An automatic system to test Li-ion batteries and ultracapacitors for vehicular applications MIRKO MARRACCI, BERNARDO TELLINI Department of Energy and Systems Engineering University of Pisa, Fac. Of Engineering L.go L. Lazzarino n. 1, 56122 Pisa ITALY mirko.marracci@ing.unipi.it, bernardo.tellini@ing.unipi.it Abstract: - In this paper we present an automatic system to test Li-ion batteries and ultracapacitors. The main goal of the system is to compare performances of different storage systems during real operating conditions in order to make possible a fair comparison. Key-Words: - Li-ion batteries, ultracapacitors, measurement system, hybrid vehicles. 1 Introduction Li-ion batteries and ultracapacitors are commonly used storage systems that can find application in different fields like, for example, electric, hybrid, and fuel cell vehicles [1]-[6]. The common judgment on these two devices is that ultracapacitors are more power oriented, while batteries are more energy oriented ; these terms are used to indicate that the device operates more efficiently with relative short (power oriented) or long (energy oriented) charge or discharge times respectively. In some important applications, like hybrid vehicles, the energy storage system is required to deliver or absorb electric power in short intervals of time, typically a few seconds or tens of seconds; therefore they need to have different characteristics from batteries for pure-electric vehicles, that are discharged in tens of minutes or hours; in this case very-high power Li-ion batteries and ultracapacitors can be both suitable for this kind of application. Ragone plots [7] are commonly used for performance comparison of various energy storing devices; these plots report the values of energy density (Wh/kg) versus power density (W/kg) defining for each device family the specific application fields. However these plots are not rigorous, but give only trends: when reporting individual device characteristics into family plots often a single point per device is used. Instead, it can be very useful to evaluate the characteristic curve for these devices through experimental tests, reporting energy density vs power density, in relation to the discharge time. Furthermore, to make a fair, useful comparison between different storage systems, they must be subjected to the same stress; in particular, their specific power must be assessed using the same discharge duration. In this paper an automatic system to make a comparison between a commercial ultracapacitor and a commercial very-high power battery is presented. The system can reproduce actual work operating conditions that devices could find during the normal use in hybrid vehicles. 2 Devices under test The photographs of the ultracapacitor and very-high power battery under test are shown in Fig. 1 and Fig. 2 respectively. Fig.1 Photograph of the ultracapacitor under test ISBN: 978-1-61804-128-9 101
Fig.2 Photograph of the buttery under test 20 *Cn. Manufacturer s documentation is available in [9]. According to this manufacturer s data, therefore, a single cell, whose mass is 226g, should be able to deliver (at 15Cn regime, for about 230 s) around 1673 W/kg, (or 1460 W/kg if a overhead of 13% for case and BMS is taken into account) that is difficult to compare with the value of 13587 W/kg of the considered ultracapacitor, much higher, but available for only a split second. These data confirm the need of a fair, useful comparison between the two different storage systems, defining rigorous laboratory tests analyzing realistic, comparable operating conditions The devices under test are both commercially available storage systems. The ultracapacitor is a 20 F 15 V system, model Maxwell BPAK0020-P015- B01 whose manufacturer declared data are available in [8]. The ultracapacitor is composed of six 2.5 V modules in series with a nominal capacitance of 120 F. The battery is one of the highest power Li-ion batteries commercially findable; it is composed of eight cells in series with a nominal (two-hour) capacity of 7.2 Ah. Performance data for this battery are available for full discharges to up to 15*Cn, while pulse discharges may be performed to up to 3 Experimental setup The schematic architecture of the realized hardware experimental setup is reported in Fig. 3. The charging system is driven by a 1500 W Toellner Power Supply (Model TOE 8872) interfaced to the PC through a GPIB port. The DC power supply allows to charge the device under test keeping both voltage or current constant. In order to enable the execution of charging phases with linearly variable current, a specific software iterative procedure has been implemented. Electronic Load A Multimeters DC Power Supply B Battery Switches Control Acquisition System IEEE 488 IEEE 488 www Fig. 3 schematic architecture of the realized hardware experimental setup ISBN: 978-1-61804-128-9 102
The discharging system is driven by a 6000 W Zentro-Elektrik Electronic Load (Mod-el EL 6000). The device allows to discharge the equipment under test in constant current mode (I-Mode), constant voltage mode (U-Mode), constant power mode (P- Mode) or constant conductance mode (G-Mode) and is fully remote controlled via GPIB standard interface. Also in this case, in order to enable the execution of discharging phases with linearly variable current, a specific software iterative procedure has been implemented. The connection of the DC power supply or the electronic load is driven by two switches (A and B in Fig. 3) controlled via the implemented software by means of digital signals. When the switch A is closed the system is in a discharge phase and the electronic load is controlled while, when the switch B is closed the system is in a charge phase and the DC power supply is controlled. Otherwise, when both A and B switches are open the system is in a pause phase. The device under test is inserted in a precision refrigerating/warming test chamber with program control (BinderTM MK 53) ranging from T = -40 C up to T = +180 C. The temperature uniformity range between ±0.8 C (@ -40 C) to ±2.0 C (@ +150 C) while the declared temperature fluctuation is limited to ±0.3 C for each selected temperature. In Fig. 4 a photograph of the climatic chamber is shown. as the total device voltage in order to assure that the voltage of each cell remains into a safety range (2.7 4.2 V). In Fig. 5 a photograph of the whole laboratory test station is reported. Fig.4 Photograph of the climatic chamber The measuring system acquire voltage at the terminals of the device under test directly while for the acquisition of the current a transducer is generally required [10]-[14]. In our case a shunt (for high currents) and a LEM current transducers (for small currents) are used respectively. For security reasons the acquisition system must be able to acquire and control each cell voltage as well Fig.5 Photograph of the whole test system Taking into account the quite high common-mode voltage range required, signals are converted into digital by means of three NI 9219 modules installed on a NI CompactDAQ USB data acquisition system. ISBN: 978-1-61804-128-9 103
Each module features 4 simultaneous sampling universal channels that can measure several signals from sensors; measurement ranges differ for each type of measurement and include up to ± 60 V for voltage and ± 25 ma for current, with 250 Vrms channel-to-channel isolation and 24-bit resolution; in this way the system can acquire simultaneously each cell voltage as well as the total voltage across and the current through the device under test. The acquisition system measures the ambient and device temperature by means of two RTD (PT100) transducers. Acquired signals can be used both to control the ambient temperature (when the device is inserted into the climatic chambers) or to generate an alarm signal and stop the test when the device temperature reach a prearranged value. The experimental setup is managed by a dedicated Virtual Instrument (VI), developed in a LabVIEW environment, running on a PC. The front panel of the realized VI during a linear current charge phase is represented in Fig. 6 4 Results and discussion One of the main results obtained from the test to compare Li-ion battery and ultracapacitor has been the specific discharge powers obtained at imposed fixed discharging currents for times typical of vehicular applications (5-240 s). The specific rated power for the ultracapacitor ranges from 25 W/kg (discharge of 240 s) to 1417 W/kg (discharge of 5 s) while the specific rated power for the Li-ion battery ranges from 1450 W/kg (discharge of 240 s) to 2100 W/kg (discharge of 5 s). The ultracapacitor is on the contrary advantaged from having nearly symmetrical charge and discharge powers while the limitation introduced to charging to I=3*Cn is a major Li-ion battery disadvantage. In this case the charge specific power is constant to 330 W/kg for the battery while ranges from 30 W/kg (charge of 240 s) to 1550 W/kg (charge of 5 s) for the ultracapacitor. Fig. 6 Front panel of the VI developed for the test system ISBN: 978-1-61804-128-9 104
In order to make a deep comparison between commercial ultracapacitor and very-high power battery, reproducing actual work operating conditions that devices could find during, for example, the normal use in hybrid vehicles, a test of the two systems using the New European Driving Cycle (NEDC) [15] is currently in progress. The New European Driving Cycle is a driving cycle consisting of four repeated urban driving cycles (ECE-15) and an Extra-Urban driving cycle, or EUDC. The NEDC is supposed to represent the typical usage of a car in Europe. Analyzing the main drive train power fluxes for a series hybrid vehicle performing a NEDC cycle, we found the typical current profile required from Rechargeable Energy Storage System (RESS) that is represented in the following Fig. 7. Current (A) 50 25 0-25 -50 0 200 400 600 800 1000 1200 Time (s) Fig.7 Current profile of the storage system for a series hybrid vehicle performing a NEDC cycle. This complex current profile reproduced in Fig. 7 can be used to analyze the storage system behavior during a realistic stress condition. For safety reasons, during the battery tests, the current profile has been obtained reducing peaks of discharge and charge current of nearly 6*Cn; This limitation does not constraint markedly vehicle operation and keep temperatures in a perfectly acceptable range. In Fig. 8 preliminary results are reported for the battery system. The figure shows the total voltage across the battery during the imposed cycle. Voltage (V) 33 32 31 30 29 28 0 200 400 600 800 1000 1200 Time (s) Fig.8 Voltage profile of the storage system for a series hybrid vehicle performing a NEDC cycle. 5 Conclusions An automatic system to test and compare Li-ion batteries and ultracapacitors has been presented. The main goal of this system is to compare performances of different storage systems during real operating conditions in order to make possible a fair comparison. The system has been used to compare performances of a real commercially available ultracapacitor and high-power Li-ion battery. References: [1] P. Thounthong, S. Raël, B. Davat, Energy management of fuel cell/battery/supercapacitor hybrid power source for vehicle applications, Journal of Power Sources, vol. 193, pp. 376-385, Issue 1, August 2009. [2] A.F.Burke, Batteries and Ultracapacitors for Electric, Hybrid, and Fuel Cell Vehicles, Proceedings of the IEEE, vol. 95, pp. 806-820, no. 4, April 2007. [3] J. Bauman, Fuel-cell-Battery-Ultracapacitor Vehicles, IEEE Transactions on Vehicular technology, vol. 57, n. 2 pp. 760-769, 2008. [4] S. Barsali, M. Ceraolo, M. Marracci, B. Tellini, Frequency Dependent Parameter Model of Supercapacitor, Measurement vol. 43, pp. 1683 1688, 2010. [5] R. Ferrero, M. Marracci, M. Prioli, B. Tellini, Simplified Model for Evaluating Ripple Effects on Commercial PEM Fuel Cell, International Journal of Hydrogen Energy (2012), vol. 37, Issue 18, September 2012, pp. 13462-13469, http://dx.doi.org/10.1016/j.ijhydene.2012.06.03 6 ISBN: 978-1-61804-128-9 105
[6] R. Ferrero, M. Marracci, B. Tellini, Impedance spectroscopy on a single PEM fuel cell for the evaluation of current ripple effects, Proc. of IEEE I2MTC 2012, Graz, Austria, May 13-16, pp. 52-56, 2012. doi:10.1109/i2mtc.2012.6229480 [7] T.Christen, M. W. Carlen, Theory of Ragone plots, Journal of Power Sources, vol. 91, pp. 210-216, Issue 2, December 2000. [8] http://www.maxwell.com/ultracapacitors/ products/ modules/ bpak0020-15v.asp [9] http://www.kokam.com/english/ product/ battery_main.html [10] S. Ziegler, R. C. Woodward, H. Ho-Ching Iu, L. J. Borle, Current Sensing Techniques: A Review, IEEE Sensors J., vol. 9, no. 4, pp. 354-376, Apr. 2009. [11] R. Ferrero, M. Marracci, B. Tellini, Analytical Study of Impulse Current Measuring Shunts With Cage Configuration, IEEE Trans. Instrum. and Meas., vol. 61, no. 5, pp. 1260-1267, 2012. [12] G. Becherini, S. Di Fraia, M. Marracci, B. Tellini, C. Zappacosta, G. Robles, Critical parameters for mutual inductance between Rogowski coil and primary conductor, Proc. of I2MTC 2009, 5-7 May, pp. 432-436, 2009. [13] M. Marracci, B. Tellini, C. Zappacosta, G. Robles, Critical Parameters for Mutual Inductance Between Rogowski Coil and Primary Conductor, IEEE Trans. Instrum. And Meas., vol. 60, no. 2, p. 625-632, Feb. 2011. [14] R. Ferrero, M. Marracci, B. Tellini, Analytical study of high pulse current shunts, Proc. of I2MTC 2011, 5-7 May, Hangzhou, China, pp. 1-6, 2011. [15] New European Driving Cycle, available at http://en.wikipedia.org/wiki/new_european_d riving_cycle ISBN: 978-1-61804-128-9 106