EVS27

Barcelona, Spain, November 17 - 20, 2013

Drivetrain design for an ultra light electric vehicle with high

efficiency
Isabelle Hofman1,2 , Peter Sergeant1,2 and Alex Van den Bossche2
1
Dept. Industrial Technology & Construction, Electrical Energy Research Group, Ghent University, Gent, Belgium
2
Dept. Electrical Energy, Systems & Automation, Electrical Energy Laboratory, Ghent University, Gent, Belgium
Isabelle.Hofman@Ugent.be

Abstract
The complete drivetrain for a single person ultra light electric vehicle (EV) is optimized towards a mini-
mal total weight and a maximal average efficiency for different driving cycles. The EV is named ELBEV,
which is an acronym for Ecologic Low Budget Electric Vehicle. The single person ultra light EV is a
tricycle, with two driven and steering front wheels and one rear wheel. The drivetrain of each front wheel
consists of an outer rotor permanent magnet synchronous motor (PMSM), a single-stage gearbox and the
power electronics with converter and controller print. The drivetrain is optimized for the New European
Driving Cycle, the New York City Cycle and the Federal Test Procedure. For the optimization of the
drivetrain analytical models are used to calculate the losses and the efficiency. The optimized parameters
of the motor are: the number of pole pairs, the number of stator slots and the outer rotor radius of the
motor. Furthermore, an analytical model for the single-stage gearbox is implemented for different gear
ratios (GRs). The optimized parameters for the gearbox are: the number of teeth and the module of each
gear combination and the total mass of the gearbox for each GR. The analytical models are fast, and
useful for designing a good PMSM in combination with a single-stage gearbox. The optimization of the
complete drivetrain is always a compromise between total average efficiency over the drive cycle and the
total mass of the drivetrain.

Keywords: Integrated design, Drivetrain, Ultra light electric vehicle

1 Introduction 1.1 Ecologic Low Budget Electric Vehi-

The main reasons why people choose for an elec-
tric vehicle (EV) is because they want to reduce
the dependence on foreign oil and they want to
mitigate the climate change. The major problem
of commercial EVs today is the high cost of the
vehicle. This is due to the large amount of bat-
teries and because of that the high weight of the
vehicle, but also because of the high cost of the
battery technology. Therefore, we want to design
an ultra light electric vehicle named the ELBEV,
acronym for Ecologic Low Budget Electric Vehi-
cle. The goal is to have a transportation method
with less energy consumption than a commercial
EV but yet much faster, more comfortable and
safer than a bike.
cle
The ELBEV (Fig.1) is a one person EV with bat-
teries, mainly for commuting purposes in the city
and in the suburbs. The vehicle is controlled by
the driver via two contactless handles [1]: a gas
and a brake handle. The low cost car has three
wheels, two driven and steering front wheels and
one rear wheel. The maximum speed of the car is
80 km/h and the range is about 100 km. The to-
tal curb weight (batteries included) is about 100
kg. The battery pack consists of Lithium Poly-
mer batteries with a high energy to weight ratio:
20 Ah, 96 V and 11 kg. The dimensions of the
car are: a total length of 2200 mm, a total width
of 1200 mm and a total height of 1300 mm.

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 1

Figure 1: Ecologic Low Budget Electric Vehicle.
1.2 Drivetrain of each front wheel of the
ELBEV
Each front wheel of the ELBEV is driven by a
motor with gearbox and power electronics. The
block diagram of the drivetrain for ELBEV is
shown in Fig. 2.
The battery (1) provides the power stage of each
brushless DC (BLDC) motor with current. The
power stage (2) is connected to the three stator
coils of the BLDC motor (3). Each of the sta-
tor coils is powered by a half bridge, two of them
contain a current sensor (4) to measure two phase
currents (Ia and Ib ) flowing through the motor.
The two phase currents are measured by the con-
trol print (5). The third phase current is equal
to the negative sum of Ia and Ib , this is done by
electronics on the control print [2]. The control
print is fed by a flyback converter on the power
stage. With the hall sensors (6), mounted on the
BLDC motor, the exact rotor position is known.
This signal is also coupled with the control print.
The signal is then fed to the power stage to switch
on the correct transistors to drive the motor. The
drive signal (7) and the brake signal (8) from the
handles are fed to the control print. An enable
button (9) is also mounted, to switch on the elec-
tronics. With the torque direction button (10) the
driving direction of the vehicle can be reversed.
The BLDC motor is coupled with a single stage
gearbox (11) in order to reduce the speed and in-
crease the torque of the motor on the front wheel
(12).
Battery
(1)
Power stage
(2)
Phase currents
(4)
Drive handle input (7)
Brake handle input (8)
Ia
Ib
Enable (9)
Control Print
(5)
Gearbox
Torque direction (10)
Hall sensors
(6)
Brushless DC
(3)
Figure 2: Drivetrain for each front wheel of the EL-
BEV.
EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium
2 Integrated design
In the past, the power electronics and controller
have been developed and tested. A gearbox has
been designed, built and tested together with
a commercial outer rotor permanent magnet
synchronous motor (PMSM). A new optimized
design for this motor is currently implemented
in combination with an improved single-stage
gearbox. The objective function in the optimiza-
tion process is shown in Fig. 3.
GR, Np, Ns and rrotor
GR
Selection of optimal
Battery
rrotor module and number
of teeth of gears
VDC
Power
Np Ns
electronics Motor Gearbox
model model model
Efficiency map Efficiency map Efficiency map
Mass Mass
Total efficiency for drive cycle:
Power electronics + BLDC motor + Single-stage gearbox
Total mass: BLDC motor + Single-stage gearbox
Total efficiency for drive cycle
Total mass
Figure 3: Objective function that is executed for gear
ratios = 12 – 17 , Np = 3 – 8, Ns = 9 – 18 and rrotor =
0.04 – 0.1 m.
The optimized parameters for the new outer
rotor PMSM are the number of pole pairs (Np ),
the number of stator slots (Ns ) and the outer
rotor radius of the motor (rrotor ). Therefore we
use an analytical model to calculate the losses
and the efficiency of the PMSM. Moreover, the
total active mass of the motor is calculated.
For the single-stage gearbox an analytical model
is implemented for different gear ratios (GRs) in
order to have a minimal weight for each GR. The
GRs that were investigated are 21 to 71 . Each of
the motor and wheel gear combinations are able
to transfer the required peak torque on the wheel.
Fig. 4 shows the mass for several possible gears
Front wheel
(motor and wheel gear combinations for the
single-stage gearbox) for a GR of 12 . Each of
(12)
the markers defines a combination of motor
and wheel gears with a different module and/or
different teeth numbers. From the figure, it is
(11)
clear that the combination of 40 teeth at motor
side and 80 teeth at wheel side with module
1, gives the lowest total mass for the different
gear combinations. This approach was used for
each GR. It is a sub optimization shown in the
objective function of Fig. 3.
The PMSM and gearbox are optimized for the
New European Driving Cycle (NEDC), the New
2
 Mass of gears in function of number of teeth and module New York City Cycle, NYCC
14
Gear ratio 1/2

Module 2.5
1.6 12
Module 2
Module 1.5
1.4 Module 1 10
Mass of gears [kg]

Vehicle speed [m/s]

8
1.2

6
1

4
0.8

2
17 2022 25 2830 40 50 60
Number of teeth at motorside
0
0 50 100 150 200 250 300 350 400 450 500 550 600
Figure 4: Weight of several gears for a single-stage Time [s]

gearbox with gear ratio 12 . All shown combinations Figure 6: New York City Cycle. Important param-
are able to transfer the required peak torque on the eters of the NYCC: average speed: 11.5 km/h, maxi-
wheel, so that the figure makes it possible to select the mum speed: 44.45 km/h, total distance: 1.903 km and
gear with the lowest mass. Note that, as even numbers duration: 598 s.
of teeth and integer multiples of teeth are not com-
mon practice in gears, the optimal combination 40-80
could be slightly modified to e.g. 41-81.
PMSM with Np = 7, Ns = 12 and rrotor = 0.04
m. The average efficiency is 93.63 % in a speed
range of 0.5 Nnom – Nnom and in a torque range
York City Cycle (NYCC) and the Federal Test of 0.5 Tnom – Tnom . The maximum efficiency is
Procedure (FTP) [4], [5]. The NEDC is shown 94.34 %. Note that the crosses in the figure show
in Fig. 5, but due to the maximum speed of 20 the working points over the NEDC. In Fig. 9 and
m/s of the ELBEV, the NEDC is cut off at that Fig. 10 the same calculated efficiency map of the
specific speed. Fig. 6 shows the NYCC and Fig. motor as in Fig. 8 is shown, but now with the
7 the FTP (dashed line is the maximum speed of working points over the NYCC and the FTP (see
the ELBEV). Fig. 9 and Fig. 10).

The goals are a minimization of total weight and

a maximization of the total average efficiency 3 Results
of the drivetrain for the different driving cycles,
consisting of the power electronics, the PMSM The objective function of Fig. 3 is executed for
and the gearbox. several GRs and for a range of the motor param-
eters Ns , Np and rrotor . For each combination
of input parameters, the total mass and average
35
New European Driving Cycle, NEDC efficiency of the drivetrain are computed for the
different driving cycles. The total mass of the
30 drivetrain (motor + single-stage gearbox) is the
same for each driving cycle, and is shown in Fig.
25 11. For each driving cycle (NEDC, NYCC and
FTP) the average efficiency is shown in Fig 12,
Vehicle speed [m/s]

20 Fig 13 and Fig. 14.

15

10
Table 1: Results total drivetrain with single-stage
5
gearbox for three driving cycles for gear ratio 13

0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 GR: 1/3, rrotor : 0.055 m, Np : 7, Ns : 12
Time [s]

Property NEDC NYCC FTP

Figure 5: New European Driving Cycle, the dashed
Av. total eff. 85.72 % 71.15 % 85.37 %
line is the maximum speed of the ultra light electric
vehicle. Important parameters of the NEDC: average Driving range 112.98 km 325.77 km 125.61 km
speed: 33.6 km/h, maximum speed: 120 km/h, total Total mass 8.9912 kg
distance: 11.017 km and duration: 1180 s [3].
From all these solutions, the following trends
are observed: For a drivetrain (motor + power
Fig. 8 shows the calculated efficiency map of a electronics + gearbox) with Np = 7, Ns = 12,

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 3
 30
Federal Test Procedure, FTP
Calculated efficiency map of motor
4.5

25
4

0.84 0.82
3.5

0.5

4
0.92

0.9
0.86
0.88

0.91

3
0.8

0.9
20
3

0.9
Torque [Nm]
Vehicle speed [m/s]

2.5
15
0.94
2
10 1.5 0.93
0.92
1 0.91
5 0.9
88
0.5 860.
0.84 0.
0.80.82
0 0.5
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000 0 1000 2000 3000 4000 5000
Time [s]
Speed [rpm]
Figure 7: Federal Test Procedure, the dashed line is Figure 9: Calculated efficiency map of the motor
the maximum speed of the ultra light electric vehicle. (GR = 17 , Np = 7, Ns = 12 and rrotor = 0.04 m).
Important parameters of the FTP: average speed: 34.2 The crosses show the working points over the NYCC.
km/h, maximum speed: 91.09 km/h, total distance:
17.787 km and duration: 1874 s.
Calculated efficiency map of motor
4.5
Calculated efficiency map of motor 4
4.5
3.5 0.82 0.8
0.86.84
4

3
0.5

0.88

0.92
0.91

0.9

0.94
0.9
0
3
0.82 0.8

3.5
0.86.84

Torque [Nm]
0.88

4
0.91
0.92

0.9
0.5

0.9
0.9

2.5
3 0.94
Torque [Nm]

2
2.5
0.94 1.5 0.93
2 0.92
1 0.91
1.5 0.93 0.9
0.88
0.92 0.5 0.84 0.
86
1 0.82 0.8
0.91 0.5
0.9 0
0.5 0.88 0 1000 2000 3000 4000 5000
0.86
0.82 0.84 0.8 Speed [rpm]
0 0.5
0 1000 2000 3000 4000 5000
Speed [rpm] Figure 10: Calculated efficiency map of the motor
(GR = 17 , Np = 7, Ns = 12 and rrotor = 0.04 m). The
Figure 8: Calculated efficiency map of the motor crosses show the working points over the FTP.
(GR = 17 , Np = 7, Ns = 12 and rrotor = 0.04 m).
The crosses show the working points over the NEDC.

is that for almost the same average efficiency

over the NEDC (85 %) the total mass of the
Table 2: Results total drivetrain with single-stage complete drivetrain is 3.59 kg higher when using
gearbox for three driving cycles for gear ratio 17 a drivetrain with Np = 7, Ns = 12, rrotor = 0.04
m and a GR of 17 . The total mass of the drivetrain
GR: 1/7, rrotor : 0.04 m, Np : 7, Ns : 12 in Table 2 is 12.58 kg (mass 1 gearbox = 5.31 kg,
mass 1 motor = 0.98 kg), and again the average
Property NEDC NYCC FTP efficiency and thus the driving range depend on
Av. total eff. 85.56 % 76.18 % 85.57 % the drive cycle.
Driving range 112.77 km 348.80 km 125.91 km
Total mass 12.5820 kg
4 Conclusions
rrotor = 0.055 m and a GR of 13 the total mass of By optimization of the complete drivetrain
the drivetrain is 8.99 kg (mass 1 gearbox = 2.33 over the different driving cycles it is possible
kg, mass 1 motor = 2.16 kg), but the average to choose a good combination of motor and
efficiency and thus the driving range depend on single-stage gearbox in order to have high
the drive cycle, see Table 1. When comparing average efficiency and a minimal weight.
this result (Table 1) with Table 2, the conclusion

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 4
 Total mass drivetrain [kg] Average efficiency over NYCC
0.1 19.967 0.1
4 21.0 22 24
18.8 799 .19 23.305 .41 0.
16 17 549 25 75 69

0.
.62 .74 0.
72 19

76
0.80
98 24

0.7

0.
81 6

0.76

42
80
15 22 2

8
19

281
.51 .19

044

04
73 18.8 .96 21.0799

428
25

4
0.09 549 74 0.09

2
14 17.7
.40 424

0.6
48 16.6 19
298

919
13 .9
.29 67
15.5 18.8549 4

6
23 173
0.7
Outer motor radius (rrotor) [m]

Outer motor radius (rrotor) [m]

0.08 14.4 0.08 64
12 048 17.7424 2
.1 16 0. 8
11 7 97 .62 80
.0 98 04
13.2 4

0.5
67 923

0.
0.6 48
2 15 0.

80
83
9. .51 76

196

0.7

0
0.6
95 73 16 42

44
0.07 46 0.07

28 0.
.6 8

4
12.1

558
8 14 29

12 69
797 .40 8
11.0 48 15
672 .5
13 17

19
.2 3
92

6
3

0.4.402
0.06 9.9 0.06

0
0.8

14
546

0.5.475
8.84 00

38 68

0.51116
8

40

0 84
215 0.7 44

47
7.

8
12.1 0. 0.7 64
72

58 0.6 28

13
797 28

32
9

34 19 0.6 12

.2
62

9
8 64 55
11

23
0.05 0.05 8 0.6
.0

919 0.80
044
67

8. 6
84
2

21 0. 0 0 0

12
0.7642
5.

7.7 25 .2 .33 .36

9.

6.6 5 8
50

0.
.17
29 80 94 03 65 0.7281
9

1 0.4 0 10.54

22
71
45

38 0.4.5 2
54

6 4 2 6 2 7 0
2 84 75 116 32 0.5834.619640.6558

97

18
7

0.691
68

8
8 96
0.04 0.04
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Gear ratio (GR) Gear ratio (GR)

Figure 11: Total mass of drivetrain with single-stage Figure 13: Average efficiency over the NYCC. The
gearbox for each drive cycle. ridge (80.04 %) shows the combinations of GR and
outer motor radius that yield the highest average effi-
Average efficiency over NEDC
ciency.
0.1
0.
0. 64
0.7

66
0.

79
0.7

0.
79

72

67 6
0.
0.
0.85

74
0.8

3
68
30
36
98

18

55

Average efficiency over FTP

60

4
16

13

1
444

0.1
9

0.09
8

0.
0. 66

0.
68 20

0.84

0.7

73
0.
50 9

0.
0.8

7
99

10
0.
0. 6

53

70
59

77

2
70

95
22

80
9
42

69

4
92
0.7 7 0.09

7
Outer motor radius (rrotor) [m]

0.08 0. 0. 23
7 79 74 04
0. 1
36 76 81
0.

05
0.8

79

8
0.

81
35

81

3
67

69

0.7
0.85
0.

Outer motor radius (rrotor) [m]

0.07 0.08 0. 31
85

0. 75 02
444

79
44

0. 39
99 77
4

69 9
0.7 5

0.
793 7
0.8

82
0.7 6

29
981
45
0.8

0.8 3

2
0.8 0.07
9

0.06 35 16
45
0.

67 9
0.8
81

9
69

35
67

0.8 0.7
544
0.

4 999
85

0.8 5
44

0.05 0.83 0.06 22

0.7 69

0.8 92
4

567
0.

0.8 5

459
77
99 7
229
9
0. .7720485

0.7 0.8
0.741302475
0 . .6
00

0.60851

54
7

98
2

0.8 4
77 8

13 0.8 356 0.85 4

0.

1
93

69 7 444
84
6

0.05
59

0.04
1 2 3 4 5 6 7 0.845
9
Gear ratio (GR)
0. .663961

0.7
0 .661
06. 8520191
0.

07. 0806
0

5 39 0.7 0
73 04

9 999 .82292 0.84

10

5 59
2
3

Figure 12: Average efficiency over the NEDC. The 0.04

1 2 3 4 5 6 7
Gear ratio (GR)
ridge (85.44 %) shows the combinations of GR and
outer motor radius that yield the highest average effi- Figure 14: Average efficiency over the FTP. The ridge
ciency. For a low GR of 12 , the highest average effi- (84.59 %) shows the combinations of GR and outer
ciency is obtained with a motor of double radius com- motor radius that yield the highest average efficiency.
pared to a high GR of 17 . For a low GR of 12 , the highest average efficiency is
obtained with a motor of double radius compared to a
high GR of 17 .
Optimization for different driving cycles will re-
sult in an optimal combination of motor and gear-
box for that driving cycle. tronics and Applications (EPE 2011), Pub-
lished: 2011.
[3] Regulation No 101 of the Economic Com-
References mission for Europe of the United Nations
(UN/ECE) Uniform provisions concern-
[1] A. Van den Bossche and P. Sergeant, “Eddy ing the approval of passenger cars powered
current based, contactless position trans- by an internal combustion engine only, or
ducer for a gas handle,” IEEE, Proceed- powered by a hybrid electric power train
ings of the 2010 Vehicle Power and Propul- with regard to the measurement of the emis-
sion Conference (VPPC 2010), Published: sion of carbon dioxide and fuel consump-
2010. tion and/or the measurement of electric en-
ergy consumption and electric range, and of
[2] A. Van den Bossche, DV. Bozalakov, T. categories M 1 and N 1 vehicles powered
Vyncke and VC. Valchev, “Programmable by an electric power train only with regard
Logic Device Based Brushless DC Motor to the measurement of electric energy con-
Control,” IEEE, Proceedings of the 2011- sumption and electric range, Official Jour-
14th European conference on Power Elec- nal L 138 , 26/05/2012 P. 0001 0077.

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 5
 [4] T.J. Barlow, S. Latham, I.S. McCrae, P.G.
Boulter, “PPR354,” in A reference book of
driving cycles for use in the measurement
of road vehicle emissions, TRL, 2009.
[5] W. Courtois. (2013, Aug. 20). Dynamome-
ter Drive Schedules [Online]. Available:
http://www.epa.gov/dynamometer.htm
Authors
Isabelle Hofman received the
master degree in electromechanical
engineering in 2010 from the Uni-
versity College of Ghent, Belgium.
In 2010, she became a Ph.D. re-
searcher in a project entitled Eco-
logic Low Budget Electric vehicle
at the University College of Ghent.
Her research is in the field of elec-
tric drives for light electric vehi-
cles.
Alex P. M. Van den Bossche re-
ceived the M.Sc. and the Ph.D.
degrees in electromechanical en-
gineering from Ghent University
Belgium, in 1980 and 1990 respec-
tively. He has worked there at the
Electrical Energy Laboratory (EE-
LAB). Since 1993, he is professor
at the same university in the same
field. His research is in the field of
electrical drives, power electronics
on various converter types and pas-
sive components and magnetic ma-
terials. He is also interested in re-
newable energy conversion. He is
an author of the book Inductors and
transformers for power electron-
ics. He was a starter of the spin-
off companies Inverto n.v. (1990)
and recently Alenco n.v. (2009).
He is an IEEE member since 2000
and a senior member of IEEE since
2003.

Peter Sergeant received the M.Sc.

degree in electromechanical engi-
neering in 2001, and the Ph.D.
degree in engineering sciences in
2006, both from Ghent University,
Ghent, Belgium. In 2001, he be-
came a researcher at the Electri-
cal Energy Laboratory of Ghent
University. He became a postdoc-
toral researcher at Ghent Univer-
sity in 2006 (postdoctoral fellow
of the Research Foundation - Flan-
ders) and at Ghent University Col-
lege in 2008. Since 2012, he is
associate professor at Ghent Uni-
versity and Ghent University Col-
lege. His current research interests
include numerical methods in com-
bination with optimization tech-
niques to design nonlinear elec-
tromagnetic systems, in particular,
electrical machines for sustainable
energy applications.

EVS27 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 6