IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO.

1, JANUARY 2005 509

A Controller of Brushless DC Motor for

Electric Vehicle
Hong-xing Wu, Shu-kang Cheng, and Shu-mei Cui

Abstract—A controller of brushless dc motor (BLDCM) for

electric vehicle is developed based on the mathematical model of
BLDCM and special working condition. The drive and control
circuit is composed of digital signal processing and EXB841.
The paper analyzes the hardware structure of system in detail.
Proportional–integral–differential control strategy is adopted in
the system. An arithmetic variable velocity prelabeled value is
presented, in order to solve regulating rotate speed in large range
accurately. The paper also analyzes flux-weakening control in
high speed and method of energy regenerative braking when the
vehicle is decelerating.
Index Terms—BLDCM, DSP, flux-weakening control.
Fig. 1. Equivalent circuit diagram of BLDCM.
I. INTRODUCTION
for the motor which air gap field shape is sine. The transforma-
P EOPLE pay more attention to zero-polluting electrical ve-
hicles (EV) with the development of environmental pro-
tection consciousness. The dynamical system of an electrical
tion is feasible if multiparametric coordinate theory is used. But
it is very complex. Fig. 1 shows equivalent circuit diagram of
BLDCM. Calculation error is great if the fundamental wave is
vehicle is composed of storage battery and electromotor. DC
transformed. By contrast, it is easy and convenient to built the
motor, ac motor, and reluctance motor are usually used in the
mathematical model [2], [3] of the motor using intrinsic phase
electric vehicle. The brushless dc motor (BLDCM) has high ef-
variable of motor. The assumption is as follows.
ficiency and good performance of timing. It also has some strong
points of simple structure, high reliability, and convenient main- 1) The stator winding is concentrative winding equispaced
tenance. So it is used widely. Mechanical commutation is re- with 60 . It is “Y” connection.
placed electronic commutation in BLDCM. Electronic commu- 2) The winding is distributing on the smooth surface of stator
tator is made of high-power switch transistor. The switch tran- equably.
sistors are on or off according to the position of the rotor, so the 3) The magnetic saturation is neglected. The loss of vortex
BLDCM must combine with the driver. The paper introduces and magnetic hysteresis are neglected.
a driver of BLDCM based on the mathematical model of the 4) The armature reaction is neglected and the distribution of
air gap field is uniform distribution.
motor, and analyzes hardware and software of the driver. The
The voltage equation of three-phase BLDCM is
core circuit is composed of a digital signal processor (DSP) [1],
[2]. The rotate speed control of BLDCM is inducted a novel
control strategy of variable velocity prelabeled value, to realize
the rotate speed accurately in large range. The paper analyzes a
weak-flux control strategy at high speed for BLDCM. BLDCM
combined with the drive can work reliably. The controller has
(1)
low cost and preferable stability. It can better realize energy re-
generative braking on an electric vehicle.

II. MATHEMATICAL MODEL OF BLDCM (2)

The characteristic of BLDCM is that the shape of counter- (3)
electromotive force (EMF) is trapezium, which means that the where
mutual inductance between stator and rotor is nonsine. It is dif- , , phase winding voltage of stator (V);
ficult to transform the equation of BLDCM to the equation of , , phase winding current of stator (A);
and axis, because the equation of and axis is suitable , , phase winding electromotive force of stator;
phase winding inductance (H);
Manuscript received December 19, 2003. This work was supported in part by phase winding mutual inductance (H);
the Hi-tech Research and Development Program of China (2001AA501522). differential operator .
The authors are with Department of Electrical Engineering, Harbin In- So, electromagnet torque is
stitute of Technology, Harbin 150001, China (e-mail: whx0422@sina.com;
cskee@hit.edu.cn; cuism@hit.edu.cn).
Digital Object Identifier 10.1109/TMAG.2004.839304 (4)

0018-9464/$20.00 © 2005 IEEE

510 IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO. 1, JANUARY 2005

Fig. 4. Block diagram of PI regulation of speed.

to K1, and so on. The count of timer is optimum matching with

speed calculation. Speed precision may achieve 0.1%.

Fig. 2. Block diagram of electric vehicle structure. B. PI Control of Speed Loop

The difference value of speed is gained through speed feed-
back value subtracting reference value of speed. It is calculated
new reference current value using a simple PI arithmetic [5], [6]
(5)
where
output of speed regulator;
coefficient of proportionality;
integrating coefficient;
Fig. 3. Block diagram of variable velocity ore-labeled value. sampling period of speed.
The sampling frequency of speed is attained basing practical
state. This time, the sampling frequency of speed is equal to 62.5
III. CONTROL STRATEGY OF BLDCM IN EV s. Fig. 4 shows PI regulation of speed.
The motor used in driving system of EV should have the fol-
lowing attributes: simple structure, small volume, lightweight, C. Control of Current Loop
high performance, convenient maintenance, easy control, and The process of current regulation is also process of
large torque. Fig. 2 shows the simple structure of EV. There are pulsewidth modulation (PWM) signal production. The av-
some units in the electric vehicle. Most permanent-magnet ma- erage value is adjusted
chines have sinusoidal flux density in the air-gap and sine-phase
current. The permanent-magnet motor of square wave electro- (6)
motive is more advantageous than the others in electric vehicle duty cycle duty cycle (7)
[4]. through adjusting width of PWM signal. The value that refer-
The control method of permanent-magnet motor is difficult, ence current subtract from back-check current decides the width
which limits the application of this motor. After using square of PWM.
wave motor, the control and driving become easy. Because the is the parameter which connects with motor parameter and
motor of electric vehicle is always in the electrodynamic and de- main voltage. It is a constant when the motor and driver are
celerating state when the vehicle accelerates, brakes, upslopes, shapely.
and downgrades, the special timing system should be adapted. A width PWM signal is formed when goes through the
The motor is used as an electromotor when the vehicle is run- proportional-integral-differential (PID) regulator. The value of
ning; contrarily it is used as an electrical generator to charge the decides the width of PWM. The width of PWM is invariable
storage battery when the vehicle is decelerating. According to when the value of is equal to zero. If the value of is very
the order of vehicle management unit, the controller and driver large, that means reference current is bigger than practical cur-
of the motor control the charging current of battery. rent, the width of PWM wave is the chopping period. The speed
increases rapidly. If the value of is very little (negative), the
A. Arithmetic of Variable Velocity Prelabeled Value width of PWM wave will diminish and maybe equal to zero.
The system calculates speed using variable timer of prela- The speed may slow down over time. Fig. 5 shows the current
beled value, as shown in Fig. 3. It will improve the precision of regulation using PID control.
the system response. Speed calculation of the system is based on
the rotor position signal. Because the alnico of rotor is symmet- D. Flux-Weakening Control of BLDCM
rical, timer can calculate speed accurately at some speed range. When the EMF value of BLDCM reaches the voltage limit
But, it also produces error in the whole speed range, which provided by power, the system regulates the direct axis current
is not in favor of PI control of speed loop. Difference value and quadrature axis current in order to expand the
of speed named equals to reference value of speed named range of speed. It is the flux-weakening control method of
and feedback value named difference. If and motor. BLDCM, in which EMF wave is sine, increases the
, prelabeled value of Timer2 equals to K0 in order direct axis current realizing the flux-weakening control. In the
to realize high-precision speed control function of the system. BLDCM system, in which the EMF wave is trapezium, it is
If and , prelabeled value of Timer2 equals unable to control the phase position of stator winding current
WU et al.: A CONTROLLER OF BRUSHLESS DC MOTOR FOR ELECTRIC VEHICLE 511

Fig. 5. PID regulation of current loop in the controller.

directly. The equifinal method is to control the phase position

of stator winding voltage. The phase position of stator coil
current is controlled through controlling phase position of
stator winding voltage. It also realizes flux-weakening control
of BLDCM. Controlling lead angle of voltage is the equifinal Fig. 6. Vector diagram of flux-weakening control of BLDCM.
flux-weakening control strategy of BLDCM.
The method is to change phase frequency characteristic of
voltage detector, and the phase shift is 90-Y. That means lead the
triggering angle of BLDCM is Y. As we know, that difference
value of voltage on the winding and EMF of winding decides the
current in the winding. At the same time, the amplitude value
of EMF is direct ratio to rotate speed of motor. It can be seen
from the mathematical model of BLDCM that the lead angle of
BLDCM changes the voltage on the equivalent resistance and
counteracts the induced electromotive force. In order to meet
Fig. 7. Equivalent circuit diagram of regenerative braking.
the equation of mathematical model, it makes the rotate speed of
motor increase when changing lead angle makes voltage on the
equivalent resistance larger. It can realize flux-weakening con-
trol. The voltage wave of phase winding shows the superficies of
flux-weakening control. Because the voltage of phase winding
equal to the induced electromotive force when the switching
tube is close and the voltage of phase winding equal to voltage
of power when the switching tube is open. The induced electro-
motive force is bigger than voltage of power when the system
implement flux-weakening control. It is very useful for elec-
tric vehicle in high speed. Fig. 6 shows the vector diagram of
flux-weakening control of BLDCM. A system of equations and
equation about of is derived by Fig. 6

tag (8) Fig. 8. Block diagram of control system of BLDCM.

(9) The circuit current equation is

E. Regenerative Braking for EV (11)

is electrical time constant of the circuit.
An electric vehicle needs regenerative braking when it decel- The essential of regenerative braking is that kinetic energy of
erates or limits speed on the downgrade. At that time, the motor vehicle will translate into magnetic energy and store in the in-
rotate speed is under the no-load rating rotate speed. The paper ductance , when special switch tubes are open. The
discusses the line voltage when two phase is in series is bigger kinetic energy of vehicle and the energy stored in the induc-
than voltage provided by dc power. The method is named regen- tance translate into electric energy through voltage boosting of
erative braking. In the state, the only three switch tubes which inductance and storage battery. At the same time, because the
locate the same half bridge is on or off in the inverter circuit, direction of current of motor is contrary to the current of motor
the other is always close. Fig. 7 shows the equifinal schematic working in whole period, it may hold electromagnet torque.
diagram of controller at regenerative braking. So it will realize regenerative braking. The energy absorbed by
The circuit voltage equation is storage battery can be expressed by

(10) (12)
512 IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO. 1, JANUARY 2005

Fig. 11. Graph of speed response at 1420 r/min.

Fig. 9. Flowchart of system and PID control.

Fig. 12. Graph of brake current at 800 r/min.

Fig. 10. Flowchart of PID control interrupt. logic circuit decides the polarity of PWM pins when compare
matching takes place. The system uses the full compare unit
PWM mode. PWM wave is made of lots of serial pulses with
IV. REALIZATION OF PRACTICAL CIRCUIT varying width. Its period is regulated by the timer1 repeatedly.
The control circuit utilizes the special purpose chip for The value of period is regulated through the software. So the
control motor MS320F240. Fig. 8 shows the block diagram width of PWM wave may change through changing the value
of the control system. The function units are denoted in the of comparator in the current regulator and the voltage regulator.
dashed line. The chip may provide high-efficiency digital signal Then it will control the motor.
processing and control hardware of high stability. It gathers The system of BLDCM has three Hall sensors. The position
the high-speed calculation and professional motor control of motor rotor can be detected through line Hall sensors that is
technology. It can revise the parameter and control strategy, named A3517s. Its power of chip is 5 V. Its output is line signal
carry out fault diagnosis, communication function, and so on. whose range is 0–5 V and its zero point is 2.5 V. The output is
The software flow of the system is expressed in Figs. 9 and 10. the detection signal of three phase positions (Ha, Hb, Hc). These
The rotor position signal of BLDCM is captured through the signals are handled through filtering and signal disposing
capture unit. When there are saltus signals at the pins of capture circuit. The CAP of DSP is connected with the output of Hall
unit, the count of selected timer is locked in a buffer. DSP will sensors in order to catch rotor position. The Hall sensors cross
carry out time-lag and interrupt and initialize logic circuit. The 120 electrical degrees. It will form six signals whose phases
WU et al.: A CONTROLLER OF BRUSHLESS DC MOTOR FOR ELECTRIC VEHICLE 513

Fig. 15. Graph of phase current.

Fig. 16. Graph of phase current.

Fig. 13. Graph of brake current at 1500 r/min.

parameter of PID regulation optimized. Fig. 16 is the graph of

phase current after parameter of PID regulation optimized.

V. CONCLUSION
The driving system of BLDCM used in EV is studied in this
paper. The core of the driving system is the TMS320F240 chip.
The driver circuit consists of the driver chip EXB841 and pe-
ripheral circuits. The intelligent power module is used as the
main power circuit. The system has characteristics of compact
structure, advantageous control, and low cost. In the control
strategy, typical PI control and arithmetic of variable velocity
prelabeled value are applied. The flux-weakening control and
energy regenerative braking control are also analyzed. The fea-
sibility and high efficiency of the driving system are approved
through experiment.

REFERENCES
[1] S. C. Persyn, M. McClelland, M. Epperly, and B. Walls, “Evolution of
digital signal processing based spacecraft computing solutions,” in Proc.
20th DASC Conf., vol. 2, 2001, pp. 8C3/1–8C3/10.
Fig. 14. Simulated graph of phase current. [2] P. Crnosija, Z. Ban, and R. Krishnan, “Application of model reference
adaptive control with signal adaptation to PM BLDCM drives,” in Proc.
IEEE Int. Symp. ISIE, vol. 3, 2002, pp. 689–694.
[3] A. Rubaai, D. Ricketts, and M. D. Kankam, “Development and imple-
are variable through detection the change of output signal. The mentation of an adaptive fuzzy-neural-network controller for brushless
system will interrupt because of the detected signal, and execu- drives,” IEEE Trans. Ind. Appl., vol. 38, no. 2, pp. 441–447, Mar./Apr.
tive interruption program. 2002.
Fig. 11 shows the correspondence of the motor after receiving [4] S. Gui-Jia, J. W. McKeever, and K. S. Samons, “Modular PM motor
drives for automotive traction applications,” in Proc. 27th Annu. IEEE
the order from the vehicle manage unit. Figs. 12 and 13 show the Conf. IECON, vol. 1, 2001, pp. 119–124.
graph of the motor when electric vehicle is braking. The front [5] Y. Sozer, H. Kaufman, and D. A. Torrey, “Direct model reference adap-
of wave denotes electric drive. The axis is the time axis. The tive control of permanent magnet BLDCMs,” in Proc. IEEE Int. Conf.
system begins braking at 0.3 s. So the total current of controller Control Applications, 1997, pp. 633–638.
[6] , “Direct model reference adaptive control of permanent magnet
is negative value. Fig. 14 is the phase current in the motor at BLDCMs,” in Proc. IEEE Int. Conf. Control Applications, 1997, pp.
identical conditions. Fig. 15 is the graph of phase current before 633–638.