Determination of Arrester Energy Handling Capability - Testing Investigation Surge Protective Devices Committee Spring 2005 Meeting

Transcription

1 Determination of Arrester Energy Handling Capability - Testing Investigation Surge Protective Devices Committee Spring 2005 Meeting Raymond C. Hill, PE

2 Introduction Conflicting opinions exist concerning the method of rating the energy handling capability of MOV arresters. The top contenders comprise the following methods: Joule kj / kv mcov Coulomb charge transfer I-squared-t action integral

3 Introduction A project proposal was made to NEETRAC by an SPDC officer and member of the NEETRAC Technical Advisors to investigate. In the fall of 2001, NEETRAC launched Phase I of a Baseline Research Project funded by the membership in order to investigate and provide input to the IEEE SPDC.

4 Introduction Phase I involved a test program utilizing one size of a single MOV arrester block and was completed at the end of Phase II was then proposed to investigate additional MOV arrester blocks with different aspect ratios.

5 Introduction In the fall of 2003, NEETRAC launched Phase II of the investigation. Phase II was a test program utilizing several sizes of a single MOV arrester block with different aspect ratios. NEETRAC has recently completed Phase II of the investigation.

6 Purpose & Scope Phases I & II To investigate the energy handling capability of single MOV arrester blocks To determine which method(s) of energy rating will describe this parameter across the board providing an equal energy rating system for all concerned.

7 Purpose & Scope Phases I & II Various sizes of blocks with different aspect ratios were tested in order to investigate the relationship of height, width, volume, and cross-sectional area to energy handling capability. Various impulse waveforms were used to investigate the relationship between waveshape and energy handling capability of the various sizes and aspect ratios of MOV arrester blocks. A total of more than 160 MOV blocks were surged during this investigation.

8 Test Procedure New MOV arrester blocks were surged (individually) to determine the single-surge-to-failure level using various current impulse waveforms. The surge current and voltage waveforms were digitized and recorded. Calculations were then made to determine the Joule, Coulomb, and I 2 t values of the surges just prior to and after destruction.

9 Test Procedure The three methods of energy rating were then analyzed and compared utilizing the various parameters of the MOV arrester blocks. A determination was then made as to which method was the most consistent and accurate.

10 Test Procedure Waveforms, Generators, and Fixtures Standard Wave 8 / 20 µs Long Tail Wave 9 / 170 µs Triangle Wave / 410 µs / 370 µs

11 Test Procedure Waveforms, Generators, and Fixtures Impulse Current Waveshapes ( Normalized to 1kA) Current (A) Time (us) 180/410 9/170 10/350 8/20

12 Test Procedure Waveforms, Generators, and Fixtures 8 / 20 µs generator and SF 6 pressure vessel

13 Test Procedure Waveforms, Generators, and Fixtures SF 6 Pressure Vessel

14 Test Procedure Waveforms, Generators, and Fixtures Sample Fixture within SF 6 Pressure Vessel

15 Test Procedure Waveforms, Generators, and Fixtures Long Tail and Triangle Wave Generator

16 Test Procedure Waveforms, Generators, and Fixtures Sample Holder for the Long Tail and Triangle Waves

17 Sample Sets for Phases I & II Quantity Height (inches) Height (mm) Width (inches) Width (mm) Table 1 Aspect Ratio (width/height) Volume (mm3) Crosssectional Area (cm2) MCOV (kv) approx. Sample Designation A1 - A B1 - B C1 - C D1 - D E1 - E F1 - F G1 - G H1 - H A - ZZZZ MOV blocks were provided by Cooper Power Systems and Hubbell / Ohio Brass

18 Sample Sets for Phases I & II A E B F C G D H Complete Set

19 Failure Criteria Individual samples were surged only once each until the failure level was achieved for that block design. Several data sets above the failure level and below (withstand) were then acquired. The one block - one surge only protocol was maintained throughout. A failure was defined as any physical damage to the block.

20 Failure Criteria A flashover across the edge of the block was not considered a failure. Examples of failures:

21 Waveshape Triangle (180/410) Phase I Results Table 2 Current Current Duration Capacitor (I**2)*t Front Tail or Time- Charge Peak (Ampere- Block Time Time to- Voltage Current I*t squared- V*I*t Withstand / Sample (us) (us) Failure (kv) (ka) (Coulombs) seconds) (kj) Fail Comments CC w/stand Epoxy collar bubbled DD fail Very small hole and crack at edge EE w/stand Epoxy collar bubbled FF w/stand Epoxy collar bubbled GG w/stand Epoxy collar bubbled MMM w/stand NNN w/stand SSS fail Channel part way down the edge TTT w/stand UUU w/stand VVV w/stand WWW w/stand XXX w/stand YYY w/stand ZZZ w/stand AAAA w/stand BBBB w/stand CCCC 172 n/app fail Failed near edge; blew off portion of block; epoxy collar bubbled DDDD w/stand EEEE 178 n/app fail Blew off chip on the edge

22 Waveshape Long Tail (9/170) Sample Current Front Time (us) Current Tail Time (us) Duration or Time- to- Failure Phase I Results Table 3 Capacitor Charge Voltage (kv) Peak Current (ka) I*t (Coulombs) (I**2)*t (Amperesquaredseconds) V*I*t (kj) Block Withstand / Fail HH fail II fail JJ fail KK w/stand LL w/stand MM w/stand NN w/stand OO fail PP w/stand QQ w/stand PPP fail QQQ w/stand FFFF w/stand GGGG fail HHHH fail Comments Channel halfway down the edge Channel halfway down the edge Channel halfway down the edge, then blew off portion of the block Channel halfway down the edge Puncture near center; split block Puncture 5 mm from edge; blew off portion of the block Channel part way down the edge

23 Phase I Results Table 4 Waveshape Sample Standard (8/20) Current Front Time (us) Current Tail Time (us) Duration or Time- to- Failure Capacitor Charge Voltage (kv) Peak Current (ka) I*t (Coulombs) (I**2)*t (Amperesquaredseconds) V*I*t (kj) Block Withstand / Fail Comments Surged under SF6; failed during current reversal; failed near edge; blew off portion of the block; left a tree-like pattern on the side ZZ , fail XX , w/stand Surged under SF6 CCC , w/stand Surged under SF6 Surged under SF6; failed along the edge burning a tree-like EEE , fail pattern on the side FFF , w/stand Surged under SF6

24 Phase I Results A first review of the impulse data values for Coulombs, I 2 t, and Joules reveals no obvious correlation among the three different waveforms. However, when the total duration of the surge or time-to-failure is taken into account a correlation appears. This is shown graphically in the following figures.

25 Phase I Results Arrester Block Single-Shot Energy Capability Kilojoule Method y = e x 100 Kilojoules Pulse Duration or Time-to-Failure

26 Phase I Results Arrester Block Single-Shot Energy Capability Action Integral (Ampere-Squared-Seconds) Method 1.0E+06 y = 45682e x 1.0E+05 Ampere-Squared-Seconds 1.0E E E Pulse Duration or Time-to-Failure (us)

27 Phase I Results Arrester Block Single-Shot Energy Capability Coulomb Method y = e x 10 Coulombs Pulse Duration or Time-to-Failure (us)

28 Phase I Results When the pass/fail data for all three waveforms is plotted versus total pulse duration or time-to-failure for each energy method, a definitive border appears among the pass/fail data. In general, any current surge with an energy content and duration which falls below this border will not cause an MOV block failure for the design tested. This border or trend line can be defined as shown in the following equation:

29 Phase I Results β e -mt β represents the y-intercept, which is the maximum energy achieved with an infinitesimally small pulse width. The coefficient m represents the slope of the line and t is time (pulse width). With this method, the borderline of failure can be described by defining these two coefficients.

30 Phase I Results This equation is waveform independent for durations less than 1000 microseconds. Each energy rating method would require a unique set of coefficients. At this point, the equation only applies to one MOV block size. This equation fits the Coulomb or charge transfer method better than the others which exhibit more of an upturn for times less than 200 microseconds. Obviously, further investigation is required.

31 Phase I Results Additional Observations Worth Noting The slope of the Coulomb equation was found to be small, such that, when considering the data scatter (and with further investigation), one might conclude that a single Coulomb value could be assigned to the MOV arrester block design utilized in this project in lieu of an equation. With the exception of Sample PPP, all of the block puncture locations were along the edge or within 5 mm of the edge.

32 Phase II Results - Tabulated Table 5 Waveshape 8/20 Sample Current Front Time (us) Current Tail Time (us) Duration or Until Failure or Flashover Capacitor Charge Voltage (kv) Peak Current (ka) I*t (Coulombs) (I**2)*t (Amperesquaredseconds) V*I*t (kj) Block Withstand / Fail / Flashover (f / o) Comments MCOV (kv) approx. Aspect Ratio (width/ height) B , f / o 20 psig SF C , w/stand 20 psig SF C , f / o 20 psig SF C , w/stand 25 psig SF C , f / o 20 psig SF D , w/stand 15 psig SF D , w/stand 15 psig SF D , w/stand 25 psig SF E , f / o 20 psig SF E , f / o 20 psig SF E , f / o 20 psig SF E , f / o 20 psig SF UU , w/stand 25 psig SF Width (mm) Height (mm)

33 Phase II Results Tabulated Table 6 Waveshape 9/170 Sample Current Front Time (us) Current Tail Time (us) Duration or Until Failure or Flashover Capacitor Charge Voltage (kv) Peak Current (ka) I*t (Coulombs) (I**2)*t (Amperesquaredseconds) V*I*t (kj) Block Withstand / Fail / Flashover (f / o) Comments MCOV (kv) approx. Aspect Ratio (width/ height) C , fail C , w/stand C , fail C , fail C , fail C , w/stand C , fail C , fail C , fail C , fail C ,497 n/av w/stand no voltage trace C , fail C , fail C , fail C , w/stand C , fail C , w/stand C , fail C , w/stand C , fail C , fail C , w/stand C , w/stand B , fail B , w/stand B ,835 n/av fail B , fail B , fail B , fail B , w/stand B , w/stand B , w/stand E , w/stand gen. can't fail E , w/stand gen. can't fail Width (mm) Height (mm)

34 Phase II Results - Tabulated Waveshape Triangle Wave Sample Current Front Time (us) Current Tail Time (us) Duration or Until Failure or Flashover Capacitor Charge Voltage (kv) Peak Current (ka) I*t (Coulombs) Table 7 (I**2)*t (Amperesquaredseconds) V*I*t (kj) Block Withstand / Fail / Flashover (f / o) Comments MCOV (kv) approx. Aspect Ratio (width/ height) B , f / o B n/av n/av n/av n/av n/av f / o can't determine when f/o ocurred C w/stand gen. can't fail C w/stand gen. can't fail C , w/stand gen. can't fail C , w/stand gen. can't fail C , w/stand gen. can't fail C , w/stand gen. can't fail C , w/stand gen. can't fail Width (mm) Height (mm)

35 Phase I & II Results with AE I Borderline Equation Arrester Block Single-Shot Energy Capability Kilojoule Method by Aspect Ratio and Pulse Duration Kilojoules w/s 1.4 AEI w/s 1.4 AEI fail 1.5 w/s 2.1 w/s 2.1a w/s 2.1b w/s 2.1a fail 2.1b fail 2.7 w/s 2.7a w/s 7.1 w/s 7.1 fail 2.1c w/s 7.1a w/s AEI Eq Pulse Duration

36 Phase I & II Results with AE I Borderline Equation Arrester Block Single-Shot Energy Capability Action Intergal Method by Aspect Ratio and Pulse Duration Ampere-squared-seconds 1.0E E E E w/s 1.4 AEI w/s 1.4 AEI fail 1.5 w/s 2.1 w/s 2.1a w/s 2.1b w/s 2.1a fail 2.1b fail 2.7 w/s 2.7a w/s 7.1 w/s 7.1 fail 2.1c w/s 7.1a w/s I sqrt t AEI 1.0E Pulse Duration

37 Phase I & II Results with AE I Borderline Equation Arrester Block Single-Shot Energy Capability Coulomb Method by Aspect Ratio and Pulse Duration Coulombs w/s 1.4 AEI w/s 1.4 AEI fail 1.5 w/s 2.1 w/s 2.1a w/s 2.1b w/s 2.1a fail 2.1b fail 2.7 w/s 2.7a w/s 7.1 w/s 7.1 fail 2.1c w/s 7.1a w/s AE1 Eq Pulse Duration

38 Phase II Results As can be seen, the borderline equation from Phase I requires adjusting in order to take into account the variance caused by the different MOV block sizes. Many different aspects of MOV block size were compared with the three energy rating methods. The correlations of predominant interest are the following:

39 Phase I & II Results Arrester Energy in Kilojoules per Unit Diameter Kilojoules / mm diameter w/s AE I fail AE I w/s AE II fail AE II AEI Eq Pulse Duration (us)

40 Phase I & II Results Arrester Energy in Kilojoules per kv of MCOV 100 kilojoules per kv of MCOV w/s 2.1 fail 7.1 w/s 7.1 fail 2.1c w/s 7.1a w/s 1.4 AE I w/s 1.4 AE I fail Duration (us)

41 Phase I & II Results Arrester Energy in Kilojoules per kv of MCOV per Unit Diameter 1.00 kj / kvmcov / mm 0.10 w/s AE I fail AE I w/s AE II fail AE II Pulse Duration

42 Phase I & II Results Arrester Energy in "I squared t" per Unit Diameter Ampere-squared seconds / mm diameter w/s AE I fail AE I w/s AE II fail AE II I sqr t AEI Pulse Duration (us)

43 Phase I & II Results Arrester Energy in "I - squared - t" per Kilovolt of MCOV Ampere-squared seconds / kv of MCOV w/s AE fail AE w/s AE fail AE Pulse Duration (us)

44 Phase I & II Results Arrester Energy in Coulombs per Kilovolt of MCOV 10.0 C / kvmcov 1.0 w/s AE I fail AE I w/s AE II fail AE II Pulse Duration (us)

45 Phase I & II Results Arreser Energy in Coulombs per Unit Diameter 1.00 Coulombs / mm diameter 0.10 w/s AE I w/s AE II fail AE I fail AE II AE I Borderline AE I & AE II Borderline Threshold Pulse Duration (us)

46 Phase I & II Results Arrester Energy in Coulombs per Unit Cross-sectional Area 1.00 Coulombs per square cm 0.10 w/s AE I fail AE I w/s AE II fail AE II AE I Borderline AE I & AE II Borderline Threshold Pulse Duration (us)

47 Phase I & II Results Arrester Block Single-Shot Energy Capability Coulomb Method by Cross-sectional Area 10.0 Coulombs 1.0 AE I Withstand AE I Fail AE II Withstand AE II Fail Cross-sectional Area (square-cm)

48 Discussion / Conclusions The kilojoule per unit diameter data exhibits the most scatter of the three methods, with the most scatter occurring at the longer durations. The I 2 t per unit diameter data exhibits less scatter than the kilojoule method. The Coulombs per unit diameter and per unit cross-sectional area data also exhibit less scatter than the kilojoule method. (Converting the diameter to cross-sectional area allows for square MOV blocks.)

49 Discussion / Conclusions When inspecting the Coulombs per unit crosssectional area graph, one could conceivably rate these arrester blocks at 0.09 or 0.1 Coulomb per square centimeter. A flat rating such as this would then be MOV recipe dependent along with any other presently unknown physical or dimensional limits. It is understood that two manufacturers provided samples, which could cause some scattering of the data due to different recipes. Any rating from this investigation is still limited to total pulse durations less than 1000 microseconds.

50 Future Work It is understood that further investigation is required before any solid conclusions can be reached. More data sets are needed. This data is presented to the IEEE Surge Protective Devices Committee as a challenge to industry leaders and other laboratories to help investigate further the possibilities presented here in order to obtain a more consistent method of rating arrester energy capability across the board for all concerned.