Electric Go-Kart with Battery-Ultracapacitor Hybrid Energy Storage System

W. O. Avelino*, F. S. Garcia**, A. A. Ferreira***, and J. A. Pomilio*

*University of Campinas, Brazil, **Ekion Electric Vehicles Technologies, ***Federal University of Juiz de Fora, Brazil Email: wellingtonavelino@ymail.com, fsgarcia@ekion.com.br, andre.ferreira@engenharia.ufjr.br, antenor@fee.unicamp.br

Abstract-An electric go-kart was designed and built. It uses a hybrid energy storage system composed of a battery and an ultracapacitor. To integrate the two energy storage devices, a multiple-input DC-DC converter was used to implement a control strategy that divides the power between them. The main objectives of using ultracapacitors in complement to batteries are: improving performance (i.e. acceleration), increase the system efficiency (including through the use of more regenerative braking) and extend the battery life. This paper presents the design of the electric go-kart and the experimental results obtained with the prototype.

Keywords-Batteries; DC-DC Power Converters; Electric Vehicles; Supercapacitors; Ultracapacitors.

INTRODUCTION

As part of a broader trend of transportation electrification, electric go-karts are being used as a viable alternative to conventional (i.e. gasoline) go-karts. As important advantages, electric go-karts have lower energy cost (because of their higher efficiency) and are more reliable, resulting in lower life-cycle costs. Also, electric go-karts are silent and do not release pollutants [1]. The noise produced by conventional go-karts cause disturbance to the neighborhood, limiting their use to remote areas. Because of the emission of smoke and pollutants, it is onerous to use conventional go-karts in indoor circuits and they are usually only used in outdoor circuits.

The electric go-kart solves the main drawbacks of the conventional go-kart. Nevertheless, its design should take into account the need for good performance and high reliability at low cost. With the objective of achieving these improvements, this paper evaluates the use of a battery-ultracapacitor hybrid energy storage system (HESS).

The battery accounts for a significant fraction of the cost of the go-kart [2]. Usually the go-karts are used for many hours a day and the battery perform many charge-discharge cycles during this period. Therefore, in order to reduce the maintenance costs, it is important to take the life of the batteries into consideration during the design of the go-kart.

As for many electric vehicles, it is of fundamental importance to reduce the weight of the go-kart, as this implies in a reduction of the power demand (for a given speed and acceleration) and of the total energy consumption. Therefore, a high specific power and high specific energy source is desirable [3]. It is well know that, considering the present technology, ultracapacitors have low specific energy and high

specific power, while batteries have higher specific energy and lower specific power. Also, ultracapacitors have much better cycle lives, can operate over a broader temperature range and are more efficient than batteries [4] [5].

The hybrid energy storage system is used with the objective of taking advantage of the best characteristics of each device, creating a system that is superior, in this application, than any of the devices used alone. It is expected that the coordination of the energy storage devices will result in increased system efficiency [6]. The combination of an ultracapacitor and a battery is more suitable to applications with fast transients and high peak-to-average power ratio, allowing the ultracapacitor to act as an energy buffer for the battery. In this scenario, the instantaneous battery power will be kept closer to the average power demand, thus reducing its RMS current and the internal losses.

ELECTRIC GO-KART DESIGN

Fig. 1 presents the speed profile of one lap of a go-kart in a racing circuit, as logged in a conventional go-kart with a professional pilot.

Using a standard model (see [3], [7], and [8]) and the specifications of the go-kart presented in TABLE I, a simulation was performed with MATLAB/Simulink to calculate the instantaneous power demand.

Fig. 1. Simulated power demand of a go-kart

978-1-4799-0148-7/13/$31.00 ©2013 IEEE

TABLE I

GO-KART SPECIFICATIONS

The power demand shown in Fig. 1 was used as the starting

point in the electric go-kart powertrain design. It is clear that the power demand during a race have a high peak-to-average ratio. The instantaneous power demand is high during acceleration and becomes negative during braking, which suggests that this application can benefit from the use of the ultracapacitor as an energy buffer.

The peak power is 9.22 𝑘𝑊 during acceleration and −11.7 𝑘𝑊 during braking (in a electric go-kart, the braking power demand may be met by regenerative braking plus the power dissipated on mechanical brakes). The average power is 1.1 𝑘𝑊, while the average power, considering only the positive part of the signal (that is, neglecting any regenerative braking power), is 3.0 𝑘𝑊.

The electric go-kart prototype is shown in Fig. 2., where (A) is the battery, (B) the ultracapacitor, (C) is the custom-built DC/DC converter (as described in the next section), (D) is the inverter, and (E) is the electric motor, which is connected to the rear axle using V-belts and pulleys. The specifications of the main components used in the electric go-kart are presented in TABLE II. The transmission ratio from the motor to the rear axle is 4.5:1.

TABLE II COMPONENTS’ SPECIFICATIONS

MULTIPLE INPUT DC-DC CONVERTER

Many structures have been proposed to integrate batteries and ultracapacitors, including the simple parallel connection (e.g. [9]), semi-active configurations (e.g. [10], [11], and [12]) and fully active configurations (e.g. [13], [14], [15], and [16]). Still other connection structures are possible (e.g. [17] and [18]). References [19] and [20] present some interesting discussions about the advantages and drawbacks of each configuration.

In this work, a fully active configuration was used, that is, a multiple input DC-DC converter can control, independently, the power flow of the battery and of the ultracapacitor. Also, the DC-DC converter allows to regulate the voltage at the DC-link, making it possible to maintain the DC-link voltage stable (despite variations in the sources’ voltages) or to adjust dynamically the DC-link voltage according to some strategy that increases the system efficiency (see reference [21]).

The simplified electrical circuit of the DC-DC converter used in this work is shown in Fig. 3. The converter is a 3-phase bidirectional boost converter. One phase was connected to the battery, while the other two phases were connected to the ultracapacitor. The phases of the ultracapacitor are out of phase by 180°, in order to reduce the input and output current ripples.

The output of the three phases is the common DC-link, which is also shared by the inverter. That is, the load seen by

Weight (without the pilot) 150 kg Aerodynamic drag coefficient (estimated) 0.9 Frontal area of kart 0.6 m² Wheel Radius 0.14 m

Motor Type AC, induction, 3-phase Rated Power 5 kW Nominal Voltage (at the DC-link) 84 V Manufacturer/Model FIMEA / N20

Inverter Nominal Input Voltage 84 V Peak input current 300 A Frequency range Up to 200Hz Manufacturer/Model ZAPI/AC-2

Battery Type Lithium (LiFePO4) Cell configuration (Series/Parallel) 10S/1P Nominal Voltage 32 V Operational Voltage 28 V ~ 40 V Nominal Capacity 100 Ah Weight 35 kg Model TS-LFP100AHA Cell Manufacturer Winston Batteries

Ultracapacitor Rated Capacitance 165 F Rated Voltage 48 V Manufacturer Maxwell Technologies

Fig. 2. Electric go-kart prototype

the DC-DC converter is the inverter (that feeds the three-phase induction motor), which is modeled in Fig. 3 by a current source.

Naturally, the DC-link voltage must be higher than the battery voltage and the ultracapacitor voltage. As this structure allows the ultracapacitor voltage to fluctuate in a reasonable range without affecting the DC-link voltage, most of the ultracapacitor energy is available to the system.

The switches {𝑇1, … ,𝑇6} are N-type MOSFETs, model IXFN230N20T, rated200 𝑉, 230 𝐴. The inductors have toroidal cores made of a special high-frequency iron-powder material and are rated 150 𝐴, 120 𝜇𝐻. The switching frequency is 50 𝑘𝐻𝑧. The current in the inductor of each phase is measured using a LEM LA200 hall-effect sensor. The battery, ultracapacitor and output voltages are measured with using a LEM LV 20-P hall-effect sensor. The measured signals are then sampled by the microcontroller, where all control loops are digitally implemented.

The input current at each phase was limited (by software) to the range of −100 𝐴 to 100 𝐴. One phase was connected to the battery and two phases were connected to the ultracapacitor. Considering the battery voltage of 32 𝑉, this implies that a maximum of 3.2 𝑘𝑊 can be extracted from (or stored in) the battery, while, considering the ultracapacitor voltage of40 𝑉, a maximum of 8 𝑘𝑊 can be extracted from (or stored in) the ultracapacitor. Combined, the two devices can supply (or store) up to 11.2 𝑘𝑊.

The multiple-input DC-DC converter was designed and built specially for this electric go-kart. The internal assembly of the converter is shown in Fig. 4. This converter is also shown in the go-kart, indicated as part (C) in Fig. 2.The power devices, the busbars and the inductors are at the bottom of the case, as shown in Fig. 4. (A), while the microcontroller, the auxiliary power supplies, the protection circuits, the signal conditioning circuits and the isolated gate drivers are in the upper part of the case, as shown in Fig. 4 (C). An aluminum plate protects the control circuitry from possible interference from the power circuitry, as shown in Fig. 4 (B). The converter with the closed case, as it was installed in the go-kart, is shown if Fig. 4 (D).

The control algorithms were implemented in a Texas Instruments microcontroller, model TMS320F28335. The

signals of the sensors (currents, voltages and temperatures) are connected through shielded cables to a signal conditioning circuit and then delivered to the microcontroller. A hardware-only protection was used for over-current, over-voltage or high heat sink temperature, to avoid damages in the case of a microcontroller or software failure.

DC-DC CONVERTER CONTROL

The performance of the HESS is fundamentally determined by the energy management algorithm, which determines the power extracted from each energy storage device, for a given load demand.

The efficiency of storing and retrieving energy in the ultracapacitor is higher than in the battery [5]. On the other hand, the amount of energy stored in the ultracapacitor is small when compared to the energy in the battery.

One effective energy management algorithm consists in dividing the power demand in the frequency domain, so that the battery supply the low frequency components of the power demand and the ultracapacitor supplies the high frequency components of it. This algorithm, which was implemented in this work, was described in [15]. Other control algorithms have been presented, e.g. [14], [22], and [23]. The objective of the implemented control strategy is to improve the performance and energy efficiency, while extending the battery life.

While the main control algorithm used in this work is described in detail in reference [15], a brief revision of it follows here.

The input current of each phase {𝑖𝑏𝑎𝑡, 𝑖𝑢𝑐1, 𝑖𝑢𝑐2} is controlled by an internal loop, as usually done in current mode control [24]. The current controller is a conventional PI controller with a low pass filter, digitally implemented with a sampling rate of 50 𝑘𝐻𝑧, with saturation of the output duty cycle (from 0 to 0.75) and a dynamic saturation of the integral part, implemented as described in reference [25]. The gains were calculated using the k-factor method [26] (in this method, the controller used is called a Type 2 controller). Fig. 5 shows the control diagram of the system.

Fig. 4. Internal and external views of the DC-DC converter

Fig. 3. DC-DC converter circuit (simplified)

A B

C D

The current controllers are shown in Fig. 5

as�𝐺𝑖𝑐𝑏𝑎𝑡(𝑠),𝐺𝑖𝑐1(𝑠),𝐺𝑖𝑐2(𝑠)�, They control directly the duty cycle of each phase {𝛿𝑏𝑎𝑡,𝛿𝑢𝑐1,𝛿𝑢𝑐2}. The cutoff frequency was chosen as2 𝑘𝐻𝑧 and margin phase as 70°.

The output voltage controller, 𝐺𝑣𝑐_𝑜(𝑠), generates the current references {𝑖𝑟𝑒𝑓𝑢𝑐1 , 𝑖𝑟𝑒𝑓𝑢𝑐2} to the phases connected to the ultracapacitor. Therefore, when there is a disturbance in the output voltage (say, because of a load transient) the ultracapacitor is the first device to react to this change. The output voltage controller 𝐺𝑣𝑐_𝑜(𝑠) is also is a conventional PI controller with a low pass filter, digitally implemented with a sampling rate of 50 𝑘𝐻𝑧, with saturation of the current reference (from −100𝐴 to 100𝐴) and a dynamically saturable integrator, implemented as described in reference [25]. The gains were calculated using the k-factor method [26] (in this method, the controller used is called a Type 2 controller). The cut-off frequency was chosen as 200 𝐻𝑧, with a phase margin of 85°.

The ultracapacitor voltage controller 𝐺𝑣𝑐_𝑢𝑐(𝑠) has the function of restoring the ultracapacitor voltage to its setpoint voltage (in this work, 42V), after transients. But, as this controller is very slow, the ultracapacitor voltage can fluctuate in a reasonable range, allowing the ultracapacitor to realize its role of supplying and storing energy. For a more complete description of how this controller works, see [15].

The ultracapacitor voltage controller 𝐺𝑣𝑐_𝑢𝑐(𝑠) generates the battery current reference 𝑖𝑟𝑒𝑓_𝑏𝑎𝑡. This controller has the same basic structure of the output voltage, but was designed with a cutoff frequency of 0.01 𝐻𝑧 and a margin phase 50°. Its bandwidth is limited to a frequency much lower than the output voltage controller bandwidth, consequently the ultracapacitor is the first device to be affected by a change in load demand. Only after the ultracapacitor voltage is disturbed, battery current will react to, slowly, restore ultracapacitor voltage to the reference level.

As it was shown in [15], this controller structure results in a division of the power demand where the low frequency components are supplied by the battery and the high frequency components are supplied by the ultracapacitor. It can be said, in a simplified way, that the power demand related to acceleration and braking have mostly high frequency components (notice that, in this context, high frequency means 𝑓 > 0.01 𝐻𝑧), while the power related to

“dissipative” components (aerodynamic drag, rolling resistance etc.), have mostly low frequency components.

The crossover frequency of the ultracapacitor voltage controller 𝐺𝑣𝑐_𝑢𝑐(𝑠) will determine the range of frequencies that are supplied by the ultracapacitor and the range of frequencies that are supplied by the battery, which, in turn, defines the sharing of the power demand between the sources.

EXPERIMENTAL RESULTS

The experimental results presented in this paper were obtained in a go-kart track in the city of Campinas, Brazil, as shown in Fig. 6. This track is located in a city park (Taquaral) and was shut down about 6 years ago as a result of a legal action by the neighbors that complained about the noise produced by the conventional go-karts.

The experimental results presented in this paper were obtained by driving the go-kart for two laps, following the course shown by the arrows in Fig. 6. Each lap has about 550𝑚, with the total route distance of about 1.1 𝑘𝑚.

The following signals were recorded with a battery-operated, 4-channel oscilloscope (Tektronix TPS2014): (1) battery voltage; (2) battery current; (3) ultracapacitor voltage; and (4) ultracapacitor current. The speed and position were logged with an app for a mobile phone with a GPS receiver. The data was plotted with MATLAB.

From the voltage and current at each energy storage device, the instantaneous power and the state of charge were calculated. The experimental data was used to plot the graphs of Fig. 7 and to calculate the parameters of TABLE III.

In Fig. 7, it is shown the go-kart speed, as measured by the GPS receiver. It must be said that the profile does not exactly represent the racing condition, as during the experiments the electric go-kart was alone in the track and the performance was not pushed to the limit. Fig. 7 also shows the battery

Fig. 5. Control diagram of the DC-DC converter

Fig. 6. Go-kart test track

power (𝑃𝑏𝑎𝑡), the ultracapacitor power (𝑃𝑢𝑐) and the total power (𝑃𝑡𝑜𝑡𝑎𝑙 = 𝑃𝑏𝑎𝑡 + 𝑃𝑢𝑐).

The initial state of charge of the battery (90%) was estimated from the open circuit voltage and the variation of the battery state of charge (SoC) was calculated as the integral of the battery current in Ah divide by the nominal battery capacity (100Ah).

The ultracapacitor SoC was estimated from its terminal voltage, using the formula of the energy stored in the ultracapacitor as 𝐸𝑢𝑐 = 0.5𝐶𝑣𝑢𝑐2. The total (nominal) energy that can be stored in the utracapacitor, considering 𝐶 = 165𝐹 and 𝑣𝑢𝑐 = 48𝑉 is 190 𝑘𝐽. The ultracapacitor SoC is then given as the fraction of the instantaneous energy stored (as estimated from the terminal voltage) over 190 𝑘𝐽.

Fig. 7 indicates that, as expected, the battery supplies low frequencies components of the power demand and the battery power is close to the total average power. The ultracapacitor power averages close to zero, but it has an important function in supplying transient power related to acceleration.

TABLE III PARAMETERS CALCULATED FROM EXPERIMENTAL DATA

CONCLUSIONS

The ultracapacitor-battery hybridization with the control strategy proposed provides significant contributions to make good use of the complementary characteristics of each source.

The power converter developed is responsible for regulating the DC-link voltage and restoring ultracapacitor voltage after transients.

As the system acts as a low-pass filter for battery current, the RMS current in the battery is reduced (compared with an only-battery system), and a higher efficiency is expected. Moreover, lower rates of discharge and attenuation of high frequency components in the battery current should result in less heating and longer life battery.

The research on the lifetime of batteries shows that higher temperatures reduce the battery life, therefore it is expected that reducing the peak and RMS currents in the battery will have a significant positive effect on battery life and efficiency [23] [27].

Nevertheless, further research and experiments are needed to quantify the increase in battery life that can be obtained in a system using ultracapacitors and batteries compared to a system with batteries only.

ACKNOWLEDGMENTS

This work was supported by Ekion Electric Vehicles Technologies, FAPESP (the São Paulo state research support agency), CNPq and CAPES. Texas Instruments supported the research with microcontrollers and programmers.

Fig. 8. Load Profile for the of Electric Go-kart

Total Energy Consumption 359 𝑘𝐽 = 0.1 𝑘𝑊ℎ Distance 1.1 𝑘𝑚 Distance / Total Energy Consumption 11 𝑘𝑚/𝑘𝑊ℎ Maximum Power (Battery) 2.70 𝑘𝑊 Maximum Power (Ultracapacitor) 2.82 𝑘𝑊 Maximum Power (Total) 5.35 𝑘𝑊 Average Power (Battery) 1.51 𝑘𝑊 Average Power (Ultracapacitor) −0.05 𝑘𝑊 Average Power (Total) 1.46 𝑘𝑊

Fig. 7. Experimental results obtained with the go-kart prototype

REFERENCES

[1] K. Vitols, N. Reinberg, A. Sokolovs, and I. Galkin, "Drive selection for electric kart," in Power Electronics and Motion Control Conference (EPE/PEMC), 2010 14th International, September 2010, pp. T9-15 - T9-18.

[2] N. Shidore, J. Kwon, and A. Vyas, Trade-off between PHEV fuel efficiency and estimated battery cycle life with cost analysis, pp. 669 - 677 , October 2009.

[3] M. Ehsani, Y. Gao, S. E. Gay, and A. Emadi, Modern Electric, Hybrid Electric, and Fuel Cell Veihcles.: Boca Raton, FL: CRC Press, 2004.

[4] John M. Miller, U. Deshpande, T. J. Dougherty, and T. Bohn, "Power Electronic Enabled Active Hybrid Energy Storage System and its Economic Viability," in Applied Power Electronics Conference and Exposition, APEC, Washington, DC, 2009, pp. 190-198.

[5] A. Virtanen, H. Haapala, S. Hännikäinen, T. Muhonen, and H. Tuusa, "Calorimetric efficiency measurements of supercapacitors and lithium-ion batteries," in Applied Power Electronics Conference and Exposition (APEC), 2011 Twenty-Sixth Annual IEEE, Fort Worth, 2011, pp. 1367-1373.

[6] S. M. Lukic, J. Cao, R. C. Bansal, F. Rodriguez, and A. Emadi, "Energy Storage Systems for Automotive Applications," Trans. on Industrial Electronics, vol. 55, pp. 2258–2267, June 2008.

[7] D.W. Gao, C. Mi, and A. Emadi, Modeling and Simulation of Electric and Hybrid Vehicles, pp. 729 - 745 , April 2007.

[8] M. Steve. Hybrid-Electric Vehicle Model in Simulink. [Online]. http://www.mathworks.com/matlabcentral/fileexchange/28441-hybrid-electric-vehicle-model-in-simulink

[9] R. A. Dougal, S. Liu, and R. E. White, "Power and Life Extension of Battery–Ultracapacitor Hybrids," IEEE Trans. on Components and Packaging Technologies, vol. 25, pp. 120–131, Mar. 2002.

[10] Jian Cao and A. Emadi, A New Battery/UltraCapacitor Hybrid Energy Storage System for Electric, Hybrid, and Plug-In Hybrid Electric Vehicles, pp. 122 - 132, May 2011.

[11] J. Dixon, I Nakashima, and E. F. Arcos, "Electric Vehicle Using a Combination of Ultracapacitors and ZEBRA Battery," IEEE Trans. on Industrial Electronics, vol. 57, pp. 943-949, Mar. 2010.

[12] B. A. Niemoeller and P. T. Krein, "Battery-Ultracapacitor Active Parallel Interface with Indirect Control of Battery Current," in Power and Energy Conference at Illinois (PECI), Illinois , 2010, pp. 12-19.

[13] L. Solero, A. Lidozzi, and J.A. Pomilio, Design of multiple-input power converter for hybrid vehicles , pp. 1007-1016, September 2005.

[14] A. A. Ferreira, J.A. Pomilio, G. Spiazzi, and L. de Araujo Silva, "Energy Management Fuzzy Logic Supervisory for Electric Vehicle Power Supplies System," Power Electronics, IEEE Transactions on, vol. 23, no. 1, pp. 107-115, Jan 2008.

[15] F.S. Garcia, A.A. Ferreira, and J.A. Pomilio, Control Strategy for Battery-Ultracapacitor Hybrid Energy Storage System, pp. 826 - 832, February 2009.

[16] Mamadou B. Camara, Brayma Dakyo, and Hamid Gualous, "Polynomial Control Method of DC/DC Converters for DC-Bus Voltage and Currents Management—Battery and Supercapacitors," IEEE Trans. on Power Electronics, vol. 27, no. 3, pp. 1455-1467, March 2012.

[17] S. Colton, "A Simple Series Battery/Ultracapacitor Drive System for Light Vehicles and Educational Demonstration," in EVER ’09, Monaco, 2009.

[18] C-H. Tu and A. Emadi, "A novel series-parallel reconfigurable hybrid energy storage system for electrified vehicles," in Transportation Electrification Conference and Expo (ITEC), 2012 IEEE, Michigan, USA , 2012, pp. 1-4.

[19] Garrett Nielson and Ali Emadi, "Hybrid energy storage systems for high-performance hybrid electric vehicles," in Vehicle Power and Propulsion Conference (VPPC), Chicago, USA, 2011, pp. 1-6.

[20] A. Kuperman and I. Aharon, "Battery-ultracapcitor hybrids for pulsed current loas: A review," Renewable and Sustainable Energy Reviews, Elsevier, vol. 15, pp. 981-992, February 2011.

[21] Jorge O. Estima and Antonio J. M. Cardoso, "Efficiency Analysis of Drive Train Topologies Applied to Electric/Hybrid Vehicles," IEEE Trans. on Vehicular Technology, vol. 61, no. 3, pp. 1021-1031, March 2012.

[22] "Polynomial Control Method of DC/DC Converters for DC-Bus Voltage and Currents Management—Battery and Supercapacitors," IEEE Trans. on Power Electronics, vol. 27, no. 3, pp. 1455-1467, March 2012.

[23] R. Carter, A. Cruden, and P. Hall, Optimizing for Efficiency or Battery Life in a Battery/Supercapacitor Electric Vehicle, May 2012.

[24] B.H. Cho, H.S. Bae, and J.H. Lee, Review of current mode control schemes and introduction of a new digital current mode control method for the parallel module DC-DC converters, pp. 202 - 210, October 2009.

[25] S. Buso and P. Mattavelli, Digital Control in Power Electronics. USA: Morgan & Claypool, 2006.

[26] H. D. Venable, The k-factor: A New Mathematical Tool for Stability Analysis and Synthesis, pp. 22-24, March 1983.

[27] N. Omar, J. Van Mierlo, B. Verbrugge, and Van den Bossche, "power and life enhancement of a battery-electrical double layer capacitor for hybrid electric and charge-depleting plug-in vehicle applications," Electrochim. Acta, vol. 55, no. 25, pp. 7524-7531, October 2010.