AbstractThis paper presents a simple and computationally

efficient algorithm for mapping magnet loss within field-

orientated controlled surface-mounted brushless AC PM
machines over a wide range of operating conditions. Induced
eddy current loss in rotor magnets can be a significant
proportion of the overall loss and should not be overlooked
when analysing thermal behaviour and efficiency over the
required torque-speed envelope. Employing finite element
analysis (FEA) to determine magnet loss at every load point
would be computationally intensive. Here, a technique is
proposed, where from a limited FEA the magnet loss over the
entire torque-speed envelope is derived, catering for both
rated flux and field weakened operation. The method requires
four discrete time-step FEA solutions accounting for open-
circuit, rated current in the quadrature axis, rated current in
the direct axis and reduced current in the direct axis. The loss
predictions from the FE analyses are then used to define a
functional representation of the magnet loss. The proposed
method has been validated on two surface-mounted brushless
AC PM machine designs showing good agreement with direct
FE predictions of the PM power loss.

Index Terms permanent magnet loss, surface-mounted
brushless PM machines, computationally efficient models.
I. INTRODUCTION
he accurate prediction of loss and its variation with
load is an important element in the design of electrical
machines [1]. Vehicle propulsion applications are
particularly demanding as the understanding of machine
efficiency over the entire working envelope and under
specific control is usually required [2-4]. The loss
derivation, in such cases, is a time demanding and
computationally intensive process requiring numerous
analyses to predict various loss components.
In general, there are two main loss components present
within an electric machine: mechanical and
electromagnetic. The mechanical power loss component is
attributed to the frictional effects within the bearing
assembly (bearing loss) and fluid dynamics or
aerodynamics effects within the motor body (windage or
drag loss) [5]. The electromagnetic loss component is
usually associated with active parts of the motor assembly
and includes the iron, winding and permanent magnet (PM)
loss components. These loss components result from
various electromagnetic interactions within the machine [7-
9].
Recently, there has been increased interest in more
universal techniques and methodologies for accurate and

Xiaopeng Wu, Rafal Wrobel and Phil H. Mellor are with the
Department of Electrical Engineering, the University of Bristol, Bristol,
BS8 1TR, UK (e-mail: xw13661@bristol.ac.uk, r.wroble@bristol.ac.uk,
p.h.mellor@bristol.ac.uk).
Chengning Zhang is with the Department of Vehicular Engineering,
Beijing Institute of Technology, Beijing, 100081, China (e-mail:
mrzhchn@bit.edu.cn).
computationally efficient derivation of the electromagnetic
loss components [1, 6, 31]. Of particular interest is the
automated generation of loss/efficiency maps and have
received some attention in the literature [1, 6]. Proposed
techniques for the calculation of iron and winding loss
components are based on functional representations of the
analysed loss components, where the loss function
parameters are informed from experiment and/or theoretical
analyses. A common approach makes use of a limited
number of finite element analyses (FEAs) to populate the
loss function parameters, and allows for rapid and accurate
loss derivation at multiple operating points across a
machines working envelope.
The power loss associated with the permanent magnet
(PM) rotor assembly is particularly important as excessive
rotor temperature may result in premature failure. High
rotor temperature will lead to a reduction in the torque and
in some severe cases irreversible demagnetization of the
PM array. Since heat is not easily dissipated from the
rotating PM assembly either the magnet loss has to be kept
at a manageable level or enhanced means of rotor cooling
need to be introduced. Moreover, the rotary rotor assembly
does not allow for a simple and reliable temperature
monitoring and protection. The continuous drive towards
high power-density and compact PM machine solutions
imposes the requirement of elevated temperature operation
to fully utilise physical properties of the active materials
used.
When reviewing the existing techniques of predicting
magnet loss, two main methods have emerged: numerical
and analytical [10-31]. Numerical time-step or frequency
domain FEA is commonly used to calculate the induced
eddy currents in the magnets from which corresponding
Joule losses are determined. Two-dimensional (2D) FEA is
used predominantly in the analysis of radial-flux machines.
For other less common machine topologies, e.g. axial-flux
and transverse-flux, and laminated PM array constructions,
three-dimensional (3D) FEA is required. The FE approach
is time consuming and computationally intensive, in
particular when 3D analysis is needed. This makes the FE
based approach difficult to be adopted within automated
design approaches and in multi-physics thermal analysis.
A variety of analytical techniques have been proposed
for predicting magnet loss. These are based on simplifying
assumptions of the field distribution and their applicability
is limited to selected machine topologies. The
simplifications made in order to find a solution for the
magnet loss mean that only a single loss mechanism can be
catered for. For example, some of the analytical techniques
account for the loss resulting from the stator slotting only
[13-19], whereas the other methods deal with armature
reaction [20-30]. Consequently whilst the analytical
approaches allows for a rapid solution and time effective
use their application is limited.
A Computationally Efficient PM Power Loss
Derivation for Surface-Mounted Brushless AC
PM Machines

Xiaopeng Wu, Rafal Wrobel, Phil H. Mellor, Chengning Zhang

T
978-1-4799-4775-1/14/$31.00 2014 IEEE 17


There have been some developments into hybrid
techniques combining a simplified magneto-static FEA
with an analytical formulae for estimating the magnitude of
the induced eddy current loss [31]. This approach benefits
from both methods providing accurate PM loss prediction
in a timely manner. However, a degree of proficiency in
using FEA is required to fully benefit from the hybrid
approach.
This paper proposes a computationally efficient
methodology for determining magnet loss in field-
orientated controlled surface-mounted brushless AC PM
machines over a wide operating range. The idea behind the
technique is similar to that discussed in [1, 6], where a loss
function is defined whose coefficients are found from a
limited number of time-stepping FEA solutions. The
magnet loss functional representation has been constructed
from observations over the manner in which the magnet
loss varies with operating point, obtained from exploratory
FEA. The variation in loss is identified as being dependent
upon three variables: the rotational speed n, the quadrature-
axis current Iq and direct-axis current Id.
II. MACHINE EXEMPLARS
Two machine design exemplars have been used to
evaluate the proposed approach. Motor version A is a
radial-flux, integer-slot, distributed-wound internal-rotor
PM machine topology. To minimise stator slotting effects
the design incorporates semi-permeable slot bridges, Fig.
1a). As a result the magnet loss will primarily stem from
the armature reaction effects.
Motor version B is a radial-flux fractional-slot outer-
rotor PM topology with an open-slot stator design, Fig. 1b).
Here, a significant magnet loss component stemming from
the stator slotting is expected. Selected details of the motor
designs are given in Table I.

TABLE I. MOTOR DESIGN DATA
Motor version A B
Number of poles 8 16
Number of slots 48 18
Number of phases 3 3
Rated speed 3600rpm 4000rpm
Rated torque 928Nm 35Nm
Rated power 350kW 14.7kW
Maximum speed 6000rpm 6000rpm
Motor outer diameter 480mm 175mm
Active length 210mm 55mm
Air-gap thickness 7.5mm 1mm
PM material SmCo NdFeB
Electrical resistivity of PM 86cm 180cm

Motor version A has a relatively large air-gap. This is to
accommodate for mechanical retention of the PM array and
particular operating conditions of the machine. Overall the
magnet loss for the motor version A is expected to be
relatively low due to the large air-gap and quasi-closed slot
construction. In contrast, motor version B is expected to
exhibit relatively high magnet loss. Both machines feature a
three-phase star-configured winding. The winding
arrangement is indicated in Fig. 1. Motor version A has a
double-layer short-pitched distributed winding whereas a
double-layer concentrated winding construction is used for
motor version B.
The combinations of poles and slots are p = 8, q = 48
and p = 16, q = 18 for motor versions A and B respectively.
Both machines have similar speed requirements and ratios
of the maximum to base speed, 1.7 and 1.5 for motor
versions A and B respectively. The laminated core packs
for both machines are made of SiFe (M300-35A), and the
PM material of motor version A is SmCo, whereas NdFeB
for motor version B. Whilst the strength of NdFeB is
greater than SmCo, e.g; Br = 1.30T, Hc = 987kA/m for
NdFeB (N42UH) as compared with Br = 1.16T, Hc =
865kA/m for SmCo (Recoma 33E), SmCo is more suited
for high temperature operation and is less prone to
demagnetisation. Also, SmCo provides better thermal
stability of Br, -0.04%/
0
C against -0.12%/
0
C for NdFeB
resulting in more consistent machine performance over a
wide range of temperatures. However the choice of NdFeB
would result in lower magnet loss because of its higher
electrical resistivity, 180cm against 86cm for
SmCo.
III. FEA MODELS
Since the two motor versions considered are of a radial-
flux topology 2D FEA can be used. At this stage of the
research, non-segmented PM array constructions are
considered only. Fig. 1 outlines geometries of the analysed
motor versions. To provide a more broad evaluation an
internal-rotor and an external rotor machine topologies have
been characterised. Due to rotational symmetry only a
section of the complete motor cross-section needs to be
modelled. The reduced FE model encompasses one octant
for motor version A and half the complete motor cross-
section for the motor version B.

a) Motor version A
A+ A+
A+
C-
C-
C-
C-
B+
B+
B+
B+
A-
Stator
Rotor
PM

b) Motor version B
B+
A-
A+
A+
A-
A-
A+
C-
C+
B-
B-
B+
B+
B-
C+
C-
C-
C+
Rotor
Stator
PM

Fig. 1. Outline of the analysed PM motor designs

Since the 2D FEA analysis does not account for end-
effects, the FE magnet loss predictions are likely to be
overestimated [31]. This is particularly significant for
machine designs with a low aspect ratio of active length to
18


outer diameter, or where the magnets are axially
segmented. There are various analytical correction
techniques that can be used to account for end-effects in the
magnet loss predictions [31]. Alternatively, a 3D FEA
would need to be employed. In this paper, no correction for
the end-effects in loss estimation is used. The main purpose
of the work is to illustrate the methodology and show that a
complete loss characterisation is possible from a limited
number of FEA solutions. If required the 2D FEA
calculations could be substituted with 3D FEA or end-effect
corrections incorporated.
The temperature of the PM array also has an important
influence on the loss predictions as the electrical
conductivity of PM material is temperature dependant. As
the electrical conductivity decreases with temperature rise,
a reduction of the PM power loss is expected at elevated
operating temperature. In this analysis a fixed temperature
for the PM array equal to 20
0
C has been assumed.
IV. PM POWER LOSS DERIVATION
The eddy current loss generated in the magnet array
stems from two effects. The first results from the
permeance variation caused by stator slotting (PPM-SE) and
the second from the armature reaction field (PPM-AR) [13-
30].

. (1)
The armature reaction loss is a consequence of the
higher order spatial harmonics of the winding distribution,
and, in the case of non-sinusoidal phase currents, temporal
harmonics. This loss component strongly depends on
control scheme and operating mode of the electrical
machine. Here, the phase current will be assumed to be
sinusoidal with any current control or high-frequency PWM
effects neglected. The relative importance of the two
magnet loss components can be evaluated by means of the
FE analysis.
A. Maximum Torque per Ampere Operation
In the constant torque operation region, the motor is
usually controlled at rated flux to yield maximum torque
per Ampere. With the non-salient rotor designs considered
here this operation corresponds to the phase current (Iph)
being aligned to the quadrature-axis, i.e. Id = 0, Iq = Iph. I
d
,
Iq nomenclature refers to the dq0 machine model
representation [32]. It has been shown in [13] that the
magnet loss from the slotting effect is proportional to the
rotational speed square (n
2
). Similarly the magnet loss from
the stator current reaction is proportional to the rotational
speed square (n
2
) and phase current square (Iph
2
) [25].
Consequently, in the maximum torque operating region
where Iph = Iq, (1) reduces to the functional relationship:

. (2)

In order to inform the functional representation of the
magnet loss given by (2), an initial FEA is required. The
first component of the right hand side in (2) can be derived
from an open-circuit FEA, whereas the second component
can be found from an FEA at rated excitation with Iph = Iq
and the same rotational speed as for the open-circuit
analysis. In the case of the second FEA the magnet loss
component from the armature reaction only can be
separated by subtracting the magnet loss at open-circuit
operation obtained from the first FEA.
The stator slotting and stator reaction magnet loss
components can be then scaled according to (2) for any
operating speed and excitation current:

, (3)

where the coefficients d and a are determined as follows:

=

, =

. (4)

and nR and IqR are the rated rotational speed and current at
which the PPM-SE and PPM-AR were derived from FEA.

Motor version A


b) Motor version B

Fig. 2. Magnet loss versus Iq and speed maximum torque per Ampere
operation

Fig. 2 presents the magnet losses calculated using FEA
over a range of quadrature axis current and speed operating
points. These are compared to the loss trends given by the
functional relationship (3). The loss function correlates well
with the directly derived data. The small discrepancy
present at higher excitation currents is a consequence of
magnetic saturation of the core material affecting the
amplitude of the stator reaction field. The magnetic
saturation softens the severity of change of the magnetic
flux seen by the PM array and consequently resulting in
reduced magnet loss. It is possible to include the magnetic
saturation effect in (3), but this would require additional
0 100 200 300 400 500 600 700
0
50
100
150
200
250
300
350
P

[
W
]
I
q
[A
rms
]


n=3600
n=3000
n=2000
n=1000
Scaling from (2)
Results from FEA
0 50 100 150 200 250 300 350
0
1000
2000
3000
4000
5000
6000
P

[
W
]
I
q
[A
rms
]


n=4000
n=2000
n=1000
Scaling from (2)
Results from FEA
19


FEA to define the form of the saturation relationship.
In Fig. 2 the parameters a, d of the magnet loss function
(3), were obtained from the following operating points:

1) motor version A open-circuit operation at
nR = 3600rpm and rated excitation at IqR = 450A
rms
and
nR = 3600rpm;
2) motor version B open-circuit operation at
nR = 4000rpm and rated excitation at IqR = 177A
rms
and
nR = 4000rpm.

As anticipated, the magnet loss for the motor version A
is low considering the machine rated power output. The
slotting effect was found to be negligible. For motor
version B, the overall magnet loss would be very high with
the slotting effect now being the dominant mechanism.
Clearly in this case the magnet loss is excessive and means
of loss mitigation would need to be introduced in the
design, for example introducing axial segmentation in the
PM array construction. The effect of segmentation is to
increase the path length for the induced eddy currents in the
magnet, thereby reducing the eddy current magnitude and
loss. Therefore segmentation is unlikely to affect the form
of the loss relationship (2), and could be accounted for by a
geometric scaling of the coefficients. Treatment of
segmentation requires 3D FEA and will be investigate by
the authors in the future.


Fig. 3. Magnet loss versus Iq at n = 4000rpm for different values of Id
motor version A


Fig.4. Magnet loss versus Id at n = 4000rpm for different values of Iq
motor version A

a) Id = 50Arms

b) Id = 100Arms

c) Id = 300Arms

Fig. 5. Magnet loss versus Iq trends for differing values of speed and Id
motor version A
B. Field Weakening Operation
Section A proposes a loss scaling rule (3) for the
constant torque region with maximum torque per Ampere
operation. This section considers the form of the magnet
loss in the constant power, field weakened region of the
torque-speed envelope commonly used in traction
applications [3, 4].
In the constant power region the resultant stator
magnetic flux is weakened by injecting a direct-axis current
component which opposes the magnet excitation. The stator
current will now contain both a torque producing
quadrature axis current and field controlling direct axis
0 50 100 150 200 250 300 350 400
0
20
40
60
80
100
120
140
160
180
200
I
q
[A
rms
]
P

[
W
]


Id=100A
rms
Id=200A
rms
Id=300A
rms
Id=400A
rms
50 100 150 200 250 300 350 400
0
50
100
150
I
d
[A
rms
]
P

[
W
]


Iq=0A
rms
Iq=100A
rms
Iq=200A
rms
Iq=300A
rms
50 100 150 200 250 300 350 400
0
50
100
150
200
250
300
350
400
I
q
[A
rms
]
P

[
W
]


Scaling from (5)
n=4000rpm
n=5000rpm
n=6000rpm
50 100 150 200 250 300 350 400
0
50
100
150
200
250
300
350
I
q
[A
rms
]
P

[
W
]


Scaling from (5)
n=4000rpm
n=5000rpm
n=6000rpm
50 100 150 200 250 300 350 400
0
50
100
150
200
250
300
350
I
q
[A
rms
]
P

[
W
]


Scaling from (5)
n=4000rpm
n=5000rpm
n=6000rpm
20


current, Iph (Iq, Id), Id 0. The magnet loss trend over this
region of operation was evaluated for the two machine
versions through FEA over a range of Iq and Id excitation
currents, and speeds. Figs. 3 and 4 present selected results
of the FEA magnet loss calculation for motor versions A at
a speed of 4000rpm. The results indicate that an increase of
Iq and/or Id results in rise of the motor loss, Fig. 3. A
characteristic minimum magnet loss point is evident at low
values of Id, Fig. 4. Similar trends have been observed at
higher operating speeds for both analysed machines. Here,
for reasons of conciseness the loss trends have been
illustrated on the motor version A only.

a) n = 4200rpm

b) n = 5000rpm

Fig. 6. Magnet loss versus Id trends for differing values of speed and Iq
motor version B

From observation of the magnet loss trends with n, Iq and
Id, a new functional representation of the PM loss
accounting for both constant torque and constant power
region is proposed in (5).

, (

). (5)

where (

) is a second order function of Iq and Id.

In (5) the armature reaction loss component has been
defined as a quadratic polynomial combining Iq and Id.
Further analysis of the loss trends indicated the following
functional relationship for the magnet loss:

. (6)
To fully define the coefficients in (6) four individual
time-stepping FEAs are required at a particular reference
speed nW in the field weakening region. To confirm the
accuracy of the proposed polynomial magnet loss function
a series of operating points have been assessed for both
motor variants over a range of values of Iq, Id and speed
using FEA. These are compared to the polynomial loss
trend in Figs. 5 and 6.
The coefficients b to c in (6) have been informed through
FEA at the following operating points for motor version A:

1) Iq = 0, Id1W = 45A
rm
, nW = 4000rpm;
2) Iq = 0, Id2W = 450A
rm
, nW = 4000rpm.

Similarly the magnet loss function for motor version B
was derived from the following operating points assumed in
the FEAs:

1) Iq = 0, Id1W = 20A
rms
, nW = 5000rpm;
2) Iq = 0, Id2W = 177A
rms
, nW = 5000rpm.

Table II lists the coefficients of the loss function (5).

TABLE II. QUADRATIC POLYNOMIAL COEFFICIENTS
Motor version

Coefficient
a [W/A
2
] b [W/A
2
] c [W/A] d [W]
A 6.3e-04 5.9e-4 -0.099 12
B 0.026 0.027 -20.0 4700

Figs. 5 and 6 compare PM loss derived directly from
FEAs and proposed scaling procedure. Various operating
points in the constant power region have been evaluated for
both machines. The scaled loss data show good agreement
with the FE loss predictions. Some discrepancy between the
results is seen at higher current excitation and as discussed
previously this is attributed to magnetic saturation.
V. COMPLETE PM LOSS SCALING PROCEDURE
This section summarises the generic procedure for
mapping the magnet loss over the entire torque-speed
according to (6).

i.) The magnet loss component from slotting effect at the
reference speed (nR) is calculated from the open-circuit
FEA and then the coefficient d is derived:

ii.) The magnet loss at the reference speed (nR) and rated
current for maximum torque per Ampere operation (IqR) is
FE calculated and coefficient a is then derived

where:

=

. (7)

The magnet loss trends suggest the variant with Iq is
independent of Id. Hence the coefficient a has been
0 20 40 60 80 100 120 140 160 180 200
1000
1500
2000
2500
3000
3500
4000
I
d
[A
rms
]
P

[
W
]


Scaling from (5)
Iq=50A
rms
Iq=100A
rms
Iq=150A
rms
0 20 40 60 80 100 120 140 160 180 200
1500
2000
2500
3000
3500
4000
4500
5000
5500
I
d
[A
rms
]
P

[
W
]


Scaling from (5)
Iq=50A
rms
Iq=100A
rms
Iq=150A
rms
21


assumed here to be constant over the entire torque-speed
envelope.

iii.) The magnet loss for two working points with Id current
only (Id1W) and (Id2W) at the reference speed in field
weakening region (nW) is FE calculated and then
coefficients b and c are derived.

,

where Id1W is assumed to be equal to 10% of rated current,
and Id2W is equal to rated current. The coefficients b and c
are derived from set of two equations.

, (8)

where:

, (9)

=

,


=

.


iv.) The complete functional representation of PM loss
catering for both constant torque and constant power
operating regions can be then written in the following form

.

The reference speed (nW) at which the loss coefficients
are evaluated should be set within the field weakened
regime of operation. Depending on a particular application
and operating points of interests operation with injected
direct axis current may not be of concern and the simplified
version of the magnet loss scaling function according to (3)
might be more applicable.
VI. CONCLUSIONS
A simple and computationally efficient approach for
mapping magnet loss within a field oriented controlled
surface-mounted brushless AC PM machine has been
presented, catering for both rated flux and field weakened
operation. The proposed technique requires input from four
discrete time-step finite element field solutions,
corresponding to i) open-circuit operation, ii) rated current
with Iq only at maximum torque per Ampere operation, iii)
rated current and iv) 10% of rated current with I
d
only,
during field weakening operation. The finite element
analyses are used to obtain parameters for the proposed
functional representation of the magnet loss accounting for
the slotting effect and armature reaction. This approach
alongside with the standard dq0 circuit analysis has been
shown to provide a high fidelity loss map over the entire
machine working envelope.
The proposed technique of predicting motor loss has
been validated against detailed FEA on a 48 slots and 8
poles AC surface mounted PM internal rotor integer slot
machine and 18 slots and 16 poles AC surface mounted PM
outer rotor fractional slot machine, showing good
agreement over the majority of the operating envelope. Due
to the saturation, a small discrepancy is visible at higher
excitation currents. Accuracy of the proposed loss function
can further improved by considering a saturation
coefficient. In rotor designs that employ a segmented PM
array construction the 2D FEA loss calculations could be
substituted with values obtained using 3D FEA or through
end effect/segmentation correction factors. These additional
elements will be investigated in the future work.
VII. REFERENCES
[1] P.H. Mellor, R. Wrobel, D. Holiday, "A Computationally Efficient
Iron Loss model for Brushless AC Machines that Caters for Rated
Llux and Field Weakened Operation," IEEE International
Conference on Electrical Machine and Drives, 2009, IEMDC09, pp.
490-494.
[2] J. Goss, R. Wrobel, P.H. Mellor, D. Staton, "The Design of AC
Permanent Magnet Motors for Electric Vehicles: A Design
Methodology," IEEE International Conference on Electric Machines
and Drives, 2013, IEMDC13, pp. 871 878.
[3] R. Wrobel, J. Goss, A. Mlot, P.H. Mellor, "Design Considerations of
a Brushless Open-Slot Radial-Flux PM Hub Motor," IEEE Trans.
Ind. Appl., vol. 50, no. 3, pp. 1757-1767, 2014.
[4] J. Goss, P.H. Mellor, R. Wrobel, D.A. Staton, M. Popescu, "The
Design of AC Permanent Magnet Motors for Electric Vehicles: A
Computationally Efficient Model of the Operational Envelope," 6
th

IET International Conference on Power Electronics, Machines and
Drives, 2012, PEMD12, pp. 1 6, 2012.
[5] R. Wrobel, G. Vainel, C. Copeland, T. Duda, D. Staton, P.H.
Mellor, "Investigation of Mechanical Loss and Heat Transfer in An
Axial-flux PM Machine," Energy Conversion Congress and
Exposition, 2013, ECCE13, PP. 4372 4379.
[6] R. Wrobel, D.E. Salt, A. Giffo, P.H. Mellor, "Derivation and Scaling
of AC Copper Loss in Thermal Modeling of Electrical Machines,"
IEEE Trans. Ind. Elect., vol. 61, no. 8, pp. 4412 4420, 2014.
[7] D. A. Howey, A. S. Holmes, K. R. Pullen, Measurement and CFD
Prediction of Heat Transfer in Air-Cooled Disc-Type Electrcal
Machines, IEEE Trans. Ind. Appl., vol. 47, no. 4, pp. 1716 1723,
2011.
[8] D. A. Howey, A. S. Holmes, K. R. Pullen, Measurement of Stator
Heat Transfer in Air-Cooled Axial Flux Permanent Magnet
Machines, 35
th
IEEE Industrial Electronics Annual Conference,
2009, IECON09, pp. 1197 1202.
[9] A. C. Malloy, R. F. Martinez-Botas, M. Jaensch, M. Lamperth,
Measurement of Heat Generation Rate in Permanent Magnet
Rotating Electrical Machines, 6
th
IET International Conference on
Power Electronics, Machines and Drives, 2012, PEMD12, pp. 1 6,
2012.
[10] X. F. Ding and C. Mi, "Modeling of Eddy Current Loss in the
Magnets of Permanent Magnet Machines for Hybrid and Electric
Vehicle Traction Application," Vehicle Power and Propulsion
Conference, 2009, pp. 419-424.
[11] K. Yoshida, Y. Hita, K. Kesamaru, "Eddy-Current Loss Analysis in
PM of Surface-mounted-PMSM for Electric Vehicles," IEEE Trans.
Magn., vol. 36, no. 4, pp. 1941-1944, 2000.
[12] M. Nakano, H. Kometani, M. Kawamura, "A Study on Eddy-Current
Losses in Rotors of Surface Permanent Magnet Synchronous
22


Machines," IEEE Trans. Ind. Appl., vol. 42, no. 2, pp. 429-435,
2006.
[13] L. J. Wu, Z. Q. Zhu, D. Staton, M. Popescu, and D. Hawkins,
"Analytical Modelling and Analysis of Open-circuit PM Power Loss
in Surface-mounted Permanent Magnet Machines," IEEE Trans.
Magn., vol. 48, no. 3, pp. 1234-1246, 2011.
[14] Z. X. Fang, Z. Q. Zhu, L. J. Wu, Z. P. Xia, "Simple and Accurate
Analytical Estimation of Slotting Effect on PM Power Loss in
Fractional-Slot Surface-Mounted PM Machines," IEEE International
Conference on Electric Machine, 2012, ICEM12, pp. 464-470.
[15] D. A. Wills and M. J. Kamper, "Analytical Prediction of Rotor Eddy
Current Loss due to Stator Slotting in PM Machines," Energy
Conversion Congress and Exposition, 2010, ECCE10, pp. 992-995.
[16] M. Markovic and Y. Perriard, "A Simplified Determination of the
Permanent Magnet (PM) Eddy Current Losses due to Slotting in A
PM Rotating Motor," IEEE International Conference on Electrical
Machines and Systems, 2008, ICEMS08, pp. 309-313.
[17] J. Alexandrova and H. Jussila, "Comparison between Models for
Eddy-current Loss Calculations in Rotor Surface-mounted
Permanent Magnets," IEEE International Conference on Electrical
Machines, 2010, ICEM10, pp. 978-982.
[18] S. M. Sharkh, A. Ali Qazalbash, N. T. Irenji, R. G. Wills. "Effect of
Slot Configuration and Air-gap and Magnet Thicknesses on Rotor
Electromagnetic Lloss in Surface PM Synchronous Machines," IEEE
International Conference on Electric Machine and Systems, 2011,
ICEMS11, pp. 1-6.
[19] F. Caricchi, F. Giulii, F. Crescimbini and L. Solero Capponi,
"Experimental Study on Reducing Cogging Torque and Core Power
Loss in Axial-Flux Permanent-Magnet Machines with Slotted
Winding," 37
th
IEEE Annual Industry Applications Conference,
2002, vol. 2, pp. 1295-1302.
[20] N. Schofield, K. Ng, Z. Q. Zhu, and D. Howe, "Parasitic Rotor
Losses in A Brushless Permanent Magnet Traction Machine," IEEE
International conference on Electrical Machine and Drives, 1997,
ICEMS97, pp. 200-204.
[21] H. Polinder and M. J. Hoeijmakers, "Eddy-current Losses in the
Permanent Magnets of A PM Machine," IEEE International
Conference on Electrical Machines and Drives, 1997, IEMDC97,
pp. 138-142.
[22] H. Polinder and M. J. Hoeijmakers, "Eddy-current Losses in the
Segmented Surface-mounted Magnets of A PM Machine," IEE Proc-
Electric Power Appl., vol. 146, no. 3, pp. 261-266, 1999.
[23] Z. Q. Zhu, K. Ng, N. Schofield, and D. Howe, "Analytical Prediction
of Rotor Eddy Current Loss in Brushless Machines Equipped with
Surface-mounted Permanent Magnets. I. Magnetostatic field model,"
IEEE International Conference on Electrical Machines and System,
2001, ICEMS01, pp. 806-809.
[24] N. Schofield, K. Ng, Z. Q. Zhu, and D. Howe, "Parasitic Rotor
Losses in A Brushless Permanent Magnet Traction Machine," IEEE
International Conference on Electrical Machines and Drives, 1997,
IEMDC97, pp. 200-204.
[25] K. Atallah, D. Howe, P. H. Mellor, and D. A. Stone, "Rotor Loss in
Permanent-magnet Brushless AC Machines," IEEE Trans. Ind. Appl.,
vol. 36, no. 6, pp. 1612-1618, 2000.
[26] D. Ishak, Z. Q. Zhu, and D. Howe, "Eddy-current Loss in the Rotor
Magnets of Permanent-magnet Brushless Machines having a
Fractional Number of Slots per Pole," IEEE Trans. Magn., vol. 41,
no. 9, pp. 2462-2469, 2005.
[27] J. Wang, K. Atallah, R. Chin, W. M. Arshad, and H. Lendenmann,
"Rotor Eddy Current Loss in Permanent Magnet Brushless AC
Machines," IEEE Trans. Magn., vol. 46, no. 7, pp. 2701-2707, 2010.
[28] K. Yanazaki and A. Abe, "Loss Investigation of Interior Permanent
Magnet Motors Considering Carrier Harmonics and Magnet Eddy
Currents," IEEE Trans. Ind, Appl., vol. 45, no. 2, pp. 659-665, 2009.
[29] N. Bianchi and S. Bolognani, "An Overview of Rotor Losses
Determination in the Three-phase Fractional-slot PM Machines,"
IEEE Trans. Ind, Appl., vol. 42, no. 2, pp. 429-435, 2010.
[30] W. Y. Huang, A. Bettayeb, R, Kaczmarek, and J. C. Vannier,
"Optimization of Magnet Segmentation for Reduction of Eddy-
Current Losses in Permanent Synchronous Machine," Trans. Energy
Convers., vol. 25, no. 2, pp. 381-387, 2010.
[31] P. Zhang, G.Y. Sizov, J. He, D.M. Ionel, N.A.O. Demerdash,
"Calculation of Magnet Losses in Concentrated-Winding Permanent-
Magnet Synchronous MachinesUsing a Computationally Efficient
Finite-Element Method," IEEE Trans. Ind. Appl., vol. 49, no. 6, pp.
2524 - 2532 , 2013.
[32] R. Krishnan, "Permanent Magnet Synchronous and Brushless DC
Motor Drives," Virginia, 2010.
VIII. BIOGRAPHIES
Xiaopeng Wu received B.Eng. degree from Beihang University, Beijing,
China in 2006. He is currently working towards an Eng. D degree with the
Beijing Institute of Technology and working as a visiting scholar in the
University of Bristol. His research interests include the design and
modelling of electrical machines in traction application.

Rafal Wrobel received the M.Sc. Eng. degree from the Technical
University of Opole, Poland, and the Ph.D. degree from the Technical
University of Lodz, Poland, in 1998 and 2000 respectively. In 2013 he
received Habilitation degree from the Technical University of Opole,
Poland.
From 2001 he was with the Technical University of Opole as an Assistant
Professor. In 2002, he joined the University of Bristol, UK, as a Research
Fellow. He is currently a Senior Research Fellow within the Electrical
Energy Management Group, University of Bristol, UK. His research
interests include development and application of techniques and
methodologies for multi-physics design-optimisation of electrical
machines, transformers and passive wound components.

Phil H. Mellor received the B.Eng. and Ph.D. degrees in electrical
engineering from the Department of Electrical Engineering, Liverpool
University, Liverpool, U.K., in 1978 and 1981, respectively.
He is currently a Professor of electrical engineering with the Department
of Electrical and Electronic Engineering, University of Bristol, U.K. Prior
to this, he held academic posts at the University of Liverpool, Liverpool,
U.K. (1986 to 1990) and the University of Sheffield, Sheffield, U.K. (1990
to 2000). His research activities include high-efficiency electric drives, and
actuation and generation systems for application in more electric aircraft
and hybrid-electric vehicles.

Chengning Zhang received the M.E. degree from control theory and
control engineering and the Ph.D. degree in vehicle engineering from the
Beijing Institute of Technology, Beijing, China, in 1989 and 2001,
respectively.
He is currently a Professor and Vice Director of the National Engineering
Laboratory for Electric Vehicles, Beijing Institute of Technology. His
research interests include electric vehicles, vehicular electric motor drive
systems, battery management systems, and chargers.










23
Powered by TCPDF (www.tcpdf.org)