Documente Academic
Documente Profesional
Documente Cultură
Fig. 3. Rate factor for the lithium-ion battery (Sony US18650). The Fig. 5. Temperature factor for the lithium-ion battery (Sony US18650). The
reference current is 0.7 A, corresponding to a 0.5C rate. reference temperature is 23 C.
(5)
(6)
Fig. 9. Rate dependence of the battery voltage for currents of (from top to
bottom) 0.28 A, 0.7 A, 1.0 A, 1.4 A, and 2.8 A at a constant temperature of
Fig. 7. Circuit for validating the battery model. The battery is discharged
23 C.
through a constant current load, while heat energy is transferred rapidly to the
ambient.
Fig. 13. Battery charge characteristics at 23 C and 1.0 A. Fig. 15. Battery voltage under different cooling conditions.
TABLE I
PARAMETERS FOR SIMULATION OF US 18650 BATTERY
TABLE II
HYBRID POWER SOURCE PARAMETER SETUP
Fig. 19. CDMA discharge test for a single cell of Sony US18650 Battery.
pulsed current. The advantages of such a system are that the Comparison of the simulation result to the experimental data.
power and discharge life of the battery can be boosted and the
overall system power density can be increased [31]. The prin-
IV. DYNAMIC SIMULATIONS
ciple of operation of the system is that the battery recharges
the ultracapacitor during the pulse-off period and the ultraca- To further validate the model, extensive dynamic simula-
pacitor then augments the battery capacity during the pulse on tions were conducted following some industry standards, for
period. This delivers a higher power to the load, with lower in- example, the CDMA cellular load cycle [34]. Fig. 18 shows
ternal losses. The ultracapacitor model used here, based on that the pulsed current discharge profile of CDMA2, in which the
of Miller et al. [32], [33], represents the Maxwell PowerCache battery discharges at 1.4 A for 250 ms and then at 0.2 A for the
PC100 and was previously validated. Table II lists the system rest of the 2.55 s pulse period. The battery runs until it cuts off
parameters. at 2.5 V.
Fig. 17 shows the battery current, the ultracapacitor current In Fig. 19, the simulation result for the battery voltage is pre-
and the load current for one period (4 s) in steady state oper- sented and compared against the experimental test. The exper-
ation. Zero seconds on the time axis refers to the beginning of imental data were obtained by testing a brand new US 18650
the data acquisition period, not the beginning of the experiment. cell taken from a consumer electronics item. The pulsed current
Notice that the battery discharges (positive current) both while discharge mode was set up by using an electronic load and con-
the load pulse is on and off, while the ultracapacitor discharges trolled by the LabView software. To protect the battery, the cut
(positive current) when the load is on but charges (negative cur- off voltage in the test was set to 3.0 V. Since the pulse rate is high
rent) when the load is off. Three sets of data are shown. The and the experiment duration is in the range of hours, the pulses
experimental data are indicated by open circles. Simulation re- are compressed into bands and the details of the pulses can not
sults obtained from the battery model described here are shown be viewed. As can be seen, over the entire discharge course, the
by crosses. Simulation results obtained from a simplified battery simulation result (shallow color band) and the experimental data
model that had all of the features of the current model except for (dark color band) agree well, though differences exist. To com-
capacitance are shown with the solid line. The complete battery pare the pulse details, the experiment was repeated using faster
model gave very good fidelity whereas the model that did not in- data acquisition speed. Results are presented in Figs. 2023.
clude the effective capacitance deviated significantly at the be- Fig. 20 shows the voltages for the pulse between 3004.602
ginning of each load pulse. s and 3007.152 s. The measured and computed pulse ampli-
GAO et al.: DYNAMIC LITHIUM-ION BATTERY MODEL 503
Fig. 20. Single pulse comparison for the time between 3004.602 s and
3007.152 s.
Fig. 22. Single pulse comparison for the time between 7029.602 s and
7032.152 s with R = 80 m
, R = 40 m
.
Fig. 21. Single pulse comparison for the time between 3004.602 s and
3007.152 s with R = 80 m
, R = 40 m
. Fig. 23. Single pulse comparison for the time between 12 879.602 s and
12 882.152 s with R = 80 m
, R = 40 m
.
tudes differ by about 40 mV during the 1.4 A discharge. Like
any other electronic components, the characteristics of batteries good agreement in internal resistive drop was obtained between
depend not only on their design, material composition and ge- the model and the experiment according to these comparisons.
ometry, but also on many other factors, such as defects, shelving The differences shown by the transient processes (Figs. 2123)
time, aging effect and service history. Therefore the character- indicate that the effective capacitance of the battery under test is
istics may differ from one battery to another even for the same smaller than that of the model and it actually varies as a function
production batch. The characterization in our model was based of the state of discharge. Indeed, both the double layer capaci-
on the manufacturers data typical to the same type of batteries, tance and the diffusion capacitance of the porous electrode are
but it may slightly differ among individual ones. Thus, the char- strong functions of the charge density and the thermodynamic
acterization shall be focused on the major features, which is ful- temperature [24]. As a result, the effective capacitance increases
filled by our model. Nonetheless, it is easy to make up the dif- during the course of a battery discharge. The variable capaci-
ference by simply adjusting the internal resistance. Fig. 21 is the tance effect is not included in the present version of the model,
comparison result after the total internal resistance in the model but it will be included in further model development. Fig. 24
was reduced to 120 m (with m , m ). As shows the temperature of the battery for the discharge profile of
a result, a better fitting for internal loss was obtained. CDMA2. The effective cooling coefficient was set to 6 W/m K
Using the new values for and , the simulation results (simulating natural convection cooling condition) and the initial
for the pulse between 7029.602 s and 7032.152 s, the pulse be- and ambient temperature was 296 K. As can be seen, the total
tween 12 879.602 s and 12 882.152 s were also compared to the temperature increase is about 1.5 degree. An experimental re-
experimental data, as shown by Figs. 22 and 23 respectively. A sult for the temperature was not available for comparison.
504 IEEE TRANSACTIONS ON COMPONENTS AND PACKAGING TECHNOLOGIES, VOL. 25, NO. 3, SEPTEMBER 2002
REFERENCES
[1] S. Megahed and W. Ebner, Lithium-ion battery for electronic applica-
tions, J. Power Sources, vol. 54, pp. 155162, 1995.
[2] C. Fellner and J. Newman, High-power batteries for use in hybrid ve-
hicles, J. Power Sources, vol. 85, pp. 229236, 2000.
[3] B. Kennedy, D. Patterson, and S. Camilleri, Use of lithium-ion batteries
in electric vehicles, J. Power Sources, vol. 90, pp. 156162, 2000.
[4] T. Horiba, K. Hironaka, T. Matsumura, T. Kai, M. Koseki, and Y. Mu-
ranaka, Manganese type lithium ion battery for pure and hybrid electric
vehicles, J. Power Sources, vol. 97-98, pp. 719721, 2001.
[5] G. Bruce, P. Mardikian, and L. Marcoux, 50 to 100 Ah lithium-ion cells
for aircraft and spacecraft applications, J. Power Sources, vol. 65, pp.
149153, 1997.
[6] T. Iwahori, I. Mitsuishi, S. Shiraga, N. Nakajima, H. Momose, Y.
Ozaki, S. Taniguchi, H. Awata, T. Ono, and K. Takeuchi, Development
of lithium ion and lithium polymer batteries for electric vehicle and
home-use load leveling system application, Electrochem. Acta, vol.
45, pp. 15091512, 2000.
[7] M. Majima, K. Hanafusa, Y. Oka, G. Tanaka, H. Yoshida, E. Yagasaki,
and T. Tada, Development of 1 kWh (300 Ah) class lithium-ion bat-
tery, J. Power Sources, vol. 68, pp. 448450, 1997.
Fig. 24. Battery temperature increase for CDMA2 test. [8] M. Majima, S. Ujiie, E. Yagasaki, K. Koyama, and S. Inazawa, De-
velopment of long life lithium ion battery for power storage, J. Power
Sources, vol. 101, pp. 5359, 2001.
[9] M. Doyle, T. T. Fuller, and J. Newman, Modeling of galvanostatic
V. CONCLUSION charge and discharge of the lithium/polymer/insertion cell, J. Elec-
trochem. Soc., vol. 140, p. 1526, 1993.
We have described here a complete behavioral model of a [10] T. F. Fuller, M. Doyle, and J. Newman, Simulation and optimization of
lithium-ion battery that is suitable for portable power system the dual lithium ion insertion cell, J. Electrochem. Soc., vol. 141, no. 1,
pp. 110, 1994.
studies. The model was formulated in a general sense, but was [11] M. Doyle and J. Newman, Comparison of modeling predictions with
coded specifically for use in the Virtual Test Bed computational experimental data from plastic lithium ion cells, J. Electrochem. Soc.,
environment. The method accounts for rate- and temperature- vol. 143, no. 6, pp. 18901903, 1996.
[12] P. Arora, M. Doyle, and R. E. White, Mathematical modeling of the
dependence of the capacity, thermal dependence of the equilib- lithium deposition overcharge reaction in lithium-ion batteries using
rium potential and transient response. A 4-step modeling pro- carbon-based negative electrodes, J. Electrochem. Soc., vol. 146, no.
cedure, based on use of manufacturers data, allows the model 10, pp. 35433553, 1999.
[13] L. Song and J. W. Evans, Electrochemical-thermal model of lithium
to have both good accuracy and the flexibility to represent dif- polymer batteries, J. Electrochem. Soc., vol. 147, no. 6, pp. 20862095,
ferent types of batteries. Construction and validation of a spe- 2000.
cific model of a Sony US18650 was described. Simulation re- [14] W. B. Gu and C. Y. Wang, Thermal-electrochemical modeling of bat-
tery systems, J. Electrochem. Soc., vol. 147, no. 8, pp. 29102922,
sults show that the model agrees well with experimental data in 2000.
both static and dynamic characteristics. [15] S. C. Hageman, Simple pspice models let you simulate common battery
The present model primarily uses simple representations for types, EDN, pp. 117132, Oct. 1993.
[16] V. H. Johnson, A. A. Pesaran, and T. Sack, Temperature-dependent
the potential loss and transient process modeling. This is not suf- battery models for high-power lithium-ion batteries, Battery Thermal
ficient in all cases. The model deviates from the experimental Manag. Assessment, 2002.
data at low temperatures and at high discharge rates. The in- [17] D. A. Calahan, Computer-Aided Network Design, Rev. Ed. New York:
McGraw-Hill, 1972.
ternal losses in these two cases are apparently the result of many [18] M. R. Lightner and S. W. Director, Computer-aided design of electronic
intricate processes. In addition to ohmic loss, the loss due to circuits, in Electronics Engineers Handbook, 3rd ed, D. G. Fink and
surface kinetics becomes important at low temperature regime; D. Christiansen, Eds. New York: McGraw-Hill, 1989, sec. 27.
[19] R. A. Dougal, C. W. Brice, R. O. Pettus, G. Cokkinides, and A. P. S.
while the resistances due to both the surface phenomena and Meliopoulos, Virtual prototyping of PCIM systemsThe Virtual Test
the concentration limitation are also significant in the high rate Bed, in Proc. PCIM/HFPC 98 Conf., Santa Clara, CA, Nov. 1998, pp.
regime. The effect of and perhaps the temperature, on 226234.
[20] (2000) Lithium-Ion Batteries: Individual Data Sheet CGR18650.
the transient process was also observed, which, according to Panasonic. [Online]. Available: http://www.panasonic.com/industrial/
the experimental data, shows an increasing effective capacitance battery/
during discharge. All these effects are not included in the present [21] (2001) Hard Carbon Lithium Ion Rechargeable Battery. Sony. [Online].
Available: http://www.sony.co.jp/en/Products/BAT/ION/Catalog-e.pdf
model, but they will be considered in the future model develop- [22] J. S. Newman, Electrochemical Systems, 2nd ed. Englewood Cliffs,
ment so that the model can be applied to studies of high power NJ: Prentice-Hall, 1991.
systems. In addition, the present model is validated by using [23] , Electrochemical Systems, 2nd ed. Englewood Cliffs, NJ: Pren-
tice-Hall, 1991.
mostly discharge data and its validity for representing charging [24] , Electrochemical Systems, 2nd ed. Englewood Cliffs, NJ: Pren-
processes is unknown. tice-Hall, 1991.
[25] B. E. Conway, Transition from Supercapacitor to Battery behavior
in electrochemical energy storage, J. Electrochem. Soc., vol. 138, no.
6, pp. 15391548, 1991.
ACKNOWLEDGMENT [26] D. Bernardi, E. Pawlikowski, and J. Newman, A general energy balance
for battery systems, J. Electrochem. Soc., vol. 132, no. 1, pp. 512,
1985.
The authors wish to thank C. Holland for conducting the [27] R. A. Dougal, S. Liu, L. Gao, and M. Blackwelder, Virtual Test Bed for
physical experiments to which the simulations were compared. advanced power sources, J. Power Sources, 2002.
GAO et al.: DYNAMIC LITHIUM-ION BATTERY MODEL 505
[28] G. Nagasubramanian, R. G. Jungst, and D. H. Doughty, Impedance, Shengyi Liu, (M96) received the B.Sc. and M.Sc.
power, energy and pulse performance characteristics of small commer- degrees from Tsinghua University, Beijing, China, in
cial Li-ion cells, J. Power Sources, vol. 83, pp. 193203, 1999. 1982 and 1985, respectively, and the Ph.D. degree in
[29] K. Kezuka, T. Hatazawa, and K. Nakajima, The status of sony Li-ion electrical engineering from University of South Car-
polymer battery, J. Power Sources, vol. 97-98, pp. 755757, 2001. olina, Columbia, in 1995.
[30] Y. Nishi, Lithium ion secondary batteries; past 10 years and the future, Prior to joining the Department of Electrical Engi-
J. Power Sources, vol. 100, pp. 101106, 2001. neering, University of South Carolina, as a Research
[31] R. A. Dougal, S. Liu, and R. E. White, Power and life extension of Professor, he was Senior Research and Development
battery/ultracapacitor hybrids, IEEE Trans. Comp. Packag. Technol., Engineer from 1995 to 1999 at InnerLogic, Inc.
vol. 25, pp. 120131, Mar. 2002. Research interests include application and control
[32] J. R. Miller, Battery-capacitor power source for digital communication study of advanced power sources and systems,
applications: Simulations using advanced electrochemical capacitors, modeling and simulation of interdisciplinary systems under virtual test bed
in Proc. Electrochem. Soc., vol. 95-29, Oct. 1995, pp. 246254. computational environment. His interests also include modeling, design and
[33] A. F. Burke and J. R. Miller, Test procedures for high energy den- applications of physical electronics-based devices, power semiconductor
sity, electrochemical capacitors, in Proc. Electrochem. Soc., vol. 95-29, devices, and converters.
1995, pp. 280297.
[34] (2002). [Online]. Available: http://www.arbin.com/Products/GSM_
CDMA/GSM.pdf.