WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

Study of Supercapacitor in the Application of Power Electronics

Yi cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang
School of Electronics and Information
Tongji University, Shanghai
4800, Cao’an Road, Shanghai
P.R. China
kzjc@263.net, weili1029@gmail.com, sxj999000@tongji.edu.cn, hugoliang@sohu.com

Abstract: - The purpose of this paper is to carry out a comprehensive study of supercapacitor in the application
of power electronics. According to the practical demand and feasibility of power electronics, the modelling,
voltage management and topology of supercapacitor-based power system have been discussed. Further more, in
order to study the terminal behaviour of supercapacitor and voltage balancing strategy for the application of
power electronics, a test bench based on HP-VEE has been built. The measurements on supercapacitor’s
capacitance, ESR (Equivalent serial resistance) and the consistency of a group of serial supercapacitors,
concerning practical application have been suggested. As an example, Maxwell’s BCAP0120 supercapacitors
have been selected for a 1.5kW hybrid supercapacitor- fuel cell power system. Tests have been carried out and
the results show that the ESR is higher and the capacitance is lower compared with the values supplied by
manufacturers; and the consistency of them is good, thus the voltage initialization strategy can be used for
voltage balancing.

Key-Words: - Supercapacitor, Power Electronics, Voltage-balancing, Capacitance, ESR

1 Introduction cells is necessary to obtain higher voltage. However,

Supercapacitor is a kind of electrical energy the unequal distribution of cell voltage will affect the
storage device. The advantages of supercapacitor are performance and lifetime of the cell. Reference [3]
high power density, high efficiency, fast charging has recommended 37several voltage balancing
and discharging speed, long cycle life, wide strategies. Another way to overcome the problem is
operating temperature range and environment so-called Voltage initialization described by
friendly. It has become an ideal option for high- Okamura [4]. It has yet to be decided that the
power applications, such as hybrid power systems, consistency of supercapacitors in order to choose a
regenerative energy systems and instantaneous back- proper voltage balancing strategy.
up power source [1-2]. The paper has carried out a comprehensive study
When designing a supercapacitor-based power of supercapacitor from the view of power
system, building a proper model for supercapacitor, electronics. Methodologies on modelling, voltage
taking the appropriate methodology of voltage management and topology of supercapacitor-based
management, choosing a right topology of power power system have been discussed. A test bench
system, and knowing the dynamic terminal based on HP-VEE has been built. Experiments on
behaviour are very important factors for the deciding the dynamic behaviour of supercapacitor
performance of power system. Power electronics and studying the consistency of supercapacitors for
device, such as DC/DC converter, is an power electronics application have been carried out.
indispensable part for the power system.
In power electronics applications, we concern
more about the dynamic parameters of 2 Supercapacitor Description
supercapacitors for they are often used for high duty 2.1 Principle of EDLC
cycle applications. Regarding that parameters on Below the decomposition voltage, while the current
data sheet supplied by manufacturers are the static does not flow, an electric double layer occurs at the
value. Hence, a supercapacitor testing method is boundary of electrode and electrolyte. The electrons
needed in order to test the dynamic characteristic are charged across the double layer and form a
parameters. capacitor. Energy is stored in the double-layer
Due to the low cell voltage of supercapacitor capacitor as charge separation in the double-layer
(0.9~3.3V), a series connection of supercapacitor

ISSN: 1109-2734 508 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

formed at the interface between the solid electrode characterized by making pores in the carbon with the
material surface and the liquid electrolyte in the ions in the electrolyte solution which has an energy
micro pores of the electrodes. Figure 1 shows the density of 20~60Wh/kg, while EEStor claims their
structure of electric double layer capacitor. examples will offer capacities on the order of 200 to
300Wh/kg. For comparison, a conventional lead-
acid battery is typically 30~40Wh/kg, modern
lithium-ion batteries are about 150~200Wh/kg.
The power density of supercapacitors in
commercial use is around 4~7kW/kg, which is much
higher than other energy-storage devices, for
example 1.8kW/kg for lithium-ion batteries. High
power density combined with long life cycle makes
supercapacitors ideal devices for peak power
application.
Maxwell (USA), NESSCap (koren), Panasonic
(Japan), Epcos (Germany), ECOND (Russia), and
NEC (Japan), Ao-wei (China), Shuangdeng (China)
are main manufactures of supercapacitor. Saft
Fig. 1 Principle of a double layers capacitor (France), Batscap (France) Superfarad (Sweden),
Okamura lab (Japan) have been doing research on
The electrodes are fabricated by high surface area, supercapacitors.
porous material having pores of diameter in the Manufacturers, such as Maxwell, NessCap,
nanometre range. Charge is stored in the micro pores Epcos, Econd, offer to sell supercapacitor module,
at or near the interface between the solid electrode which is a package of a group of supercapacitors in
material and the electrolyte. Double-layer capacitor series and contains supercapacitor management
electrodes have been fabricated using carbon black equipment, giving much more feasibility in
and carbon aero gel and carbon cloth. application.
There are two kinds of electrolyte: aqueous and
organic electrolyte. Aqueous electrolyte, such as
potassium hydroxide or sulfuric acid, has a much
smaller resistance and thus a larger power density is 3 Equivalent model for power
easy to be obtained; Organic electrolyte based on electronics application
propylene carbonate or acetonitrile whose ions are In the application of power electronics, an equivalent
bigger, usually has higher resistance and smaller model which reflects the terminal behaviour of
power density (but higher energy density due to the supercapacitors is desired in simulation with the
voltage used). Up to now, capacitor rated voltage purpose of further studying the characteristics of
with an aqueous electrolyte is about 0.9V per cell supercapacitor based power system.
and with organic electrolyte is 2.3 to 3.3 V for each
cell. When developing supercapacitors, the electrode
material and electrolyte characteristics should be 3.1 RC serial model
considered jointly and not separately. The simplest equivalent model for EDLC is the RC
The capacitance is dependent primarily on the serial model as shown in Figure 2 [7]. This method
characteristics of the electrode material (surface area assumes ideal behaviour of supercapacitor and it
and pore size distribution). The resistance of the neither reflects physical aspects, nor the influence of
supercapacitor cell is strongly dependent on the voltage or temperature on supercapacitors. However,
resistivity of the electrolyte used and size of the ions this model can be applied in the low required
from the electrolyte that diffuse into and out of the precision.
pores of the micro porous electrode particles [5].

2.2 Recent development of supercapacitor Fig.2 RC Equivalent model

In terms of energy density, existing commercial
electric double-layer capacitors range from around
0.5 to 10Wh/kg. Especially, the Okamura lab has 3.2 Three branch model
developed the Nanogate capacitor[6] which is

ISSN: 1109-2734 509 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

Another model shown as Figure 3 which is based on engineering is proposed by reference [8].
the physical aspects and the desire of practical

Fig.3 Three branch model

Fig.4 Time domain model

The model has three well distinct RC time distribution shows self-discharge. It is possible to
constants covering the desired time range. Each of determine the parameters of the model using
the three branches has a distinct time constant measurements at the DLC terminals which is very
differing from the others in more than an order of practical for engineering.
magnitude which will result in an easily measurable This model shows good agreement with
model. experimental data and is possible for gaining the
The first branch, with the elements Ri,Ci0, and the parameters. Reference [9] has added temperature
voltage-dependent capacitor Ci1(in F/V), dominates influences to this model and making it an efficient
the immediate behaviour of the DLC in the time way to study its application in power electronic
range of seconds in response to a charge action. The circuits and automotive applications.
second branch, with parameters Rd and Cd,
dominates the terminal behaviour in the range of
minutes. Finally, the third branch, with parameters 3.3 Time domain model
Rl and Cl, determines the behaviour for times longer A model shown as Figure 4 which takes into
than 10 minutes. account frequency, voltage and temperature
The equivalent circuit model reflects the physics dependencies of capacitance, series resistance,
of the double-layer charge distribution. First, the redistribution of electrical charges on the electrode
resistive element represents the resistivity of carbon surface and leakage current is proposed by [10]. It is
particles. The capacitive element represents the also on the base of second model.
capacitance between carbon and electrolyte. Second, “Circuit 1” takes into account the electrolyte ionic
the capacitance of the double-layer charge resistance temperature dependence in the low
distribution depends on the potential difference frequency range. The parallel capacitance Ci has
across the material, and according to measurements, been used to cancel the contribution of Ri(T) in high
in the practical voltage range of the device, the DLC frequency range. For low frequency, the circuit 1
capacitance varies linearly with the capacitor behavior is close to that of resistance Ri(T). The
terminal voltage. Third, double layer charge

ISSN: 1109-2734 510 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

relationship between Ri and the temperature can be includes a parallel RL resistance, which gives the
established from experimental results by using EIS. long time leakage current contribution.
“Circuit 2”is introduced to increase the value of It depends on the practical application to decide a
capacitance of the average frequencies. Their proper model which meets the requirement.
behaviour is the one of a phase shifter. Generally speaking, the RC serial model can be
“Circuit 3” describes the leakage current and the used in a system with low precision; the three
internal charge redistribution. The self discharge branch model is fit for the application which has
behaviour of supercapacitors is an important factor requirement on dynamic characteristics; while the
because it determines the duration time of stored time domain model is often used for precise study. It
energy on open circuit. The supercapacitor self- is suggested that the three branch model is the most
discharge is also a function of temperature. It is appropriate choice for power electronics
necessary to use two different time constant circuits applications.
RC by elements Rp1Cp1, Rp2Cp2 which depend on the
voltage and on the operating temperature. It also

Fig.5 Passive Balancing

Fig.6 Active Balancing

4 Management of supercapacitor active balancing. Another way to overcome the

problem caused by unbalance of voltage is so-called
power systems voltage initialization described by Okamura[4].
When connecting many capacitors in series, the
issue of voltage balancing inevitably comes into
4.1 Passive balancing
A passive balancing system is designed to
play [11]. Basically there are two reasons for an
overwhelm the inherent variations in leakage current
imbalance of voltages in a serial string of
by installing a resistor in parallel with each other.
supercapacitors: (i) deviations from the nominal
The resistor is typically sized at 10 times the
capacitance of the capacitors and (ii) deviations in
average leakage current of the cell. The benefits to
self discharge performance. While the first topic is
this balancing method are simplicity and low cost.
mainly important during dynamic performance of
The drawback of this technique is slow response due
the capacitor string, the latter topic dominates for
to the linearity of leakage current with voltage and
static capacitor performance during constant voltage
high parasitic losses due to the 10-time additional
phases. A cell management circuit maximizes the
leakage current. Passive balancing is mainly used in
performance and life of supercapacitors installed in
low-duty cycle applications such as in backup
series [12-14].
power systems. Fig 5 shows a simple balancing
Generally speaking, there are two ways in voltage
network with resistors.
balancing, one is passive balancing, and the other is

ISSN: 1109-2734 511 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

4.2 Active balancing power system [16-18]. The non-isolated Buck-Boost

In contrast to passive solutions, an active balancing bidirectional DC/DC converter shown as Figure 8 is
circuit behaves nonlinearly and works to force the the first choice for that. The advantage of this
cells to have an equal voltage, resulting in the most topology is high efficiency, high reliability, low
effective use of the supercapacitor string. Fig.6 power loss, less expensive and small in size.
shows a simplified diagram of an active balancing
circuit incorporating a comparator. In this
configuration, each circuit stretches across two cells,
comparing their voltage and moving charge to
equalize the two cells. A number of schemes are
used to achieve active balancing and many are
patented. Active balancing circuit is required in high
duty-cycle applications and where low parasitic
losses are necessary.

4.3 Voltage Initialization

The principle of voltage initialization is that all Fig.8 Supercapacitor energy storage system
capacitors are balanced at the upper voltage limit of topology
the capacitor module. As a consequence, when the This bidirectional Buck-Boost DC/DC converter
module is discharged, the individual capacitors will allows the power transfer in both directions. This
adopt different voltages on a lower level. When feature enables the process of charging and
recharged to the upper voltage, all the capacitors discharging through one unit. The current from the
will be balanced again. Provided that the supercapacitor is fully controlled by bidirectional
capacitances of individual capacitors change slowly DC/DC converter, and the voltage of DC link is
with time, an occasional initialization of the module dependent on the control result of it. When charging
will keep the capacitors balanced at the upper the supercapacitor bank, the DC/DC converter
working voltage. This is shown by Figure7. works in Buck mode, and supplies a constant charge
current. The power flows from DC-link to
supercapacitor bank. When discharging, the DC/DC
converter works in Boost mode, and keeps the
voltage of DC-link constant. The power flows from
supercapacitor bank to DC-link.

6 Supercapacitor measurement
Regarding that parameters on data sheet supplied by
manufacturers are the static value. In power
electronics applications, we concern more about the
Fig.7 Principle of voltage initialization dynamic parameters of supercapacitors for they are
This method is much more feasible and less often used for high duty cycle applications. Hence, a
expensive. Reference [15] has implemented it with supercapacitor testing method is needed in order to
an occasional initialization of all capacitors, which test the dynamic characteristic parameters.
has achieved a nice result. However, it demands In this section, a test bench based on HP-VEE has
sound consistency and high quality of been built firstly. Then, methods on testing
supercapacitor. supercapacitor’s capacitance, ESR (Equivalent
serial resistance) and the consistency of a group of
serial supercapacitors, concerning practical
5 Topology of supercapacitor energy application have been suggested. As an example,
storage system Maxwell’s BCAP0120 supercapacitors have been
The terminal voltage of supercapacitor changes tested for a project which was to build up a hybrid
significantly when being charged or discharged. For supercapacitor-fuel cell power system of 1.5kW.
this reason, a power electronics device —DC/DC
converter is necessary to form a supercapacitor

ISSN: 1109-2734 512 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

6.1Review of testing methods for 6.3 Terminal behavior tests

supercapacitors 6.3.1 Scheme design for measuring Capacitance
Currently, electrochemical impedance spectroscopy and ESR
(EIS) and constant current charge and discharge are The capacitance of supercapacitor is voltage-
two main ways in research of supercapacitors. EIS depended, and it is also affected by discharging
is used to characterize electrode material for current when the initial voltage is the same.
supercapacitors in frequency domain, and Generally speaking, the step of measurement is to
professional electrochemical equipment is required keep the supercapacitor at rated voltage for 30~60
for doing this test [19-21]. However, for power minutes, and then discharge it by a constant current
electronics applications, we concern more about the I, and the value of capacitance and ESR is
terminal behaviour. Hence, constant current charge calculated by the following formulas:
and discharge is used for this study. C = I ⋅ ∆t / ∆U (1)
ESR = ∆U / I (2)
Among which,
6.2 Test bench based on HP-VEE C－ the capacitance at rated voltage；
The test bench is based on HP-VEE, which is a
graphical programming language optimized for I－the constant discharge current；
designing test and measurement applications, and U－the absolute value of voltage variation during
programs with operator interfaces. Through HP-
VEE, we can design a test procedure; communicate t；
with test equipment over a general purpose interface
bus (GPIB), and record data of measurements. Table V－the voltage leap occurred at the instant when a
1 shows the main equipment for building this test discharge current is applied or taken out.
bench. According to technique guide of manufacturers,
Table 1: Equipment for building the test bench. Necsscap chooses a U=(0.7~0.3)Urated with
Equipment Parameters discharge current at 1mA/F[21]; EUCAR—an
Programmable Current 0~40A, European association of vehicle manufactures—
DC electronic Load: voltage 0~80V, choose a U=(0.6~0.4) Urated with discharge current
Chroma 6310 resolution 1mA/10mV at 5mA/F[22]; Maxwell choose a U=(1~0.5) Urated
Programmable Current 0~40A, with discharge current at 1mA/F[23].
Power Source : voltage 0~80V, The above scheme of measurement is to test the
ITech 6121 resolution 1mA/10mV static characteristics of supercapacitor. The
Agilent34970A Data Acquisition Unit, procedure of keeping voltage constant and
61/2-digit multimeter discharging by a small current has given enough
accuracy time for the ions from the electrolyte to diffuse into
Measurements have been done on a Maxwell and out of the pores of the micro porous. However,
Technologies BCAP0120 supercapacitor, with a in practical power electronics applications,
nominal capacitance of 120F, ESR of 5mhom, and a supercapacitors are often used for supplying high
rated voltage of 2.5V. Figure 9 shows the test bench. power, which means a very large current, often
hundreds of ampere , will pass though it. Thus,
there is not enough time for ions’ diffusing, which
will lead to a smaller dynamic capacitance. This
dynamic capacitance will affect the performance of
the power system.

Fig9. Test bench for terminal behavior

ISSN: 1109-2734 513 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

Test I: balance each cell at 2.45V, charge the

capacitor group with Icharge that equals 2A, discharge
them with Idischarge that equals 4A, and last for 7
hours.
Test II: change the current, with Icharge that equals
4A, Idischarg that equals 10A, and last for 7 hours.

6.4 Test Results and Analysis

Based on this test bench, measurements were carried
out for a project which was to build up a hybrid
Fig10. Procedure of Measurement for supercapacitor-fuel cell power system of 1.5kW.
Capacitance and ESR Supercapacitor BCAP120 of Maxwell was chosen.
From the above discussion, the scheme of The following is the test result.
measurement for supercapacitor considering 6.4.1 Capacitance and ESR
practical operation situation such as duty cycle, and The application of the power system requires a rated
charging current is suggested by the author. A current Iw of 20A, and rated voltage of 2.3 V for a
testing procedure, which is shown by figure 10, is single cell. According to former principle, three
proposed as follows: supercapacitors were carried out. Table 2 shows the
1st. From 0 to t0, discharge supercap to Vmin with test result. From the above result that the
current Iw. capacitance is lower and the ESR is larger than that
2nd. From t0 to t1, set I=0; of the value on the datasheet.
3rd. From t1 to t2, charge supercap to Vmax at Iw; Table 2. Test result of capaciatance and ESR
4th. From t2 to t3, set I=0 for a period, record the BCAP0120/num 1 2 3
voltage at t3; Capacitance/F 110 115 113
5th. From t3 to t4, discharge supercap to Vmin at ESR/mohm 5.2 5.3 5.5
constant current Iw;
6th. From t4 to t5, set I =0 for a period, record the
voltage at t5; 6.4.2 Cycling test result
7th. Repeat this cycle for 10 times and calculate the
average value.
Calculation
C = I w ⋅ (t 4 − t 3 ) / Vd (3)
ESRDC = (Vt 5 − Vt 4 ) / I w (4)
Vd is the Vt3 minus Vt5.
Iw is the rated current of supercapacitor in practical
application. Vmax is the rated voltage of
supercapacitor in practical application. The cycle of
charging and discharging is the rated duty cycle in
practical application.

6.3.2 Consistency Test for a group of

supercapacitors
The purpose for this test is to check the consistency
of the selected supercapacitors in order to choose a
proper voltage-balancing strategy.
Cycling test was carried out to test the consistency
of supercapacitors. Three BCAP0120
supercapacitors in serial connection were tested.
This test is to check the voltage change of each cell
after voltage initialization in long term cycle. The
test consists of two parts:

ISSN: 1109-2734 514 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

0.2 0.20
(V1-V2)
0.15 (V1-V2)
0.1
0.10
∆volt(V)

∆volt(V)
0.05
0.0
0.00
-0.1 -0.05

Time(Hour) -0.10
-0.2
0 1 2 3 4 5 6 7 -0.15
Time(Hour)
0.2 -0.20
(V1-V3) 0 1 2 3 4 5 6 7

0.1 0.20 (V1-V3)

∆volt(V)

0.15
0.0 0.10

∆volt(V)
0.05
Time(Hour)
-0.1
0 1 2 3 4 5 6 7 0.00

0.2 -0.05
(V2-V3)
-0.10 Time(Hour)
0.1
∆volt(V)

0 1 2 3 4 5 6 7

0.0 0.20
(V2-V3)
0.15
-0.1
0.10
Time(Hour)
∆volt(V)

-0.2 0.05
0 1 2 3 4 5 6 7
0.00

-0.05
Fig.11 Voltage deviation among cells in Test I
-0.10
-0.15
Fig.11 shows the voltage variation of the cell in test Time(Hour)
-0.20
I (V1, V2, V3 is the terminal voltage of 0 1 2 3 4 5 6 7
supercapacitor ). During the whole cycle, the
Fig.13 Voltage deviation among cells in Test II
voltage deviation of each cell is less than 0.1V, and V1
doesn’t change with time. The largest voltage 2.5 V3
variation is between V1and V3 caused by the
Voltage(V)

2.0
difference of ESR. Fig.12 is the enlarged view of 1.5
V1, V3 and (V1-V3). We can see that, voltage
1.0
variation is bigger when charging and discharging, Time(Hour)
0.5
and it is smaller when cell voltage reaches the 6.55 6.60 6.65
initialized value. 0.2 ( V1-V3)
Voltage(V)

2.5 V1
0.1
Voltage(V)

2.0 V3
0.0
1.5 -0.1 Time(Hour)
1.0 6.55 6.60 6.65
Time(Hour)
0.5
6.55 6.60 6.65 Fig.14 Enlarged view of V1, V3 and (V1-V3) in Test II
0.10
Voltage(V)

( V1-V3)
0.05 Fig.13 shows the voltage deviation of the cell in test
0.00 II, and the maximum voltage deviation is less than
-0.05 Time(Hour) 0.2V, which is also a constant with time. The largest
6.55 6.60 6.65 voltage variation is between V1and V3. Fig.14 is the
enlarged view of V1, V3 and (V1-V3). The voltage
Fig.12 Enlarged view of V1, V3 and (V1-V3) in Test I variation is bigger when charging and discharging,
and smaller when cell voltage reaches the initialized
value. However, voltage deviation of test II is larger

ISSN: 1109-2734 515 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

than that of test I, resulted of higher discharging [3] Dirk Linzen, Stephan Buller, Analysis and
current. evaluation of charge-balancing circuits on
From the result, we can conclude that the voltage performance, reliability, and lifetime of
variation among cells after voltage initialization has supercapacitor systems, IEEE Transactions on
relationship with charge and discharge current and industry applications, Vol.41, No.5, 2005,
the initialized voltage. There is no apparently pp.1135-1141
voltage change in long time cycle, which means the [4] M. Okamura, ECaSS: System Improvements &
BCAP0120 supercapacitors show good consistency Discussion, proceedings of the 13th
during cycling. Hence, an occasional voltage International Seminar on Double Layer
initialization of all capacitors can be adopted for Capacitors and Hybrid Energy Storage
voltage management. Devices, 2003, pp.1-12
[5] H. Gualour, D. Bouquain, Experimental study
of supercapacitor serial resistance and
7 Conclusion capacitance variations with temperature.
The paper focuses on the study of supercapacitors Journal of power sources, Vol.123, 2003 pp.86-
for power electronics applications. Methodologies 93
on modeling, voltage management and topology of [6] M.Okamura, K.Mitsui, Production status of
supercapacitor-based power system have been nanogate capacitors and integrated electronics-
discussed. A test bench has been built, and the Part I, The 14th international seminar on
scheme of measurement for supercapacitor’s double layer capacitors and similar energy
characteristic parameters considering practical storage devices, 2004, pp.1-8
operation situation such as duty cycle, and [7] R.L Spyker, Discharge characteristics of high
discharging current is proposed. Cycling test for energy storage double layer capacitors, IECEC-
studying the consistency of supercapacitor has also 97, Proceedings of the 32nd Intersociety, Vol 1,
been carried out. According to the test result, 1997, pp.292-296
Maxwell BCAP0120 supercapacitor has shown [8] Luis Zubieta, Richard Bonert, Characterization
good consistency, thus voltage initialization strategy of double-layer capacitors for power electronics
can be applied in its management. applications, IEEE transaction on industry
applications, vol.36, 2000, pp.199-205
[9] H.Gualour, D.Bouquain, Experimental study of
supercapacitor serial resistance and capacitance
variations with temperature. Journal of power
Acknowledgement sources ,Vol.123, 2003, pp86-93
This paper is supported by Chinese National Natural [10] F.Rafik, H.Gualous, Frequency, thermal and
Science Foundation (No.50877054). voltage supercapacitor characterization and
The authors are grateful to the center for energy and modeling, Journal of power sources,Vol.165,
process (CEP) of ecole des mines de Paris (EMP) in 2007, pp 928-934
Sophia-Antipolis, France. The author would like to [11] N.Rizoug, P.Bartholomeüs, B.Vulturescu,
thank P. Achard, R. Metkemeijer, S. Berthon-Fabry Voltage sharing in supercapacitor
and P. Leroux for valuable discussions and help modules:experimental study, 35th Annual
during the study in France. IEEE power electronics specialists conference,
2004 pp.690-696
[12] D.linzen,S.Buller,E,Karden, Analysis and
Reference： evaluation of charge-balancing circuits on
performance, reliability, and lifetime of
[1] P. Thounthong, S. Raël, B. Davat, Fuel cell and supercapacitor systems, IEEE Transactions on
supercapacitors for automotive Hybrid industry applications, Vol.41, No.5, 2005,
Electrical system, ECTI Transactions on pp.1135-1141
Electrical Eng., Electronics, and [13] R.Kötz, J-C.Sauter, Voltage balancing of a 250
Communications, vol.3, No.1, 2005, pp.20-30 Volt supercapacitor module for a hybrid fuel
[2] A. Rufer, P. Barrade, A supercapacitor-based cell vehicle, International seminar on double
energy-storage system for elevators with soft layer capacitors, 2007, pp.79-86
commutated interface, IEEE Transactions on [14] N.Rizoug, P.Bartholomeüs, Voltage sharing in
Industry Applications, Vol.38, 2002, pp.1151- supercapacitor modules: experimental study,
1159

ISSN: 1109-2734 516 Issue 6, Volume 8, June 2009

 WSEAS TRANSACTIONS on CIRCUITS and SYSTEMS Yi Cheng Zhang, Li Wei, Xiaojun Shen, Haiquan Liang

35th Annual IEEE power electronics specialists

conference, 2004, pp.690-696
[15] R.Kötz, J-C.Sauter, Voltage balancing of a 250
Volt supercapacitor module for a hybrid fuel
cell vehicle, International seminar on double
layer capacitors, 2007, pp.79-86
[16] H.Y.Zhang, T.Z. Wei, Study on ultracapacitor
energy storage. Power system technology,
Vol.30, 2006, pp.92-95
[17] Rufer A, Hotelher D, Barrade P. A
supercapacitor-based energy storage substation
for voltage compensation in weak
transportation networks. Delivery IEEE
Transactions, Vol.19, 2004, pp.629-636
[18] Sng E K K, Choi S Skian, Analysis of series
compensation and DC-link Voltage controls of
transformer self-discharging dynamic voltage
restorer. IEEE Transactions on Power Delivery,
Vol.19, 2004, pp.1511-1518
[19] F. Rafik, H.Gualous, Frequency, thermal and
voltage supercapacitor characterization and
modeling, Journal of power sources, Vol.165,
2007, pp.928-934
[20] Eckhard Karden, Stephan Buller, A frequency
domain approach to dynamical modeling of
electrochemical power sources, Electrochimica
Acta, Vol.47, 2002, pp.2347-2356
[21] Nesscap ultracapacitor technical guide,
Nesscap Co., Ltd. 2008
[22] P. Kurzweil, B. Frenzel, Capacitance
characterization methods and ageing behaviout
of supercapacitors, The 15th International
seminar on double layer capacitors, 2005
[23] Ultracapacitor product guide, Maxwell Co.,
Ltd. 2008

ISSN: 1109-2734 517 Issue 6, Volume 8, June 2009