Andreas EILENBERGER, Manfred SCHROEDL

Technical University of Vienna

Extended Back EMF model for permanent magnet synchronous

machine with different inductances in d- and q-axis
Abstract. This paper discusses a reluctance-dependent back electromotive force (EMF) model for encoderless vector driven permanent magnet
synchronous machines (PMSM) compared to well known, non reluctance-dependent back EMF models. Established PMSM could have considerable
varieties in direct and quadrature inductances. So it makes sense to consider this behaviour in an extended back EMF model. On closer examination
it will turn out that a reluctance dependent EMF model has a better behaviour at low speed as the standard model. Also the derivation of the
extended model will be illustrated. Furthermore some simulation results and practical measurements will be discussed.

Streszczenie. W artykule przedstawiono wpływ zależnej od reluktancji siły elektromotorycznej (EMF) dla bezczujnikowego sterowania maszyny
synchronicznej z magnesami trwałymi (PMSM) w porównaniu z powszechnie znanym modelem nie uwzględniającym tego wpływu. Założony model
PMSM może mieć różne warianty. Z tego powodu rozważono zachowanie układu z rozszerzonym modelem siły elektromotorycznej EFM. Z badań
wynika, że model zależnej od reluktancji siły elektromotorycznej ma lepsze właściwości przy małych prędkościach obrotowych niż model
standardowy. Pokazano wyprowadzenie zależności modelu rozszerzonego. Przedstawiono także wybrane wyniki badań symulacyjnych i
laboratoryjnych. (Rozszerzony model EMF dla maszyny synchronicznej z magnesami trwałymi z różnymi indukcyjnościami w osiach d- i q)

Keywords: Electrical Drive, Permanent magnet motor, Synchronous motor, Reluctance drive, Sensorless control.
Słowa kluczowe: napęd elektryczny, silnik z magnesami trwałymi, silnik synchroniczny, napęd reluktancyjny, sterowanie bezczujnikowe.

Introduction ⎡ψ ⎤
There are a lot of publications on the field of sensorless γ = arg ⎡⎣ψ m ⎤⎦ = arctan ⎢ mβ ⎥ =
driven PMSM, which means replacing the mechanical ⎣ψ mα ⎦
position encoder by mathematical models. Low speed (3)
⎧⎪ ⎣⎡usα (τ ) − rS iSα (τ ) ⎦⎤ dτ − lS iSα (τ ) ⎫⎪
models combined with a back EMF model for higher speed
= arctan ⎨
∫ ⎬
⎩⎪ ∫ ⎣ s β ( ) S S β ( ) ⎦
enables an operation in the whole speed range. The lower ⎡u τ − r i τ ⎤ dτ − lS iS β (τ ) ⎪
the speed the more sensitive the parameters of the EMF ⎭
model are. Such parameters are stator inductance and
stator resistance. In most cases there is a significant Extended back EMF model
difference between direct and quadrature inductance. Using well known two-axis theory [4], [5] the flux linkage
Hence, this paper describes an extended EMF model with ψS in the dq rotor-oriented reference frame is given by
regard to the saliency to achieve highly dynamical operation
in a wide speed range, especially at low speed. (4) ψ s , dq = ψ m , dq + ld id + jlq iq ⋅e jγ

The classical back EMF model ψ s ,αβ = ψ m,αβ + ( ld id + jlq iq ) ⋅ ( cos γ + j sin γ )
In the subsequent analyses for determining the
operational behavior of the given machine, steady-state ψ s ,αβ = ψ m ,αβ + ld id cos γ − lq iq sin γ
+ j ( lq iq cos γ + ld id sin γ )
operation in conjunction with constant rotor angular velocity
ωK is supposed. All data are given in normalized values.
The following calculations deal with the classical back EMF which yields to
model [1], [2] and [6]. The stator flux linkage ψS is derived
from the stator voltage equation (5) ψ mα = ψ sα − ld id cos γ + lq iq sin γ
d (6) ψ mβ = ψ sβ − ld id cos γ − lq iq sin γ
(1) uS = rS ⋅ iS + ψ S + jωKψ S
dτ
The rotor position γ follows with simplified notation to
by integration
⎧⎪ ( uα − rs iα ) dτ − ld id cos γ + lq iq sin γ ⎫⎪
∫
ψ S (τ ) = ∫ ⎡⎣ uS (τ ) − rS ⋅ iS (τ ) ⎤⎦ dτ . (7) arctan ⎨ ⎬.
⎪⎩ ∫ ( uβ − rs iβ ) dτ − ld id cos γ − lq iq sin γ ⎭⎪
Furthemore flux linkage ψm due to permanent magnets is
The block diagram of the realized extended EMF model
calculated from stator flux linkage ψS is shown in fig. 3. Furthermore an extended observer
(2) ψ S = lS ⋅ iS + ψ m structure as discussed in [3] is implemented.

Experimental setup
and yields in the αβ stator-oriented reference frame to For verification of the novel back EMF model, an
equations outerrotor PMSM was used, shown in fig. 2. Furthermore a
voltage source inverter with only a DC-link measurement is
ψ mα (τ ) = ∫ ⎡⎣usα (τ ) − rS ⋅ iSα (τ ) ⎤⎦ dτ − lS ⋅ iSα (τ ) used. A short overview of the given hardware is shown in
figure 1. The subsequent table 1 specifies the
ψ mβ (τ ) = ∫ ⎡⎣usβ (τ ) − rS ⋅ iS β (τ ) ⎤⎦ dτ − lS ⋅ iS β (τ )
characteristics of the used PMSM.
Following fig. 2 shows the used outer-rotor PMSM with
Normally, the integration is stabilized by a certain
assembled encoder and belt to load-machine.
feedback. The argument of flux linkage vector ψm is the
searched rotor position γ

106 PRZEGLĄD ELEKTROTECHNICZNY (Electrical Review), ISSN 0033-2097, R. 85 NR 7/2009

Fig. 1. Structure of the realized encoderless drive with an outer-rotor PMSM

Fig. 3. Schematic of the realized extended back EMF model with

extended observer structure [3]

Simulation results
For simulation of both back EMF models the simulation
tool MATLAB-Simulink is used. In the following, two
scenarios were simulated. First scenario shows results of
the calculated rotor position with the well known EMF model
with constant inductances and extended EMF model at
constant quadrature-axis current iq,ref. To verify results, the
reference rotor position is mapped. This simulation starts at
Fig.2. Picture of the used outer-rotor PMSM with encoder and load- standstill and accelerates rotor speed slowly. As be can
machine seen in fig. 4 the extended back EMF model provides a
useful angular rotor position at first while increasing the

PRZEGLĄD ELEKTROTECHNICZNY (Electrical Review), ISSN 0033-2097, R. 85 NR 7/2009 107

rotor speed. At higher speed, the standard back EMF model speed. Due to an offset error of the measured voltage
also yields an angular rotor position. The significant offset space phasor us,αβ the calculated rotor position with
between reference and calculated EMF angular rotor standard back EMF model has a visible positive offset.
position is a result of the integrator feedback. An offset on Subsequent fig. 7 depicts the calculated angular rotor
voltage space phasor usα is used in the simulation position with the mentioned two back EMF models at
representing voltage measurement errors of the constant rotor speed and quadrature current. The extended
experimental setup. model shows a better behaviour as the novel back EMF
The second scenario shows the behaviour of the EMF model.
models (fig. 5) at a constant rotor speed ω = 0.2 with a
rising quadrature current iq,ref.

Fig. 4. Characteristics of calculated rotor position with standard

back EMF model (blue) and extended back EMF model (yellow),
furthermore encoder position (red) and rising rotor velocity (green) Fig. 7. Measured characteristics with iq,ref = 1 and ! _ 0.15, Ch1:
with iq,ref = 0.8. extended model (22.5/Div.), Ch2: novel model (22.5/Div.), Ch3:
encoder position (22.5/Div.)

Conclusion and outlook

This paper demonstrates an extended back EMF model
with improved low-speed properties compared to the well
known back EMF model with constant inductances, both
with an observer structure. It is also shown that measured
results with the experimental setup shows expected
simulation results. Furthermore, the extended EMF model
was successfully implemented in a drive for light vehicles.

The authors are very much indebted to the Austrian

Science Foundation (FWF) which generously supports the
Rys. 5. Characteristics of calculated rotor position with standard work at the Institute of Electrical Drives and Machines at the
back EMF model (blue) and extended back EMF model (yellow), Vienna University of Technology.
encoder position (red) and the rising quadrature current iq,ref
REFERENCES
(green) with constant ω = 0.2 [1] Schroedl M., Hofer M., Staffler W. ”Combining INFORM method,
Voltage model and mechanical observer for sensorless control
of PM Synchronous Motors in the whole speed range including
standstill”, PCIM, Nuernberg (Germany), (2006).
[2] Rieder U.-H.: ”Optimierung der sensorlosen Regelung von
permanentmagneterregten Aussenlaeufer-
Synchronmaschinen”, PhD thesis, Vienna University of
Technology, 2005.
[3] Schroedl M., Hofer M., Staffler W. ”Sensorless control of PM
Synchronous Motors in the whole speed range including
standstill using a combined INFORM/EMF model”, EPE-PEMC,
Portoroz (Slovenia), (2006).
[4] Park R.H., ”Two-Reaction Theory of synchronous machines,
Generalized Method of Analysis”, AIEE Trans., pt. I, Vol.48, pp.
716, July 1929.
[5] Jones C. V., ”The unified theory of electrical machines”,
Butterworth & Co. (Publishers) Ltd.,pt. II, London, 1967.
[6] T.M. Wolbank, M.A. Vogelsberger, R. Stumberger, ”Adaptive
Flux model for commissioning of signal injection based zero
speed sensorless flux control of induction machines”
Rys. 6. Measured characteristics with a rising rotor speed and Proceedings of IEEE International Conference on Power
constant iq,ref = 0.5, Ch1: extended model (22.5/Div.), Ch2: novel Electronics and Drive Systems, Bankok (2007)
model (22.5/Div.), Ch3: encoder position (22.5/Div.)
Authors: Dipl.-Ing. Andreas Eilenberger, Technical University
Measurements results of Vienna, Institute for Electrical Drives and Machines,
The measured results verify the results of the two Gusshausstrasse 25-29/E372, A-1040 Vienna, E-mail:
simulated scenarios. First diagram (fig. 6) depicts the [email protected];
Dipl.-Ing. Dr. Manfred Schroedl,
measurement according to fig. 4 with increasing rotor
E-mail: [email protected]

108 PRZEGLĄD ELEKTROTECHNICZNY (Electrical Review), ISSN 0033-2097, R. 85 NR 7/2009