U.S. patent number 3,684,951 [Application Number 05/096,343] was granted by the patent office on 1972-08-15 for methods of corona location utilizing three-phase voltage distribution patterns.
This patent grant is currently assigned to Westinghouse Electric Corporation. Invention is credited to Ronald T. Harrold.
United States Patent |
3,684,951 |
Harrold |
August 15, 1972 |
METHODS OF CORONA LOCATION UTILIZING THREE-PHASE VOLTAGE
DISTRIBUTION PATTERNS
Abstract
A method of locating a source of corona discharges in an
electrical winding, comprising the steps of stressing the winding
with a first voltage to provide a first voltage distribution
pattern to ground and noting the magnitude of the first voltage at
corona inception and corona extinction, and stressing the winding
with a second voltage to provide a second voltage distribution
pattern to ground, different than the first, and noting the
magnitude of the second voltage at corona inception and corona
extinction. The first and second voltage distribution patterns at
the noted first and second voltages, respectively, for either
corona inception or corona extinction, are then compared to
determine where in the winding the voltages are identical on the
two distribution curves.
Inventors: |
Harrold; Ronald T.
(Murrysville, PA) |
Assignee: |
Westinghouse Electric
Corporation (Pittsburgh, PA)
|
Family
ID: |
22256923 |
Appl.
No.: |
05/096,343 |
Filed: |
December 9, 1970 |
Current U.S.
Class: |
324/536; 324/523;
324/509; 324/547 |
Current CPC
Class: |
G01R
31/62 (20200101); G01R 31/1227 (20130101) |
Current International
Class: |
G01R
31/06 (20060101); G01R 31/02 (20060101); G01r
031/06 (); G01r 031/08 () |
Field of
Search: |
;324/51,52,54,55 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
ganger et al., Ionization Measurements on Transformers Offprint
from the Brown Boveri Review 1967, No. 7, pp. 3-15..
|
Primary Examiner: Strecker; Gerard R.
Claims
I claim as my invention:
1. A method of locating a source of corona discharges in a
three-phase delta connected electrical winding having first, second
and third terminals, comprising the steps of:
stressing the three-phase winding with an applied voltage to
provide a first voltage distribution pattern in which the winding
is stressed uniformly to ground,
changing the magnitude of the applied voltage in a predetermined
direction,
detecting corona discharges in the winding, noting the phase in
which they occur, and noting the magnitude of the applied voltage
when the corona condition of the winding changes,
stressing the three-phase winding with an induced voltage to
provide a second voltage distribution pattern, different than the
first voltage distribution pattern, which equally and uniformly
stresses the phases of the winding turn-to-turn,
changing the magnitude of the induced voltage in the same direction
as the applied voltage,
detecting corona discharges in the winding and noting the magnitude
of the induced voltage when the corona condition of the winding
changes,
and comparing the first and second voltage distribution patterns at
the noted magnitudes of the applied and induced voltages,
respectively, determining where in the noted phase winding the
voltages are of like magnitude in the first and second voltage
distribution patterns.
2. The method of claim 1 wherein the first and second voltages are
changed to increase their magnitude, an the corona condition of the
winding at which the magnitudes of the first and second voltages
are noted is the inception of corona in the windings.
3. The method of claim 1 wherein the first and second voltages are
changed to decrease their magnitude, and the corona condition of
the winding at which the magnitudes of the first and second
voltages are noted is the magnitudes at corona extinction.
4. The method of claim 1 wherein the steps are repeated a plurality
of times and the magnitude of the first and second voltages at
inception are averaged, to set the first and second voltage
distribution patterns which are compared to locate the source of
corona.
5. A method of locating a source of corona discharges in a
three-phase delta connected winding having first, second and third
terminals, comprising the steps of successively testing the winding
with a different terminal grounded during each test, with each test
including the steps of inducing a voltage into the winding,
changing the voltage magnitude in a predetermined direction, and
noting the magnitude of the winding voltage at which the corona
condition of the winding changes, locating the phase in which the
corona discharge is occurring and the location of the discharge
within the phase by comparing the voltage distribution pattern of
each phase when a terminal of the selected phase is grounded, with
the voltage distribution pattern of the phase when its terminals
are not grounded,
and determining in which phase, and where in that phase, the
voltages of the distribution patterns are of like magnitude.
6. The method of claim 5 wherein the first and second voltages are
changed to increase their magnitude, and the corona condition of
the winding at which the magnitudes of the first and second
voltages are noted is the voltages at corona inception.
7. The method of claim 5 wherein the first and second voltages are
changed to decrease their magnitude, and the corona condition of
the winding at which the magnitudes of the first and second voltage
are noted is corona extinction.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates in general to testing electrical inductive
apparatus, such as electrical power transformers, and more
specifically to locating sources of corona discharges within such
apparatus.
2. Description of the Prior Art
Certain types of high voltage electrical apparatus, such as liquid
filled power transformers, are tested after manufacturing and prior
to shipment to non-destructively overstress the electrical
insulation between the phases, between the conductor turns of the
phases, and from the phase windings to ground, to a predetermined
factor of safety. If the electrical apparatus passes these tests,
it assures that the insulation will perform as expected at the
intended operating voltage of the apparatus.
If corona or partial discharges are detected during these tests, it
is important to measure the amount of energy in the discharges and
to locate their source.
Partial discharges are a rapid discharge of electrical energy which
ionizes, deteriorates, and may eventually break down the
surrounding insulation through the heat produced by the electrical
discharge, and resulting chemical changes in the organic
insulation. There are many locations in high voltage electrical
apparatus, such as power transformers, where a corona discharge may
initiate, some of which may cause rapid deterioration of the
surrounding insulation, and others which may cause little or no
damage, even over a long period of time. Thus, it is important to
test high voltage electrical inductive apparatus for the presence
of corona discharges, during the normal test procedures for such
apparatus, but it is not sufficient to simply determine if corona
discharges are present in the apparatus. The specific location of
the corona discharge source in the inductive apparatus should also
be determined, to enable design engineering personnel to evaluate
the type and severity of the partial discharge, to determine the
corrections which may be necessary in the apparatus under test, and
also to provide information useful in developing future insulating
structures.
If it is necessary to drain the oil from the electrical apparatus,
remove the winding from the tank, and unstack the coils, in order
to repair the defect causing the corona, it is very important to
have the corona source pinpointed before disassembly of the
apparatus, as it is usually difficult to visually observe the
corona site after disassembly of the apparatus. Therefore, accurate
methods of locating the site or source of corona discharges are
continually sought.
The methods conventionally used to locate corona within electrical
windings are the sonic methods and the electrical methods. In the
sonic methods, mechanical to electrical transducers are placed on
the sidewalls of the electrical apparatus, and they detect
mechanical disturbances within the apparatus resulting from corona
discharges. The location of the corona source is narrowed to a
specific location within the apparatus by triangulation. In the
electrical methods, the corona is measured at each end of the
winding, and corona located by the amount of attenuation of the
corona discharge signals as measured at the winding ends. In other
words, if the signals are of equal magnitude at each end of the
winding, it is assumed that the corona source is at the midpoint of
the winding.
With both the sonic and the electrical methods, sensitivity is a
problem. For example, the sonic methods are useful for locating
corona sources having magnitudes in the range of about 500 to 1,000
.mu. V, and up, and the electrical methods for sources greater than
100 .mu. V. Further, the electrical methods require measurements at
opposite terminals of a winding, and often the signal will be
measurable at only one terminal of the winding, which then rules
out the use of the electrical method.
It would be desirable to be able to locate sources of corona
discharges having energy contents down to the corona free level,
defined as less than 10 pC, which for most power transformers will
be approximately 5 .mu. V, the "ambient" reading for most radio
noise meters used in corona testing.
SUMMARY OF THE INVENTION
Briefly, the present invention is new and improved methods of
locating the site of corona discharges in electrical windings,
which method is sensitive enough to locate any corona site that can
be detected. Depending upon the type of corona source, either the
corona inception voltage at the corona site, or the corona
extinction voltage at the corona site, with respect to ground, will
be repeatable on successive tests. The invention uses at least two
winding tests, which tests are selected to stress the winding with
different voltage distribution patterns with respect to ground. The
voltage distribution patterns for each test, using the test voltage
magnitudes at which corona inception, or extinction, was detected,
are then examined to determine where in the winding the two
different patterns have the same voltage to ground, which pinpoints
the corona source. The two different voltage distribution patterns
may be obtained by the applied potential and induced potential
tests, which are required by the USAS transformer standards for
delta connected transformer windings, or by the proposed new
induced tests for transformers in which different phases of a delta
connected winding are grounded in sequence. Thus, the disclosed
methods may utilize the results of the tests currently conducted on
high voltage electrical inductive apparatus, making it unnecessary
to add additional tests, or testing apparatus and procedures, and
making it unnecessary to provide special training for test
personnel. Further, the tests do not require a shielded room, and
it is not necessary to make absolute measurements of the amount of
energy in the corona discharges, as it is only necessary to
accurately measure the winding test voltage when corona inception
and corona extinction is noted.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention may be better understood and further advantages and
uses thereof more readily apparent, when considered in view of the
following detailed description of exemplary embodiments, taken with
the accompanying drawings, in which:
FIG. 1 is a partially schematic and partially block diagram,
illustrating an applied potential test on an electrical power
transformer, which test is a step in a method of locating a corona
source according to a first embodiment of the invention;
FIG. 2 is a partially schematic and partially block diagram,
illustrating an induced potential test on the electrical power
transformer shown in FIG. 1, which test is another step in the
method of locating a corona source according to the the first
embodiment of the invention;
FIGS. 3 and 4 are vector diagrams of the electrical power
transformer shown in FIG. 2, for the induced potential test;
FIG. 5 is a graph illustrating the phase voltage distribution with
respect to ground for the induced potential test shown in FIG. 2
for the test voltage magnitude at which corona extinction was
observed;
FIG. 6 is a vector diagram illustrating the location of a corona
source external to the high voltage winding of the transformer
shown in FIGS. 1 and 2;
FIG. 7 is a vector diagram illustrating the location of a corona
source between the high and low voltage windings of the transformer
shown in FIGS. 1 and 2;
FIGS. 8A, 8B and 8C are partially schematic and partially block
diagrams illustrating the steps of another method of locating a
corona source, according to the teachings of the invention, using
only induced potential tests;
FIGS. 9A, 9B and 9C are graphs which illustrate the voltage
distribution to ground for each phase during the test shown in FIG.
8A, at the winding voltage at which corona extinction occurred;
FIGS. 10A, 10B and 10C are graphs which illustrate the voltage
distribution to ground for each phase during the test shown in FIG.
8C at the test voltage at which corona extinction occurred;
FIG. 11 is a graph which compares the two different voltage
distributions across one of the phase windings shown in FIG. 8A,
taken from the graphs in FIGS. 9A and 10A, illustrating that the
corona source is not in that specific phase; and
FIG. 12 is a graph comparing the two different voltage
distributions across another of the phases, taken from the graphs
shown in FIGS. 9C and 10C, and illustrating that the corona source
is in this phase, and the specific location of the corona source
within the phase winding.
DESCRIPTION OF PREFERRED EMBODIMENTS
The present invention discloses new and improved methods of
locating corona sources, which methods may utilize the presently
used applied potential and induced potential tests, or the proposed
induced potential tests in which different winding terminals of a
delta connected winding are successively grounded.
The new and improved methods are based on the realization that
either the corona inception voltage at the corona site, or the
corona extinction voltage at the corona site, will be identical
from test to test, with certain types of corona sources producing
corona extinction voltages at the site which are of like magnitude
from test to test, and other types of corona sources providing
corona inception voltages at the corona site of like magnitude from
test to test. For example, corona producing voids located between
pressboard barriers in oil, produce corona extinction voltages, at
the corona site, which are consistently the same from test to
test.
If the winding test voltage at which corona extinction is noted is
not consistently repeatable from test to test, using the same type
of test, the corona inception values may be used, and when the
corona inception values of the test voltage are not substantially
similar from test to test, using similar tests, accurate results
have been obtained by taking the average of the noted test voltages
at which corona inception occurred for four or five similar
tests.
Broadly, the invention utilizes the repeatability of the corona
extinction or inception voltages, at the corona site, along with
the steps of stressing the winding to be tested to provide two
different voltage distribution patterns to ground across the
winding, with the distribution patterns being developed by using
the specific test voltage for each distribution pattern at which
corona inception was noted, or corona extinction, as hereinbefore
stated. Then, the two voltage distribution patterns are compared to
determine where in the winding the voltage is of like magnitude in
the two patterns, which will accurately pinpoint the corona source.
The comparison of the voltage distribution patterns may be
accomplished graphically, or vectorially, as desired.
The new and improved methods of locating corona sources may be more
easily understood by considering specific examples, with a first
embodiment of the invention developing the required information for
locating corona sources according to the teachings of the
invention, by using the standard applied potential and induced
potential tests, now used to test delta connected power
transformers. The applied and induced potential tests produce
different voltage distribution patterns across the winding tested,
with respect to ground, and thus these patterns may be used to
provide the at least two different voltage distribution patterns
required by the method.
Referring now to the drawings, and FIG. 1 in particular, there is
illustrated a partially schematic and partially block diagram of a
step in a first embodiment of the invention, illustrating the
applied potential test, which uniformly stresses the insulation of
the three phases of a power transformer 10 to ground. Power
transformer 10 includes a tank or casing 12 containing three-phase
high and low voltage windings 14 and 16, respectively, which are
disposed in inductive relation with a magnetic core (not shown) and
immersed in a suitable fluid insulating and cooling dielectric,
such as mineral oil, askarel, or SF.sub.6 gas. The low voltage
winding 16 includes phase windings 18, 20 and 22, connected in wye,
in this example, with the outwardly extending ends of the phase
windings 18, 20 and 22 being connected to the encased ends of low
voltage bushings 24, 26 and 28, respectively, and the common or
neutral terminal of the wye connection is grounded at 30. The high
voltage winding 14 includes phase windings 32, 34 and 36, connected
in delta, in this example, with the line terminal 39, to which
phase windings 32 and 34 are connected, being connected to the
encased end of high voltage bushing 38, line terminal 37, to which
phase windings 34 and 36 are connected, is connected to the encased
end of high voltage bushing 40, and line terminal 41, to which
phase windings 32 and 36 are connected, is connected to the encased
end of high voltage bushing 42. High voltage bushings 38, 40 and 42
commonly have bushing taps 46, 48 and 50, respectively which are
capacitively coupled to the main conductor studs of the high
voltage bushings, and corona discharges may be conveniently
detected at these bushing taps.
While the first embodiment of the invention is most applicable to
delta connected high voltage windings, it is to be understood that
it may also be applied to any type of winding, as long as care is
taken to insure that the voltages developed within the winding
during test do not exceed the maximum allowable voltage for all
parts thereof.
In the applied potential test, shown in FIG. 1, a source 52 of low
frequency alternating potential, commonly 60 or 180 hz., is applied
to the winding terminals 38, 40 and 42 which voltage stresses all
of the phases uniformly to ground. The source voltage 52 is
increased in magnitude, toward a predetermined maximum test
voltage, which depends upon the BIL voltage rating of the
apparatus. The winding 14 is monitored for partial discharges or
corona, during this test. While the method may be used with any
means which indicates corona inception and extinction, it is
preferable to use the most sensitive corona detecting means
available, such as a radio noise meter, which reads the radio
frequency energy detected in .mu.V, or a picocoulomb meter, which
measures electrical charge per cycle. A pulse counter may be used
in conjunction with these devices, if desired.
The corona may be detected by connecting the corona detecting
means, indicated by reference numeral 54 in FIG. 1, to the bushing
tap of any of the high voltage bushings 38, 40 and 42, such as
indicated by the solid line 56 between bushing tap 46 and corona
indicating means 54. Suitable isolating chokes, for shorting line
frequency voltage from the bushing tap to ground, and a tuning
capacitor connected from the bushing tap to ground, may be used in
conjunction with the corona indicating means, in a manner well
known in the art. If the bushing taps are not provided on the high
voltage bushings, any other suitable means for coupling the corona
detecting means 54 to the winding may be used, such as by using an
auxiliary capacitive pickup with the NEMA test circuit.
If corona discharges are detected as the magnitude of source 52 is
gradually increased towards the maximum test voltage, the magnitude
of source 52 at corona inception is noted, and the reading of the
corona indicating means, such as .mu. V or picocoulombs, is noted.
The magnitude of source voltage 52 is then gradually reduced until
corona extinction occurs, with the magnitude of source 52 at corona
extinction being noted.
The next step of the new and improved method of corona location is
shown in FIG. 2, which provides a different voltage distribution
pattern across the phase windings of the transformer 10, with
respect to ground. This different voltage distribution pattern is
obtained by using the induced potential test, in which a
three-phase low frequency source 62 of alternating potential is
applied to the outwardly extending end of the low voltage bushings
24, 26 and 28, and the stressing voltage for the high voltage
winding 14 is induced therein. The source voltage is gradually
increased towards a predetermined maximum test voltage, and the
bushing taps are monitored for corona discharges at each voltage
step, either by switching from tap to tap with a single indicating
means, or by using three separate indicating means. When corona
discharges are detected by corona detecting means 54, before
reaching the maximum test voltage, the phase voltage of the high
voltage winding 14 at which corona starts is noted. The source
voltage is then gradually reduced until corona extinction occurs,
and the phase voltage of the high voltage winding 14 is again
noted. The corona readings at the three bushing taps will usually
indicate by their relative magnitudes, in microvolts (.mu.V), or
picocoulombs (pC), in which phase the corona source is located, and
often in which half of the phase it is located.
In the induced potential test, the turn-to-turn insulation
throughout each phase is equally stressed, but the voltage to
ground will vary throughout each phase winding, from a maximum at
the ends of the windings to a minimum at the winding center. The
exact voltage to ground at different positions along each phase may
be determined by considering an "effective ground" for the high
voltage winding 14, for the transformer as energized under no-load
conditions and with the windings have substantially equal ground
capacitance. If the windings do not have equal capacitance to
ground, the ground, for purposes of a vector diagram, may be found
by measuring the voltage to ground at two terminals.
FIG. 3 is a vector diagram of the transformer 10 shown in FIGS. 1
and 2, with vectors L1, L2, and L3 illustrating the primary or low
voltage phase windings 18, 20 and 22, respectively, and vectors H1,
H2 and H3 illustrating the secondary or high voltage phase windings
34, 36 and 32, respectively. The effective ground for winding 14 is
indicated at 64. Thus, the high voltage and low voltage vectors,
with respect to ground, for transformer 10, are as shown in FIG.
4.
From the vector diagrams shown in FIGS. 3 and 4, the actual phase
voltage distribution, with respect to ground, for one high voltage
phase winding and one low voltage phase winding, may be determined,
for any phase voltage at which corona extinction or inception
occurs. For example, the maximum voltage to ground of a selected
phase, such as vector H2', would be equal to one-half of vector H3
divided by sine 60.degree., and the minimum voltage, indicated by
dotted line 66, would be equal to one-half of vector H3 divided by
tan 60.degree.. As an example, the phase voltage distribution with
respect to ground, during an induced test of a 115 kv delta
connected winding having 64 discs or winding sections per phase, in
which 185 kv is the maximum test voltage for both the applied and
induced tests, the insulation would be uniformly stressed at 185 kv
to ground for the applied test, while for the induced test the
maximum voltage to ground would be 107 kv i.e., 185 divided by 2
sine 60.degree., at the ends of the winding, and 53.5kv i.e., 185
divided by 2 tan 60.degree., at the center of the winding, as shown
in FIG. 5. FIG. 5 is a graph which plots kilovolts to ground versus
the number of discs or winding sections for the winding of this
example, with curve 70 illustrating the voltage to ground across
the high voltage winding during the induced test. Curve 72
illustrates the voltage to ground for the low voltage phase
winding, which is zero at its grounded end and 13.2 kv at its other
end.
Since the voltage to ground at which corona inception or extinction
occurs in the applied test is the magnitude of the applied voltage
at that time, and since the voltage to ground pattern may be
determined for a phase winding during the induced test, at the
corona inception or corona extinction voltage, the location on the
induced voltage pattern which is the same as the noted applied
voltage, will locate the corona source.
For example, assume that with the applied test corona extinction
consistently occurred at a winding voltage of 69 kv, that corona
extinction occurred at 161 kv phase voltage during the induced
potential test, and that from the .mu.V or picocoulomb readings at
the bushing taps it was determined that the corona source is in
phase H3, and in the half of phase H3 which is directly connected
to phase H2. Because the corona was consistently extinguished at 69
kv during the applied test, it can safely be assumed that the
corona was extinguished at 69 kv during the induced test. Thus, it
is necessary to determine where in the winding a stress of 69 kv
occurs when the phase voltage is 161 kv during the induced test.
One location for the stress will be outside the high voltage
winding, and another location, if the transformer has
concentrically adjacent high and low voltage windings, will be
between the high voltage and low voltage windings. These locations
may be determined graphically, or vectorially, as desired.
The vectors for a corona source external to the high voltage
winding are shown in FIG. 6. The minimum voltage is 46.5 kv for a
phase voltage of 161 kv, and the maximum voltage is 93 kv to
ground. The angle .PHI., and the vector y at which corona
extinction occurs is 69 kv. The angle .PHI. may be found by cosine
.PHI. = 45.5/69 , or about 48.degree.. The angle .theta. is thus
about 12.degree. degrees and the voltage vector b from the corona
source to the end of the winding is equal to 2y sine .theta. or
29.8 kv. Since there are 64 discs in the winding, the location of
the corona source may be found by multiplying the ratio of 29.8/161
times the number of discs 64, which locates the corona
approximately 12 discs from end 37 of the phase winding H3.
To determine where the 69 kv stress occurs between concentrically
adjacent high and low voltage windings, the vector diagram shown in
FIG. 7 is used, which illustrates that the primary voltage to
ground (9 kv in this example) is vectored with the 69 kv, to
provide a voltage of 22.6 kv from the end 37 of the phase winding,
which when multiplying the ratio of 22.6/161 by the number of discs
64 indicates that the corona source is in the ninth disc from the
end 37 of the phase winding. Thus, the search for the corona source
has been narrowed to the location between the ninth and twelfth
discs from end 37 of the phase winding.
The method of noting the test voltage applied to a winding at
corona inception and corona extinction, on successive tests, and
using either the inception or extinction voltages to find the
corona source, may also be used with the winding voltage
distributions, rather than voltage vectors. This approach will
become especially important if the USAS transformer standards are
changed. As hereinbefore stated, the present USAS standards require
that a delta connected transformer must be low frequency tested by
means of both applied and induced potentials. Since the applied
test stresses all parts of the winding to the full test voltage,
and since in practice under the worst fault condition the voltage
at the midpoint of the winding can reach only about 80 percent of
the voltage at the line end, the winding must be insulated to
withstand a test which will never be duplicated in operating
practice. Thus, a new test procedure is under consideration which
would allow insulation added to the midpoint of the winding, merely
to pass the applied test, to be eliminated. This proposed new test
is an induced test, which duplicates the worst possible operating
condition during the test by successively grounding the three line
terminals of a delta connected high voltage winding.
FIGS. 8A, 8B and 8C are schematic diagrams which illustrate the
proposed new testing method, with FIG. 8A illustrating a
transformer 80 having a primary or low voltage winding 82, and a
secondary or high voltage winding 84. Low voltage winding 82 has
phase windings 86, 88 and 90 connected in wye, with the outwardly
extending ends 92, 94 and 96 of phase windings 86, 88 and 90 being
connected to a source 100 of alternating potential, such as 60 or
180 hz., and their commonly connected ends are grounded at 98.
High voltage winding 84 has phase windings 102, 104 and 106
connected in delta, having line terminals 108, 110 and 112. The
high voltage winding 84 is tested by successively grounding the
line terminals 108, 112 and 110 as illustrated in FIGS. 8A, 8B and
8C, respectively, increasing the induced voltage in winding 84 to
the maximum test voltage.
In this embodiment of the invention, the bushing taps associated
with the ungrounded terminals are monitored by corona indicating
means 116, i.e., such as a radio noise meter or a picocoulomb
meter, on two successive tests, indicated by solid line 118 and
dotted line 120, or simultaneously by using two similar corona
indicating means. The source voltage 100 is gradually increased,
and if corona is detected, the phase voltage of winding 84 is
noted, and then the voltage is gradually reduced until corona
extinction occurs, at which time the phase voltage of winding 84 is
again noted. The voltage distribution to ground, across each phase
winding, at the corona inception or corona extinction voltage,
whichever is selected, as hereinbefore explained relative to the
first embodiment of the invention, is plotted on a graph for each
test in which corona discharges are detected, for the three tests
shown in FIGS. 8A, 8B and 8C, and then the voltage distribution
curves for like phases compared. The intersection of two curves for
a phase indicates the location of a corona site within the
phase.
For purpose of example, assume that a corona site is located within
phase winding 102 at a site which is 20 percent of the winding
length from line terminal 112. The first test, shown in FIG. 8A,
will detect corona discharges at the bushing tap for terminal 112,
and it will be assumed that the phase voltage to ground at corona
extinction was measured at 100 kv. Since corona discharges were
detected during this first test, shown in FIG. 8A, the voltage
distribution across each phase of winding 84 is plotted, using the
100 kv corona extinction voltage as the terminal voltages of the
winding, as illustrated in FIGS. 9A, 9B and 9C. FIG. 9A plots the
voltage to ground across the length of phase winding 106, which
will be 100 kv at each end of the winding and 87 kv at the center
of the winding. FIG. 9B plots the voltage across phase winding 104,
which is ground or zero voltage at one end and 100 kv at the other.
FIG. 9C plots the voltage across phase winding 102, which is ground
or zero voltage at one end and 100 kv at the other end.
The ground 114 is then removed from terminal 108, a ground 114' is
applied to terminal 112, and the second test, shown in FIG. 8B, is
performed, while detecting corona at the bushing taps for terminals
108 and 110. In this test, no corona discharges are detected at
either terminal, even at the maximum test voltage of 185 kv to
ground, since the corona source is near terminal 112, which is
grounded, and the voltage at the corona site thus never reaches the
corona threshold level. Since no corona discharges are detected in
test 2, it will not be necessary to plot voltage distribution
patterns for test 2, and it should now be apparent that the corona
source is closer to terminal 112 than it is to terminals 108 or
110, and that the site must be located in phase winding 102 or in
phase winding 106.
The ground 114' is then removed from terminal 112, a ground 114" is
applied to terminal 110, and the third test, shown in FIG. 8C is
performed while detecting corona discharges at the bushing taps for
terminals 108 and 112. During this test, corona is detected at the
bushing taps for both terminals 108 and 112, and corona extinction
was measured at a terminal voltage of 88 kv to ground. Since corona
discharges were detected during this third test, the phase voltage
distribution to ground was plotted for each phase, with FIGS. 10A,
10B and 10C being graphs which illustrate the voltage magnitudes to
ground across phase windings 106, 104 and 102, respectively. FIG.
10A plots the voltage to ground across the length of phase winding
106, which is ground at one end and 88 kv at the other end, FIG.
10B plots the voltage to ground across phase 104, which is also
ground at one end and 88 kv at the other and FIG. 10C plots the
voltage to ground across phase winding 102, which is 88 kv at each
end of the winding and 76.5 kv at the center thereof.
The next steps are to plot the voltage distribution pattern to
ground for phase 106, during each of the tests in which corona was
detected, with the curves being plotted on the same graph, and also
to plot the voltage distribution to ground for phase 102 for each
of the tests in which corona was detected, with these curves being
plotted on the same graph. The voltage distribution patterns are
only plotted for these phases, since it was determined that they
are the only phases in which the corona site could be located.
FIG. 11 is a graph which combines the two voltage distribution
curves shown in FIGS. 9A and 10A, and since the two curves do not
intersect, it is obvious that the corona site is not in phase
106.
FIG. 12 is a graph which combines the two voltage distribution
curves shown in FIGS. 9C and 10C. These curves intersect at point
P, and since the extinction voltage at the corona site is identical
on successive tests, the corona location can only be at the
intersection of these two voltage distribution curves. Therefore,
from the three tests performed, illustrated in FIGS. 8A, 8B and 8C,
it is definitely established that a corona site is located in phase
102, 20 percent of the winding length from terminal 112.
In summary, there has been disclosed new and improved corona
locating methods for electrical windings which locate corona by
developing at least two different voltage distribution patterns to
ground for the winding under test, with the patterns being
established for the winding voltage at corona inception, or corona
extinction, and with the corona being located by determining where
identical voltages exist in the two voltage distribution patterns.
The disclosed methods are extremely sensitive, locating any corona
source which can be detected. The methods do not require a shielded
room or special tests, as the standardized tests for high voltage
electrical inductive apparatus provide all of the information
required to locate corona sites, according to the teachings of the
invention. Absolute measurements of corona discharge energy are not
required. All that is necessary is to measure the test voltage at
which corona inception and corona extinction occurs. Relative
values of the corona energy will indicate which phase, and which
half of the phase, the corona site is located. Still further, while
the methods have been described relative to locating a single
corona source, it will be apparent that more than one source of
corona discharge may be located, if they have different inception
or extinction voltages. While the two embodiments of the invention
have been described as alternative methods of locating a corona
source, they may both be used on a single transformer, if desired,
if more than two voltage distribution patterns are desired to
increase the accuracy of the method.
* * * * *