See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/322947351

Modern Electrical Machine Design Optimization: Techniques, Trends, and Best

Practices

Article  in  IEEE Transactions on Industrial Electronics · February 2018

DOI: 10.1109/TIE.2018.2801805

CITATIONS READS
13 6,490

4 authors, including:

Gerd Bramerdorfer Juan A. Tapia

Johannes Kepler University Linz University of Concepción
57 PUBLICATIONS   399 CITATIONS    68 PUBLICATIONS   880 CITATIONS   

SEE PROFILE SEE PROFILE

J. Pyrhönen
Lappeenranta – Lahti University of Technology LUT
437 PUBLICATIONS   5,292 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Magnetic properties of permanent magnets - IEEE Trans Magn View project

Modeling of the nonlinear characteristics of electric machines View project

All content following this page was uploaded by J. Pyrhönen on 02 April 2018.

The user has requested enhancement of the downloaded file.

 Modern Electrical Machine Design Optimization:
Techniques, Trends, and Best Practices
Gerd Bramerdorfer, Member, IEEE, Juha Pyrhönen, Senior Member, IEEE, Juan A. Tapia, Member, IEEE,
Andrea Cavagnino, Senior Member, IEEE
Abstract—Disruptive innovations in electrical machine design Millions of motors are embedded in consumer products, while
optimization are observed. Emerging trends were the motivation modern internal combustion engine car may have 60 individual
for this study. Improvements in Mathematics and Computer electric motors for all kinds of auxiliary systems. Electric
Science enable more detailed optimization scenarios which
cover evermore aspects of physics. In the past, machine design motors for industrial applications use more than 40% of the
was equivalent to investigating electromagnetic performance. total energy produced worldwide [2], [3].
Nowadays thermal, rotor dynamics, power electronics and From power generation to final user, electrical machinery
control aspects are included. Material and engineering science vary in terms of power from several megawatts to few mil-
have introduced new dimensions on the optimization process and liwatts; in size, from several meters diameter to just few
impact of manufacturing and unavoidable tolerances should be
considered. Consequently, multi-faceted scenarios are analyzed millimeters; according to the use particular restrictions on
and improvements in numerous fields take effect. This article efficiency, reliability, controllability and availability are set.
is a reference for both academics and practicing engineers Also requirements on the energy conversion process (constant
about recent developments and future trends. It comprises or variable speed, linear or rotational motion) increase the
the definition of optimization scenarios regarding geometry diversity of machine designs.
specification and goal setting. Moreover, a materials-based
perspective and techniques for solving optimization problems The squirrel cage induction motor and the wound rotor syn-
are included. Finally, a collection of examples from literature chronous machine are still the main industrial machine types
is presented and two particular scenarios are illustrated in detail. used. The induction motor is even called the workhorse of
Index Terms—electric machines, multidimensional systems, op- the industry. Lately, the authorities have launched regulations
timization, Pareto optimization, genetic algorithms, evolutionary that followed new challenges to the efficiency development
computation, particle swarm optimization, reliability, robustness, [4]–[6]. Efficiency classes, i.e. IE1 (Standard Efficiency); IE2
metamodeling (High Efficiency); IE3 (Premium Efficiency); IE4 (Super Pre-
I. I NTRODUCTION mium Efficiency), were introduced. Present-day IE3 efficiency
requirements for direct-on-line (DOL) induction motors are
E LECTRICAL machinery research and development is
facing the problem that the field of electromechanics is
considered mature and no big needs for future research are
already fairly tight, and there are IE4, IE5 yet to come.
Permanent magnet synchronous machines, synchronous reluc-
seen by the financiers. However, the field of applications is tance machines, switched reluctance machines, flux switching
so widespread that only local optima can be found and new (permanent magnet) machines, etc. can be shown optimal for
research and development efforts must be continuously taken. certain applications and purposes.
The following gives a brief overview and introduction why Selection becomes even more complicated considering that
optimization of electrical machines is continuously required all types may be realized as inner rotor, outer rotor, axial flux
and related research is a very up-to-date topic. versions or variants with translational movement and the con-
The total amount of electric energy generated globally pro- structions can feature single rotor and stator or multiple ones.
jected by year 2020 is 26000 TWh and it is about to increase Motors for DOL applications and speed-controlled power-
towards 30000 TWh [1]. Great proportion of this energy electronic applications should be different because of adverse
will be produced by electromechanical conversion process effects caused by the converter. Insulation is more stressed,
considering the many different types of mechanical power more losses are generated and bearing currents are more severe
sources available, like renewable (wind, marine, geothermal, in PWM-supplied systems. Electric vehicles demand their own
etc.), coal, natural gas, nuclear, with several topologies for gen- dedicated motors and especially the driving cycles of traction
erators. After the energy was transmitted to the place where it motors differ much from the industrial motor applications
is utilized and distributed according to the voltage requirement, cycles such that optimal designs may be far off the industrial
significant portion is converted back to mechanical using the machine ideal counterpart.
same principle. There are about 300,000,000 industrial electric What was written above already shows that optimizing an
motors installed and the amount is increasing by 15% annually. electrical machine is a question where no universal optimum
can be found. We must remember that utilizing machines for
Manuscript received July 11, 2017; revised October 19 and December 27, different applications, e.g., for high-speed, sets totally different
2017; accepted January 15, 2018. This work was partly conducted within LCM boundary conditions than a typical industrial drive at normal
GmbH as a part of the COMET K2 program of the Austrian government. The
projects are kindly supported by the Austrian and Upper Austrian government. speed ranges. While in former times optimal designs were
G. Bramerdorfer is with Johannes Kepler University Linz, Austria, e-mail: found by just investigating the electromagnetic part, nowadays
[email protected] ; J. Tapia is with the Unviersity of Concepcion, rotor dynamics, heat transfer and fluid-flow analyses need to
Chile, e-mail: [email protected] ; J. Pyrhönen is with the Lappeenranta Uni-
versity of Technology, Finland, e-mail: [email protected] ; A. Cavagnino be included following an optimal design that is highly likely
is with the Politecnico di Torino, Italy, e-mail: [email protected]. far off any “traditional optimum”.
 This article is organized as follows. Section II gives a brief O PTIMIZATION PROBLEM
Efficiency
introduction in techniques for evaluating machine designs for Cogging torque
Torque ripple
getting started. Consequently, Section III is about techniques Definition of Objectives Power factor
and Constraints Voltage
for optimizing machine designs. It comprises the definition of Cost
Mass
the optimization scenario by considering objectives and con- Construction volume
Machine topology
straints as well as the investigated search space defined by the • Parametrized geometry

applied geometry definition, materials selection and associated • Non-parametric geometry

Materials for
boundaries. Section IV is about recent examples from the field Definition of the search space • Conductors
• Permanent magnets
of electric machine optimization. As the optimization scenarios • Stator and rotor core material
Environmental conditions
get more and more comprehensive, the more developments and • Temperature

achievements in associated fields impact the machine design • Pressure

Design candidate modeling

optimization. For instance, both improvements in the power • Linear vs. nonlinear
• 2D vs. 3D
electronics, e.g., in the field of new switching devices, as Design candidate evaluation
well as achievements in Mathematics and Computer Science, Exploration of the search space / • FE-based
• Analytical
by allowing to efficiently consider robustness and reliability obtaining the solution space • Surrogate-modeling based
Search strategy
during optimization, take effect on the optimality of designs. • Deterministic
• Stochastic
Hence, an extensive but still not exhaustive collection of recent • Hybrid

examples graded according to major trends is presented. To Design selection

Evaluation and inter- • Automatized vs. interactive
round out this survey, two particular examples are illustrated pretation of the results Post-optimization verification
in more detail highlighting current research in the field. • Sensitivity / tolerance analysis

S OLUTION
II. T ECHNIQUES FOR EVALUATING MACHINE DESIGNS
Fig. 1: Typical structure of an optimization process comprising main
Selecting a dedicated technique for modeling and evaluation stages (colored boxes) and exemplary tasks (white boxes); dashed
is crucial for the investigation and comparison of different lines give loops for possibly required iteration(s).
machine design candidates.
In the early beginning, analytical equations were applied
III. T ECHNIQUES FOR OPTIMIZING MACHINE DESIGNS
to characterize the performance of machine configurations by
relating electric, magnetic, and mechanical variables. Many The range of problems in terms of selected objectives
modeling strategies based on analytical equations were con- and parameters and applied techniques for design candidate
tinuously adjusted and improved based on experimental results modeling, evaluation, and optimization is very widespread.
such that the therewith predicted machine performance often Citing all different scenarios thus would go beyond the scope
is of high accuracy. On the other hand, those models usually of this article. Instead, the authors try to give the reader
focus on a single machine topology. Thus, the developed an overview about integral components and most popular
techniques lack universality. In addition, emphasis is laid on methods for defining and solving such kind of problems in
deriving accurate performance criteria at the price of not the field of electric machine design.
providing the opportunity to study local effects. Nevertheless, In Fig. 1 a typical structure of an optimization process is
those techniques are still very useful, especially if there is illustrated. To take respect of the diversity of the optimization
need for a very rapid evaluation of a huge number of different problems considered in the field, an abstract overall setup is
machine designs, e.g., [7]. presented. The typical work flow is indicated by arrows on
The nonlinearity of the magnetic circuit makes even the bold black lines and the four relevant steps are introduced as
electromagnetic optimizing tasks complicated and it is often
• definition of objectives and constraints,
not possible to analytically calculate the torque or losses in
• definition of the search space,
motors with sufficient accuracy for optimization. In order
to allow for investigating machine designs in more detail, • exploration of the solution space and the
finite element analyses (FEA) are adopted. Due to limits in • evaluation and interpretation of the results.
computational power, most often two-dimensional scenarios Practicing engineers know that very often solving an opti-
are applied. Thanks to dramatically increased performance mization problem is an iterative process instead of a purely
of computers, nowadays electric machines are often analyzed sequential procedure. Due to obtained findings, the problem
in 3D comprising the simultaneous investigation of different might be reconfigured, e.g., by reformulating the objectives
physical aspects at once, e.g., the combination of electro- and constraints, redefining the search space for machine design
magnetic and thermal aspects as well as mechanical stress candidates, or modifying the applied technique(s) for the ex-
[8]. Moreover, the analysis often comprises the entire drive ploration of the solution space. Most often, the reconfiguration
to consider interactions of control, power electronics and the is done by engineers themselves. However, at least some
machine design to find the overall best configuration [9]. Using reconfiguration parts can also be automatized, as for instance
FEA makes it possible to obtain sufficient accuracy with some a dynamic change of the design parameter discretization.
restrictions. For example, assessing iron losses accurately in Possible iterative loops are illustrated using thin dashed lines
FEA still builds a challenge. with arrows in Fig. 1.
 In the following, crucial methods and techniques used for 1
solving optimization problems are itemized and explained.
This comprises basic mathematical definitions and equations. 0.9
Engineers often avoid studying pure mathematical explana-

efficiency
0.8
tions and definitions as they prefer gaining hands-on experi-
ence. Nevertheless, understanding the basic ideas of typically 0.7
applied techniques is important for feasibly configuring the
optimization process. To accomplish this balancing act, the 0.6 Pareto points /
fundamental mathematical principles are briefly described pro- non−dominated points
dominated points
viding references where they are adopted in the field of electric 0.5
0 0.1 0.2 0.3 0.4 0.5
machine design. This should enable the interested readers to relative cost
get a good overview about this particular field.
Fig. 2: Pareto front for an optimization problem with two objectives
relative cost and efficiency.
A. Definition of Objectives and Constraints
To efficiently run and solve an optimization problem and
definition is that standard optimization techniques developed
to obtain reliable results, it is crucial to correctly define
for single-objective scenarios can be applied. Moreover, design
the problem. This comprises the variables (continuous or
candidates are characterized by a single measure allowing for
discrete), the objectives to be maximized (or minimized) and
an easily comparison. On the other hand, the definition of
the conditions (constraints) under which the problem has
the weighting factors has a crucial impact on the results and
to be solved. The relationship of variables, objectives, and
objectives are very different, e.g., efficiency and cost. Hence,
constraints further allows for classifying problems.
defining these factors is a difficult task as each weighting
1) Single objective: For sake of completeness and as an results in different optimum and it requires wide experience
introduction, a general definition of a single-objective problem about the considered application field.
is presented. This comprises an objective function F (x) and Instead, most of the engineers today are choosing the right-
vectors for equality and inequality constraints, g(x) and h(x): hand definition presented above, i.e. a vector-based approach is
applied. This requires more complex optimization algorithms,
min F (x) subject to g(x) ≤ 0 and h(x) = 0 , which will be presented in section III-C, and normally the
x
runtime is increased. Typically some objectives are contradic-
where x is the design vector which specifies the design under tory in electric machine design. Thus, Pareto fronts of optimal
optimization. In electric machine design, typical objectives are solutions are obtained. A 2D-example is given in Fig. 2. As a
cost, efficiency, torque or power density, Back-EMF waveform, common optimization problem is about maximizing efficiency
and reliability regarding mechanical stress, thermal aspects, while minimizing cost, the upper and more left a point is
etc. Examples can for instance be found in [10], [11]. While in located, the better. The red crosses give the Pareto-optimal
the past a single operational point was investigated, nowadays points which often are called non-dominated points. The latter
analyses evermore comprise multiple load points, e.g., for designation implies that there is no solution that is better in
electric vehicles with particularly defined driving cycles or all objectives compared to a Pareto-optimal point. From this
other applications with operating cycles [12]–[20]. it follows that all the Pareto points are optimal and it is up to
All the listed quality indexes can also be considered as the engineer’s choice to select one of those designs by trading
constraints, e.g., motor efficiency must not be less than y % or cost vs. efficiency. By contrast, the blue crosses in Fig. 2 give
cost should not be more than z USD. Usually, some constraints dominated points, which are not Pareto-optimal.
are determined by the construction space, the maximum Even though some extra effort is required when using
operational temperature in specific parts of the motor (e.g., the vector-based multi-objective optimization approach, it
permanent magnets), the minimum value of the magnetic field gives beneficial ancillary information about the relationship
in the permanent magnets to avoid demagnetization, etc. of different objectives. The engineer can investigate these
It comes as no real surprise that optimizing electric coherences and can specify the weighting based on these find-
machine designs typically comprises the consideration of ings. The additional expense is especially justified if different
several performance measures. Thus, the multiple objectives weightings need to be analyzed. While the first definition of
approach must be adopted. the multi-objective problem necessitates a separate run per
each weighting factor set, using the second definition only
2) Multiple objectives: A multi-objective problem is de-
post-processing of the results needs to be run multiple times.
fined as
min qT F(x) or min F(x) ,
x x B. Definition of the search space
while the same definition of constraints as for the single- Besides selecting appropriate objectives, the definition of
objective case can be applied. the search space provides the basis for the results that can be
Using the definition on the left, introducing a vector of obtained. When optimizing electric machine designs, the def-
weighting factors q is required. This vector is of same size inition of the respective geometry is crucial. In the following,
as F which comprises all objectives. The advantage of this the two most prominent formulations are presented.
 π ple of synchronous reluctance machines typically works best
pz
with a particular number of barriers and respective shapes
(rectangular or round) depending on the number of stator
slots, the stator winding configuration and the number of rotor
dr
e
poles. By contrast, maximizing the mean torque and thus the
er output power often requires a different number of barriers
d ri hPM and barrier shape. If the engineer fixes those two barrier-
related parameters in advance, the obtained solution space
might not comprise the best possible solutions with regard to
both performance parameters torque ripple and mean torque.
r
h

Hence, parametrized geometries are applied if the geometry

Fig. 3: Parametrized rotor geometry. is fixed to particular shapes anyway or if expert knowledge
for the considered scenario is already available. Otherwise,
non-parametric geometries are utilized with increasing
1) Definition of the geometry: The geometry of machine frequency which will be explained in the following.
designs must be characterized with regard to the analyses
that are carried out during optimization, e.g., magnetostatic, b) Non-parametric geometry: If no expert knowledge
thermal, mechanical, etc., and the level of detail which about the ideal machine design structure for particular opti-
is required for the investigations. Basic definitions mization scenarios is available, non-parametric geometries can
usually encompass the dimensions and shape of stator be a good choice to find an optimal machine design. Moreover,
and rotor components (or equivalent components for such geometry representations are also used if new topologies
actuators) comprising the material distribution of (laminated) should be discovered. Even though those geometry definitions
ferromagnetic components, material for the conductors and are designated as non-parametric, they feature parameters for
for the permanent magnets. defining the structure of the assembly. However, just a few
classical parameters, e.g., dimensions, are used, while the
a) Parametrized geometry: An example for a other parameters have rather less physical meaning. In the
parametrized geometry is given in Fig. 3, where a rotor following, the most prominent methods used for optimizing
with embedded magnets is presented. The topology of the electric machine designs are presented which typically are
geometry is predefined, but changing the parameters, e.g., the summarized by the umbrella term topology optimization (TO).
rotor outer or inner diameter, takes effect on the shape of the One prominent example for an optimization comprising
rotor design. The parameters can comprise both continuous a non-parametric geometry definition is binary optimization
variables like dimensions as well as discrete parameters, such (BO). Figure 4 gives the basic idea of BO illustrated for the
as number of pole pairs or the number of slots for the rotor definition of the rotor of a synchronous reluctance machine.
of an induction machine.
Even though design parameters for such geometries can
be arbitrarily defined, the definition of the geometry has a
significant impact on its applicability and also on the run-
time of optimization problems. For instance, the rotor inner
and outer diameter can be independently defined by using
parameters dri and dre , respectively. It is obvious that dri < dre
must hold. However, a breach of this inequality constraint is
possible if both parameters are varied in a wide range requiring
overlapping parameter ranges. One possibility to overcome this (a) Rotor surface dis- (b) Particular aspect: Detached
problem is to use an alternative definition for dri , for example cretization material elements
dri = α×dre . Hence, dri is scaled by α with regard to the rotor
outer diameter. Thus, an alternative parameter (α) is varied
during optimization.
A parametrized geometry offers possibilities for finding
the best machine design configuration for an optimization
problem. However, by selecting a particular geometry like the
rotor in Fig. 3, the engineer restricts the potential solution
space to, e.g., rotors with buried magnets. Thus, even if a
rotor with surface-mounted magnets would feature superior
performance, using the geometry of Fig. 3 will not allow to (c) Sample FE-model (d) Sample flux density distribu-
derive machine designs of this rotor type. tion
While this conclusion seems obvious, very often more
sophisticated limitations with less priorly available knowledge Fig. 4: Rotor modeling using discretized surface and element-wise
have to be considered. For instance, minimizing torque rip- material assignment.
The rotor surface bounded by some rotor inner and outer functions, e.g., Gaussian [23], [24]. As by applying such
diameter is discretized into a particular number of elements. optimization scenarios very often rotor structures with thin
The surface can feature equal discretization in size or number ribs are derived, a computation of the mechanical stress in the
of elements in both radial and circumferential direction or rotor is favorable. Moreover, as the mass-production of electric
could also be defined by a non-uniform discretization. A machines follows a degradation of the material properties
certain number of material properties is defined and usually along the cutting edges, it is beneficial to also cover those
to any element one of those material properties can then be effects, as the authors of [25] did.
assigned to. Even though defining more than two distinct prop- Besides binary optimization, where the analyses of dif-
erties follows more than two possible states for an element, ferent rotor structures are independently computed, further
the term binary optimization is also frequently used in that non-parametric approaches for the optimization of geometries
case. For the synchronous reluctance machine presented in exist. Common examples are gradient-based techniques. At
Fig. 4, the two material properties laminated steel and air first, a typically nonlinear evaluation of a machine design is
were selected. A particular aspect that has to be considered performed. Using the obtained element-wise results, it requires
are detached particles of a material which is illustrated in Fig. low computational effort to derive the sensitivities of the
??. Techniques for deriving feasible designs both online or as outputs of the FE-analyses with regard to a change of the
post-processing action after the optimization were developed permeability of any element. Thus, one can derive which
in the past. While the first one is comparably more difficult to finite element featuring properties of laminated steel should
join with the optimization technique, the latter one can follow be changed to air or vice versa.
that the optimization algorithm focuses on non-feasible design Strictly speaking, the change of one element would require
regions while the optimal feasible designs cannot be obtained. another nonlinear computationally expensive analysis and,
Another crucial aspect to keep in mind is the huge number consequently, deriving the sensitivities again. However, very
of possible designs when dealing with useful discretization often a particular number of elements is changed at once to
levels. The overall number of potential design candidates nd save computation time. This is justified by the assumption
for number of materials and elements nmat and nel , respec- that changing a comparably low number of elements does
tively, can be computed using not take much effect on the overall sensitivities. The number
of elements per step is a crucial parameter for this iterative
nd = nnmat
el
. (1)
optimization technique. Typically, two different materials are
For the two-materials example, an exemplary computation for considered. While some researchers apply techniques that just
different discretization levels is provided by Table I. As can be allow for changing from laminated steel properties to air or
noticed, a discretization of 15 × 15 follows 5.4 × 1067 possible the other way round, other techniques allow for continuously
design combinations. Even though today’s optimization algo- changing in both directions.
rithms allow for significantly reducing the number of required A significant advantage of this technique is the low com-
design candidate evaluations compared to the overall number putational effort compared to binary optimization, as deriving
of possible designs, one has to be aware of these numbers the sensitivity information comes at low computational cost.
and consider them with regard to the run time, the number of The drawback, as with any gradient-based technique, is that it
evaluated designs, etc. In addition, any way of reducing the is prone to get stuck in local minima. Hence, the analysis of
number of degrees of freedom is beneficial, e.g., by defining multiple attempts with different or even adaptive numbers of
a symmetric rotor structure or by just discretizing the rotor elements to be changed per iteration step, different starting
surface close to the air gap very fine and using a coarse grid points, etc. is required in order to increase confidence for
for the remaining rotor surface. particular solutions. An example can be found in [26].
Examples for binary optimization applied for deriving op- Another technique is to define a finite element mesh and
timal electric machine designs can for instance be found in consider the respective nodes to be moved within the design
[21], [22]. Some authors consider the reduction of number domain. Hence, almost free shapes can be obtained following
of degrees of freedoms by keeping the definition of the rotor optimal contours of components. Recently, Li et al. presented
structure quite flexible. For instance, they use a fine discretiza- this technique applied to electromechanical devices, e.g., for
tion, but the material assignment of the elements is not totally switched reluctance motors [27], [28].
independent, but based on parametrized spatial distribution
2) Definition of the materials: A typical machine design
comprises soft magnetic components as core material, copper
# rows # columns # elements # designs or aluminum for the conductors and, depending on the ma-
5 5 25 2 = 3.4 × 107
25 chine type, possibly permanent magnets are employed. Nowa-
10 10 100 2100 = 1.3 × 1030 days, the optimization of electric machine designs usually
15 15 225 2225 = 5.4 × 1067 comprises the minimization of material cost. While in the past
20 20 400 2400 = 2.6 × 10120 particular material types were selected and the overall volume
was subject to minimize, nowadays often multiple winding
25 25 625 2625 = 1.4 × 10188
materials, permanent magnet materials and/or laminated steels
TABLE I: Number of possible machine design variants for BO as are simultaneously investigated and solving the optimization
function of the discretization level for two material properties. problem also requires selecting the best material [29].
 Besides optimizing machine designs itself, the characteris-
tics of materials are particularly optimized for electric machine 1
application. In [30], [31], optimization of permanent magnet
materials was reported, while [32] gives status and trends of

x3
0
soft magnetic materials in electric machines comprising soft
magnetic composites and cobalt iron. We also have novel −1 1
−1 0 x2
heat tolerant Samarium-Cobalt permanent magnet materials 0 1 −1
x1
(e.g., RECOMA [33]) that can operate at high temperatures
around 300◦ C and maintain high remanence. Besides better Fig. 5: Box-Behnken based selection of designs investigated in the
magnetic circuit materials, conductive materials much better sampling phase - blue circles give designs considered, while red
than copper seem to be within reach for the future development crosses denote parameter combinations that are ignored.
step of electric machinery [34]. For instance, carbon nanotube
materials at the moment feature the specific conductivity at the
of experiments. Originally, the development of these tech-
same level as aluminum. The big future promise in carbon-
niques was inspired from experimental practices. The max-
nanotube-based conductors is based on the fact that individual
imum number of experiments is always bounded to obtain
best nanotubes demonstrate conductivity levels higher than
results at reasonable effort and within acceptable time. Even
100 MS/m corresponding to resistivity levels of 1×10−8 Ωm.
though the simulation of a device is much faster and cheaper
These kind of conductors enable no practical space for eddy
than manufacturing and measuring a prototype, limitations
currents. As a consequence, a new machine design paradigm
regarding computational power and analysis time still apply.
is needed when they will be available [35]–[37].
Thus, DOE techniques are also utilized in electric machine
Cooling of the machine is always a central aspect that affects
design for simultaneously varying multiple design parameters
machine performance and material selection [38]–[40]. There
and retrieving the impact of each parameter on the objective
is no globally optimal cooling means for a machine. For in-
functions. If a suitable technique is applied, it allows for sig-
stance, the totally-enclosed fan-cooled squirrel-cage induction
nificantly reducing the number of required design evaluations
motor does not offer the highest possible efficiency, it is heavy
compared to a full grid analysis while providing comparable
and its power factor is low at least for multiple-pole machines.
informative content. Thus, the overall effort and runtime for
But the machine is cheap and satisfies many of the boundary
solving optimization problems can be minimized. Famous
conditions set by manufacturers and customers. Hence, present
DOE techniques are Latin-Hypercube-Sampling, Fractional or
solutions in practice typically constitute optimal designs with
Full Factorial tests, Central Composite Design or the Box-
regard to particularly set cooling options.
Behnken sampling [46], [47].
The definition of the materials involves a computationally-
Fig. 5 gives an example for the Box-Behnken approach.
efficient and highly-accurate formulation and evaluation of
A grid of three design parameters with three levels each is
their properties, including the impact of manufacturing [41]–
considered, thus analyzing all possible combinations would
[43] and the evaluation of associated losses [44].
require 33 = 27 evaluations. By contrast, with the Box-
Behnken approach less calculations are required, i.e. the
C. Techniques for efficiently exploring the search space / combinations defined by red crosses in Fig. 5. Any other result
obtaining the solution space is then calculated based on the obtained results and, e.g., a
consequently created surrogate model. Fig. 6 gives an example
Due to limitations in computational power, the number of
of Latin-Hypercube-Sampling. The design parameter ranges
evaluated candidates for a typical machine design process was
of x1 and x2 are discretized and every parameter section
rather limited in former times. Hence, usually a grid search of
is used once (or multiple times, but usually same frequency
a few most relevant parameters was applied for finding the best
is applied). The crosses in the figure denote the (randomly)
machine design. The amount of data was easily manageable
defined combinations of the sections.
and engineers normally did not care about topics like efficient
exploration of the design space.
Nowadays, this topic is highly relevant in machine design. 1.0
0.9
Due to (economic) competition, engineers try to find optimal 0.8
solutions for particular applications and try to consider 0.7
0.6
multiple design parameters, physical domains, and objectives
x2

0.5
and constraints. Thus, optimization problems are very 0.4
0.3
complex and techniques for exploring the parameter space 0.2
efficiently are developed by considering the minimization of 0.1

time-consuming evaluations. In the following, some major 0.0

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

aspects are described. x1

Fig. 6: Latin-Hypercube-Sampling for two parameters: The nor-
1) Design of Experiments: Design of experiments (DOE, malized range of all parameters is discretized to a finite number
e.g., in [45]) is a field of research that is about retrieving of sections. Consequently, (random) section combinations for all
as much information as possible from a minimum number parameters are selected.
 Very often, design of experiments is used as initial stage techniques related to (a) the field of electric machine design,
during the optimization process in order to derive useful (b) Mathematics and / or (c) the field of Computer Science
information about the problem and to obtain a practicable are presented in the following:
initial situation for applying an optimization technique.
a) Symmetries and Discretization: Very often, symme-
2) Deterministic optimization: This concept refers to tries in the machine design are not applied when using finite
techniques with no random component, i.e., running the same element analyses. As this might carry no weight for a single
optimization problem with exactly the same settings multiple run, usually solving an optimization problem comprises the
times always leads to the same results. Prominent examples evaluation of thousands of different designs. Those problems
are the aforementioned Grid search and Derivative-based take hours to weeks to be solved with FEA to be the most
optimization like Newton’s method. If properly configured, considerable part and bottleneck of the analysis. Hence, any
such methods can converge very fast to local extrema. A information about symmetries of the assembly or in the
crucial drawback of these techniques is that they are usually machine characteristics (i.e. periodic behavior) with regard to
not feasible for problems with multimodal characteristics, as the rotor angle / translational degree of freedom should be
they get stuck in local extrema. Today’s problems typically utilized. This was also pointed out in [57].
have very complex functional relationships and they are Moreover, engineers should always consider the tradeoff
unknown beforehand. accuracy vs. runtime of the optimization problem when spec-
ifying a project. E.g., if only the mean value of the load
3) Stochastic optimization: This section encompasses tech- torque, or particular low-order harmonics are relevant, or if
niques with some random aspects. This should allow to evolve it is sufficient to have a coarse approximation of iron losses,
and release the optimization technique to not hang on to local there is no need for a fine discretization of the rotor movement.
extrema. Another aspect is to scan the promising regions in Even though those measures can be designated as well-
the design space finer than areas with less good prospects. known facts, they are very often disregarded. Consequently,
Many different techniques were developed and applied in the this increases the runtime many times over.
past. Thereby, evolutionary population-based algorithms that
are typically inspired by mechanisms in nature gained most b) Surrogate models: If the analysis of design candidates
attraction. Examples are Genetic Algorithms (GAs) like the takes considerable time, interpolation techniques are often
Non-dominated Sorting Genetic Algorithm II (NSGA-II) [48] considered. Nowadays surrogate models are applied more and
and the Strength Pareto Evolutionary Algorithm (SPEA2) [49]. more frequently. They represent a (simplified) model of the
Further important procedures are Particle Swarm Optimization original problem. Based on initial results, e.g., obtained by
(PSO) [50], [51] and Differential Evolution (DE) [52]. DOE, a model is found that maps the input-output functional
Also hybrid algorithms comprising multiple stochastic relationship of (parts of) the optimization problem. Reconsid-
optimization techniques, multiple deterministic techniques, ering step 3 of Fig. 1, surrogate modeling can be utilized to
or combinations of both were introduced in the past. Thus, replace parts or the entire evaluation process of that step. For
the attempt is to gain an additional performance increase by instance, it can be applied to directly map the impact of the
a perfect combination and matching of different techniques. design parameters on the objectives and constraints. Relevant
However, in machine learning and optimization it can be examples of this approach applied to electric machine design
concluded from the well known no-free-lunch-theorem that can for instance be found in [58]–[63].
for certain types of problems the averaged computational cost On the other hand, surrogate models are frequently applied
for finding a solution over all problems of that type is the to replace parts of the individual analysis of machine designs.
same for any optimization technique. Thus, if particularly For instance, the nonlinear machine characteristics can be
adapted techniques perform best on certain problems, they defined using a surrogate model. This approach typically is
probably will do worse on other problems in the field. considered if the design candidates are investigated using FE-
Hence, selecting a suitable algorithm with regard to the simulations. Based on initial results obtained using FEA, a
requirements based on the optimization problem is one of the model is developed that reproduces the attributes of the design.
most crucial aspects in solving optimization tasks. Examples Any further evaluation can be conducted by using that model.
of evolutionary algorithms applied to electric machine design Hence, the number of required finite element computations
can, e.g., be found in [53]–[56]. can be significantly minimized. Interesting implementations
are, e.g., presented in [64]–[67]. Fig. 7 gives an example for
4) Reduction of computational effort and runtime: interpolating the load torque characteristics as function of the
In the previous subsections of III-C the focus was on d- and q-axis current based on just 12 sample points.
efficiently scanning the design space by minimizing the From a mathematical point of view, prominent techniques
number of required evaluations of design candidates. By applied for surrogate modeling or interpolation are the re-
contrast, this section is about how to minimize the evaluation sponse surface methodology (RSM) [68], artificial neural
time for any single design candidate. As usually the networks (ANNs) [69], Kriging [70], [71], support vector
candidates are evaluated by means of finite element (FE-) machines [72] and radial basis functions (RBFs) [64], [73],
simulations, obtaining the performance of a considered design [74]. A comprehensive fitness approximation survey can be
requires significant computational effort. Hence, most relevant found in [75].
 y
optimal
robust

Fig. 8: 1D-explanation of optimality vs. robustness. The red square

denotes the overall optimum if the target is to minimize y. However,
a more robust solution is given by the green square. This is indicated
by a normal distribution around the two respective x-values and their
corresponding target distributions along the y-axis.
Fig. 7: Surrogate modeling of the torque as function of the d-/q-axis
current. Black squares give simulation results while colored plane
represents results of a radial basis function based surrogate model. were carried out with regard to the robustness of designs
in context of optimization. Typical keywords are sensitivity,
c) Computer cluster and parallel computing: In order robustness, reliability, tolerance analysis, etc. Hence, during
to speed up the optimization run, parallel computing is a optimization the tradeoff rated performance vs. sensitiveness
viable technique. Parallel computing can be used for solving must be considered. By way of explanation, a 1D-example is
particular tasks, as for instance a finite element simulation, by presented in Fig. 8. Some recent examples from literature can
distributing a single problem to multiple cores / computers. be found in [49], [90]–[95].
More often, the analysis of multiple designs or of multiple cur- In the following, two representative optimization problems
rent vectors of a design is analyzed using parallel computing. that recently have been published are briefly introduced. While
This follows that independent tasks are run on the processors in subsection IV-A the focus is on multi-physics optimization of
while the host computer defines the designs to be analyzed and an induction machine with severe boundaries, e.g., regarding
the ambient conditions, in subsection IV-B scenarios are
the tasks to be solved. Moreover, the optimization algorithm is
run at the host. A dedicated software is required to distribute presented where the cost versus efficiency tradeoff is analyzed.
the tasks to a cluster. A prominent example is HTCondor [76]. For the latter, DOE techniques, evolutionary algorithms, sur-
rogate modeling, and a computer cluster were applied. More-
over, symmetries of the machine designs were automatically
IV. R ECENT EXAMPLES FROM LITERATURE
detected and utilized and the discretization in the finite element
Various optimization scenarios were investigated in the past. software was automatically adapted. While these two examples
Thus, the following list is not intended to be exhaustive. In- cover all relevant topics discussed throughout this article, the
stead, it should serve as starting point for interested researchers comprehensive collection of references allows for delving into
and practitioners to delve into electric machine optimization the wide field of electric machine optimization.
and its multifaceted aspects.
Some typical optimization scenarios where evolutionary
A. Solid-rotor induction machine for natural gas compression
algorithms are applied can be found in [77]–[79]. A collection
of speed improvements comprising an exemplary optimization Recently, an optimization process for a multiple-step natural
problem is presented in [57]. A trend is to investigate machine gas compressor motor with 8MW power at 12000 min−1 was
designs from a multiphysics point of view, e.g., by adding presented. The optimization process comprised
mechanical or thermal aspects or aspects related to the power • the selection of a rugged enough machine topology for a
electronics or control to the optimization of electric machine rough environment [96]
designs. Thus, more system-level based optimization scenarios • the design of an active magnetic bearing system
[9], [14], [80], [81] are considered which requires enhanced • the investigation of rotor (and stator) dynamics, e.g.,
computational frameworks. A major trend, e.g., for traction (AMB) eigenfrequencies
applications, is further to combine magnetic gears with electric • the investigation of cooling of the machine by the com-
machine designs, for instance investigated in [82]–[84]. pressed gas itself and
Due to improvements in power electronics, SiC and GaN • the analysis and selection of suitable materials for high
MosFETs came into the picture. They have the capability pressure gas and high temperature humidity [97], [98].
to replace low-voltage IGBTs in some fields [85]–[89]. New As a consequence, such a scenario requires a multi-physics
machine design optimization is required to profit at best. approach where electrical engineers have to understand the
When dealing with electric machine design optimization, the boundary conditions set by the machine driven. It is important
obtained arrangements are typically highly-utilized. This is a that the “optimized” electrical machine itself represents a
consequence of objectives like power or torque density, cost, suitable compromise satisfying the conditions of the scenario.
etc. As tolerances in dimensions or material characteristics are The flow chart of the optimization process is presented in
unavoidable, the performance of the design should be insen- Fig. 9, while the first prototype integrated with the compressor
sitive to such variations. Lately, more and more investigations is shown in Fig. 10. The assembly should withstand very high
 ✁✂✄ ☎✆ ✝✞✟✆☎✟✠✡☛☞✞ ✟✞✌✍✁✟✞✠✞☛✄✂
✎✡✄✞✟✁✡✏ ✡☛✡✏✑✂✁✂
✗☎✄☎✟✒✑☛✡✠✁☞✂ ✡☛✡✏✑✂✁✂✙ ✛✞✄✡✁✏✞✒ ✏☎✂✂✞✂ ✡☛✒ ☞☎☎✏✁☛✔ ✡☛✡✏✑✂✁✂
✎☎✄☎✟ ✄✑✝✞ ✂✞✏✞☞✄✁☎☛
✢✂✄✁✠✡✄✞✒ ✄✘✞ ✂✄✡✄☎✟ ✄☎✄✡✏ ✏✞☛✔✄✘
✎☎✄☎✟ ✠✡✁☛ ✒✁✠✞☛✂✁☎☛✂
✚✘✡☛✔✞ ✄✘✞ ✟☎✄☎✟ ✞☛✔✄✘ ✟✞✌✍✁✟✞✒ ✆☎✟ ✖✞✡✟✁☛✔✂
✓✡☛✔✞☛✄✁✡✏ ✂✄✟✞✂✂ ✂✞✏✞☞✄✁☎☛
✒✁✠✞☛✂✁☎☛✣ ✞✂✝✞☞✁✡✏✏✑ ✄✘✞ ✧☛✡✏✑✫✞ ✄✘✞ ✦☎✟✪✁☛✔ ✄☎☎✏ ✁✠✝✡☞✄ ✼☎
✏✞☛✔✄✘ ✄☎ ✒✁✡✠✞✄✞✟ ✟✡✄✁☎ ✛☎ ✡✏✏ ✝✡✟✄✂ ✟✞✠✡✁☛ ✍☛✒✞✟
✏✁✠✁✄✁☛✔ ✄✞✠✝✞✟✡✄✍✟✞✤
✼☎
✼☎☛ ✆✞✡✂✁✖✏✞ ✽✞✂
✚✡☛ ✄✘✞ ✟✁✂✪ ☎✆ ✼☎
✕✞✡✂✁✖✁✏✁✄✑ ✂✄✍✒✑ ✂✍✝✞✟☞✟✁✄✁☞✡✏ ✟☎✄☎✟ ✖✞ ✥✂ ✄✘✞ ✟☎✄☎✟ ✂✍✖☞✟✁✄✁☞✡✏✤
✄✡✪✞☛✤
✛✞✄✡✁✏✞✒ ✟☎✄☎✟✒✑☛✡✠✁☞✂ ✡☛✡✏✑✂✁✂
✕✞✡✂✁✖✏✞ ✽✞✂
✽✞✂
✗☎✄☎✟ ✆✁✟✂✄ ✠✞☞✘✡☛✁☞✡✏ ✡☛✡✏✑✂✁✂✙ ★✞✡✟✁☛✔ ✂✞✏✞☞✄✁☎☛✙
✓✍☛✞ ✄✘✞ ✠✞☞✘✡☛✁☞✡✏ ✒✞✂✁✔☛✙
✚✞☛✄✟✁✆✍✔✡✏ ✆☎✟☞✞ ✗☎✏✏✁☛✔ ✖✞✡✟✁☛✔
✳ ✚☎☛✂✁✒✞✟ ✖✞✡✟✁☛✔ ✝☎✂✁✄✁☎☛
✯✴ ✞☛✔✄✘ ✄☎ ✒✁✡✠✞✄✞✟ ✟✡✄✁☎ ✂✞✏✞☞✄✁☎☛ ✕✏✍✁✒ ✆✁✏✠ ✖✞✡✟✁☛✔
✚☎☛✂✁✒✞✟ ✧✎★ ✠✞✡✂✍✟✞✠✞☛✄ ✏☎☞✡✄✁☎☛✂
✵ ✧☞✄✁✜✞ ✠✡✔☛✞✄✁☞ ✖✞✡✟✁☛✔
✶ ✩☎ ✖✡☞✪ ✄☎✙
✱ ✻ ✭✂✡✆✞✄✑ ✖✞✡✟✁☛✔
✲ ✖✞✡✟✁☛✔ ✒✞✂✁✔☛ ✡☛✒ ✒✞✄✡✁✏✞✒ ✆✟✡✠✞
✸✹ ✹✰✵
✳ ✳✴ ✛✞✂✁✔☛ ☎✆ ✄✘✞ ✠☎✄☎✟ ✡☞✄✁✜✞ ✝✡✟✄✂✙ ✒✞✂✁✔☛ ☎✟ ✂✄✡✟✄ ✆✟☎✠ ✄✘✞ ✖✞✔✁☛☛✁☛✔
✷
✵ ✰ ✢✏✞☞✄✟☎✠✡✔☛✞✄✁☞ ✒✞✂✁✔☛✣ ✛✕✎ ✥☛☞✟✞✡✂✞ ✝☎✦✞✟ ✒✞☛✂✁✄✑
✴✶ ✺ ✎✡☞✘✁☛✞ ✆✟✡✠✞ ✒✞✂✁✔☛
✵✶ ✵ ✚☎☎✏✁☛✔ ✠✞✄✘☎✒ ✂✞✏✞☞✄✁☎☛
✳ ✴✶ ✥✂ ✄✘✞ ✟☎✄☎✟✒✑☛✡✠✁☞✂
✴✯ ✵✶
✳ ✠✡☛✡✔✞✡✖✏✞✤ ✼☎
✲
✱ ✼☎
✰ ✼☎ ✽✞✂
✯ ✛☎ ✒✁✠✞☛✂✁☎☛ ✂✡✄✁✂✆✑ ✄✘✞
✮ ✛☎✞✂ ✄✘✞ ✟☎✄☎✟ ✄☎✏✞✟✡✄✞ ✄✘✞
☞✞☛✄✟✁✆✍✔✡✏ ✆☎✟☞✞✂✤ ☎✟✁✔✁☛✡✏ ✟✞✌✍✁✟✞✠✞☛✄✂✤ ✎✡☛✍✆✡☞✄✍✟✁☛✔✣ ✡✂✂✞✠✖✏✑✣ ☞☎✠✠✁✂✂✁☎☛✁☛✔
☎✂✂ ✠✁☛✁✠✁✫✡✄✁☎☛
✽✞✂ ★✞✄✄✞✟ ✠✡✄✞✟✁✡✏ ✬ ✒✞✂✁✔☛
✎☎✟✞ ✞✆✆✁☞✁✞☛✄ ☞☎☎✏✁☛✔

Fig. 9: Flow chart of the optimization process for the motor for a natural gas compressor.

IM (4poles)
200
PMSM 12 08 Cost1
175 PMSM 12 08 Cost2
PMSM 06 04 Cost1
costs [Euro] 150 PMSM 06 04 Cost2
125 IE2 IE3 IE4

100

75
Fig. 10: Prototype of the optimized machine design.
50

25

pressure of 130 bar. Hence, the frame was designed as a strong 0

85 87.5 90 92.5 95
pressure vessel where frame thickness was set to 100 mm efficiency [%]
at 750 mm stator stack outer diameter. This further ensured Fig. 11: Material cost vs. efficiency comparison of IMs and PMSMs.
that the AMB control is not adversely affected by the frame
eigenfrequencies, as they are at very high levels.
200
SurfM.Ferrite.Co
The investigation of the rotor dynamics was a crucial SurfM.Ferrite.Alu
175 SurfM.NdFeB.Co
part of the analysis. The final design features 6.6kV having SurfM.NdFeB.Alu
Material cost in Euro

Embedded.NdFeB.Co
prefabricated windings and open stator slots. Semimagnetic 150 Embedded.NdFeB.Alu
SyncRM.Ferrite.Co
wedges offer a solution to mitigate permeance harmonics’ 125
SyncRM.Ferrite.Alu
SyncRM.NoMagnet.Co
effects on the solid rotor surface and, thus, minimizing the 100
SyncRM.NoMagnet.Alu

rotor surface losses. As the magnetizing inductance is inversely

75
proportional to the square of the number of pole pairs p, p was
set to 1 to get an acceptable power factor. The final design has 50

an electrical efficiency of 96.5%, while the power factor and 25

ηIE−3
ηIE−2
the rated slip are 0.7 and 0.7%, respectively.
80 82.5 85 87.5 90 92.5 95
Efficiency in percent
B. Optimization and Comparison of Different Machine Fig. 12: Material cost vs. efficiency of PMSMs and SyncRMs for
Topologies different magnet materials and positions and conductor materials.
In recent time, optimization scenarios comprising a com-
parison of different machine topologies were presented. As
mentioned in the introduction, efficiency-driven machine de- power supply, materials, it soon becomes apparent that it is
sign optimization gets more and more important. On the other impossible to derive a universal solution. Instead, the optimal
hand, the price of the designs is crucial to stay competitive on machine designs obtained are as diverse as the optimization
the market and thus the best tradeoff is sought by researchers tasks considered.
and practitioners. Investigating different scenarios in terms of Fig. 11 gives an example for a comparison of induction ma-
requirements, i.e., boundaries and objectives, e.g., efficiency, chines (IMs) with permanent magnet synchronous machines
cost, construction space, ambient conditions, operating mode, (PMSMs). The two objectives considered are efficiency and
material cost and Pareto-optimal solutions are shown in the on the optimization of electromagnetic electrical machines is
figure. As recently the raw materials for permanent magnets fully demonstrated by the recently publication of the Special
were subject to significant price fluctuations, two different cost Section on Optimization of Electric Machine Designs in
scenarios were investigated regarding their respective costs. IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
By contrast, Fig. 12 gives a comparison of PMSMs and [100], [101]. This successful editorial project includes 46
synchronous reluctance machines (SyncRMs) with regard to papers related practically to all the types of electrical ma-
magnet position (surface mounted or buried), magnet type (fer- chines - namely, the synchronous reluctance machine, interior
rite or Neodymium-Iron-Boron) as well as material selected permanent magnet machine, permanent magnet synchronous
for the conductors (copper or aluminum). Again, objectives machine, induction machine, switched reluctance machine,
material cost and efficiency were considered, while some flux switching machine, magnetic-geared motor, magnetic
constraints, e.g., on maximum permissible torque ripple, were gear, and total enclosed fan cooled electrical machine. Also,
set. Both studies were conducted for rated power of 3kW and various optimization algorithms have been deeply discussed.
rated speed of 1500rpm [29], [99]. From this point of view, it is worth to cite the following
methods: genetic algorithm, differential evolution, response
V. C ONCLUSION AND O UTLOOK surface methodology, grid search, particle swarm optimization,
multi-dimensional secant method, sequential quadratic pro-
This article is about the state-of-the-art and emerging trends gramming, global search, tabu search, evolutionary algorithm,
in the field of electric machine design optimization. Thanks heuristic algorithm, Hooke-Jeeves algorithm, sequential-stage
to advances in Mathematics and Computer Science, nowadays optimization strategy, gradient-based strategy, social network
electric machine optimization scenarios can be investigated optimization.
at very high level of detail. Besides traditional electromag-
netic analyses, the currently considered problems evermore R EFERENCES
comprise rotor dynamics, thermal aspects, impact of power [1] U.S. Energy Information Administration,
https://www.eia.gov/outlooks/ieo/electricity.php, visited June 16, 2017.
electronics and control strategy, etc. Dozens of variables are [2] Electric Motor Systems Annex EMSA, Website
used to define a design under investigation. http://www.motorsystems.org, visited July 5, 2017.
[3] H. De Keulenaer, R. Belmans, E. Blaustein, D. Chapman, A. De Almeida,
Apart from optimizing geometrical aspects, material science B. De Wachter, and P. Radgen, “Energy efficient motor driven systems,”
European Copper Institute, Tech. Rep., 2004.
finds way in the scenarios and optimized material distributions [4] European Union, “Commission regulation (EC) no 640/2009 of 22 July
or material characteristics are object of research. Manufactur- 2009 implementing Directive 2005/32/Ec of the European Parliament and
of the Council with regard to ecodesign requirements for electric motors,”
ing aspects are incorporated by either improving the applied Official Journal of the European Union, 2009.
models via additional layers for the zone of manufacturing [5] ——, “Commission regulation (EC) no 4/2014 of 6 January 2014 amending
regul. (EC) no 640/2009 implementing Directive 2005/32/Ec of the Euro-
influence, or, globally, by conducting a tolerance analysis or pean Parliament and of the Council with regard to ecodesign requ. for electr.
investigating the design performance with regard to reliability. motors,” Official Journal of the European Union, 2014.
[6] I. O. for Standardization (ISO), “Rotating electrical machines - Part 30: Eff.
Most scenarios deal with multiple objectives and con- classes of single-speed, three-phase, cage-induction motors,” 2000.
[7] A. Boglietti, A. Cavagnino, M. Lazzari, and S. Vaschetto, “Prelim. induction
straints. Besides classical machine performance ratings like motor electromagnetic sizing based on a geometrical approach,” IET Electric
efficiency, torque ripple, power factor, power or torque density, Power Appl., vol. 6, no. 9, pp. 583–592, Nov. 2012.
[8] X. Jannot, J. C. Vannier, C. Marchand, M. Gabsi, J. Saint-Michel, and
installation space, and cost, further objectives like radiated D. Sadarnac, “Multiphysic mod. of a high-speed interior permanent-magnet
noise, robustness, etc. are investigated. synchr. mach. for a multiobjective opt. design,” IEEE Transactions on
Energy Conversion, vol. 26, no. 2, pp. 457–467, June 2011.
Hence, both scenarios and respective solutions are multi- [9] G. Lei, T. Wang, J. Zhu, Y. Guo, and S. Wang, “System-level design
optim. method for electr. drive systems - robust appr.” IEEE Transactions
faceted and rich in variety. As a consequence, optimal designs on Industrial Electronics, vol. 62, no. 8, pp. 4702–4713, Aug 2015.
must be found case by case and no universal machine design [10] J. A. Tapia, A. Parviainen, J. Pyrhönen, P. Lindh, and R. R. Wallace,
“Optimal design proc. for an ext. rotor permanent-magnet mach.” in 2012
exists that can be considered as global optimum. As electric XXth Int. Conf. on Electrical Machines, Sept 2012, pp. 2695–2701.
machine design optimization encompasses multiple topics [11] J. A. Tapia, J. Pyrhönen, J. Puranen, P. Lindh, and S. Nyman, “Optimal de-
sign of large permanent magnet synchronous generators,” IEEE Transactions
of Physics, Mathematics and Computer Science, innovations on Magnetics, vol. 49, no. 1, pp. 642–650, Jan 2013.
[12] E. Carraro, M. Morandin, and N. Bianchi, “Traction pmasr motor optimiza-
from various fields influence this field of research. One can tion according to a given driving cycle,” IEEE Transactions on Industry
await with bated breath new developments to optimize in Applications, vol. 52, no. 1, pp. 209–216, Jan 2016.
[13] C. Lu, S. Ferrari, and G. Pellegrino, “Two design procedures for pm syn-
greater detail, faster, and for obtaining better and better results. chronous machines for electric powertrains,” IEEE Transactions on Trans-
However, even though significant advances in multiple fields portation Electrification, vol. 3, no. 1, pp. 98–107, March 2017.
[14] E. Silvas, T. Hofman, N. Murgovski, L. F. P. Etman, and M. Steinbuch, “Rev.
allow for optimizing more and more in detail, engineers should of opt. strateg. for system-level design in hybr. electr. veh.” IEEE Trans. on
Vehicular Technology, vol. 66, no. 1, pp. 57–70, Jan 2017.
always find the right balance regarding computational effort [15] R. Vincent, V. Emmanuel, G. Lauric, and G. Laurent, “Optimal sizing of an
made and modeling accuracy. electrical machine using a magnetic circuit model: application to a hybrid
electrical vehicle,” IET Electrical Systems in Transportation, vol. 6, no. 1,
Even in the long-term view, solving electric machine design pp. 27–33, 2016.
optimization scenarios at the touch of a button is not within [16] M. Degano, E. Carraro, and N. Bianchi, “Selection criteria and robust
optim. of a traction pm-assisted synchr. reluct. motor,” IEEE Transactions
sight. Thus, it requires experts from electrical engineering to on Industry Applications, vol. 51, no. 6, pp. 4383–4391, Nov 2015.
[17] L. Chen, X. Chen, J. Wang, and P. Lazari, “A computationally efficient multi-
properly address electric machine design optimization prob- physics optimization technique for permanent magnet machines in electric
lems, to develop and evaluate techniques for solving the vehicle traction applications,” in 2015 IEEE International Electric Machines
Drives Conference, May 2015, pp. 1644–1650.
scenarios, and to check reasonableness of the obtained results. [18] J. Aubry, H. B. Ahmed, and B. Multon, “Sizing opt. methodology of a
Last, but not the least, the interest of the power and surface pm machine-converter system over a torque-speed operating profile:
Application to a wave energy converter,” IEEE Transactions on Industrial
industrial electronics researcher community on topics related Electronics, vol. 59, no. 5, pp. 2116–2125, May 2012.
[19] A. Fatemi, N. A. O. Demerdash, T. W. Nehl, and D. M. Ionel, “Large-scale [49] E. Zitzler, M. Laumanns, and L. Thiele, “SPEA2: Improving the Strength
des. optim. of pm mach. over a target operating cycle,” IEEE Transactions Pareto Evolutionary Algorithm for Multiobjective Optimization,” in Evolu-
on Industry Applications, vol. 52, no. 5, pp. 3772–3782, Sept 2016. tionary Methods for Design, Optimisation and Control with Application to
[20] C. Krasopoulos, M. Beniakar, and A. Kladas, “Velocity and torque limit Industrial Problems (EUROGEN 2001), 2002, pp. 95–100.
profile optimization of electric vehicle including limited overload,” IEEE [50] J. Kennedy and R. Eberhart, “Particle swarm optim.” in IEEE Intern. Conf.
Transactions on Industry Applications, vol. PP, no. 99, pp. 1–1, 2017. on Neural Netw., 1995, vol. 4, Nov 1995, pp. 1942–1948.
[21] N. Takahashi, T. Yamada, S. Shimose, and D. Miyagi, “Optimization of rotor [51] J. H. Lee, J. W. Kim, J. Y. Song, D. W. Kim, Y. J. Kim, and S. Y.
of actual ipm motor using on/off method,” IEEE Transactions on Magnetics, Jung, “Distance-based intelligent particle swarm optimization for optimal
vol. 47, no. 5, pp. 1262–1265, May 2011. design of permanent magnet synchronous machine,” IEEE Transactions on
[22] N. Takahashi, T. Yamada, and D. Miyagi, “Examination of optimal design of Magnetics, vol. PP, no. 99, pp. 1–1, 2017.
ipm motor using on/off method,” IEEE Transactions on Magnetics, vol. 46, [52] K. Price, “Differential evolution: a fast and simple numerical optimizer,”
no. 8, pp. 3149–3152, Aug 2010. in 1996 Biennial Conference of the North American Fuzzy Information
[23] S. Sato, T. Sato, and H. Igarashi, “Topology optimization of synchronous Processing Society, 1996. NAFIPS., Jun 1996, pp. 524–527.
reluctance motor using normalized gaussian network,” IEEE Transactions [53] N. Bianchi and S. Bolognani, “Design optim. of electr. mot. by genetic
on Magnetics, vol. 51, no. 3, pp. 1–4, March 2015. algor.” IEE Proceedings - Electric Power Appl., vol. 145, no. 5, pp. 475–
[24] T. Sato, K. Watanabe, and H. Igarashi, “Multimaterial topology optimization 483, Sep 1998.
of electric machines based on normalized gaussian network,” IEEE Transac- [54] B. Mirzaeian, M. Moallem, V. Tahani, and C. Lucas, “Multiobj. optim.
tions on Magnetics, vol. 51, no. 3, pp. 1–4, March 2015. method based on a gen. algor. for switched reluct. mot. design,” IEEE
[25] Y. Hidaka and H. Igarashi, “Topology optim. of synchronous reluctance Transactions on Magnetics, vol. 38, no. 3, pp. 1524–1527, May 2002.
motors considering localized magnetic degradation caused by punching,” [55] G. F. Uler, O. A. Mohammed, and C.-S. Koh, “Design optimization of elec-
IEEE Transactions on Magnetics, vol. PP, no. 99, pp. 1–1, 2017. trical machines using genetic algorithms,” IEEE Transactions on Magnetics,
[26] P. Gangl, S. Amstutz, and U. Langer, “Topology optimization of electric vol. 31, no. 3, pp. 2008–2011, May 1995.
motor using topological derivative for nonlinear magnetostatics,” IEEE [56] F. Messine, B. Nogarede, and J. L. Lagouanelle, “Opt. design of elec-
Transactions on Magnetics, vol. 52, no. 3, pp. 1–4, March 2016. tromechanical actuators: a new method based on global optimization,” IEEE
[27] Y. Li and D. C. Aliprantis, “Opt. design of electromech. devices using a Transactions on Magnetics, vol. 34, no. 1, pp. 299–308, Jan 1998.
hybrid finite element/air-gap element method,” in 2013 IEEE Power and [57] G. Bramerdorfer, A. C. Zavoianu, S. Silber, E. Lughofer, and W. Amrhein,
Energy Conference at Illinois (PECI), Feb 2013, pp. 106–113. “Possibilities for speeding up the FE-based optimization of electrical ma-
[28] Y. Li, S. Ravi, and D. C. Aliprantis, “Tooth shape optimization of switched chines - a case study,” IEEE Transactions on Industry Applications, vol. 52,
reluctance motors for improved torque profiles,” in 2015 IEEE Intern. Electr. no. 6, pp. 4668–4677, Nov 2016.
Mach. Drives Conf. (IEMDC), May 2015, pp. 569–575. [58] D. K. Lim, D. K. Woo, H. K. Yeo, S. Y. Jung, J. S. Ro, and H. K. Jung,
[29] G. Bramerdorfer, S. Silber, G. Weidenholzer, and W. Amrhein, “Compre- “A novel surrogate-assisted multi-objective optimization algorithm for an
hensive cost optimization study of high-efficiency brushless synchronous electromagnetic machine design,” IEEE Transactions on Magnetics, vol. 51,
machines,” in 2013 Intern. Electr. Machines Drives Conference, May 2013, no. 3, pp. 1–4, March 2015.
pp. 1126–1131.
[59] Y. Duan and D. M. Ionel, “A review of recent developments in electrical ma-
[30] O. Gutfleisch, “Controlling the properties of high energy density permanent chine design optimization methods with a permanent-magnet synchronous
magnetic materials by different processing routes,” Journal of Physics motor benchmark study,” IEEE Transactions on Industry Applications,
D: Applied Physics, vol. 33, no. 17, p. R157, 2000. [Online]. Available: vol. 49, no. 3, pp. 1268–1275, May 2013.
http://stacks.iop.org/0022-3727/33/i=17/a=201
[31] O. Gutfleisch, M. A. Willard, E. Brck, C. H. Chen, S. G. Sankar, and J. P. [60] R. C. P. Silva, M. Li, T. Rahman, and D. A. Lowther, “Surrogate-based
Liu, “Magnetic materials and devices for the 21st century: Stronger, lighter, moea/d for electric motor design with scarce function evaluations,” IEEE
and more energy efficient,” Advanced Materials, vol. 23, no. 7, pp. 821–842, Transactions on Magnetics, vol. 53, no. 6, pp. 1–4, June 2017.
2011. [Online]. Available: http://dx.doi.org/10.1002/adma.201002180 [61] A.-C. Zăvoianu, E. Lughofer, G. Bramerdorfer, W. Amrhein, and E. P. Kle-
[32] A. Krings, A. Boglietti, A. Cavagnino, and S. Sprague, “Soft magnetic ment, “An effective ensemble-based method for creating on-the-fly surrogate
material status and trends in electric machines,” IEEE Transactions on fitness functions for multi-objective evolutionary algorithms,” in SYNASC
Industrial Electronics, vol. 64, no. 3, pp. 2405–2414, March 2017. 2013 - 15th International Symposium on Symbolic and Numeric Algorithms
[33] Arnold Magnetic Technologies, Website for Scientific Computing, 2013.
http://www.arnoldmagnetics.com/en-us/Products/RECOMA-SmCo- [62] A.-C. Zăvoianu, G. Bramerdorfer, E. Lughofer, S. Silber, W. Amrhein, and
Magnets-en, visited June 16, 2017. E. P. Klement, “Hybridization of multi-objective evolutionary algorithms
[34] F. Lin, R. Qu, D. Li, and K. Xie, “A fully supercond. homopol. dc mach.” and artificial neural networks for optimizing the performance of electrical
IEEE Trans. on Appl. Supercond., vol. 27, no. 4, pp. 1–5, June 2017. drives,” Engineering Applications of Artificial Intelligence, vol. 26, no. 8,
[35] V. Rallabandi, N. Taran, D. M. Ionel, and J. F. Eastham, “Coreless multidisc pp. 1781–1794, 2013.
axial flux pm machine with carbon nanotube windings,” IEEE Transactions [63] ——, “A hybrid soft computing approach for optimizing design parameters
on Magnetics, vol. 53, no. 6, pp. 1–4, June 2017. of electrical drives,” in Soft Computing Models in Industrial and Environ-
[36] ——, “On the feasibility of carbon nanotube windings for electrical ma- mental Applications, 2013, vol. 188, pp. 347–358.
chines - case study for a coreless axial flux motor,” in 2016 IEEE Energy [64] G. Weidenholzer, S. Silber, G. Jungmayr, G. Bramerdorfer, H. Grabner, and
Conversion Congress and Exposition, Sept 2016, pp. 1–7. W. Amrhein, “A flux-based pmsm motor model using rbf interpolation for
[37] J. Pyrhönen, J. Montonen, P. Lindh, J. J. Vauterin, and M. Otto, “Replacing time-stepping simulations,” in 2013 International Electric Machines Drives
copper with new carbon nanomat. in electr. mach. windings,” Intern. Review Conference, May 2013, pp. 1418–1423.
of Electrical Engineering (IREE), vol. 10, no. 1, 2015. [65] G. Bramerdorfer, W. Amrhein, S. M. Winkler, and M. Affenzeller, “Iden-
[38] M. K. Yoon, C. S. Jeon, and S. K. Kauh, “Efficiency increase of an induction tification of a nonlinear PMSM model using symbolic regression and its
motor by improving cooling performance,” IEEE Transactions on Energy application to current optimization scenarios,” in IECON 2014 - 40th Annual
Conversion, vol. 17, no. 1, pp. 1–6, Mar 2002. Conference of the IEEE Industrial Electronics Society, Oct 2014, pp. 628–
[39] A. J. Rix and M. J. Kamper, “Radial-flux permanent-magnet hub drives: 633.
A comparison based on stator and rotor topologies,” IEEE Transactions on [66] G. Bramerdorfer, S. M. Winkler, M. Kommenda, G. Weidenholzer, S. Silber,
Industrial Electronics, vol. 59, no. 6, pp. 2475–2483, June 2012. G. Kronberger, M. Affenzeller, and W. Amrhein, “Using FE calculations and
[40] M. Galea, G. Buticchi, L. Empringham, L. de Lillo, and C. Gerada, “Design data-based system identification techniques to model the nonlinear behavior
of a high-force-density tubular motor,” IEEE Transactions on Industry of PMSMs,” IEEE Transactions on Industrial Electronics, vol. 61, no. 11,
Applications, vol. 50, no. 4, pp. 2523–2532, July 2014. pp. 6454–6462, Nov 2014.
[41] M. Bali and A. Muetze, “Modeling the effect of cutting on the magnetic [67] G. Bramerdorfer, G. Weidenholzer, S. Silber, W. Amrhein, and S. Lanser,
properties of electrical steel sheets,” IEEE Transactions on Industrial Elec- “Measurement-based nonlinear modeling of PMSMs,” in IECON 2015 - 41st
tronics, vol. 64, no. 3, pp. 2547–2556, March 2017. Annual Conference of the IEEE Industrial Electronics Society, Nov 2015, pp.
[42] Z. Gmyrek, A. Cavagnino, and L. Ferraris, “Estimation of the magnetic 2036–2041.
properties of the damaged area resulting from the punching process: Ex- [68] I. Brown and R. Lorenz, “Response surface methodol. for the design of ind.
perimental research and fem modeling,” IEEE Transactions on Industry machine self-sensing rotor pos. saliencies,” in Intern. Conf. on Electrical
Applications, vol. 49, no. 5, pp. 2069–2077, Sept 2013. Machines and Systems (ICEMS), Oct 2007, pp. 1354–1359.
[43] A. Cavagnino, R. Bojoi, Z. Gmyrek, and M. Lefik, “Stator lamin. geom. [69] A. Fatemi, D. M. Ionel, N. A. O. Demerdash, S. J. Stretz, and T. M. Jahns,
influence on the building factor of synchr. reluctance motor cores,” IEEE “Rsm-de-ann meth. for sens. analys. of active mat. cost in pm mot.” in 2016
Transactions on Industry Applications, vol. PP, no. 99, pp. 1–1, 2017. IEEE En. Conv. Congr. and Expos., Sept 2016, pp. 1–7.
[44] G. Bramerdorfer and D. Andessner, “Accurate and easy-to-obtain iron loss [70] J. Gong, F. Gillon, J. T. Canh, and Y. Xu, “Proposal of a kriging output
model for electric machine design,” IEEE Transactions on Industrial Elec- space mapping technique for electromagnetic design optimization,” IEEE
tronics, vol. 64, no. 3, pp. 2530–2537, March 2017. Transactions on Magnetics, vol. PP, no. 99, pp. 1–1, 2017.
[45] D. C. Montgomery, Design and analys. of experim. J. Wiley&S., 2008. [71] D. K. Lim, Y. S. Cho, J. S. Ro, S. Y. Jung, and H. K. Jung, “Opt. design of an
[46] M. Cavazzuti, Optimization Methods, S.-V. B. Heidelberg, Ed., 2013. axial flux pm synchronous motor for the electric bicycle,” IEEE Transactions
[47] G. E. P. Box and D. W. Behnken, “Some new three level on Magnetics, vol. 52, no. 3, pp. 1–4, March 2016.
designs for the study of quantitative variables,” Technometrics, [72] R. Zhong, Y. Xu, Y. Cao, X. Guo, W. Hua, S. Xu, and W. Sun, “Accur. mod.
vol. 2, no. 4, pp. 455–475, 1960. [Online]. Available: of switched reluct. mot. based on indir. meas. method and least squ. support
http://amstat.tandfonline.com/doi/abs/10.1080/00401706.1960.10489912 vector mach.” IET Electr. Power Appl., vol. 10, no. 9, pp. 916–922, 2016.
[48] K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan, “A fast and elitist multi- [73] H. Sun, Y. Mi, Y. Dong, Y. Zheng, and Y. Dong, “The nonlin. model. of srm
objective genetic algorithm: NSGA-II,” IEEE Transactions on Evolutionary with improved rbf network,” in 2009 Second Intern. Conference on Intelli.
Computation, vol. 6, no. 2, pp. 182–197, 2002. Networks and Intell. Systems, Nov 2009, pp. 54–57.
 [74] B. Zhang, A. Wang, and M. Doppelbauer, “Multi-objective optimization of
a transverse flux mach. with claw-pole and flux-concentr. structure,” IEEE Gerd Bramerdorfer (S’10-M’14) received the PhD
Transactions on Magnetics, vol. 52, no. 8, pp. 1–10, Aug 2016. in Electrical Engineering in 2014 from Johannes
[75] Y. Jin, “A comprehensive survey of fitness approximation in evolutionary Kepler University Linz, Austria. He is currently an
computation,” Soft Computing, vol. 9, no. 1, pp. 3–12, Jan. 2005. Assistant Professor with the Department of Electri-
[76] T. Tannenbaum, D. Wright, K. Miller, and M. Livny, “Condor – a distributed
job scheduler,” in Beowulf Cluster Computing with Linux, T. Sterling, Ed. cal Drives and Power Electronics, Johannes Kepler
MIT Press, October 2001. University Linz. His research interests comprise the
[77] J. Baek, S. S. R. Bonthu, and S. Choi, “Design of five-phase pm assisted design, modeling, and optimization of electric ma-
synchronous reluctance motor for low output torque ripple applications,” IET chines as well as magnetic bearings and bearingless
Electric Power Applications, vol. 10, no. 5, pp. 339–346, 2016. machines. Gerd Bramerdorfer is a member of IEEE,
[78] G. Y. Sizov, D. M. Ionel, and N. A. O. Demerdash, “Multi-objective IEEE Industrial Electronics Society and its Electric
optimization of PM AC machines using computationally efficient - FEA and
differential evolution,” in 2011 IEEE International Electric Machines and Machines Technical Committee, and the IEEE In-
Drives Conference (IEMDC), May 2011, pp. 1528–1533. dustry Applications Society. He constantly serves for the scientific community,
[79] A. Stuikys and J. K. Sykulski, “An efficient design optimization framework e.g., as guest editor, track chair, topic chair, by organizing special sessions,
for nonlinear switched reluctance machines,” IEEE Transactions on Industry and as a reviewer for journals and conferences. He is an Associate Editor of
Applications, vol. PP, no. 99, pp. 1–1, 2017. the IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS.
[80] S. Silber, W. Koppelstätter, G. Weidenholzer, and G. Bramerdorfer, “MagOpt
- Optimization Tool for Mechatronic Components,” in 14th International
Symposium on Magnetic Bearings (ISMB14), 2014.
[81] S. Silber, G. Bramerdorfer, A. Dorninger, A. Fohler, J. Gerstmayr, W. Kop-
pelstätter, D. Reischl, G. Weidenholzer, and S. Weitzhofer, “Coupled opti-
mization in MagOpt,” Proc. of the Institution of Mechanical Engineers, Part
I: Journal of Systems and Control Engineering, 2015. Juan A. Tapia (M’03) received the B.Sc. and M.Sc.
[82] Q. Wang, S. Niu, and S. Yang, “Design optimization and comparative study degrees in Electrical Engineering from University
of novel magnetic-geared permanent magnet machines,” IEEE Transactions of Concepcion, Chile, in 1991, 1997, respectively
on Magnetics, vol. PP, no. 99, pp. 1–1, 2017.
[83] J. M. Crider and S. D. Sudhoff, “An inner rotor flux-modulated pm synchr. and the Ph.D. degree from University of Wisconsin,
machine for low-speed high-torque applications,” IEEE Transactions on Madison in 2002. From 2010 to 2014, he was
Energy Conversion, vol. 30, no. 3, pp. 1247–1254, Sept 2015. a FiDiPro Fellow from Academy of Finland at
[84] X. Zhu, B. Yan, L. Chen, R. Zhang, L. Quan, and L. Mo, “Multi-obj. optim. Lappeenranta University of Technology where he
design of a magnetic planetary geared pm brushless machine by combined conducts research on PM machine on LUT-Energia
design of experiments and response surface methods,” IEEE Transactions on
Magnetics, vol. 50, no. 11, pp. 1–4, Nov 2014. Group. Currently, he is an Associate Professor with
[85] P. Shamsi, M. McDonough, and B. Fahimi, “Wide-bandgap semiconductor the Department of Electrical Engineering, University
technology: Its impact on the electrif. of the transportation industry.” IEEE of Concepcion, Chile. His primary research areas are
Electrification Magazine, vol. 1, no. 2, pp. 59–63, Dec 2013. electrical machines design, numerical method for electromagnetic fields, DSP-
[86] K. O. Armstrong, S. Das, and J. Cresko, “Wide bandgap semiconductor based electric machine control, and renewable energy.
opportunities in power electronics,” in 2016 IEEE 4th Workshop on Wide
Bandgap Power Devices and Applications, Nov 2016, pp. 259–264.
[87] R. Singh and M. Pecht, “Commercial impact of silicon carbide,” IEEE
Industrial Electronics Magazine, vol. 2, no. 3, pp. 19–31, Sept 2008.
[88] D. Han, J. Noppakunkajorn, and B. Sarlioglu, “Comprehensive efficiency,
weight, and volume comparison of sic- and si-based bidirectional dc-dc
converters for hybrid electric vehicles,” IEEE Transactions on Vehicular Juha J. Pyrhönen (M’06, SM’17) born in 1957 in
Technology, vol. 63, no. 7, pp. 3001–3010, Sept 2014. Kuusankoski, Finland, received the Doctor of Sci-
[89] D. Han, Y. Li, and B. Sarlioglu, “Analysis of sic based power electronic ence (D.Sc.) degree from Lappeenranta University
inverters for high speed machines,” in 2015 IEEE Applied Power Electronics of Technology (LUT), Finland in 1991. He became
Conference and Exposition, March 2015, pp. 304–310.
[90] H. R. Choi and G. H. Choe, “A multiobjective parametric optimization for Professor of Electrical Machines and Drives in 1997
passenger-car steering actuator,” IEEE Transactions on Industrial Electron- at LUT. He is engaged in research and development
ics, vol. 57, no. 3, pp. 900–908, March 2010. of electric motors and power-electronic-controlled
[91] C. T. Krasopoulos, M. E. Beniakar, and A. G. Kladas, “Rob. optim. of high drives. Prof. Pyrhönen has wide experience in the
speed pm mot. des.” IEEE Trans. on Magnetics, vol. PP, no. 99, pp. 1–1, research and development of special electric drives
2017.
[92] J. Jang, S. gil Cho, S.-J. Lee, K.-S. Kim, J.-M. Kim, J.-P. Hong, and T. H. for distributed power production, traction and high-
Lee, “Reliability-based robust design optim. with kernel density estim. for speed applications. Permanent magnet materials and
electric power steering motor considering manufact. uncertainties,” IEEE applying them in machines have an important role in his research. Currently
Transactions on Magnetics, vol. 51, no. 3, pp. 1–4, March 2015. he is also studying possibilities of using carbon-based materials in electrical
[93] G. Bramerdorfer, “Computationally-efficient tolerance analysis of brushless machines.
PMSMs,” in 2016 International Conference on Electrical Machines (ICEM),
Sept 2016, pp. 1–6.
[94] G. Bramerdorfer and A. C. Zavoianu, “Surrogate-based multi-objective
optimization of electrical machine designs facilitating tolerance analysis,”
IEEE Transactions on Magnetics, vol. PP, no. 99, pp. 1–1, 2017.
[95] G. Bramerdorfer, “Computationally-efficient tolerance analysis of the cog-
ging torque of brushless pmsms,” IEEE Transactions on Industry Applica- Andrea Cavagnino (M’04-SM’10) was born in
tions, vol. PP, no. 99, pp. 1–1, 2017. Asti, Italy, in 1970. He received the M.Sc. and
[96] J. Pyrhönen, J. Nerg, P. Kurronen, and U. Lauber, “High-speed high-output Ph.D. degrees in electrical engineering from the
solid-rotor induct.-motor technol. for gas compress.” IEEE Transactions on
Industrial Electronics, vol. 57, no. 1, pp. 272–280, Jan 2010. Politecnico di Torino, Turin, Italy, in 1995 and
[97] V. Sihvo, Insulation system in an integrated motor compressor, 2000, respectively. He is currently a Full Professor
ser. Acta Universitatis Lappeenrantaensis, 2010. [Online]. Available: with the Dipartimento Energia of the Politecnico
https://books.google.at/books?id=vlVDAwEACAAJ di Torino. .He has authored or coauthored more
[98] T. Aho, V. Sihvo, J. Nerg, and J. Pyrhönen, “Rotor materials for medium- than 200 papers published in technical journals
speed solid-rotor induction motors,” in 2007 IEEE International Electric and conference proceedings. His research interests
Machines Drives Conference, vol. 1, May 2007, pp. 525–530.
[99] G. Bramerdorfer, A. Cavagnino, and S. Vaschetto, “Cost-optimal ma- include electromagnetic design, thermal design, and
chine designs fulfilling efficiency requirements: a comparison of IMs and energetic behavior of electrical machines. Prof. Cav-
PMSMs,” in 2017 IEEE International Electric Machines Drives Conference agnino is a Past Chair of the Electrical Machines Technical Committee of
(IEMDC), May 2017, pp. 1–6. the IEEE Industrial Electronics Society, and a past Associate Editor of the
[100] A. Cavagnino, G. Bramerdorfer, and J. A. Tapia, “Optimization of elec- IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS and the IEEE
tric machine designs-part I,” IEEE Transactions on Industrial Electronics,
vol. 64, no. 12, pp. 9716–9720, 2017. TRANSACTIONS ON INDUSTRY APPLICATIONS; he is a reviewer for
[101] A. Cavagnino, G. Bramerdorfer, and J. A. Tapia, “Optimization of elec- several IEEE TRANSACTIONS and other international journals. From 2011
tric machine designs-part II,” IEEE Transactions on Industrial Electronics, at present, Prof. Cavagnino was the Guest Editor of six Special Sections for
vol. 64, no. 12, pp. 9716–9720, 2017. the IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS.

View publication stats