Notice of Violation of IEEE Publication Principles

"On-line Insulation Diagnostic System and Off-line PD Monitoring with HVAC Testing," by Yong-Sung Choi; Ju-Ho Yun; Kyung-Sup Lee, in the Proceedings of the 2008 International Conference on Condition Monitoring and Diagnosis, 21-24 April 2008, pp 85 - 88 After careful and considered review of the content and authorship of this paper by a duly constituted expert committee, this paper has been found to be in violation of IEEE's Publication Principles. This paper is a near duplication of the original text from the papers cited below. The original text was copied without attribution and without permission. Due to the nature of this violation, reasonable effort should be made to remove all past references to this paper, and future references should be made to the following articles: "What to Do After the Warning of an On-line Insulation Diagnostic System? Off-line PD Monitoring with HVAC Testing!" by Kubler, B.; Hauschild, W, in the Proceedings of the 2004 Power Systems Conference and Exposition, 10-13 Oct. 2004, vol. 2, pp. 822 - 828

Yong-Sung Choi1*, Ju-Ho Yun1 and Kyung-Sup Lee1

1Department of Electrical Engineering, Dongshin University, 252 DaeHo-dong, Naju 520-714, Korea *E-mail : yschoi67@dsu.ac.kr

On-line Insulation Diagnostic System and Off-line PD Monitoring with HVAC Testing

�Abstract--The paper considers the relation between on-line

monitoring and diagnostics on the one hand and high-voltage (HV) withstand and partial discharge (PD) on-site testing on the other. HV testing supplies the basic data (fingerprints) for diagnostics. In case of warnings by on-line diagnostic systems, off-line withstand and PD testing delivers the best possible information about defects and enables the classification of the risk. Because alternating voltage (AC) is the most important test voltage, the AC generation on site is considered. Frequency tuned resonant (ACRF) test systems are best adapted to on-site conditions. They can be simply combined with PD measuring equipment. The available ACRF test systems and their application to electric power equipment –from cable systems to power transformers – is described.

Index Terms--Cable, Frequency tuned resonant, High voltage withstand, Off-line PD monitoring, On-line insulation diagnostic,

I. INTRODUCTION

onitoring and diagnostics of power apparatus and systems are applied for cost reduction by avoiding of

defects during service (with consequences up to blackout), application of condition based maintenance, extension of lifetime, etc. This introduction shall demonstrate that diagnostics of electric insulation is not an isolated method, but only successful considering interactions with high-voltage (HV) testing and previous test results.

The electric breakdown is a weak point phenomenon caused by a defect in the insulation. The defect might be the result of a production failure (to be detected by the routine test in factory) or a too high stress during transportation or an assembling fault (to be detected by the on-site test). But also the normal aging process of the whole insulation causes degradation as a volume phenomenon and only finally a weak point leading to breakdown. The described physical breakdown process enables the following classification of HV tests and measurements:

Whether a certain defect is dangerous or not at a fixed test voltage level can only be decided by a HV withstand test which is therefore characterized as a ”direct” test. This direct test needs no interpretation, the result interprets itself directly. Each measurement of a different quantity and the conclusion for critical defects and breakdown requires a more or less physically or technically based knowledge rule. Therefore measurements deliver always an “indirect” test result of higher uncertainty than a direct test. Last but not least partial

This work was finally supported by MOCIE program (I-2006-0-092-01).

discharge (PD) measurements which measure also a weak point phenomena deliver a less uncertain result related to critical defects than dielectric measurands (e. g. tan�) which are volume phenomena. The latter give only a rough indication if the aging process could lead to a critical defect (weak point) or not.

On-line monitoring has the big advantage that it indicates certain measurands of apparatus without interruption of service. The measurands might be related to volume phenomena or to weak point phenomena (e. g. PD). But in any case, they are indirect results requiring interpretation by knowledge rules. Such knowledge rules are based on basic research, but should also be related to results of previous type, routine and on-site tests. A remarkable improvement of the conclusions is reached if in case of warning by the on-line diagnostic system an off-line withstand test combined with PD measurement (this means a combination of a direct test with an indirect measurement) is added. Such an off-line, onsite test enables PD measurements in a wide range of voltages, higher and lower than the operation voltage. This means on-line diagnostics and HV testing on site must not be considered as separate or even contrary methods. They complete each other in an excellent way.

The optimum interpretation of on-line monitoring and offline PD test results requires not only a knowledge rule but also information of all previous data from the test object. Therefore Fig. 1 shows the schematic life cycle of power apparatus with special respect to tests and test results. Modern information technology enables to establish the life cycle record from all relevant data semi-automatically or – for important and expensive equipment – even automatically. It should be mentioned that it will be helpful if the different tests are performed under comparable conditions, e. g. related to the test voltage shape. The life cycle data deliver the important information about the trend of the measurands and are the basis for a decision after the warning respectively after the subsequent on-site test.

It is good practice for nearly 100 years and a basic principle of insulation coordination [1] that the test voltage shall represent stresses in service. The horizontal standards for HV testing in test laboratories [2], [3] consider this principle by defining power frequency (45 to 65 Hz) lightning and switching impulse voltage (applicable for AC equipment). With respect to weak point detection by withstand voltage testing and PD measurement, voltages which represent power frequency are by far the most important. The application of

M

2008 International Conference on Condition Monitoring and Diagnosis, Beijing, China, April 21-24, 2008

978-1-4244-1622-6/08/$25.00 ©2007 IEEE

direct or very low-frequency (VLF) voltage (Fig. 2) to AC insulation causes resistive instead of capacitance voltage distributions. Consequently the PD measuring results are much less representative than those measured at an AC voltage which is representative for power frequency. IEC Standards (e. g. [4], [5]) allow for AC on-site testing wider frequency ranges. This is considered by a frequency range 10 Hz to 500 Hz in the latest draft for the horizontal IEC Standard for HV testing on site [6]. Using AC voltages of a certain frequency range of that overall range will deliver test results in close relation to the stresses in service. Therefore AC voltage generation on site will be considered more in detail.

Fig. 1. Life cycle and test / monitoring data.

Fig. 2. VLF voltage in comparison to 50 Hz voltage and voltages 20 to 300 Hz used for cable testing [7].

II. HVAC TEST SYSTEMS FOR WITHSTAND AND TESTING AND

PD MONITORING

A. Frequency-tuned resonant test systems Most test objects are capacitances and therefore HVAC test

systems based on series resonant circuits have tremendous advantages compared with conventional test transformer. A resonant circuit is an oscillating circuit with the natural frequency

(1)

which is excited with exactly that frequency. Because the capacitance CT of the test object is fixed, there a two ways to

reach resonance - by tuning the inductance LH in such a way that fr becomes equal to the frequency of the supply power (50/60 Hz): inductance-tuned (ACRL) system (Fig. 3 (a));

- by the power supply via a frequency converter with exactly fr: frequency-tuned (ACRF) system (Fig. 3 (b)).

(a) Inductance-tuned (ACRL) system

(b) Frequency-tuned (ACRF) system Fig. 3. Series resonant circuits of variable inductance (a: ACRL, 50 Hz) and variable frequency (b: ACRF, 20 to 300 Hz).

The ACRL system operates at power frequency whereas the ACRL system needs a wider frequency range (e. g. 20 to 300 Hz). This certain disadvantage is more than compensated by the following series of advantages: Because of lower frequency at maximum load (20 Hz), ACRF systems have a higher equivalent test power at identical current. A reactor of tuneable inductance (ACRL) has higher losses than one of fixed inductance (ACRF), consequently the quality factor of an ACRF system is about twice of that of an ACRL system. The necessary feeding power is remarkable lower and can be supplied from a three phase system. Also the weight-to-power ration, which is an indication for the transportability, is about three to five-times better for ACRF systems than for ACRL systems. This is completed by an ACRF load range up to 10 times larger than that of ACRL system. The high robustness of ACRF systems comes from components without moving parts, whereas ACRL systems contain a reactor with a moving core and a conventional regulator transformer. B. PD measurement at variable frequency

The static frequency converter (Fig. 4) (FC) of the ACRF system generates four switching pulses (of few 10 microseconds each) which may influence the PD measurement. In modern PD measuring devices (PD) they can be suppressed by triggered windowing (tw). But in many cases they can also

remain on a PD pattern (Fig. 5, D) because they are simply identified as noise signals.

Fig. 4. ACRF test circuit for PD measurement.

Fig. 5. Signal-to-noise relation for PD measurement at a frequency different from power frequency.

PD measurement at variable frequency is not different from the well-known at power frequency. With respect to onsite measurements, there is an advantage of PD measurement at frequencies different from power frequency, because the noise signals are triggered to the latter. Fig. 5 shows the onsite measurement of a 400-kV-voltage transformer at 81 Hz in an open-air substation. Whereas the PD signals are in a stable phase position of the 81 Hz voltage (characterized by the red colour (C) which means high PD pulse rate), the up to tentimes higher noise signals (E) are related to 50 Hz, but deliver a “grey background” of low density. Even such a high noise cannot prevent a successful PD pattern recognition.

III. ON-SITE, OFF-LINE PD MONITORING TESTING

Quite often withstand testing of power equipment in service is not applied, because utilities are afraid that it could damage or even destroy the equipment. But when the test is performed with a well selected AC voltage amplitude (in relation to the standardized test voltage level for routine tests) and combined with a sensitive PD measurement, then this fear is not justified and the advantage of a direct test becomes evident. Because testing means classification into equipment safe for service, equipment which must stay under observation and equipment critical for service. If the HVAC withstand test

results in a breakdown the equipment would have failed in service soon. The whole test should be performed as a step test (Fig. 6 shows an example) and all the time be observed by PD measurement. In addition to the selection of the withstand voltage level (Uw = n . Uo; usually 1,3 < n < 2) a voltage level shall be chosen below that PD should not exist (UPD = mUo; e. g. m = 1.1 ...1.2).

Fig. 6. Voltage-time characteristic of a combined PD and withstand test.

On-site testing of machines must be connected with PD and tan��measurement. Whereas the PD behaviour is similar in a sufficiently wide frequency range [8], the dissipation factor depends on frequency by definition (power loss/capacitive apparent power). But more important than tan��itself is its increase with voltage (Fig. 10) which is nearly independent of frequency. Therefore ACRF test systems completed with PD and tan��measuring equipment are efficient tools for generator testing on site [8].

The first application of ACRF test systems has been the on-site testing of GIS starting about 20 years ago [9]. The relevant standard [5] defines a very wide range of acceptable frequencies (10 ... 300 Hz), but mainly a range 50 ... 300 Hz is used with the advantage that for a higher test frequency the voltage transformers can be included into the GIS test. Meanwhile ACRF on-site testing of GIS is common practice [10], [11].

IV. CONCLUSIONS

With respect to on-line monitoring and diagnostics, off-line testing and PD measurement is not unnecessary, but an important completion in case of a warning by the on-line diagnostic system. Off-line withstand testing in combination with PD measurement enables a classification of defects in connection with the life cycle record.

It seems necessary to say that on-line monitoring and diagnostics have to take the experience of one century of HV insulation testing into account, especially when standards are under preparations. Vice versa, also the standards of HV testing should be modified with respect to later monitoring, because HV testing shall supply the necessary basic data (fingerprints) for diagnostics.

Fig. 7. Tan� measurement of 137 MVA generator at 50, 42 and 26 Hz.

V. REFERENCES

[1] IEC 60071:1993: Insulation coordination. Part 1: Definitions, principles and rules.

[2] IEC 60060-1:1989: High voltage test techniques – Part 1: General definitions and test requirements.

[3] IEEE Standard 4 – 1995: Techniques for high-voltage testing. [4] IEC 60840:1999 and 62067:2002: Cables with extruded insulation and

their accessories for rated voltages 30 to 150 kV respectively 150 up to 500 kV.

[5] IEC 62271-203:2003: HV switchgear and controlgear. Part 203: Gas insulated, metal-enclosed switchgear for rated voltages above 52 kV.

[6] Draft IEC 60060-3: High voltage test techniques – Part 3: Definitions and requirements for on-site tests.

[7] W. Hauschild; W. Schufft; R. Plath et. al: The technique of AC on site testing of HV cables by frequency-tuned resonant test systems CIGRE Sessions (2002), Report 33-304.

[8] J. R. Weidner; P. Gradinarov; W. Hauschild; P. Coors: Diagnostic Testing of stator coils of large rotating machines by ACRF test systems (in German). ETG-Conference on Diagnostics of Electric Power Equipment, 2004, pp. 191-196

[9] W. Zaengl, et. al.: Experience of AC tests with variable frequency using light-weight, on-site series-resonance devices. CIGRE Session 1982, Paris, Report 23-07.

[10] U. Schichler: On-site testing of GIS, HIS and GIL (in German). HIGHVOLT Kolloquium ’03, pp. 243-248.

[11] R. Pietsch; P. Coors; W. Hauschild: Advantages and applications of frequency-tuned resonant test systems for HV testing of GIS and their components. 12th ISH Bangalore (2001).