Bushing Test Errors

1350

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 21, NO. 3, JULY 2006

Analysis of Some Measurement Issues in Bushing Power Factor Tests in the Field S. Zhang, Member, IEEE

Abstract—The power factor measurement is widely used in the evaluation of bushing insulation conditions in the field. Some measurement issues occur in the field tests which give abnormal readings of bushing 1 and/or 2 power factors. This paper provides an engineering analysis of the possible measurement issues in the bushing field tests such as testing a bushing on a transformer or in its shipping crate, or a bushing with wet/contaminated voltage/test tap or wet/dirty porcelain surface. Index Terms—Bushing, 1 , 2 , crate, porcelain, power factor, test tap, transformer, voltage tap. Fig. 1. Illustration of bushing insulation. (a) Equivalent circuit. (b) Phasor diagram.

I. INTRODUCTION

T

HE measurement of power factor is widely used in the evaluation of bushing insulation conditions in the field. The insulation of a bushing can be illustrated with a capacitor and a resistor in parallel (Fig. 1). With the voltage applied to and resistive current the bushing, both capacitive current flow through the bushing insulation. The bushing power factor is the cosine of the angle , by which the total current leads the applied voltage . The magniis the ratio of the magnitude tude of bushing power factor of the resistive current to that of the total current (1) The of a bushing is very small and is very close to 90 . For example, an oil-impregnated paper-insulated bushing , which gives . typically has The measurement of is very sensitive to the small changes of the bushing insulation as well as the testing environment. The leakage currents could flow into the test circuit and cause measurement issues. The “guard circuit” of some general portable measurement equipment is used to have unwanted currents bypass the measurement circuit. The measurement errors could occur if testers do not follow certain instructions when testing bushings on transformers. Some conditions such as a wet/contaminated bushing voltage tap or test tap, or wet/dirty porcelain surface could cause measurement issues. Testing a bushing in its shipping crate could also give abnormal readings. This paper analyzes the measurement issues that occurred in these cases. For simplicity, we use capacitors to represent and of bushing insulation in our discussions below. Manuscript received June 20, 2005; revised December 1, 2005. Paper no. TPWRD-00347-2005. The author is with PCORE Electric Company, Inc., LeRoy, NY 14482 USA (e-mail: [email protected]). Digital Object Identifier 10.1109/TPWRD.2006.874616

Fig. 2. Testing C of bushing H1 on the transformer using the UST method (two phases of a three-phase transformer or a single-phase transformer are shown).

II. TESTING BUSHINGS ON TRANSFORMERS It is well known that we must follow the following instrucinsulation on a transformer using tions when testing bushing the ungrounded specimen test (UST) method (Fig. 2) [1].

0885-8977/$20.00 © 2006 IEEE

ZHANG: ANALYSIS OF MEASUREMENT ISSUES IN BUSHING POWER FACTOR TESTS

1351

• Terminals of transformer windings at the same voltage as the bushing under test should be connected together to the equipment voltage source (e.g., H1 and H2 in Fig. 2). • Terminals of windings at other voltages should be grounded (e.g., X1 and X2 in Fig. 2). • All of the covers of bushing voltage tap or test tap should be closed except for the bushing under test. We see from time to time that testers do not follow these instructions and the test results turn out to be wrong. The logic behind these recommendations is not well explained in any published document. With the inductance of the transformer windings and the stray capacitance between the transformer windings and the bushing foils, leakage currents flow into the test circuit and could cause measurement errors during bushing tests. The following sections present the engineering analysis of how these steps minimize the effect of the inductance of transformer windings when followed correctly. A. Stray Capacitances Inside Transformer As most transformers are core-form transformers, we use core-form transformers for discussion in this paper. Usually, the high-voltage (HV) coil and the low-voltage (LV) coil in one phase of a core-form transformer are wound concentrically around the grounded core inside the transformer (Fig. 2). Since the windings are physically very close to each other and very tall, the capacitance between the HV and LV coils,1 , and the capacitance between the LV coil and grounded core could be very large compared to bushing capacitances. Depending on the actual dimensions, the capacitance between could be very large the HV coil and the grounded tank wall too. The capacitance-graded bushings have conductive foils in the bushing core extended below the grounded flange. Therefore, a exists between the bushing foils and transstray capacitance former windings. When the bushing is tested on a transformer, the applied voltage will be also impressed on the windings. Then, there will be a leakage current flowing from the trans, and into the former winding, through the stray capacitance bushing core (Fig. 2). Since this leakage current flows directly into the bushing core ( insulation) rather than the guard circuit, the guard circuit cannot eliminate it from the test circuit. Therefore, it could cause measurement errors. B. Terminals of Transformer Windings at the Same Voltage as the Bushing Under Test Should be Connected Together to the Applied Voltage During Bushing Tests As discussed before, when we apply a voltage to the bushing for test, the stray capacitance carries a leakage current from the winding into the bushing core and, hence, into the test circuit. If the terminals of windings at the same voltage as the bushing under test are not connected together during bushing tests, the leakage current could cause measurement errors. We will discuss the single-phase transformer first and then multiphase transformers next. 1It is a U.S. convention to use the letter “X” rather than “L” to represent transformer LV windings and terminals.

Fig. 3. Testing bushing H1 without connecting winding terminals together causes measurement error (shown in two terminals of one winding in a single-phase transformer). (a) Leakage current flows into the measurement. (b) Equivalent circuit (C of bushing H1 consists of C and C in series). (c) Phasor diagram and power factor change.

1) On a Single-Phase Transformer: Fig. 3 shows the case of testing bushing H1 on a single-phase transformer without connecting terminals of the windings at the same voltage as the bushing under test. For simplicity, we only show the two terminals of one winding and put the effect of the other windings (as well as the bushings connected to those windings) into the capacitors and . In the equivalent circuit, the bushing is represented with two capacitors and in series, and the winding inductance is represented with two inductors and in series. The points separating and winding inductance are the two terminals of the lumped stray capacitance .

1352

The transformer windings have resistance. However, for simplicity, we use inductors to represent the windings, and in the phasor diagram, the current through a winding lags the applied voltage across it by nearly 90 . The currents through and will flow into the ground (which is the same connection as the guard circuit in the UST test) and cannot influence the measurement. Since the reactance is much smaller than that of the of the winding inductance , capacitances between windings and bushings we can use a simplified circuit in Fig. 3(b) for analysis, where is the total of , and . Since is much smaller than , the current is much smaller than . Therefore, and have almost the same phase and magnitude. Since leads by nearly leads by nearly 90 leads by 90 and nearly 180 . Because the reactance of is much smaller than that of is much smaller than . Since the combination and is the applied voltage and have of almost the same phase and magnitude. Hence, both and lead by nearly 90 . The difference between them is the leakage current . The actual current through the bushing insulation is (Fig. 1). Without leakage current , the measured current . However, flows into the test circuit and makes the measured different from the actual current , which causes current the measurement error.2 The phasor diagram shows that the measured is smaller than the actual ; hence, the measured power factor is higher than the actual value of the bushing insulation. changes the It should be noted that the leakage current measured values of both resistive and capacitive currents. Even though is mainly capacitive, it affects the measurement of resistive current (Fig. 1) more than it does to the capacitive cur(Fig. 1) because . Therefore, the testers would rent notice the abnormal resistive current (or power loss) during the bushing power factor test. Depending on the value of the stray capacitance and the leakage path, the effect of the leakage current might also be so insignificant that the testers do not notice its effect. This leakage current cannot be eliminated using the method, such as the guard circuit, as was discussed before. However, we can connect two terminals of the winding together to the applied voltage to minimize its effect. Fig. 4 shows that when we connect the two terminals of the winding together, we actually make the currents through the , and flow into the guard capacitors circuit (ground) and these currents cannot influence the measurement. Therefore, we can simplify the equivalent circuit in Fig. 4(b), where the effective inductance is actually the resultant inductance of and in parallel.

2Strictly speaking, current i changes after the introduction of the leakage current i , and is no longer equal to current i. However, we assume that i does not change here to simplify the diagram, and this assumption does not affect the analysis results.


Fig. 4. Connecting terminals together minimizes the effect of the leakage current and makes the measured power factor closer to the actual value of the bushing insulation. (a) Connecting terminals to minimize the effect of the leakage current. (b) Equivalent circuit (C of bushing H1 consists of C and C in series). (c) Power factor is closer to the actual value.

is smaller than , and the reactance of smaller than that of . Therefore, the branch of

-

is much in the


simplified circuit is capacitive and its reactance is almost the . Hence, the leakage current flows only same as that of through the capacitance and it leads the voltage as well as the total applied voltage by nearly 90 . For comparison, we put the phasor diagrams of both before and after connecting the winding terminals together in Fig. 4(c). It is clear that by connecting the two terminals of the winding together to the applied voltage, the measured is much closer to the actual than not connecting the terminals; hence, the measured power factor is much closer to the actual value of the bushing. Therefore, connecting the terminals of the winding together to the applied voltage is to make the leakage current more capacitive. As the leakage current is much smaller than the capacitive current through the bushing insulation, the effect of this leakage current is minimized. 2) On a Multiphase Transformer: When we test a bushing on a three-phase transformer with -windings connected to the bushing under test, each winding has two terminals connected to two phases as well as two bushings. Each phase of the -windings is similar to a single-phase transformer. Therefore, the single-phase transformer case applies to this case. When testing a bushing on a three-phase transformer with Y-grounded windings connected to the bushing under test, the neutral point should be ungrounded during the bushing tests. The reason is that by connecting the three phases together to the applied voltage during the bushing test, the applied voltage is zero-sequence voltage since all three phases have voltages of the same polarity. With the neutral point still grounded, in case the secondary winding has a good zero-sequence path (such as the -winding), large zero-sequence current will flow through the circuit. The limiting factor of this zero sequence will be the short-circuit impedance of the measuring equipment and the ground resistance. Therefore, this current will be considerably higher than the maximum current that the measuring device can provide. As we disconnect the neural point from the ground and connect it to the applied voltage, such as the phase terminals of the windings, the situation is similar to the case of a single-phase transformer. When testing bushings on a three-phase transformer with ungrounded Y windings, the analysis is similar to the condition on a single-phase transformer. By connecting the three-phase terminals of the windings together, the effect is similar to that which we discussed for the single-phase transformer. C. Terminals of Transformer Windings at Different Voltages from the Bushing Under Test Should be Grounded During the Tests When testing a bushing on a transformer, the terminal of transformer windings at different voltages from the bushing under test should be grounded. Otherwise, when we apply the voltage to bushings as well as windings, these ungrounded terminals will have floating potentials. As shown in Fig. 5, without grounding X1–X2 terminals, there is some voltage on X1–X2 during the tests. It certainly is unsafe if touched by testers.

1353

Fig. 5. Not grounding X1-X2 when testing bushing H1 will apply voltage on X1 and X2 and is an unsafe condition to testers.

Fig. 6. Leakage current through the wet tap increases the C power factor.

D. Covers of Voltage Tap or Test Tap Should be Closed for Bushings Not Under Test The covers of voltage tap or test tap of bushings not under test should be closed during the tests. Energizing the bushing under test will also energize some other bushings not under test; hence, leaving the cover open could leave very high voltage on ) and damage the bushing insulation. the bushing tap (and It is also unsafe to testers. III. EFFECT OF WET/CONTAMINATED TAP If the voltage tap or test tap of a bushing is wet and/or contaminated, a leakage current will flow from the tap pin, through the wet/contaminated tap, and to the grounded bushing flange during the bushing test. This would affect the test results. The leakage current is resistive, and the leakage path is in parallel with the bushing . Fig. 6 shows the effect of the wet/contaminated tap on bushing and tests. In the UST test for bushing , the leakage current actually flows into the guard cirpower factor cuit (ground). Therefore, the measured value of is not affected by the tap condition.

1354


Fig. 7. Salty water is sprayed onto the tap to simulate the wet and contaminated tap condition. (a) Wet bushing tap. (b) Bushing in the test. TABLE I EFFECT OF WET AND CONTAMINATED TAP ON BUSHING TEST Fig. 8. Effect of wet/dirty porcelain on the bushing UST test for (a) Equivalent circuit. (b) Phasor diagram.

C

.

IV. EFFECT OF WET/DIRTY PORCELAIN SURFACE

In the grounded specimen test (GST) for bushing , the resistive leakage current flows into the measurement circuit. power factor is much Therefore, the measured value of the higher than the actual value. We soaked the tap insulator in boiling water, put salt on the tap insulator, and sprayed the salty water onto the tap insulator to simulate the condition of wet tap (Fig. 7). The test results afterwards on several bushings are shown in Table I. Due to the tap condition, the tester may not be able to get any test because the protection of the test readings during the equipment could be triggered by the high leakage current. It should be pointed out that the condition of the wet/contaminated tap may occur together with a wet/dirty porcelain surface which will be addressed in the next section. Usually, the wet/contaminated tap is caused by some moisture on the outside surface of the tap. This condition may be corrected by cleaning and drying the tap before any tests in order to have the normal power factor. When using the tap adaptor in reading of the the bushing test, it is very important to keep the tap adaptor from being contaminated.

If the bushing porcelain surface is dirty and wet with contamination and moisture, leakage currents will flow on the porcelain surface. If a bushing is tested in this condition, measurement issues could occur. Fig. 8 shows the effect of the wet/dirty porcelain surface on test using the UST method. The contamination the bushing is usually evenly distributed on the porcelain surface, which has an effect of two equivalent resistors in parallel with bushing and , respectively; and, hence, the two corresponding leakage is so small that we currents and . The voltage drop across during the can ignore which is the actual current crossing UST test. The leakage current flows into the test circuit and increases the measured power factor while flows away from the test circuit and decreases the measured power factor. , the measured is smaller than the Therefore, when actual which gives a higher value of bushing power factor ; when , the measured is bigger than the actual , which gives a lower value of bushing power factor . In the case that and the difference between them is greater than the actual resistive current crossing bushing , the , which gives both negative power factor and measured negative resistive current (power loss). with a wet/dirty porcelain surface, When we test bushing the analysis is shown in Fig. 9 and is similar to that of bushing . The two equivalent leakage currents and , as well as the resistive current crossing bushing , will together alter the power factor. Therefore, we may expect measured bushing power factor to be higher than the actual value, the measured lower than the actual value, or even negative. To see the effect of the wet/dirty porcelain surface, we sprayed salty water on bushing porcelains and performed the bushing tests. The test results are shown in Table II. The test results show that the spray condition is not repeatable and the results are random from test to test. However, the wet/dirty porcelain surface does have a significant effect on the


Fig. 9. Effect of wet/dirty porcelain on the bushing GST test for (a) Equivalent circuit. (b) Phasor diagram.

C

.

TABLE II EFFECT OF WET/DIRTY PORCELAIN SURFACE ON BUSHING TESTS

test results and could cause the measurement issues on both and power factors. Therefore, it strongly suggests that we should clean and dry the porcelain surfaces before the bushing tests. A similar situation was discussed regarding testing bushings with resistance-graded (RG) porcelain [2]. As we pointed out before, the condition of a wet/dirty porcelain surface may occur together with a wet/contaminated tap. Therefore, the tap should also be cleaned and dried. V. TESTING BUSHING IN CRATE Testers may incorrectly test a bushing in its shipping wood crate. The resistance of the wood is not as high as that of bushing porcelain. Therefore, there could be some leakage current flowing through the wood during the bushing test. Howand are different. ever, the effects of wood on bushing

1355

Fig. 10.

Effect of crate on the bushing UST test for C .

Fig. 11.

Effect of crate on the bushing GST test for C .

If we test bushing in the crate using the UST method, the equivalent circuit and diagram are shown in Fig. 10. The crate wood has contacts with the bushing porcelains and flange, which make a path for leakage current from part of the bushing through the crate wood and then to the grounded flange. As the is very small, the leakage current is voltage drop across formed from the voltage across the leakage resistor and is in phase with . This resistive leakage current will turn the curfrom . Since rent away from and, hence, the voltage the test equipment only measures the angle between the total applied voltage and the current flowing into the test circuit , the measured angle is greater than the actual , which reduces the power factor . Further analmeasured value of bushing current is also reduced which makes the ysis shows that the capacitance lower than the actual value. measured value of However, during the GST test for bushing , the leakage current flows into the measurement circuit (Fig. 11). Therefore, has a phase shift from the actual curthe measured current rent and its magnitude is greater than . Therefore, both the measured power factor and capacitance are higher than the actual values of the bushing insulation. We tested a 115-kV bushing (POC550G0800S) in its crate (Fig. 12), and the test results are shown in Table III. The bushing was stocked outside in the crate, and experienced a scattered shower and sunshine afterwards but before the tests. The crate appeared dry. A bushing is usually shipped with plastic covers. The test shows that if the testers test the bushing in the crate with the plastic bags still on the bushing, the test results will be far different from the actual values of the bushing. A similar result was reported when a capacitor was hung with a crane sling which gave a negative power factor in the UST test [3]. It should be pointed out that the effect of the crate depends largely on the condition of the wood. The testers may not see significant differences if they test the bushings in the crate. In

1356


2)

3)

4) Fig. 12.

Experimental test of the bushing in crate.

TABLE III EXPERIMENTAL TESTS OF THE BUSHING IN CRATE

5)

windings at the different voltages from the bushing under test are not grounded, there will be some potential on some bushings as well as transformer terminals, which is unsafe to the testers. When testing a bushing on a transformer, all of the covers of bushing voltage/test taps should be closed except for the bushing under test. Otherwise, these bushings could be damaged and it is also unsafe to the testers. The wet and/or contaminated tap makes the bushing power factor higher than the actual value but has no effect power factor. The on the test results of the bushing tap should be cleaned and dried in this situation; the tap adaptor, if used in the tests, should have no contamination. In most cases, the issue of wet/contaminated tap occurs together with the wet/dirty porcelain surface. The wet and/or dirty porcelain surface can cause random and power factors measurement issues on both during bushing tests. Sometimes the test values are even negative. We should clean and dry the porcelain surface before bushing tests. power Testing bushing in crate could give a lower power factor than the actual values factor but a higher of the bushing insulation. A bushing should not be tested in its shipping crate. ACKNOWLEDGMENT

The author would like to thank the valuable discussions with M. Foster of Doble Engineering Company during the revision of this paper. some cases, even if the bushing has no contact with the wood crate, the stray capacitance could be large enough to cause measurement errors.3 Therefore, the bushing should not be tested lying on the crate. VI. CONCLUSION We draw the following conclusions from the discussions above. 1) When testing a bushing on a transformer without connecting terminals of transformer windings at the same voltage as the bushing under test to the applied test voltage, the leakage current between the transformer winding and bushing insulation could cause measurement errors. The testers could see the abnormal power loss in this situation. If the terminals of transformer 3Personal communication

with Mary Foster of Doble Engineering Company.

REFERENCES [1] Type M2H Instruction Manual—Electrical Insulation Testing, Watertown, MA, 1988. Doble Eng. Co. [2] D. Zeng, “An improved method of measuring C power factor of resistance-graded bushings,” IEEE Trans. Power Del., vol. 14, no. 2, pp. 437–442, Apr. 1999. [3] L. Pong, “Review negative power factor test results and case studies—Analysis and interpretation,” presented at the Int. Conf. Doble Clients, Boston, MA, Mar. 2002.

Shibao Zhang (S’99–M’02) received the B.S. degree in electrical engineering and the M.S. degree in high voltage engineering from Tsinghua University, Beijing, China, in 1994 and 1997, respectively, and the Ph.D. degree in electrical engineering from Kansas State University, Manhattan, in 2001. Currently, he is a Senior Engineer with PCORE Electric Company, Inc., LeRoy, NY. His responsibilities include bushing design, development, test, and repair. He was an Electrical Engineer with Beijing Xuji Electric Company, Beijing, China, for one year.

Bushing Test Errors

Recommend Documents