Thermal Design of 25ka Current Injection Transformer (CIT) With Finite Element Method
Thermal Design of 25ka Current Injection Transformer (CIT) With Finite Element Method
Thermal Design of 25ka Current Injection Transformer (CIT) With Finite Element Method
Where
max
T is the maximum temperature,
20
is
resistivity at , C 20
and
20
is temperature
coefficient of resistivity at C 20
.
2
The wire sizes are selected from standard wire
tables, which normally specify resistance in
m
at 20 C
.
The total resistive losses for all the windings are
( )
2
A
) A . J ( MLT N
RI P
n
1 i
i
2
i i 2
cu
Where
The voltage drop
X
V is constant in the y direction
and the total foil voltage drop is obtained from [4]:
2
1
X
X
z
X
foil oil f
(8) dx ) x ( V l
H
1
N V
3.2-Primary Windings
Primary winding of current injection transformer
consist of considerable stranded conductor connected
in series. The extents of the wires are smaller than skin
depth corresponding to the applied frequency.
However, the extents of the cross-section of the entire
winding may exceed . As exceeds the dimension
of the wires, the current density may be assumed to be
constant within the cross section of each wire, because
the wires are connected in series, within the cross-
section of the entire winding. The current density
str
J
is related to the current
str
I supplying the stranded
conductor, by
(9) I
D
N
J
str
str
str
str
With
str
N the number of turns and
str
D is the surface
of the cross section
str
of the entire winding. The
voltage drop along the winding is as:
( ) 10 d )) y , x ( A j
f
J
(
D
l . N
V
Z
str
str
str
Z str
str
str
With
str
f the fill factor accounting for the correction of
the overall conductivity made necessary by the
presence of insulation and gaps and
z
l length of
winding [4].
4-Results of 2-D Finite Element Analysis
The transformer modeling in this work has been
carried out using an electromagnetic FEM analysis
softwares, Ansys 5.4 and Cosmos/M [5, 6]. Although
modern leakage field in CIT generally needs to use
3
3-D modeling and computation, in this
particular case the 2-D approach is used because
the leakage field responsible for elements, nodes
and constraints is restricted, detailed modeling
with each conductor representation is carried out
for only the winding whose circulating current is
to be estimated [7].
In all transformers, only half of the core width is
modeled and boundary conditions are placed on
all nodes on the outer edge of the core half, i.e.
the centerline of the core. This is based on the
assumption that the centre of the core is placed
at zero magnetic potential.
In all transformers, full load conditions are
established in all the windings in the model, by
proper external excitation [8].
This transformer is modeled with 4 areas as the
core, the air gaps, and the primary and
secondary windings.
Results of 2-D analysis are as below:
Fig (1) shown flux density of tree-limb of The
CIT.
Fig.1, Core flux density distribution of the current
injection transformer
,
_
2
m
wb
The foil conductor model is applied to simulate
short-circuit operation of a dry-type, single-
phase foil-winding CIT. The current distribution
in the secondary winding is shown in Fig (2)
Fig.2, Current distribution of the CIT for 7 foils-winding
parallel together, solved with Ansys5.4 software
The results of FEM analysis of the circulating-current
loss are shown in Fig (3) and Fig (4)
Fig.3, FEM analysis on the CIT to find leakage flux losses,
solved with Cosmos/M software
,
_
2
m
wb
4
Fig.4, Eddy current distribution obtained from
leakage flux in the CIT secondary winding
,
_
2
m
A
,
solved with Cosmos/M software
Copper and core losses of the CIT are estimated
with consider of current density and volumes of
the windings in the table (1). Losses for
secondary winding are calculated with two
softwares. In the case of A, we use current
density that solved with Ansys5.4, and obtain
losses of secondary winding And B is sum of
main load current losses and eddy currents
losses in secondary winding that obtained with
Cosmos/M software. Output current of the CIT
can melt foil windings, for preventing of this, we
use 7 foils-winding that are parallel together,
current is divided to the foils and temperature of
secondary windings will be limited.
For core losses calculation we use losses curve
of 30M5 core lamination that shown in fig (5)
Fig.5, 30M5 w/kg curve of core laminations used for
calculation of iron losses
Copper and iron losses of the CIT is estimated with
consider of current density and volumes of the core
and windings that shown in table (1).
Table 1-Current injection transformer losses
Core
Primary
winding
Secondary
winding(A)
Secondary
winding(B)
Losses
[kw]
0.18 0.56 1.345 1.41
The warm and cooling of the CIT is analogous to the
process taking place in a three-mass composite body.
The three masses in this case are the winding of the
transformer, core and the air. In the process of
warming up, one part of the heat developing in the
winding and the core increases their heat contents, and
the other part is transferred to the air through the
surfaces of the winding and core in contact with the
air. Starting from the beginning of the warming period,
when t=0 the temperature rise of the winding is
cu
,
that of the core is
core
and that of the air is
air
after time t [9].
The following three differential equation can be
written for the winding, the core and the air:
( ) (11) dt P dt . k A d C m
core core core o core core core core
+
( ) (12) dt P dt K A d C m
cu cu cu o cu cu cu cu
+
( )
( ) (13) dt K A
dt . k A dt K A d C m
cu cu o cu
core core o core air air air air air air
+
+
Writing the above equation in expounded form:
( ) 14 ), t ( g Ax
dt
dx
Required data for thermal analysis are as:
[ ]
[ ]
1
]
1
1
]
1
C Wskg C
kg m
C Wskg C
kg m
o
cu
cu
o
core
core
1
1
390
313
470
190
[ ]
( )
( ) w P P
w P P
C Wskg C
kg m
core
cu
o
ins
ins
180
1900
160
14 / 7
2
1
1
1
]
1
air
core
cu
x
x
x
3
2
1
Where
ins ins
C and m are mass and thermal capacitance
of winding insulations respectively.
The set of inhomogeneous linear differential equations
of first order is as follows:
5
1
1
1
]
1
1
1
1
1
1
1
]
1
1
1
1
]
1
1
1
1
1
1
1
1
]
1
0
2
2 . 8
10
6 8 - 3.4 1 4 1
.08 0 0.08 0
.6 0 0 0.6
10
3
3
2
1
3
3
2
1
x
x
x
dt
dx
dt
dx
dt
dx
With solving of sate space equations,
temperature raise of the core and windings
obtained as below:
Fig (6) warming-up of the CIT under its normal
losses
In Fig (6) ( ) C 60
cu
and ( ) C 18
core
.
Due to short duration of short-circuit, practically
no heat is transferred to the air. After test
duration the heat transfer begins and, after a time
corresponding to a few winding time constants,
the winding assumes its normal service
temperature again
The losses arising during a short circuit in a
conductor of length l is converted into an
increase of the heat content of the conductor
without transferring it to the air if heat stored in
the interturn insulation be neglected temperature
rise in winding is obtain as below[9]:
(15) ) C 235 (
e e
C 31
C 235
. e C 31
w
2
w
1
1
]
1
,
_
+
Where is
( ) 16 t
10
J
392 . 0
2
8
,
_
Where e is eddy-current loss expressed as a percentage
of the D.C. loss at T
max
and
w
is the average
temperature of the winding at incidence of the short-
circuit; and is temperature rise after short-circuit
current,
J is current density and t is short-circuiting time
duration [5].
As the maximum current distribution in the secondary
winding is
,
_
2
8
m
A
10 607 . 0 we have:
4333 / 0
10
10 607 / 0
392 / 0
10
392 / 0
8
8
2
8
1
1
]
1
,
_
3
2
t
J
C 71 ) C 235 (
e e
C 31
C 235
. e C 31
w
2
w
1
1
]
1
,
_
+
C 98 27 71
hotspot
+
Fig (7) shown Temperature rise time curve of hot spot
temperature of the CIT winding during a short-circuit
and it's cooling after test duration.
Fig (7) Temperature rise time curve of hot spot temperature
of the CIT secondary winding during test operation and it's
cooling of the same after test duration
5-CONCLUSIONS
Finite element method for solving of main problem of
transformers is an effective method. Short-circuit time
duration is strongly dependent to amount of output
current and it is one of main limitations for CIT
operation. Temperature of secondary windings of the
CIT limited with 7 parallel foils. Eddy current losses
are prominent factor for temperature rise of the CIT.
6
We estimate secondary losses of the CIT with
two method and the results are equal together.
State space model shown average temperature
rise of windings and core of the transformer. Air
temperature rise is neglected because it circulate
in room test with installed fan on the transformer
This transformer must cool about 2.5 hours with
forced air cooling to be serviceable for another
25KA type test.
REFERENCES
[1]Glenn Swift Senior, Tom.S.Molinski, "A
Fundamental Approach to Transformer Thermal
Modeling-part I-Theory and Equivalent
Circuit", IEEE TRANS ON POWER DELIVERY,
VOL.16, NO.2, APRIL 2001
[2]William Gerard Hurley, Werner Hugo
Wolfle,"Optimized Transformer Design Include
of High-Frequency Effects,"IEEE
TRANSACTIONS ON POWER
ELECTRONICS, VOL.13, NO.4, JULY 1998
[3] R. Petkov, Optimum design of a high-
power high-frequency transformer,IEEE Trans.
Power Electron, vol. 11, no. 1, pp. 3342, 1996.
[4]Herbert De Gersem, Kay Hameyer,"A Finite
Element Model for Foil Winding
Simulation,"IEEE TRANSACTION ON
MAGNETICS, VOL.37, NO.5, SEPTEMBER
2001
[5]"ANSYS Simulation Software", ver5.4, 1997
[6]"Cosmos/M simulation Software", ver1.75A,
1996
[7] D.A.Koppikar, S.V.kulkarni,
G.Ghosh"Circulating-Current Loss in
Transformer Windings"IEE Proc.-
Sci.Meas.Technol, Vol.145, NO.4, July 1998
[8] Gnter F. Meckler and Remiss S. Girgis,
Fellow, IEEE
"Magnetic Flux Distributions in Transformer
Core
Joints" IEEE TRANSACTIONS ON POWER
DELIVERY, VOL. 15, NO. 1, JANUARY 2000
[]K.KARSAI, L.KISS D.Sc."Large Power
Transformers" Amsterdam 1987