness and the thermal conductvty for the component respectvely. Smlarly, the composte electrcal conductvtes can also be expressed by Equatons 6 and 7 n characterzng the electrcal feld of the cell. For a gven cell wth properly defned boundary condtons for heat transfer, Equaton 4 can be solved usng FEA approach. In ths study, the electro-thermal behavor of the L-on battery cells and modules wll be characterzed usng 3D-FEA model developed at AVL. The approach of the cell modelng used n AVL battery electro-thermal model s llustrated n Fgures 7 and 8. Fgure 7 shows the couplng of governng equatons characterzng the electrcal feld and the temperature feld of the cell. The cell electrcal behavor s characterzed wth Posson equaton for the cell voltage potental, whch may be understood as Ohm s law n a dfferental form. The Ohm s law equaton and the Fourer equaton characterzng the thermal behavor of the cell are coupled through the current densty under a gven cell termnal current. Fgure 8 shows the procedure of characterzng the electro-thermal behavor of the cell. For a gven cell, ts electrcal behavor s characterzed wth the cell voltage potental V = V(DOD,A,T) and the cell nternal resstance R = R (DOD,B,T), where A and B are parameters related to the cell chemstry. Because the Ohm s law equaton does not contan tme, the modelng of the electrcal feld of the cell n a dscharge process nvolves the processes of ntalzaton, localzaton of the bulk cell propertes obtaned from the cell performance data, and contnuous update of the local parameters governng the local SOC (or DOD) and local nternal resstance durng the cell (6) SIMULIA Customer Conference

7 dscharge or charge. The spatal and temporal characterzatons of the electrcal feld of the cell are carred out by two user subroutnes. Due to the couplng of the electrcal feld and the temperature feld, varatons of the electrcal feld of the cell also nduce changes n the cell temperature dstrbuton. Fgure 7. Couplng of electrcal feld wth temperature feld of the cell. Fgure 8. Procedure for characterzng electro-thermal behavor of the cell. 3. Battery cell modelng In ths secton, the focus s on technques n modelng for cells wth dfferent geometres. For smplcty, drect ar coolng s assumed for all the cells. Valdaton of the model predctons wth 2012 SIMULIA Communty Conference 7

8 the avalable test data wll be dscussed n the next secton where the focus s on the coolng of the battery module. All battery cells studed n ths paper are L-on battery cells. 3.1 Pouch Cell In pouch cells the temperature gradent across the cell thckness s generally small and neglgble n comparson to those n the other two dmensons. Hence, pouch cells can be modeled wth only one equvalent battery layer. Fgure 9a shows a smplfed model for a pouch cell. The model ncludes the current collectng tabs and equvalent electrodes. Detals n modelng technques for a pouch cell can be found n the authors prevous work [11]. Fgure 9b shows the selected smulaton results for two pouch cells: an 8Ah cell (dmensons = 142mm 115mm 8.5mm) wth the postve and negatve termnal tabs arranged on the opposte sdes of the cell and a 5Ah cell (dmensons = 190mm 108mm 7mm) wth the two termnal tabs arranged on the same sde. The cells are ar cooled wth Heat Transfer Coeffcent (HTC) correspondng to that n channel flows. The smulated results represent cell temperatures at 80% DOD n a dscharge process from a fully charged state (DOD = 0) under 13.5C rate for the 8Ah cell and 11C rate for the 5Ah cell. It s seen that the cell temperature dstrbutons for pouch cells are nfluenced greatly by termnal tab desgns. Fgure 9. (a) Smplfed FEA model for pouch cell; (b) Smulaton results. 3.2 Prsmatc Cell Prsmatc cells are generally enclosed n a metal case. The thckness of a prsmatc cell s not too much smaller than the cell length and heght. Thus, the cell may be modeled wth multple equvalent battery layers n order to better smulate the maxmum dfferental temperature across the cell thckness. Fgure 10a shows the smplfed model for a 6Ah prsmatc cell (dmensons = 112mm 70mm 27mm). The model ncludes a metal case, two termnal poles, current collectng tabs and equvalent electrodes for three equvalent battery layers. The cell s ar cooled wth HTC correspondng to that n a channel flow. Fgure 10b shows the selected smulaton results for both cell surface and core temperatures at 80% DOD n a dscharge process from a fully charged state (DOD = 0) under 10C rate. The three-layer model predcts that the maxmum cell temperature s n the center of the cell for ths prsmatc cell and the cap of the cell case has the lowest temperature snce t has no drect contact wth the cell SIMULIA Customer Conference

9 Fgure 10. (a) Smplfed FEA model for prsmatc cell; (b) Smulaton results. 3.3 Cylndrcal Cell All cylndrcal cells have metal cases and the cell termnals are generally arranged on opposte ends of the cell. Due to the spral wound structure, the surface area of a battery layer vares wth the radus of the layer poston. Hence the number of current collecton tabs also vares wth the surface area of the battery layer n the cell [13]. Several methods have been proposed to model the spral wound structure of the cylndrcal cell [14-17]. The technques developed prevously for modelng cylndrcal cells have been revewed by the authors of ths study [11]. Fgure 11 shows the smplfed models for a 2.3Ah cylndrcal cell (dameter = 26mm; heght = 65 mm): (a) oneequvalent-layer model and (b) three-equvalent-layer model. Fgure 11. (a) One-equvalent-layer model; (b) Three-equvalent-layer model SIMULIA Communty Conference 9

10 Each equvalent battery layer has three concentrc rngs: the nner rng s an equvalent cathode wth a postve current collector connectng t wth the postve termnal of the cell; the outer rng s an equvalent anode wth a negatve current collector connectng t wth the negatve termnal of the cell; the mddle rng represents the separator. The cell s assumed to be cooled under natural convecton wth 25 C ambent ar and HTC = 10 W/m 2 - C. The selected smulaton results for cell temperatures at 90% DOD n a dscharge process from a fully charged state (DOD = 0) under 10C rate are also presented n Fgure 11. It s seen that the maxmum cell temperature predcted by the three-equvalent-layer model s about 1 C lower than that predcted by the one-equvalentlayer model under the specfed heat generaton and coolng condtons. Ths s because the cell heat dstrbuton s more unform n the three-equvalent-layer model than n the one-equvalentlayer model. It s also apparent that the maxmum-cell-temperature locaton shfts to the nner adabatc surface of the cell as the number of the equvalent battery layers n the model ncreases. 4. Battery module modelng 4.1 Battery module wth ndrect lqud coolng Fgure 12 shows a three-cell module wth ndrect lqud coolng. In the module, three 80Ah pouch cells (dmensons = 240mm 260mm 11.5mm) are arranged n seres electrcally wth ther termnal tabs connected wth busbars. Each cell n the module s cooled va a 1.5-mm thck alumnum coolng fn whch s n contact wth a cold plate located on the opposte sde of the cell termnal tabs. The sde surface of the cell that s not n contact wth the coolng fn s thermally nsulated from the next coolng fn by an elastomerc thermal pad. The cells were modeled wth a one-equvalent-layer model as llustrated n Fgure 12. Fgure 12. FEA model for three pouch cell module wth ndrect lqud coolng SIMULIA Customer Conference

11 Fgure 13 compares the measured and smulated temperatures at 90% DOD n a dscharge process from a fully charged state (DOD = 0) under 2C rate. Durng the test, the cold plate surface temperature was mantaned at 25 C. Because the termnal tab and busbar temperatures were not measured, only the cell temperatures are compared. Snce the three cells n the module are thermally symmetrc, only the results for the mddle cell n the module are presented. The measured cell temperatures are presented n a contour plot wth 0.5 C as the nterval between two contour lnes. It s seen that the smulated cell temperatures correlate reasonably well wth the measurements both n the maxmum cell temperature and the cell temperature dstrbuton. Ths ndcates that the model can reasonably characterze the thermal behavor of the cells n the module. Fgure 13. Comparson of measured and smulated cell temperatures. 4.2 Battery module wth drect ar coolng Fgure 14a shows the FEA model for a module assembled wth 44 A123-ANR26650M1A cylndrcal cells (dameter = 26 mm; heght = 65 mm) wth drect ar coolng. The module has an 11P4S confguraton,.e., the Ah cells are arranged n 4 groups that are connected n seres. The 11 cells n each of the groups are connected n parallel. The module was taken from an A123 Hymoton L5 PCM battery pack consstng of 14 dentcal modules connected n seres [18,19]. For smplcty, the one-equvalent-layer model dscussed n Secton 3.3 was used to characterze the electro-thermal behavor of the cells n the module. In order to obtan more accurate thermal boundary condtons for the cells n the module, 3D Computatonal Flud Dynamc (CFD) smulatons on the ar flow dstrbuton n the module were performed usng AVL FIRE [20]. Several analyss teratons were performed between FEA and CFD where the wall temperature (as boundary condton for CFD analyss), the HTC and the ar temperature (as boundary condton for the FEA analyss) were updated n successve teratons. More detaled dscusson regardng the CFD smulatons on the ar flow dstrbuton n the module were reported n a prevous paper by 2012 SIMULIA Communty Conference 11

12 the authors [11]. The sectonal vew n Fgure 14b shows the smulated module temperatures at 90% DOD n a dscharge process from a fully charged state (DOD = 0) under 5C rate. Fgure 15 compares the smulated cell temperatures wth the measurements under 5C dscharge rate for three selected cells: cell A at the module entrance and cells B and C at the module ext wth B at the center and C on the sde. Overall, the smulated cell temperatures and the measurements agree reasonably well. The relatvely poor predcton for cell B was due to the nfluence of the ar leakage through the two bolt holes along the centerlne of the module n the test, whch generated addtonal coolng over the busbar surfaces. The busbar surface coolng was not consdered n the smulaton. Fgure 14. (a) FEA model for the 44 cylndrcal-cell module; (b) Smulaton result. Fgure 15. Comparson of the smulated and measured cell temperatures SIMULIA Customer Conference

13 4.3 Battery modules wth ndrect ar coolng and heatng Coolng analyss Fgure 16 shows the FEA models for two pouch-cell modules wth ndrect ar coolng. The two modules are dfferent only n the structures of the ar coolng channels: one wthout fn nserts (Fgure 16a) and the other wth fn nserts (Fgure 16b). Both modules are stacked wth twelve 8Ah pouch cells (dmensons = 140mm 190mm 8.5mm) connected n seres. The cells are cooled ndrectly through sx 1-mm thck alumnum coolng plates sandwched between each par of cells. Each of the coolng plates has an extended coolng fn exposed n the ar flow channel. In the space between each par of cells wthout a coolng plate, a 1-mm thck elastomerc thermal pad s nserted. For a better vew of the coolng unts n the module, the module frame and electrcal connectors are removed from the dsplay. The secton of the coolng plate through the module plastc frame and the fn secton of the coolng plate (where ar coolng takes place) s each 19 mm wde. The coolng-plate ptch (.e., the space for ar flow between the two coolng plates) s also 19 mm. Because the length of the ar coolng channel s hydraulcally short (channel-length/plateptch < 10), some coolng ar n the center of the channel may not be effectvely nvolved n coolng of the plate (.e. short crcuted). To mprove the effectveness of ar utlzaton, a model wth fn nserts placed n spaces between the coolng plates (Fgure 16b) s also evaluated. The fn nserts have an equvalent fn densty of 12 fns/n, wth a total of 10 fns nserted n the 19 mm space between the coolng plates. Fgure 16. FEA models for 12-pouch-cell modules and smulaton results: (a) wthout fn nserts n ar channel; (b) wth fn nserts n ar channel SIMULIA Communty Conference 13

14 Thermal behavor for both modules was evaluated under 5C dscharge rate from the fully charged condton (DOD = 0) to 80% DOD. Boundary condtons for heat transfer n the ar flow channels are T ar = 35 C and HTC = 60 W/m 2 - C (correspondng to channel flow). The smulaton results plotted to the same scale are also presented n Fgure 16. As seen n Fgure 16, temperatures of the fn nserts are close to the ar temperature, suggestng that the fn nserts mprove the ar utlzaton consderably. Fgure 17 plots the temperature dstrbutons along the center lnes of the coolng plate n the mddle of the module for the two models. The followngs are apparent: (1) the temperatures of the fn nserts are close to the ar temperature, suggestng that the fn nserts mprove the ar utlzaton and heat transfer effectveness consderably, and (2) the temperature dstrbuton among coolng plates s nfluenced sgnfcantly wth changes n heat transfer on coolng-fn surfaces. Fgure 17. Temperature varatons along centerlne of coolng plate at DOD = 80% Heatng analyss The same two battery modules as outlned and shown n Fgure 16 were used n the warm-up analyss. In ths case t s assumed that the battery pack s thermally soaked n -20 o C envronment. The module s dscharged under a constant 1C rate to provde electrcal power to the heater. Hot ar boundary condton of T ar = 40 o C and HTC = 60 W/m 2 - o C are appled to the ar flow channels for the pack warm-up. The pack s fully functonal only after the mnmum cell temperature becomes 0 o C. The tme t takes for the cell temperature to reach 0 o C was hence used as a crteron for comparng the effectveness of the two coolng fn desgns. Because of the small temperature dfference between the cell and the coolng plate as prevously observed, the cell warm-up s evaluated based on changes n the temperature of the alumnum plates. Fgure 18 shows temperatures n the modules (plotted to the same scale) wthout and wth fn nserts n the ar channels at the tme when the mnmum plate temperature reaches 0 o C at the adabatc end. Fgure 19 shows the transent temperatures at the adabatc end of the alumnum SIMULIA Customer Conference

15 plate durng the heatng process. The smulaton results ndcate that the fn nserts mprove the heatng effcency sgnfcantly. Wthout and wth fn nserts, t takes 820 seconds and 480 seconds respectvely for the adabatc end of the alumnum plates to reach 0 o C. Wth the fn nserts, the warm-up tme s 340 seconds shorter. The shorter the warm-up tme, the lower the energy requred for the pack warm-up. It also means that the battery pack can be fully functonal that much sooner. Fgure 18. Results of heatng smulaton at the tme the adabatc edge reaches 0 o C for module wthout and wth fn nserts. Fgure 19. Transent temperature at the adabatc edge of the coolng plate durng warm-up. Fgure 20 shows the heat flux to each cell together wth the nternal heat generated durng the heatng process. The results show the effectveness of the fn nserts n extractng energy from the heatng ar. It also shows that cell self heatng contrbuted a small percentage of the overall heat durng the cell heatng process SIMULIA Communty Conference 15

16 Fgure 20. Heat flux to each cell and cell self heatng durng warm up. 5. Summary AVL has developed a method to characterze the electro-thermal behavors of L-on battery cells. 3D FEA models were used to analyze the thermal behavor of commercally avalable cylndrcal, prsmatc and pouch battery cells. Characterstcs of these three types of cells may be summarzed as follows. (1) Pouch cells have neglgble temperature dfference across ther thcknesses. However they can have large dfferental cell temperatures across ther cell surfaces n hgh current applcatons. For ths type of cells, maxmum cell temperatures are typcally near the termnal tabs where local heat fluxes are hgh due to hgh local current denstes. (2) For cylndrcal cells, cell core temperatures can be sgnfcantly dfferent from cell surface temperatures. Ths s because of the adabatc condton at the cell core and decreasng heat transfer resstance wth ncreasng radus. (3) For prsmatc cells, dfferental temperature across the cell thckness must be consdered. However, large thermal mass for ths type of cell may mtgate the cell temperature rse. It s found that for cased cells (cylndrcal and prsmatc cells) the maxmum cell temperature s n the core of the cell, not near the cell termnals as observed for the pouch cells. Smulaton results were compared wth avalable test data for battery temperature measurements coverng pouch-cell and cylndrcal-cell modules under dfferent coolng condtons. Good agreement between the smulatons and measurements were observed, ndcatng that the 3D electro-thermal model employed n ths study reasonably characterzes the electrothermal behavor of the L-on battery cells SIMULIA Customer Conference

17 6. References Yuft, V., Shearng, P., Hamlton, R.W., Lee, P.D., Wua, M. and Brandon, N.P., Investgaton of Lthum-on Polymer Battery Cell Falure Usng X-Ray Computed Tomography, Electrochem. Commun., Vol.13, pp , Lnden, D. and Reddy, T.B., Handbook of Batteres, 3rd ed., McGraw-Hll, Brodd, R.J., "Lthum-Ion Cell Producton Processes," Chapter 9, Advances n Lthum-Ion Batteres, W.A. Van Schalkwjk and B. Scrosat (ed.), Kluwer Academc / Plenum Publshers, Thomas, K.E., Newman, J. and Darlng, R.M., "Mathematcal Modelng of Lthum Batteres," Chapter 12 n Advances n Lthum-Ion Batteres, Van Schalkwjk, W.A. and Scrosat, B. (ed.), Kluwer Academc / Plenum Publshers, Smula, Abaqus/Standard documentaton, 8. Idaho Natonal Engneerng & Envronmental Laboratory, PNGV Battery Test Manual, DOE/ID-10597, Idaho Natonal Engneerng & Envronmental Laboratory, Battery Test Manual for Plug-In Hybrd Electrc Vehcles, INL/EXT , USCAR, USABC Manuals, avalable onlne va Storage-TLC. 11. Ma, Y., Teng, H. and Thellez, M., "Electro-Thermal Modelng of a Lthum-on Battery System," SAE Paper No , Ma, Y. and Teng, H., "Comparatve Study of Thermal Characterstcs of Lthum-on Batteres for Vehcle Applcatons," SAE Paper No , Chu, A.C., Gozdz, A.S., Rley, G.N. and Hoff, C.M., "Battery Cell Desgn and Method of Its Constructon," U.S. Patent 2007/ A1, November 22, Inu, Y. Kobayash, Y. Watanabe, Y. Watase, Y. Ktamura, "Smulaton of Temperature Dstrbuton n Cylndrcal and Prsmatc Lthum Ion Secondary Batteres," Energy Converson and Management, Vol.48, pp , S. Al-Hallaj, H. Malek, J. S. Hong, J.R. Selman, "Thermal Modelng and Desgn Consderatons of Lthum-on Batteres," J. Power Sources, Vol.83, 1 (1999). 16. T. D. Hatchard, D. D. MacNel, A. Basu, J. R. Dahn, "Thermal Model of Cylndrcal and Prsmatc L-on Cells," J. Electrochem. Soc., Vol.148, 7, T. I. Evans, R. E. Whte, " A Thermal Analyss of a Sprally Wound Battery Usng a Smple Mathematcal Model," J. Electrochem. Soc., Vol.136, 2145, Iu, H.Y. and Smart, J., "Determnng PHEV Performance Potental User and Envronmental Influences on A123 Systems Hymoton Plug-In Converson Module for the Toyota Prus," EVS24 Internatonal Battery, Hybrd and Fuel Cell Electrc Vehcle Symposum, Stavanger, Norway, May 13-16, AVL GmbH, AVL-FIRE, Theory, Verson 2009, AVL, Graz, SIMULIA Communty Conference 17