International Journal of Advanced Computer Science, Vol. 2, No. 11, Pp. 435-440, Nov., 2012.

Manuscript
Received:
30,Jan., 2012
Revised:
15,Jun., 2012
Accepted:
19,Aug., 2012
Published:
15,Dec., 2012

Keywords
Electrical,
Generator,
I nsulation,
Slot,
Stator,

Abstract It should be noted that most
important part of an electrical machine is the
Stator insulation. This fact reveals the
requirement for inspection of the electrical
machine insulation along with the electrical
tensions. Therefore with respect to insulation
system improvement of stator, the HV
generator can be optimized. Electrical
conductivity of groundwall insulation is
major design consideration for insulation
system of HV generator. A very powerful
method available to analyze electromagnetic
performance is Finite Element Method
(FEM) which is used in this paper. These
processes of optimization have been done
according the proposed optimization
algorithm. In this algorithm the technical
constraints have been considered. This paper
describes the process used to perform
classical design and improvement analysis of
stator slots insulation with respect to
objective function and constraints.


1. Introduction
Stator winding insulation is one of the most critical
components for reliable operation of ac electric machines of
all types and sizes. Since the stator insulation is
continuously exposed to an electrical stress, gradual
degradation of the insulation is inevitable [1]. Industrial
researches show that troubles initiated in the stator winding
insulation are one of the primary root causes of electric
machine failures [2, 3 and 4]. It is shown in [5, 6] that
30-40 % of ac machine failures are stator related and also
shown in [7] that 60-70 % high voltage machine failures
result from stator insulation troubles. The winding and core
of stator integrity plays essential role in the reliability of the
alternator [8-12].
The electrical insulation system of high voltage
rotating machines is one of the most important parts with
respect to the manufacturing costs as well as to the
maintenance and life time aspects [13]. Almost one-third of


This work was supported by the Iran University of Science and
Technology.
Diako Azizi is with the University of Science and Technology, Tehran,
Iran (corresponding author to provide phone: +989128195358; e-mail:
azziizii@iust.ac.ir).
Dr. Ahmad Gholami is with University of Science and Technology,
Tehran, Iran. He is now with the Department of electrical engineering,
(e-mail:gholamii@iust.ac.ir).
the forced outages of large generators in generating stations
and industrial plants are caused by the failure of the
insulation system in the stator winding [14]. The stator
insulation system of generators consists of groundwall
insulation, phase-phase insulation as well as turn and the
strand insulation. The groundwall insulation is the main
insulation that separates the high voltage copper conductors
from the grounded stator core, and the turn or strand
insulation prevents shorts between the turns or the strands in
a coil. Modern machines use mica flake or mica paper tapes
bonded with epoxy or polyester resin as groundwall
insulation [13, 15].
Thus, performing the optimal design consist of
electromagnetic study in core, winding and insulation of
stator slots is necessary. To attain an improved design,
electromagnetic analysis with observe to real condition is
performed which completes pervious researches.
The reason of applying the method is investigating the
effects of electrical and magnetically stresses on insulation
parts. The main idea involves finding the optimum electrical
conductivity of stator slots groundwall insulation in order
to minimize the electrical stress. FEM analysis is used to
simulate the performance of generator. The main
supremacies of the simulations in comparison with other
researchers are:
Simulating ordinary rotation of rotor.
Considering magnetic saturation of core.
Considering of technical constraints.
Considering core losses.
Considering insulation losses.
Considering copper losses.
Considering ambient conditions.
According to this method, electrical conductivity of
stator slot groundwall insulation is optimized to investigate
the possibility of improvement high voltage generator
characteristics.
2. Case study
To study a sample, it is required to determine or design
the sample first. Therefore, in order to do field analysis and
to study the generator insulation Status, a generator with a
given profile has been selected. The selected generator is
synchronous, three phase, 4 poles that have 24 slots in
stator. Rated frequency, voltage and power are respectively
50Hz, 13.8 kV and 10 MVA. Wiring in this study is
form-wound multiturn type and has many layers insulation
with different specification (Fig. 1). In this study turn
Minimization of Electrical Field in Stator Slot
Groundwall Insulation of HV Generators
Diako Azizi & Ahmad Gholami
International Journal of Advanced Computer Science, Vol. 2, No. 11, Pp. 435-440, Nov., 2012.
International Journal Publishers Group (IJPG)

436
insulation and the strand insulation are the same, i.e. nylon
type. Ground wall insulation type is PMMA and semi
conductive coating is Si(c) with characters identified in
table 1.


Fig. 1 Cross sections of stator slot insulation layers for form-wound
multiturn coils in selected HV generator

TABLE 1
ELECTRICAL AND THERMAL CHARACTERISTICS OF INSULATION LAYERS
Quantities Insulation layers
Nylon PMMA Si(c)
heat capacity
[J/(kg*K)]
1700 1420
700

youngs modulus
[Pa]
2e9 3e9 170e9
Thermal expansion coeff.
[1/K]
280e-6 70e-6 2.6e-6
relative permittivity
4

3

11.7
thermal conductivity
[W/(m*K)]
0.26 0.19 130
Density
[kg/m^3]

1150

1190

2329
Electrical conductivity
[S]
0 0 0

3. Electromagnetic model
Amperes law is the main part to derive
electromagnetic system equation [13].

t
D
J B v E
t
D
J H
e
c
c
+ + + =
c
c
+ = V o o
(1)
Where:
- E is the electric field intensity
- D is the electric displacement or electric flux
density
- H is the magnetic field intensity
- B is the magnetic flux density
- J is the current density
- J
e
is the externally generated current
- is the electrical conductivity
- v is the velocity
Time variant-harmonic fields effect can be introduced
by equations (2) and (3):
A B V =
(2)
t
A
V E
c
c
V =
(3)
Amperes law is rewritten by equations (2) and (3)
Combining with constitutive relationships B=
0
(H+M) and
D=
0
E+P , as:
( ) ( )
( ) ( ) P j J V j A v
M A A j
e
e ec o o
c e eo
+ = V + + V
V V +

0
1
0 0
2
(4)
In which ,
0
,
0
, M and P respectively refer to
Angular frequency, Relative permittivity, Relative
permeability, magnetization vector and electric polarization
vector.
In the case of 2-dimensional-plane, there are no
variations in z-direction, so the electric field is parallel to
z-axis. , therefore is written as V/L, where V is the
potential difference over the distance L. Now these
equations are simplified to:

( )
z
e
z
z z
x
y
z
P j J
L
V
A j A v
M
M
A
e o
c e eo o
+ +
A
=
|
|
.
|

\
|
+ V +
(

V V

0
2 1
0
. .
(5)
In the ax-symmetric case, another form of the electric
potential gradient has been used (

) as the
electric field is only present in the azimuthally direction.
The above equation, in cylindrical coordinates, becomes:
| |
( )
P j J
r
V
u V u j r
z
u
r
u
v r
M
M
u
z
z
u
r
u
r
z r
e loop
r
r
z
e
t
o
o c e e o o

+ + =
+ +
|
|
|
.
|

\
|
(
(

c
c
c
c
+
|
|
|
.
|

\
|
|
|
|
.
|

\
|
(

+
(
(

c
c
c
c
c
c
c
c

2
2 .
0
.
0
2
1
0
1
0
(6)
The dependent variable u is the nonzero component of
the magnetic potential divided by the radial coordinate r, so
that:
r
A
u

=
(7)
The application mode performs this transformation to
avoid singularities on the symmetry axis.

Diako Azizi et al.: Minimization of Electrical Field in Stator Slot Groundwall Insulation of HV Generators.
International Journal Publishers Group (IJPG)

437
4. Optimization algorithm
The optimization algorithm in this paper builds on the
proven SNOPT package developed by Prof. Philip Gill
(University of California, San Diego) along with Profs.
Walter Murray and Michael Saunders (Stanford University).
SNOPT is a general-purpose system for large-scale
nonlinearly constrained optimization.
The optimization section builds on a flexible data
structure, the opt structure, which contains the entire
optimization problem within a single variable. The main
optimization function, selects the proper solver algorithm
depending on properties of the objective function and the
constraints. Linearity, for example, allows it to take
appropriate shortcuts.
For optimization involving fields of physics and
geometric properties, such as shape optimization, it can be
connected the optimization section to FEM to perform
optimization on a finite element model.
Numerical optimization can be described as the art of
finding an optimum set of parameters or
optimization-variable values that minimize an objective
function subject to a number of constraints. The objective
function (also known under different names such as goal
function, cost function, or quantity of interest) can be an
explicit function of the optimization variables, or which is a
more interesting case an output quantity from a simulation
parameterized by the optimization variables.
While the optimization application mode can indeed
solve classical pure optimization problems, its real strength
is its ability to compute accurate gradients also when the
objective function and constraints depend on the solution to
a multi physics model. This enables the use of an efficient
gradient-based optimization algorithm, reducing the number
of multi physics model evaluations required to find an
optimal solution.
To define and solve a complete optimization problem,
it will be needed to perform four tasks. In addition to those
that apply when solving a standard PDE problem:
Define a scalar objective function
Select optimization variables
Set additional constraints
Compute optimal values of the optimization variables
The optimization application mode provides an
interface for handling all four tasks.
The optimization application mode is built around a
general minimization problem formulation, where it will be
supplied an objective function to be minimized and
constraints on the optimization variables. On top of this, the
complete multi physics model and its constraints serve as an
additional condition on the optimization problem. For this
reason, the procedure is sometimes referred to as
PDE-constrained optimization.
After it have been set up a complete optimization
model, the goal of the optimization process can be stated
roughly as find the values of the optimization variables
that minimize the objective function under the condition
that u is a solution to the multi physics model and no
constraints are violated. More formally, this can be stated as
follows:

(8)

The objective function Q(u, ) is a function of the
optimization variables { i } both directly and indirectly
through u, which is the solution to the general multi physics
problem L(u, ) = 0. The objective function is defined as the
sum of integral contributions from all dimensions k in the
d-dimensional multi physics model. The optimization
procedure is presented in the Fig. 2.


Fig. 2 Flow chart of the optimization procedure
5. Simulation and results
Fig. 3 shows the primary generator. Fig. 4 shows this
structure at zoomed view. Voltage distribution along the
one of the stator slots for this 3 phase synchronous
generator that has 24 slots in stator and 2 winding layers in
each slot has been presented at Fig. 5. Optimized electrical
conductivity for groundwall insulation layer has been
showed at Fig. 6. Electrical potential distribution for
optimized sketch at groundwall insulation layer has been
presented in Fig. 7. The optimization procedure leads to
International Journal of Advanced Computer Science, Vol. 2, No. 11, Pp. 435-440, Nov., 2012.
International Journal Publishers Group (IJPG)

438
plan which corresponding electrical field distribution
presented in Fig. 8. It can be seen that electrical field
distribution in groundwall insulation has been improved a
lot.


Fig. 3 Primary generator


Fig. 4 Typical stator slot of generator



Fig. 5 Electrical field distribution for primary sketch


Fig. 6 Optimized electrical conductivity for groundwall insulation layer
Diako Azizi et al.: Minimization of Electrical Field in Stator Slot Groundwall Insulation of HV Generators.
International Journal Publishers Group (IJPG)

439

Fig. 7 Electrical potential distribution at groundwall insulation layer for
optimized sketch


Fig. 8 Electrical field distribution at groundwall insulation layer for
optimized sketch
6. Conclusion
Selection of proper insulation materials in several
layers of stator winding can decrease these stresses.
Different schemes with regard to different construction have
different characters and specification. The SNOPT
optimization algorithms with respect to finite element
electromagnetic analyses are useful to improve the
insulation system of generator which conquests the previous
problems in classical design. In this paper, according to
optimization algorithm, optimized groundwall insulation
presented. According to optimized plan, maximum value of
electrical field decreases and electrical field distribution
becomes more uniform along the groundwall insulation
layer in stator slots compared to the primary mode. It should
be noted that additional costs due these changes in
insulation structure are negligible for manufactures.
References
[1] F. Jinkyu Yang, Sang Bin Lee, Jiyoon Yoo, Sanghoon Lee,
Yongmin Oh and Changho Choi, A stator winding insulation
condition monitoring technique for inverter-fed machines
IEEE Transactions on Power Electronics, Vol. 22, No. 5, Sep,
2007.
[2] F. Tim Emery and Dennis Pavlik, Electrostatic field analysis
of high voltage generator stator insulation systems, 2000
conference on electrical insulation and dielectric phenomena,
IEEE.
[3] Muhammad Arshad, Abdul Khaliq and Syed M. Islam,
Turbo generator stator winding condition assessment,2004
international conference on power system technology,
POWERCON 2004, Singapore, 21-24 November 2004.
[4] Sang Bin Lee, Jinkyu Yang, Karim Younsi and Raj Mohan
Bharadwaj, An On-Line groundwall and phase to phase
insulation quality assessment technique for AC machine
stator winding, IEEE Transactions on Industry Application,
vol. 42 no.4, July/August 2006.
[5] P. ODonnell, Report of large motor reliability survey of
industrial and commercial installations: Part I, IEEE Trans.
Ind. Appl., vol. IA-21,no. 4, pp. 853864, 1985.
[6] P. ODonnell, Report of large motor reliability survey of
industrial and commercial installations: Part II, IEEE Trans.
Ind. Appl., vol. IA-21, no. 4, pp. 865872, 1985.
[7] P. F. Albrecht, J. C. Appiarius, and D. K. Sharma,
Assessment of reliability of motors in utility
applicationsUpdated, IEEE Trans. Energy Convers., vol.
EC-1, no. 1, pp. 3946, Mar. 1986.
[8] G. C. Montanari and M. Cacciari, A probabilistic insulation
life model for combined thermal-electrical stresses, IEEE
Transactions on Electrical Insulation vol. EI-20 no.3, June
1985.
[9] G. J. Anders, SM J. Endrenyi, F G.L. Ford, M G.C. Stone,
SM, A probabilistic model for evaluating the remaining life
of electrical insulation in rotating machines, lEEE
Transactions on Energy Conversion, vol. 5, no. 4, December
1990.
[10] S.B.Pandey, Estimation for a life model of transformer
insulation under combined electrical & thermal stress, IEEE
Transaction on reliability. vol.41, no.3, 1992 September.
[11] H.S.Endicott, B.D.Hatch, R.G.Sohmer, Application of
eyring model to capacitor aging data, IEEE Trans.
Component parts, vol 12, 1965 Mar, pp 34-41.
[12] T.S.Ramu, On the estimation of life power apparatus
insulation under combined electrical and thermal stress,
IEEE Trans. Electrical insulation, vol EI-20, 1985 Feb, pp
70-78.
[13] M. Kaufhold, K. Schaefer, K. Bauer, A. Bethge and J. Risse,
Interface Phenomena in Stator Winding Insulation
Challenges in Design, Diagnosis and Service Experience,
IEEE Electr. Insul. Mag., Vol. 18, No. 2, pp. 27-36, 2002.
[14] B.K Gupta and W.T Fink, A Proposed Type Test for
Interturn Insulation in Multi-turn Coils, IEEE Intern.
Sympos. Electr. Insul. (ISEI), Montreal, Quebec, Canada, pp.
International Journal of Advanced Computer Science, Vol. 2, No. 11, Pp. 435-440, Nov., 2012.
International Journal Publishers Group (IJPG)

440
235-238, 1996.
[15] G.C. Stone, B.K. Gupta, J.F. Lyles and H.G. Sedding,
Experience with Accelerated Aging Tests on Stator Bars and
Coils, IEEE Intern. Sympos. Electr. Insul., Toronto, Canada,
pp. 356-360, 1990.
[16] Diako Azizi, Ahmad Gholami, Abolfazl Vahedi, Analysis of
the deterioration effects of stator insulation on the its
electro-thermal property, International Journal of Electrical
Power and Energy Systems Engineering., vol. 2, no. 3, 2009.

Diako Azizi was born in 1985. He has
received B.Sc. degree in Electrical
Engineering from Tabriz University,
Tabriz, Iran in 2007. And he received the
Master degree in Electrical Power
Engineering from the University of
Science and Technology, Tehran, Iran in
2009. He is presently pursuing the Ph.D.
degree in Electrical Power Engineering,
Iran University of Science and
Technology. His research interests are aging of insulations in
electrical machines.


Ahmad Gholami has received his B.Sc.
Degree in electrical engineering from
IUST, Tehran, Iran, in 1975, the M.Sc.
and PhD. Degrees in electrical
engineerin from UMIST, Manchester,
England, in 1986 and 1989 respectively.
He is currently an associate professor in
the Electrical Engineering Department of
Iran University of Science and
Technology. His main research activities are high voltage
engineering, electrical insulation, insulation coordination,
transmission lines and substations planning.