ARC FLASH TESTING OF TYPICAL 480-V UTILITY

2y ago
9 Views
2 Downloads
2.38 MB
15 Pages
Last View : 2d ago
Last Download : 2m ago
Upload by : Jerry Bolanos
Transcription

ARC FLASH TESTING OF TYPICAL 480-V UTILITY EQUIPMENTCopyright Material IEEEPaper No. ESW-XX (do not insert number)Marcia L. Eblen, P.E.Member, IEEETom A. ShortSenior Member, IEEEElectric Power Research Institute (EPRI)801 Saratoga RdBurnt Hills, NY 12027USAtshort@epri.comPacific Gas & Electric3400 Crow Canyon RdSan Ramon, CA 94583USAMLS2@pge.comII.Abstract – A test program was completed to measure arcflash incident energy from actual 480-V utility equipment todetermine the most appropriate flame resistant clothing forutility workers.The equipment tested included selfcontained and CT-rated meters, padmounted transformersecondary cubicles, power panels, and network protectors.Testing was performed to determine the sustainability of lowvoltage arcs in actual utility equipment, the most appropriatecalculation method to predict the measured incident energy,and to identify any key variables that would effect bothduration and heat from this type of equipment.480-V THREE PHASE SELF-CONTAINEDMETERSA.Test SetupStandard seven jaw (type 16S) self-contained three-phase480-V meter bases were mounted to a back board as shownin Fig. 1. Incident energy was measured by nine coppercalorimeters on stands spaced 8” apart and positioned 18”away from the meter jaw. Calorimeters were built andcalibrated according to ASTM specifications [7, 8].Index Terms — Incident Energy, Heat Flux Rate, 480-VMeters, Power Panels, Padmounted Transformers, NetworkProtectors.I.INTRODUCTIONSeveral investigators have tested arc-flash scenarios atvoltages less than 1000 V. The pioneering work of Doughtyet al. [1,2] provided test data that was further expanded inIEEE Std. 1584-2002 [3]. The Doughty et al. tests and theadditional IEEE tests were based on a box open at the frontwith vertical electrodes. Stokes and Sweeting [4] and Wilkinset al. [5] showed that electrode orientation makes asignificant difference in the direction in which the energy fromthe arc projects and a significant difference in incident heatenergy. Wilkins et al. [6] showed that arc sustainability andincident energy are also impacted by insulating barriers.Based on these results and internal testing of various arcgap configurations, most accurate results were expectedfrom testing actual equipment.Predicting incident energy from utility systems’ low-voltageequipment includes many variables. When using worst caseassumptions, predictions of incident energies often resultedin extremely high values. These values were not consistentwith the injuries or events that have happened in the past. Inmany instances, the arcing fault magnitude is below anyprotective device setting resulting in arc durations that areonly limited by the ability of the arc to self extinguish. Theauthors embarked on a testing project to determinemaximum arc durations and incident energies for major utilityequipment e.g. meters, power panels, padmount transformersecondary cubicles, and network protectors. The testingproject used actual utility equipment and enclosures toimprove the assumptions made for utility low voltage arcflash calculations.Fig. 1. Self-contained meter test setupThe line side of the meter base was shorted across threephases with a solid 12-AWG wire, as shown in Fig. 2.Testing was conducted at four different values of bolted faultcurrent: 6.6 kA, 12.7 kA, 25.7 kA, and 44 kA. All test eventswere allowed to continue until the arcing self extinguished.Fig. 2. Shorting wire used to start the fault1

Table 1Measured incident energy for self-contained metersB.Test DataMeasuredincident energy,ArcingMeasured2cal/cmCurrent,Duration, heat rate,2AvgMaxmsecTestkAcal/cm /sec244 kA bolted fault, IEEE 1584 flux rate 82.7 cal/cm /sec,2Commercial program 58.1 cal/cm 029.7Test data is shown in Table 1 for incident energiesmeasured at 18 inches from the calorimeters. As shown inFig. 3, the majority of the events ended when all of thecopper from the line side meter jaws had vaporized alongwith the 1/0 copper lugs that attached the power source tothe meter base.225.7 kA bolted fault, IEEE 1584 flux rate 50.3 cal/cm /sec,2Commercial program 20.7 cal/cm 6126.541.21292.37.515.8176.042.5Fig. 3. Meter base after a faultThe IEEE1584 calculation results in Table 1 were madeusing the IEEE 1584 spreadsheet calculator with the boltedfault current, equipment type MCC/panel, grounded system,and working distance of 18”. The commercial programresults were made using 1 inch arc length, 480V systemvoltage, the available bolted fault current, 18’ distance to thearc, and no multiplier.212.7 kA bolted fault, IEEE 1584 flux rate 26.2 cal/cm /sec,2Commercial program 7.2 cal/cm lysis of Test DataAll events provided a focused energy stream. Thecalorimeter directly in front of the meter often had at leasttwice the incident energy of the calorimeters on either side(offset by eight inches). The middle calorimeter takes muchmore of a direct hit from the hot plasma generated by thearcing. Lower currents led to higher incident energy becauseof longer duration, as shown in Fig. 4. At higher currents, theavailable copper at the top of the meter base vaporizedquickly, and the event self-extinguished in well under 200msec; see Fig. 5. The highest incident energies wererecorded at the 12.7-kA tests, with 20.1 cal/cm2 being thehighest. The cutting-torch action of the 12.4-kA arc may bespraying copper into the arc channel, increasing thelikelihood that the fault continues to arc. At 6 kA, the faultsapparently are not energetic enough to keep the arc channelsufficiently ionized to maintain the fault. The meter testresults highlight the importance of self clearing at 480 V.26.6 kA bolted fault, IEEE 1584 flux rate 14.3 cal/cm /sec,2Commercial program 2.8 cal/cm .92

Fig. 6. Measured vs. calculated heat flux for self-contained480-V three phase metersFig. 4. 480-V meter arc duration vs. arcing fault currentIII. 480-V THREE PHASE TRANSFORMER RATEDMETERSA. Test SetupStandard thirteen jaw (type 9S) transformer rated threephase 480-V meter bases were mounted to a back board.All meters were wired with a 10-AWG wire between the mainpower panel and the meter panel. Incident energy wasmeasured by nine copper calorimeters mounted on standsspaced 8” apart and positioned 18” away from the meter jaw,Arcing was initiated at two different locations in the meterenclosure. The meter base was shorted with 22-AWG wirefor one set of tests as shown in Fig. 7, and then the switchblock was shorted with 14-AWG wire as shown in Fig. 8.Fig. 5. 480-V meter incident energy vs. arcing fault currentComparing the measured maximum heat flux rates to twodifferent calculation methods, IEEE 1584 and commerciallyavailable software shows that the IEEE 1584 more closelypredicts the heat flux rate at lower available fault currents.Neither modeling approach predicts the leveling off of heatrate seen in the measured data between the 25kA and 44kAbolted fault currents, as shown in Fig. 6.Fig. 7. Shorting wire in the meter base3

Table 2Measured incident energy for transformer-rated metersMeasuredincident energy,ArcingMeasured2cal/cmDuration, heat rate,Current,2AvgMaxTestkAmseccal/cm /sec244 kA bolted fault, IEEE 1584 flux rate 91.7 cal/cm /sec,2Commercial program 58.1 cal/cm 46.620.1225.7 kA bolted fault, IEEE 1584 flux rate 55.8 cal/cm /sec,2Commercial program 20.7 cal/cm 41.32.415.2157.714.9212.7 kA bolted fault, IEEE 1584 flux rate 29.1 cal/cm /sec,2Commercial program 7.2 cal/cm 0.75.9443.824.026.6 kA bolted fault, IEEE 1584 flux rate 15.9 cal/cm /sec,2Commercial program 2.8 cal/cm 2.94.11530.71.42.6256.85.6Fig. 8. Shorting wires in the switch blockTesting was conducted at four different values of boltedfault current: 6.6 kA, 12.7 kA, 25.7 kA, and 44 kA. All testevents were allowed to continue until the arcing selfextinguished.B.Test DataTest data is shown in Table 2. Almost all of the faults selfcleared very quickly. Because voltage is brought to thesemeters through a relatively small wire, the voltage lead actedas a fuse. Consequently, all events were short duration anddid not generate much incident energy.C.Analysis of Test DataIn all cases the arcing current was severely limited by the 10AWG wire providing power to the transformer rated meterpanel. This effectively extinguished the arc rapidly in all 452events. Only one of the events measured above 8 cal/cm ,test number 166 which measured a peak value of 10.7cal/cm2. The remaining 44 events measured less than 2.52cal/cm , as shown in Fig. 9.Fig. 9. Arcing current vs. incident energy for 480-V CTrated meters4

1541550.00.30.10.62.94.120.459.4Fig. 12 shows a progression of high-speed video framestaken. The event lasted less than two cycles as shown inFig. 13. Note that the event progressed from a line-to-linefault to a three-phase fault in less than a quarter cycle.2.610.2IV. 480-V THREE PHASE PADMOUNTEDTRANSFORMERSA. Test SetupFig. 10 shows the spacing of the secondary terminalconfiguration of the unit used for testing. The internals of thetransformer were removed, and voltage was supplied to thesecondary terminals from the back side from the 480-V faultcurrent source.Fig. 12. 1200-fps video snapshots through an infraredpassing filterEvent 28840Currents, kA20Fig. 10. 480-V padmounted transformer secondary terminalsB. Test Data0 20Out of 35 tests, there were no cases of sustained arcing.Most arcs self-extinguished in less than 2.5 cycles with amaximum of 12 cycles. Incident energies were mostly lessthan 1 cal/cm2 with the highest at 4.0 cal/cm2.Fig. 11 shows a typical fault test initiated with a pair of vicegrips laid across phases. The phase spacing in thisconfiguration is approximately 2.75 in. This phase gap wasprogressively shortened by adding plates to see if tighterspacing would cause the arc to sustain. 400.080.090.100.110.120.13Time, secFig. 13. Current waveforms for a wrench eventFig. 14 shows several fault initiations that were tested.Both phase-to-ground and phase-to-phase faults wereattempted. Some common observations include:1) Fault progressionFaults generally became three-phase faults within a halfcycle (with the exception of event 289 where the mainpoint of the arc initiation was well away from otherphases).2) Gap lengthEven to distances as close as two inches, faults could notsustain. The arcs grow into the open space around theelectrodes until they cannot sustain. Arcs were firstinitiated at a distance of 3.7 inches and shortened untilreaching two inches.3) Fault currentThis did not seem to change arc duration characteristics.Bolted fault currents of 13, 28, and 53 kA were tested.Fig. 11. Vise grip initiated test setup5

4) Blanket coveringsOne reason the arcs clear quickly in the secondarycompartment is because of the open space. To see ifcovering would restrict the arc and lead to sustainedarcing, we initiated a phase-to-phase fault under a blanket,either with a 12-AWG fuse wire or a wrench. See event305 in Fig. 14 for one example. In three such tests, arcsdid not sustain any longer.The longest-duration arcing occurred for a configurationwhere we used a 500-kcmil conductor from either the phaseor the ground and looped it around and touched it to aterminal block. This is to replicate the condition in the fieldwhere a worker accidentally touches a conductor to thewrong phase, and that conductor is either energized orgrounded at the other end. See event 308 in Fig. 14 for anexample where the conductor is solidly grounded to theneutral bushing and then touched to the connector. In thiscase, the fault lasts longer than the fuse wire or wrenchtests, but it still clears quickly.Fig. 15 shows damage after test 310, a phase-to-phaseconnection similar to test 308 in Fig. 14. This test was mademore severe by taping the incoming cable to adjacentconductor stubs to prevent cable movement. This eventcleared in less than 12 cycles. We see that the cable andaluminum alloy connector both burned away, apparently untilthe gap was large enough for the arc to self clear. As thiswas the most severe event found so far, this fault scenariowas tried at other spacings and fault currents. Fig. 16 showsan example tested at a spacing of less than two inches.Faults still cleared within 12 cycles. Fig. 17 showswaveforms for some of the longer-duration events.285289305303Fig. 15. Results after event 310308Fig. 14. Fault variations tried6

normal conductor spacings. Because the duration is short,incident energies are low, with no event exceeding 4 cal/cm2.108count6Fault typeTool or wire faultMiswired cableMiswired cable, tied together42Fig. 16. Event 317: two-inch gap between phases0024681012Fault duration, cyclesFig. 18. Fault duration histogramEvent 310402015010 40Fault typeTool or wire faultMiswired cableMiswired cable, tied togethercountCurrents, kA 20 60Event 312540200001234Incident energy, cal/cm2 20Fig. 19. Incident-energy histogram 40V.0.100.150.20480-V THREE PHASE POWER PANELS0.25A. Test SetupTime, secA variety of 480-V power panels were tested under similarconditions to the meter tests. The calorimeter configurationwas altered to give more data at the center of the arcgenerated plasma.A. Test ResultsTable 3 shows results from tests of 50-A and 100-A ratedpanels. The arc event usually ended when the bus bracingfailed which allowed the bus bars to separate and increasethe arc length beyond the sustainable limit. The maximum2incident energy measured was 14.3 cal/cm . At higheravailable fault currents, faults self-cleared faster and led tolower incident energies as shown in Fig. 20 and Fig. 21. TheFig. 17. Waveforms from the longer duration eventsC. Data AnalysisFig. 18 summarizes the fault durations observed from the35 tests with fault type shown by color. The “miswired cable”indicates the tests with the 500-kcmil conductor jumperingground to phase or phase to phase, either tied down toadjacent stubs or free. Fig. 19 shows distributions of incidentenergies measured at 21 inches from the fault location.The durations show that the arcs cannot sustain long insecondary compartments with typical or even tighter-than-7

panels included several styles and ratings, so an exactcomparison is difficult.For two higher-rated panels, results were much different—faults did not self clear like they did with meters and smallerpanels. Fig. 22 shows the 250-A panel prior to testing.Table 3Results from 50A and 100A Power PanelsTest rageincidentenergy2(cal/cm )Peakincidentenergy2(cal/cm 23.6181.21.622810020044.014.41362.33.3Maximum incident energy, cal/cm 221815Fig. 22. 250-A power panel prior to testFig. 23 shows a video frame from this event. The event2caused an incident energy of almost 48 cal/cm at an 18-inworking distance which is considerably more thanmeasurements on any of the smaller panels or meters. Thistest configuration had an available bolted fault current of 25.7kA and a measured average arcing current of 14.8 kA.This event lasted for 0.74 sec and did self clear. However,it did not clear in the same manner as the smaller panels.One of the incoming 480-V leads at the bottom of the panelburned free, probably from rubbing against the frame. Thisacted like a fuse that helped clear the fault. Based on thevideo evidence and the condition of the bus work in thecabinet, it is likely that the fault would have continued to arc ifthe incoming lead had not burned free.Bolted fault12 kA25 kA44 kA105005101520Average arcing current, kAFault duration, secFig. 20. Maximum incident energy from 50-A and 100-ApanelsBolted fault12 kA25 kA44 kA1.5Fig. 23. 250-A power panel for test 2321.0Fig. 24 shows an identical panel after another test. Thisfault heavily damaged the calorimeter array, and the panelenclosure suffered significant damage. This event did notself clear. The fault was cleared by laboratory protection justprior to two seconds. Incident energies likely exceeded140 cal/cm2. Fig. 25 summarizes the measurements duringthis test. This fault had almost three times the arc energy inthe fault as test 232. The event showed no signs of selfclearing as the smaller panels and likely would havecontinued until all the bus had been consumed.0.50.005101520Average arcing current, kAFig. 21. Fault duration from 50-A and 100-A panels8

The burn rate estimated from Stanback’s equation is:Y 1.519E-6 Iarc1.5 1.519E-6 (16,400)1.5 3.2 in3/secConsidering the duration of just less than two seconds, theburn rate based on the test is:Stanback estimate of volume consumed/phase 6.4 in3The bus configuration and spacing in the 250-A panel wasenough to allow the arc to sustain for a significant period oftime. With vertical bus bar oriented flat to each other, the arcgap does not elongate as the event continues. In smallerpanels and in other equipment, either the volume was low orthe spacing was large enough for 480-V arcs to self clear ina relatively short period of time. Equipment with bus barsimilar to the 250-A panel will likely have extremely longduration arcing events, or they must rely on systemprotection to end the fault in a realistic time frame. Withoutrelay protection, equipment of this type will be extremelyhazardous to work on in an energized condition.Fig. 24. 250-A panel after test 233VI. 480-V NETWORK PROTECTORSA. Test SetupFig. 26 shows the network protector used during tests. Theinternal operating mechanisms have been stripped from theunit. The unit is energized from the top, which is the networkside of the unit. A common work procedure is removing thefuses on the network feed. Bus bars from the top wereincluded in the box. The unit is fed by a 480-V source that’scapable of supplying a bolted fault current of 52 kA.Fig. 25. Measured incident energy for test 233Notes:1. Calorimeter destroyed (melted)2. Highest measured value, actual likely exceeded 140 cal/cm2B.Data AnalysisStanback [9] derived bus burn rates for 480-V faults basedon fault tests. For copper and aluminum bus bar, heproposed the following equations:Fig. 26. Network Protector Test SetupCopper bus barY 0.7230E-6 Iarc1.51.5Aluminum bus barY 1.519E-6 IarcWhere,Y, burn rate of bus, in3/secIarc, rms arcing fault current, AB. Test ResultsFault events were normally initiated with a 12-AWG copperwire connected between bus bars. Vice grips were also usedto initiate faults. With open space below the bus bars, faultsself extinguished. With the wide open spacings, arcs selfextinguish. The magnetic forces push the arc towards thebottom of the enclosure, and the arc balloons out in theprocess, reaching a length where the fault cannot sustain.The amount of bus bar burned in the panel in test 233matched closely with the Stanback equation. The panel had3.25 0.25 inch aluminum bus bar, and approximately eightinches was burned from each phase. Based on thesemeasurements, the bus volume consumed was:3Actual Volume consumed/phase 3.25 0.25 8 6.5 in9

In order to asses how confinement impacts arc sustainability,a metal ground plane was added behind and below the busbar, as shown in Fig. 27. In this configuration, faults wereable to sustain and did not clear until the laboratoryprotection tripped the circuit.Event 265Currents, kA40200 20 400.080.090.100.110.120.13Time, secFig. 29. Current waveforms with micarda spacersC.Data AnalysisThis section documents many of the calorimeter incidentenergy readings obtained during the network protector tests.Unless otherwise stated, these incident energies weremeasured 18 in from the arc initiation point. Fig. 30 showshow incident energy varied with duration. For the cases witha bolted fault current of 44 kA, the incident energy wasreasonably linear. Fig. 31 shows the relationship betweenarc energy and incident energy, and it is reasonably linearacross fault current ranges. Fig. 32 shows that the ratiobetween arc energy and incident energy is not stronglyduration dependent, meaning that the incident energy isdirectly related to arc energy without an extra effect causedby duration. Fig. 33 shows that arc power and incident heatrate also track linearly. These graphs support two basicassumptions used in IEEE 1584 and other analysis: (1)incident energy increases linearly with fault duration, and (2)incident energy is linearly related to arc energy.Fig. 27. Configuration with metal ground planeNetwork protectors normally have micarda dividers thatseparate bus bars. Fig. 28 shows a test setup with a micardadivider separating one bus bar from the other two bus barsthat are shorted with a 12-AWG copper wire.35Fig. 28. Test configuration with micarda dividersIncident energy, cal/cm230Fig. 29 shows that even with the micarda divider (event 265),the fault escalated to phase C in less than one half cycle.With a reduced gap between the micarda divider (event 266)and the ground plane, the fault still escalated in less than ½cycle.25Bolted faultcurrent, kA25.727.84452.62015105200300400500600Duration, msecFig. 30. Incident energy vs. duration10700800900

2727035256602520253Bolted faultcurrent, kAa 26a 28a 44a 5327626726815264265266105Heat rate, cal/cm2/secIncident energy, cal/cm230279257258247255 251269248271245277278500Range ofduration, msec(0,100](100,200](200,300](300,1e 03]5040Bolted faultcurrent, kA2628445330201010001500200025006000Arc energy, watt hours80001000012000Fig. 31. Comparison of incident energy to arc energyIn many tests, calorimeter measurements were taken atdifferent distances as shown in Fig. 34 with the closestcalorimeters at 18 in and the back calorimeters at 24 in fromthe arc initiation point. The measurements at each locationtrack closely as shown in Fig. 35. The slope of the linear fitto Fig. 35 is 0.52, which equates to a distance factor of 2.3.This is higher than the distance factor of 1.473 used in IEEE1584 for low-voltage switchgear. Note that in thisconfiguration, the front calorimeters are located such thatthey may have shielded the back calorimeters. This shieldingmay have reduced the energy to the back calorimetersenough to produce an artificially high distance factor.Ratio of incident energyto arc energy, cal/cm2/kWh201510520040060016000Fig. 33. Arc power vs. incident heat rateFig. 31 test notes:245: The test did not have the plate behind the bus bars(only below), possibly allowing more of the blast togo down the enclosure rather than out the front.256: The shelf below the bus bars blew out; themaximum readings were on the bottomcalorimeters which was unusual.257: Arc power was underestimated some because themiddle-phase voltage was lost for two out of sevencycles.267 & 268: Configuration had a larger bus gap: 4" and6", so the plasma may have been directeddifferently.272: Arc energy was underestimated by about 10%;only 90% of the waveform was captured.014000Arc power, kW800Fig. 34. Calorimeter arrangement1000Fault duration, msecFig. 32. Energy transfer ratio vs. fault duration11

see less energy, but the flaming arc plasma coversa larger area.IEEE 1584 comparison A351030Incident energy, cal/cm2Calorimeter at 24"1555101520253025Bolted faultcurrent, kA25.727.84452.6201510355Calorimeter at 18"10Fig. 35. Incident energy measured at different distances 70Incident energy, cal/cm23025Bolted faultcurrent, kA25.727.84452.62015105The main findings of the network protector tests are asfollows: 6035Both comparisons show that IEEE 1584 generally overpredicts incident energies. Actual measurements aregenerally 30 to 75% of the IEEE 1584 prediction. 50IEEE 1584 comparison BComparison B – Fig. 37, meant to replicate the defaultIEEE 1584 calculation – Bolted available fault currents areused along with the default gap distance of 1.25 inspecified in IEEE 1584 for low-voltage switchgear. 40Fig. 36. Measured incident energy vs. calculated (with IEEE1584 using gap 2.5” and actual arcing current)Comparison A – Fig. 36, meant to replicate the test moreprecisely – Actual fault current for each test is used alongwith the network protector bus bar gap of 2.5 in. 30IEEE 1584 prediction, cal/cm2Fig. 36 and Fig. 37 shows two different evaluations of theIEEE 1584 calculated incident energy estimates and thenetwork protector test results. The y-axis values are themeasured incident energies. The x-axis values come fromIEEE 1584 estimates using the duration and currents fromthe test. The differences are as follows: 20102030405060IEEE 1584 prediction, cal/cm2Fig. 37. Measured incident energy vs. calculated (with IEEE1584 using gap 1.25” and bolted fault current)Although some faults did not sustain, we think thatsustainable arcs are certainly possible in networkprotectors.The calorimeter incident energy is linear with arcenergy.The heat rate stays the same with fault duration(you double the duration, the incident energydoubles).Micarda dividers are not effective at containing theplasma from the arc.Measurements were generally 30 to 75% of thedefault IEEE 1584 prediction.The arc plasma from a network protector failure isless focused than the meters. For a given arcenergy, a single point in front of the equipment mayOne question to consider is how representative the testedfault scenarios are to real-life operation. Our test enclosurehad the network protector innards removed. We think thatthe innards may change how the fireball propagates, butoverall, we don’t think it will change findings significantly. Theinnards will fill up more airspace and make sustainable arcsmore likely as faults were more sustainable in confinedareas.VII. CONCLUSIONSFig. 38 compares the network protector, power panel, andself-contained meter test results. This graph shows howmuch of the energy is transmitted from the arc to the12

measuring calorimeter. The major conclusions that can bedrawn from this testing are:Self-clearing – Faults in meters and small panels will selfextinguish. Faults in large panels and network protectorsmay not.Energy focusing – The small meter housing focuses the arcenergy straight out of the box in a relatively tight pattern. Thelarger network protector enclosure has less of a focusingeffect, but the incident energy impacts a much wider area.Fig. 39 and Fig. 40 show two typical events. The shape ofthe enclosure and the magnetic fields determine how the arcenergy is released.Fault current – For meters and small panels, incidentenergy decreases with higher fault current because the faultsself extinguish faster. For network protectors and large panelboards, the incident energy increases with higher faultcurrent because the faults may not self clear.Fig. 40. Typical network protector arc flashIncident energy, cal/cm24030The test results generally support using single-layer flameresistant clothing for padmounted transformers and CT-typemeters, double-layer clothing for self-contained meters andsmall panels, and flash suits for large panels and networkprotectors.Test set480 V panelsNetwork protectorsSelf contained meters20VIII. ACKNOWLEDGEMENTSThe authors with to acknowledge the invaluableassistance provided by Ralph Seban, Dan Kaufman, and thetechnicians of the Pacific Gas & Electric High CurrentLaboratory.100500 1000 1500 2000 2500 3000IX. REFERENCESArc energy, WhFig. 38. Comparison of incident energy and arc energy[1][2][3][4][5]Fig. 39. Typical meter arc flash[6]13Doughty, R. L., Neal, T. E., Dear, T. A., and Bingham,A. H., “Testing update on protective clothing andequipment for electric arc exposure,” Petroleum andChemical Industry Conference, Industry ApplicationsSociety 44th Annual, Banff, Alberta., Canada, pp. 323–336, Sept. 1997.Doughty, R. L., Neal, T. E., and Floyd, H. L.,“Predicting incident energy to better manage theelectric arc hazard on 600-V power distributionsystems,” IEEE Transactions on Industry Applications,vol. 36, no. 1, pp. 257–269, Jan./Feb. 2000.IEEE Std 1584-2002, IEEE Guide for Performing ArcFlash Hazard Calculations.Stokes, A. D. and Sweeting, D. K., "Electric arcing burnhazards," IEEE Transactions on Industry Applications,vol. 42, no. 1, pp. 134-41, 2006.Wilkins, R., Allison, M., and Lang, M., "Effect ofelectrode orientation in arc flash testing," IEEE IndustryApplications Conference, 2005.Wilkins, R., Lang, M., and Allison, M., "Effect

IEEE Std. 1584-2002 [3]. The Doughty et al. tests and the additional IEEE tests were based on a box open at the front with vertical electrodes. Stokes and Sweeting [4] and Wilkins et al. [5

Related Documents:

Arc Flash Facts Arc Flash Fact Sheet Brady Arc Flash Training Aids Promote awareness of the dangers associated with arc flash accidents and make sure your workers know how to protect themselves! Poster Highlights the common causes of arc flash and provides safe work practices and personal protection equipment requirements

130.5 Arc Flash Risk Assessment. An arc flash risk assessment shall be performed and shall: (1) Determine if an arc flash hazard exists. If an arc flash hazard exists, the risk assessment shall determine: a) Appropriate safety-related work practices b) The arc flash boundary c) The PPE to be used within the arc flash boundary

Arc-flash analysis has been performed for this site (calculations, labeling and Arc-flash hazard analysis. arc-flash boundaries). Electrical safety training for operators and maintenance personnel. Arc-flash hazard training for operators and maintenance personnel. Arc-flash hazard training focusing on selecting and using the proper PPE.

Arc Flash Testing with Current Simulation Figure 7. Arc Flash Element Settings Figure 8. Simulation Settings Follow these steps to set up the simulation mode and test the sensors: 1. Navigate to Setpoints Control Arc Flash Arc Flash 1 and set the Function to "Trip". Enable the light sensors according to their connection.

Bussmann series arc flash relay system 1. Introduction Eaton's BussmannTM series arc flash relay (EAFR) system is a combination of arc flash relay modules and sensors designed to detect and clear arc flash events in low and medium voltage electrical assemblies. An arc fault is the most devastating type of fault in medium

130.7(C)(15)(b) specify arc flash PPE category 1 or 2b Arc Flash Suit A total clothing system consisting of arc-rated shirt and pants and/or arc-rated coveralls and/or arc flash coat and pants (clothing system minimum arc rating of 40) Situations where a risk assessment indicates that PPE is required and where Table 130.7(C)(15)(a) and table .

provides an overview of arc flash hazards, arc flash calculations, and suggest a means of reducing the arc flash hazard level through faster detection and clearing of arc flash electrical faults. I. INTRODUCTION The last few years has seen a great increase in the

The work permit shall include the results of an Arc Flash Risk Assessment The Arc Flash Risk Assessment shall determine 130.5(1) – Appropriate safety related work practices Including Maintenance Switching – Arc Flash Boundary – PPE to be used within the arc flash boundary NEC 110-16 requir