MedicineLakex

Topic: a, b, c, d, e or f

The Second TSME International Conference on Mechanical Engineering 19-21 October, 2011, Krabi Investigation of Load Profiles of Lithium-ion batteries
for Electric Vehicle Applications
Chadchai Srisurangkul National Metal and Materials Technology Center, 114 Thailand Science Park, Phahonyothin Rd. Klong 1, Klong Luang, Pathumthani 12120 Thailand *Cor esponding Author: Tel: +66 2 564 6500 ext. 4731, Fax: +66 2 564 6332, E-mail: [email protected] Abstract

Vehicle electrification become and continues to be the major trend in the automotive industry and a main part of this is batteries. Unfortunately the battery testing presently does not represent the real-
world condition, in which the bat ery load profile comprises complex charge-discharge wave. Therefore in
this study, the load profile of bat eries for electric vehicles was investigated. An integrated simulation and
testing approach is used to predict the bat ery load. In simulation the bat ery load profile is calculated
based on the utilization of vehicle and battery model. The present vehicle and bat ery model are
developed using a vehicle dynamics technique and equivalent circuit modeling technique respectively,
whereas the model development is focused on a specific vehicle which converts conventional propulsion
into pure electric propulsion. The results are demonstrated through the current profile of advanced
bat eries subjected to standard driving cycles. With this approach, the battery loads correspond to actual
driving condition or any driving cycle can be predicted and can be used for bat ery testing. This creates a
new aspect to evaluate advanced bat eries for power train application.
Keywords: Current Profile, Vehicle Modeling, Bat ery Modeling, Driving Cycle
1. Introduction
some bat ery parameters but still not suite to All major car manufacturers are planning investigate, for example, the bat ery life or heat to bring out electric cars in the near future. This generation. It is not suitable to use as tool for trend drives the development of high power design bat ery subsystem such as bat ery bat eries such as lithium-ion bat eries. Although thermal management. lithium-ion bat eries are widely used for electric This paper describes an approach to vehicles, there is no testing standard for those determine the bat ery load profile comprising bat eries [10]. The bat ery test profiles presently complex charge-discharge wave from real-world base on constant current or step response technique which does not represent the real- world condition. They may be used to investigate

The Second TSME International Conference on Mechanical Engineering 19-21 October, 2011, Krabi 2. Vehicle Dynamics Modeling
speed, x is the vehicle speed, and Jred is the The vehicle dynamic model can be reduced mass moment of inertia at wheel. derived from Newton's second law considering The vehicle considered in this work is all the forces applied upon the vehicle. It should the two year used HONDA Jazz that will be be noted that only the longitudinal dynamic is converted to a pure electric vehicle. The existing sufficient to investigate the bat ery load profile. engine will be replaced with a 40kW new With this the complex and uncertain parameters developing motor. The vehicle uses a 100 cell as well as the simulation time can be reduced. lithium-ion bat ery (type: LiFeYPO4) as power The total resistance forces acting on the source. The total capacity is 60Ah. The vehicle vehicle are rolling resistance, aerodynamic uses the original transmission modified to use resistance, and grade resistance force, as the 3rd gear ratio. The vehicle specification is demonstrated in Fig. 1. shown in Table 1. Table. 1 Vehicle specification Value/Description 2 Weight distribution 62/38 (%) 4 Drag coef icient Fig. 1 Resistance force acting on the Hence, the vehicle dynamic equation can be obtained as 3.461, 1.869, 1.303 F  F  F  F  F Eq. (1) 8 Final drive ratio F  mg( f  f x)sgn(x)  mgsin  A (x  x )2 sgn(x)  (m continuous power roll is rolling resistance, Fad is the aerodynamic resistance, F resistance, m is the vehicle mass, g is the 12 Battery Capacity 60 (Ah) natural acceleration,  is the angle of grade, f 13 Number of cells 14 Battery weight 1 are the rolling coef icients,  is the air density, Cd is the aerodynamic drag coef icient, Af is the vehicle frontal area, x is the wind simulations there are various types of mathematical models available. In this work the

The Second TSME International Conference on Mechanical Engineering 19-21 October, 2011, Krabi vehicle model was created by using a transient effort-flow refer to [1]. Ef ort-flow in the model refers to the combinations of torque/angular speed, force/velocity and voltage/current. The basic configuration of the electric vehicle used for simulation is illustrated in Fig. 2. Overall system is modeled in a MATLAB/ Simulink environment. Fig. 3 Efficiency map of the electric motor 3. Battery Modeling
3.1. Battery Model
A bat ery model calculates the bat ery variables like current, voltage, and the battery temperature. The bat ery is modeled as an equivalent circuit with a bat ery internal resistance, as shown in Fig. 4. The model uses Fig. 2 Electric vehicle configuration data from experiment described in the next Vehicle system models typically contain a mix of empirical data, engineering assumptions, and physical based parameter and algorithms. Driven by the need for fast simulation times, complex components such as motor are typically simulated using empirical data in form of lookup maps. For example, the electric motor is modeled using an efficiency map handled as a 2-D lookup table indexed by Fig. 4 Bat ery Model rotor speed and torque. This map is generally The model consists of voltage source obtained via experiments. In our case the generating an open-circuit voltage (OCV) and a efficiency map of electric motor is unknown resistor representing the internal resistance (R). because the motor is presently constructing. Both parameters can be described as a function Therefore the scaling of torque to provide the of state of charge (SOC) and temperature. The required maximum power has been done. This internal resistance is considered separately type of scaling is valid only for the same motor between charge and discharge process. type and in the neighborhood near the actual The state of charge was estimated by parameter. The efficiency map of the electric coulomb counting (current based SOC motor is shown in Fig. 3. estimation). The SOC is obtained by measuring the current flowing into and out of a battery and

The Second TSME International Conference on Mechanical Engineering 19-21 October, 2011, Krabi integrating this current over time and then The system consists mainly of power subtracting it from the charge in a fully charged supply and electronic load which connected to bat ery [4]. The other methods of SOC PC via GPIB. The output voltage range and the determination can be found in [5]. maximum current rating of those machines are The temperature is obtained by using 0-60Vdc and 120A respectively. Labview was bat ery thermal model which describes the heat used to define the test procedure and log data. transfer process. The heat is generated from The tested specimen is a 5 cell series bat ery due to electrochemical reactions and connected LiFeYPO4 with a capacity of 60 Ah. resistive heating which causes increased bat ery The bat ery core is covered by poly-propylene temperature [6]. In practice the heat dissipation case which has low thermal conductivity. During is best estimated from values of the current and the test the bat ery was placed in a temperature- voltage [7], according to control cabinet. The test cycles is shown in Fig. where V is the instantaneous cell potential, Voc is the open-circuit voltage. The heat dissipation is conducted from the internal bat ery to the bat ery case and then convected from the case's surface to the air. 3.2. Experiment Setup
To characterize battery parameters for its model, the step response technique [8] was employed. The tests were carried out on the Deep of Discharge, DoD (%) testing system as shown in Fig. 5. Fig. 6 Bat ery test cycle as function of DoD Fig. 7 Bat ery test cycle as function of time The test started with the discharge Fig. 5 Testing system configuration process. For each test cycle the constant current of 0.5C (30A) is applied and was paused for one The Second TSME International Conference on Mechanical Engineering 19-21 October, 2011, Krabi minute at every 10% change in state of charge. 4. Simulation
Measurements were taken every 200ms. The 4.1 Driving Cycle and Battery Load Cycle
test was performed at temperature of 5°C, 25°C Through the battery load profile of real- and 40°C. Some initial cycles were applied to world driving cycles is the aim of this work, the ensure that the bat ery reaches stable behavior. standard driving cycles with transient The voltage response during the pause characteristic was used to test our model. The period is used in order to determine the battery transient cycles give more dynamic changing in internal resistance. An example of the voltage load and are based upon real-world data. UDDS, response during pause period is shown in Fig. 8. NYCC and CSC cycles are three examples of transient driving cycle used to define the bat ery load profile. They were chosen because they are based upon real-world driving and content a high portion of with non-aggressive urban and suburban routes that are suitable for electric vehicles. UDDS cycle describes an urban route with cold start and transient phase, NYCC cycle represents an urban route through New York Fig. 8 Example of the voltage response and CSC cycle describes a city suburban route with lower average speed than UDDS cycle. Fig. 1, V2, and V3 in Fig. 8 are easily measured. Both the resistance and the open-circuit voltage 10 shows UDDS cycle as an example. (OCV) are assumed to be constant over the pulse period. Voltage dif erent at the beginning of the pulse is also assumed at the end of the pulse. The internal resistance after analysis the test data is shown in Fig. 9. For other methods for resistance estimation can be found in [11]. Fig. 10 An example UDDS cycle A bat ery load profile used by the vehicle can be determined by using one of these driving cycles as an input to the simulation. The load profile results are shown in Fig. 11-13. The simulation was set up with a full battery and a temperature of 25°C. Deep of Discharge, DoD (%) Fig. 9 Internal resistance in case of discharge The Second TSME International Conference on Mechanical Engineering 19-21 October, 2011, Krabi the vehicle dynamic simulation must be filtered while still maintaining the typical load characteristic. The filter model with filter and stair-step function was developed for this purpose. It performs an online analysis to meet the requirement of hardware in the loop simulation in the future. Fig. 14 shows a close up view of a portion of the filtered and unfiltered Fig. 11 Battery load profile, UDDS load profile for UDDS cycle. Fig. 14 Unfiltered and filtered battery load profile Fig. 12 Bat ery load profile, NYCC The peak current of unfiltered and filtered profile are 95.06 A and 96 A, while the average current are 18.85 A and 18.64 A respectively. This guarantees correctness and stability of the filter model. Fig. 15-17 shows the distribution of the filtered bat ery load profile for all three driving Fig. 13 Bat ery load profile, CSC 4.2 Cycle Filtration
In the practice the high dynamic load cannot use directly due to the limited capability of bat ery testing machines. To enable the use of widely available battery testing machines the minimum sampling period of load profile for input to the testing machine should be 2 seconds and the profile should be rectangular wave. To meet Filtered battery current (A) these requirements the battery load profile from Fig. 15 Load distribution of UDDS cycle The Second TSME International Conference on Mechanical Engineering 19-21 October, 2011, Krabi been done to provide input parameters for the The vehicle dynamic model and the bat ery model are a core of this vehicle system simulation. In the vehicle dynamic model only the longitudinal dynamic is sufficient to investigate the battery load profile, while an equivalent circuit with an internal resistance is considered in the bat ery model. Filtered battery current (A) The experiment has been conducted to Fig. 16 Load distribution of NYCC cycle get the bat ery properties. The test methodology relies on the current pulse relaxation technique to determine the battery characteristics. Accordingly, using this methodology, the bat ery load profile correspond to actual driving condition or any driving cycle can be predicted and can be used for bat ery testing. This creates a new aspect to evaluate advanced batteries for power train application. 6. References
Filtered battery current (A) Fig. 17 Load distribution of CSC cycle [1] Gao, D.W., Mi, C. and Emadi, A. (2007). All three driving cycles have the most Modeling and Simulation of Electric and Hybrid common current value with in the same range. Vehicles, Proceedings of the IEEE, vol.95(4), For UDDS cycle the bat ery current varies widely April 2007, pp. 729 – 745. between -15 and 96 A, whereas the most [2] Johnson, V.H., Pesaran, A.A. (2000). common value lies between 2 to 6 A. NYCC and Temperature-Dependent Battery Models for CSC cycles show narrower current distribution High-Power Lithium-Ion Batteries, paper than the UDDS cycle. The vehicle takes current presented in the 17thElectric Vehicle Symposium, from the bat ery at 2 to 6 A through about half of Montreal, Canada. [3] Johnson, V.H. (2002). Battery performance 5. Conclusion
models in ADVISOR, Journal of Power Sources, The present paper examined the bat ery vol.110, pp. 321 – 329. load profile of a vehicle driving in the real world. [4] Bergveld, H.J., Kruijt, W.S., Notten, P.H.L. To predict the bat ery load a vehicle system (2002). Battery Management Systems: Design by simulation was developed while many tests have Modelling, Philips Research Book Series Volume 1, ISBN: 1402008325, Springer The Second TSME International Conference on Mechanical Engineering 19-21 October, 2011, Krabi [5] Pop, V., Bergveld, H.J., Danilov, D., Regtien, [9] Duvall, M.S. (2005). Battery Evaluation for P.P.L., Notten, P.H.L. (2008). Battery Plug-In Hybrid Electric Vehicles, Vehicle Power Management Systems: Accurate State-of-Charge and Propulsion, IEEE Conference, pp. 338 – Indication for Battery-Powered Applications, Philips Research Book Series Volume 9, ISBN: [10] Morita, K., Akai, M., Hirose, H. (2009). 1402069448, Springer Development of Cycle Life Test Profiles of [6] Pesaran, A.A. (2002). Battery thermal models Lithium-ion Batteries for Plug-in Hybrid Electric for hybrid vehicle simulations, Journal of Power Vehicles, EVS24 International Battery, Hybrid Sources, vol.110, pp. 377 – 382. and Fuel Cell Electric Vehicle Symposium [7] PCM Thermal Control of Nickel-Hydrogen [11] Schweiger, H.G., Obeidi, O., Komesker, O., Batteries, Division of Technical Services and Raschke, A., Schiemann, M., Zehner, C., Planning, PL-TR-93-1075, Energy Science Gehnen, M., Keller, M., Birke, P. (2010), Laboratories Inc., Final Report 1993 Comparison of Several Methods for Determining [8] Doerffel, D. (2007). Testing and the Internal Resistance of Lithium Ion Cells, Characterisation of Large High-Energy Lithium- Sensors 2010, 10, ISSN 1424-8220, pp. 5604 – Ion Batteries for Electric and Hybrid Electric Vehicles, PhD Thesis, University of Southampton

Source: http://rcme.engr.tu.ac.th/TSME-ICoME%20Web/full_paper_file/ETM/ETM18%20920-1611-1-RV.pdf