1664 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 54, NO. 3, JUNE 2007
Sensorless Indirect-Rotor-Field-Orientation Speed
Control of a Permanent-Magnet Synchronous
Motor With Stator-Resistance Estimation
Mohamed Rashed, Peter F. A. MacConnell, A. Fraser Stronach, and Paul Acarnley
Abstract—Efficient and precise sensorless speed control of a
permanent-magnet synchronous motor (PMSM) requires accu-
rate knowledge of rotor flux, position, and speed. In the litera-
ture, many sensorless schemes have been presented, in which the
accurate estimation of rotor flux magnitude, position, and speed
is guaranteed by detecting the back electromotive force (EMF).
However, these schemes show great sensitivity to stator resistance
mismatch and system noise, particularly, during low-speed opera-
tion. In this paper, an indirect-rotor-field-oriented-control scheme
for sensorless speed control of a PMSM is proposed. The rotor-flux
position is estimated by direct integration of the estimated rotor
speed to reduce the effect of the system noise. The stator resistance
and the rotor-flux speed and magnitude are estimated adaptively
using stable model reference adaptive system estimators. Simple
stability analysis and design of the estimators are performed using
linear-control theory applied to an error model of the PMSM
in a synchronous rotating reference frame. The convergence of
rotor position- and speed-estimation errors to zero is guaranteed.
Experimental results show excellent performance.
Index Terms—Parameters identification, permanent-magnet
synchronous motor (PMSM), sensorless control, speed control.
NOMENCLATURE
ψ
r
Rotor-flux space vector.
i
s
Stator-current space vector.
u
s
Stator-voltage space vector.
ω
r
Rotor angular speed in electrical radians per second.
ρ
r
Rotor position.
ω
e
Synchronous rotating-reference-frame speed in electri-
cal radians per second.
ρ
e
Position of synchronous rotating reference frame in elec-
trical radians.
ψ
r
Actual rotor (magnet)-flux magnitude.
R
s
Stator-winding resistance.
L
s
Stator-winding inductance.
Manuscript received April 6, 2005; revised January 3, 2006.
M. Rashed is with the Electrical Engineering Department, Mansoura Univer-
sity, Mansoura 35516, Egypt (email: m.rashed@eng.abdn.ac.uk).
P. F. A. MacConnell is with the School of Engineering and Physical Sciences,
University of Aberdeen, Aberdeen, AB24 3UE, U.K.
A. F. Stronach was with the School of Engineering and Physical Sciences,
University of Aberdeen, Aberdeen, AB24 3UE, U.K. He is now with Genbase
Information Systems, Aberdeen, AB24 3UE, U.K.
P. Acarnley is with the School of Engineering, Robert Gordon University,
Aberdeen, AB10 1FR, U.K., and also with the University of Aberdeen,
Aberdeen, AB24 3UE, U.K.
Digital Object Identifier 10.1109/TIE.2007.895136
x, s Subscripts, real, and imaginary components in synchro-
nous rotating reference frame.
α, β Subscripts, real, and imaginary components in stator
reference frame.
I. I NTRODUCTION
P
ERMANENT-MAGNET synchronous motors (PMSMs)
are known to provide higher torque per unit volume and
better efficiency than induction motors, while improvements in
the properties of permanent-magnet materials have increased
their viability. Recently, sensorless PMSM drives have received
increasing interest for industrial applications where there are
limitations on the use of a position sensor. The elimination
of the position sensor reduces the cost of the drive and
increases the overall system ruggedness and reliability. High-
performance operation of sensorless PMSM drives mainly re-
lies on accurate knowledge of the rotor-magnet flux magnitude,
position, and speed. The sensorless rotor-position estimation
techniques can be classified into two major groups: the motor-
model-based and the rotor-saliency-based techniques. The latter
are suitable only for the interior PMSM (IPMSM). Most of
the motor-model-based techniques detect the back-EMF vec-
tor, which holds information about the rotor position and
speed, using either open-loop estimators [1]–[3] or closed-loop
estimators/observers [4]–[9]. In other motor-model-based tech-
niques, the rotor-flux vector is directly estimated [10], [11].
Moreover, adaptive observers have been used to estimate the
stator current, the rotor speed, and the rotor position [12]–[15].
Extended Kalman filters (EKF) have also been proposed for ro-
tor speed and position estimation [16], [17]. Although the EKF
algorithm is stable and well known, the EKF is computationally
intensive and requires proper initialization. Furthermore, model
reference adaptive system (MRAS) has been used extensively
for combined rotor flux and speed estimation in induction
motors [18].
In [1], open-loop back-EMF-based position and speed
estimators have been considered. The scheme is sensitive to
the system noise and to the stator resistance mismatch. In
[4], the rotor-flux vector is estimated by direct integration of
the calculated back EMF. The estimated rotor-flux vector is
corrected adaptively by utilizing the error between measured
and estimated stator currents to reduce the effect of integration
drift. However, the scheme is not suitable for low-speed
operation. On the other hand, in [12], an identity observer
has been proposed to estimate the rotor speed by utilizing the
0278-0046/$25.00 © 2007 IEEE
Authorized licensed use limited to: Newcastle University. Downloaded on June 23,2010 at 13:37:29 UTC from IEEE Xplore. Restrictions apply.