EVS27 Barcelona, Spain, November 17-20, 2013 Lithium-ion Battery Electro-thermal Moel an Its Application in the Numerical Simulation of Short Circuit Experiment Chengtao Lin 1, Can Cui 2, Xiaotian Xu 3 1 The State Key Laboratory of Automotive Safety an Energy, Tsinghua University, Beijing, China, 100084,lct@tsinghua.eu.cn Abstract As a key issue in EVs (Electric Vehicles) evelopment, Li-ion battery s thermal safety is focuse on in which the thermal abuse moel is an important tool. This paper has establishe a 3-D electro-thermal moel taking sie reaction into account, in orer to stuy on MGL100Ah lithium manganite battery s safety issue. This paper firstly states the mechanism of heat generation, types of thermal moel, an the mesh moel, then analyzes the simulation on the moel uner the short circuit conition, one of thermal abuse conition. On one han, the paper verifies the liability of the moel, so as to apply the moel on other thermal abuse conition in the future; on the other han, the paper iscusses in etail the internal temperature istribution, voltage an current changes uner the short circuit conition. Analyzing this typical safety issue in etails, this paper can help reuce cost, raise esign efficiency an avoi thermal abuse in future evelopment of battery. Keywors: Lithium battery, Thermal Management, Short Circuit, Moeling 1 Introuction The vehicle inustry has evelope rapily an become biggest inustry in many large inustrial countries since the 21st century. In 2013, many Chinese cities were threatene by fog an haze, which can raise probability of lung cancer an harm people s health. Traitional fuel vehicle emission is consiere to be the culprit of that poor weather. Nowaays non-renewable energy like petroleum is going to the en ay by ay. Due to the worl s sustainable strategy, as a result of solving both energy an pollution problems, EV (Electric Vehicle) is uner lots of stuies in many countries. Storage battery has won high attention as is the heart of EV. Differing from orinary cell phone or laptop battery, the large-scale battery in EV must possess 3 main performances: 1) Goo safety; 2) High specific energy an specific power; 3) Low cost. These are the basic elements for application an promotion of future EV. Li-ion battery is one of the most high-potential energy sources in the future, which is the high-quality battery satisfying any working conition of EVs. Li-ion batteries have high specific energy an specific power, but they are oubte in safety for many times. In 1995 an 1997, there were two li-ion battery explosion accients in Japan which cause large loss of property an personal injuries. Every accient reveale problems that nee to be solve relate to li-ion safety. Many experiments an analyses tell that uner short circuit, large current over-charge an high temperature etc. conitions the li-ion battery was in unpreictable anger. In orer to iscovery these thermal abuse conition, nowaays, each country has EVS27 International Battery, Hybri an Fuel Cell Electric Vehicle Symposium 1
built unifie battery safety stanar, for instant, IEC, JIS, GB, EN, etc. In these stanars, thermal abuse behaviors are state an stanarize generally, an they are use to test li-ion batteries safety. Many researchers from ifferent institutes has stuie much on Li-ion battery s abuse: Hatchar et al mainly focuse on small-scale Li-ion batteries (e.g. 18650 cyliner battery) an their abuse simulation; Kim et al from National renewable energy laboratory extene that into 3D abuse moel an carrie out simulation in epth on oven test. However, in this fiel the stuy on large-scale battery has not been carrie out yet, especially uner external short circuit conition. This paper will buil a 3D electro-thermal moel which can simulate thermal abuse performance of 100Ah lithium manganate battery well. Combining experiments with simulations, this paper mainly stuies on the safety issue uner short circuit conition, an preicting the thermal behavior uner thus conition. Using this moel we expect that the cost in battery evelopment will be ecrease an the esign efficiency will be raise, what s more, achieve the goal of reucing thermal abuse by controlling temperature. 230mm 80mm 40mm 25mm Fig.2 Battery unit s layer moel 40mm 150mm 40mm 230mm 80mm 300um 40mm 590um 22mm 320mm Fig.3 MGL100Ah battery s spatial size 25mm 10mm 2 Battery Introuction 2.1 Battery structure an size The object of the stuy is MGL100Ah lithium manganite battery prouce by Citic Guoan MGL Power Sources Co, Lt. It s a large-scale storage battery fitting Evs. In micro view, the battery is arrange in stacks. One single stack unit contains 5 layers: current collector on positive electroe, positive electroe, separator, negative electroe an current collector on negative electroe, as is shown in Fig.2. The whole battery is mae up of 88 fairly same stacks. In macro view, MGL100Ah battery is 295mm(X) x 320mm(Y) x 22mm (Z). The sizes of every etail are shown in Fig.3 an Fig.4. Fig.1 MGL100Ah Li-on battery 10mm 320mm 10mm 10mm Fig.4 MGL100Ah battery shell s spatial size 2.2 Battery materials Current collector is mae with homogeneous an isotropous metal, aluminum by the positive sie an copper by the negative sie. Positive electroe material is LiMnO3 while negative one is graphite. Accoring to the pre-stuy, the separators are fille with electrolyte, where separator material is PE/PP/PE, electrolyte is LiPF6, organic solvent is DEC+DMC+EC; positive electroe mainly contains positive active material, biner an electrolyte while negative electroe contains negative active material, biner an electrolyte. Accoring to integrative compute an experiments, this paper obtaine the thermal physical parameter of the three parts mentione above. Next, all these part together will be calle one unit. Besies, one positive an one negative lug are extene out from the two kins of current collectors, an these two lugs are connecte with external circuit. EVS27 International Battery, Hybri an Fuel Cell Electric Vehicle Symposium 2
Inner Resistance(mohm) OCV(V) 2.3 Battery electric performance After HPPC (Hybri Pulse Power Characteristic) experiment, we collecte the battery s OCV (Open Circuit Voltage) an equivalent internal resistance changes with SOC (State of Charge), which are shown in Fig.5 an Fig.6. 4.2 4 3.8 3.6 3.4 0 0.5 SOC 1 Fig.5 OCV changes with SOC 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0 0.5 1 SOC Fig.6 battery s internal resistance changes with SOC The OCV will grow when SOC grows, while the equivalent internal resistance will escen. Through fitting calculation, the paper obtaine 2 fitte curves, which are shown in Eq.1 an Eq.2. Eq.1 Eq.2 3 Establishment of 3D electrothermal moel 3.1 Thermal physical parameters Through experiments an research, this paper obtaine the thermal physical parameters, they are shown in Table.1. 3.2 Mesh moel The mesh moel for follow-up numerical simulation is shown in Fig.7, which is mae up of triangle meshes. The very thin parts are meshe with ajuste laxity egree an the whole moel consists of 67836 gris. Fig.7 Battery s mesh moel 3.3 Thermal moel 3.3.1 Heat equilibrium equation The transient 3D energy transfer formula in Li-ion battery is shown in Eq.3, an this formula can fit any tiny unit element in the battery, where is the ensity of the unit, is equivalent specific heat capacity, is temperature, is equivalent heat conuctivity coefficient (which is a matrix in 3D moel as shown in Eq.4), is heat generation rate, an is surface s heat iffusion rate. ( ) Eq.3 kxx 0 0 k 0 k 0 yy 0 0 k zz Eq.4 One popular heat generation moel is raise by D. Bernari in 1985 [ i ] as shown in Eq.5, which can explain the whole battery s heat generation conition, where Q is the battery s heat generation rate, Q I is irreversible heat generation rate, Q r is reversible heat generation rate, I is the current through the battery, E is the equilibrium voltage, U is the working voltage, T is its absolute temperature. Q Q i Q r I(E U) IT E Eq T In orer to get a more accurate moel, we take sie reaction heat generation into consier, it becomes Eq.6 which is similar to Sato heat generation rate moel[1]: Q Q r Q p Q s Q j nft ( E 0 ) nf(e T 0 U L ) Q s Q j Eq.6 Where Q r is the heat from battery reaction, Q j is the Joule heat, Q p is the polarization heat an Q s is the sie reaction heat. Unit s heat iffusion rate q is can be ivie into two parts: the convection heat transfer rate q an the raiative heat transfer rateq r, as shown in Eq.7. Eq.7 Equation of the convection heat transfer is shown in Eq.8, where is the surface s heat transfer EVS27 International Battery, Hybri an Fuel Cell Electric Vehicle Symposium 3
Table.1 Temophysical parameters in every layer material Current collector of P.E Positive Electroe Separator Negative electroe Current collector of N.E Material aluminum lithium manganate PP PE PP graphite copper Density (kg.m-3) 2.70 10 3 2.80 10 3 1.30 10 3 1.20 10 3 8.92 10 3 Specific heat capacity 880 839 1980 1064 390 (J.kg-1.K-1) Thermal conuctivity 237 3.9 0.6 3.3 397 (W.m-1.K-1) Resistivity in X (Ω.m) 2.83 10-8 33.2 1.26 1.75 10-8 Resistivity in Y (Ω.m) 2.83 10-8 33.2 1.26 1.75 10-8 Resistivity in Z (Ω.m) 2.83 10-8 33.2 0.25 1.26 1.75 10-8 coefficient, is the surface area, T is the surface s temperature, T ir is the atmosphere temperature. Equation of the raiative heat transfer is shown in Eq.9, where is Stefan-Boltzmann constant, is emissivity, T is the ambient temperature. The q r is always much less than q, so the term q r can be ignore except when it s in an oven test. Q (T T ir) Eq.8 ( ) Eq.9 3.3.2 Sie reaction heat generation moel Sie reactions mainly consist of SEI (Soli Electrolyte Interface) ecomposition, Negativesolvent reaction, positive-solvent reaction an electrolyte ecomposition. Batteries uner special conition may cause ifferent egrees of sie reactions. So the heat generation rate from sie reactions nee to be taken into account in special working conitions. Base on the stuies from National Renewable Energy Laboratory s researcher Gi-Heon Kim et al [ii], we hereby expan the equations into our Li-ion batteries. Following the formulation of Hatchar et al [ iii ], the 4 sie reactions can be expresse in the following equations: (4) SEI (Soli electrolyte interface) ecomposition reaction: ( ) Eq.10 Eq.11 Eq.12 where s i is the imensionless amount of metastable SEI, R(s -1 ) an A(s -1 ) are frequency parameters, E (J mol -1 ) is the activation energy of the reaction, heat release, s i (J g -1 ) is the SEI-ecomposition (g m -3 ) is the volume-specific carbon content. (2) Negative-solvent reaction: (T s i) ( ) E T Eq.13 Eq.14 Eq.15 where an s i is the imensionless amount of lithium in the carbon/sei, R(s -1 ) an A(s -1 ) are frequency parameters, E (J mol -1 ) is the activation energy of the reaction, (J g -1 ) is the heat release, (g m -3 ) is the specific carbon content. (3)Positive-solvent reaction: E p (T ) p ( ) T Eq.16 p p p p Eq.17 p Eq.18 where is the conversion egree, R(s -1 ) an A(s -1 ) are frequency parameters, E (J mol -1 ) is the activation energy of the reaction, p (J g -1 ) is the heat release, p(g m -3 ) is the specific positive-active content. (4)Electrolyte ecomposition: ( ) Eq.19 Eq.20 Eq.21 where is the imensionless concentration, R(s -1 ) an A(s -1 ) are frequency parameters, E (J mol -1 ) is the activation energy of the reaction, (J g -1 ) is the heat release, (g m -3 ) is the specific electrolyte content. 4 Experiment This paper mainly iscusses about external short circuit numerical simulation which hasn t been stuie much on. For follow-up ifferent conitions simulation, the experiment is esigne for knowing about the battery s thermal behavior as well as comparing with following simulation results. First, we nee to obtain the real conition results uner external short circuit. Following is the process of the experiment. 4.1 Experiment process (1)Device reay: Test evices, ensure reliability an safety; EVS27 International Battery, Hybri an Fuel Cell Electric Vehicle Symposium 4
(2)Design for battery ata collector: As is shown in Fig.8, set 6 ata collectors an connect the evice with the battery; (3)Start: Start collecting ata, close the circuit to short-circuit the battery, keep on 10min after the voltage stop ropping; (4)Finish: Power off evices, stop collecting. Fig.10 Voltage an current changes uring short circuit Fig.8 Short circuit Experiment s ata collectors First of all, there is no significant physical change observe from outsie look, no plumping or exploing, as is shown in Fig.9. Fig.9 (a) Before short circuit Fig.9 (b) After short circuit 4.2 Experiment results an analysis Through processing the collecte ata, the current an voltage changes with time are shown in Fig.10. The whole short circuit perio laste until about 7 secon when the voltage rop to 0V. The maximum current is nearly 1800A. Fig.11 Temperature changes uring short circuit The temperature changes are shown in Fig.11. Since the short circuit perio was short, the battery boy temperature rise was small, no significant physical changes occurre on the boy. But the lug temperature rise was relatively huge, even after short circuit stoppe; the lug s temperature still rose fast for a while. It was because the temperature of the furthest point of the lug an of external circuit was far higher than point 6, so that they coul convey heat to the lug even if the short circuit stoppe. Thus, the lug s highest local temperature was higher than that in the ata, an it occurre on the ege of the lug where connecte with external circuit. The local high temperature cause metal fusing which stoppe the short circuit perio. It is obvious that some safety problems may lie in this high local temperature status. In general, the short circuit perio, temperature changes, current an voltage changes were characteristics stating battery external short circuit, an they provie the comparison source. 5 Simulation results an analysis 5.1 Simulation results Through builing 3D electro-thermal moel, this paper carrie out the numerical simulation on battery external short circuit. During simulation, this perio was 7.2s, an then the voltage suenly roppe to 0V. We obtaine the current change uring the simulation which is shown in Fig.12. The maximum EVS27 International Battery, Hybri an Fuel Cell Electric Vehicle Symposium 5
of current was nearly 1800A at the beginning an then slowly went own, that fitte the experiment ata well. on the lugs. The error comes from external circuit evice s heat transfer, for there is only the battery built in the simulation without external circuit evices. Uner this conition, the external circuit s heat transfer can harly influence the temperature of the battery boy, but can only influence the ege of the lugs. Due to the ifferent heat generation rate between lugs an external circuit evices, the graients shown in Fig.14 of 2 lug curves are not exactly the same. This error is in reasonable range. Fig.12 Current uring short circuit Fig.14 Surface temperature changes uring short circuit Fig.13 Temperature istribution after short circuit Using postprocess program, we obtaine the temperature istribution of the battery. Fig.13 shows the temperature istribution at the time of 7.2s. It s easy to ientify that 2 lugs have biggest temperature rises while the boy has no significant change in temperature. The highest temperature point is at the en of the positive lug (Aluminum) where it fuse. The process resembles the experiment very much. 5.2 Simulation results analysis Firstly, the temperature comparison between ata collecte an simulation ata at the same point, as is shown in Fig.14. Center_Lab an center_sim show temperature curves at point 5 from experiment an simulation respectively, meanwhile, lug_lab an lug_sim show temperature curves at point 6 from experiment an simulation respectively. Both simulation curves trens are very similar to that in experiment, an also the lugs temperature rise is nearly the same, far more than the temperature rise in the battery boy. The only ifference is the temperature rise rate is not same Fig.15 Voltage rop uring short circuit Fig.16 Inner temperature istribution at 7.2s The voltage change is shown in Fig.15. The trens are basically same but the experiment value is a little smaller. To analyze this little error, we start from the metal s resistance increasing with the temperature rising. So every point in lugs has ifferent increase in resistance, the further from the battery boy the greater change it has, namely, the internal resistance of the whole battery is getting bigger. Thus the external circuit s voltage is getting smaller. However, the internal resistance we use in simulation inclues EVS27 International Battery, Hybri an Fuel Cell Electric Vehicle Symposium 6
the lugs resistance ignoring the little change in lugs resistance which finally cause the little ifference. The sie reactions are not significant, as is shown in Fig.16, but they have inee occurre a little, as is shown in Fig.17, where hr_sei expresses the heat generation of SEI ecomposition, hr_ne expresses the heat generation of Negative-solvent reaction, hr_pe expresses the heat generation of Positivesolvent reaction, hr_ele expresses the heat generation of Electrolyte ecomposition, an all the units are translate into temperature rise of the whole battery (K/min). It is shown that these4 sie reactions can be escribe well by this moel, however, the level of reactions are very low ue to the very small temperature rise in the battery boy. Only SEI ecomposition an Negative-solvent reaction starte a little. Compare with that in oven test simulation, these can be ignore. 1) 3D moel can help know eeply an irectly about the temperature istribution in the battery uner any working conition, this is an effective way to evaluate battery s safety. 2) The simulation results in this paper basically fit the real experiment s ata, incluing potential angers. The moel has great application value in guiing the battery esigns. 3) In the very short time perio of external short circuit, the current is huge, thus the heat impulse is huge on lugs which are potential safety hazar points, while the internal thermal safety mainly lies in battery materials an structure. 4) The moel stuie in this paper can be applie in other thermal abuse conitions (e.g. over charge, oven test, etc.). This moel will make great contributions in methos to the future Li-ion battery thermal management technology in epth. Acknowlegements Gratitue must be extene to those who ha ever helpe in the stuy of this paper: NSFC project (51006056), Inepenent fune project of Tsinghua University (2011THZ01004), Professor Guangyu Tian an Jiangbo Zhang from Tsinghua University, staffs of Citic Guoan MGL Power Sources Co, Lt. Fig.17 Heat generation rate of sie reactions Thus, the main heat source is Joule heat from circuit uner short circuit conition. The inner temperature rise is little so that sie reactions activities are rather low. No sie reaction occur, it is goo for battery safety. In general, the 3D electro-thermal moel establishe here can simulate the physical fiel s changes well, an the errors are in reasonable range without any problems in simulation. Results from both experiment an simulation show that the lugs are playing important roles uner external short circuit conition. The local high temperature may bring anger. To avoi the unnecessary loss of property, suggestions are mae: 1. Set the lugs as metal with low melting point which can play as a fuse; 2. Set aitional short circuit protection circuit. 6 Conclusion This paper built a 3D electro-thermal moel accoring to MGL100Ah lithium manganite battery which is commonly use. An the paper carrie out the experiment an simulation uner external short circuit one of the typical thermal abuse conitions. The results show the stuy metho is feasible an tells: References [1] D. Bernari, E. Pawlikowski, J. Newman. A General Energy Balance for Battery Systems. J. Electrochem. Soc., 1985, 132(1): 5-12. [2] Sato, N., Thermal behavior analysis of lithiumion batteries for electric an hybri vehicles. power sources, 2001(77): p. 71 [3] Gi-Heon Kim, Ahma Pesaran, Robert Spotnitz. A three-imensional thermal abuse moel for lithium-ion cells. Journal of Power Sources 170 (2007) 476 489 [4] T.D. Hatchar, D.D. MacNeil, A. Basu, J.R. Dahn, J. Electrochem. Soc., 148 (7) (2001) A755 A761. Authors Can Cui, Postgrauate The State Key Laboratory of Automotive Safety an Energy Department of Automotive Engineering, Tsinghua University Beijing 100084, P.R.China Tel: +86 1381-159-1351, E-mail: cuican0118@139.com Chengtao Lin, Ph.D, Associate Professor EVS27 International Battery, Hybri an Fuel Cell Electric Vehicle Symposium 7
The State Key Laboratory of Automotive Safety an Energy Department of Automotive Engineering, Tsinghua University Beijing 100084, P.R.China Tel: +86-10-62785599, Fax: +86-10-62784616, E- mail: lct@tsinghua.eu.cn Xiaotian Xu, Postgrauate The State Key Laboratory of Automotive Safety an Energy Department of Automotive Engineering, Tsinghua University Beijing 100084, P.R.China Tel: +86-1521-056-0430 EVS27 International Battery, Hybri an Fuel Cell Electric Vehicle Symposium 8