supercapacitor enhanced battery traction systems – … from the traction system working as a...

The World Electric Vehicle Journal, Vol 2, Issue 2

Supercapacitor Enhanced Battery Traction Systems – Concept Evaluation© 2008 WEV Journal

32

Supercapacitor Enhanced Battery Traction Systems – Concept Evaluation

Frederik Van Mulders*, Jean-Marc Timmermans**, Zach McCaffrey**, Joeri Van Mierlo**, and Peter Van den Bossche**

* Erasmushogeschool Brussel, IWT, Nijverheidskaai 170, B-1070 AnderlechtPhone: 32-2-629-33-96Fax: 32-2-629-36-20 Email: [email protected]

** Vrije Universiteit Brussel, IR-ETEC, Pleinlaan 2, B-1050 ElsenePhone: 32-2-629-28-04Fax: 32-2-629-36-20

This review paper gives an overview of state-of-the-art technology regarding supercapacitors, briefly discusses the aspect of supercapacitor balancing and compares and assesses two system designs incorporating supercapacitors, batteries and in one case with an interconnected converter. Supercapacitors offer a high power density and long life cycle and could improve a battery-only setup subjected to peak loads. An electric kart test setup will be used to evaluate a simple direct parallel connection between the batteries and the supercapacitor bank. Measurements on a more advanced setup, incorporating an interconnected converter between the batteries and supercapacitors, will illustrate the effect of such an implementation compared to a plain parallel connected setup.

Keywords: Supercapacitors, Batteries, Converter, Peak Power, Power Management

1. INTRODUCTION

A typical traction system (shown in Figure 1) consists of two main components, an energy reservoir (e.g. a battery or generator) and an energy converter (e.g. an electric motor). In many cases traction systems are subjected to dynamic loads as the energy converter should be able to deliver the required energy at any time. If only one energy reservoir is available, optimal use becomes difficult. In order to respond to energy peak demands these energy reservoirs are often overdimensioned, resulting in suboptimal energy yield in regime conditions.

Energy reservoirs are defined by two key parameters, the energy density and the power density. Ideally an energy reservoir should offer both a high energy and a high power density. A typical energy reservoir, such as a battery, has a high energy density, but a limited power density, whereas new components, such as supercapacitors, offer low energy, but high power density [1]. Supercapacitors are electrochemical energy storage components that offer new possibilities in various applications. The question at hand is if the

combination of a supercapacitor and relatively low power energy reservoir can result in a better energy storage system.

In systems using a battery or a fuel cell as energy storage unit, the use of a supercapacitor peak power unit (PPU) could result in significant improvements [2, 3]. As traction batteries were designed for low power density applications, high load conditions result in large internal losses and a short longevity. Because a supercapacitor unit can be used to relieve the battery system from peak loads, the battery unit can be dimensioned and optimized for regime load conditions, resulting in better energy management and longer battery cycle-life.

The supercapacitor unit has an added bonus; not only can the supercapacitors supply high-power peak demands, they can also be charged very quickly. This energy can come from the available energy reservoir or from the traction system working as a generator. This energy saving regenerative feature can be implemented in many ways, such as recovery of braking energy in vehicles, elevators, etc.

2. SUPERCAPACITOR CHARACTERISTICS

The mentioned key characteristics of supercapacitors are illustrated in Figure 2 and Table 1, next to common battery types. Differences can be found in specific energy and power, efficiency, cycle life and cost. Consequently it would be convenient to create a combination with both a potential high power and energy density, as well as a longer cycle life.

ISSN 2032-6653Page 0120



33

Figure 1: A typical traction system

Sp

ecifi

c p

ower

(W

/kg)

Specific energy (Wh/kg)

Figure 2: Ragone chart comparing battery types and supercapacitors [4]

Table 1: Battery and Supercapacitor characteristics [5]

ISSN 2032-6653Page 0121



34

Key point when using supercapacitors in an energy system is the usable energy. The total energy stored in a capacitor, Etot, is a product of the capacitance, C, and the square of the voltage, V:

(1)In a practical energy-storage system that combines

batteries and supercapacitors, the supercapacitor’s total stored energy will almost never be fully used due to the fact that the converter necessary to accommodate such a wide potential difference range would be too complex for most applications.

Moreover, when discharging a supercapacitor to 50% of his nominal voltage, already 75% of the energy stored in the supercapacitor will have been released [eq. 1]. With the addition of the necessary voltage boundaries the usable energy is given by:

(2)Still, with the introduction of new higher power

density battery technologies, the question of the supercapacitors’ added benefit is an interesting one. The power density of these newer battery technologies is not on a par with that of a supercapacitor [Table 1], but some of the supercapacitors’ appropriated applications could possibly benefit from the new battery technologies. Reconsidering these applications could thus proof to be beneficial.

3. THE SUPERCAPACITORS’ LIFETIME

A supercapacitors’ lifetime can be as high as 10 years under normal operation, but due to an acceleration of continuous parasitic electrochemical reactions at increased temperatures and voltages, the capacitance decreases, while the internal resistance and self-discharge rate rise, lowering the lifetime of the

totE C V= ⋅ 212

max minE C V C V= ⋅ − ⋅2 21 12 2

supercapacitors [6].

Normally, manufacturer datasheets supply a chart displaying the influence of cell voltage and temperature on the degradation of the supercapacitors. An example [6, 7] results in a general applicable rough estimation for the lifetime that states that the lifetime is halved for each 100mV above nominal voltage and for each 10°C above the nominal temperature of 25°C.

These rough estimates can be expressed as an exponential function, with TL,N the life expectancy under nominal circumstances (nominal voltage VN and nominal temperature θN) and cV and cθ constants depending on the device (technology and manufacturing process), with VC the voltage over the capacitor (figure 3) and θ the temperature [8]:

(3)As deduced by Kötz et al. [9] (when considering the

leakage current proportional to the rate of the aging processes), the aging of the capacitor is accelerated by a factor cV of 1.5–2 for an increase of the maximal cell voltage by 100mV and an increase of the temperature of 10°C results in an accelerated aging factor cθ between 1.7 and 2.5. These values hold for a capacitor with a nominal voltage of 2.5V. However, they are still applicable for the last generation supercapacitors with a higher nominal voltage [8].

4. SUPERCAPACITOR CELL VOLTAGE BALANCING

Figure 3 shows a simplified equivalent circuit of a supercapacitor, this simple model will enable to address some of the reasons for the voltage imbalances. The value of the components is function of temperature, age and subject to manufacturing tolerances.

Aging increases the series resistance Rs and leakage

C N NV VmV K

L L,N vT T c cq q

q

− −− −

= ⋅ ⋅100 10

Rs Equivalent series resistanceRp Equivalent parallel resistanceC Supercapacitor capacitance Vcell Supercapacitor cell voltageVRs Voltage over RsVC Voltage over the capacitor (C) Icell Current through the cellIleak Leakage currentIC Displacement current

Figure 3: Simplified equivalent circuit of a supercapacitor

ISSN 2032-6653Page 0122



35

current Ileak, while the capacitance C diminishes. This process is not identical for each cell, so when various cells are connected in series, differences in cell voltage (caused by the differences in the equivalent series resistance) and state of charge (caused by the differences in self discharge rate due to the differences in the equivalent parallel resistance) surface. This effect emphasizes the need for a cell balancing system [8, 10].

Voltage balancing principles can be classified as passive or active. With passive voltage balancing the voltage difference of every cell is instantaneously balanced, which causes high power losses. Active voltage balancing normally is implemented by using a controller with an individual charging system.

There are two passive voltage balancing concepts which are most often used, namely fixed and differential voltage balancing. Fixed voltage balancing switches a resistor parallel to the supercapacitor cell on when a upper threshold voltage is reached and opens the circuit when a lower voltage threshold is reached. Differential voltage balancing also works with a parallel resistor, but will compare a cell’s voltage to that of its neighbors, discharging the cell over the parallel resistor when the difference becomes too big.

Bohlen et al. found a 180% increase in life expectancy for the weakest cell while maintaining a good efficiency when using a fixed voltage balancing simulation model. Differential voltage balancing does not protect the weakest cell from harmful voltage stress (it only balances the time average of the cell voltages). The simulation model thus showed a lower increase in life expectancy for the weakest cell (60%) together with a higher decrease of efficiency. However, because the operational voltage scope of a differential balancing system is much broader then that of a fixed voltage balancing system, it is able to balance cells at lower voltages and thus allows the system to operate before a critical cell condition is reached while also reducing the losses. With this in mind the differential voltage balancing system can even be more efficient.

Other, more elaborate, passive balancing systems can incorporate threshold voltage temperature compensation, eliminate the balancing systems activation during short voltage spikes at the end of charging in some applications and predictive balancing [8].

5. SYSTEM DESIGN AND INTEGRATION

When dimensioning and integrating a supercapacitor pack in a PPU, the final characteristics will depend

on the desired functionality and specifications. These will be formulated in terms of efficiency, power and energy specifications, the optimal energy management strategy, cost, lifetime and safety guaranties and will define a peak power system for the specific application. Designing a hybrid power system is thus often an iterative process adjusting the parameters, that eventually converges to a ‘best’ fit design.

There are several ways to integrate supercapacitors into an electric power system. Two of them will be compared in practical examples to illustrate the effect of one or both combinations (displayed in Figures 4 and 5). One topology, as displayed in Figure 4, can incorporate a DC/DC converter. This will allow to vary the voltage over the supercapacitor pack independently of the battery DC-busvoltage.

Figure 4: Propulsion system withDC/DC converter

Figure 5: Propulsion system withoutDC/DC converter

6. ELECTRIC KART BASED TEST SYSTEM

This first testing setup consists of a dismantled [11] electric kart with preservation of the electric power system, as displayed in Figures 5 and 6. The karts were initially dimensioned with only batteries in mind. The energy to propel the kart is supplied by four VRLA batteries [12] (each 12.8V and C5=55Ah). A current controlled DC/DC converter [13] is positioned between the batteries and the electric motor to regulate vehicle power. The controller also enables regenerative braking, so energy can flow back to the energy reservoir for recuperation.

ISSN 2032-6653Page 0123



36

Figure 6: The electric kart test setup

This system is subjected to two kinds of tests. First the effect of several mild and full throttle accelerations is evaluated for the system with and without supercapacitors. Second, the systems (with and without supercapacitors) are evaluated during two driving cycles. In this stage of the research only the first step will be executed. These preliminary acceleration tests give a good indication of the supercapacitor’s peak power absorption capabilities and its possible limitations.

The kart motor controller was configured to draw a maximum current of 240A from the energy supply. The tests where conducted with little mechanical load applied to the motor (overcoming only the inertia of the motor and load-machine). The controller regulates the current to the motor by potentiometer. For mild acceleration the potentiometer was connected to a 2.5V potential difference and for full throttle acceleration this potential difference was 6V. Logically, the higher this value, the higher the load on the energy supply.

The motor accelerates up to 2915 rpm at which point the motor controller limits the rotation speed (stabilisation). Data acquisition was done for the energy system’s potential difference and currents through the battery branch and supercapacitor branch.

The first measurements are depicted in Figure 7, these measurements are taken with a mild acceleration (top) and with full throttle acceleration (bottom) when using

only batteries. Figure 8 illustrates the measurements taken with a mild acceleration (top) and with full throttle acceleration (bottom) in case of batteries and supercapacitors connected in parallel. These results will give an indication of the supercapacitor’s added value and effectiveness when directly connected in parallel with the batteries.

Each figure can be divided in three subsections. In the first subsection, the motor is accelerated from 0 rpm to 2915 rpm and the current is gradually increasing. As can be seen in Figure 8, due to the voltage drop, the supercapacitor clearly aids during the acceleration process. In the second subsection, the motor rotation speed is held constant by the controller and the current drops and stabilises to a constant current value needed to maintain the rotation speed. Figure 8 illustrates that a voltage increase caused by the relaxing batteries results in a negative current flow through the supercapacitors, which recharges the supercapacitors. And finally in the third subsection, the motor speed and current return to zero due to the zero set point for the throttle. The battery relaxes further causing a further battery voltage increase, and current flow from the batteries to the supercapacitors.

6.1 Comparison

Table 2 shows the comparison of some key values during acceleration, namely the voltage drop and

ISSN 2032-6653Page 0124



37

Vol

tage

(V

), C

urr

ent

(A)

Time (s)

Figure 7: Batteries only mild (top) and full (bottom) throttle acceleration up to 2915rpm followed by stabilisation at 2915 rpms and no-throttle

Vol

tage

(V

), C

urr

ent

(A)

Time (s)

Table 2 : Key measurements comparison* Taken at the same time point as the cross point where the total current for the ‘Batteries and supercapacitor’ setup matches Ibatt,max

for the ‘Batteries only’ setup.

maximal drawn battery current difference. The added benefit of the supercapacitors is indicated by their relative (%) influence. In case of the voltage drop for instance, this figure indicates how much percent the voltage drop decreases by using supercapacitors. The effects on the other factors (maximum current and

acceleration time) are similar.

The results from Table 2 show a definite positive effect on the voltage drop and drawn battery current, thus battery load.

ISSN 2032-6653Page 0125



38

Figure 8: Batteries and supercapacitors mild (top) and full (bottom) throttle acceleration up to 2915rpm followed by stabilisation at 2915 rpms and no-throttle

Vol

tage

(V

), C

urr

ent

(A)

Time (s)

Vol

tage

(V

), C

urr

ent

(A)

Time (s)

The supercapacitor current delivery assistance (current delivered by the supercapacitors against the total delivered current) in function of the acceleration progress is illustrated by Figure 9. This figure provides clear evidence that the supercapacitor bank is delivering higher efficiency during high loads of the acceleration cycle (from zero to maximal speed).

As can be seen in Figure 10, only a relatively small percentage of the available energy in the supercapacitor pack is used. This is due to the small voltage drop. Using more of the energy stored in the supercapacitors is possible when increasing the voltage variance. This would imply the implementation of a DC-DC controlled supercapacitor system.

As indicated by Table 2 and Figure 9, the parallel connected supercapacitor-battery combination can substantially relieve the VRLA battery pack in peak

load situations.

7. HYBRID ELECTRIC VEHICLE TEST BENCH

The second test permits to load the energy system of a vehicle as if it was integrated in a real vehicle subjected to on-road tests. The functioning is based on a motor-generator coupling in which the load, delivered by the generator, can be regulated independently of the speed of the series motor [14, 15, Figure 11]. The system also permits recuperation of kinetic energy (regenerative braking).

The setup consists of eight 12V traction-batteries [16] with a capacity C20 of 50Ah each, connected in series. These batteries (about 4x12V in total) are the main energy supply. Besides the batteries there is also a supercapacitor pack. This pack consists of forty cells connected in series (1500F and 2.7Vdc each) [17],

ISSN 2032-6653Page 0126



39

Acceleration progress (%)

Su

per

cap

acit

oras

sist

ance

(%

)

Figure 9: Supercapacitor current delivery assistance in function of the acceleration progress

Su

per

cap

acit

oras

sist

ance

(%

)

Acceleration progress (%)

Figure 10: Percentile supercapacitor energy depletion in function of the acceleration process

Figure 11: Principle diagram of the test setup

ISSN 2032-6653Page 0127



40

with its total capacity amounting to 37.5F and a total maximum voltage of 108V.

The system is subjected to a simple test. Data acquisition makes use of a National Instruments system [15]. The tests purpose is to illustrate the influence of using a DC/DC-converter between the supercapacitors and the DC-bus, as opposed to a system using a direct connection. Both layouts are illustrated in Figures 4 and 5.

For this purpose an increased torque, lasting forty seconds, is applied during operation at constant velocity. Both energy-systems have to support the corresponding current pulse during forty seconds. Comparing both energy system responses should reveal the influence of the converter. Measured quantities are the current (A) through the batteries and the supercapacitors (in and out), the voltage (V) over the batteries and the supercapacitors, the torque (Nm) and the velocity (km/h).

7.1 Parallel Supercapacitor Pack Connection

Figure 12 illustrates the effect of directly coupling the supercapacitors to the batteries and shows how the supercapacitors assist in dealing with the edge of the rising pulse.

This behaviour is mainly determined by the voltage-drop over the batteries and supercapacitors. Because the limited voltage-drop stabilizes quickly, the energy contributed by the supercapacitors will also be limited,

Vol

tage

(V

), C

urr

ent

(A)

Time (s)

Figure 12: System characteristics without the use of a DC/DC converter

En

ergy

(W

h)

Time (s)

Figure 13: Cumulative delivered energy without the use of a DC/DC converter

as well in magnitude as in time [Eq. 1, Figures 12 and 13]. The current provided by the supercapacitors returns to zero when the bus-voltage stabilizes and as a result the batteries have to supply all power.

When the trailing edge follows the battery gets ‘unloaded’ and is able to ‘relax’, the bus-voltage goes up again (due to a lower voltage drop over the internal resistance of the batteries). As a result, the batteries will recharge the supercapacitors. The charge that was delivered by the supercapacitors after the leading edge of the pulse returns to the supercapacitors after the trailing edge [Figure 12, areas 1&2 and Figure 13].

7.2 With DC/DC-converter Interconnection

With indirect coupling of the supercapacitors and batteries by means of a DC/DC converter, which allows to vary the voltage over the supercapacitor pack independently of the battery DC-busvoltage, it is clearly noticeable that the supercapacitors not only catch sudden changes in current, but can also limit the current delivered by the batteries for a limited period of time (depending on the supercapacitor packs’ capacitance) [Figure 14].

Figure 14 clearly shows that the voltages over the supercapacitor pack (usc) and over the batteries (ubat) are different. Given their independence and equation 2, the exchangeable energy can be regulated by the DC/DC converter in the PPU. In addition Figure 14 shows how current and power delivered by the batteries are limited. The converter enables a dynamic

ISSN 2032-6653Page 0128



41

Vol

tage

(V

), C

urr

ent

(A)

Time (s)

Figure 14: System characteristics with the use of a DC/DC converter

En

ergy

(W

h)

Time (s)

Figure 15: Cumulative delivered energy with the use of a DC/DC converter

power distribution between both systems, while the system without converter and control strategy is fully dependent on the weakly dynamic and system typical battery voltage.

Summarized, Figure 14 shows that the illustrated PPU is not only able to smoothen out the current through the batteries, but is also able to limit it.

The usc drop in Figure 14 and the cumulative energy delivered by the PPU displayed in Figure 15, indicate that the supercapacitors are not fully recharged after the pulse. The converter in combination with an intelligent control strategy enables to discharge and recharge the supercapacitor pack at the most appropriate moment (e.g. recharging during braking). This is in shrill contrast to the non-intelligent system without an interconnected converter.

The internal resistance of a supercapacitor is generally a lot smaller than that of a battery. This implies that supercapacitors are more efficient (certainly at high currents). Limiting the battery current and storing recuperated energy directly in supercapacitors could thus imply an improvement in efficiency. However, the efficiency of the DC/DC converter should not be forgotten when analyzing the efficiency of the overall system.

8. LOAD SIMULATION FOR THE ELECTRIC KART SETUP

The next phase of our research involves implementing a dynamic road-load hardware simulation for the kart motor. The kart motor’s drive-cycle load will

be simulated by an electric generator. The electric generator will act as the load that incorporates forces due to acceleration, air resistance, rolling resistance, slope resistance [5]:

(4)and internal resistances such as friction and inertia

resistances with,

- FR resistance force (N)- α road incline (°)- v vehicle speed (km/h)- vw average wind speed(km/h)- ρ air density (kg/m3)- S frontal surface of the vehicle (m2)- Cx air resistance coefficient- fr frictional coefficient (dependant of tire pressure, road type, …)- M total weight (kg)- g gravitational constant (9.81 m/s2)

The acceleration force is given by:

(5)with m the mass of the object to be accelerated (kg).

8.1 Control of the Load Machine

To apply a speed profile corresponding to the use in a vehicle, the load machine will be voltage controlled. The load machine is a separately excited DC machine

wR x r

v vF . .S.C M.g.f .cos( ) M.g.sin( )

.r a a

+ = + +

212 3 6

aF m a= ⋅

ISSN 2032-6653Page 0129

force Fa (eq. 5). The total resistive force FR acting on a vehicle can be expressed as the sum of the rolling resistance force Fr, the aerodynamic force Fv and the climbing force Fc. The total resistive torque TR can be found by multiplying FR by the radius of the wheel R:

(7)The acceleration torque (Ta) is calculated based on the

vehicle weight, wheel radius (R) and the acceleration (av). The mass of the vehicle (M) is increased with a coefficient mf to take into account the inertia of the rotating parts of the drive train (consisting of the motor and the load machine, but no differential):

(8)The torque that should be developed by the load

machine (Tload) to simulate the vehicle driving on the road can now be calculated out of total torque (Ttot), but should be reduced with a term corresponding to the torque required to accelerate or decelerate the inertia of rotating masses of the test bench:

(9)With J the inertia of the rotating masses of the test

bench, seen from the side of the motor of the electric drive train. k is the transmission ratio of the coupling between the motor and the load machine and ω being the angular acceleration of the shaft of the motor of the electric drive train.

The torque to be applied by the load machine (not to be confused with the torque to be applied to the wheels) can be expressed in function of the measured rotational speed of the load machine ωload and can be simplified to:

(10)Out of the value of the torque of the load machine Tload,

the armature current Ia can be calculated. Figure 17 illustrates a possible control strategy.

8.3 Driving Cycles

The system is able to simulate a drive cycle consisting of several accelerations and braking events (for example a heavy duty cycle), both with and without the use of supercapacitors. Two cycles will be used as a comparison basis. The first cycle is based on the US06



42

(Figure 16). The variable voltage source is connected to the armature of the load machine and can regulate the speed of the load machine (Eq. 6).

(6)

a a aload

U R IC

w− ⋅

=⋅Φ

( )R R r v cT F R R F F F= ⋅ = ⋅ + +

( )a f vT M m .R.a= +

The excitation of the load machine is constant, so Ф can be considered as a constant and the speed of the load machine is depending on the armature voltage value and armature current. An external reference signal is generated by the real time controller and applied to the DC power supply. The measurement of the rotational speed and of the torque of the shaft of the load machine (ωload and Tload) will be used to calculate the value of the required reference signal when the motor of the vehicle drive train is following a predefined speed profile. Different road conditions and trip or vehicle parameters can be easily adapted through the reprogrammable real time controller. This gives the test bench a large flexibility. Predefined load conditions can be repeated multiple times and allow using the test rig to evaluate different hybrid electric power systems and components. Different power flow strategies can be evaluated and different modes of operation can be considered in a repetitive way.

8.2 Simulation of the Vehicle and the Road Conditions

The electric motor of the vehicle’s drive train can now be loaded in accordance to the speed profile of the vehicle and to the road conditions. The load torque of the electric motor will depend on the total resistive force FR acting on the vehicle and on the acceleration

( )load tot mot a R motT k.T J. k. T T J.w w= − = + −

load load loadT A B. C.w w= + +2

Figure 16: Drive trains’ motor and the load machine

ISSN 2032-6653Page 0130



43

Figure 17: Possible load simulation control strategy for the electric kart setup

Supplemental Federal Test Procedure (SFTP) [18], which represents aggressive, high speed and/or high acceleration driving behaviour, rapid speed fluctuations, and driving behaviour following start-up. The cycle was rescaled to fit the karts speed profile. It is not a typical kart driving cycle, but will supply information on the systems response during heavy duty.

The rescaled US06 cycle represents a 3.7 km route with an average speed of 22 km/h), maximum speed 37 km/h and duration of 596 seconds [Figure 17].

Furthermore, the second test cycle, a rescaled version of the standardized ECE-15 speed cycle, is implemented on the system with and without the use of supercapacitors [Figure 18, 19].

Data and cycle times of both cycles will then be used to evaluate the effects of the supercapacitor enhanced electric kart system compared to a battery-only system.

Sp

eed

(km

/h)

Time (s)

Figure 18: The heavy duty cycle imposed on the system

9. FUTURE WORK

Primary research for now was limited to evaluating the effect of directly connected supercapacitors on the electric go-kart’s acceleration cycle. Finishing the electric kart load simulation setup (including regenerative braking simulation) and extensive testing of the different energy systems, as well as creating a simulation tool to evaluate PPU design in various applications is planned as future work.

The integration and evaluation of a battery-supercapacitor interconnected converter will conclude this future work.

10. CONCLUSION

Preliminary test results indicate a potential for a parallel connected battery and supercapacitor system. Further research will investigate the effect of different supercapacitor capacitance values, as well as the effect

ISSN 2032-6653Page 0131



44

Sp

eed

(km

/h)

Time (s)

Figure 19: Rescaled ECE-15 speed cycle

of different battery capacities (dimensioning Ah versus F). These evaluations could help in optimizing the system and will be the subject of future papers.

Nevertheless, a definite intermediate conclusion can be drawn. The optimal combination of both a battery unit and a supercapacitor unit is not realized by just connecting them in parallel. The useful stored energy of the supercapacitor is defined by the potential difference. For optimal use of the available energy, a bidirectional DC-DC converter between the supercapacitor unit and the battery-system is advisable.

The resulting ability to limit the main energy supply’s current, for example with batteries, and relieve it from peak demands, positively influences battery life and Ah specifications. Also, when applying a supercapacitor based PPU to a fuel cell powered traction drive, the response to dynamic loads could improve.

Hence, the use of supercapacitors can definitely proof beneficial in multiple applications. Due to the high power density, high efficiency and long lifetime it can be an attractive durable option when creating a hybrid system. The main setbacks currently curbing the use of supercapacitors are its cost and volume requirements (especially in non-stationary applications). An electric kart propulsion system topology based on supercapacitors and an interconnected converter will be the subject of further research to evaluate not only the effect on lifetime, but also on the energy efficiency.

REFERENCES

[1] R. Kötz, M. Carlen, Principles and applications of electrochemical capacitors, Electrochimica Acta, The

Journal of the International Society of Electrochemistry, ISSN 0013-4686, Volume 45, Number 15, 3 May 2000, pp. 2483-2498(16).[2] Faggioli E.; Rena P.; Danel V.; Andrieu X.; Mallant R.; Kahlen H., Supercapacitors for the energy management of electric vehicles, Journal of Power Sources, ISSN 0378-7753 ,Volume 84, Number 2, December 1999, pp. 261-269(9).[3] Nigel Schofield, Heng Tian Yap, Hybrid Energy Sources for Electric Vehicle Propulsion, proceedings of EET-2007 European Ele-Drive Conference, Brussels, Belgium, 2007.[4] P. Van den Bossche, F. Vergels, J. Van Mierlo, J. Matheys, W. Van Autenboer, SUBAT: an assessment of sustainable battery technology, Journal of Power Sources, vol. 162, nr. 2, pp. 7, 2006, Impact factor: 2.770, impact year: 2005.[5] J. Van Mierlo, Simulation software for comparison and design of electric, hybrid and internal combustion vehicles with respect to energy, emissions and performances, PhD thesis promoted by Gaston Maggetto, Vrije Universiteit Brussel, 2000.[6] J. Schiffer, D. Linzen, D. U. Sauer, Heat generation in double layer capacitors, Journal of Power Sources, In Press, Available online since February, 7th 2006.[7] EPCOS: UltraCapTM Double Layer Capacitors, A New Energy Storage Device for Peak Power Applications, Product Profile, 2002.[8] Oliver Bohlen, Dirk Uwe Sauer, How to achieve the maximum lifetime for supercapacitors - A quantitative analysis of cell balancing systems and a proposal for improved cell protection concepts, EVS-22 Yokohama, Japan, 2006.[9] R. Kötz, M. Hahn, R. Gallay, Temperature behavior and impedance fundamentals of supercapacitors, Journal of Power Sources, 154 (2006) 550-555.

ISSN 2032-6653Page 0132



45

[10] Yonghua Cheng, Van Mierlo Joeri, Philippe Lataire, Method of identifying voltage difference of super capacitors and principle of voltage balancing, proceedings of EET-2007 European Ele-Drive Conference, Brussels, Belgium, 2007.[11] ASMO, http://ds1.dreifels.ch/asmo2/index.asp, 2007.[12] ODYSSEY, model PC1500, C5 = 55Ah, SoC 100% = 12.84V, Sealed dry cell valve regulated lead acid (VRLA) gas recombination technology battery type.[13] ASMO motor controller, Type MD 304 - 48 V 300 A.[14] Y. Cheng, J. Van Mierlo, P. Lataire, D. Nuttall and A. Forsyth, The Method of Optimization and Validation for Hybrid Fuel Cell Vehicles, proceedings of EET-2007 European Ele-Drive Conference, Brussels, Belgium, 2007.[15] Y. Cheng, J. Van Mierlo, P. Lataire and G. Maggetto, Test Bench of Hybrid Electric Vehicle with the Super Capacitor based Energy Storage, proceedings of ISIE 2007.[16] Exide Maxxima 900 Battery, http://automotive.exide.nl/downloads/Exide_Maxxima_UK.pdf, 2007.[17] Maxwell Boostcap, BCAP1500P, 1500F, 2,7VDC, http://www.maxwell.com/ultracapacitors/products/large-cell/bcap1500.asp, 2007.[18] Federal Register, Motor Vehicle Emissions Federal Test Procedure Revisions, Supplemental Federal Test Procedure (SFTP), 1996, http://www.epa.gov/otaq/sftp.htm, 2007.[19] New European Driving Cycle (NEDC), ECE-15, European Parliament, C5-0028/1999.

AUTHORS

Ing. Frederik Van Mulders graduated in 2005 as a Mechanical Industrial Engineer at the Erasmus University College Brussels and was invited to be a Ph.D. student at the ETEC-department for the Vrije Universiteit Brussel and the Erasmus University College Brussels. There, his main research involves the analysis and implementation of communication

and diagnostic systems in higher order automotive hybrid drivetrain systems, but is currently shifting towards supercapacitor based peak power units.

Ir. Jean-Marc Timmermans graduated in 2003 as an Electromechanical Engineer at the Vrije Universiteit Brussel. His master’s thesis dealt with the development of a test bench for electric bicycles. As an academic assistant, he is involved in projects about the evaluation of the environmental impact of

conventional and alternative road vehicles and also about the development of electric postal bikes. Further main research goes to the evaluation of hybrid electric drive trains for road vehicles.

M.Sc. Zach McCaffrey is an electrical engineering Ph.D. student at Vrije Universiteit Brussel. His current research focuses on electric driveline technologies for electric and hybrid vehicles.

Prof. dr. ir. Joeri Van Mierlo obtained his Ph.D. in Engineering Sciences from the Vrije Universiteit Brussel. Joeri is now a full-time lecturer at this university, where he leads the ETEC research team on transport technology. His research interests include vehicle and drive train simulation, as well

as the environmental impact of transportation.

Prof. dr. ir. Peter Van den Bossche graduated as civil mechanical-electrotechnical engineer from the Vrije Universiteit Brussel, and got involved in the research activities on electric vehicles at that institution. Since its inception in 1990, he has been coordinating the international association CITELEC, more particularly

in the field of electric and hybrid vehicle research and demonstration programmes. Furthermore, he has a particular research interest in electric vehicle standardization issues, in which he finished a Ph.D.

ISSN 2032-6653Page 0133

supercapacitor enhanced battery traction systems – … from the traction system working as a...

Documents