Transformer and Circuit Breaker Monitoring in Substations
PATAKI, ANDRÁS M.Sc PhD Student
DÁN, ANDRÁS PhD, Associate professor
Department of Power Engineering
Budapest University of Technology and Economics
Budapest 1117 Egry József utca 18
HUNGARY
Abstract: - The paper deals with the introduction of the comparative method of the state estimation of transformers and
circuit breakers using continuous monitoring of voltage, current, oil-tank vibration and displacement of breaker
contacts if function of time. The paper gives details of the hardware and software of the measuring equipment
Key-Words: - Monitoring, Test on site, Vibration, Displacement, Harmonics, Special hardware
1 Introduction
2 Problem Formulation
The privatisation of the electric sector brought a profit
oriented thinking in the whole area of the electrical
energy bussiness. It means that not only the production
and delivery of electrical energy must be kept on the
lowest price, but the cost of maintenance must be
decreased as well.
In order to decrease the cost of maintenance, instead of
the periodically predefined planned maintenance, the
necessity of maintenance is introduced.
It means that e.g. a transformer- circuit breakerdisconnector unit of a substation will be in operation
until the weekest element needs the maintenance.
However this policy requires a sophisticated monitoring
system and a CAM (Computer Aided Maintenance
software) which is able to make decision on time of
necessity of maintenance. If the system works well, the
cost of maintenance can decrease while the quality of
supply remains on the desired level.
The paper is focusing on the transformer oil-tank
vibration and on the movement of circuit breaker
contacts. These data are suitable to develop a softwarebased comparison of the sequenced time functions- for
decision making.
There are known methods [1],[2] to measure transformer
oil-tank vibration, but without measuring the current
simutaneously with vibration, the decision can be false.
Below the suggested method will be discussed.
Regarding circuit breakers, the new type of breakers are
equipped optionally with monitoring. However the older
breakers need monitoring as well. Below a cheep and
rigid solution will be introduced, suitable to maintain the
old type breakers as well.
2.1 Transformer
Because power transformers are expensive machines,
thus the number of maintenances should be the possible
minimal over its lifetime. To reach this goal two main
parts must be examined (windings and core) by making
tests on the site. Vibration analysis extended with the
measurement of currents and voltages offers a robust
method to plan maintenance. The task is to find
correlation between current and vibration which
represents correctly the state of the transformer winding.
While vibration generated by the windings caused by the
electrodynamic forces, the magnetostriction effect[1],[3]
makes the core to vibrate. Separation of these sources
would be advantageous, but it is essential to find out if
they are correlated or not? This method hopefully gives
an important benefit, after an accurate state estimation
the transformer may be repaired on site if it is needed.
With more than one accelerometer, it is possible to
predict which winding should be repaired in case of
three phase transformer. In this case repairment on the
site is enough. To remove a transformer from a
substation, and to transport it for maintenance is a
complex and very expensive process and requires a lot of
time too.
2.2 Circuit breaker
Circuit breakers are very important elements in the
power transmission line so monitoring the events gives a
knowledgebase to determine time to the next
maintenance.
In contrast with transformers, the measurement of
vibration is not the appropriate method for state
estimation, instead of that displacement-time function of
the breaker contacts must be recorded.
3 Problem Solution
3.1 Transformer
Whilst in the previous section the problem was
described, this section gives the details and answers
regarding the minimum number of accelerometers, their
placement on the transformer tank, the role of the
harmonic currents and vibrations and hardware solution
for the measurement. This method compares previous
results carried out on the same transformer with the
actual ones. Since the initial vibration spectrum is
unknown for older transformers, the first series of
measurement serves as a reference for later tests.
Information is resulted partly from the appearance of
such harmonics which were not relevant in the previous
spectrum and from the relative changes of the spectrum.
Two main types of vibration tests are possible:
Switching measurement may be separated into
subtasks. (By using synchronous switch, two
measurements on the transformer have taken in different
times become comparable.)
 Characterising the vibration spectrum
generated by the core (Magnetostrictive
effect) can be decided after transients
having dumped.
In Fig. 1 a 400/120kV power transformer can be seen.
There are three points which indicate where to place the
accelerometers to capture the vibrations.
The vibration measurement requires a special equipment,
and a remote unit, because the vibration can be captured
only in the vicinity of the transformer, but currents and
voltages are measured in the control room. To solve this
problem a dedicated remote vibration measurement unit
was developed to the existing TR-16 equipment.
The TR-16 (Fig. 2) is a 16 channel portable Class A
power system analyzer[4]. It can be used with various
configurations of input signal conditioners for different
measurement tasks. It is capable to perform both short or
longterm monitoring functions or the evaluation of
voltages, currents, active and reactive powers,
impedances, voltage outages, flicker harmonics on 16
channels simultaneously
.

Localising a virtual vibration source for
the transformer is calculated from the
time delay of the accelerometer’s
transient signals.(There are more than
one accelerometer) Since not only the
vibration is recorded but currents and
voltages too, more accurate results are
expected.
The other type of test is the load current and
vibration measurement at high currents. In this case
estimation can be made regarding the state of the
winding.
Fig. 1 Marked points on the power transformer
Fig. 2 Picture of TR-16 equipment
There are often 200-300 meter distance (in a larger
substation) between the transformer where vibrations
are picked up, and the control room where the
secondary of the current and voltage transformers are
accessible. On the other hand voltage output
accelerometers can not be used with cables longer
than 100 meters, but charge output ones can be used
with just not longer than 10 meters of cables [1]. To
resolve this difficulty a special fiber based
communication line has been adopted to avoid
disturbances resulted by electromagnetic fields. This
hardware ensures the galvanic isolation from high
potential parts as well.
Fig. 3 Schematic diagram of TR-16 remote unit
In Fig. 3 the principle of the remote unit is shown. There
are four input channels in TR-16 to retrieve power
transformer-sensor-remote unit signals. There is also a
reverse directioned communication between the daisychained input cards and the remote unit. The master card
collects the range setting from the slaves and transmits to
the remote unit. The measurement range can be changed
by software before the measurement. Fig. 4 Shows the
to be taken into account at the hardware design phase.
An other requirement –equipment separation from high
potential parts- involves the problem.
In Fig. 5 there is the simplified schematic diagram of the
breaker monitoring equipment. Two light sources emit
10m light ray which are interrupted by the rotating
disc. Size of 10 mil width holes are manufactured to the
disk to chop the light. This pattern secures the
appropriate resolution of speed. Receivers sense square
pulse which carries the velocity. In this figure only one
phase is shown. It must be mentioned that at high
voltage breakers three set of sensor unit needed, but for
medium voltage breakers one set is enough.
The capture circuit may be connected to a PC to
Fig. 5 Schematic diagram of circuit breaker
realisation of an input card.
download the stored data, and with a special software
further analysis can be made. The capture circuit has the
responsibility to communicate with the SCADA system
if any kind of problem concerning the circuit breakers
would begin to develop. Breaker trigger signal are
connected into the capture circuit to act as a start signal
for the measurement Fig. 6 shows a solution for the
communication.
Fig. 4 Realisation of the input card
3.2 Circuit breaker
There are numerous methods to examine circuit
breakers, but now, just two from them being covered
here. A low current test and a high current one. The
given results show significant difference in the shape of
the displacement-time function of the breaker. While in
case of low current test, there is no swing in the
displacement-time diagram, in the high one a swing
might be observed which means that (because of
electrodynamic forces) the breaker contacts are slowing
down or reverses their movement direction for a short
time in an open-close sequence. This phenomenom has
Fig. 6 Capture unit PC communication
4 Measurement results
4.1 Transformer test results
In order to test the method mentioned above,
measurements have been made in Encs substation on a
single phase 2MVA, 26/0.7kV transformer supplying
thyristor switched static VAr compensation. The
compensation serves 3rd, 5th 7th harmonic filter as well.
Fig. 7 shows the time function of the primary current
(black line) in which more than 80A current changes can
be observed. These changes are sudden and rapid
according the railway’s demand. Below in Fig. 8 and
Fig. 9 the switching transient and the corresponding
vibration-time function is shown. There are three
relevant peaks in the vibration RMS time function,
which is correlated with the current of secondary side.
This means that the amplitude of vibration is mostly
determined by the electrodynamic force.
Fig. 10 Harmonics of u25
The next step is to find special patterns which may be
useful. First let us take a closer look to the time function
and FFT spectrum of vibration and current.
Fig. 7 Iszk Current RMS–time function
The switching measurement must be done by applying
the synchronous switch because the needed worst case
can be ensured. In this case just small transient resulted
since the prviously mentioned switch has missed.
Fig. 11 Time function of vibration
In Fig. 11 and Fig. 12 that is shown. It is interesting that
the most significant harmonic of the vibration is the
200Hz component. The even order harmonics must
originate from the core because the exciting current has
no such kind order of component (just odd order) as
shown in Fig. 13 and Fig. 14. Only u25 has 200Hz
component in the frequency specturm. (Fig. 10)
Fig. 8 vibration RMS-time function
Fig. 12 Harmonics of vibration
Fig. 9 Primary voltage RMS-time function
Fig. 13 Time function of iszk
Fig. 16 Graph vibration(3)-iszk(3)
Fig. 14 Harmonics of iszk
Just based on this information accurate decision is
impossible to make, therfore further signs must be find.
A possible solution is to draw the current harmonic in
the function of –vibration harmonic. The following
diagrams show interesting patterns. The practical
meaning of these patterns is not understood yet, but they
must be investigated further. One approach may be if the
correlation ratio changes test by test is should mean that
somekind of problem has beggun to develop.
Fig. 17 Graph vibration(5)-iszk(5)
4.2 Breaker test results
Fig. 15 Graph vibration(1)-iszk(2)
Each graph has is own special shape, thus one idea could
be the monitoring of the changes of this pattern. In Fig.
15 an almost straight line can be observed, but in Fig. 16
the shape is loop like.
In Fig. 18 an open-close sequence of a circuit breaker is
drawn. The closing time approximatly 57 msec while the
opening time is 49msec. This experiment carried out by
ICMET High Power Laboratory in Craiova. It can ce
seen that this trial was made at 50A which is relative low
current. At high currents if there is a swing in the
displacement-time function of the breaker contacts, that
predicts the blowing up of the breaker. Therefore if this
kind of swing begins to develop, the monitoring system
has to send an alarm signal to the SCADA system. That
is why it is necessary to make direction sesnsitive
hardware.
[4]
RST Ltd, Transanal-16 user’s manual,
Budapest, 2001
Fig. 18 Laboratory test of circuit
breaker at low current (making and
breaking)
5 Conclusion
5.1 Transformer
This paper tried to introduce such kind of monitoring
methods which are suitable to predict time to the next
mintenance both for power transformers and circuit
breakers. Hardware solutions were proposed too. It is
seen that state estimation is a very difficult and complex
task. For transformers not only the minimum number of
sensors must be determined, but the locations on the
transformer’s oil-tank where by placing them the most
informative results can be extracted. Measurement
results show an interesting correlation between current
and vibration harmonics that worth for further analysis.
In this state of the research is is seen that for transformer
monitoring a physical modell is essential to be able
interpreting the results. Furthermore a comparative
knowledgebase is also very important to recognize the
individual errors.
5.1 Circuit breaker
At circuit breakers the displacement- time function
represents the mechanical state. If a swing begins to
develop actions have to be made to avoid the blowing up
of the circuit breaker.
References:
[1]
Timár P. L., Fazekas A., Kiss L., Miklós A.,
S.J. Yang, Villamos gépek zaja és rezgése. Műszaki
Könyvkiadó, Budapest, 1968.
[2]
L. Tóth, T. Tóth, „Short-circuit tests on
power transformers with pressure measurements,”
Powertech 99, pp, Budapest
[3]
S. L. Foster, E. Reiplinger „Characteristics
and control of transformer sound,” IEEE
Transactions on Power Apparatus and Systems, Vol.
PAS-100, No.3, March 1981