See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/282227917

Mathematical Modeling and Simulation of an Electric Vehicle

Article in International Journal of Mechanical Sciences · July 2015

DOI: 10.15282/jmes.8.2015.6.0128

CITATIONS READS

71 17,349

3 authors:

Tengku Azman Tengku Mohd Mohd Khair Hassan

Politeknik Kuala Terengganu Universiti Putra Malaysia
13 PUBLICATIONS 126 CITATIONS 136 PUBLICATIONS 1,343 CITATIONS

SEE PROFILE SEE PROFILE

Wmk Wmk
Northeast Forestry University
4 PUBLICATIONS 108 CITATIONS

SEE PROFILE

All content following this page was uploaded by Tengku Azman Tengku Mohd on 16 February 2016.

The user has requested enhancement of the downloaded file.

Journal of Mechanical Engineering and Sciences (JMES)
ISSN (Print): 2289-4659; e-ISSN: 2231-8380; Volume 8, pp. 1312-1321, June 2015
© Universiti Malaysia Pahang, Malaysia
DOI: http://dx.doi.org/10.15282/jmes.8.2015.6.0128

MATHEMATICAL MODELING AND SIMULATION OF AN ELECTRIC

VEHICLE

T.A.T. Mohd1, M.K. Hassan1,2 and WMK. A. Aziz1

1
Department of Electrical and Electronic Engineering
Faculty of Engineering, Universiti Putra Malaysia
43400 Serdang, Selangor, Malaysia
Email: [email protected]
Phone : +603-89464367; Fax : +603-89466327
2
Institut Teknologi Maju, Universiti Putra Malaysia
43400 Serdang, Selangor, Malaysia
Email: [email protected]

ABSTRACT

As electric vehicles become promising alternatives for sustainable and cleaner energy
emissions in transportation, the modeling and simulation of electric vehicles has
attracted increasing attention from researchers. This paper presents a simulation model
of a full electric vehicle on the Matlab-Simulink platform to examine power flow during
motoring and regeneration. The drive train components consist of a motor, a battery, a
motor controller and a battery controller; modeled according to their mathematical
equations. All simulation results are plotted and discussed. The torque and speed
conditions during motoring and regeneration were used to determine the energy flow,
and performance of the drive. This study forms the foundation for further research and
development.

Keywords: Mathematical modeling; simulation; electric vehicle; Matlab-Simulink.

INTRODUCTION

Environmental concerns and energy issues have led to the mass transfer of effort in the
automotive industry from the internal combustion engine (ICE) vehicle to an electrical
vehicle (EV) as the prime source of transportation. The issues of air pollution in urban
areas, ICEs being the second highest contributor to global warming with approximately
21% of greenhouse gasses emissions [1], and the depletion of fossil fuels and their
increasing prices, have significantly amplified interest in EVs [2-4]. This is due to the
advantages of EVs as clean and silent technology, as well as their offering better
efficiency than ICE vehicles and electricity being a cheaper energy source than fuel [5].
Numerous simulation and modeling package have been developed to study the
operation of electric and hybrid powered trains, such as CarSim from AeroVironment
Inc., SIMPLEV from the DOE’s Idaho National Laboratory, MARVEL (Argonne
National Laboratory), V-Elph (Texas A&M University) and ADVISOR (DOE’s
National Renewable Energy Laboratory, US). Butler, Ehsani, & Kamath [6] in their
study present a four-vehicle drive train vehicle modeling, simulation, and analysis
package for an EV, parallel hybrid EV (HEV), series HEV, and conventional ICE, using
Matlab-Simulink. The aim is to investigate fuel economy, efficiency and emissions. The

1312
 Mohd et al. / Journal of Mechanical Engineering and Sciences 8(2015) 1312-1321

use of visual programming allows the user to quickly modify parameters, architectures
and graphically examine the output data.
Much effort has been put in by past researchers in modeling an EV [7-12]. The
model by Husain & Islam [7] focuses on the electric propulsion unit and the drive
evaluation to meet performance desires using a switched reluctance motor (SRM). The
study showed the effective use of computer tools in the preliminary design stage of an
EV. A popular vehicle modeling package written in Matlab-Simulink, ADVISOR has
been promoted by Markel et al. [8]. They clarify that ADVISOR is a tool for evaluating
and quantifying the vehicle level impacts of the advanced technologies applied to
vehicles. Simulations and analysis of a series hybrid EV (SHEV) using the correct
matching of the vehicle powertrain was carried out by Jiang-Wen & Liang [9] on the
ADVISOR platform. The results satisfied the vehicle performance requirements and
improved the vehicle driving range. A study by Kaloko, Soebagio, & Purnomo [10]
examined the performance of small EVs and analyzed the power flow of required
electrical energy. A Matlab-Simulink model was developed in order to identify the best
power flow for the EV. The driving range and battery usage are determined from the
required battery capacity and EV specifications. A comprehensive study of modeling
full EVs (FEVs) or battery EVs (BEVs) was undertaken by Schaltz [11] and Luigi &
Tarsitano [12]. Schaltz [11] modeled and designed a BEV which fulfilled both the
power and energy requirements for a given driving cycle while Luigi & Tarsitano [12]
developed a simulation model comprising multidisciplinary domains such as the
electric, mechanical, thermal, power electronic, electrochemical and control domains.
Both studies used the iterative process in designing the procedures and all component
losses were included to attain realistic energy calculations for the vehicle. In the next
section, the modeling of the EV will be thoroughly explained. Then, the simulation
results will be presented and discussed while the final section concludes the study.

ELECTRIC VEHICLE MODELING

For modeling purposes, the recommended EV drive train is as shown in Figure 1. The
drive train consists of six components: the electrical motor, power electronics, battery,
motor controller, battery controller and vehicle interface. The vehicle interface provides
the interface for the sensors and controls which communicate with the motor controller
and battery controller. The motor controller normally controls the power supplied to the
motor, while the battery controller controls the power from the battery. The battery is
for energy storage, usually lithium-ion cells which provide more than 200 V and high
current to the power electronics. The power electronics manipulate the voltage, current
and frequency provided to suit the motor requirements.
By considering both directions of operation (clockwise and anti-clockwise) and
both modes (acceleration and deceleration), the motor’s operation can be described in
four quadrants of operation. This can be visualized by plotting the motor speed and the
applied torque on the x–y axis as shown in Figure 2. The drive train is in motoring mode
when the speed and torque values having the same polarity (1st Quadrant & 3rd
Quadrant), and in regenerating mode when the speed and torque values differ in polarity
(2nd Quadrant & 4th Quadrant). In the 1st Quadrant, with both positive polarities, the
motor moves forward, but in the 3rd Quadrant, the motor moves backward. In the 2nd
Quadrant, when the torque is positive and speed is negative, the motor is decelerating
returning energy to the battery in reverse braking, while in the 4th Quadrant, the energy
returns to the battery during forward braking [13]. The battery energy is decreased

1313
 Mathematical modeling and simulation of an electric vehicle

during motoring mode, but is increased in regenerating mode during regenerative

braking when the motor is operating as a generator.

Power
Battery Electronics Motor

Battery Motor
Controller Controller

Vehicle Interface

Figure 1. EV Drive Train [14].

Torque, T

Quadrant 2 Quadrant 1

Reverse Forward
T (+ve) T (+ve)
Regeneration Motoring
N (–ve) N (+ve)
Speed, N
T (–ve) T (–ve)
Reverse Forward
N (–ve) N (+ve)
Motoring Regeneration
Quadrant 3 Quadrant 4

Figure 2. Four Quadrant Drive Operation.

To model an EV, all mathematical equations to represent each component in the

EV drive train were determined. The motor, battery, motor controller and proportional-
integral (P-I) controller were modeled on the Matlab-Simulink platform into individual
block diagrams to form an EV drive system using the following equations. For a DC
motor, the torque developed in the motor, Td is proportional to the armature current, Ia;

Td = Km · Ia (1)

where, Km is the motor constant depending on its winding construction.

Voltage developed in the motor, Vd is proportional to armature speed, ωd;

Vd = Km · ωd (2)

1314
 Mohd et al. / Journal of Mechanical Engineering and Sciences 8(2015) 1312-1321

Voltage at the high side of the motor (terminal voltage), VH is given by;

VH = IH · Ra + LH · di(t)/dt + Vd (3)

where, IH is the current at the high side (terminal current), Ra is the armature resistance
value, and LH is the inductor value at the high side.
By assuming that there is no friction loss and no inertia loss, the electrical torque
developed, Td, is equal to the output mechanical torque, Tmech. Hence, the developed
electrical power is equal to the developed mechanical power. A simple motor controller
is used to maintain the input power equal to the output power. The controller is assumed
to be ideal with zero loss and no time lag.

High side voltage (input),

VH = K · VL (4)

High side current (input),

IH = (1/K) · IL (5)

where K is the controller gain value, VL is voltage at the low side (output), and IL is
the current at the low side (output).
The battery is modeled as the voltage source, EB and internal power loss in the
battery resistance, RA.

VL = IL · RA + EB (6)

The required battery’s internal voltage is calculated using the current and
voltage from the motor controller. The difference between the calculated EB (EB
(calculated)) and the actual EB (EB (actual)) represents the battery voltage error, BErr to be used
by the P-I controller for gain adjustment.

BErr = EB (actual) - EB (calculated) (7)

The (P-I controller employs the values of the proportional gain, KP and integral
gain, KI to compute the motor controller’s, K value.

K = ( KP + s·KI ) · BErr (8)

In the drive cycle the road was modeled in the computer simulation to help
reduced the expensive on-road test [14]. For the driving test and simulation purposes,
the vehicle speed values were established for a drive cycle of 100 s. Normally, the
torque value is acquired from the speed value and the vehicle dynamics. However, since
the vehicle dynamics are not included in the model, it is assumed that the torque values
are known for the simulation. The speed and torque data were respectively added into
the drive cycle subsystem using look-up tables. The developed drive cycle consisting of
the speed and torque values is as shown in Figure 3.

1315
 Mathematical modeling and simulation of an electric vehicle

Required Road Speed

Road Speed (rpm)
3000

2000
1000

0
0 10 20 30 40 50 60 70 80 90 100
Time(sec)
Required Road Torque
400
RoadTrq(Nm)

200

-200

0 10 20 30 40 50 60 70 80 90 100
Time(sec)
5
x 10 Required Road Power
2
Road Power(Watts)

Figure 3. Speed and torque values(-T)*(+S)=

(+T)*(+S)= +P = 1st Quadrant = Motoring
for the-Psimulation.
= 4th Quadrant = Regeneration
1

The required road speed is a plot of a combination of step by step increases and
0
decreases with a partly constant positive speed, while the required road torque consists
of the plots
-1
0
of both10the positive
20
and
30
negative
40
sides.
50
The
60
drive70cycle 80
is designed
90
in100
such a
way as to observe the power flow during both motoring and regeneration.
Time(sec)

RESULTS AND DISCUSSION

Figure 4 illustrates the EV drive simulation model produced from mathematical Eqs.
(1)–[12] which are represented by each subsystem block. From the model, five
simulation points were selected and added to the output scopes in order to determine
and illustrate the energy flow, performance and efficiency of two important elements of
the EV drive train: (i) the motor, and (ii) the battery. The simulation points are labelled
in Figure 4 (in blue) as below:

1. Road speed, torque and power

2. Motor voltage, current and power
3. Required battery voltage, current and power
4. Battery error
5. Gain value

The power requirement during the 100 s simulation time is calculated from the input
road speed and road torque data from the driving cycle and is plotted in Figure 5. The
positive power is acquired when both speed and torque are positive, note that the motor
is operating in forward motion (1st Quadrant operation). However, when the torque
becomes negative during positive speed, the motor switches into the 4th Quadrant
operating region and acts as a generator. This operation could be graphically identified
either by the negative y-axis region on the power curve or the negative value of power.
These operations are based on the four quadrant operation of the motor as shown in
Figure 2.

1316
 Mohd et al. / Journal of Mechanical Engineering and Sciences 8(2015) 1312-1321

Required Road Speed

Road Speed (rpm)

3000

2000
1000

0
0 10 20 30 40 50 60 70 80 90 100
Time(sec)
Required Road Torque
400
RoadTrq(Nm)

200

-200

0 10 20 Figure
30 4. EV40drive simulation
50 60 model.
70 80 90 100
Time(sec)
5
x 10 Required Road Power
2
Road Power(Watts)

(+T)*(+S)= +P = 1st Quadrant = Motoring (-T)*(+S)= -P = 4th Quadrant = Regeneration

1

-1
0 10 20 30 40 50 60 70 80 90 100
Time(sec)

Figure 5. Required power.

Figure 6 illustrates the voltage, current and power developed in the motor. The
figure is very similar to the previous figure of speed, torque and power. The voltage
curve follows the speed curve while the current follows the torque. This is obvious due
to their relationship as in Eqs. (1) and (2). Finally, the power curve in the third plot,
once again confirms the motor’s operation in the motoring and regeneration modes. The
power flows from the battery to the motor during motoring operation, but returns back
to the battery during regenerative braking.
Figure 7 shows the voltage, current and power draws from the battery. From the
figure, the battery current curve follows the shape of the motor current and required
torque curves. The rise in battery current is subsequently due to the increase in torque
demand as in Figure 3. From the battery power curve, power is drawn from the battery
to the load during motoring and returns to the battery during regeneration. When the
polarity of the voltage and current are equal, the motor operates in regular motoring
mode. However, the motor is switched into regenerative braking mode (generator) when
the current turns negative, and the power flows in the opposite direction. The battery
energy consumed during motoring and regeneration is 745 Watt-hour and -413 Watt-
hour, respectively.

1317
 Mathematical modeling and simulation of an electric vehicle

Motor Voltage (Volts) Motor Voltage

600
400

200

-200
0 10 20 30 40 50 60 70 80 90 100
Time(sec)
Motor Current
Motor Current (Amps)

400

200

-200

-400
0 10 20 30 40 50 60 70 80 90 100
Time(sec)
5
x 10 Motor Power
2
Motor Power (Watts)

(+V)*(+I)= +P Motoring (+V)*(-I)= -P Regeneration

1

-1
0 10 20 30 40 50 60 70 80 90 100
Time(sec)

Figure 6. Motor voltage, current and power.

Battery Voltage
Motor Voltage (Volts)

300

200

100

0 10 20 30 40 50 60 70 80 90 100
Time(sec)
Battery Current
Motor Current (Amps)

400

200

-200

-400
0 10 20 30 40 50 60 70 80 90 100
Time(sec)
5
x 10 Battery Power
3
Motor Power (Watts)

Batt Energy Motoring = 745WattHour Batt Energy ReGeneration = -413WattHour

2
(+V)*(+I)= +P Motoring (+V)*(-I)= -P Regeneration
1

-1
0 10 20 30 40 50 60 70 80 90 100
Time(sec)

Figure 7. Battery voltage, current and power.

Figure 8 shows the battery voltage error; the difference between the actual
battery internal voltage and the one calculated from the motor voltage and current

1318
 Mohd et al. / Journal of Mechanical Engineering and Sciences 8(2015) 1312-1321

values. This error output was employed as the input to the P-I controller. From the
figure, the maximum negative simulation error of -200 V which appears at 0 sec is
normal and can be negligible. The controller promptly recovers the system and
compensates the maximum error of 23.5 V during motor starting. The performance of
the controller is acceptable because for most of the time the error is zero. However, the
controller could still be improved by using various intelligent methods. The controller
gain, K is presented in Figure 9. From the figure, the value of K varies proportionally to
the speed demand where the rise in speed demand will increase the value of K. During
modeling, the integrator in the P-I controller block is preset with 0.1 initial conditions to
avoid an algebraic loop error during simulation. The P-I controller compensates the
error through the selected proportional and integral constants of Kp and Ki to keep the
system healthy. The value of the constants was established from the tuning process in
the Matlab-Simulink environment, providing the respective minimum and maximum K
values of 0.1 and 2.93.
Battery Voltage Error
100
Battery Voltage Error (Volts)

50
Max Positive Battery Error = 23.5
0

-50

-100

-150
Max Negative Battery Error = -200
-200
0 20 40 60 80 100
Time(sec)

Figure 8. Battery voltage error.

Controller Gain (K)
3.5
Max K value = 2.93
3
2.5
2
Gain

1.5
1
0.5
Min K value = 0.1 Set by PI
0
-0.5
0 20 40 60 80 100
Time(sec)
F
Figure 9. Controller gain

CONCLUSIONS

Modeling and simulation in Matlab-Simulink has been shown to be of great value in

investigating the energy flow, performance and efficiency of the EV drivetrain. In this
study, the simulation was performed and analyzed in both motoring and regeneration
mode. The operation mode of the motor is determined either by the road speed and
torque requirements or by the polarity of the motor current and voltage. The energy
flows from the battery to the load during motoring but in the opposite direction during

1319
 Mathematical modeling and simulation of an electric vehicle

regeneration. The EV’s performance depends on the performance of the controller in

removing error from the system. This work utilized a simple controller to maintain the
identical input–output power of battery and the P-I controller to compensate for the
voltage error. The design of the EV model presented in this paper is indeed a basic
model. There are still many opportunities for augmentation in order to establish a good
EV model which will form the foundation for further research and development.
Modeling and simulation are very important for automotive designers in order to find
the best energy control strategy and exact component size, and to minimize the use of
energy, because prototyping and testing are expensive and complex operations. Good
design leads to a good compromise among flexibility, model simplicity, computational
load and detailed representation of the components.

ACKNOWLEDGEMENTS

The authors would like to express highest gratitude to the Faculty of Engineering,
Universiti Putra Malaysia (UPM) for providing the effective facilities and efficient
learning environment in conducting the research. The research is supported by the
research grant of 03-01-13-1245FR from Ministry of Higher Education, Malaysia.

REFERENCES

[1] Trigg T, Telleen P, Boyd R, Cuenot F, D’Ambrosio D, Gaghen R, et al. Global

EV outlook: understanding the electric vehicle landscape to 2020. International
Energy Agency. 2013:1-40.
[2] Situ L. Electric vehicle development: the past, present & future. 3rd
International Conference on Power Electronics Systems and Applications. 2009;
1-3.
[3] Rahmat MS, Ahmad F, Mat Yamin AK, Aparow VR, Tamaldin N. Modeling
and torque tracking control of permanent magnet synchronous motor (PMSM)
for hybrid electric vehicle. International Journal of Automotive and Mechanical
Engineering. 2013;7:955-67.
[4] Salleh I, Md. Zain MZ, Raja Hamzah RI. Evaluation of annoyance and
suitability of a back-up warning sound for electric vehicles. International Journal
of Automotive and Mechanical Engineering. 2013;8:1267-77.
[5] Pereirinha PG, Trovão JP. Multiple energy sources hybridization: the future of
electric vehicles?2012.
[6] Butler KL, Ehsani M, Kamath P. A Matlab-based modeling and simulation
package for electric and hybrid electric vehicle design. IEEE Transactions on
Vehicular Technology. 1999;48:1770-8.
[7] Husain I, Islam MS. Design, modeling and simulation of an electric vehicle
system. SAE Technical Paper No. 1999-01-1149; 1999.
[8] Markel T, Brooker A, Hendricks T, Johnson V, Kelly K, Kramer B, et al.
ADVISOR: a systems analysis tool for advanced vehicle modeling. Journal of
Power Sources. 2002;110:255-66.
[9] Xu JW, Zheng L. Simulation and analysis of series hybrid electric vehicle
(SHEV) based on ADVISOR. International Conference on Measuring
Technology and Mechatronics Automation. 2010; 354-7.

1320
 Mohd et al. / Journal of Mechanical Engineering and Sciences 8(2015) 1312-1321

[10] Kaloko BS, Soebagio MHP, Purnomo MH. Design and Development of Small
Electric Vehicle using MATLAB/Simulink. International Journal of Computer
Applications. 2011;24:19-23.
[11] Schaltz E. Electrical Vehicle Design and Modeling. In: Soylu S, editor. Electric
Vehicles - Modelling and Simulations: InTech; 2011.
[12] Mapelli FL, Tarsitano D. Modeling of full electric and hybrid electric vehicles.
INTECH Open Access Publisher; 2012.
[13] Sanita CS, Kuncheria JT. Modelling and Simulation of four Quadrant operation
of Three phase Brushless DC Motor with Hysteresis Current Controller.
International Journal of Advanced Research in Electrical, Electronics and
Instrumentation Engineering. 2013;2:2461-70.
[14] McDonald D. Electric Vehicle Drive Simulation with MATLAB/Simulink.
Proceedings of the 2012 North-Central Section Conference; 2012.

1321

View publication stats