IEEE TRANSACTIONS ON MAGNETICS, VOL. 49, NO.

7, JULY 2013 3953

Robust Design Optimization of PM-SMC Motors for Six Sigma Quality

Manufacturing
Gang Lei , J. G. Zhu , Y. G. Guo , J. F. Hu , Wei Xu , and K. R. Shao
School of Electrical, Mechanical and Mechatronic Systems, University of Technology, Sydney, NSW 2007, Australia
College of Electrical and Electronic Engineering, Huazhong University of Science and Technology, Wuhan, 430074, China
Platform Technologies Research Institute,RMIT University, Melbourne, VIC 3001, Australia

In our previous work, soft magnetic composite (SMC) material was employed to design cores for two kinds of permanent magnet (PM)
motors, namely transverse flux machine (TFM) and claw pole motor. Compared with motors designed by traditional silicon steel sheets,
these motors require 3D flux design with new material and new manufacturing method. Meanwhile, the performances of these motors
highly depend on the material and manufacturing parameters besides structure parameters. Therefore, we present a robust design op-
timization method for high quality manufacturing of these PM-SMC motors to improve their industrial applications. Thereafter, from
the design analysis of a PM-SMC TFM, it can be found that the proposed method can significantly improve the manufacturing quality
and reliability of the motor, and reduce the manufacturing cost.
Index Terms—PM transverse flux machine, robust optimization, six sigma quality manufacturing, soft magnetic composite.

I. INTRODUCTION manufacturing quality and cost are needed to investigate for the
industrial applications of these motors. Firstly, there are many

R ECENTLY, interest in soft magnetic composite (SMC)
and its application in permanent magnet (PM) electrical
machines have increased significantly [1]–[3]. SMC is a new
type of soft magnetic material made of iron powder particles
which are separated with electrically insulated layers. SMC has
been employed to design cores for several kinds of motors, such
as PM transverse flux machine (TFM) and claw pole motor [4].
Compared with traditional silicon steel sheets, SMC cores
have the following advantages. Firstly, SMC cores are isotropic
both mechanically and magnetically, so they are suitable for the
design of 3-D flux path. Secondly, SMC cores are manufactured
by modules, so they are suitable for the motors with complex
structures. Thirdly, SMC cores have low eddy current loss as
SMC powders are separated with electrically insulated layers.
Fourthly, SMC is a cheap material. By using SMC, the material
cost will be reduced significantly. Finally, SMC is easy to re-
cycle. Therefore, SMC is a promising material for the design of
PM motors with complex structure and 3-D flux path, such as
PM TFM and claw pole motor [4], [5].
In our previous work, two kinds of PM-SMC motors, TFM
and claw pole motor have been designed, fabricated and tested
[4]–[7]. From the discussion, it was found that these motors can
make full use of the unique magnetic characteristics of SMC
and provide good performances. And their performances highly
depend on the material and manufacturing method, which are
different from traditional motors with silicon steel sheets. For
the design optimization of motors with traditional silicon steel
sheets, most works only investigate the structure parameters
[8], [9]. However, besides structure parameters, the material
and manufacturing parameters should also be investigated for
PM-SMC motors. Furthermore, two issues with respect to the
noise factors in the manufacturing process of these motors. Sec-
ondly, besides the material costs, manufacturing costs should
also be considered for the industrial applications of these mo-
tors. Therefore, robust design optimization method with inves-
tigation of all these issues is presented in this work to achieve
good performances and high quality manufacturing for these
PM-SMC motors.
II. MANUFACTURING OF PM-SMC MOTORS
Considering the manufacturing of PM-SMC motors, two is-
sues, namely manufacturing quality and cost, are needed to in-
vestigate as they are different from traditional motors with sil-
icon steel sheets. For the manufacturing, there are many noise
factors in the manufacturing process of PM-SMC motors. Be-
sides the structure parameters, the material parameters, namely
the type of SMC material and the density of SMC core are also
important issues for the motor’s performance.
Fig. 1 illustrates three magnetization curves for different den-
sity values of a kind of SMC core [1]. From this figure, it can
be found that there are significant differences of B-H data be-
tween different density values. Actually, the density of SMC
core depends on the manufacturing condition, namely the tons
of the stamping press. Therefore, all these parameters and is-
sues should be taken as design optimization factors for the in-
dustrial applications of these motors to improve their manufac-
turing quality.
Considering the manufacturing condition and material char-
acteristic of SMC cores, the design optimization model of
PM-SMC motors can be defined as

Manuscript received November 05, 2012; revised January 21, 2013; accepted (1)
January 21, 2013. Date of current version July 15, 2013. Corresponding author:
G. Lei (e-mail: gangleimc@gmail.com).
where , and are the structure, material and man-
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. ufacturing parameters respectively. and are the lower
Digital Object Identifier 10.1109/TMAG.2013.2243123 boundary and upper boundary of respectively, is the

0018-9464/$31.00 © 2013 IEEE

3954
Fig. 1. B-H curves with respect to different SMC density values.
Fig. 2. Sigma level and its equivalent probability for normal distribution.
number of constraints. For the six sigma quality manufacturing,
the design model can be obtained as (2) within the framework
of design for six sigma technology [9], [10].
(2)
where and are the mean and standard deviation of the corre-
sponding terms; LSL and USL are the lower and upper specifi-
cation limits; is the sigma level which is generally with respect
to a probability value of a standard normal distribution as shown
in Fig. 2. In our work, the designed SMC motors are expected
to achieve six sigma manufacturing quality, so will be defined
as 6. For industrial manufacturing and management, six sigma
level manufacturing quality means 0.002 defects per million for
the “short term sigma quality”, and 3.4 defects per million for
the “long term sigma quality” [9], [10].
III. DESCRIPTION OF A PM-SMC TFM
A PM-SMC TFM is investigated to illustrate the efficiency
of the proposed method. Fig. 3 shows the magnetically relevant
IEEE TRANSACTIONS ON MAGNETICS, VOL. 49, NO. 7, JULY 2013
Fig. 3. Magnetically relevant parts of PM-SMC TFM (a) rotor, (b) stator.
Fig. 4. Region for the 3D field analysis of PM-SMC TFM.
TABLE I
MAIN DESIGN MATERIAL AND PARAMETERS
parts of this machine. It is designed to deliver a power of 640
W at 1800 r/min. Fig. 4 shows the finite element analysis model
for this machine [5]. And we can see that 3D flux path design
is needed for this machine. Table I lists several parameters and
materials for this machine. From our design experience, eight
structure parameters are significant to the performance of this
machine. They are and : circumferential angle and axial
width of PM; to : circumferential width, axial width and
radial height of SMC tooth; and : number of turns and
diameter of copper wire winding; and : air gap.
Now we consider the material parameter and manufacturing
condition. As the SMC core is compressed by module, SMC
core’s density is calculated from the compacting pressure ap-
plied on the core’s surface and the pressure is related to ton’s
value of the used stamping press. Furthermore, the manufac-
turing cost of the SMC core directly depends on the selected
stamping press. Fig. 5 shows the manufacturing cost and pro-
ductivity of this SMC core by using different stamping presses.
Therefore, ton’s value is also selected as a design as well as a
noise factor for the robust manufacturing quality optimization.
The optimization model of this motor can be defined as,
LEI et al.: ROBUST DESIGN OPTIMIZATION OF PM-SMC MOTORS FOR SIX SIGMA QUALITY MANUFACTURING
Fig. 5. Manufacturing cost and productivity for SMC cores.
TABLE II
OPTIMIZATION RESULTS FOR TFM
(3)
where is a vector of design parameters which include eight
structure parameters and a manufacturing parameter; cost in ob-
jective mainly includes material costs of PM, SMC core, wire
winding, steel and manufacturing cost; and are cost and
output power of the initial design scheme [5]; and (unit:
W) in and are the motor’s efficiency and output power
respectively; and (unit: A/mm ) in and are the fill
factor and current density of the winding respectively.
Then with the robust optimization framework of (2), we can
get the robust optimization model of (3).
(4)
It should be noted that the optimization parameters in (3) and (4)
are discrete values, and their step sizes are shown in Table II. For
the calculation of core loss, an improved method proposed for
the core losses in 3D flux SMC motors [11] is used in this paper.
For the other performance parameters, such as and , their
calculation methods can be found in [5].
IV. DISCUSSION AND RESULTS
In the implementation, each parameter is defined to follow a
normal distribution with standard deviation as 1/3 of its man-
ufacturing tolerance. The tolerance values of the seventh and
3955
TABLE III
PERFORMANCE PARAMETERS FOR TFM
TABLE IV
SIGMA LEVELS FOR CONSTRAINTS AND POF FOR TFM
ninth motor parameters are defined as 1% of their mean values.
Other parameters’ tolerance values are the same as their step
sizes as shown in Table II. To illustrate the performance of dif-
ferent methods, probability of failure (POF) is taken as a crite-
rion, which is defined as
(5)
Tables II and III show the motor’s optimization results by two
methods, namely deterministic (Det. row) design optimization
of (3) and robust (Rob. row) design optimization of (4). Table IV
shows the robust levels for all constraints, and the POF values
for this motor. From these results and comparison, we can draw
the following conclusions.
1) For the deterministic design optimization, the obtained per-
formance parameters of TFM are listed as follows. Cost
is $27.4, output power is 731 W and efficiency is 0.82.
For the robust design optimization, cost is $28.4, output
power is 701.6 W and efficiency is 0.83. The robust design
scheme has higher cost and efficiency and lower output
power. Moreover, these values are better than those of the
initial design scheme, which are $34.1, 640 W and 0.795
respectively [5].
2) Considering the manufacturing condition, 200 ton
stamping press is suggested by deterministic method,
the corresponding manufacturing cost is $0.5 for each
SMC core, and its density is 7.26 g/cm . However, 100
ton stamping press is suggested by robust method, man-
ufacturing cost is $0.2, and core’s density is 6.60 g/cm .
Therefore, lower manufacturing condition and cost are
requested for robust design scheme.
3) After Monte Carlo analysis, we can get the sigma levels
for all constraints and POF values for both methods. For
the deterministic scheme, the reliability of constraint is
0.4697, and the corresponding sigma level is only 0.6275.
Actually, the current density is 6.0 A/mm , which is the
same as the limit of this constraint. As a result, the motor’s
POF is 53.03%. For the robust scheme, the sigma levels for
all constraints are larger than 6 and motor’s POF is almost
0. Therefore, robust method can produce high reliability
3956 IEEE TRANSACTIONS ON MAGNETICS, VOL. 49, NO. 7, JULY 2013

Fig. 6. Cost distributions of deterministic and robust optimization solutions.

Fig. 8. Current density distributions of deterministic and robust solutions.

lower manufacturing condition and cost. Though the cost of

robust design scheme is higher than that of deterministic de-
sign scheme, the motor’s efficiency, the standard deviations of
output power and current density are better than those of deter-
ministic design scheme. In summary, robust design optimiza-
tion presents products with higher reliability and manufacturing
quality, and will improve the industrial applications of these
PM-SMC motors.

ACKNOWLEDGMENT
This work was partially supported by the National Natural
Fig. 7. Output power distributions of deterministic and robust solutions.
Science Foundation of China (NSFC) under Grant 51207059.

products and this is very important for industrial manufac- REFERENCES

turing and applications.
4) Figs. 6 to 8 show the distributions of cost, output power
and current density respectively for both methods. From
Figs. 6 and 7, we can see that the standard deviations of cost
and output power of robust design scheme are smaller than
those of deterministic scheme. From Fig. 8, it can be found
that all distribution points of robust design scheme are
satisfied with the condition of “no larger than 6.0 A/mm ”;
while many points of deterministic design scheme are not
satisfied with the condition. Therefore, the reliability and
sigma level of this constraint of deterministic method is
very low, and the POF of motor is high. Actually, the lower
cost of deterministic design scheme is obtained at the cost
of low reliability and robustness.
V. CONCLUSION
Considering the differences of material and manufacturing
method between PM-SMC motors and motors with traditional
silicon steel sheets, structure, material and manufacturing pa-
rameters are needed to investigate for the design optimization
of PM-SMC motors. A robust design optimization method is
presented for these motors to achieve six sigma manufacturing
quality in this work. From the investigation of a PM-SMC TFM,
it can be found that the proposed method can significantly im-
prove the reliability and manufacturing quality of the motor with
[1] Höganös AB, SOMALOY Technology for Electric Motors [Online].
Available: www.hogonas.com
[2] G. Gvetkovski and L. Petkovska, “Performance improvement of PM
synchronous motor by using soft magnetic composite material,” IEEE
Trans. Magn., vol. 44, no. 11, pp. 3812–3815, 2008.
[3] D. H. Kim, J. K. Sykulski, and D. A. Lowther, “The implications of
the use of composite materials in electromagnetic device topology and
shape optimization,” IEEE Trans. Magn., vol. 45, no. 3, pp. 1154–1157,
2008.
[4] J. G. Zhu et al., “Development of PM transverse flux motors with soft
magnetic composite cores,” IEEE Trans. Magn., vol. 47, no. 10, pp.
4376–4383, 2011.
[5] Y. G. Guo, J. G. Zhu, P. A. Watterson, and W. Wu, “Development of
a PM transverse flux motor with soft magnetic composite core,” IEEE
Trans. Energy Conver., vol. 21, no. 2, pp. 426–434, 2006.
[6] Y. G. Guo, J. G. Zhu, and D. Dorrell, “Design and analysis of a claw
pole PM motor with molded SMC core,” IEEE Trans. Magn., vol. 45,
no. 10, pp. 4582–4585, 2009.
[7] Y. K. Huang et al., “Design and analysis of a high speed claw pole
motor with soft magnetic composite core,” IEEE Trans. Magn., vol.
43, no. 6, pp. 2492–2494, 2007.
[8] L. Jolly, M. A. Jabbar, and Q. H. Liu, “Design optimization of PM
motors using response surface methodology and genetic algorithms,”
IEEE Trans. Magn., vol. 41, no. 10, pp. 3928–3931, 2005, Member.
[9] X. J. Meng, S. H. Wang, J. Qiu, Q. H. Zhang, J. G. Zhu, Y. G. Guo, and
D. K. Liu, “Robust multilevel optimisation of PMSM using design for
six sigma,” IEEE Trans. Magn., vol. 47, no. 10, pp. 3248–3251, 2011.
[10] P. N. Koch, R. J. Yang, and L. Gu, “Design for six sigma through ro-
bust optimization,” Struct. Multidiscip. Optim., vol. 26, no. 3–4, pp.
235–248, 2004.
[11] Y. G. Guo et al., “An improved method for predicting magnetic power
losses in SMC electrical machines,” Int. J. Appl. Electromagn. Mech.,
vol. 19, pp. 75–78, 2004.