Induction Motor Design Methodology based

on Rotor Diameter Progressive Growth

A. Boglietti, Senior Member, IEEE, A. Cavagnino, Senior Member, IEEE, M. Lazzari,
A. Miotto, and S. Vaschetto, Student Member, IEEE
Politecnico di Torino, Dipartimento di Ingegneria Elettrica, ITALY

Abstract – A methodology to design three-phase induction motors and so on). In fact, when the machine has to be designed
is presented. The procedure is based on a simple geometrical starting from a fresh sheet of paper, a tool able to provide
approach and self-consistent equations, without any use of geometrical dimensions, slot shapes, and turn number of the
empirical/corrective coefficients. In the proposed design tool, the winding is requested. For this reason, it is an authors’ opinion
machine geometrical dimensions are automatically computed
starting from few design specifications imposed by the user, such
that the FEM has to be considered and used as a powerful and
as rated voltage and torque, electric and magnetic material most valuable tool for the analysis and the optimization of
solicitations. In the paper the design procedures and algorithms preliminary solutions defined by means of suitable design
are described step-by-step and, for each step, the used equations tools. This opinion is also confirmed when, in order ‘to design’
are reported in detail, in order to allow readers to develop own the machine, the FEM geometrical model is defined in a
‘self-do-it’ software for induction motor electromagnetic design. parametric way. This because an initial geometry must be in
The authors developed this software in Microsoft Excel electronic any case defined and, in addition, arbitrary variation ranges of
sheet using the embedded Visual Basic for Applications. To prove the dimensions have to be assumed. Obviously, the same
the goodness and accuracy of the proposed design tool, a considerations can be done to software for electrical machines
comparison with two commercial software for electrical machine
design and analysis has been done and discussed.
analysis/design based on analytical approaches and self
consistent equations, where the lamination dimensions,
Keywords: induction motor, electromagnetic design, sizing winding characteristics and material properties are requested as
equations, Finite Element Method. input data.
In the paper the attention is focused on the electromagnetic
I. INTRODUCTION
design of induction machines using a geometrical approach,
The induction motor electromagnetic design is a well known providing all the used algorithms and equations. As a
problem and, in the technical literature, many approaches to consequence, the readers can develop own ‘self-do-it’ software
solve this problem have been already proposed [1]-[7]. In for induction motor electromagnetic design. The authors
addition, several commercial packages are available for design developed this tool in Microsoft Excel electronic sheet, using
induction motors (e.g. [8]) and, in many cases, the company the embedded Visual Basic for Applications (VBA), but,
R&D offices have developed own software. Roughly speaking, obviously, any other programming language can be used.
the methodologies commonly used to design induction
machines can be classified in two main types of automatic II. PROBLEM DEFINITION
tools: tools based on a general theoretical approach and tools
based on past experiences and historical data-base collection. The approach to the induction motor electromagnetic design
The first approach allows a very good flexibility to the designer presented in this paper is not based on the use of empirical
and it is typical of commercial software and research activities equations got by the designer experience and historical design
developed by academia. The second approach, welcome in activities, as often made by several manufacturers. In the
industrial areas, uses tools based on the past experience. This proposed method the motor is designed through an iterative
approach has a very low flexibility because new designs which process based on a progressive increase of the rotor diameter
are far from the old ones are not supported by previous results. until the specified rated torque is obtained. The requested
In the last 20 years, the use of Finite Element Method (FEM) is specifications to define and to start the design process are of
become a standard tool [1]. Thanks to FEMs it is possible to different kinds and they are listed in Table I, where the rotor
perform very accurate analyses of electromagnetic devices, aspect ratio (1) is defined as the ratio between the rotor pack
modeling in depth several phenomena and performances [9]- axial length (Lcore) and the outer rotor diameter (Dr). This ratio
[12]. In fact, magnetic, thermal and mechanical aspects can be defines the shape (stocky or extended) of the motor.
λ  core
studied, both one-by-one and as coupled field problems. L
(1)
Unfortunately, even if FEM allows a detailed analysis of Dr
electromagnetic structures, the magnetic structure must be The rotor skewing ratio (2) is defined as the ratio between the
known in each its detail (dimensions, material properties). In rotor skew pitch i and the pole pitch p. Both the pitches are
other words, FEM is not able to define the machine measured around the airgap perimeter as shown in Table I.
Ri  i
geometrical dimensions and the turn number of the winding
starting from the well known design specifications (e.g. rated (2)
voltage and torque, electric and magnetic material solicitations, p

978-1-4577-0541-0/11/$26.00 ©2011 IEEE 3104

 TABLE I
LIST OF THE DESIGN SPECIFICATIONS
A. Induction motor rated data B. Permissible active materials stress
Vr Rated phase voltage Bag Maximum air-gap flux density
fr Rated frequency Bts, Bys Maximum flux density (stator teeth and stator yoke)
Npole Number of poles Btr, Byr Maximum flux density (rotor teeth and yoke)
Tr Rated torque Js, Jr Rms current density (stator and rotor conductors)
C. Geometric specifications
λ = Lcore/Dr Rotor aspect ratio
Dsh Shaft diameter
h11s, h12s, wos Dimension of the stator slot enclosure
h11r, h12r, wor Dimension of the rotor slot enclosure
hag Thickness of the airgap

D. Pack and filling coefficients E. Stator winding specifications

Kir Lamination pack coefficient q Number of slots per pole and phase
Kcs, Kcr Filling coefficient of the stator and rotor slots nr Winding pitch reduction (in slot number)

cos  Rated power factor

F. Rotor winding specifications G. First approximation quantities

Ri = i/p Rotor slot skewing ratio

Nsr Rotor slot number
ΔV Stator voltage drop at rated load
Ktor Corrective coefficient for the torque computation

The shaft diameter can be defined on the basis of the For this reasons can be useful to highlight some particular
mechanical torque transmitted by the motor to the mechanical aspects concerning the slot type to be used in the motor
load, or in particular cases (for example hollow shaft) it can be realization. The rotor slot shape can be rectangular (slot with
chosen as an independent variable. Also the airgap thickness parallel edges) or trapezoidal (tooth with parallel edges), as
can be fixed ‘a priori’ at this design stage, but its value has to shown in Fig.1.
be checked in the final project verification, such as the imposed The trapezoidal slot is usually chosen in die cast rotor cage,
rated power factor and the stator voltage drop at rated load. because it allows a better magnetic exploit of the rotor tooth.
As well known, it is mandatory to take care at the definition of The rectangular slot is better when the rotor cage is realized
the permissible active materials stresses. In fact, high values of using copper bars.
the current densities can lead to excessive conduction losses
and to a consequent motor overheating and high flux density
values can lead to high iron losses and high value of the no- Rectangular slot Trapezoidal slot
load current with a low motor power factor.
About the flux density values the authors suggest to respect the

 1  1.8  1  1.5
B t s, r 
following inequality relationships:
B yr
Bag Bag Fig. 1. Example of the slot shapes.
On the basis of the used active materials and of their insulation
it is possible to define the lamination pack coefficient and the
filling coefficient of the stator and rotor conductors. Slot top width wt
At this point, one might think that the imposed specification
number is too high. Anyway, after a quick analysis of these Slot opening wo
specifications, it is easy understand that many of these values
are ‘in the mind’ both of expert and inexperienced designers. h1 Slot enclosure height

III. ROTOR LAMINATION DESIGN

As first attempt, the rotor diameter can be defined a little bit hu Slot useful height
higher than the shaft diameter (Dr > Dsh). The rotor lamination

L core  λ  D r
length is defined by the rotor aspect ratio, as shown in (3).
Au Slot useful
(3) area
The preliminary problem to be solved is the determination of
the rotor slots dimensions (width and height of the slot) taking Slot bottom width wb
into account the magnetic and geometrical constrains Fig. 2. Slot dimensions.
previously defined.

3105
 Dr
hur
A2

wrotor tooth /2
A3 wtr
wbr
P

O sr

Fig. 3. Complex shape rotor slot characterization. Dsh Dext_yr

hrotor yoke Dr -h1

In addition, since the rotor slots are usually closed or semi-

closed, it is necessary to define a radial bulk h1 imposed by the Fig. 4. Rotor lamination design for trapezoidal rotor slot.
slot enclosure border. Furthermore, the slot opening w0 has to
be defined too (Fig. 2). The values h1 and w0 will not be change Dr
during the loop for the increase of the rotor dimensions.
Often the rotor slot shape does not cope the two cases
previously described. Since the rotor slot shape is a decisive
factor for the motor electromechanical behaviors during the wrt /2
starting operations (starting torque and current) the rotor slots wtr
are always shaped with narrowing or enlarging along the radial
height for getting the requested skin effect during the starting.
In presence of particular rotor slot shapes (Fig. 3), the rotor O sr wrt /2
design can be done using a simpler slot shape (rectangular or hyr
trapezoidal) and doing a correction of the natural filling Dsh Dext_yr Dr -h1
coefficient Kcr of the rotor conductor using a factor lower than Fig. 5. Rotor lamination design for triangular rotor slot.
one, as shown in (4).
In (4), A3 is the area of the shaped slot and A2 is the area of the
At this stage, the rotor lamination drawing can be defined as
simplified one. In this way it is possible to take into account
shown in Fig. 4. Let define:
α sr  2π N sr
the incomplete use of the available area of the shaped slot.
K'cr  3  K cr
A rotor slot angular pitch;
D ext_yr  Dsh  2h rotor yoke
(4)
A2 rotor yoke external diameter.
A. Trapezoidal slot With reference to Fig. 4 and by simple geometrical
In the case of trapezoidal slot the geometrical computation can considerations the following geometrical dimensions of the
be summarized as it follows. The fundamental harmonic of the rotor slot can be obtained.
D r  D ext_yr
2  D r  L core h ur   h1
flux for each pole can evaluated using the equation:
Φ pole  Bag
Slot useful height
(5) 2
 
N pole
The maximum flux in the rotor yoke and in the rotor teeth can w br   D ext_yr  rotor tooth  tan sr
sin α sr 2 
w α

Slot bottom width
be determined using (6) and (7) respectively. 2
D L
Φ rotor yoke   Bag r core (6) w tr  w br  2  h ur  tan α sr
Φ pole
Slot top width
2 N pole 2
π  D r  L core
Φ rotor tooth  Bag
It is important to remark that these dimensions have to be
(7) verified in term of geometrical compatibility. In particular, two
N sr
cases have to be considered.
By the rotor yoke and rotor tooth flux values and on the basis
of the imposed yoke and tooth flux density (Byr and Btr), the Case 1: hur < 0. In this case this design attempt has to be left

increased of a defined step ( D r  D r old  D r ) or with a

thickness of the rotor yoke and tooth can obtained by: and the computation must start again with a new rotor diameter

h rotor yoke 
Φ rotor yoke
(8)
K ir B yr L core higher flux density value in the rotor yoke (Byr).

w rotor tooth 
Φ rotor tooth Case 2: The trapezoidal slot bond needs that the slot bottom
(9) width has to be greater than zero, as shown hereafter.
K ir B tr L core

3106
 wbr=wtr
wrt
w br  0  D ext_yr  rotor tooth
sin α sr 2 
w
h1

If this condition is not verified the rotor slot shape will be

triangular, but this does not mean that the design should be hur
unacceptable, as shown in Fig. 5. In fact, this condition occurs Dr
if the rotor tooth width is very large or if the rotor yoke width
is very narrow.
In this case it is necessary to redesign again the slot and to
determine the rotor yoke starting from the condition of Dext_yr
degenerated slot (triangular). As a consequence, the yoke
width, the slot height and the slot width on the top have to be Fig. 6. Rectangular slot design.
compute following the hereafter procedure.

w br  0  D ext_yr 
sin α sr 2
Useful area for the
w rotor tooth rectangular slot.
Dr

D ext_yr  Dsh
h rotor yoke 
wrotor tooth /2

D r  D ext_yr  2h1
2 Dext_yr

h ur 
Dcritical
2 B

w tr  2  h ur  tan sr
α Dsh
P1
P2
2
sr
P
In this case, the rotor yoke flux density will be lower than that O
the imposed one.
hrotor yoke
B. Rectangular slot h1
With a rectangular slot the tooth width is not constant and it hur
has a variable thickness, as shown in Fig. 6. Fig. 7. Rotor lamination design for rectangular rotor slot.
Since the flux density Btr has been defined as the maximum
1D 
h ur_opt  PB   r  h1  rotor tooth 
2sin α sr 2 
value reached in the tooth, this value will be verified in the
1 w
2 2
bottom of the tooth. In the other tooth sections the magnetic
solicitations will be, of course, lower than the imposed one. 2
Also in this case, the yoke height and the tooth width will be In this case, the optimal external rotor yoke diameter results
determined starting from the flux in the yoke and in the tooth. equal to:
D ext_yr, opt  rotor tooth  r  h1
2sin α sr 2 2
These quantities can be calculated using (6), (7), (8), (9). w D
The external diameter of the rotor yoke Dext_yr and the slot (12)
useful height hur are defined by (10) and (11).
D ext_yr  Dsh  2h rotor yoke
This condition is possible only if the rotor yoke external
(10)
D r  D ext_yr
diameter computed by (10) is lower than the diameter

h ur   h1
evaluated by (12).
(11) Let’s define the choice between these two values of the rotor
2
yoke external diameter (when possible); the rectangular slot
These preliminary geometrical determinations have to be
D r  2h1  D ext_yr
geometry can be computed by (13) and (14).
verified taking into account the possibility to get the
rectangular slot. In particular, the useful height of the slot must h ur  (13)
be positive (hur > 0). If this condition is not verified, the 2
procedure proposed for the trapezoidal slot can be followed.
  α 
w br  w tr   D ext_yr  rotor tooth  tan sr 
The previous relations allow defining an area where it is

sin α sr 2    2 
w

possible to include the drawing of the rectangular slot, as (14)
shown in Fig.7.
Nevertheless, in the obtained area it is possible to include
several rectangular slots, as shown in Fig. 7. The concept of IV. TORQUE COMPUTATION
“best shape” for the rotor slot can be conditioned by the With the determination of the rotor lamination dimensions, it is
necessity of deep slot (deep bar: P1 in the picture) or by the possible to compute the torque available with this design
necessity of a maximum area slot (minimum bar resistance: P2 attempt. The rotor dimensions (length and diameter) will have
in the picture). In particular, the slot with the maximum area to to increase step by step until the obtained torque will be equal
be included in the available area can be obtained when the slot or greater than the requested one. Hereafter the procedure for
has an ‘optimal height’ equal to half of the segment PB. the electromagnetic torque computation is reported.

3107
 wor h11 geometrical skewing coefficient Kgi, it is possible to define the
net area of the conductive material and, as a consequence, from
A1 h12 h1 the defined current density it is possible to get the rms current
in the rotor bar, as shown in (20).
I r, rms  J r  K cr  K gi  A sr (20)
Aur hur
C. Flux density distribution around the airgap.
Let’s number the rotor slots with a progressive index k from 1
wbr to Ncr and let’s consider the fundamental distribution of the air-
wtr gap flux density and slot current only (see Fig.10). The active
airgap flux density in correspondence of the rotor slot k, in a
Fig. 8. Rotor slot design. generic time instant, can be written in the following way:
 N pole 
Bk  Bag  sin   α k  where α k  k 
Lcore 2π
 2  N sr

Rotor bar
If the rotor bars are skewed, the flux density to take into
A’ur account for the electromechanical interaction is the average one
i= Ri along the bar, as shown in the following equation.
 N pole 
p Aur
Bm, k   Bag sin   α k 
 
i (21)
2
Fig. 9. Effect of the rotor slot skewing. In the previous equation:
 R π 
 sin  i 
2
Ri  π  2 
A. Rotor slot useful area is the electric skewing coefficient;
i
After the definition of the rotor slot shape and its dimensions, it
is possible to get the useful slot area. This area represents the
Ri·π is the electric skewing angle.
available slot portion for the winding conductors.
In general way, the useful rotor slot area Aur can be defined D. Current distribution in the rotor slots
(both for trapezoidal and rectangular slot) by the following The fundamental distribution of the rotor current induced in the

A sr  A ur  A1  w br  w tr  ur  A1
relation: rotor conductors by the fundamental flux density previously
h described can be written by means of (22).
 N pole 
(15)
I r, k  2  I r, rms  sin   α k  
2
 2 
The added term A1 is the contribution of the closing slot zone (22)
to the total available area, as shown in Fig. 8. It is important to
underline that the value of the closing zone height h12 and h11 The angle is the bar current delay angle with respect to the
can be defined ‘a priori’ during the design process. As a voltage induced in the bar. This angle is function of the slip
frequency and it can be determined as follow:

A sr  w br  w tr  ur  w or  w tr  12 tan    slip 
consequence it is possible to get (16).


h h L , bar
(16)
2 2 R bar
B. Effect of the rotor slot skewing where L , bar and Rbar are the leakage inductance and resistance
If the rotor slots are skewed, the available area is reduced in of the conductive cage bar respectively. At rated load condition
accordance to (17), depending on the geometrical skewing it is possible to assume that the slip is very small and the angle

A'sr  K gi A sr
coefficient Kgi. This condition is highlighted in Fig.9. can be considered equal to zero.
(17)
The geometrical skewing coefficient Kgi is defined by (18), Ir,k
where i is the skew pitch. Bm,k
K gi 
L core
L core  i2
(18)
2

αk α
The skew pitch i can be imposed, at the level of design
specification, as a fraction Ri of the motor pole pitch (see (2)).
2π  D r
i  Ri  p  Ri  (19)
N pole
Fig. 10. Flux density and rotor bar current distributions around the airgap.
By the filling coefficient of the rotor conductors Kcr and the

3108
E. Torque determination
The motor torque can be evaluated by the interaction between
the air-gap flux density distribution (sinusoidal) with amplitude
Bag and the rotor bar currents. The torque can be computed as
the addition of the contributions of each bar. Vr
C   Bk  I r, k  core r
N sr Er Er =Vr -ΔV
L D
(23) Is
k 1 2 Is
The addition of the contributions can be developed in the I’r I’r
following way:
N sr 1 N  N 
C  Bag  I r, rms  L core  D r  sin  α k   sin   α k   
2
φ
pole pole

   2 
2
i
k 0 2 φ

  N sr  Bag  I r, rms  L core  D r  cos

2
i
4
Putting in evidence the stress indexes in the active materials
and the rotor geometrical coefficients, the following equation
can be written. Im Im
C design  K tor  N sr  i  Bag  J r  K cr  K gi  A sr  L core  D r
2 Fig. 11. Vector diagram in rated condition (left side) and the adopted
4 approximation (right side).
The coefficients Ktor is less than 1 (generally it is in the 0.9 – Then, the number of conductors in series per phase Zph for

 the phase delay between the rotor current wave and the
0.95 range) and it takes into account: getting the airgap flux density distribution able to produce the
requested E.M.F. is:

 the motor friction losses; Z ph  

flux density one; N pole
2E r
2π  f r  K w  i  Bag D r L core
 the additional torques linked to the higher spatial
(27)

harmonics. As a consequence, the number of conductors in series per slot

results:

Zc 
V. FINAL DESIGN STEPS
Z ph
q  N pole
After the rotor dimensions Dr and Lcore which allow the (28)
production of the requested torque Cdesign have been defined, it
is possible to start with the design of other motor parts. The conductor number Zc should be approximated to an integer
number or to a simple fraction of the pole number if the
A. Cage ring section winding will be realized connecting in series or in parallel the
The cage short circuit ring section can be evaluated taking into poles (or the pole pairs).
account that between the ring current and the bar current there
is the following relation.

 
C. Stator phase rated current

I ring 1 Once the conductors per phase Zph and the rotor bar rated
(24) current Ir are known, it is possible to get in first approximation
Ir 2 sin πN pole N sr
the stator phase rated current. In fact, from the magneto-motive
Assuming to use in the ring the same current density adopted force production point of view, the rotor bar current Ir is
for the rotor bar, the ring section can be computed by (25). equivalent to a fictitious stator current Ir’ flowing in the stator

 
A ring 
K cr K gi A sr winding. This equivalent current Ir’ can by calculated by (29).
I'r 
(25)
2 sin πN pole N sr N sr
3K w  Z ph
Ir (29)
B. Stator phase winding At this time the magnetizing current Im is not still known. As a
stator winding (see section II: Vr, V, q, nr, and cos), it is
Taking into account the specifications initially imposed for the consequence, since the stator phase current includes both the
magnetizing current and the rotor current reported to the stator,
possible to calculate the number of turns of the three phase it is possible to use the rated power factor cos in order to
winding. estimate the approximated stator current (30). In terms of
The rated E.M.F. Er, induced in the stator phase winding by the vector diagrams, the adopted simplification is shown in Fig.11.
approximated as Er = Vr - V (see Fig. 11).
airgap flux density distribution of amplitude Bag, can be The errors due to this approximation can be neglected at this
design step.
Is  r
For a typical three phase winding structure with short pitch, the
I'

cosn r π 6q 
cos
stator winding coefficient Kw is defined by (26). (30)

K w  K w (q, n r ) 
2q  sin π 6q 
(26) Since the power factor at rated condition depends on several
motor quantities (leakage reactance, number of poles, airgap

3109
thickness, etc.) only when the design will be completed an
estimation of the actual power factor will be possible. For these
reasons, at this design step it is possible to use a prudential Dinn_ys
power factor value of about 0.7 - 0.85.
D. Stator slot dimensions

N ss  3q  N pole
The stator slot number is: wts wbs
(31) Q
The stator slots have to be dimensioned for supporting the O ss P
wst
requested stator current. The stator slot area for the winding
can be computed by (32). h1 hus
A ss  3
Z ph Is Dis
N ss K cs  J s
(32) Fig.13. Trapezoidal slot: useful height determination.

In the previous equation, Kcs is the stator slot fill coefficient

Let’s define the angular pitch of the stator slot α ss  2π N ss .
Trapezoidal slot
and Js is the stator current density previously defined for the
stator conductive material. The difference between the stator The slot bottom width wbs depends on the slot useful height hus
inner diameter and the external rotor diameter is given by two by the following equations (see Fig.13).
Dis  D r  2h ag
 h 
times the airgap thickness hag.
w bs  w ts  1  us 
 QP 
(33) (38)

D is  2h1
The minimum width of the stator tooth can be evaluated taking
QP  
2sin α ss 2 
into account the maximum stator tooth flux density. w st
Bag π  Dis
(39)
w st 
2
K ir  B ts N ss
(34)
In this case, the slot useful height hus has to be computed by
solving the second order equation (40).

The stator slot width at the top of the slot (close to the airgap)

π D is  2h1 
A us  w ts  h us  w ts us
can be obtained as difference, as shown in (35). 1 h2
w ts   w st 

2 QP
 rejecting the inconsistent solution
(35)
N ss

(40)

h us  2A us  QP w ts  QP 2  QP
In (35), h1 is the stator slot closing zone height, as shown in


Fig. 12. As discussed for the rotor lamination, rectangular and
trapezoidal slot shapes can be considered also for the stator.
But in this case the problem to be solved is different, because
now the slot area is known and the slot height has to be Rectangular slot
determined. With reference to the trapezoidal geometry shown The stator slot bottom width is equal to the stator slot top
width, as a consequence, the following two relations can be
A us  A ss  A1 , where A1 is the closing zone area chosen for
in Fig. 12, let’s define the useful area Aus of the slot as
w bs  w ts
easily defined.
(41)

A1  w ts  w os   h12
the stator slot. It can be calculated by (36).
h us 
1 A us
(36) (42)
w ts
2
Independently by the slot shape, the following relation has to E. Stator yoke

w  w bs
be verified, where wbs is the bottom slot width. The height of the stator yoke must support the stator yoke flux
A us  ts  h us
respecting the original maximum flux density fixed for the
(37)
D L
2 stator yoke Bys.
Aus Φ stator yoke  Bag r core (43)
N pole
wst
h stator yoke 
Φ stator yoke
hus (44)
K ir B ys L core
wts

Dinn_ys  Dis  2  h1  h us 

The stator yoke inner diameter is defined by (45).
h12
A1 h11
h1 (45)
hag
Finally, the external stator yoke diameter Dext_ys can be
computed as shown in (46).
Dext_ys  Dinn_ys  2  h stator yoke
wos Dr
(46)
Fig. 12. General layout of the stator slot.
3110
 VI. FINAL REMARKS TABLE II
PERCENTAGE DIFFERENCES BETWEEN THE PROPOSED DESIGN METHOD AND
At this step all the geometrical dimensions of the machine COMMERCIAL SOFTWARE (MEANING OF THE SYMBOLS LISTED IN TABLE I)
magnetic laminations have been determined. Modification of Design #1 Design #2 Design #3
the slot shape (such as fillets, chamfers, special shapes and so Speed Flux Speed Flux Speed Flux
Bag 0.0 -2.0 -1.6 +3.5 +2.4 -0.9
on) can be introduced at this time. The electromagnetic design
Bts -0.8 +2.0 -1.6 +1.3 +1.6 +6.0
can be now subjected to the computations for the electrical, Bys -0.9 +5.0 -0.5 -0.7 +1.3 -3.0
thermal and mechanical evaluation, in order to verify the full Btr -0.5 +1.3 -8.9 +5.1 -2.7 +7.3
specific compliances [14]-[16]. In particular, the quantities Byr -4.0 +1.4 +2.7 +8.3 +5.6 +3.2

accuracy (such as the stator voltage drop V and the angles φ

defined ‘a priori’, will have to be evaluated with greater Tr [Nm] -0.2 - -7.0 - +3.0 -
Ir [A] +0.3 - 0.0 - 0.0 -
I0 [A] +9.0 0.0 +8.0% 0.0 -6.0 0.0
and at rated conditions, see Fig. 11).
The proposed design methodology can be easily implemented
The authors have been applied with success this method to
in any programming language and, for this reason, it can be
design several induction motors with very different rated
quickly suited to optimization analyses, acting in interactive
torque and speed values.
way on the design specifications listed in Table I. As a
consequence, the electromagnetic structure can be optimized REFERENCES
from several points of view, such as geometrical dimensions, [1] A.C. Smith, “Integrating FE into induction motor design-a marriage of
weight, losses and efficiency. inconvenience?”, Current Trends in the Use of IEE Seminar on Finite
Elements (FE) in Electromechanical Design and Analysis, 2000 pp.1-7.
VII. VALIDATION OF THE DESIGN PROCEDURE [2] G. Madescu, I. Boldea, T.J.E Miller, “An analytical iterative model (AIM)
for induction motor design”, Conf. Rec. IEEE IAS’96, Vol. 1, 6-10
The proposed design methodology has been used with success October 1996, pp. 566 – 573.
for designing several types of induction motors, in particular [3] C. Tindall, P. Calvert, “Computer-Aided Synthesis and Optimization of
low voltage industrial induction motors from few kW up to 1 Induction Motor Design”, IEEE Trans. on Manufacturing Technology,
MW, very high speed machines (up to 80000 rpm) for high Vol. 6, n. 2, pp. 31 – 37, June 1977.
[4] J.R. Lopez, “The effect of induction motor design on losses and efficiency
vacuum molecular pump applications and small three-phase for adjustable-frequency drives”, Conf. Rec. ISIE’94, 25-27 May 1994, pp.
induction motors with die cast copper squirrel cage [13]. 214 – 219.
Unfortunately, due to confidentiality, the details of these [5] T.J.E Miller, M. McGilp, A. Wearing, “.Motor design optimization using
induction machine projects cannot be reported. So, in order to SPEED CAD software”, IEE Seminar on Practical Electromagnetic Design
Synthesis, 11 February 1999, pp. 2/1 - 2/5.
provide a validation of the proposed method a comparison with [6] D.G. Dorrell, “Combined Thermal and Electromagnetic Analysis of
a commercial software [8] and with a 2D magnetostatic finite Permanent-Magnet and Induction Machines to Aid Calculation”, IEEE
element method package has been done [17]. In particular, Trans. on Industrial Electronics, Vol. 55, No. 10, pp. 3566-3574, 2008.
three machines have been designed by the proposed method [7] M. Centner, U. Schafer, “Optimized Design of High-Speed Induction
Motors in Respect of the Electrical Steel Grade”, IEEE Trans. on
and the obtained solutions have been implemented in the Industrial Electronics, Vol. 57, No. 1, pp. 288-295, January 2010.

 Design #1: 2-poles, 45 kW, 146 Nm, 50 Hz, 660 V

commercial software. The three TEFC induction motors are: [8] The Scottish Power Electronics and Electric Drives Consortium, SPEED,

 Design #2: 4-poles, 90 kW, 585 Nm, 50 Hz, 660 V

Glasgow, UK [on line]. Available on http://www.speedlab.co.uk.
[9] L. Alberti, N. Bianchi, “A Coupled Thermal–Electromagnetic Analysis for

 Design #3: 6-poles, 660 kW, 6400 Nm, 50 Hz, 690 V

a Rapid and Accurate Prediction of IM Performance”, IEEE Trans. on
Industrial Electronics, Vol. 55, No. 10, pp. 3575-3582, October 2008.
The percentage differences on the most critical design [10] Z. Gmyrek, A. Boglietti, A. Cavagnino, “Estimation of Iron Losses in
quantities between the proposed design method and the Induction Motors: Calculation Method, Results, and Analysis”, IEEE
Trans. on Industrial Electronics, Vol. 57, No. 1, pp. 161-171, 2010.
commercial software are reported in Table II and they [11] J.F. Watson, D.G. Dorrell, “The use of finite element methods to improve
demonstrate the accuracy of the proposed design method. The techniques for the early detection of faults in 3-phase induction motors”,
authors want to remark that these percentage differences are IEEE Trans. on Energy Conversion, Vol. 14, No. 3, pp. 655-660,
reasonable and, in general, they can be considered more than September 1999.
[12] S. Williamson, S. L.H. Lim, M.J. Robinson, “Finite-element models for
satisfactory during the initial design stage. In addition, these cage induction motor analysis”, IEEE Trans. on Industry Applications,
differences could due to the adopted computation algorithms Vol. 26, No. 6, pp. 1007-1017, Nov.-Dec. 1990.
for taking into account the magnetic saturation phenomena. [13] A. Boglietti, A. Cavagnino, L. Feraris, M Lazzari, “Energy-efficient
motors”, IEEE Industrial Electronics Magazine, Vol. 2, No. 4, pp.32-37,
VIII. CONCLUSION December 2008.
[14] A. Boglietti, A. Cavagnino, M. Lazzari, “Computational Algorithms for
The paper reports a three-phase induction motor design method Induction Motor Equivalent Circuit Parameter Determination Part I:
based on a simple geometrical approach. Following the Resistances and Leakage Reactances”, IEEE Trans. on Industrial
proposed step-by step procedure, the machine geometry and Electronics, 2011, IEEE Early Access.
[15] A. Boglietti, A. Cavagnino, M. Lazzari, “Computational Algorithms for
the winding structures can be determined. The method requires Induction Motor Equivalent Circuit Parameter Determination Part II: Skin
few design specifications and it is based on an increase of the effect and Magnetizing Characteristics”, IEEE Trans. on Industrial
rotor dimensions until the torque specification is verified. The Electronics, 2011, IEEE Early Access.
proposed approach does not require the typical [16] A.Boglietti, A.Cavagnino, M.Lazzari, M.Pastorelli, “A Simplified Thermal
Model for Variable Speed Self Cooled Industrial Induction Motor”,
experience/historical database collection and can be easily Transaction on Industry Applications, Vol.39, No.4, 2003, pp.945-952.
implemented by the company design offices. [17] FLUX 2D, Cedrat Group, available on http://www.cedrat.com.

3111