# Performance Assessment of a Lithium-Polymer Battery for HEV Utilizing Pack-Level Battery Hardware-in-the-Loop-Simulation System

• Han, Sekyung (Dept. of Electronic and Control Engineering, Hanbat National University) ;
• Lim, Jawhwan (System Integration & Test Department, Battery System R&D, SCEKSK-continental)
• Received : 2012.11.08
• Accepted : 2013.08.07
• Published : 2013.11.01
• 122 21

#### Abstract

A pack-level battery hardware-in-the-loop simulation (B-HILS) platform is implemented. It consists of dynamic vehicle models using PSAT and multiple control interfaces including real-time 3D driving and GPS mode. In real-time 3D driving mode, user can drive a virtual vehicle using actual drive equipment such as steering wheel and accelerator to generate the cycle profile of the battery. In GPS mode, actual road traffic and terrain effects can be simulated using GPS data while the trajectory is displayed on Google map. In the latter part of the paper, several performance tests of an actual lithium-polymer battery pack are carried out utilizing the developed system. All experiments are conducted as parts of actual development process of a commercial battery pack adopting 2nd generation Prius as a target vehicle model. Through the experiments, the low temperature performance and fuel efficiency of the battery are quantitatively investigated in comparison with the original nickel-metal hydride (NiMH) pack of the Prius.

# 1. Introduction

MODEL-BASED quantitative simulation schemes such as rapid prototyping and hardware-in-the-loop-simulation (HILS) are being widely accepted by many industrial engineers as a vital tool to enhance the efficiency of a complex system design [1]-3]. Especially, HILS helps to reduce the integration time by enabling the component designer to perform the test in an actual control loop with the target device in it. The control loop consists of the actual target device and interacting models. The engineer thus can verify the operation of the target device before it is actually integrated to the system. For this reason, HILS is being utilized in various fields including the automotive industry [4-6].

Traditionally, vehicle power-train has been considered as an exclusive area for the vehicle manufacturers. Recently, however, as the electrification of vehicles proceeds, electric propulsion components, especially battery, are mainly developed in a separated company. Moreover, since the development phase usually overlaps with that of the platform components, the battery manufacturers have little choice but to rely on the stand-alone test with offline profiles, such as standardized cycle pattern and/or the drive cycles obtained from similar electric vehicles, during the development of the battery. The integrated test is often deferred until the relevant components become available. Consequently, many unexpected troubles continue delaying the overall development period.

To overcome the problem, the authors have developed a battery HILS (B-HILS) system on behalf of SK Innovation, one of the major battery-manufactures in South Korea. B-HILS enables the integrated test of a pack-level battery by establishing a closed control loop with a vehicle model. Currently, many vehicle models are available in various types [7-17] We deployed a quasi-steady vehicle model, the Power-train System Analysis Toolkit, or PSAT, from Argonne National Laboratory (ANL) for our system [9]. PSAT is built upon Simulink environment with model-based design. It contains numerous types of commercial vehicle models and enables user to build a custom model by assigning appropriate components through its configuration tool. We customized PSAT for our own purpose by directly cooperating with the PSAT team in ANL.

Neeraj in ANL and H. Dai et al have developed the similar platforms [18, 19]. Their platforms, however, accept only a fixed offline drive cycle and focus mainly on the verification of the battery management system (BMS) functions. In contrast, our B-HILS provides multiple interfaces to feed the drive profiles. For example, the 3D driving mode enables user to drive a virtual vehicle with actual drive equipment such as steering wheel and accelerator. The corresponding profiles are then fed into the B-HILS, eventually cycling the battery. In the GPS mode, the drive profile is generated from GPS data logged during an actual driving on the road. Through this mode, a real traffic situation and a specific terrain effect can be applied in the battery test.

Our B-HILS has been actually utilized in developing various commercial battery packs. In this paper, we introduce several case study results regarding a Lithium polymer (LiPo) battery pack, which has been developed for a mid-size hybrid sedan. The typical vehicle model utilized during the experiments is 2nd generation Toyota Prius. In some cases, the results are compared with those of the original nickel-metal hydride (NiMH) pack of Prius. Specifically, the experiments are performed to investigate: the fuel efficiency with respect to battery size; the performance of the battery in an extremely low temperature; the effect of BMS power restriction in a harsh driving condition such as mountain road.

Rest of the paper is organized as follows. Firstly, the configuration of the B-HILS as well as its unique features is introduced in Section 2. In the following section, the performance of SK’s LiPo battery is investigated using the B-HILS system. The experiments are organized so the features of the B-HILS can be explored. Finally, the achievements of this paper are summarized with conclusions in Section 40.

# 2. Configuration of Battery HILS

The original PSAT is a pure software package. All components are described in a software model and connected to each other via input/output ports composing a closed control loop. The replacement of the software battery model in PSAT with a real battery was achieved by directly collaborating with the PSAT team in ANL. All required input/output ports are appropriately tuned to interact with the real BMS signals. The requested electric power is conveyed to a cycler to actually drive the battery. The employed cycler is a customized product directly imported from Hitachi for the B-HILS. It accepts the control command via either of an analog voltage or CAN signal.

Physical I/O interfaces are also required to interact with the battery (BMS) and the cycler. To this end, we employed the dSpace system that supports real-time workshop in Matlab [20]. It supports Simulink based control interfaces, and thus the PSAT signal could be directly relayed without an extra conversion toolkit. Through the dSpace I/O interfaces, BMS information is fed into the PSAT and the electric power signals are conveyed to the battery cycler. Fig. 1 illustrates the overall configuration of the B-HILS systems.

Fig. 1.Configuration of the developed B-HILS system

The drive profile can be generated via multiple methods. The original PSAT fetches the drive profile from an offline data file. In our B-HILS platform, real-time 3D driving mode and GPS mode are additionally implemented.

In the 3D driving mode, the drive profile is generated in real-time as the user drives a virtual vehicle on 3D simulator. For the realistic driving, actual drive gears such as steering wheel, accelerator and brake are provided. Figs. 2 and Fig. 3 depict the equipment and the 3D simulator, respectively. On the 3D simulator, the relative information such as vehicle speed, battery state-of-charge (SOC), pack voltage, battery current, maximum allowable power, and cell temperature is displayed. The battery parameters are real data received from the BMS of the battery while the vehicle information is calculated value by the vehicle model.

Fig. 2.Driving gears for real-time 3D driving mode.

Fig. 3.The display of 3D driving simulator with relevant information.

In GPS mode, the drive profile is generated from logged GPS data. The GPS data is converted into vehicle speed and altitude information and fetched into the PSAT in real-time in accordance with the GPS timestamp. As shown in Fig. 40, the trajectory of the drive is displayed on the Google satellite map. The GPS mode is useful when simulating an actual road situation such as traffic and terrain effect on a specific road.

Fig. 4.Trajectory of the drive in GPS mode on Google map.

During the simulation, the battery is placed in a chamber to emulate a specific temperature condition. The simulation facilities are illustrated in Fig. 50.

Fig. 5.From the left-upper image in the clockwise direction, temperature chamber, battery clcyler, and the 72 cell lithimum-ion battery pack tested in the latter part of the paper.

The relevant information is logged and selected items are displayed on the monitors. For example, Fig. 60 is a dash-board interface that displays vehicle speed, generator RPM, engine RPM, state-of-charge (SOC), current and voltage of the battery. The detailed data such as individual cell voltage are logged in a file for further analysis.

Fig. 6.A dash-board interface displaying vehicle speed, generator RPM, engine RPM, state-of-charge (SOC), current and voltage of the battery.

# 3. Experiments & Case Study

In this section, several experiments are carried out to show the features of B-HILS while investigating the performance of a SK’s commercial battery for mid-size hybrid sedan. The target vehicle model is second generation (2004-2009) Toyota Prius, for which the electric specifications are summarized in Table 10 [21]. The operation CAN code for the Prius are referred from the open data [22].

Table 1.Electric specifications of the target vehicle(2nd generation Prius, 2004-2009)

The experiments are carried out using the lithium-polymer (LiPo) battery from SK Innovation in comparison with the nickel-metal hydride (NiMH) battery, which is extracted from the target vehicle, or 2nd generation Prius. The specification of the LiPo battery can be found in Table 20. The dimension of the LiPo battery is almost same as the NiMH battery, but the energy density and the power capacity is far better than the NiMH. Note that the experiments in this section are conducted as parts of SK’s commercial battery development project for hybrid vehicle.

Table 2.Specifications of the target battery pack for the experiments

## 3.1 Real-time 3D driving simulation

In general, battery experiments are conducted using predetermined offline cycle data. In our B-HILS, however, the drive profile can be generated in real-time using a virtual driving environment. Battery is cycled as the user drives a virtual vehicle using driving gears such as steering wheel, accelerate, and the brake shown in Fig. 2.

Some simulation data logged during the 3D driving mode are illustrated in0. The target battery is the one described in Table 2, with BMS and cooling system. During the operation, BMS provides real pack information to the PSAT via CAN communication. The typical battery information includes pack voltage, cell temperature, pack current, SOC, and the maximum allowable power for the battery.

Based on the BMS data and the drive profile, PSAT generates the required electric power. For example, the allowable power limits enforce the hybrid control unit (HCU) to operate the battery within the given range. The impact of the power restriction is discussed in the following section. This kind of real-time interaction is especially important in developing and verifying the BMS logic.

## 3.2 Case study 1: Reactivity in low temperature

Usually, the internal resistance of a chemical battery increases drastically at low temperature and flowing large current under this circumstance often damages the cell [23-28]. Especially, lithium deposition (plating) as a mechanism for degradation is likely at low temperature for the Lithium-ion batteries. Therefore, as we discussed above, the battery power is restricted to protect the cell by the BMS at low temperature [29].

In order to investigate the effect of the power restriction, the experiment is performed assuming the driving in a winter morning. For the experiment, the LiPo pack in Table 2 is used after storing in a temperature chamber at minus 25 ℃ for 12 hours. Then, one and half cycles of the US06 drive cycle (around 15 minutes), which is designated to test at high speeds and aggressive driving conditions [30], is applied to the B-HILS. The initial SOC is set to 65%.

The US06 drive cycle and the resulting SOC are illustrated in Fig 80. As seen in Fig 70, the HCU of Prius tries to maintain the battery SOC at around 65% under a normal driving condition. However, the SOC during the low temperature experiment drops down to 30% as depicted in Fig 80. The cause of this phenomenon can be found in Fig 9. During the first four minutes, the maximum powers for both directions are restricted to less than 5 kW. Apparently, it seems that the vehicle is suffering from the shortage of the electric power; the battery is enforced to discharge at maximum rate during this period.

Fig. 7.Example of the test result with 3D driving interface. All the data are gathered from the real battery through the simluiation interface.

Fig. 8.Ttajectory of the state-of-charge for US06 driving cycle during the low temperature experiment (- 25 ℃)

Fig. 9.Power restriction and relevant battery parameters estimated during the low temperature experiment (- 25 ℃)

As the temperature exceeds minus 10 ℃, the power limit is mitigated rapidly and the HCU operates the battery in both directions. However, it still hits the limits occasionally, especially dragging at the charge limit frequently. This seems to arise from the aggressive acceleration of the drive cycle. That is, the HCU experiences shortage of the charging power more frequently than that of the discharging power.

After around 11 minutes, as the battery temperature exceeds 10 ℃, the restrictions are drastically lessened, especially in discharge direction. The battery power thus stays far less than the discharge limit. It is, however, not because the drive cycle turned mild but because the discharge limit exceeded the inverter size. Note that the battery in this experiment is SK’s LiPo battery, while the vehicle model is 2nd generation Toyota Prius which comes with the NiMH battery. From our analysis, the original NiMH battery provides the power limits in current level. During a normal operation, the maximum current is allowed up to 100 A for both directions. Since the voltage of the original pack is around 200 V, the inverter must have been sized to fit this level, around 20 kW. Therefore, the discharge limit stays around 20 kW despite the higher capability of the battery power. As a result, it could be concluded from the HIL experiment that the tested battery is oversized for this vehicle, especially in discharge direction. The inverter size as well as HCU logic thus should be tuned to fully utilize the new battery.

## 3.3 Case study 2: Fuel efficiency vs battery size

The second experiment is to investigate the relationship between fuel efficiency and battery size. The experiment is performed for two batteries with different power limits. Both batteries are physically identical, but the available power limits are configured differently by manipulating the BMS parameters. Specifically, the power limits for the high power battery is set to 16 kW and 20 kW for charging and discharging, respectively, which is the same as the original NiMH battery of the target vehicle, or Prius. The other one, the low power battery, is configured to have half of those of the high power battery, or 8 kW for charging and 16 kW for discharging. The experiments are carried out at normal temperature (25 ℃) with the initial SOC of 65 %. For the drive profile, two cycles of Urban Dynamometer Driving Schedule (UDDS) [31] is applied. Other vehicles relative parameters are same as the previous experiment.

Fig 100 illustrates the trajectory of the electric power actually assigned to the batteries. As marked with circles, the low power battery pulls out its maximum power and drags in many places. On the other hand, the high power battery operates within far less than the maximum power seldom hitting the limit. In fact, as shown in 0Fig. 11, the energy requirement of the vehicle is almost same for the both batteries. Nevertheless, due to the lack of the power, the low power battery tends to drag at the maximum power, and hence results in the higher deviation of the state-of-charge (SOC). The SOC deviation is illustrated in Fig. 120; the maximum SOC deviation for the low power battery is 10%, while that of the high power battery is only 6%.

Fig. 10.The trajectory of electric power

Fig. 11.The trajectories of accumulated energy

Fig. 12.The trajectory of SOC

The results imply that the battery size may not properly match the motor size, at least for this type of drive pattern. For the low power battery, the battery size seems to be small for the given motor since it frequently suffers the power shortage. Conversely, the high power battery is operated at around half of its power capability being rarely required the maximum power. Actually, the drive cycle used in this experiment (UDDS) is comparably mild since it merely represents the urban driving. Thus, additional tests are required to exactly judge if the battery is oversized.

As a measure of the performance, we estimated the fuel efficiency of the both experiments. From Fig. 130, the fuel efficiency of the high power is estimated at around 31 km/l. With the low power battery, however, the fuel efficiency deteriorates more than 30% marking around 21 km/l. Obviously, the proper sizing of the battery appears as a significant factor in terms of the vehicle performance. In practice, however, performance is not the only factor to determine the battery size. Many other parameters such as profitability should be considered simultaneously. Yet, the result of this experiment is important in terms of determining the physical matching of the battery with the motor.

Fig. 13.The trajectory of fuel efficiency

## 3.4 Case study 3: Harsh driving test using GPS mode

As introduced in section 2, our B-HILS system supports a GPS mode. In this mode, actual GPS log is fetched and converted into the speed and the gradient information, from which the cycle profile of the battery is generated for a given vehicle model.

In this experiment, the GPS mode is utilized to investigate the reactivity of the batteries under harsh driving condition. Using a similar size vehicle with the target model, actual drive profile was logged by the GPS along the route illustrated in Fig. 140 that lies on the Mt. Jiri, one of the steepest mountains in South Korea. The GPS data is then fed into the B-HILS with a vehicle model of 2nd generation Prius. The experiment was performed for two different batteries; original NiMH battery and SK’s LiPo battery described in Table 2, for which the results are depicted in 0 and Fig. 16, respectively.

Fig. 14.The drive route of the test cycle

During the first 30 minutes, the vehicle climbs uphill, and thus the battery is likely to be used for propulsion rather than the regeneration. Consequently, the SOC in both tests continues to decrease until the vehicle reaches the top of the mountain. During this period, the power variance of NiMH battery is higher than that of the LiPo battery. The cause seems to be the wider variation of the battery voltage in NiMH due to the higher internal resistance. The SOC variation is, however, higher in the NiMH due to the smaller energy capacity. Specifically, from 0Table 1 and Table 2, the energy capacity of the original Prius pack is around 1.3 kWh, while the SK’s LiPo pack is around 2.0 kWh. For the same reason, the SOC increase by the regeneration during the idle period, between 30-50 minute, is higher in the NiMH battery.

A noticeable difference occurs during the downhill area. Generally, batteries are designed to have higher capability in discharge. Therefore, a typical battery experiences higher stress when charged at the same power. As seen in 0Fig. 15, the absolute value of the charging power during the downhill is not significantly higher than the discharging power during the uphill. For the aforementioned reason, however, the battery temperature in NiMH battery increases rapidly during the downhill where the battery is mainly charged for regeneration. As the temperature exceeds 55 ℃, BMS suspends the battery power to protect the cell and re-permits it only after the temperature drops below 50 ℃. On the contrary, the LiPo battery in 0Fig. 16 does not experience the temperature increase, and hence could continue the service during the entire downhill. Note that the pause of the power during 75-80 minute is not because of the power suspension. The vehicle had actually stopped the driving and stayed idle during this period due to the traffic.

Fig. 15.Battery data during harsh driving for original NiMH pack

Fig. 16.Battery data during harsh driving for SK’s LiPo pack

It is also noted that the SOC of the NiMH battery decreases during the period of the power suspension. It is because the BMS is trying to reset the SOC based on the open circuit voltage (OCV) during the idle period. This is a common technique used for a vehicular BMS to correct the accumulated SOC error using the OCV value [32-34]. The large SOC deviation, say around 10%, of the NiMH battery after an hour driving reveals the poor performance of the SOC estimation. Usually, the maximum SOC error allowed for an electric vehicle (EV) supporting all-electric driving is less than a few percent [35]. Thus, the results here prove the inadequacy of using the NiMH battery for an all-electric EV.

# 4. Conclusions

A battery manufacturer often faces absence of the test platform until actual vehicle platform is ready for the integrated test. Under a commercial mass production system, it open leads to a critical cost increase due to the unexpected trouble handling during the integration phase. To resolve the problem, we developed a B-HILS system equipped with various features. To the author’s best knowledge, battery HILS supporting other than the fixed drive profile such as real-time 3D driving mode has not yet been developed. Exploiting the developed system, the characteristics of the SK’s LiPo battery could be quantitatively assessed, especially in comparison with the original NiMh pack of the 2nd generation Toyota Prius. The system is currently being updated to include other electric components such as motor and invertor in the control loop.

#### References

1. R. J. Patton, P. M. Frank, and R. N. Clark, Issues of Fault Diagnosis for Dynamic Systems. London, U.K.: Springer-Verlag, 2000.
2. J. Gerler, M. Costin, Xiowen Fang, Z. Kowalczuk, M. Kunwer, R. Monajemy, "Model based diagnosis for automotive engines-algorithm development and testing on a production vehicle", IEEE Trans. on Control Systems Technology, vol. 3,no. 1, pp. 61-69, 1995. https://doi.org/10.1109/87.370711
3. L. Jinming, "Modeling and Control of a Power-Split Hybrid Vehicle", IEEE Trans. on Control Systems Technology, vol. 16, no. 6, pp. 1242-1251, 2008 https://doi.org/10.1109/TCST.2008.919447
4. B. Lu, X.Wu, H. Figueroa, and A.Monti, "A low-cost real-time hardwarein-the-loop testing approach of power electronics controls," IEEE Trans . Ind. Electron., vol. 54, no. 2, pp. 919-931, Apr. 2007. https://doi.org/10.1109/TIE.2007.892253
5. W. Lee, M. Yoon, and M. Sunwoo, "A cost- and timeeffective hardwarein-the-loop simulation platform for automotive engine control systems,"J. Automobile Eng., vol. 217, no. 1, pp. 41-52, 2002. https://doi.org/10.1243/095440703762702969
6. J. Du, Y.Wang, C. Yang, and H.Wang, "Hardware-inthe-loop simulation approach to testing controller of sequential turbocharging system," in Proc. 2007 IEEE Int. Conf. Autom. Logistics, Aug. 2007, pp. 2426- 2431.
7. K. B. Wipke, M. R. Cuddy, and S. D. Burch, "ADVISOR 2.1: A user-friendly advanced power train simulation using a combined backward/forward approach," IEEE Trans. Vehicular Technol., vol. 48, no. 6, pp. 1751-1761, Nov. 1999. https://doi.org/10.1109/25.806767
8. T. Markel, A. Brooker, T. Hendricks, V. Johnson, K. Kelly, B. Kramer, M. O'Keefe, S. Sprik, and K.Wipke, "ADVISOR: A systems analysis tool for advanced vehicle modeling," J. Power Sources, vol. 110, no. 2, pp. 255-266, Aug. 2002. https://doi.org/10.1016/S0378-7753(02)00189-1
9. PSAT Documentation. Available at http://www.transportation.anl.gov/; Oct. 2012.
10. PSIM Website. Available at http://www.powersimtech.com/; Oct. 2012.
11. VTB Website. Available at http://vtb.ee.sc.edu/; Oct. 2012.
12. B. Powell, K. Bailey, and S. Cikanek, "Dynamic modeling and control of hybrid electric vehicle powertrain systems," IEEE Contr. Sys. Mag., vol. 18, no. 5, pp. 17-22, Oct. 1998. https://doi.org/10.1109/37.722250
13. C. C. Lin, Z. Filipi, Y. Wang, L. Louca, H. Peng, D. Assanis, and J. Stein, "Integrated, feed-forward hybrid electric vehicle simulation in Simulink and its use for power management studies," in Proc. SAE 2001 World Congr., Detroit, MI, Mar. 2001.
14. K. L. Butler, M. Ehsani, and P. Kamath, "A Matlabbased modeling and simulation package for electric and hybrid electric vehicle design," IEEE Trans. Vehicular Technol., vol. 48, no. 6, pp. 1770-1118, Nov. 1999. https://doi.org/10.1109/25.806769
15. G. Rizzoni, L. Guzzella, and B. M. Baumann, "United modeling of hybrid electric vehicle drivetrains," IEEE Trans. Mechatronics, vol. 4, no. 3, pp. 246-257, 1999. https://doi.org/10.1109/3516.789683
16. X. He and J. W. Hodgeson, "Modeling and simulation for hybrid electric vehicles, I. Modeling," IEEE Trans. Intelligent Transportation Syst., vol. 3, no. 4, pp. 235-243.
17. X. He and J. Hodgson, "Modeling and simulation for hybrid electric vehicles-Part II," IEEE Trans. Transportation Syst., vol. 3, no. 4, pp. 244-251, Dec. 2002.
18. H. Dai, X. Wei, Z. Sun, J. Wang, "Measuring Technology and Mechatronics Automation in Electrical Engineering", Lecture Notes in Electrical Engineering. Vol. 135, pp 27-36, Springer, 2012
19. N. Shidore, H. Lohse-Busch, R. W. Smith, T. Bohn, P. B. Sharer, "Component And Subsystem Evaluation In a Systems Context Using Hardware In the Loop", Proc. Of 2007 IEEE Vehicle Power and Propulsion Conference, Arlington, TX, September 09-12, 2007.
20. dSpace Hadware-in-the-Loop Simlation. Available at http://www.dspace.com/en/inc/home/applicationfields/electric_drives/hardware_in_the_loop_simulatio.cfm; Oct. 2012.
21. Specifications of 2nd generation Toyota Prius. Available at http://www.new-cars.com/2004/toyota/toyota-prius-specs.html; Oct. 2012
22. Operation CAN code of 2nd generation Prius: http://www.vassfamily.net/ToyotaPrius/CAN/PriusCodes.xls
23. D. Szente-Varga, G. Horvath, M. Rencz, "Thermal characterization and modelling of lithium-based batteries at low ambient temperature", in Proc. 14th Internation Workshop on Thermal Investigation of ICs and Systems, Rome, Italy, 24-26 Sept. 2008
24. D. Szente-Varga, Gy. Horvath, M. Rencz: "Creating temperature dependent Ni-MH battery models for low power mobile devices", in Proc. 14th Internation Workshop on Thermal Investigation of ICs and Systems, Nice, France, 2006.
25. D. Szente-Varga, Gy. Horvath, M. Rencz: Ni-MH battery modeling for ambient intelligence applications, Proceedings of DTIP 2007, pp. 332-337., Stresa, Lago Maggiore, Italy, 25th-27th Apr 2007
26. M. Doyle, T. F. Fuller, and J. S. Newman, "Modeling of galvanostatic charge and discharge of the lithium/polymer/insertion cell," J. Electrochem. Soc., vol. 140, no. 6, pp. 1526-1533, 1993. https://doi.org/10.1149/1.2221597
27. T. F. Fuller,M. Doyle, and J. S. Neuman, "Simulation and optimization of the dual lithium ion insertion cell," J. Electrochem. Soc., vol. 141, no. 1, pp. 1-10, 1994. https://doi.org/10.1149/1.2054684
28. K. Uthaichana, R.A. DeCarlo, L.H. Tsoukalas, "Development and Validation of a Battery Model Useful for Discharging and Charging Power Control and Lifetime Estimation", IEEE Trans. on Energy Conversion, vol. 25, no. 3, pp. 821-835, 2010 https://doi.org/10.1109/TEC.2010.2043106
29. M.C. Sart, B.V. Ratnakumar, L. Whitcanack, K. Chin, M. Rodriguez, S. Surampudi, "Performance characteristics of lithium ion cells at low temperatures", IEEE Aerospace and Electronic Systems Magazine, vol. 17, no. 12, pp. 16-20, 2002
30. Emission Test Cycles: SFTP-US06. Available at http://www.dieselnet.com/standards/cycles/ftp_us06.php; Oct. 2012
31. Emission Test Cycles: FTP-72 (UDDS). Available at http://www.dieselnet.com/standards/cycles/ftp72.php; Oct. 2012
32. D. Haifeng, W. Xuezhe, S. Zechang, "State and Parameter Estimation of a HEV Li-ion Battery Pack Using Adaptive Kalman Filter with a New SOCOCV Concept", in Proc. Measuring Technology and Mechatronics Automation, pp. 375-380, Zhangjiajie, China, 11-12, April, 2009
33. S. Piller, M. Perrin, A. Jossen, "Methods for state-ofcharge determination and their applications." Elsevier. J. Pwr. Sources, vol. 116, pp. 110-117.
34. C. C. Chan, E. W. C. Lo, S. Weixiang, "The available capacity computation model based on artificial neural network for lead acid batteries in electric vehicles." Elsevier. J. Pwr. Sources, vol. 87, pp. 201-204. https://doi.org/10.1016/S0378-7753(99)00502-9
35. Tremblay, O., Dessaint, L.-A. "Experimental Validation of a Battery Dynamic Model for EV Applications." World Electric Vehicle Journal. Vol. 3 -ISSN 2032-6653-${\copyright}$ 2009 AVERE, EVS24 Stavanger, Norway, May 13 - 16, 2009.

#### Cited by

1. Novel performance of ultrathin AlPO4 coated H2Ti12O25 Exceeding Li4Ti5O12 in cylindrical hybrid supercapacitor vol.273, 2015, https://doi.org/10.1016/j.jpowsour.2014.09.090
2. Fabrication and electrochemical properties of cylindrical hybrid supercapacitor using H2Ti12O25 as anode material vol.143, 2015, https://doi.org/10.1016/j.matlet.2014.12.069