ALLOWABLE LIMITS TO SHORT TIME VOLTAGE DROPS IN HV WITHSTAND TESTS

Transcription

1 The 19 th International Symposium on High Voltage Engineering, Pilsen, Czech Republic, August, 23 28, 2015 ALLOWABLE LIMITS TO SHORT TIME VOLTAGE DROPS IN HV WITHSTAND TESTS William Larzelere Evergreen High Voltage, NY, USA * Abstract: The current versions of IEC60060 part 1 [1] and part 2 [2] give incomplete definitions about the allowable voltage drop during a HV test and give limited requirements for measuring systems necessary for tests where non-disruptive discharges occur. Currently, there is vague information in IEC60060 Part 1 as to how long the specified maximum voltage drop can persist before the test is considered a failure. IEC60060 Part 2 contains some information about the frequency response of measuring systems for AC and DC voltages but does not give guidance on the characteristics of measuring systems needed to record fast, non-disruptive discharges that may be more than 10% for DC and 20% for AC. In this paper, we would like to stimulate a discussion and encourage researchers to work on this problem to form better definitions for voltage drop. This paper also attempts to encourage researchers to study this phenomena to better define criteria for characterizing the response of measuring and recording systems used for tests where short-term, non-disruptive voltage drops are expected. By establishing better definitions and values for these parameters, researchers will be able to help improve the definitions contained in the next IEC revisions and also the requirements for digital measurement systems for recording these events as covered by the proposed new standards IEC61083 Parts 3 and 4. This information can also lead us to improved characterization of the source requirements for AC and DC HV tests. 1 INTRODUCTION The current IEC standard IEC60060 parts 1 and 2 for HV Testing and Measurement include some details about the allowable voltage drop due to the interaction of the test object, the test circuit and the test source and the measurement system. The measurement system requirements defined are for normal dry or wet test situations but do not cover the specific demands of artificial pollution tests whose procedures are covered by IEC Mention is made of faster bandwidth systems being required for pollution tests but no requirements other than the maximum voltage drop are given. During the last revisions of IEC60060 Parts 1 and 2 it was found that there was some controversy about the definitions of test source suitability and questions about whether these current definitions could assure users of the ability to achieve the goal of performing repeatable tests in different test facilities. This paper attempts to review the current definitions of allowable voltage drop for AC and DC testing, the characteristics of measuring systems, and to make recommendations for future work for standard revisions. Of particular interest is to consider how new digital recorders for measuring AC and DC test voltages should be characterized. It should be noted that these problems do not just appear during normal dry or wet testing but also are of particular concern during Artificial Pollution testing and Combined and Composite Testing [4] 2 REVIEW OF IEC AND -2 ON VOLTAGE DROP FOR NORMAL WITHSTAND TESTS The current versions of IEC and 2 give recommendations for the generation and measurement of AC and DC voltages. The following excerpts are from the latest revisions. 2.1 DC VOLTAGE REQUIREMENTS From IEC60060 Part Paragraph 5 Tests with direct voltage Paragraph voltage drop instantaneous reduction of the test voltage for a short duration of up to a few seconds NOTE: Voltage drop may be caused by nondisruptive discharges. Paragraph Tolerances For test durations not exceeding 60 s, the measured values of the test voltage shall be maintained within ±1 % of the specified level throughout the test. For test durations exceeding 60 s, the measured value of the test voltage shall be maintained within ±3 % of the specified level throughout the test. The source characteristics should be sufficient to allow charging of the capacitance of the test object

2 in a reasonably short time. In the case of wet or pollution tests, the source, including its storage capacitance, should also be adequate to supply the transient discharge currents of the test object with voltage drop of < 10 %. 2.2 SUMMARY OF DC REQUIREMENTS The DC Voltage can drop no more than 10%, instantaneously and cannot exceed a 10% drop for more than a few seconds. These statements are in contradiction. What is "a few seconds"? Repeated how often? What are the requirements of the measurement system to be sure we measure an "instantaneous" event? What rise time? Bandwidth? 2.3 COMMENTS ON DC REQUIREMENTS During DC tests it is often possible to have streamers, or non-disruptive discharges from the HV electrodes and connections or even the test object due to non-uniform electrical fields. These non-disruptive discharges are very common when testing air-insulated devices with positive polarity at voltages above 400kV-500kV, since the electric field required to produce streamers in air compared to negative polarity is reduced. [5] [6][7] If the voltage of a DC test circuit has been raised to the correct test voltage without persistent discharges from anywhere in the test circuit, then normally the voltage will be held at the test voltage and timed for the test duration required. Intermittent discharges can take place in a short period of time and can be repeated often. After a discharge, the voltage recovers to the preset level at a rate that depends on the load capacitance and the voltage regulation characteristics of the DC source. If a voltage drop of 10% is allowed, then a single, short duration of <10us discharge could be recharged in approximately 1 line period considering the following characteristics of a typical 1000kV DC generator [see Figure 1]. This means that non-disruptive discharges can occur often and may not be noticed on control desk panel meters with normal display time constants that are much longer than the discharge event. Figure 1 External Discharge in Air 2.4 TYPICAL 1000 kv DC TEST SYSTEM Generator Internal Capacitance Charging current Voltage Drop 10% Using the formula below Δt = C ΔU i C= 5nF I = 5mA ΔU = 100 kv The recharge time calculated for one time constant is t=100ms. [3] [8] It is safe to say that in several cycles or less than 0.5 to 1 second the voltage will be restored the original voltage. Without fast recording devices monitoring the output voltage, this event may not be noticed. The normal apparatus test duration is either 1- minute, 5 minutes or 1 hour. For any of these test durations "a few seconds" defined as less than 3 seconds would mean that the voltage was lower than 10% for at worst 5% of a 1-minute test. For longer tests it is a lower percentage. This is if one allowable event happened during a single test and if we agree that 3 seconds is a valid duration. In practice most test engineers would call a few second discharge a very long discharge during a DC test and probably result in a much larger voltage drop than 10% because of the energy required to sustain such a long non-disruptive discharge and the source voltage regulation. DC Pollution testing can have currents of several amperes for such long durations without flashovers. If a 3 second discharge could never realistically be sustained with normal DC test without exceeding the 10% voltage drop limit then this requirement should be changed in the IEC standard. An easier approach for all of this is to say that if you exceed a

3 10% voltage drop of any duration your test has failed. This ends the ambiguity of the current standard. A more realistic testing situation is where short duration, non-disruptive discharges (<10 us) and short recharge (<100ms) events happen randomly but can reoccur during a test. It is possible that many of these >10% voltage drop events can occur during a test and no flashover ever occurs which should mean failure. As long as these dips are less than 10% of the test voltage there would be no concern based on magnitude alone. If we agree that the voltage drops cannot exceed 10% for U >90% of the test voltage then we can define how much time the test voltage can be below Utest. In absence of guidance from the relevant technical committees, a proposed value of 1% can be considered. For a one-minute test this would be 0.6 seconds where the voltage can be below the test voltage but above 90% of the test voltage. As a note, the rise time and duration of the events we are talking about are much faster than the frequency of the ripple voltage. In IEC60060 Part 2 the only measurement system frequency characteristics specified is for having adequate bandwidth to measure the ripple voltage of a DC test source. That ripple frequency is usually between 100 and 400 Hz. Clearly, the standard measurement system is inadequate to measure fast voltage dips during DC tests that can be "instantaneous". What we really need is to know on a charging recharging cycle-to-cycle basis if the voltage has dropped but not the actual shape of the drop. The authors hope that research will continue using fast response measuring systems to accumulate data to record the actual characteristics of the non-disruptive discharges in dry, wet, pollution and combined tests. 2.5 DC SUMMARY For HV and UHV DC testing if a test circuit does not have audible or visually observable nondisruptive discharges then IEC qualified DC measuring systems are adequate. For HV and UHV DC tests where audible and visual non-disruptive discharges are present a measurement system with increased bandwidth and a digital recorder should be used to document the test results. It is of no interest to actually see the shape of a non-disruptive discharge. We are interested in the maximum drop in the DC voltage for each charging cycle. Therefore if we have a sufficient sampling rate to record the DC source charging rate (normally twice the mains frequency) for each cycle this would be sufficient to know if the DC voltage has dropped. 2.6 DIGITAL RECORDERS FOR DC HV AND UHV If the charge cycle frequency of the DC source is 2 x 50 Hz or 100 Hz, typical for voltage doublers [8], then the bandwidth of the digitizer to record the ripple voltage and to satisfy the Nyquest Theorem has to be at least 2 x and better for 10x to have safety factor. The characteristics of the digital recorder could be (ripple only): Rated Resolution Sampling Frequency Record length 1-minute test > 8 bits >1k Samples /sec 60k samples 5 minute test 300k samples 60 minute test 1 M samples The stated requirements for the digital recorder are easily met by commercially available components. It should be pointed out that faster events may need to be recorded such as voltage drop during streamers. In this case the digital recorder characteristics will approach those requirements as used for measuring Lightning Impulse voltages as required by IEC DC MEASUREMENT SYSTEM The voltage divider to derive the signal for the digital recorder and measurement system must have a bandwidth sufficient to record the ripple voltage that is normally 2X the mains frequency. Using the same criteria as for digitizers for this purpose the measurement system should have a bandwidth of DC to 2 khz minimum to detect voltage drops. The voltage divider for actually measuring the fast events must exceed the normal requirement for measuring ripple and DC magnitude and even for capturing the 10% voltage dips in normal testing. This is more interesting from a purely research standpoint. A suggestion for researchers studying this problem would be to use "universal" or RCR type dividers that meet the requirements for impulse voltage measurements. Future work may make better definitions

4 2.8 APPLICATION NOTE It should be noted here that the vast majority of DC testing is done with voltages under 500kV and requires no consideration of non-disruptive discharges in the test and measuring circuit during normal testing. It is mostly when we go above 400 to 500kV, especially with positive polarity, that these non-disruptive discharge problems exist and can impact test results if ignored. With the introduction of more HV DC transmission systems at the higher voltage levels, the issue of voltage drop will become more important in the future. 3 AC VOLTAGE REQUIREMENTS 3.1 FROM IEC60060 Part Paragraph 6 Tests with AC Voltages Paragraph voltage drop instantaneous reduction of the test voltage for a short duration of up to a few periods Paragraph Tolerances For test durations not exceeding 60 s, the measured values of the test voltage shall be maintained within ±1 % of the specified level throughout the test. For test durations exceeding 60 s the measured value of the test voltage shall be maintained within ±3 % of the specified level throughout the test. The test voltage source including the supporting capacitances should be adequate to supply the transient discharge currents also in the case of wet and pollution tests with a voltage drop of < 20 %. Paragraph Generation of the test voltage Paragraph Requirements for the transformer test circuit It is normal for apparatus to produce some current pulses since the test voltages are much higher than the operational voltages and these devices often lack large electrodes and ground shields to keep the test object electrically quiet. Since the current pulses are of short duration, voltage drops may be unrecognized by conventional AC measuring systems. The voltage stability of an AC. test system used in tests with time varying leakage current pulses can be verified by using a voltage measuring system with sufficient bandwidth. 3.2 SUMMARY OF AC REQUIREMENTS AC Voltage can drop no more than 20%, instantaneously and cannot exceed a 20% drop for more than a few line periods. We have the same ambiguity and contradiction as we have with DC. The voltage drop limit must not be exceeded at any time during the test. We have the same question as in DC: What is "a few periods"? What is "instantaneous"? Another question arises as to why we have a 20% limit for AC and a 10% limit for DC? This leads to "What are the characteristics of a measuring system to be able to determine if the requirements are met?" 3.2 COMMENTS ON AC REQUIREMENTS The following discussion relates to AC systems for Dry, Wet, Pollution, Combined or Composite testing. The AC and DC systems for pollution and wet testing have additional requirements on static and dynamic voltage regulation that normal AC or DC test systems for routine tests do not have. In general, this forces the requirements for test sources for pollution and wet testing to be higher current rated and also lower impedance or "stiffer" in source characteristics and fast voltage adjustment schemes. The same comments made about non-disruptive discharges in DC testing apply to AC testing. In the case of AC testing we have to decide the number of line periods that constitute "a few". If we say 3 or 5 we still must have a measuring system that can record the peak voltage of each line half period for comparison with the next via automated means. Again, the simple solution will be to put a limit on the amount of voltage drop and avoid the trap of how much time below the limit is allowed. Figure 2 shows the typical problem we encounter when we have streamers, which can persist for some time before they extinguish. When continuous streamers are present it is recommended that faster responding voltage measurements are made to ensure that the test voltage is held within the voltage drop limit for the duration of the test.

5 comment as was given in the DC section applies as a suggestion to researchers investigating the transient voltage drops in AC testing. However instead of RCR or "universal" voltage dividers, series damped or purely capacitive voltage dividers may be used in this case. 5 CONCLUSION The current standards do not clearly define the requirements for allowable Voltage Drop Fig. 2 AC External Discharge in Air 3.3 REQUIREMENTS OF MEASUREMENTS IN AC SYSTEMS If we wish to record the highest voltage for each half cycle of the applied AC voltage we will have a bandwidth requirement of 100 or 120 Hz. Following IEC, AC measuring systems must have a bandwidth sufficient to record up to the 7 th harmonic of the fundamental test frequency or Hz. It is normal to have some harmonics particularly with transformers supplying capacitive loads. [9] If we satisfy the Nyquist theorem we will need to sample at twice that rate with a digitizer and up to 10x to add safety factor. This 10X factor will help us in the cases where the test voltage waveform has a crest factor of up to 10% deviation from a perfect sine wave. The characteristics of the digital recorder for AC testing could be: Rated Resolution > 8 bits (1/256) Sampling Frequency Record length >2k samples /sec 1 minute test 120k samples 5 minute test 600k samples 60 minute test 7.2 M samples The requirements for the digital recorder for AC are similar to the requirements for a digital recorder for DC. The measuring system for AC systems for normal applications is inadequate for measuring the characteristics of pulses generated during Wet, Pollution, Combined or Composite tests. A similar The current standards do not clearly define the characteristics of a HV measurement system to know if the allowable voltage drop is exceeded. The normal measurement systems used for HV AC and DC test systems with voltage ratings above kv may not be sufficient to measure the output voltage correctly during periods where nondisruptive discharges occur. More research is required to provide the data to improve the existing standards. 6. RECOMMENDATIONS FOR STANDARDS REVISIONS Determine why AC and DC systems have different allowable Voltage Drop values. (Suggest 10% for both) Define the bandwidth or rise time of measurement systems when the measurement of Voltage Drop is required Define the typical current pulses found in pollution testing to aid in the characterization of test sources Define the total time that a test voltage can be in the range of 90 to 100% as a percentage of the total test time. 6 ACKNOWLEDGMENTS The authors acknowledge the assistance and comments of Dr. Wolfgang Hauschild in preparing this document. REFERENCES [1] IEC Ed. 3 "High Voltage Test Techniques Part 1 "General Definitions and Test Requirements 2010 [2] IEC Ed. 3 "High Voltage Test Techniques Part 2 Measuring Systems" [3] Larzelere, W., Gamlin, M., Acceptable Voltage Drop during AC and DC HV Testing, CIGRE WGD1.35, Kista 2012

6 [4] Hauschild, W., Lemke, E., High-Voltage Test and Measuring Techniques, Springer, Heidelberg, 2014 [5] Meek, J.M. and Craggs, J.D. Electrical Breakdown of Gases, John Wiley, New York 1978 [6] Peek, F.W., Dielectric Phenomena in High- Voltage Engineering, McGraw-Hill, New York, 1929 [7] Naidu, M.S. and Kamaraju, V. High Voltage Engineering, Tata-McGraw Hill, New York 1995 [8] Lee, Reuben, Electronic Transformers and Circuits, John Wiley & Sons, New York 1955 [9] Blume, L.F. Boyajian, A., Camilli, G. et al, Transformer Engineering John Wiley & Sons, New York 1938,1951