Proceedings of the 2011 International Conference on Power Engineering, Energy and Electrical Drives Torremolinos (Málaga), Spain.

aga), Spain. May 2011

Regenerative Braking Control Method and Optimal

Scheme for Electric Motorcycle
Wei Cui, Hang Zhang, Yan-li Ma, Yue-jin Zhang
School of Mechanical and Electronic Engineering and Automation, Shanghai University, Shanghai 200072, China
[email protected]

Abstract- This paper describes an optimal regenerative braking

control scheme for a permanent magnet brushless DC motor of I. INTRODUCTION
an electric motorcycle to achieve dual goals of the electric brake As an important opportunity of leading the cities around
and the maximal power harvest in two cases of the road,
downhill and flat, without any additional changes on the the world for green economy, Shanghai Expo 2010 delivers
hardware. Based on regenerative braking of brushless DC the theme “Better City, Better Life” to the world by
motor in half bridge modulation mode, the relationship between developing the green transportation as a priority. The clean
average regenerative current on the DC bus and PWM duty energy technologies of super capacitor vehicles, electric
cycle under different speeds was studied intensively from the vehicles, hydrogen fuel cell vehicles and hybrid vehicles are
three aspects of theoretical deduction, simulation based on
Matlab/Simulink and platform experiments, and then a simple developed to put into practice, and we learned that 300
but effective scheme considering the actual application was electric motorcycles have been used for sightseeing, security
proposed tentatively to achieve the dual goals. Since the braking and logistics in Expo Park. More detailed data of electric
kinetic energy is harvested in the form of the electric energy vehicles about the whole city have been listed in 2010
stored up in battery, the scheme could improve the driving Shanghai Comprehensive Transportation Annual Report, the
range, safeness, noise preventability and cost effective of electric
motorcycles. registrations of electric motorcycle came up to 2.48 million at
the end of 2009 with the rate 18.4% of all the non-motor
Index Terms- Permanent magnet brushless DC motor, Half vehicles, and compared to the previous year by 6.7% while
bridge modulation mode, Regenerative braking control,
Maximal power harvest. the non-motor vehicles increased by 1.3%. It can be seen
from this that the electric motorcycle is playing a more and
NOMENCLATURE more important role in the citizens’ lives.
To meet the market demand and energy consumption
Cd Large capacitor in parallel with battery standards for electric motorcycles, the scheme of energy
D1 ~ D6 Freewheeling diode conservation for electric motorcycles must be studied
E Sum of two-phase back EMF intensively. Nowadays, permanent magnet brushless DC
e A , eB , eC Back electromotive force of phase inductors motor is widely used as a drive system of electric
I10 Initial value of i1 motorcycles because of its large power density, high
I 20 Initial value of i2 efficiency and simple control strategies [1-2]. As the inertia
i1 , i2 Current of motor’s armature winding of electric motorcycle is very large, it can produce a lot of
L Inductance of two phase kinetic energy when operated. Considering the urban driving
l Ratio of T and τ cycle of Shanghai as shown in Fig.1, the riders have to brake
n Ratio of E and U the electric motorcycles from time to time. Therefore, if the
n[k ] Motor speed at the kth control period kinetic energy can be harvested by regenerative braking
R Sum of two-phase resistances [3-16], then the important energy conservation problems such
r Inner resistance of the battery as the endurance of battery and the driving range of electric
T Value of PWM cycle motorcycle can be improved. Besides, the mechanical braking
T1 ~ T6 Power devices noise pollution and driving safety of electric motorcycle are
Tb Braking torque obtained by the system becoming so serious that many cities prohibit the use of it,
Tmg sin θ Torque produced by gravity component however, the regenerative braking control scheme could
ton Conduction time of T4 make it better.
toff Duration of non-conduction of T4 i
tx Duration of t2 i1

U Terminal voltage of the storage battery

i0
Ud Voltage on the DC bus
α PWM duty cycle
t1
α [k ]
t0 t2 t
PWM duty cycle at the kth control period
Fig.1. Driving cycle Fig.2. Current waveform
α n[ k ] Optimal duty cycle under n[k ]
τ Ratio of L and R In view of the problems referred above, this paper, based
on the regenerative braking theory of half bridge modulation

978-1-4244-9843-7/11/$26.00 ©2011 IEEE

mode without making any additional changes to the relation of HALL states and conductive power devices when
hardware[6-15], is dedicated to studying the relationship forward braking is shown in Tab.2.
between average regenerative current on the DC bus and Tab.1
PWM duty cycle under different speeds in simulation and CORRESPONDING RELATION OF HALL STATES AND CONDUCTIVE POWER
DEVICES WHEN THE MOTOR IS FORWARD DRIVING IN HALF BRIDGE
experiment respectively, and then tentatively proposing a MODULATION MODE
simple but effective scheme to harvest the most regenerative
HALL
braking energy in the form of the electrical one stored up in 110 010 101 100 001 011
States
battery. Considering the battery may be injured by the surge
Power T1 T1 T3 T3 T5 T5
recovery current of large negative-accelerate duration and the
States T2 T6 T4 T2 T4 T6
brake process may be rough for a sudden change of the PWM
duty cycle, this paper presents two cases, downhill roads and Tab.2
CORRESPONDING RELATION OF HALL STATES AND CONDUCTIVE POWER
flat roads, in which the scheme is optimized reasonably. The DEVICES WHEN THE MOTOR IS FORWARD BRAKING IN HALF BRIDGE
optimal scheme has the properties of the regenerative electric MODULATION MODE
braking and essential mechanical braking to increase the HALL
110 010 101 100 001 011
States
driving range, safeness, noise preventability, and cost Power
effective of the electric motorcycle. Device
T4 T4 T6 T6 T2 T2
II. ANALYSIS OF REGENERATIVE BRAKING CONTROL As shown in Tab.1 and Tab.2, under the same HALL states,
METHOD the direction of the armature current when the motor is
A. Principle of Regenerative Braking forward braking is opposite to that when the motor is forward
The control of regenerative braking depends on buck-boost driving, so the effect of braking is achieved.
converter [6-8]. Fig.2 shows the current waveform in one C. Deduction of the Relationship Between Average
switching period of t0 ~ t2 . In the period of t0 ~ t2 , defined Regenerative Current on the DC Bus and PWM Duty
as the continuous current state, the power device is switched Cycle
on, the magnetic field energy is stored in the inductors of the It is assumed that HALL state is 010 now, and T4 is
stator windings and the armature current increases; in the controlled by PWM. When T4 is switched on, the circuit is
period of t1 ~ t2 , defined as the charging state, the power working under the continuous current state of regeneration
device is switched off, the magnetic field energy previously braking, and the inductors of phase A and phase B store the
stored in the inductors is released and superposes the back magnetic field energy. In this case, the current flow of the
EMF, resulting in charging to storage battery and the circuit is illustrated in Fig.3 (To simplify the analysis and
decrease of the armature current. calculation, the phenomenon that three phases are conductive
B. Control Method of Regenerative Braking in Half Bridge in the period of non-commutation when eC < 0 is ignored).
Modulation Mode When T4 is switched off, the circuit is working under the
Based on the operating principle of brushless PM motor, charging state, and the inductors of phase A and phase B
the direction of motor output torque is reversed by changing release the magnetic field energy, as illustrated in Fig.4.
the direction of the stator current under the same poles. While T1 T3 T5 T1 T3 T5
r D1 D3 D5 r D1 D3 D5
maintaining the same rotation direction, motor output torque Cd Cd
A B C A B C
is reversed by changing the direction of the stator current to Ud
T4 D4 T6
T2
Ud
T4 D 4 T6 T2
D6 D2 D2
achieve the effect of braking. To reverse the direction of the D6

stator current, the conduction sequences of the power devices r1 r1

r1 r1 r1 r1
of three-phase inverter are only changed [3]. eA eB eC eA eB eC
When brushless PM motor is forward driving in half bridge
modulation mode, six-step 120D drive scheme is used, that Fig.3. Continuous current state Fig.4. Charging state
is, two-phase power devices which belong to high bridge arm Since the actual value of inductance L of the circuit can not
and low bridge arm of the inverter respectively are conducted be infinite, the circuit shown in Fig.4 can work under two
at any time, and the power device at the low bridge arm is types of modes, armature discrete current mode (DCM) and
controlled by PWM in speed control of motor. Tab.1 shows armature continuous current mode (CCM) [4], as shown in
the corresponding relation of HALL states and conductive Fig.5 and in Fig.6 respectively.
power devices when the motor is forward driving. When the
i i
motor is forward braking in half bridge modulation mode, i1 i1
i2 i2
only three power devices ( T2 , T4 , T6 ) at the low bridge arm I 20
are switched on and off at a controlled duty cycle for 120 o t on t1 t x t2 I 20
t I 10 I 10
electrical degrees respectively while the other three power t off o t on t off t
devices ( T1 , T3 , T5 ) at the high bridge arm are always T
switched off, and the conductive time of T2 , T4 , T6 is that of T
T5 , T1 , T3 (forward driving) respectively. The corresponding Fig.5. Armature discrete current Fig.6. Armature continuous current
 To establish the mathematical model of the regenerative Because the armature current is discrete, the average
braking system, the following simplification is taken: the regenerative current on the DC bus I ave is
induced back electromotive force (EMF) and inductor αT
− αT
coefficients are regarded as a constant in a PWM period; 1 t L U − Ee τ −
I ave = ∫ 2 i2 dt = 2 [( E − U ) ln + E (1 − e τ )] (9)
on-state voltage drops of the power devices in the inverter are T ton RT U −E
ignored; due to the effect of a large capacitor in parallel with −
αT
αT
the storage battery, the inner resistance of the battery is dI ave E 2 1 − e τ −
= ( )⋅e τ >0 (10)
dα α T
regarded as zero. R −
Based on two working states of the circuit shown in Fig.4 U − Ee τ
(armature discrete current mode and armature continuous Therefore, I ave is a monotonically increasing function,
current mode), mathematical expression of the relationship i.e., as the value of PWM duty cycle increases, the average
between average regenerative current on the DC bus and regenerative current on the DC bus I ave also increases.
PWM duty cycle will be deducted in the following section. 2 Continuous Current Mode
1 Discrete Current Mode Using the original condition i1 = I 20 at t = ton , i2 = I10
When T4 is switched on (shown in Fig.2), the voltage at t = T (see Fig.6), we can obtain the current equations for
equation is written as: the continuous current mode according to formulas (2) and (4)
di1 as:
L + Ri1 = E (1) t
− on
t
− on
dt τ E
I 20 = I10 e + (1 − e τ ) (11)
Solving the equation, we have: R
t t toff toff
− E − − E −U −
i1 = I10 e τ + (1 − e τ ) (2) τ τ
I10 = I 20 e + (1 − e ) (12)
R R
When T4 is switched off (shown in Fig.4), the voltage For the above two simultaneous equations, we have
equation is written as: −
toff

di2 E 1− e τ U 1 − e− β l U
L + Ri2 = E − U (3) I10 = −( ) = ( n − ) (13)
dt R −
T
R 1 − e−l R
1− e τ
Solving the equation, we have:
t T
t −ton t −ton − on −
−
τ E −U −
τ E e τ −e τ U e −α l − e −l U
i2 = I 20 e + (1 − e ) (4) I 20 = −( ) = (n − ) (14)
R R −
T
R 1 − e−l R
1− e τ
Using the original condition I10 = 0 (see Fig.4) and
i1 = I 20 at t = ton , we can obtain the current i2 as: Set β = 1 − α and linearize the above two formulas using
αT t Taylor series [4]:
E −U U τ E −
i2 = + ( e − )e τ (5) U
R R R I10 = I 20 = (n − β ) (15)
In the above formula, i2 = 0 at t = t2 , hence, R
ton The above formula represents the average armature current
e τ −n I 0 when inductance L is infinite [4], hence
t2 = τ ln( ) (6) U E − (1 − α )U
1− n I 0 = (n − β ) = (16)
t R R
− on
1 − ne τ The average regenerative current on the DC bus is
t x = τ ln (7)
1− n E − (1 − α )U
I ave = (1 − α ) I 0 = (1 − α )[ ] (17)
When t x < toff , the circuit is working under the armature R
discrete current state. Using the condition as t x < toff , we I ave is a quadratic parabola function with down opening,
obtain the following formula as: so it has a maximum I max . Using dI ave dα = 0 , we can
L U (8) obtain this maximum I max and corresponding PWM duty
α < 1− ln R
 α0
RT − T
U − E (1 − e L ) cycle α max as:
Formula (8) shows that when PWM duty cycle is very E
α max = 1 − (18)
small, i.e. α < α 0 , the armature current is discrete, and when 2U
PWM duty cycle is very large, i.e. α ≥ α 0 , the armature E2
I max = (19)
current is continuous. Therefore, the working state of circuit 4UR
depends on α , thereby the deduction of mathematical To sum up, when brushless PM motor is working under
expression of the relationship between average regenerative half-bridge-modulation regenerative braking state, then if the
current on the DC bus and PWM duty cycle is also based on armature current is discrete, i.e., α < α 0 , the average
α. regenerative current on the DC bus I ave is a monotonically
increasing function, and if the armature current is continuous, the DC motor for electric operation and the brushless PM
i.e., α ≥ α 0 , I ave is a quadratic parabola function with a motor for generation operation. Under the condition, the
maximum. Therefore, one optimal duty cycle which makes oscillogram of phase current and the charging current when
average regenerative current maximum exists, and formulas the brushless PM motor is braking is illustrated in Fig.7(b)
(18) and (19) show that the optimal duty cycle decreases and and Fig.8(b) respectively.
the maximum average regenerative current increases as the Comparing with Fig.8(a) and Fig.8(b), one can know that
back EMF increases, i.e., the speed of the motor increases. the charging current is above the zero baseline, even the rise
of continuous stage current and the fall of charging stage
D. Simulation and Experiment current are clearly showed every PWM cycle. Besides,
Tab.3 Fig.7(b) is nearly the same with Fig.7(a). Therefore, the
EXPERIMENTAL SETUP PARAMETERS theoretical deduction and simulation results are proved to be
Parameter Symbol Value correct.
Stator resistance( Ω ) R1 0.38
E. Data Processing and Analysis
Stator inductance( mH ) L1 0.3 To study the relationship between average regenerative
Ratio of back EMF and speed( V / r ⋅ min −1 ) E1 / n1 0.045 current on the DC bus and PWM duty cycle under different
Number of pole pairs p 19
speeds, as shown in Fig.9(a) and Fig.9(b), four kinds of speed
are selected both in simulation and experiment: 50r/min,
Rated voltage( V ) UN 24
100r/min, 150r/min and 200r/min.
Rated speed( r / min ) nN 200

(a) (b)
Fig.9. Average regenerative current Vs PWM duty cycle under four speeds.
(a) Simulation. (b) Experiment.
Fig.7. Waveform of phase current. (a) Simulation. (b) Experiment.
As shown in Fig.9(a) and Fig.9(b), when the brushless PM
motor is working under the half-bridge-modulation
regenerative braking state, one optimal duty cycle which
makes average regenerative current maximum exists.
Furthermore, the higher the speed is, the greater average
regenerative current is; as the speed increases, the optimal
duty cycle which makes average regenerative current
maximum decreases. Therefore, the deduction of the
relationship between average regenerative current on the DC
bus and PWM duty cycle is proved to be consistent with the
Fig.8. Waveform of charging current. (a) Simulation. (b) Experiment. experiment.
To verify the correctness of the theoretical deduction, a III. OPTIMAL BRAKING CONTROL SCHEME BASED ON
simulation mode1 of regenerative braking system was MAXIMAL HARVEST DUTY-CYCLE MODEL
established based on Matlab/simulink [2]. The parameters of
In the actual application, there are four constraints to the
brushless PM motor used in the simulation are given in Tab.3.
optimization of regenerative braking control scheme in
The simulation waveform of phase current and charging
electric motorcycle presented in this paper. 1) The optimal
current of battery when the motor is braking are illustrated in
PWM duty cycle is selected to obtain the maximum average
Fig.7(a) and Fig.8(a) respectively.
regenerative current according to the brake initial speed, 2)
The experimental system of regenerative braking for The average regenerative current can not exceed the
electric motorcycle consists of a 24V storage battery, a maximum allowable charge limited current of valve-regulated
three-phase MOSFET inverter, a brushless PM motor, a DC lead-acid (VRLA) [16], 3) The braking torque should be
motor and control circuits. The parameters of brushless PM suitably controlled considering the riding feeling and safeness
motor used in the experiment are the same as that used in the especially on downhill roads, 4) Mechanical braking is
simulation. In the experiment, a brushless PM motor and a essential in emergency.
DC motor compose a reversible drive system to simulate This paper defines three braking modes “downhill roads
half-bridge-modulation regenerative braking condition, i.e., mode”, “flat roads mode” and “emergency mode”. The
“downhill roads mode” is switched by turning on an on-off signal of dynamic PWM duty cycle is obtained from
switch when the electric motorcycle goes into downhill roads equations (21) in which α b1 and α b 2 is the tolerance of
[17], and the handlebar is designed bi-directional in order to duty cycle for two cases respectively and this paper sets
drive forward in clockwise direction and regenerative brake lower limit motor speed nmin = (1 − δ )n0 and upper limit
in anticlockwise direction. In additional, the mechanical
motor speed nmax = (1 + δ )n0 considering the riding feeling
handbrake is equipped with a pressure sensor to check
whether the electric motorcycle is going to work in and safeness as the maximum tolerance rate of downhill
“emergency mode”. speed δ , i.e. D Downhill (n1 ) .
α [ k + 1] = α [ k ] + α b1 , ifα [k ] < α n[ k ] & nmin < n[k ] < nmax
A. Maximal Harvest Duty-cycle Model 
As the voltage of battery is basically changeless in the α [ k + 1] = α [ k ] − α b1 , ifα [ k ] > α n[ k ] & nmin < n[k ] < nmax (21)

experiment, so when the average regenerative current reaches α [ k + 1] = α [ k ] + α b 2 , ifn[k ] ≥ nmax
a maximum, the regenerative feedback power reaches a α [ k + 1] = α [ k ] − α , ifn[ k ] ≤ n
 b2 min
maximum too. Therefore, Maximal Harvest Duty-cycle At the beginning of downhill roads, the rider turns on the
Model [18] is tentatively established to realize the simple but switch to work the motorcycle in downhill roads mode at n0 ,
effective regenerative braking control scheme considering the and the optimal duty cycle α n0 making average
actual application. It focuses on the dual goals of the safe regenerative current maximum is selected by way of a certain
braking process and the maximal power harvest [5]. transition from 0 to α n 0 to keep the process smooth. When
According to the three braking modes, the Maximal the transition comes to the end, the motor reaches a new
Harvest Duty-cycle Model is established as follows, where speed n[1] , and then the equations (21) come into effect.
n1 is the motor speed, and the symbol Ω represents the From this model, the PWM duty cycle α of next control
Brake Depth varied with the handlebar turned in period is always selected to approach the one which makes
anticlockwise direction. average regenerative current maximum according to the
 D Downhill (n1 ) , Downhill Roads Mode speed of current control period.
 (20)
 D Flat (Ω ) , Flat Roads Mode 2 Flat Roads Mode & Emergency Mode
 0 , Emergency Mode The two modes are different from the downhill roads mode

The output value of Maximal Harvest Duty-cycle Model is for no gravitational potential energy. When the electric
the PWM duty cycle α . motorcycle works in “flat roads mode”, the rider brakes the
electric motorcycle by turning the handlebar in anticlockwise
1 Downhill Roads Mode direction, i.e. Brake Depth Ω , and the PWM duty cycle
It aims at the smooth and reliable speed limit on downhill would be varied linearly to realize the regenerative braking.
roads [9], beside the goal of the maximal power harvest. Assume k is the positive slop of D Flat (Ω ) , thus, the model
Therefore, the speed fluctuation range and the dynamic duty of “flat roads mode” is established as follows:
cycles to harvest the maximal power are concerned in the D Flat ( Ω ) = k ⋅ Ω (22)
model D Downhill (n1 ) . While in “emergency mode”, the rider must press the
handbrake to stop the electric motorcycle immediately in
emergency. In this case, the negative-accelerate may be so
large that the too large braking torque at high speed might
result in a danger of turn over and the surge recovery current
would injure the VRLA. Therefore, the output value of
Maximal Harvest Duty-cycle Model in “emergency mode”
equals ZERO, and the electric motorcycle is stopped by
mechanical braking only.
(a) (b) B. Regenerative Braking Control Scheme
Fig.10. (a)Speed Vs PWM duty cycle making the current maximum under Fig.11 shows the configuration of the proposed
the constraints. (b) Prototype of electric motorcycle regenerative braking scheme based on Maximal Harvest
Eliminate the data points dissatisfied CONSTRAINT 2 in Duty-cycle Model.
Fig.9, and then the trajectory of n1 - α shows the ∆ n[ k + 1]
n*[k +1]
relationship between the downhill speed and the PWM duty α [k + 1] Tb
cycle making average regenerative current maximum as n[k ]
Tmg sin θ
Fig.10(a). This paper hopes that when the downhill speed is α [k ]
α
n[k ]
n[k ]
varied, the PWM duty cycle of next control period is also n

varied to harvest the maximal power all the time.

Fig.11. Configuration of the proposed regenerative braking control scheme
According to the curves in Fig.10(a), when the electric
motorcycle works in “downhill roads mode”, the control The actual speed signal of the kth control period n[k ] is
fed back and it is compared with the reference speed of the [1] K. T. Chau, C. C. Chan, and C. Liu, “Overview of permanent-magnet
brushless drives for electric and hybrid electric vehicles,” IEEE Trans.
k + 1th control period n[k + 1] based on Maximal Harvest Ind. Electron., vol. 55, no. 6, pp. 2246-2257, Jun. 2008.
Duty-cycle Model. The other two inputs of Maximal Harvest [2] S. K. Safi, P. P. Acarnley, and A. G. Jack, “Analysis and simulation of
Duty-cycle Model are the braking mode and the PWM duty the high-speed torque performance of brushless DC motor drives,” IEE
Proc. Electr. Power Appl., vol.142, no.3, pp.191-200, 1995.
cycle α [k ] which makes average regenerative current [3] Braun D H., GiImore T P., MaslowskiW A. “Regenerative Converter
maximum according to the speed of current control period for PWM AC Driver,” IEEE Trans. on Ind. Appl. 1994. 30 (5) ,
n[k ] . pp:1176-1184.
[4] D. W. Hart, Introduction to Power Electronics. Englewood Cliffs, NJ:
Fig.10(b) shows the overall view of the electric motorcycle Prentice-Hall, 1997.
designed in this paper. The electric motorcycle is driven by [5] Xiaoyu LIU, Chuang LIU, Ming LU, “Regenerative Braking Control
Strategies of Switched Reluctance Machine for Electric Bicycle,”
one BLDCM. Compared to most products, this prototype has Electrical Machines and Systems, 2008. ICEMS 2008. International
only improved the software of controller and the manual Conference on Publication Year: 2008 , pp: 3397 -3400.
handle without too many changes on hardware. Therefore, the [6] M. Marchesoni and C. Vacca, “New DC–DC converter for energy
storage system interfacing in fuel cell hybrid electric vehicles,” IEEE
cost effective is obvious and it is convenient to update the Trans.Power Electron., vol. 22, no. 1, pp. 301–308, Jan. 2007.
general ones to the novel ones with low cost. [7] J. W. Dixon and M. E. Ortlizar, “Ultracapacitors + DC–DC converters
Some basic driving experiments have been only tested in regenerative braking system,” IEEE Aerosp. Electron. Syst. Mag., vol.
17,no. 8, pp. 16–21, Aug. 2002.
qualitatively to verify the feasibility of the proposed optimal [8] K. Takahashi, H. Seki, and S. Tadakuma, “Safety driving control for
regenerative braking control scheme employed in this electric power assisted wheelchair based on regenerative brake,” in
prototype. When the electric motorcycle worked on long Proc.IEEE Int. Conf. Ind. Technol., Dec. 2006, pp. 2492–2497.
[9] Hirokazu Seki,Kenji Ishihara,and Susumu Tadakuma, “Novel
downhill roads, the process was relatively smooth and the Regenerative Braking Control of Electric Power-Assisted Wheelchair
temperature rising of VRLA can not be felt obviously. for Safety Downhill Road Driving,” IEEE TRANSACTIONS ON
Another experiment was tried on flat roads, the rider turned INDUSTRIAL ELECTRONICS, VOL. 56, NO. 5, MAY
2009,pp.1393-1400.
the handlebar in clockwise direction first to reach a certain [10] Ming-Ji Yang, Hong-Lin Jhou, Bin-Yen Ma, “A Cost-Effective Method
speed, and then turned it in anticlockwise direction to brake. of Electric Brake With Energy Regeneration for Electric
Owing to the optimal regenerative braking control scheme, Vehicles,”IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS,
JUNE 2009, 56(6), pp. 2203-2212.
the riding feeling was satisfactory and the noise usually made [11] Ying-Shing Shiao, Yong-Long Chen, Rong-Wen Huang,and
by the general mechanical braking was reduced in effect. Chung-Hong Huang, “Design and Implementation of a Bidirectional
Converter for Battery-Powered Motorcycles,” in Proceeding of the 23th
IV. CONCLUSION ROC Symp. on Elect. Pow. Eng., 2002, pp. 1194-1199.
[12] Y. M. Liu and L. K. Chang, “Design and Implementation of Energy
This paper has studied intensively the relationship between Regeneration in Brushless DC Motor with the Simplest Topology,” in
average regenerative braking current and PWM duty cycle in Proceeding of the 23th ROC Symp.on Elect. Pow. Eng., 2002, pp.
half bridge modulation mode of the regenerative braking 1447-1451.
[13] S. Lu, K. A. Corzine, and M. Ferdowsi, “A unique ultracapacitor direct
scheme for electric motorcycle. The proposed method could integration scheme in multilevel motor drives for large vehicle
only change the switching signals of motor driver to control propulsion,” IEEE Trans. Veh. Technol., vol. 56, no. 4, pp.
the inverse torque during the braking period and achieves the 1506–1515,Jul. 2007.
[14] Y. P. Yang and T. H. Hu, “A new energy management system of
goal of the maximum energy harvest without additional directlydriven electric vehicle with electronic gearshift and regenerative
converter or changing windings of the motor stator as the braking,” in Proc. Amer. Control Conf., Jul. 2007, pp. 4419–4424.
methods proposed by other papers. The feasibility of the [15] M. C. Tsai, M. C. Chou, and C. L. Chu, “Control of a variable-winding
brushless motor with the application in electric scooters,” in Proceeding
proposed method was proved by simulation results and some of the IEEE-IEMDC2001, 2001, pp.922-925.
basic experiments. Furthermore, a simple but effective [16] E.M.Valeriote, T.G.Chang, D.M.Joehim, “Fast Charging of Lead-Acid
regenerative energy harvest scheme for electric motorcycle Batteries,” Proeeedings of the Annual Battery Conference, Jan11-13,
1994.
has been tentatively proposed based on a novel Maximal [17] Sanyo electric co., “Electric Bicycle,” Japan Patent, JP2004149001A.
Harvest Duty-cycle Model under four constraints. Employing [18] Suchart Janjornmanit, Samart Yachiangkam, and Aswin Kaewsingha,
the scheme to electric motorcycle, the problems such as “Energy Harvesting from Exercise Bicycle,” Power Electronics and
Drive Systems, 2007, 27-30 Nov. 2007,pp.1138-1140.
driving range, safeness, noise pollution and cost effective
would be improved. Our future work will pay attention to the
actual effect and some important parameters for riding feeling
and safeness in actual application.
ACKNOWLEDGMENT
The authors would like to thank for the project supported
by The Natural Science and Technology Major Project
(2009ZX04010-033) and The Natural Science Foundation of
Shanghai (09ZR1411500).
REFERENCES