Proceedings of the 2nd World Congress on Electrical Engineering and Computer Systems and Science (EECSS'16)

Budapest, Hungary – August 16 – 17, 2016

Paper No. EEE 126
DOI: 10.11159/eee16.126

Stator Shape Optimization for Electrical Motor Torque Density

Improvement
Hea-Joong Kim, Hee-Deuk Jun, Youn Hwan Kim, Jae-Won Moon
Rotating Machinery Center / Korea Testing Certification
22 Heungan-daero 27beon-gil, Gunpo-si, Korea
motor@ktc.re.kr; jhdeuk@ktc.re.kr; younhwan@ktc.re.kr; moon@ktc.re.kr

Abstract - The shape optimization of the stator and the rotor is important for electrical motor design. Among many motor design
parameters, the stator tooth and yoke width are a few of the determinants of noload back-EMF and load torque. In this study, we proposed
an equivalent magnetic circuit of motor stator for efficient stator tooth and yoke width shape optimization. Using the proposed equivalent
magnetic circuit, we found the optimal tooth and yoke width for minimal magnetic resistance. To verify if load torque is truly maximized
for the optimal tooth and yoke width indicated by the proposed method, we performed finite element analysis (FEA) to calculate load
torque for different tooth and yoke widths. From the study, we confirmed reliability and usability of the proposed equivalent magnetic
circuit.

Keywords: Torque Density, Power Density, Electric Motor, Stator structure, Equivalent Circuit

1. Introduction
As the electronics of today become highly efficient and precise, there are more demands and attempts to improve the
performance via design optimization [1-2].
For electric motor design, deciding stator tooth and yoke width is cumbersome yet very important. When designed
with inappropriate tooth and yoke width, the motor may not operate at its operation region due to flux density saturation of
the yoke, and not reach the maximum torque for given tooth and yoke width due to lack of maximum Back-EMF. Thus, the
proper tooth and yoke width must be selected for electric motor design. One of the most widely used methods for electric
motor analysis is finite element analysis (FEA). However, FEA requires a number of component modelling, complex pre-
process steps, and long analysis time [3-5].
Thus in this study, instead of suffering slight accuracy compared to FEA, we propose an equivalent magnetic circuit
of stator, which allows fast and simple calculations for finding tooth and yoke width that optimize load torque [6-8].
Generally Back-EMF and load torque show the greatest values for the tooth and yoke width ratio that has the smallest
magnetic resistance to magnetomotive force of permanent magnet. Therefore, to find the tooth and yoke width with minimum
magnetic resistances for the number of slots per pole of concentrated winding electric motor and distributed winding electric
motor, we formulated the magnetic stator resistance for one pole. Finally, we confirmed the reliability of the equivalent
magnetic circuit method by comparing the optimal tooth and yoke width found by FEA and the equivalent magnetic circuit,
and by the proposed method.

2. Finding Optimal Tooth-Width/Yoke-Width Using Stator Equivalent Magnetic Circuit

For electric motor design, when desired output power and current are decided as electric motor specifications, current
density can be decided considering the cooling method, and once the current density is decided, wire diameter can be decided.
The calculation of armature winding turn for the required Back-EMF is decided for voltage limit and magnetic flux per pole,
and once armature winding turn is decided, the cross-sectional area slot can be determined considering the outer diameter of
the stator given as an electric motor specification.
When rotor diameter is designed considering Torque per unit Rotor Volume(TRV), slot area and stator outer diameter
are already determined and therefore, widths of the tooth and yoke cannot be increased by much. Therefore, it is necessary
to find tooth and yoke width ratio that maximizes Back-EMF within the size restrictions.

EEE 126-1
 To optimize the tooth and yoke width, we formulated the equivalent magnetic circuit (EMC) of the stator core. To find
the tooth and yoke width with the maximum Back-EMF and maximum torque, we found magnetic resistance of the stator.
Since the magnetic flux on one pole flows to the neighboring poles on both sides, to simplify the calculation, flux path on
both sides are not considered, but only the flux path though one side is considered for finding the magnetic resistance.
The assumptions for finding the magnetic resistance of stator are as follows:
1. Relative permeability of tooth and yoke is the same.
2. Saturation effects of tooth and yoke are not considered
3. Tooth tip is not considered.
4. The outer diameter and inner diameter of stator are determined by the initial shape.
Figure 1 shows the equivalent magnetic circuit formed by flux of each pole for distributed winding 6-pole 36-slot and
concentrated winding 6-pole 9-slot.

lyoke Dh lyoke

Ryoke Dh
Ryoke

Dy Rtooth
Rtooth
ltooth
ltooth Dy
x x
Ds Ds
(a) Distributed winding (b) Concentrated winding
Fig. 1: Equivalent magnetic circuit of stator.

The area for a slot can be formulated as the following:


As    Dy2  Ds2   x   Dy  Ds  (1)
4ns

x represents a half of tooth width, AS the cross-sectional area of a slot, ns the total number of slots, Dy the outer diameter
of a slot, and Ds the inner diameter of the stator.
For all analysis models, the slot areas have to be constant, and Dy can be found using the root of 2nd order polynomials.

2
2 xns  2 xns  4
Dy      Ds  ns  xDs  As 
2
(2)
    

For distributed winding, the equivalent magnetic circuit equation for magnetic resistance changes if different values
are used for the number of slots per pole. Magnetic resistance for 3 slot per pole:

 
1 1 1
Rtotal  2  2a 1  b   a
1
b a , b (3)
Rtooth Ryoke
,

Magnetic resistance for 6 slot per pole:

EEE 126-2
 1
 1

 a 
 1

1

 b  a
1
Rtotal  2 1
 b  a  b  b (4)
    
 
 

As the number of slots per pole increases, the magnetic resistance of equivalent magnetic circuit for stator shows in a
long continued fraction. The magnetic resistance of the equivalent magnetic circuit for concentrated winding electric motor
is as follows:

Rtotal  Rt  Ry (5)

The magnetic resistance change for tooth widths change is shown in Figure 2.
Since the slot areas are the same, the yoke width decreases with tooth width increase and the magnetic resistance of
stator changes accordingly. For tooth width increase, the magnetic resistance value first shows a steep drop, a relatively flat
curve, and after a certain region, an increase again. From the figure, we can find the tooth and yoke widths which minimize
the magnetic resistance.

5
Magnetic resistance

0 2 4 6 8 10 12
Tooth width [mm]
Fig. 2: Magnetic resistances (6pole 36slot) of stator for different tooth widths.

3. Finite Element Analysis For Stator Tooth/Yoke Widths

To validate the proposed method, FEA was conducted for 4 types of models: SPM distributed winding model 6pole
36slot and 8pole 48slot, IPM concentrated winding model 6pole 9slot and 10pole 15slot. For each model, we found the
model that gives the most Back-EMF from FEA result.
Table 1 summarizes the number of armature series turns per phase and pole width and the materials of electrical steel.
Since all of four models have the same stator, rotor outer diameter, number of turn per slot, and slot cross-sectional area, the
model with a wider tooth width has a relatively narrower yoke width. No-load Back-EMF is computed based on 3000rpm at
room temperature, when electric motor is operating, torque is analyzed for every 20Arms of input current ranging from
20Arms to 80Arms in case of distributed winding model and is similarly analyzed for every 50Arms of input current ranging
from 50Arms to 200Arms in case of concentrated winding model. In addition, Id=0 control method, a current vector current
method, is applied. Figure 3 shows the shape and dimension of the review model each with the number of poles and slots.

EEE 126-3
 (a) 6pole 36slot (b) 8pole 48slot

(c) 6pole 9slot (d) 10pole 15slot

Fig. 3: Review model shape.

Table 1: Specification of models.

Distributed Concentrated
Winding(SPM) Winding(IPM)
Item
6pole 8pole 6pole 10pole
36slot 48slot 9slot 15slot
Stator/rotor
outer 356.4/153.6 85/45
diameter [mm]
Stack length
50
[mm]
Series turns per phase 60 120 10 10
Residual flux density
1.4
[T]
Magnetic core
50PN470
material
Slot area [mm2] 679.0 614.6 163.5 89.6

The no-load Back-EMF and load torque of distributed winding and concentrated winding review model, each with
different number of poles and slots, is analyzed using FEM for different tooth and yoke widths. Figure 4 shows the FEA
result of no-load Phase back-EMF and load torque changes in the 6pole 36slot model for tooth and yoke width changes.
As shown in Table 2, the maximum Back-EMF and maximum torque were observed for tooth width with 9.7[mm] for
distributed winding 6pole 36slot model, tooth width with 7.2[mm] for 8pole 48slot model, 9.0[mm] concentrated winding
6pole 9slot mode, 7.0[mm] for 10pole 15slot model. This means that tooth/yoke width at maximum no-load Back-EMF is
equal to tooth/yoke width at maximum load torque.
Table 3 compares the tooth width obtained using FEA on each of review model, distributed winding model and
concentrated winding model, and the proposed equivalent magnetic circuit of stator. It is shown that the tooth widths obtained
using FEA and the proposed method are almost the same.

EEE 126-4
 20Arms 40Arms 60Arms 80Arms
180 80

BackEMF@3000rpm [V]

Torque @ Current [ Nm ]
70
170
60
160
50

150 40
30
140
20
130
10
120 0
6 7 8 9 10 6 7 8 9 10
Tooth width [mm] Tooth width [mm]
(a) Phase back EMF (b) Load torque
Fig. 4: Phase back EMF and load torque(6pole 36slot) for stator tooth widths.

Table 2: Tooth width to maximize back-EMF and load torque (FEA).

Tooth width
Tooth width to
to
maximize
Pole/slot maximize
load torque
back-EMF
[mm]
[mm]
6pole
9.7 9.7
36slot
Distributed winding
8pole
7.2 7.2
48slot
6pole
9.0 9.0
Concentrated 9slot
winding 10pole
6.8 6.8
15slot

Table 3: Comparison of FEA and EMC.

Optimized
Optimized
tooth width
Pole/slot tooth width
_EMC
_FEA [mm]
[mm]
6pole
9.7 9.7
Distributed 36slot
winding 8pole
7.2 7.2
48slot
6pole
Concentrate 9.0 8.8
9slot
d
10pole
winding 6.8 6.6
15slot

4. Tooth Width/Yoke Width Determination Considering Armature Reaction

Using the equivalent magnetic circuit of stator, tooth width and yoke width were determined. But the method used
previously assumes no-load, and therefore, in case of the model with large armature reaction, tooth and yoke width that
maximize the load torque may change. To check if the method applies to this case, we compared tooth and yoke widths that

EEE 126-5
maximize load torque for traction electric motor which has a large armature reaction and that minimize magnetic resistance
of stator considering armature reaction. Also, we compared results of FEA and the proposed EMC analysis to check the
validity of the proposed equivalent magnetic circuit of stator. Figure 5 and Table 4 show the shape, dimension and
specification of stator and rotor of continuous rated power 5kW traction electric motor.

Fig. 5: Review model shape.

Table 4: Specification of model.

16pole 24slot
Item (Concentrated winding,
IPM)
Stator/rotor outer
250/155.4
diameter [mm]
Stack length [mm] 40
Series turns per phase 42
Residual flux
1.4
density [T]
Magnetic core
35PN230
material
Slot area [mm2] 391

The stator was assumed to have outer diameter of 250mm, stack length of 40mm, residual flux density 1.4T for room
temperature. For the proposed traction electric motor model, under the same slot area constraint, FEA was conducted to
analyze load torque change for tooth/yoke width change, and tooth/yoke width for maximum load torque was found. Figure
6 demonstrates this result. The maximum load torque was shown for tooth width with 14.8mm.
In general, if either tooth or yoke has excessively large tooth magnetic resistance, the total magnetic resistance of
stator increases and resulting load torque decreases. In that case, the flux density of either of tooth and yoke increases. We
then studied how flux density and load torque are related. Figure 7 shows the change of flux density waveform on tooth and
yoke as rotor rotates. For each tooth width, the flux density waveform is analyzed. As indicated in the figure, with the increase
in tooth width, yoke width decreases, and consequently, flux density on the tooth reduces and that on the yoke increases.

EEE 126-6
 54.2

54.1

Torque [Nm]
54.0

53.9

53.8
13.5 14.0 14.5 15.0 15.5
Tooth width [mm]
Fig. 6: Variation of load torque with tooth width.

2.0 Tooth 2.0 Tooth Tooth

2.0 2.0 Tooth
Yoke Yoke
Flux density [Wb/m ]

1.5 1.5 Yoke Yoke

2
Flux density [Wb/m ]

1.5 1.5
2

Flux density [Wb/m ]

Flux density [Wb/m ]

2

2
1.0 1.0
1.0 1.0
0.5 0.5
0.5 0.5
0.0 0.0
0.0 0.0
-0.5 -0.5 -0.5 -0.5
-1.0 -1.0 -1.0 -1.0
-1.5 -1.5 -1.5 -1.5
-2.0 -2.0 -2.0 -2.0
0 60 120 180 240 300 360 0 60 120 180 240 300 360 0 60 120 180 240 300 360 0 60 120 180 240 300 360
Moving angle [deg.] Moving angle [deg.] Moving angle [deg.] Moving angle [deg.]

(a) Tooth width 13.8mm (b) Tooth witdh 14.4mm (c) Tooth width 14.8mm (d) Tooth width 15.2mm
Fig. 7: Flux density waveform on tooth and yoke.

2.5
Tooth 5.0
Yoke
Flux density [Wb/m ]
2

Magnetic resistance

4.5
2.0

4.0

1.5
3.5

1.0 3.0
13.5 14.0 14.5 15.0 15.5 13.5 14.0 14.5 15.0 15.5
Tooth width [mm] Tooth width [mm]
Fig. 8: Maximum value of flux density fundamental wave on tooth and yoke. Fig. 9: Variation of magnetic resistance of stator with
tooth width.

In Figure 8, for tooth and yoke width changes, the trend of change in tooth fundamental waveform maximum value of
flux density using FFT(Fast Fourier Transform) on tooth and yoke flux density waveform observed in Figure 7. As in Figure
7, tooth width increases with decrease in yoke width, and this causes the maximum value of fundamental waveform of flux
density on tooth to decrease and that of yoke to increase. The tooth width with maximum load torque using FEA was 14.8mm,
and this is close to the value that flux densities of tooth and yoke converge to each other. This shows that maximum load
torque was observed for the tooth and yoke widths that have similar flux density values. Using the proposed tooth and yoke
width determination method, we found the tooth width that minimizes the magnetic resistance of stator for traction electric
motor.

EEE 126-7
 Figure 9 shows the stator magnetic resistance change for the tooth/yoke width change in traction electric motor.
Magnetic resistance of stator in Figure 9 was computed using the proposed equivalent magnetic circuit. As in the figure,
tooth width with minimum magnetic resistance of stator was 14.4mm, and this was slightly different from the width found
by FEM: 14.8mm. This shows that the application of proposed stator equivalent magnetic circuit may be limited in case of
electric motors with significant amount of armature reaction. To resolve the problem, equivalent magnetic circuit design
needs to consider the electric motor seen from the stator side.
Figure 10 describes the equivalent magnetic circuit of stator seen from the stator side. Assuming air-gap and rotor
region as magnetic equipotential region, we form the corresponding area as a short circuit. For simplification of calculation,
we replace the Y-shaped wiring in the center of the circuit to the △-shaped wiring. Then we calculate magnetic resistance of
stator seen from a single magnetomotive force. In Figure 10, equivalent magnetic circuit of stator is equivalent magnetic
circuit corresponding to a period of electric motor flux density distribution and Ry3 is magnetic resistance of yoke
corresponding to the boundary between the periods.
Figure 11 shows stator load torque change using FEA, and magnetic resistance change using equivalent magnetic
circuit of stator seen from the stator for the tooth width changes. As in the previous, for the same slot area, the tooth width
changes with yoke width. As implied in Figure 11, tooth width that gives the minimal magnetic resistance of stator is 14.8mm,
the same value obtained using FEA. Thus, when seen from the stator side, tooth width with minimal magnetic resistance of
stator coincides with the result confirmed by FEA.

S ta to r R t1

R o to r R t1 F=N i R t1
F=N i R y1 R y1 R t2 y 1
an d
a ir- g a p R y3 R y3 R y1y1 R y3
(L in e a r R t2 R t2
M o d e l) R y2 R y2 R t2 y 2

R t3 R t3 R t3
F=N i

Fig. 10: Equivalent magnetic circuit of stator.

0.80 54.2
Torque(FEA)
Magnetic resistance(EMC)
0.75
Magnetic resistance

54.1
Torque [Nm]

0.70

0.65 54.0

0.60
53.9
0.55

0.50 53.8
13.5 14.0 14.5 15.0 15.5
Tooth width [mm]
Fig. 11: Variation of equivalent magnetic resistance with tooth width.

5. Conclusion
Compared to FEA, it is much simpler to use the equivalent magnetic circuit to find the optimal tooth and yoke width:
the width minimized the magnetic resistance of stator. In this study, we proposed equivalent magnetic circuit of stator to
calculate the tooth and yoke width that maximize no-load Back-EMF and load torque. In addition, we compared the optimal
teeth widths obtained using FEM and the proposed method. The similarities in the value confirmed the reliability and
feasibility of the proposed stator equivalent magnetic circuit. Furthermore, we proposed a compensatory method for
equivalent magnetic circuit of stator to account for armature reaction and confirmed the validity of the method by comparison
with the FEA result.

EEE 126-8
Acknowledgements
This work was supported by the Korea Evaluation Institute of Industrial Technology (Keit) grant funded by the Korea
government (Ministry of Trade, Industry and Energy) (No. 10060295)

References
[1] L. Liu, W. Liu, and D. A. Cartes, “Permanent Magnet Synchronous Motor Parameter Identification using particle Swarm
Optimization,” International Journal of Computational Intelligence Research, vol. 4, no. 2, pp. 211-218, 2008.
[2] S.-B. Lim, H. Kwon, S.-Y. Kwon, S. Choi, S.-H. Baek, and J. Lee, “Permanent Magnet overhang Effect in Permanent
Magnetic Actuator Using 3 Dimension Equivalent Magnetic Circuit network Method,” KIEE International Transactions
on Electrical Machinery and Energy Conversion System, vol. 5-B, no. 2, pp. 123~128, 2005.
[3] J.-Y. Lee, S.-H. Lee, G.-H. Lee, J.-P. Hong, and J. Hur, “Determination of Parameters Considering Magnetic
Nonlinearity in and Interior Permanent Magnet Synchronous motor,” IEEE Trans. Magn., vol. 42, no. 4, pp. 1303-1306,
2006.
[4] Z. Q. Zhu, Y. Pang, D. Howe, S. Iwasaki, R. Deodhar, and A.Pride, ”Analysis of Electromagnetic Performance of Flux-
Switching Permanent-Magnet Machines by Nonlinear Adaptive Lumped Parameter Magnetic Circuit Model,” IEEE
Trans. Magn., vol. 41, no. 11, pp. 4277-4287, 2005.
[5] A. H. Isfahani and S. Sadeghi, “Designof a Permanent Magnet Synchronous Machine for the Hybrid Electric Vehicle,”
International Journal of Electrical, Computer, and Systems Engineering, vol. 2, no. 1, 2008.
[6] L. Sang-Ho, K. Soon-O, L. Jeong-Jong, and H. Jung-Pyo, “Characteristic Analysis of Claw-Pole Machine Using
Improved Equivalent Magnetic Circuit,” IEEE Trans. Magn., vol. 45, no. 10, pp. 4570-4573, 2009.
[7] B. Sheikh-Ghalavand, S. Vaez-Zadeh, and A. H. Isfahani, “An improved Magnetic Equivalent Circuit Model for Iron-
Core Linear Permanent-Magnet Synchronous Motors,” IEEE Trans. Magn., vol. 46, no. 1, pp. 112-120, 2010.
[8] J. K. Kim, S. W. Joo, S. C. Hahn, and J. P. Hong, “Static characteristics of linear BLDC motor using equivalent magnetic
circuit and finite element method,” IEEE Trans. Magn., vol. 40, no. 2, pp. 742-745, 2004.

EEE 126-9