Professional Documents
Culture Documents
Abstract:
Magnetic Balance test is a most commonly used proactive test to detect faults in the core
and/or the windings of the transformer at early stage of manufacturing at works. Although
considered as the simplest test to conduct, sometimes it is difficult to interpret the results
conclusively because of some unpredictable variations in test results. Invariably, for
authenticating the analysis of results, we have to conduct some supplementary tests also.
Nevertheless, at times we do not find suitable technical explanation for certain strange test
results, leaving us clueless about the type of abnormality in the transformer.
This paper discusses typical trends of induced voltages in un-excited limbs of transformer,
observed during magnetic balance test and the measurement of magnetizing currents. It also
attempts to give explanation for the observed trend such as lesser magnetizing current in the
centre limb than the extreme limbs, the effect of delta connected windings on distribution of
3-ph magnetizing currents and the combined voltage magnitude in the un-excited limbs in
excess of the applied voltage, etc.
Through various case studies supported by data analysis, the author shares his extensive
testing experiences about magnetic balance and magnetizing current tests and suggests
possibilities of leveraging the knowledge for detection of various faults and abnormalities
within the windings and magnetic core of the transformer. The case studies present step by
step method for analysis of test results to help diagnosis and location of faults.
Findings of the paper will help gain an insight in to the subtle quality problems in the
windings and core during manufacturing of transformer or the in-service problems developed
during operation of the transformers, and enrich readers knowledge encouraging optimum
deployment of magnetic balance test and the complementary magnetizing current
measurements, which are recognized as very effective tests to detect typical problems like:
a) Inter-turn shorting;
b) Inter-strand shorting;
c) Any external loops around the core;
d) Abnormal magnetizing current due to unequal turns in winding sections connected in
parallel;
e) Wrong interleaving joints in windings.
The paper discusses several aspects of core and winding configurations with a view to
address some of the most sought-after answers on the subject.
When we apply voltage to individual limb, and measure the voltages in other two limbs, ideal
expected conditions are,
Table:1
Measured Voltage in Volts Total Voltage induced Magnetizing
in other 2 phases Current
2U-2N 2V-2N 2W-2N Phase
249 (100%) 178 (71.5%) 71 (28.5%) 100.0% U 158.0 mA
125 (50.2%) 249 (100%) 124 (49.8%) 100.0% V 118.6 mA
71 (28.5%) 178 (71.5%) 249 (100%) 100.0% W 158.3 mA
Table:2
From the foregoing we can say that DC test affects both the test results (voltage distribution
as well as magnetizing current trend) substantially. Hence it is advisable to perform this test
before conducting any DC test on the transformer.
In large rating transformers it is observed that ideal condition of magnetic balance is not
achieved, if this test is performed from HV winding (normally HV winding is placed away
from the core). If low voltage is applied to HV winding, its large no of turns results in poor
V/T in the windings,
which results in development of poor flux and the test results obtained may not be reliable.
Hence it is advisable to conduct this test from the winding closer to the Core in case of large
rating transformers. There are some cases where as low as 4 to 5% voltage induction is
measured in the extreme limb. In such cases the test results measured from LV winding
should be considered for analysis. An example of measurements on HV winding of 100
MVA, 220/66/11 kV System Transformer is given in Table: 4
Table:4
Measured Voltage in Volts Total Voltage Magnetizing
induced in other Current
2 phases
1U-1N 1V-1N 1W-1N Phase
224.7 (100%) 214.4(95.4%) 8.80 (3.92%) 99.3 % U 2.85 mA
111.2 (49.7%) 223.6 (100%) 112.4(50.3%) 100.0 % V 2.23 mA
8.6(3.83%) 215.2(95.9%) 224.3(100%) 99.8 % W 2.95 mA
During commissioning stage some people perform this test from all windings and at all tap
positions. Practically there is no need to perform this test from different windings and at
different tap positions, as
when we apply voltage to the LV winding; every turn of each windings of the particular
phase gets proportionate voltage induced in it. Moreover due to higher voltage induction in
HV windings it is as good as performing this test at higher voltage level. If there is any
problem with any of the turns of any winding, it will reflect in the test results due to flux
choking in that particular limb.
Although it is advisable to perform this test from LV side there are some exceptions to this
concept, like in case of Furnace Transformers, Rectifier Transformers and Transformers with
higher voltage ratio like Generator Transformers. In such cases, if MBT is performed from
LV side then voltage induced in HV windings may exceed the safe limits due to higher
ratios. Thus, safety aspect becomes more crucial when this test is performed at
manufacturing stage or on transformer in un-tanked condition where HV line leads are in
open condition. Moreover, in LVs of Furnace and Rectifier Transformers, handling of very
high currents is also not possible.
Sometimes very surprising test results are experienced in case of MBT test from delta
connected tertiary windings in large rating autotransformers. In such cases the sum of
voltages induced in other two windings exceeds the voltage applied normally by 5% to 10%.
However, in some extreme cases this voltage even exceeds 30% to 35%. Experience shows
that these transformers, having satisfactorily passed all the performance tests, serving in the
field satisfactorily for several years. Hence this phenomenon is considered to be a normal in
case of autotransformers. Results of the measurements carried out from Tertiary winding of
315MVA, 400/220/33 kV, YNa0d11 connected autotransformer are given in Table: 5 as
example.
Table:5
Measured Voltage in Volts Total Voltage Magnetizing
induced in other 2 Current
phases
3U- 3W 3V-3U 3W-3V Phase
239.2(100%) 238.3(99.6%) 75.2(31.4%) 131.1% U 22.9 mA
124.0(52.1%) 238.2(100%) 114.8(48.2%) 100.2% V 15.7 mA
80.2(33.4%) 236.2(99.2%) 238.1(100%) 132.9 % W 23.0 mA
Above measurements are carried out on the autotransformer at final stage where Transformer
is in tanked condition with bushings mounted. In contrast to above the results performed on
same
transformer in un-tanked condition the sum of voltage induced in other two phases is found
normal but the magnetizing current is higher by about 30% to 40% as shown in Table:6
below.
Table:6
Measured Voltage in Volts Total Voltage induced Magnetizing
in other 2 phases Current
3U- 3W 3V-3U 3W-3V Phase
249.0(100%) 216.0(86.7%) 35.1(14.1%) 100.8 % U 36.1 mA
125.0(50.0%) 250.0(100%) 125.0(50.0%) 100.0% V 25.6 mA
37.2(14.9%) 215.0(86.0%) 250.0(100%) 100.8% W 36.7 mA
Table:7
Practically it is seen that, magnetizing current is more sensitive to the abnormal conditions or
faults than voltage. Hence it is always advisable to perform & analyze both the tests
simultaneously.
In normal conditions, magnetizing current measured in centre limb is comparatively lesser
than that of extreme limbs. This variation w.r.t. extreme limbs is of the order of 62 % to 80%.
While performing this test from the outermost winding (mainly HV of large Power
Transformers) this variation may be of the order of 82% to 88%. But this variation is quite
normal and is a result of lesser reluctance path experienced by flux in centre limb.
This phenomenon can be explained as under1.
The reluctance (S) of a magnetic path is directly proportional to its length (l) and inversely
proportional to its area (A). The reluctance is also inversely proportional to
the absolute permeability () of the magnetic material.
Thus, S = l / (*A)
The permeability of a magnetic medium is a measure of its ability to support magnetic flux
and it is equal to the ratio of flux density (B) to magnetizing force (H)
Thus, = B / H
,
Hence,
S = l / ((B/H) * A)
i.e.
Also,
m.m.f. (H) = (N * I) / l . (Where N = No of turns; I = Current; l = length of flux path)
Therefore,
l * ((N * I) / l))
S=
(B * A)
(N * I)
S= ------------ (2)
(B * A)
S l and S I
Hence I l
(i.e. Magnetizing current is directly proportional to the length of the magnetic reluctance
path)
Thus we can conclude that, due to asymmetrical core geometry, more reluctance path is
offered to the flux produced in extreme limbs which is mainly responsible for drawing higher
magnetizing current than centre limb.
This is to be noted that for delta connected winding also, magnetic section corresponding
to V phase required least magnetizing current, but the phasor addition of currents of two
phases results into a condition that current in V phase equals the current of one of the
extreme phases. This trend is observed at different excitations and confirmed that it
remains same at any excitation.
This test is widely used by Indian end users as a powerful diagnostic tool for checking
healthiness of transformer at site post fault occurrence events during operation.
Manufacturers also use this test to ascertain healthiness of windings at different
manufacturing stages. General understanding about this test is that, only inter-turn failures
are detected by this test but some case studies discussed below confirm that this test can
detect other faults like shorting of parallel conductors and presence of external metallic loop
around the core limb. This implies the need for proper study of results of MBT test along
with the complementary magnetizing current measurement tests.
Following are some case studies which prove this test as one of the powerful techniques to
locate the abnormalities in the transformer windings.
Case Study: 1
50 MVA, 220/33 kV, YNyn0 connected Power Transformer
At pre-connection stage of manufacturing, following test results were measured as given in
Table:10.
Table:10
Measured Voltage in Volts Total Voltage induced Magnetizing
in other 2 phases Current
2U-2N 2V-2N 2W-2N Phase
239.0 (100%) 4.6 (1.92%) 233.8 (87.8%) 99.4% U 23.9 mA
118.6 (49.8%) 238 (100%) 119.0 (50.0%) 99.8% V 2.49 Amp
233.4 (97.8%) 3.47 (1.45%) 238.3 (100%) 99.5% W 23.3 mA
Here voltage trend shows that there is very less voltage is induced in the limb V, i.e. very
poor flux linkage with the winding turns in the limb V. In other terms we can say that flux is
not able to enter the V limb and getting the path through W limb (as evident from voltage
trend). Magnetizing current trend shows high current in V limb, which indicates there is a
closed loop in the V limb which is acting as a localized load, drawing the load current from
the source.
The abnormally high magnetizing current drawn by the limb suggests that there is a inter turn
short circuit fault in the winding. To locate this fault, circulating current was checked on the
winding discs by clamp meter and gauss meter in which circulating
current was sensed in uppermost disc of HV winding. Based on these findings, visual
inspection of top disc was carried out in which it is found that main lead take off is having
sharp bend and this bend touched the 2nd disc. This further crushed its insulation during
pressing activity and resulted in to short circuiting of the turns.
Case Study: 2
Results given in Table:11 are of MBT test conducted from LV side of the transformer before
commencing the routine tests on the transformer
Table:11
Measured Voltage in Volts Total Voltage induced Magnetizing
in other 2 phases Current
2U-2N 2V-2N 2W-2N Phase
226 (100%) 196(86.7%) 30.4 (13.5%) 100.2% U 61.3 mA
97(42.7%) 227(100%) 130(57.3%) 100.0% V 39.8 mA
27(11.9%) 198(87.6%) 226(100%) 99.6% W 49.0 mA
Prior to this test, resistance measurement test was carried out at pre-tanking stage.
Considering this fact above test results could have been accepted. However, still to confirm
the healthiness, all other routine tests were performed on this transformer. While the
transformer withstood the dielectric tests like separate source & induced over voltage
withstand successfully, the no- load losses were found erratic and on higher side (almost 1.8
times the expected value). DGA in oil was conducted before & after routine tests, which
showed no abnormality.
To investigate it further, 1-phase loss measurement was carried out. Rated phase voltage
applied from LV side to individual phases in which power loss value of U limb was found
substantially higher as compared to other two limbs. With this it was concluded that there
might be an inter-strand shorting in the winding, forming a closed loop within the winding,
leading to circulating current with consequent increase in the losses in U limb. To locate the
fault the core coil assembly of the transformer was un-tanked. Crimping lug of the lead was
cut and discontinuity was checked between the winding strands, in which both parallel
strands of regulating winding were found shorted at ID side at one of the transpositions.
Case Study: 3
Test results given in Table:12 are the results when this test was conducted from Tertiary side
of the transformer at pre-connection stage. For this test temporary delta connection of
Tertiary winding was formed.
Table:12
Measured Voltage in Volts Total Voltage induced Magnetizing
in other 2 phases Current
3U-3W 3V-3U 3W-3V Phase
226.4(100%) 70(30.9%) 171.6(75.8%) 106.9 % U 50.8 mA
112.4(49.4%) 227.6(100%) 115.8(50.9%) 100.2% V 90.3 mA
171.8(75.3%) 69.6(30.6%) 228.3(100%) 105.7% W 50.8 mA
As evident in this case, there is no significant variation in absolute values of voltages &
magnetizing currents, but the trend of magnetizing current is abnormal.
These types of test results are tricky and confusing and need to be tackled carefully.
As magnetizing current in V limb was higher, to analyze it further, discontinuity was test
performed between individual strands of each winding of V limb. In this case, out of 15
parallel strands of CTC conductor, 2 strands in common winding were found shorted at inner
transposition due to scissoring action. This type of failure was detected at very early stage by
MBT test due to which heavy rework was avoided.
Similarly one more case of inter-strand shorting gave following test results tabulated in
Table:13. In which LV CTC coil of 100 MVA, 220/66-33/ 11 kV, YNyn0d11 connected
System transformer had a inter-strand failure (confirmed by discontinuity test) which was
detected at pre-connection stage.
Table:13
This phenomenon can be explained as under. Difference in the magnitude of leakage flux
linkage in different strands produces different voltages in the strands of the conductor. The
diagram below illustrates the phenomenon. Say, is the flux linkage with conductor at
position 3, /2 is with middle conductor at position 2 & 0 is of the conductor at position
3.
The alternating leakage flux linkage experienced by different strands being different,
different potential would be induced in these strands. In case of shorting of any of the strands
at different potentials due to insulation damage would nullify the effect of transposition and
cause flow of circulating current within the shorted loop of strands. This would eventually
reflect in to higher magnetizing current.
Further, when the transformer with this type of defect is charged at the rated voltage for no-
load loss measurements, we observe abnormally higher no-load loss as a result of the
circulating currents, which may almost twice the expected value of no-load losses. This is
already discussed in Case Study-.2
Case study : 4
2 x 8789 / 2 x 2 x 6214 KVA, 11 kV / 2 x 2 x 0.3986 kV, 3 Phase Rectifier transformer
Following are the test results when this test was conducted from HV side of the rectifier
transformer at post-connection stage. (Refer Table:14)
Table:14
Measured Voltage in Volts Total Voltage Magnetizing
induced in other Current
2 phases
1U-1V 1V-1W 1W-1U Phase
230(100%) 184(80.0%) 45.1(19.6%) 99.6% U 565 mA
33(14.4%) 229(100%) 201.6(88.03%) 102.4% V 117 mA
13(5.68%) 216.5(94.7%) 228.6(100%) 100.4% W 123 mA
Following are the test results when this test was conducted from Tertiary side of the
transformer at post-connection stage during manufacturing of the transformer. (Refer
Table:15)
Table:15
Measured Voltage in Volts Total Voltage induced Magnetizing
in other 2 phases Current
3U-3W 3V-3U 3W-3V Phase
213(100%) 200.4(94.1%) 12.4(5.8%) 99.9% U 140 mA
185.7(8.7%) 213(100%) 27.4(12.8%) 100.0% V 130 mA
54.3(25.5%) 159(74.5%) 213.3(100%) 100.0% W 750 mA
Test results indicated problem in the windings of W- phase. While checking the circulating
current in the shorting leads of HV Regulating winding (comprising 4 parallel sections)
circulating current was observed, which had resulted in higher magnetizing current in W-
phase. Parallel shorting leads were cut apart and MBT was repeated with satisfactory results.
Further, to diagnose the fault, ratio test between HV and individual step of Regulating
winding was performed in which one turn was found more than specified in one of the steps.
This extra one turn was causing the potential difference between the parallel coils, resulting
in circulating current. To rectify this problem this extra one turn was removed from the
winding and MBT was repeated after all connections which found normal.
Case study 6:
As discussed earlier, in single phase transformers, only magnetizing current can be measured,
hence it becomes essential to compare magnetizing currents of such transformers with similar
units.
While testing this transformer at pre-connection stage phase angle was found higher during
ratio measurement. Also magnetizing current was found 26mA from HV side as against 4mA
value measured in earlier units. To locate the fault, presence of circulating current in the
winding checked with the help of clamp on milli-ammeter, which indicated the circulating
current in bottom most disc. To rule out possibility of shorting of bottom Static End Ring
(SER), SER was physically checked, which was found healthy. All cross-overs in that zone
were checked for any damage of conductor paper insulation due to scissoring action and
these were also found intact.
Later, Ratio measurement was carried out on individual strand of the HV coil which is
interleaved disc winding. Test results indicated that out of two strands, one strand has less
number of turns compared to other, which is a clear indication of one interleaved joint made
wrongly (i.e. dummy turn brazed to dummy turn), thus forming a short circuited turn.
Physically this fault was located at the winding start and could be repaired by interchanging
the main and dummy conductor at the interleaved joint.
Case study 7:
167 MVA, 400/ 3 / 220/3 / 33 kV, 1 Phase Autotransformer
At pre-connection stage, high magnetizing current of the order of 10A was measured at 230V
when tested from Tertiary winding side as against 40mA measured earlier on similar units.
To confirm further with the help of clamp on type milli-ammeter, circulating current checked
through out the series winding axially. In series winding between disc no. 61 & 62 high
circulating current was observed. While inspecting those discs it is observed that while
making interleaved joint in the winding, unscrupulously the winder had brazed dummy
conductor to dummy conductor, which resulted in a short circuited turn. This turn was acting
as a localized load to the supply source and drawing high current.
Conclusion:
Magnetic Balance test when supplemented with measurement of magnetizing current, proves
to be a strong diagnostic test to detect Faults like
a) Inter-turn shorting;
b) Inter-strand shorting;
c) Any external loops around the core;
d) Abnormal magnetizing current due to unequal turns in winding sections connected in
parallel;
e) Wrong interleaving joints in windings.
To obtain accuracy in the test results, it is recommended to perform this test before resistance
measurement test or after proper demagnetization of the core as remnant DC component in
Core substantially affect test results of MBT.
Reference books:
Acknowledgements:
Author is thankful to the EMCO Management for granting permission to publish this paper.