Efficiency analysis of a

three-phase power transformer

By R Gouws and O Dobzhanskyi, North-West University

Industries are concerned about the cost of energy; and the lower efficiency of the transformer owing to energy that is lost in it.

ransformers are key elements in the industrial processes into • Voltmeter

T which they are integrated. Reliability is crucial to ensure unin-

terrupted power supply to motors, furnaces and smelters used
in a wide variety of applications including primary aluminium and
•
•
Two wattmeters
Connection wires

steel plants, mines, pump storage power plants, rail networks etc. For Before discussing the methods of how the transformer parameters
example, referring to ‘references’ in this article – in [1] authors discuss are calculated, it is important to explain the important principles of
an importance of efficient transformers feeding electric railways. In machine operation and its equivalent circuit.
[2] the authors touch a subject of transformers’ efficiency in petro- The behaviour of transformers can be considered by assum-
leum industries. Article [3] discovers a use for efficient transformers ing that it has an equivalent ideal transformer. The imperfections,
in the cement industry. Authors in [4] focus on energy saving using losses, magnetic leakage and an imperfect iron core, of an actual
efficient transformers in such industries as the iron-steel sector, non transformer are then drawn into the equivalent circuit by means of
ferrous metal sector, a paper and pulp company, chemical industrial additional circuits or impedances inserted in between the primary
enterprise etc. Owing to a growing number of transformers used source and secondary load [9]. The approximate equivalent circuit
nowadays, the problem of their efficiency is a concern for many of the transformer is shown in Figure 1 [10].
researchers. Efficient use of energy is one of the main problems of There are basically two types of constructions that are in common
each industry [5]. use with transformers – namely shell and core type. The core type’s
The efficiency of a three-phase power transformer is affected by windings are wound around the two outside legs of the magnetic core
power losses. There are two main sources of losses: Winding and and the shell type is wound in the middle of the magnetic core [9].The
core losses which contribute to the total losses of the electrical system alternating current flowing through the primary winding produces an
[5]. Core losses consist of the hysteresis losses in the magnetic core alternating magnetic flux in the transformer’s core.
of the transformer.
Winding losses consist of the losses in the primary and secondary
windings. They depend on the load current and are found as I2R[5].
There are associated losses owing to harmonics but they can be
neglected assuming that the supply voltage of the transformer is not
distorted [6, 7, 8]. That is why it is crucial to operate a transformer as
close as possible to its rated load condition.
Figure 1: Equivalent circuit of the transformer.

Materials and method This magnetic flux by itself induces Electromotive Force (EMF) in the
The materials which are required to conduct the practical tests at any winding placed at the secondary side. The frequencies of the supply
industry and in a heavy current laboratory are: voltage and induced EMF are the same. Owing to induced EMF in
• Three-phase transformer the secondary winding, current flows to the external load which is
• Three-phase voltage supplier connected to its terminals. This way the power is transformed from
• Ammeter or multi-meter primary to secondary winding [11].
• Current transformer Transformers can be connected in numerous ways such as either

E+C SPOT ON • August 2015

Y/Y, ∆/∆, Y/∆ or ∆/Y. The efficiency of a transformer can be calculated Soc = VocIoc (4)
by gaining the ratio of the output power (Pout) to the input power
(Pin) [10]: where Voc and Ioc – are open-circuit voltage and current respectively.
(Pout)
 =  (1) The lagging power factor angle can be calculated by using the ap-
(Pin)
parent power calculated [(S]oc) and the active power (Poc) read from
Pout and Pin are found: the wattmeter [10]:
Pout = Re[V2I2] (2)
1
Pin = Re[V1I1] (3)  =  (5)
cos [ PS ]
oc
oc

where V1, V2, I1 , I2 – voltages and currents of the primary and sec-
ondary windings The reactive power can easily be calculated with Pythagoras [10]:

It is important to note that no transformer will have an efficiency Qoc = Soc

2  Poc
2 (6)
of 100 %. This introduces the possibility of a non-ideal transformer
which consists of losses and effecting factors. It has been required to The core-loss resistance and magnetising reactance can then be cal-
determine the unknown parameters of a given transformer by way of culated by rewriting P=V2R and using values already calculated [10]:
using the open and short circuit tests and performing calculations on
2
the results gained. Thereafter, the calculated results must be used to RcL = Voc (7)
determine the efficiency of the transformer. Poc
2
XmL = Voc (8)
Three-phase power transformers play significant Qoc

roles in industrial sectors in terms of energy saving. For the short-circuit test, the low voltage side of the circuit is con-
nected as a short-circuit, while the high voltage side’s voltage is
slowly incremented from zero V until the low voltage side reaches
Determining transformer parameters using open and
its rated current.
short-circuit tests
This test is designed to determine the winding resistances and
To perform an open-circuit test, one winding of the transformer is left leakage reactance. Rated current in each winding ensures a proper
open while the other is excited. Availability of lower voltage sources, simulation of the leakage flux pattern associated with that winding.
cause the low voltage side to be excited and all measurement equip- [9]. Figure 3 shows the connections that were made for the short-
ment is connected on the same side as source. circuit test. The total resistance as referred to the high voltage side
Even with the transformer experiencing no-load, rated volt- can be calculated by rewriting P=I2R and using the readings from the
age must be applied carefully. Figure 2 shows the connection of wattmeter and the ammeter [10]:
the transformer and shows how the ammeter, voltmeter and two
wattmeters are connected. As shown, the two wattmeter method is ReH = Psc (9)
used so that the three-phase power can be calculated and not only Isc2
the per-phase power. where Psc – active power at short-circuit test, Isc – short-circuit current.

Figure 2: Connection diagram for open-circuit test.

The apparent power is given by the ammeter and voltmeter read- Figure 3: Connection diagram for the short circuit test.
ings [10]:

E+C SPOT ON • August 2015

The total impedance is calculated by using Ohm’s law [10]: The input power and the output power are given mathematically by the
equations (2) and (3). The copper losses are calculated as follows [10]:
ZeH = Vsc (10) Pcu = Ip2Re1 (16)
Isc

where Vsc – short-circuit voltage where Ip – is the current in the primary winding

The total leakage reactance as referred to the high voltage side is The magnetic losses are found by [2]:
easily calculated by using Pythagoras [10]: Pm = Ipn2 Re1 (17)

XeH = Ze2H  Re2H (11) This can be summarised with the following equation [10]:
Pin = Pout + Pcu (18)
The following equations can be used to segregate the winding resist-
ances and the leakage reactance in order to draw an exact equivalent
Experimental results
circuit [10]:
ReH = RH + a2RL (12) In Figure 2, the voltage was taken between points A and B in the star
XeH = ZH + a2XL (13) configuration, it is, thus, the line to line voltage. To get the phase
RH = a2RL = 0,5ReH (14) voltage the line to line voltage is divided by √3. The current measured
XH = a2XL = 0,5XeH (15) is the per-phase current, but to calculate the power per-phase the
power measured has to be divided by three. The Table 1 shows the
where a – is the ratio of number of turns on the low and high per-phase measurements for the open circuit test. The parameters
sides of the transformer; RH, RL, XH, XL – resistances and reactances discussed are included in Table 1.
of the winding on the high and low sides of the transformer
Table 1: Per-phase open circuit results.
Figures 2 and 3 show the approximate circuits and the way the
Parameter Value
transformers must be connected in order to do the two tests. During
the open-circuit test, as shown in Figure 2, the wattmeter measures Voltage (V) 11,55

the core loss in the transformer. It is important to conduct this test Current (A) 1,5
on the low voltage side of the transformer because it is safer and low Power (W) 9,16
voltage power sources are more common. From Figure 2 it can be ReH(Ω) 3,5
seen that the power source supplies an excitation current under no 2 450
ZeH(Ω)
load. The excitation current is responsible for the core-loss and the
XeH(Ω) 2,87
required magnetic flux in the core [9].
The short-circuit test, as shown in Figure 3, is mainly conducted RH(Ω) 1 707,32

to determine the winding resistances and the leakage reactance of

the transformer. It is important to be extremely careful while doing For the short circuit test (see Figure 3), the voltage was taken between
this test because the applied voltage is only a fraction of the rated points A and B in the star configuration, it is, thus, the line to line
voltage. This concludes that core-loss and the magnetising currents voltage. To get the phase voltage the line to line voltage is divided by
are so small that they can be neglected. The test is done on the high √3. The current measured is the per-phase current, but to calculate the
voltage side for safety purposes. Here the wattmeter shows copper power per-phase the total power measured has to be divided by three.
loss at full load [12]. As has been mentioned, efficiency is the ratio Table 2 shows the per-phase measurements for the short-circuit test:
of the output and input power. In the analysed transformer there are
Table 2: Per-phase short circuit results.
two types of losses: Magnetic loss and copper loss. Magnetic loss is
core-loss/fixed loss and is the result of eddy-current and hysteresis Parameter Value

loss. Copper loss is variable loss and is I2R loss [9]. These losses can Voltage (V) 11,5

be shown through a power flow diagram (see Figure 4) [10]: Current (A) 1,5
Power (W) 9,16

ReH(Ω) 4,07

ZeH(Ω) 7,7

XeH(Ω) 6,54

RH(Ω) 2,04

XH(Ω) 3,27

RL(Ω) 2,04

XL(Ω) 3,27
Figure 4: Power losses diagram of the transformer.

E+C SPOT ON • August 2015

For all the efficiency calculations a per-phase load voltage V2 of 220<0°
is used. All calculations performed for the per-phase circuit. The load
current I2 is given by the following calculation:
s
3
I2 = rated% · θ (19)
V2

where θ is given by the inverse cosines of the power factor, in this

case 30°. The formulas will be shown completely for the 60 % rated
load. The turns ratio a for the Y-Y configuration is 220/220 = 1.

At 60 % the per-phase load current in the primary winding is:

s
3
I2 = rated% · θ (20)
V2 Figure 5: The exact equivalent circuit of the transformer.

The induced EMF in the secondary winding is:

The core-loss resistance and the magnetising reactance is much big-
E2 = V2 + [I2 (R]L + jXL) (21)
ger than the winding resistances and the leakage reactance.
The reactance of the low voltage and high voltage is as expected
The induced voltage in the primary winding is given by:
since the transformer has a one-to-one ratio and a Y-Y configuration
E1 = aE2 30 (22)
was used.
The practical test was done on a 1 kVA, 380/380 V transformer.
The current in the primary winding of the transformer is given by:
The following measurements were taken down during the laboratory
I2 test (see Table 4):
Ip = a
30 (23)
Table 4: The laboratory readings.
The per-phase source current is thus given by: Parameters Open Short
circuit circuit
I2 = Ip + E1 ( R1c + jX1 ) (24) Voltage (V) 120 20
m
Current (A) 0,05 1,5
where Rc – is resistance of the core Power 1 (W) 4 12,5
Power 2 (W) 2 15
The per-phase voltage supplied by the source:
Total Power (W) 16 27,5
V1 = E1 + [I1 (R]H + jXH) (25)

As seen in Table 4, the short-circuit test was only done at a rated volt-
The power input is calculated as follows:
age and current and care was taken to not pass the rated current of
Pout = Re[V2 I2] (26)
the transformer. As expected, there is a great current at a low voltage.
The efficiency of the transformer was calculated at a rated load
Lastly, the efficiency calculated by means of equation:
of 60 % to 90 % in 5 % increases and can be seen in Table 5.
(Pout)
 =  (27) Table 5: Transformer efficiency at different rated loads.
(Pin)
Rated Efficiency (%)
The calculated data for the transformer is enclosed in Table 3. load (%)
60 88,160
Table 3: Transformer parameters.
65 88,720
Parameters Impedance (Ω)

Rc 2 450 70 89,188

Xm j1707,32 75 89,580
RH 2,04
80 89,909
RL 2,04
85 90,186
XH j3,27

j3,27 90 90.419
XL

E+C SPOT ON • August 2015

Graphically, the change of efficiency at different loads is shown in [5] Haggerty NK, Malone TP, Crouse J. Justifying the use of high
Figure 6. efficiency transformers. Petroleum and Chemical Industry Confer-
ence. USA. September 1996.
[6] Massey GW. Estimation methods for power system harmonic ef-
fects on power distribution transformers. IEEE Transactions on
Industry Applications. Kansas City, March 1994.
[7] Sharifian MBB, Faiz J., Fakheri SA, Zraatparvar A. Derating of
distribution transformers for non-sinusoidal load currents using
finite element method. ICECS 2003, Proceedings of the 2003
10th IEEE International Conference on Electronics, Circuits and
Systems. December, 2003.
[8] Jayasinghe NR, Lucas JR, Perera KBIM. Power system harmonic
effects on distribution transformers and new design considera-
tions for K factor transformers. IEEE Sri Lanka Annual Sessions.
September 2003.
[9] Hughes E. Electrical technology, 5th ed. New York, USA: Longman
Figure 6: Transformer efficiency curve at different loads.
publishers. 1977.
[10] Guru BS and Hiziroglu HR. New York, United States of America:
The transformer shows an efficiency of between 88 % and 99,5 %
Oxford University Press. 2001.
when operated between 60 % and 90 % of the rated load.
[11] Einstein College of Engineering, Lecturer notes. http://www.
einsteincollege.a-c.in
Conclusion [12] Theraja BL and Theraja AK. A textbook of electrical technology,
5th ed. India: S Chand & Company. 2006.
The results show that the open and short-circuit tests are an effec-
[13] Carroll & Meynell. Carol & Meynell Transformers. http://www.
tive way to calculate the parameters of a non-ideal transformer. The
carollmeynell.com/technicaltransforme-rs.htm
efficiency that was worked out, at certain percentages of the rated
load, is in the range of 88 % to 90 %. The maximum efficiency of a
1 kVA should be in the range of 94 % [13]. The lower efficiency of
the transformer can be ascribed to the inaccuracy of the equipment
(ammeter, wattmeter and voltmeter) and to human error – reading off
from the equipment. The difference can be ascribed to the saturation
of the core as it is made out of magnetic material and previous uses
can affect the core. As there is only a small difference it can be said
that the parameters that were calculated with the measurements of the
tests are correct and, thus, that the tests were successful. The graph
of efficiency changing when the transformer operates at different
loads demonstrates clearly how important it is to use the transformer
at its rated load. Power losses of the transformer increase when the
transformer operates out of its rated load. This causes the efficiency Rupert Gouws holds a Ph.D. degree in Electrical and Electronic
Engineering from the North-West University (Potchefstroom
to go down. For industries it is important to know this phenomena,
campus). He has consulted to a variety of industry and public
since when efficiency gets lower, energy is lost in the transformer.
sectors in South Africa and other countries in the fields of en-
ergy engineering and engineering management. Currently he
References is appointed as an associate professor specialising in energy
engineering, electrical machines and control at the North-West University. The
[1] Zhengqing , Jianzhong W, Shibin L. Study on protection scheme Engineering Council of South Africa (ECSA) registered him as a Professional
for traction transformer of high-speed railway. Power and Energy Engineer and the Association of Energy Engineers (AEE) certified him as a
Engineering Conference (APPEEC). March 2011. Certified Measurement and Verification Professional (CMVP).
[2] Mitchell, GF. Application of transformers for the petroleum in-
dustry. IEEE Trans. on industry applications. September 1969. Oleksandr Dobzhanskyi holds a M.S and Ph.D degrees in
Electrical and Computer Engineering. He graduated from
[3] Kump RK. Why should the cement industry do anything with
Louisiana State University (USA) in 2012. Currently he is
their PCB transformers. Cement Industry Technical Conference.
taking his postdoctoral research at the North-West University
May 1993 in South Africa. Enquiries: Tel. 018 299 1902
[4] Hulshorst WTJ, Groeman JF. Energy saving in industrial distribu- or email [email protected]
tion transformers. May 2002.

E+C SPOT ON • August 2015