Transformer Design

(© Dr. R. C. Goel & Nafees Ahmed )

By

Nafees Ahmed
Asstt. Prof. Department of Electrical Engineering
DIT, University, Dehradun, Uttarakhand

References:
1. Notes by Dr. R. C. Goel
2. Electrical Machine Design by A.K. Sawhney
3. Principles of Electrical Machine Design by R.K Agarwal
4. VTU e-Learning
5. [Link]
6. [Link]

1 [Link] © Nafees Ahmed

OUTPUT EQUATION: - It gives the relationship between electrical rating and physical dimensions
of the machines.
Let
V1 = Primary voltage say LV
V2 = Secondary voltage say HV
I1 = Primary current
I2 = Secondary current
N1= Primary no of turns
N2= Secondary no of turns
a1 = Sectional area of LV conductors (m2)
I
= 1

a2 = Sectional area of HV conductors (m2)
I
= 2

 = Permissible current density (A/m2)
Q = Rating in KVA
We place first half of LV on one limb and rest half of LV on other limb to reduce leakage flux.
So arrangement is LV insulation then half LV turns then HV insulation and then half HV turns.

(1) For 1-phase core type transformer

Rating is given by
Q = V1 I1  10 3 KVA
= 4.44 fm N1 I1 103 KVA  V1  4.44 fm N1 
= 4.44 f  Ai Bm N1 I1 10 3
KVA -----------(1) ( m  Ai Bm )
Where
f = frequency
m = Maximum flux in the core
Ai = Sectional area of core H L L H Window H L L H
Bm = Maximum flux density in the core V V VV V V VV

Window Space Factor

Actual Cu Section Area of Windings in Window

Kw 
Window Area ( Aw ) 1-phase core type transformer with
a N  a2 N 2 concentric windings
 1 1
Aw
( I /  ) N1  ( I 2 /  ) N 2
 1 ( a1  I /1  & a2  I / 2  )
Aw
I N I N
 1 1 2 2
Aw

2 I1 N1
 ( For Ideal Transformer I1 N1  I 2 N 2 )
Aw
So

2 [Link] © Nafees Ahmed

 K w Aw 
 N1 I 1                (2)
 2 
Put the value of N1I1 form equation (2) to equation (1)

 K w Aw
Q  4.44 f A i Bm  103 KVA
2
Q  2.22 f A i Bm  K w Aw  103 KVA          (3) Window

(2) For 1- phase shell type transformer

LV LV
Window Space Factor HV HV
a1 N1  a2 N 2 LV LV
Kw 
Aw
HV HV
( I /  ) N1  ( I 2 /  ) N 2
 1 ( a1  I /1  & a2  I / 2  ) LV LV
Aw
I N I N
 1 1 2 2
Aw 1-phase shell type transformer with
2I N sandwich windings
 1 1 ( For Ideal Transformer I1 N1  I 2 N 2 )
Aw
So
K w Aw
N1I1               (4)
2
Put the value of N1I1 form equation (4) to equation (1)
 K w Aw
Q  4.44 f A i Bm  103 KVA
2
Q  2.22 f A i Bm  K w Aw  10 3 KVA        (5)
Note it is same as for 1-phase core type transformer i.e. equ (3) Window

(3) For 3-phase core type transformer

H L L H H L L H H L L H
V V VV V V VV V V VV

3-phase core type transformer with

Rating is given by concentric windings
Q = 3 V1I1  103 KVA
= 3  4.44 fm N1 I1  10 KVA
3
 V1  4.44 fm N1 
= 3  4.44 fAi Bm N1 I1  103 KVA -----------(6) ( m  Ai Bm )

Window Space Factor

3 [Link] © Nafees Ahmed
 Actual Cu Section Area of Windings in Window
Kw 
Window Area ( Aw )

2(a1 N1  a2 N 2 )

Aw
2  ( I1 /  ) N1  ( I 2 /  ) N 2 
 ( a1  I /1  & a2  I / 2  )
Aw
2( I1 N1  I 2 N 2 )

Aw
2  2 I1 N1
 ( For Ideal Transformer I1 N1  I 2 N 2 )
Aw Window
So
 K w Aw
N1I1               (7)
4
Put the value of N1I1 form equation (7) to equation (6)

 K w Aw
Q  3  4.44 f A i Bm  103 KVA
4
Q  3.33 f A i Bm  K w Aw  10 3 KVA    (8)

(3) For 3- phase shell type transformer

Window Space Factor

a1 N1  a2 N 2
Kw 
Aw
( I /  ) N1  ( I 2 /  ) N 2
 1 ( a1  I /1  & a2  I / 2  )
Aw 3-phase shell type transformer with
I N I N sandwich windings
 1 1 2 2
Aw
2I N
 1 1 ( For Ideal Transformer I1 N1  I 2 N 2 )
Aw
So
K w Aw
N1I1               (9)
2
Put the value of N1I1 form equation (9) to equation (6)
 K w Aw
Q  3  4.44 f A i Bm  10 3 KVA
2
Q  6.66 f A i Bm  K w Aw  10 3 KVA          (10)

4 [Link] © Nafees Ahmed

CHOICE OF MAGNETIC LOADING (Bm)

(1) Normal Si-Steel 0.9 to 1.1 T

(0.35 mm thickness, 1.5%—3.5% Si)

(2) HRGO 1.2 to 1.4 T

(Hot Rolled Grain Oriented Si Steel)

(3) CRGO 1.4 to 1.7 T

(Cold Rolled Grain Oriented Si Steel)
(0.14---0.28 mm thickness)

CHOICE OF ELECTRIC LOADING  

This depends upon cooling method employed

(1) Natural Cooling: 1.5---2.3 A/mm2

AN Air Natural cooling
ON Oil Natural cooling
OFN Oil Forced circulated with Natural air cooling

(2) Forced Cooling : 2.2---4.0 A/mm2

AB Air Blast cooling
OB Oil Blast cooling
OFB Oil Forced circulated with air Blast cooling

(3) Water Cooling: 5.0 ---6.0 A/mm2

OW Oil immersed with circulated Water cooling
OFW Oil Forced with circulated Water cooling

CORE CONSTRUCTION:

(a) U-I type (b) E-I type

(c) U-T type (d) L-L type

5 [Link] © Nafees Ahmed

 45o

(e) Mitred Core Construction (Latest)

EMF PER TURN:

We know
V1  4.44 fm N1              (1)
V1
So EMF / Turn Et   4.44 fm            (2)
N1
and

Q = V1 I1  10 3 KVA (Note: Take Q as per phase rating in KVA)

= 4.44 fm N1 I1 103 KVA
 Et N1 I1 103 KVA        (3)
In the design, the ration of total magnetic loading and electric loading may be kept constant
Magnetic loading = m
Electric loading = N1 I1
m m
So  cons tan t ( say " r" )  N1I1  put in eqution (3)
N1I1 r
m
Q  Et  103 KVA
r
Et
Or Q  Et  103 KVA using equation (2)
4.44 f r
Et  (4.44 f r  103 )  Q
2

Or Et  Kt Q Volts / Turn

Where K t  4.44 f r  10 3 is a constant and values are

Kt = 0.6 to 0.7 for 3-phase core type power transformer
Kt = 0.45 for 3-phase core type distribution transformer
Kt = 1.3 for 3-phase shell type transformer

Kt = 0.75 to 0.85 for 1-phase core type transformer

Kt = 1.0 to 1.2 for 1-phase shell type transformer

6 [Link] © Nafees Ahmed

ESTIMATION OF CORE X-SECTIONAL AREA Ai
We know
Et  Kt Q           (1)
Et  4.44 f m
Or Et  4.44 f Ai Bm           (2)
Et
So Ai            (3)
4.44 f Bm

Now the core may be following types

d d/√2

1-Step 2-Step 3-Step Core 4-Step Core

Or Square- Core Or Cruciform- Core

K= 0.45 0.56 0.60 0.625

d = Diameter of circumscribe circle
For Square core
d d
Gross Area    0.5 d 2
2 2
Let stacking factor
Ki  0.9
Actual Iron Area
Ai  0.9  0.5d 2
 0.45 d 2 (0.45 for square core and take ‘K’ as a general case)
 K d2
So Ai  K d 2
Ai a
Or d
K b
Graphical method to calculate dimensions of the core
Consider 2 step core

90o
 , n  No of Steps
n 1 Ө
i.e n  2 b a
So a  dCos
o
90
  30o b  dSin
2 1

Percentage fill
2-Step
Or Cruciform- Core

7 [Link] © Nafees Ahmed

 Gross Area of Stepped core Kd 2 K i
 
Area of circumcirc le d 2
4
2
0.625d / 0.9
 for 4 Step core
 2
(d )
4
= 0.885 or 88.5%
No of steps 1 2 3 4 5 6 7 9 11
% Fill 63.7% 79.2% 84.9% 88.5% 90.8% 92.3% 93.4% 94.8% 95.8%

ESTIMATION OF MAIN DIMENSIONS:

Consider a 3-phase core type transformer

W

hy

L 0.9d H
W w=
(D-0.9d)

d D

We know output equation 3-phase core type transformer

Q  3.33 f A i Bm  K w Aw  103 KVA

So, Window area
Q
Aw  m2
3.33 f A i Bm  K w  10 3

where Kw =Window space factor

8
Kw  for upto 10 KVA
30  HigherKV
10
Kw  for upto 200 KVA
30  HigherKV
12
Kw  for upto 1000 KVA
30  HigherKV
For higher rating Kw = 0.15 to 0.20

Assume some suitable range for

D = (1.7 to 2) d
Width of the window Ww = D-0.9d
Height of the window

8 [Link] © Nafees Ahmed

 Aw
L (L  Ww  Aw )
width of window(Ww )
L
Generally  2 to 4
Ww
The yoke can have same area as that of the core and can be of same stepped size as core (in this case D y=a,
hy=a). Alternatively it could be of rectangular section. In that case yoke area Ay is generally taken 10% to
15% higher then core section area (Ai), it is to reduce the iron loss in the yoke section. But if we increase the
core section area (Ai) more copper will be needed in the windings and so more cost through we are reducing
the iron loss in the core. Further length of the winding will increase, resulting higher resistance so more cu
loss.
a
Ay = (1.10 to 1.15) Ai
Depth of yoke Dy = a b
Height of the yoke hy = Ay/Dy

Width of the core

W = 2*D +0.9 d
Or W=2Ww+3x0.9d (As Ww = D-0.9d)
Height of the core b a
H = L + 2*hy

Flux density in yoke

Ai 2-Step
By  Bm Or Cruciform- Core
Ay

ESTIMATION OF CORE LOSS AND CORE LOSS COMPONET OF NO LOAD CURRENT IC:

Volume of iron in core = 3*L*Ai m3

Weight of iron in core = density * volume
=  i * 3*L*Ai Kg
 i = density of iron (kg/m3)
=7600 Kg/m3 for normal Iron/steel
= 6500 Kg/m3 for M-4 steel
From the graph we can find out specific iron loss, pcore (Watt/Kg ) corresponding to flux density Bm in core.

So
Iron loss in core =pcore*  i * 3*L*Ai Watt
Similarly
Iron loss in yoke = pyoke*  i * 2*W*Ay Watt
Where pyoke = specific iron loss corresponding to flux density By in yoke

Total Iron loss Pi =Iron loss in core + Iron loss in yoke

Core loss component of no load current

Ic = Core loss per phase/ Primary Voltage
P
Ic  i
3V1
9 [Link] © Nafees Ahmed
ESTIMATION OF MAGNETIZING CURRENT OF NO LOAD CURRENT Im:

Find out magnetizing force H (atcore, at/m) corresponding to flux density Bm in the core and atyoke
corresponding to flux density in the yoke from B-H curve
Bm  atcore / m, BY  at yoke / m 
So
MMF required for the core = 3*L*atcore
MMF required for the yoke = 2*W*atyoke

We account 5% AT for joints etc

So total MMF required = 1.05[MMF for core + MMF for yoke]

Peak value of the magnetizing current

Total MMF required
I m, peak 
3N1
RMS value of the magnetizing current
I m, peak
I m, RMS 
2
Total MMF required
I m, RMS 
3 2 N1

ESTITMATION OF NO LOAD CURRENT AND PHASOR DIAGRAM:

No load current Io V1=-E1

Io  Ic  Im
2 2
Io
Ic
No load power factor
Ic Ф0
Coso 
Io

Im
The no load current should not exceed 5% of the full load current.

E2
No load phasor diagram
ESTIMATION OF NO OF TURNS ON LV AND HV WINDING
V
Primary no of turns N1  1
Et
V
Secondary no of turns N2  2
Et

10 [Link] © Nafees Ahmed

ESTIMATION OF SECTIONAL AREA OF PRIMARY AND SECONDARY CONDUCTORS

Q  103
Primary current I1 
3V1
Q  103 N
Secondary current I2  OR 1 I1
3V2 N2
I1
Sectional area of primary conductor a1 

I2
Sectional area of secondary conductor a2 

Where  is current the density.
Now we can use round conductors or strip conductors for this see the IS codes and ICC (Indian Cable
Company) table.

DETERMINATION OF R1 & R2 AND CU LOSSES:

Let Lmt = Length of mean turn

Resistance of primary winding
L N (m)
R1, dc, 75o  0.021  10 6 mt 1 2
a1 (m )
R1, ac, 75o  (1.15 to 1.20) R1, dc, 75o
Resistance of secondary winding
Lmt N 2 (m)
R2, dc, 75o  0.021  10 6
a2 (m2 )
R2, ac, 75o  (1.15 to 1.20) R2, dc, 75o
Copper loss in primary winding  3I12 R1 Watt
Copper loss in secondary winding  3I 2 R2
2
Watt
Total copper loss  3I1 R1  3I 2 R2
2 2

 3I12 ( R1  R2' )
 3I12 Rp
R01  R p  R1  R2'
Where
 Total resistance referred to primary side
Note: Even at no load, there is magnetic field around connecting leads, tanks etc which causes additional
stray losses in the transformer tanks and other metallic parts. These losses may be taken as 7% to 10% of
total cu losses.

DETERMINATION OF EFFICIENCY:

Output Power
Efficiency 
Input Power

Output Power

Output Power  Losses

11 [Link] © Nafees Ahmed

 Output Power
  100 %
Output Power  Iron Loss  Cu loss
ESTIMATION OF LEAKAGE REACTANCES(X1 & X2):

Assumptions
1. Consider permeability of iron as infinity that is MMF is needed only for leakage flux path in the
window.
2. The leakage flux lines are parallel to the axis of the core.

Consider an elementary cylinder of leakage flux lines of thickness ‘dx’ at a distance x as shown in following
figure.

MMF at distance x
a b2
N1I1 b1
Mx  x
b1 dx
Permeance of this elementary cylinder Lc x
A L H
 o V V
L
L dx
 o mt (Lc =Length of winding ≈ 0.8L)
Lc
 1 L 1
 S  & Permeance  
 o A S
Leakage flux lines associated with elementary cylinder N1I1=N2I2
dx  M x  Permeance
NI L dx
 1 1 x  o mt
b1 Lc
Flux linkage due to this leakage flux x
d x  No of truns with which it is associated  d x MMF Distribution
NI NI L dx
 1 1  1 1 x  o mt
b1 b1 Lc
2
L x
 o N12 mt I1   dx
Lc  b1 
Flux linkages (or associated) with primary winding
2
x
b
L b 
1
L
   o N mt I 1    dx  o N12 mt I1  1 
'
1 1
2

Lc 0  b1  Lc  3 
Flux linkages (or associated) with the space ‘a’ between primary and secondary windings
L
 o   o N12 mt I 1 a
Lc
We consider half of this flux linkage with primary and rest half with the secondary winding. So total flux
linkages with primary winding
o
 1   1' 
2

12 [Link] © Nafees Ahmed

 Lmt  b1 a 
 1   o N12 I1   
Lc  3 2 
Similarly total flux linkages with secondary winding
o
 2   2' 
2
L b a
 2   o N 22 mt I 2  2  
Lc  3 2 
Primary & Secondary leakage inductance
 L b a
L1  1   o N12 mt  1  
I1 Lc  3 2 
2 Lmt  b2 a 
L2    o N 22
  
I2 Lc  3 2 
Primary & Secondary leakage reactance
L b a
X 1  2fL1  2fo N12 mt  1  
Lc  3 2 
Lmt  b2 a 
X 2  2fL2  2fo N 22   
Lc  3 2 
Total Leakage reactance referred to primary side
L  b  b2 
X 01 X P  X 1  X 2'  2f o N12 mt  1  a
Lc  3 
Total Leakage reactance referred to secondary side
L  b  b2 
X 02  X S  X 1'  X 2  2f o N 22 mt  1  a
Lc  3 
It must be 5% to 8% or maximum 10%
Note:- How to control XP?
If increasing the window height (L), Lc will increase and following will decrease b1,
b2 & Lmt and so we can reduce the value of XP.

CALCULATION OF VOLTAGE REGULATION OF TRANSFORMER:

I 2 Ro 2Cos2  I 2 X o 2 Sin2
V .R.   100
E2
R Cos2 X Sin2
 o2  100  o 2  100
E2 / I 2 E2 / I 2
 %Ro 2Cos2  % X o 2 Sin2

TRANSFORMER TANK DESIGN:

Width of the transformer (Tank)

Wt=2D + De + 2b
Where De= External diameter of HV winding
b = Clearance width wise between HV and tank
Depth of transformer (Tank)
Dt= De + 2a
Where a= Clearance depth wise between HV and tank

13 [Link] © Nafees Ahmed

Height of transformer (Tank)
Ht= H + h
Where h=h1 + h2= Clearance height wise of top and bottom

hy

L H
Ww
(D-d)

D D
a

Dt
De

b a b

Wt

h1
W

Ht
H

h2

14 [Link] © Nafees Ahmed

 Tank of a 3-Phase transformer

CALCULATION OF TEMPERATURE RISE:

Surface area of 4 vertical side of the tank (Heat is considered to be dissipated from 4 vertical sides of the
tank)

St= 2(Wt + Dt) Ht m2 (Excluding area of top and bottom of tank)

Let
 = Temp rise of oil (35o C to 50o C)
12.5St  =Total full load losses ( Iron loss + Cu loss)
Total full load losses
So temp rise in o C 
12.5 St
If the temp rise so calculated exceeds the limiting value, a suitable no of cooling tubes or radiators must be
provided

CALCULATION OF NO OF COOLING TUBES: Specific Heat dissipation

6 Watt/m2-0C by Radiation
Let xSt= Surface area of all cooling tubes 6.5 Watt/m2-0C by Convection
Then
Losses to be dissipated by the transformer walls and cooling tube
= Total losses
12.5 St  8.5 xSt   Total losses
6 W-Raditon+6.5 W-Convection=12.5 6.5*1.35 W  8.5 (  35% more) Convection only

So from above equation we can find out total surface area of cooling tubes (xSt)
Normally we use 5 cm diameter tubes and keep them 7.5 cm apart
At= Surface area of one cooling tube
 d tubeltube, mean
Hence
xS
No of cooling tubes  t
At

7.5 Cm

d= 5 Cm

Tank and Arrangement of Cooling tubes

15 [Link] © Nafees Ahmed

WEIGHT OF TRANFORMER:

Let

Wi = Weight of Iron in core and yoke (core volume* density + yoke volume* density) Kg
Wc= Weight of copper in winding (volume* density) Kg
(density of cu = 8900 Kg/m3)
Weight of Oil
= Volume of oil * 880 Kg
Add 20% of (Wi+Wc) for fittings, tank etc.
Total weight is equal to weight of above all parts.

Example 1: Estimate the main core dimensions for a 50 Hz, 3-phase, 200 KVA, 6600/500 V, Star/mesh
connected core type transformer. Use the following data:
Core limb section to be 4-stepped for which the area factor =0.62
Window space factor =0.27
Height of window
2
Width of window
Current density = 2.8 MA/m2
Volts per turn =8.5
Maximum flux density =1.25 Wb/m2.
Solution:
We know emf per turn
Et  4.44 fAi Bm ⇒ 8.5  4.44  50  Ai 1.25 ⇒Ai=0.03063 m2
For 4 stepped core
Ai=K d2 ⇒ 0.03063=0.625 d2 ⇒d = 0.2214 m
We also know
Q  3.33 f A i Bm  K w Aw  10 3
⇒ 200  3.33  50  0.03063 1.25  2.8 10 6  0.27 Aw  10 3
⇒Aw=0.0415 m2 = LxWw ---------- (1)
L
2 ---------- (2)
Ww
Solving (1) & (2) Ww = 0.144 m

L = 0.288 m

Ay =1.15 Ai (Let yoke area Ay is 15% more than area Ai)

Dy= a (Width of largest stamping)
=d Cosϴ
Dy=0.92 d (To give maximum area Ai)
Or =0.95 d ( By Graphical Method)

Selecting Dy =0.92 d ⇒Dy=0.204 m

Also assuming rectangular section for yoke

16 [Link] © Nafees Ahmed

 hy = Ay/Dy=1.15Ai/Dy (Assuming Ay=15% more than Ai)
=1.15x0.03063/0.204 ⇒hy=0.173 m
Overall height H=L+hy =0.288+0.173 ⇒ H=0.461 m

Overall width W= 2D+0.9d = 2Ww+3x0.9d=2x0.144+3x0.9x0.2214

⇒ W=0.88578 m

Example 2: Calculate no load current of a 400 V, 50 Hz, 1-Phase, core type transformer, the particulars of
which are as follows:
Length of means magnetic path =200 Cm,
Gross core section =100 Cm2,
Joints equivalent to 0.1 mm air gap,
Maximum flux density =0.7 T,
Specific core loss at 50 Hz & 0.7 T =0.5 W/Kg,
Ampere turns =2.2 per cm for 0.7 T,
Stacking factor =0.9,
Density of core material = 7.5x 103 Kg/m3.
Solution:
Find IC:
Totol core loss Specific core loss  Weight of core
Core loss component of no load current I C  
V1 V1

Specific core loss  K i  Agi  length  density  0.5  0.9 100 10 4  200 10 2  7.5 10 3
IC  
V1 400

⇒ IC=0.168 A
Find Im:
We know V1  4.44 fAi Bm N1
400  4.44  50  0.9 100 10 4  0.7  N1 ⇒ N1=286

Totol MMF MMF for core  MMF for airgap of length of 0.1mm
Magnetizing component I m  
2 N1 2 N1
1 1
MMF for core  Bm  lg 2.2  200  0.7   0.110 3
0 4 10 7  1 
Im    MMF    S  Bm  l g 
 0 
2 N1 2  286 

⇒ Im=1.226 A
So No load current
I 0  I C2  I m2  0.168 2  1.226 2

⇒ I0=1.237 A

Example 3: Design an adequate cooling arrangement for a 250 KVA, 6600/400 V, 50 Hz, 3-phase,
delta/star core type oil immersed natural cooled transformer with the following particulars:
Winding temperature rise not to exceed 500 C,
Total losses at 900 C are 5 Kw,
Tank Dimensions height x length x width = 125 x 100 x 50 (all in cm)
Oil level = 115 cm length
Sketch diagram to show the arrangement of cooling tubes.
17 [Link] © Nafees Ahmed
Solution:
Dissipating surface area of plain tank after neglecting the top and bottom
St=2(Wt+Dt)Ht=2(50+100)125=3.75x104 cm2=3.75 m2
Total full load losses 5000
   106.66 0 C
12.5 S t 12.5 3.75
But it is required that the temp rise is not to exceed 500 C. So cooling tubes are required.
Let xSt = Surface area of all cooling tubes
12.5 St  8.5 xSt   Total losses
⇒ xSt=6.25 m2
Surface area of one cooling tube (Assuming Tube dia = 5 cm, average height of tube =105 cm)
At  d tubel tube, mean  3.14  0.05 1.05  0.1649 m 2

xS t 6.25
No of cooling tubes    38
At 0.1649
Let the tubes to space 7 cm apart centre to centre, we will be able to accommodate 13 tubes on 100 cm side
and 6 tubes on 50 cm side.
Total tubes =2x13+2x6=38

7 Cm

50 Cm

100 Cm

Tank and Arrangement of Cooling tubes

18 [Link] © Nafees Ahmed