480-V Distribution Arc Flash Updates PDF

480-V Distribution Arc Flash Updates
1022002
480-V Distribution Arc Flash
Updates
1022002
Technical Update, December 2011
EPRI Project Manager

T. Short
ELECTRIC POWER RESEARCH INSTITUTE

3420 Hillview Avenue, Palo Alto, California 94304-1338 PO Box 10412, Palo Alto, California 94303-0813 USA
800.313.3774 650.855.2121 [email protected] www.epri.com
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATIONS NAMED BELOW AS AN ACCOUNT
OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)

WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER

(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY

ITS TRADE NAME, TRADEMARK, MANUFACTURER, OR OTHERWISE, DOES NOT NECESSARILY
CONSTITUTE OR IMPLY ITS ENDORSEMENT, RECOMMENDATION, OR FAVORING BY EPRI.
THE FOLLOWING ORGANIZATIONS PREPARED THIS REPORT:
Electric Power Research Institute (EPRI)
Pacific Gas & Electric (PG&E)
This is an EPRI Technical Update report. A Technical Update report is intended as an informal report of
continuing research, a meeting, or a topical study. It is not a final EPRI technical report.
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail [email protected].
Electric Power Research Institute, EPRI, and TOGETHERSHAPING THE FUTURE OF ELECTRICITY
are registered service marks of the Electric Power Research Institute, Inc.
Copyright 2011 Electric Power Research Institute, Inc. All rights reserved.
ACKNOWLEDGMENTS
The following organizations prepared this report:
Electric Power Research Institute (EPRI)

1300 West W.T. Harris Blvd.
Charlotte, NC 28262
Principal Investigator
T. Short
Pacific Gas & Electric (PG&E)

3400 Crow Canyon Road
San Ramon, CA 94583
Principal Investigator
M. Eblen
This report describes research sponsored by EPRI.
The EPRI project team would like to thank PG&E and Marcia Eblen for helping to conduct arc
flash testing. EPRI also thanks Ralph Seban, Dan Kaufman, and Ryan Sparacino of PG&E for
their dedicated work.
This publication is a corporate document that should be cited in the literature in the following
manner:
480-V Distribution Arc Flash Updates. EPRI, Palo Alto, CA: 2011. 1022002.
iii
ABSTRACT
The 2012 version of the National Electrical Safety Code (NESC) has significant changes for
distribution arc flash on secondary networks (below 1000 V). Research was targeted to address
those changes, particularly at 480 V for spot networks and self-contained meters.
Spot networks can have very high incident energies. The 2012 NESC requires an analysis, and
many spot networks can have well over 100 cal/cm2 of exposure. Information exchange and
analysis have shown that work can still be done with a combination of proper arc flash suits and
work practices. In addition to deenergizing the primary side, there are other promising options,
including network protectors with relaying and network protectors with external fuses or
disconnects. Utilities can use this information to improve their arc flash programs for 480-V spot
networks.
Exploratory arc flash tests were performed on meter sockets, a current transformer (CT) cabinet,
and an overhead scenario to evaluate the thresholds of incident energy given in the NESCs new
table 410-1. For the CT cabinet and for overhead work, the tests validated the NESC thresholds.
For self-contained meter bases, results showed that higher rated meter sockets or multi-socket
units with significant bus bars can have significantly higher incident energies than those given in
the NESC table. Utilities can use this information to adjust their arc flash programs to better
protect workers working on this type of unit.
Keywords
Arc flash
Arcs
Faults
Overcurrent protection
Power distribution
Secondary networks
v
EXECUTIVE SUMMARY
In the 2012 National Electrical Safety Code, the approach for arc flash below 1000 V has
changed. There is no longer a blanket exclusion. Now, utilities will have to review arc flash on
distribution secondary equipment. The two main changes are at 480 V for metering and spot
networks. To address these issues, the work in 2011 concentrated on 480-V arc flash.
For 480-V spot networks, research concentrated on information exchange and practices to
manage arc flash in spot networks. The main findings are:
IEEE 1584 is the predominate calculation method with utilities assuming either an 18 or
24-inch (45.7 or 61.0-cm) working distance.
Many utilities are de-energizing the feeder. This reduces fault current and energized buswork
in protectors. Note that they are not operating a primary-side, oil switch.
Work is manageable in many spot networks with heavy arc suits. 100 cal/cm2 suits are
common.
Utilities normally assume either a self-extraction time or assume that internal network
protector fuses operate. Both assumptions allow work in many spot networks with available
arc flash suits. Both assumptions also have disadvantages.
External fuses or disconnects are a promising option to reduce incident energies and
completely de-energize a network protector. These scenarios can be treated as open-air
applications if the only exposure is line to ground.
For 480-V metering, several exploratory tests were performed on different meter styles to see if
there were any units where the 20 cal/cm2 threshold would not apply. Tests showed that meters
with significant internal busbar can have long durations and incident energies much higher than
20 cal/cm2. Figure ES-1 shows two meters that were tested. Both have significant bus bars at
tight spacings. In both units, very long duration events were measured. Figure ES-2 shows a test
on a 320-A meter base (similar to the unit on the left in Figure ES-1). The fault arced for 1.4 sec
before the lab protection tripped, and the incident energy at 18 inches (45.7 cm) was
86.9 cal/cm2. Previous tests by EPRI and PG&E were on single-socket, 200-A units. In these,
arcs self extinguished to limit incident energies. In the single-socket meters with less metal, arcs
burned enough metal to self clear. These tests formed the basis of the 20 cal/cm2 threshold in the
2012 NESC. Based on these new results, utilities should not work on 480-V meters with
significant busbar that are energized without an analysis.
vii
Figure ES-1
Meter Bases with Significant Busbar
Figure ES-2
Arc Flash in a 320-A Meter Base
Tests were also performed on overhead quadraplex lines (Figure ES-3) and on a CT cabinet
(Figure ES-4). In both cases, faults self-cleared quickly, validating the thresholds for each of
these that are in the 2012 NESC.
viii
Figure ES-3
Arc Flash on Overhead 480-V Secondary
Figure ES-4
480-V Secondary CT Cabinet
Because arc flash is so equipment dependent at 480 V, more testing may need to be done on
utility equipment if it does not fit the profile of equipment tested to date. For 480-V network
protector maintenance, future work could include testing of external fuses and/or disconnects to
verify that they will self clear quickly. In addition, more industry work is needed to evaluate
relaying and other options to reduce incident energies in spot networks.
ix
CONTENTS
1 INTRODUCTION ....................................................................................................................1-1
Previous Research ................................................................................................................1-2
2012 NESC ...........................................................................................................................1-4
2 UPDATE ON 480-V SPOT NETWORK PRACTICES ............................................................2-1

Summary of Interest Group Discussions...............................................................................2-1
Analysis of Calculation Assumptions.....................................................................................2-3
Single-Phase, 277-V Scenarios ............................................................................................2-6
Summary ...............................................................................................................................2-7
3 FAULT CURRENTS IN 480-V NETWORK PROTECTORS...................................................3-1

IEEE 1584 Fault Currents .....................................................................................................3-1
Fault Currents from PG&E 480-V Network Protector Tests ..................................................3-4
4 480-V METER SOCKET TESTS ............................................................................................4-1

Review of Previous Meter Socket Testing.............................................................................4-1
Another 200-A Meter Socket Variation..................................................................................4-5
320-A Meter Sockets.............................................................................................................4-8
Multi-Meter Sockets.............................................................................................................4-18
Analysis of Meter Socket Test Results................................................................................4-28
5 480-V OVERHEAD TESTS ....................................................................................................5-1

Secondary Quadraplex Tests at 25 kA..................................................................................5-1
Analysis of Overhead Secondary Tests ................................................................................5-9
6 480-V CT CABINET TESTS ...................................................................................................6-1
7 FUTURE WORK .....................................................................................................................7-1
8 REFERENCES .......................................................................................................................8-1
xi
LIST OF FIGURES
Figure ES-1 Meter Bases with Significant Busbar ........................................................................ x

Figure ES-2 Arc Flash in a 320-A Meter Base.............................................................................. x
Figure ES-3 Arc Flash on Overhead 480-V Secondary ............................................................... xi
Figure ES-4 480-V Secondary CT Cabinet.................................................................................. xi
Figure 2-1 Spot Network Example Showing Infeeds to the Network Protector in the Top
Left .....................................................................................................................................2-4
Figure 2-2 Tight Spacings in a Sidewalk Vault ..........................................................................2-5
Figure 2-3 Example of a Network Protector with External Fuses ..............................................2-6
Figure 2-4 Example of a Fault Initiated under a Rubber Cover in the 480-V Compartment
of a Transformer.................................................................................................................2-7
Figure 3-1 Cumulative Distribution of Arcing Currents for the IEEE 1584 Test Data.................3-2
Figure 3-2 Arcing Current versus Bolted Current from the IEEE 1584 Formula ........................3-3
Figure 3-3 Arcing Current versus Bolted Current for the IEEE 1584 Test Data.........................3-3
Figure 3-4 Network Protector Test Setup ..................................................................................3-4
Figure 3-5 Cumulative Distribution of Arcing Currents for the PG&E 480-V Network
Protector Tests...................................................................................................................3-5
Figure 3-6 Cumulative Distribution of 99th Percentile RMS Currents for the PG&E 480-V
Network Protector Tests.....................................................................................................3-6
Figure 3-7 Test 278 Current Waveforms ...................................................................................3-7
Figure 3-8 Test 278 RMS Currents ............................................................................................3-8
Figure 3-9 Test 268 Waveforms ................................................................................................3-9
Figure 3-10 Test 268 RMS Currents........................................................................................3-10
Figure 3-11 Test 268................................................................................................................3-10
Figure 3-12 Test 268 Sequence Currents................................................................................3-11
Figure 3-13 RMS Currents versus Negative-Sequence Currents ............................................3-12
Figure 3-14 Cumulative Distribution of 99th Percentile Negative-Sequence Currents .............3-12
Figure 4-1 Peak Incident Energy as a Function of Fault Current for 480-V Meter Bases ..........4-2
Figure 4-2 200-A Meter Socket Styles Previously Tested .........................................................4-3
Figure 4-3 Peak Incident Energy Comparison on 200-A Meter Sockets at 18 .........................4-4
Figure 4-4 Fault Duration Comparison on 200-A Meter sockets................................................4-4
Figure 4-5 Fireball Comparison of Ringed and Ringless Meter Socket Arc Flash .....................4-5
Figure 4-6 200-A Milbank West 127TP Meter Socket................................................................4-6
Figure 4-7 Milbank 127TP Fault Initiation Points .......................................................................4-6
xiii
Figure 4-8 Damage after Test 3.................................................................................................4-7
Figure 4-9 Fault Current Waveforms .........................................................................................4-8
Figure 4-10 Incident Energies....................................................................................................4-8
Figure 4-11 320-A Milbank U2594 Meter Socket.....................................................................4-10
Figure 4-12 Calorimeter Layout ...............................................................................................4-11
Figure 4-13 320-A Meter Socket Prior to Test 6 with the Fault Initiation Point Shown ............4-12
Figure 4-14 Movie Frames from Test 6....................................................................................4-13
Figure 4-15 Infrared-Filtered Movie Frames from Test 6 .........................................................4-14
Figure 4-16 Current Waveforms from Test 6 ...........................................................................4-15
Figure 4-17 Damage after Test 6.............................................................................................4-16
Figure 4-18 Incident Energies by Calorimeter Position for Test 6 ...........................................4-17
Figure 4-19 Fault Currents from Test 4....................................................................................4-17
Figure 4-21 Three-Position Meter Socket ................................................................................4-19
Figure 4-22 Four-Position Meter Socket ..................................................................................4-20
Figure 4-23 Three-Position Meter Socket in Position for Test 7 ..............................................4-21
Figure 4-24 Fault Initiation Point on the Three-Position Meter Socket for Test 7 ....................4-22
Figure 4-25 Infrared-Filtered Movie Frames from Test 7 .........................................................4-22
Figure 4-26 Infrared-Filtered Movie Frames from Test 7 (continued) ......................................4-23
Figure 4-27 Fault Currents for Test 7.......................................................................................4-24
Figure 4-28 Incident Energies by Calorimeter Position for Test 7 ...........................................4-24
Figure 4-29 Test 7 Damage .....................................................................................................4-25
Figure 4-30 Fault Initiation for Tests 10 ...................................................................................4-26
Figure 4-33 Fault Currents for Test 12.....................................................................................4-27
Figure 4-34 Test 12 Damage ...................................................................................................4-28
Figure 4-35 Arcing Durations by Meter Type ...........................................................................4-29
Figure 4-36 Maximum Incident Energy by Meter Type ............................................................4-30
Figure 4-37 Maximum Incident Energy Rate by Meter Type....................................................4-30
Figure 5-1 Overhead Secondary Test Set-up for Test 13..........................................................5-2
Figure 5-2 Setup for Test 13 ......................................................................................................5-3
Figure 5-3 Line-to-Neutral Fault Initiation Prior to Test 13 .........................................................5-3
Figure 5-4 Damage after Test 13...............................................................................................5-4
Figure 5-5 Test 13 Fault Current................................................................................................5-4
Figure 5-6 Test 13......................................................................................................................5-5
Figure 5-7 Test 13 Incident Energies.........................................................................................5-5
Figure 5-8 Phase-to-Phase Fault Initiation for Test 14 ..............................................................5-6
Figure 5-9 Test 14......................................................................................................................5-7
xiv
Figure 5-10 Fault Currents during Test 13.................................................................................5-7
Figure 5-12 Test 14 Incident Energies.......................................................................................5-8
Figure 5-13 Test 14 Infrared-Filtered Video Frames..................................................................5-9
Figure 5-14 Incident Energy and Current-Duration Product for 480-V Overhead Tests ..........5-10
Figure 6-1 CT Cabinet ...............................................................................................................6-2
Figure 6-2 Bottom Distances in the CT Cabinet ........................................................................6-3
Figure 6-3 Test 20 Video Frames through and Infrared-Passing Filter......................................6-4
Figure 6-4 Fault Initiations .........................................................................................................6-5
Figure 6-5 Fault Currents...........................................................................................................6-6
Figure 6-6 Test 21......................................................................................................................6-7
Figure 6-7 Damage after Test 21...............................................................................................6-7
Figure 6-8 Infrared Filtered Video Frames from Test 21............................................................6-8
Figure 6-9 Calorimeter Values by Position from Test 21 ...........................................................6-9
xv
LIST OF TABLES
Table 1-1 NESC 2012 Table......................................................................................................1-5

Table 2-1 Summary of Approaches and Options Used by Several Utilities ...............................2-2
Table 2-2 Comparison of Two Sets of Assumptions for Network Protector Calculations ..........2-4
Table 4-1 Summary of 2011 Testing of Meter Sockets at a 25.7-kA Bolted Fault Current ........4-7
Table 5-1 Summary of Overhead Quadraplex Tests at 480 V...................................................5-7
Table 6-1 Summary of Overhead Quadraplex Tests at 480 V...................................................6-7
xvii
1
INTRODUCTION
The arc flash research in 2011 was focused on secondary work, mainly to help utilities prepare
for changes based on the new National Electrical Safety Code (NESC) [IEEE C2-2012]. The
main changes in the 2012 edition of the NESC are for voltages less than 1000 V. For distribution
systems, the two main changes are 480-V metering and 480-V spot networks. 480-V metering
work will require heavier clothing, and spot networks will require an analysis to determine
protective levels and incident energies can be very high.
Because of utility concern with 480-V arc flash, research was focused on utility practices for
480-V arc flash and equipment tests at 480 V. For 480-V spot networks, we reviewed industry
practices for analysis, data collection, work assumptions, and options for protecting workers and
reducing incident energies. We also identified several 480-V scenarios for test to expand the
database of arc flash in 480-V equipment.
Testing by PG&E and EPRI has found that at 480 V, arc flash is very equipment dependent. The
main factor is whether arcs will self-sustain or whether they will self-extinguish. Another factor
is the box and electrode geometry, which determines how the arc energy projects out of the box.
In equipment with relatively wide electrode spacings and open space in a compartment, arcs
cannot sustain. In 480-V padmounted transformers, arcs could not sustain. For equipment with
tight spacings and with substantial buswork, faults can sustain until externally tripped. Such
situations includes network protectors and some large panelboards. Meter sockets are in between.
Depending on fault currents, arcs can sustain, but they will normally self clear when electrodes
or incoming leads have burned clear.
At 240 V and below, arcs are much more difficult to sustain. In tests on self-contained meters of
the ringed design, PG&E could not find any sustained arcing and very low incident energies for
three-phase, 208-V meter faults.
At 480 V, we identified several areas worth further testing:

Meter sockets with internal buswork
Open air triplex
CT cabinets
All of the meter sockets tested by EPRI and PG&E have been 200-A, single-socket units. Some
320-A units and multi-socket units have busbar and tight spacings where faults may sustain
longer than the single-socket units tested. We were interested if incident energies could exceed
the 20-cal/cm2 threshold given in the 2012 NESC. As we will see, cases were found with
1-1
Introduction
incident energies much higher than this in units with internal busbar. Open air is also specified in
the NESC table. Quadraplex cable was tested as it represents an overhead application with
conductor-to-conductor spacings that are as close as possible. In all tests, incident energies were
less than 4 cal/cm2. CT cabinets are another utility scenario of interest. Based on previous tests
on padmounted transformer secondary compartments, we did not expect arcs to sustain, but the
geometries are different enough for testing to be worth exploring. Tests on a CT cabinet largely
confirmed the thresholds in the NESC.
For spot networks, research largely focused on information gathering from utilities on
approaches to reducing incident energies. This included hosting meetings of an EPRI arc flash
interest group. Utilities have generally found high incident energies in 480-V spot networks.
Utilities generally approach analysis in similar ways, but there are key differences in
assumptions. We will discuss tradeoffs in assumptions along with options for reducing incident
energies.
Previous Research
Previous EPRI research has included evaluation of arc flash analysis approaches, arc flash tests
at low and medium voltages, and exchange of utility experience and practices. This work is
documented in the following reports:
Distribution Arc Flash: Analysis Methods and Arc Characteristics

3/18/2009
http://my.epri.com/portal/server.pt?Abstract_id=000000000001018693
Evaluation of modeling and calculation approaches for utility applications
Results of 480-V tests on meters and panel boards
Overhead medium-voltage arc flash test results
Arc characteristics from monitoring
General approaches to arc flash analysis
Identification of future work
Distribution Arc Flash: Industry Practices

3/23/2009
Utility case studies
Industry statistics
Utility industry survey: sections on calculation approaches, arc energy mitigation,
PPE, and work rules
1-2
Introduction
Distribution Arc Flash: 480-V Padmounted Transformers and Network Protectors

8/31/2009
480-V padmounted transformer test results
480-V network protector test results
Distribution Dispatch, Fourth Quarter, 2009
12/14/2009
Arc flash changes on the horizon
Arc Flash FAQs
Arc Flash Phase II Work Practices Survey Report
12/18/2009
Utility survey
Update of 2008 survey; new emphasis on secondary networks
Distribution Arc Flash: Phase II Test Results and Analysis

3/8/2011
208-V network protector tests
Options for reducing incident energies in 480-V spot networks
208-V meter tests
Medium-voltage arc-in-a-box tests
Overhead medium-voltage arc flash test results with incident energies
PMH padmounted switch test results and formula for calculating incident energies
More fabric testing
Options for more advanced arc flash simulations
1-3
Introduction
Arc Flash Testing of Clothing and 480-V Meter Sockets

6/15/2011
480-V testing of 200-A ringless meter sockets
Face shield and arc hood testing
2012 NESC
The NESC table 410-1, new for the 2012 edition, is partly based on testing by PG&E and EPRI.
This modifies the blanket exemption to the requirement to do an arc flash study below 1000 V,
and replaces it with an exemption based on equipment. Table 4-1 shows NESC protective
requirements for given equipment and voltages in lieu of doing a study.
1-4
Introduction
Table 1-1
NESC 2012 Table
Table 410-1: Clothing and clothing systems (cal /cm2) for voltages 50 to 1000 V ac1 (See Rule 410A3)
Nominal Voltage Range and Calories/Cm2

Equipment Type 50 250 V 251 600 V14 601 1000 V
Self-contained meters /
42 204 308
cabinets
Pad-mounted transformers 49 49 68
CT meters and control wiring 42 45 68
Metal-clad switchgear / motor control

83 406 608
centers
Pedestals, pull boxes, hand holes 42 87 128
Open air (includes lines) 42 47 68
Network Protectors 410 Note 11 Note 11
Panel boards - single phase (all) / three

42 812 128
phase (<100 A)
Panel boards three phase
42 Note 13 Note 13
(> 100 A)
1. This table was developed from fault testing based on equipment type and is independent of fault current
unless otherwise noted.
Calculations and test data are based on an 18-in separation distance from the arc to the employee. (See
IEEE 1584-2002.)
Other methods are available to estimate arc exposure values and may yield slightly different but equally
acceptable results.
The use of the table in the selection of clothing is intended to reduce the amount or degree of injury but
may not prevent all burns.
2. Industry testing on this equipment by two separate major utilities and a research institute has demonstrated
that voltages 50 250 V will not sustain arcs for more than 2 cycles thereby limiting exposure to less than
4 calories/cm2. (See Ref 1)
3. Value based on IEEE 1584 formula for Motor Control Centers. [Gap = 2.54 cm (1 in)] (Xd = 1.641) [46
cm (18 in) distance] 51 kA (Based on a 208 V, 1000 kVA, 5.3% Z, served from a 500 MVA system)
Maximum duration without circuit protective device operation from industry testing (See Ref. 1) is 10
cycles: 46.5 cal/s/cm2 * 0.167 sec = 7.8 cal/cm2
4. Industry testing on 480 V equipment indicates exposures for self-contained meters do not exceed 20
cal/cm2.
1-5
Introduction
5. Industry testing on 480 V equipment indicates exposures for CT meters and control wiring does not exceed
4 cal/cm2.
6. Value based on IEEE 1584 formula for Motor Control Centers. [Gap = 2.54 cm (1 in)] (Xd = 1.641)[46
cm (18 in) distance] 12.7 kA at 480 V (worst-case energy value from testing). (See Ref. 2). Maximum
duration without circuit protective device operation from tests is 85 cycles: 26.2 cal/s/cm2 * 1.42 sec = 37
cal/cm2
7. Incident analysis on this equipment indicates exposures do not exceed the values in the table.
8. Engineering analysis indicates that applying a 150% multiplier to the 480 V exposure values provides a
conservative value for equipment and open air lines operating at 601 V to 1000V.
9. Industry testing on 480 V equipment indicates exposures on pad-mounted transformers do not exceed
4 cal/cm2. (See Ref. 2)
10. Industry testing on 208 V network protectors indicates exposures do not exceed 4 cal/cm2. (See Ref 1)
11. Industry testing on 480 V network protectors indicates arcs will not self- extinguish and heat flux rates will
exceed 60 cal/cm2/sec at 24 inch working distance. Perform arc hazard analysis. (See Ref 2).
12. Industry testing on 480 V panels with non-edge mounted bus bars indicates exposures do not exceed
8 cal/cm2. (See Ref 2)
13. Industry testing on panelboards with edge mounted, parallel bus bars indicates arcs will not self extinguish
and heat flux rates will exceed 60 cal/cm2/sec at 18 inch working distance. Perform arc hazard analysis.
(See Ref 2)
14. IEEE1584 original test data indicates there is no significant difference between heat flux rates for 400 V
class equipment verses 600 V class equipment.
Table References
1. 208-V Arc Flash Testing in Network Protectors. EPRI, Palo Alto, CA: 2009. March 26, 2010.
2. Eblen, M. L. and Short, T. A., Arc Flash Testing of Typical 480V Utility Equipment, IEEE Industry
Applications Society-Electrical Safety Workshop Paper No. ESW2010-05.
1-6
2
UPDATE ON 480-V SPOT NETWORK PRACTICES
480-V spot networks involve some of the highest potential incident energies due to high fault
currents and relatively long clearing times. Network protector maintenance is particularly
important to address. Continuing previous evaluations in EPRI 1022697, we update industry
practices related to spot networks.
The 2012 NESC requires a study for 480-V spot networks. We review industry analysis
calculation approaches and assumptions used. Because calculation approaches show high
incident energies for many spot network configurations, we also update industry options to
reduce energies.
Summary of Interest Group Discussions
In 2011, EPRI hosted several web meetings as part of an Arc Flash Interest Group. Two of these
meetings focused on practices in 480-V spot networks. Table 2-1 summarizes several of the key
assumptions and solutions options considered for 480-V networks based on the most recent
interest group meeting. The main differences between utilities centered around the working
distance and whether to use a self-extraction time.
The two main working distances used for calculations are 18 inches (45.7 cm) and 24 inches
(61 cm). For the utilities that used 24 inches, they either modified or confirmed that their tools
and working procedures were compatible with the 24-in distance. Distance is important because
incident energies are approximately 35% lower at 24 in (61 cm) than at 18 in (45.7 cm).
Most utilities are using or considering heavy flash suits. The heaviest suits available from most
suppliers range from 100 to 140 cal/cm2.
The majority of utilities are de-energizing the feeder of the network protector under service. In
addition to lowering the current supplied to the fault, de-energizing the primary side de-energizes
a significant portion of the buswork in a network protector. This is the major change seen from
earlier meetings. Note that utilities are de-energizing the feederthey are not operating a
primary-side switch. As utilities reported numerous times, operating a primary-side, oil-filled
switch is hazardous. Utilities have reported bad accidents caused by failures during switching.
De-energizing the feeder is normally done with SCADA. Some utilities are investigating the use
of vacuum interrupters for installation on the primary side of the network transformer to avoid
tripping at the feeder level.
2-1
Update on 480-V Spot Network Practices
Table 2-1
Summary of Approaches and Options Used by Several Utilities
Self De-
Working
Utility extraction energize Clothing Solution options considered
distance
time used feeder
A 24" Yes Yes 100 cal/cm2 CM-52, some with ARMs
100 or 140* Looking at smaller fuses; put in CM-52

B 18" No Yes
cal/cm2 Oberon with rimrack option and ARMs
C No
External CLF fuses for large spots;

D 18" No Yes 100 cal/cm2
external disconnects for smaller spots
E Yes
F 24" * *
G 18" No 65 cal/cm2 Looking at smaller fuses
H Yes
Looking at external disconnects for

I Yes the low-voltage side and a primary-
side switch
External fuses for new units; de-

J 24" Yes Yes 100 cal/cm2 energize adjacent transformers;
retooled to get 24" working distance
K 18"
L 18" Yes No
* = still under review
There were several themes common with several utilities:

IEEE 1584 was used for modeling, either using the IEEE 1584 software or a short-circuit
program that included IEEE 1584 calculations for arc flash.
External fuses in single-phase enclosures could be removed one at a time. Exposure would be
treated as short-duration or as an overhead scenario, leading to single-layer FR clothing.
As an alternative to tripping the feeder, utilities are looking at an Elastimold MVI interruptor
or a Kearney/Cooper interrupter for the high side.
Several are looking at retrofitting for external fuses.
2-2
Utilities raised several issues during discussions:

How do you get a reference voltage for protector maintenance if you trip the primary?
o One utility will clip onto the load side of protector fuse holder or removable link
bus in single-phase external compartments. Another utility will use a fused
receptacle on the wall.
Are there issues with 100-cal/cm2 suits with dark visors or fogging?
In the Eaton CM-52, only current-limiting fuses can be used in external fuse housings.
Current-limiting fuses will not coordinate for double phase-to-ground faults.
Aluminum collector bus (IWCB) and 480-V cables and splices were both raised as other
areas of possible concern.
Current-limiting fuses are not submersible, so they need to be in a box.
Does the fan noise in 100-cal/cm2 suits impede the ability to hear alarms?
o Three other utilities mentioned that they rely on other employees to monitor the
alarms.
Is there a concern with using cooling packs which are plastic and compatibility of plastic
with arc flash?
Could a 277-V single-phase scenario be treated like the 250 V column in NESC table 410-1
for future revisions of the NESC?
Would working in a vault near a bus be an open-air scenario?
Analysis of Calculation Assumptions
Most utilities calculate arc flash in spot networks using similar approaches. The main differences
center on duration. The two main camps assume one of the following:
Option 1: Internal network protector fuse operates The fault occurs on the transformer side
of the network protector fuses. The fuses operate to clear the fault.
Option 2: Self-extraction time Most commonly taken as two seconds, this assumes that a
worker can move back or to the side to limit exposure.
Table 2-2 shows IEEE 1584 calculation examples showing each of these options. Both options
are evaluated for 1500-kVA, 7% transformers and a 24-inch working distance. Both examples
assume that the feeder is open on the network protector being serviced.
2-3
Table 2-2
Comparison of Two Sets of Assumptions for Network Protector Calculations
Option #1 Option #2
Transformer Internal Z-37.5 fuse clears 2-sec self extraction
units in Bolted Duration, Duration,
parallel cal/cm2 cal/cm2
current, kA sec sec
1 unit 25.7 3.8 113.5 2 59.7
2 units 51.4 1.1 61.3 2 111.5
3 units 77.1 0.4* 32.1 2 160.7
* Duration estimated the current is beyond whats given on the fuse curve.
1 2 3
10 kA 30 kA
10 kA
1500 kVA
480 V Assume arcing current = 40% of bolted current
Figure 2-1
Spot Network Example Showing Infeeds to the Network Protector in the Top Left
The most striking difference between the assumptions is how results vary with more units in
parallel. With the 2-sec self-extraction time, incident energies increase with more units in
parallel because of the constant duration assumption. With the fuse clearing assumption, incident
energies decrease with more units in parallel because faults clear more quickly with more fault
current. If we assume a flash suit of 100 cal/cm2, both options have workable scenarios, and both
options have unworkable scenarios, but unworkable scenarios occur in different situations.
The drawback to the internal fuse assumption is that the internal fuses may not operate if the
fault starts above the fuses (fault position 3 in Figure 2-1) or propagates above the fuses.
Regardless of where the fault starts, the magnetic forces will push the arcs downward in the
2-4
network protector (this matches the assumption). If arcs last as long as those in Table 2-2 (0.4
sec or more), significant damage will occur to the bus bars and other hardware in the enclosure.
The entire enclosure will be enveloped in hot ionized gas. Restrikes may occur on the load side
of the fuses even if the fuses operate. Also, fuses may not interrupt as expected. Cholewinski and
Davis [2011] give an example where Y-25 fuse links did not operate as expected based on
manufacturers published fuse curves. Their follow-up tests showed differences between fuse
characteristics and the manufacturers fuse curves. Clearing time is dependent on fuse
characteristics. The example in Table 2-2 is for just one size and type of fuse. Larger and/or
slower fuses will allow more incident energy. Current limiting fuses have faster clearing times at
higher currents, but they can also fail to clear for a wide range of lower-level fault currents (the
fuse curves are steeper). Analyzing each fuse / spot network combination also increases
complexity of data collection and calculations.
The drawback to the self-extraction assumption is that the worker may not be physically able to
move. If conditions are cramped as in the sidewalk vault in Figure 2-2, workers may not have
room to move. In addition, if arcing occurs for many seconds or minutes in a vault, the worker
may not be able to escape the confined space before the air supply and visibility are completely
compromised.
Source: W. Deal, Distribution Network 480 V NESC 410A3 Arc Flash Compliance,
EPRI North American Dense Urban Working Group Meeting, St. Louis, MO, 2011.
Figure 2-2
Tight Spacings in a Sidewalk Vault
Because both the self-extraction and the fuse-clearing assumptions have drawbacks, the industry
should consider other ways to reduce hazards, including relaying options, remote racking, and
external disconnects or fuses.
2-5
Single-Phase, 277-V Scenarios
One issue with the 2012 NESC is that it does not directly address single-phase scenarios at
480 V (277 V line to ground) in most equipment. In spot networks, the main question is: what is
the exposure for removing external fuses in their own compartment (like those in Figure 2-3)?
Each fuse is accessed separately. The enclosures are nonconducting, so the phase-to-ground
flashover path is long. There is no exposed phase-to-phase flashover path.
For single-phase, 277 V (L-G) applications like the external fuse, arcs will self clear quickly. In
all 480-V testing we have done, to have sustained arcing, you need (1) relatively close electrode
to electrode spacings, either phase to phase or phase to ground and (2) arc confinement. If the arc
has room to push into the open, it will lengthen and self extinguish. If it is more confined, arcing
is likely to continue. In addition, with a single-phase arc, the arc will extinguish briefly and cool
at every half cycle. Arcing sustains if the arc restrikes after the current zero crossing. With a
three-phase fault (with two or three arcs), heating is continuous, so arcing is more likely to
sustain.
Courtesy of PG&E
Figure 2-3
Example of a Network Protector with External Fuses
In reviewing table 410-1 of the 2012 NESC (Table 1-1 of this report), it is somewhat unclear as to
how to interpret this work scenario. For most equipment, the table does not distinguish between
single-phase and three-phase scenarios. Presumably, three-phase equipment and exposure is
assumed, since the testing that formed the basis of the table was almost all three phase.
EPRI and PG&E have done some tests with line-to-ground fault initiations. In all cases, faults
cleared quickly unless the fault jumped to multiple phases. See chapter 5 for examples on
overhead secondary. Figure 2-4 shows the setup for one test in the secondary compartment of a
padmounted transformer. Faults were applied from line to ground with either a fuse wire or a
pair of vice grips. The faults self-cleared in 1.5 cycles or less.
2-6
Figure 2-4
Example of a Fault Initiated under a Rubber Cover in the 480-V Compartment of a
Transformer
For analysis purposes, we recommend treating scenarios like this as open air as long as exposure
is only phase to ground. That corresponds to single-layer FR clothing of 4 cal/cm2.
Summary
Overall, 480-V network protector maintenance is tricky from the point of view of arc flash.
Depending on assumptions, maintenance can be performed in many cases with heavy flash suits
(until the network protector is de-energized). Because incident energies can be well over 100
cal/cm2 in some cases, other options to reduce incident energies should continue to be explored.
External fuses or disconnects are options. Relay schemes to trip network protectors are also
worth more investigation; options could include fiber optics, heat sensors, or forward fault
current.
2-7
3
FAULT CURRENTS IN 480-V NETWORK PROTECTORS
If relaying solutions are used for 480-V network protectors, the fault current and arcing
characteristics of these faults may be an important consideration. For relaying, the minimum
arcing current from fault events may limit what relay settings can be used. The arcing current is
variable and depends on the fault current and the length of the arcs. The length of the arcs
depends on equipment and bus configuration as well as where a fault may be initiated. This
section explores low-voltage arcing currents, specifically aimed at faults in network protectors.
We evaluate arcing currents from IEEE 1584 tests and EPRI tests on network protectors.
IEEE 1584 Fault Currents
IEEE 1584 provides the following equation for estimating arcing current from the bolted fault
current:
log10 I a K 0.662 log10 I bf 0.0966V 0.000526G 0.5588V log10 I bf 0.00304G log10 I bf
where,
Ia, arc current, kA
Ibf, bolted fault current, kA
K, 0.153 for open configurations
0.097 for box configurations (enclosed equipment)
V, system voltage, kV
G, distance between buses, mm
For low-voltage switchgear, IEEE 1584 provides a default distance between buses (G) equal to
32 mm.
The IEEE 1584 equation for arcing current was developed based on test data from several
laboratories. Figure 3-1 shows cumulative distributions of arcing currents as a percentage of the
bolted fault current for all tests below 1000 V. Most of these were arc-in-a-box tests with vertical
electrodes; 60% had open-circuit voltages of 600 V, and about 25% had open-circuit voltages of
400 V. Note the wide range of values for arcing current between different testing groups.
3-1
Fault Currents in 480-V Network Protectors
1.0
0.8
less than the value given
Portion of xaxis values
0.6
Paper Vol 361
Kinectrics 2001
0.4
Paper 9735
0.2
0.0 Square D 2002
20 40 60 80
Arcing current as a percentage of bolted current
Figure 3-1
Cumulative Distribution of Arcing Currents for the IEEE 1584 Test Data
Theoretically, as a percentage of bolted currents, arcing currents should drop as the bolted
available fault current increases. As available fault current increases, the arcing current increases.
With more arcing current along the same path, the resistance across the arc increases relative to
the source impedance. For 480-V switchgear, the IEEE 1584 formula predicts the arcing current
to bolted current relationship shown in Figure 3-2, and Figure 3-3 shows data from the IEEE
1584 data set. For 480-V spot networks, network protectors, fault currents can be quite high, so
this relationship is important.
3-2
55

50
45
40
50 100 150 200
Bolted fault current, kA
Figure 3-2
Arcing Current versus Bolted Current from the IEEE 1584 Formula
80
CLF Paper
KEMA 2001
60 Kinectrics 2001
Paper 9735
Paper Vol 361
Square D 2002
40
0 20 40 60 80 100
Bolted fault current, kA
Figure 3-3
Arcing Current versus Bolted Current for the IEEE 1584 Test Data
3-3
Fault Currents from PG&E 480-V Network Protector Tests
In this section, we evaluate arcing currents from EPRI 480-V network protector tests at PG&Es
San Ramon test facility [Eblen and Short, 2010; EPRI 1020210, 2009]. Figure 3-4 shows the
network protector used during tests. The internal operating mechanisms have been stripped from
the unit. The unit is energized from the top, which is the network side of the unit. A common
work procedure is removing the fuses on the network feed. The unit is fed by PG&Es 480-V
source thats capable of supplying a bolted fault current of 44 kA.
Figure 3-4
Network Protector Test Setup
Figure 3-5 shows the statistical of arcing currents as a percentage of bolted current for all tests
with sustained arcing on 480-V network protectors. The values for arcing current came from the
average of the three phases. All led to three-phase faults, event if they were initiated phase-to-
phase or phase to ground. Each phase current was the average value during the event.
3-4
1.0

0.8
Portion of xaxis values 0.6
0.4
0.2
0.0
40 50 60 70
Figure 3-5
Cumulative Distribution of Arcing Currents for the PG&E 480-V Network Protector Tests
Because of the variability of arcing currents, the peak or near-peak currents during an event are
of interest. Figure 3-6 shows the 99th percentile current of all three phases during the fault event.
This is based on a rolling one-cycle rms evaluation of the waveform that only includes the 60-Hz
component. If an instantaneous relay element is used, the statistical distribution in Figure 3-6
may be more appropriate than that shown in Figure 3-5.
3-5
1.0

0.8
0.6
0.4
0.2
0.0
50 60 70 80 90 100
Figure 3-6
Cumulative Distribution of 99th Percentile RMS Currents for the PG&E 480-V Network
Protector Tests
Most of these tests were done with a bolted fault current of 44 kA. Because bolted fault currents
in spot networks are often well above this, we may expect that the arcing fault currents will be
somewhat lower than the distribution of currents in Figure 3-6.
Figure 3-7 shows current waveforms from test 278, the event with the lowest percentage arcing
current given in Figure 3-5. Because of the intermittency of the arcing, the overall averages of
currents on each phase shrank artificially. Figure 3-8 shows rolling rms currents based on these
waveforms. For a fast-acting relay, a trip setting of 50% of the bolted current would catch the
fault on two of the phases, and phase B has a peak value above 40% of the bolted fault current.
3-6
Arcing currents, kA
20
10
0
10
A
20
30
40
20
0
B
20
40
60
20
0
C
20
40
60
0.0 0.1 0.2 0.3 0.4 0.5
Time, sec
Figure 3-7
Test 278 Current Waveforms
3-7
Arcing current in percent of bolted current
60
50
40
A
30
20
10
0
40
30
B
20
10
0
50
40
30
C
20
10
0.0 0.1 0.2 0.3 0.4 0.5
Time, sec
Figure 3-8
Test 278 RMS Currents
Figure 3-9 shows waveforms from test 268, the test with the lowest average arcing current on a
percentage basis. Figure 3-10 shows rms currents from test 268.
The probable reason for the low arcing currents is that this test had the longest distance between
the network protector bus bars and a ground plane added below. In most tests, this distance was
two inches (5 cm) or less, but in this test, it was six inches (15.2 cm). Figure 3-11 shows the
configuration for test 268 along with an infrared video frame captured during the event. The
distance between the bus bars and the bottom ground plane impacts the arc lengths. With longer
arcs, the arc voltage is higher, and the arc resistance is higher. As arc voltage increases, the
arcing current decreases. This event probably has longer arc lengths than would exist in a typical
network protector.
3-8
Arcing currents, kA
40
20
A
20
20
0
B
20
40
40
20
C
20
0.0 0.1 0.2 0.3 0.4 0.5
Time, sec
Figure 3-9
Test 268 Waveforms
3-9
Arcing current in percent of bolted current
80
60
40
A
20
60
40
B
20
0
70
60
50
40
C
30
20
10
0
0.0 0.1 0.2 0.3 0.4 0.5
Time, sec
Figure 3-10
Test 268 RMS Currents
Figure 3-11
Test 268
3-10
Figure 3-12 shows sequence currents for test 268. The negative-sequence current is an interesting
option for relaying solutions. Although more chaotic than the phase currents, a negative-
sequence element may be more sensitive than a phase relay because negative-sequence currents
are normally small.
Figure 3-13 shows how negative-sequence currents compare to rms currents. Each point in this
graph shows the 99th percentile rms current for the three phases of that event versus the 99th
percentile negative-sequence current for that event. Figure 3-14 shows a cumulative distribution
of negative-sequence currents. Based on that, a negative-sequence relay element set to 10% to
15% of the bolted fault current could catch almost all arc flash events.
Sequence currents in percent of bolted current
50
40
30
s1
20
10
15
10
s2
0
5
3
s0
0.0 0.1 0.2 0.3 0.4 0.5
Time, sec
Figure 3-12
Test 268 Sequence Currents
3-11
negativesequence current, kA
15
99th percentile
10
20 25 30 35 40 45
99th percentile rms current, kA
Figure 3-13
RMS Currents versus Negative-Sequence Currents
1.0
0.8
0.6
0.4
0.2
0.0
15 20 25 30 35 40
Negativesequence current as a percentage of bolted current
Figure 3-14
Cumulative Distribution of 99th Percentile Negative-Sequence Currents
3-12
When using this data for relaying solutions for network protector applications, keep the
following in mind:
The network protector tests were done on a unit with the guts removed. They were designed
to mimic faults near the fuses at the top. Arcs in complete network protectors may have
different arc lengths, and this depends on where the fault initiates.
Only bolted fault currents up to 52 kA were tested. Arcing currents may drop with higher
available fault currents. Based on the IEEE 1584 formula, increasing the bolted fault current
from 50 to 150 kA, reduces the arcing current from 48% to 40% of the bolted value.
Because of the nonlinearity and variability of arcing currents, one must be careful when applying
relaying solutions. Some options to check include:
Use a relay test set to feed arcing current waveforms into a relay to make sure relays respond
as expected given the nonlinearities.
Test relaying solutions in an offline mode to see if they false trip from load, switching, or
other events.
3-13
4
480-V METER SOCKET TESTS
For the 2012 edition of the NESC [IEEE C2-2012], the requirements for arc flash under 1000 V
were changed. An arc flash study is not required if the equipment and voltage meet the
requirements of table 410-1. For self-contained meters, a study is not required if a 20-cal/cm2
protective level is provided. This value was largely based on prior PG&E and EPRI testing
where faults would self extinguish in self-contained meter sockets at 480 V. Units tested
previously were 200-A, single-socket meter bases. Self clearing of faults happens largely
because the metal in the equipment burns back enough such that spacings are large enough that
arcs tend to self clear. Smaller boxes and more metal are more likely to have longer durations.
In this chapter, we describe exploratory testing on meter sockets with more substantial bus bars
that may prolong fault durations before self clearing. Multi-socket meter bases and higher-rated
meter bases have substantial bus bars with the potential for longer arcing durations than we
found with single-socket, 200-A meter bases. As we will see in this chapter, the multi-socket and
higher-rated meter bases did allow longer arcing and higher incident energies. These are
situations where utilities should consider de-energizing for work or providing other protection.
Review of Previous Meter Socket Testing
In 2008, PG&E tested many 480-V meter sockets. The maximum incident energies from 40 tests
are summarized in Figure 4-1. In all cases, the faults self cleared, and the duration was a key
component that limited incident energies. The worst case was at a bolted fault current of 12 kA
(with typical average arcing currents of 7 kA). At higher currents, faults self-cleared faster. For
more information on these meter base tests, see EPRI 1018693 [2009] and Eblen and Short
[2010]. These test results formed the bases for NESC subcommittee 8 to recommend a minimum
of 20 cal/cm2 for work on self-contained 480-V meter bases. Since then, PG&E has gathered
data from tests of additional self-contained meter sockets.
4-1
480-V Meter Socket Tests
16S Meter Base Tests, 18"
25
20
Bolted Fault
2
Heat Energy cal/cm
15 6 kA
12 kA
25 kA
44 kA
10
0
0 5 10 15 20 25 30
Current (kA RM S)
Source: Eblen and Short [2010]

Figure 4-1
Peak Incident Energy as a Function of Fault Current for 480-V Meter Bases
EPRI and PG&E have tested the two types three-phase 480-V meter sockets shown in Figure
4-2. Both were seven-jaw, type-16S self-contained units. The ringed-style meter sockets are
quite different than the ringless style. The ringless style has locking jaws with a bypass lever.
The main differences that may impact arc flash exposure are:
The cover must be removed on the ringless design to remove the meter or test voltages. This
creates a larger opening.
The ringless design with the bypass lever has a larger enclosure.
The locking jaws have more metal.
The connections for the incoming leads are different.
4-2
Ringed Ringless
Figure 4-2
200-A Meter Socket Styles Previously Tested
Figure 4-3 summarizes peak incident energies measured on 200-A meter sockets. The peak
incident energy is the maximum value measured from nine calorimeters. There are some
differences between the ringed and ringless designs:
The ringless sockets do not have as much of a rise in incident energies near 7 kA. The 12-kA
bolted fault case (approximately 7-kA arcing current) still gives the highest incident energies,
but the difference is less.
For the 12-kA bolted fault case, much less peak incident energy was measured on the ringless
sockets.
Another item to note is that peak incident energies were above 20 cal/cm2 for several of the
ringed meter sockets. The 20 cal/cm2 was based on four tests at a bolted fault current of 12 kA
submitted to the NESC subcommittee 8. Since then, several additional units were tested, and
some had peak incident energies above 20 cal/cm2.
4-3
25
Maximum incident energy, cal/cm 2

20
Meter socket
Ringed
Ringless
15
Bolted fault, kA
6
10 12
25
44
5
5 10 15 20 25
Current, kA
Figure 4-3
Peak Incident Energy Comparison on 200-A Meter Sockets at 18
1.0
0.8 Meter socket

Ringed
Ringless
Duration, sec
0.6
Bolted fault, kA
6
0.4 12
25
44
0.2
0.0
5 10 15 20 25
Current, kA
Figure 4-4
Fault Duration Comparison on 200-A Meter sockets
4-4
The enclosure geometry is the main reason for less peak incident energy with the ringless meter
compared to the ringed design. With the cover off in the ringless meter style, the arc energy is
much less directed. See Figure 4-5 for a comparison. The ringless design has a larger enclosure,
and since the cover will be off, the fireball expands in many directions.
Ringed meter socket Ringless meter socket
Figure 4-5
Fireball Comparison of Ringed and Ringless Meter Socket Arc Flash
Another 200-A Meter Socket Variation
Figure 4-6 shows a 200-A meter socket design with more bus bar than units tested previously.
The line and load connections are both at the bottom, and busbars connect these to the meter
jaws through bus bars. For the arc flash tests, three-phase faults were initiated in two locations:
once across the meter jaws and once across the bus bars (Figure 4-7).
4-5
Figure 4-6
200-A Milbank West 127TP Meter Socket
Test 2 Test 3
Figure 4-7
Milbank 127TP Fault Initiation Points
4-6
Figure 4-8 shows damage after both tests. The meter socket was reused after test 2 because the
damage was minimal. Figure 4-9 shows that both arc flash events self cleared in less than 10
cycles. Based on the damage patterns, the arcs tended to shoot off of the meter jaws. This
movement tended to cause arcs to clear.
For both of these tests, the bolted fault current was set at 25.7 kA. Figure 4-10 shows incident
energies measured at 18 in (45.7 cm) in front of the meter socket. During both tests, the
maximum incident energy was 5.4 cal/cm2. This is well below the 20 cal/cm2 threshold in NESC
table 410-1, and this is in line with other single-socket meter bases with a 200-A rating.
Figure 4-8
Damage after Test 3
4-7
Test 2 Test 3
Figure 4-9
Fault Current Waveforms
Test 2 Test 3
Figure 4-10
Incident Energies
In the two tests of this meter socket, faults self-extinguished quickly, and incident energies were
fairly low. Please note that because only two tests were performed, we cannot determine if
incident energies would be significantly worse with additional testing (especially with other fault
currents and/or other fault initiation modes or locations).
320-A Meter Sockets
320-A meter sockets were chosen for testing because some have bus bar and significantly more
metal than 200-A meter sockets, yet conductor-to-conductor spacings and conductor-to-ground
spacings are still very tight. Two Milbank 320-A meter sockets were tested. Figure 4-11 shows
pictures and nameplate information for these units. Other details of the tests include:
Calorimeter distance: 18 inches (45.7 cm) from the face of the meter socket
Bolted fault current: 25.7 kA
4-8
Source side: 350-kcmil Al leads

Meter bypass switch: closed
Load side: terminated with bolts to mimic cable terminations
Fault initiation: #12 Cu fusewire connected across all three phases
Faults on all meter bases were started as three-phase faults, normally with #12 Cu. Although this
is not a realistic way faults would start, we initiated faults in this manner to make the tests more
consistent and repeatable. Faults are most likely to start phase to phase or phase to ground. Based
on other tests and field experience, faults that start between two points (phase to phase or phase
to ground) can jump to involve all three phases in less than one half cycle.
The bolted fault current of 25.7 kA was chosen to investigate a case where high incident energies
could occur if arcing were to sustain. This is approximately the bolted fault current on a
500-kVA transformer with low impedance. For 200-A meter bases, the worst cases were for a
bolted fault current of 12 kA. In meter bases with busbar and tight spacings, we thought that
faults might sustain much longer at 25 kA (and they did).
4-9
Figure 4-11
320-A Milbank U2594 Meter Socket
4-10
Incident energies were measured by nine copper calorimeters on stands spaced 6 in (15.2 cm)
apart with 8 in (20.3 cm) center to centersee Figure 4-12 and positioned 18 in (45.7 cm) away
from the face of the meter bases. Calorimeters were built and calibrated according to ASTM
specifications [ASTM E 457 08, 2008; ASTM F1959, 2006].
Figure 4-12
Calorimeter Layout
The two units were involved in three tests. Results varied. In test 5, arcing self cleared in
11.5 cycles, and in test 6, arcing sustained after restriking for 84 cycles. The highest incident
energies were measured on test 6 with 86.9 cal/cm2 on one calorimeter. During this test, the
arcing did not self clear; the labs protection tripped to clear the fault. In test 5 and test 6, the
same unit was tested. Because of the short duration in test 5, there was minimal damage on the
meter socket, so another fault was initiated, and the unit was retested. Figure 4-13 shows the
fault initiation point for test 6. Once the circuit is energized during the test, the #12 copper
shorting wire quickly vaporizes, and the arcs can move. Normally, the magnetic forces push the
arc away from the source. In this meter base, the arcs traveled clock-wise around the bus bars.
4-11
Figure 4-13
320-A Meter Socket Prior to Test 6 with the Fault Initiation Point Shown
Figure 4-14 shows high-speed video frames from test 6. The fault self-cleared after
approximately 0.2 sec (12 cycles). The fault reignited 48 cycles into the event. The reignition
may be from mechanical movement (broken leads or bus bars) causing another short circuit.
Figure 4-15 shows video frames captured through an infrared-passing filter (only the first part of
the event is shown smoke and dust on the lens obscured the last part of the test). The infrared-
filtered video highlights the hottest parts of the event. Note that the arcs danced around and did
not stay in one location. Arcing on the load-side terminals was common, but arcs often reignited
upstream of that.
4-12
0.0 2.4 4.6 cycles
7.8 11.0 12.6 cycles
47.8 51.0 65.0 cycles
70.8 104.8 121.0 cycles

Figure 4-14
Movie Frames from Test 6
4-13
0.0 0.1 0.3 0.7 1.1 cycles
3.0 5.1 6.8 8.9 11.1 cycles
47.9 48.1 49.7 53.2 56.1 cycles

Figure 4-15
Infrared-Filtered Movie Frames from Test 6
Figure 4-16 shows current waveforms from test 6. The arcing current was chaotic and depended
on where the arcing occurred. During the first burst of arcing (the left panel in Figure 4-16), the
red phase was intermittent, and during a period where the red phase was off, the fault completely
self cleared. The red phase is phase A, the left phase at the incoming terminals.
4-14
Figure 4-16
Current Waveforms from Test 6
The meter base was extensively damaged by arcingsee Figure 4-17. The meter socket holder
burned completely free and dropped to the ground. The top row of meter jaws completely burned
up. Based on the damage pattern, the arcs spent most of the time on the bus bars above the meter
jaws (these burned completely through) and the top meter jaws; the bolts attached on the load-
side terminals also saw significant damage from arcing.
Figure 4-18 shows incident energies measured at each calorimeter during test 6. The center
calorimeter measured the highest incident energy (86.9 cal/cm2).
4-15
Figure 4-17
Damage after Test 6
4-16
Figure 4-18
Incident Energies by Calorimeter Position for Test 6
Figure 4-19 shows fault currents from test 4, the other significant event on the 320-A meter
sockets. The currents show a similar pattern as test 6. The arcing reignited, but in this test, the
fault did not sustain. Again, the red phase was most intermittent. Figure 4-20 shows damage after
the test. The arcing concentrated on the bus bars above the meter jaws. The top meter jaws and
the bolts on the load-side connectors also showed arcing damage.
Figure 4-19
Fault Currents from Test 4
4-17
Figure 4-20
Damage after Test 4
Multi-Meter Sockets
Multi-meter sockets are another category of meter bases that have significant bus-work and tight
spacings. We tested two multi-socket units: see Figure 4-21 for a view and nameplate for the
three-position Landis & Gyr unit, and see Figure 4-22 for the four-position Milbank unit. Test
parameters were similar to the 320-A meter socket tests:
Calorimeter distance: 18 inches (45.7 cm) from the face of the meter socket
Meter bypass switch: closed
Load side: terminated with bolts to mimic cable terminations
Fault initiation: #12 Cu fusewire connected across all three phases
4-18
Figure 4-21
Three-Position Meter Socket
4-19
Figure 4-22
Four-Position Meter Socket
4-20
Test 7 was the first test of the multi-meter sockets, and it turned out to be the most severe event.
A three-phase fault was initiated in the meter compartment to the far right with the cover off
(Figure 4-24). The remaining covers were left on (but without meters) as shown in Figure 4-23.
The bypass switch was engaged on that meter socket.
The fault began in the right-most compartment, but it propagated to adjacent compartments until
reaching the left-most compartment. The series of video frames in Figure 4-26 shows this
propagation. These video frames are taken through an infrared-passing filter. After about
40 cycles, the video became too cloudy to see anything.
Figure 4-23
Three-Position Meter Socket in Position for Test 7
4-21
Figure 4-24
Fault Initiation Point on the Three-Position Meter Socket for Test 7
0.0 0.1 0.3 cycles
0.5 0.9 2.0 cycles
3.1 6.8 7.1 cycles
Figure 4-25
Infrared-Filtered Movie Frames from Test 7
4-22
8.9 11.2 11.6 cycles
11.8 12.2 13.7 cycles
14.1 13.7 17.7 cycles
18.6 20.5 22.2 cycles
23.3 25.1 26.2 cycles
28.8 30.6 32.5 cycles
30.6 34.9 36.4 cycles
Figure 4-26
Infrared-Filtered Movie Frames from Test 7 (continued)
4-23
The fault current sustained for two seconds until it was cleared by the laboratory protection. The
fault did not self clear. Figure 4-27 shows the fault currents from this event. The incident
energies for this event were not particularly high, with a peak of 12.1 cal/cm2 (Figure 4-28). The
calorimeters were aligned in front of the right compartment. Because the arcing moved to other
compartments, the calorimeters did not absorb much energy. Based on incident energy rates from
other tests, we expect that incident energies to the front (and possibly the side) of the left
compartment exceeded 50 cal/cm2.
Figure 4-27
Fault Currents for Test 7
Figure 4-28
Incident Energies by Calorimeter Position for Test 7
Figure 4-29 shows damage from test 7. Several inches of bus bar were burned off from the left
compartment along with a sizeable hole in the enclosure. The meter socket dropped free.
4-24
Figure 4-29
Test 7 Damage
In tests on the second multi-socket unit, arcing was more intermittent. In the first test on the unit,
test 10, arcing lasted for 17 cycles. In test 11 with a different fault initiation point, arcing lasted
4-25
for 10 cycles, and in test 12, arcing lasted for 57 cycles. The fault initiation points for each test
are shown in Figure 4-30 through Figure 4-32.
Figure 4-30
Fault Initiation for Tests 10
Figure 4-31
4-26
Figure 4-32
In test 12, the longer duration event, arcing extinguished then reignited as shown in Figure 4-33.
Test 12 was the only phase-to-phase fault initiation in this series of tests. Note that the fault
jumped to the third phase in less than one half cycle. Figure 4-34 shows damage from this event.
Most of the arcing occurred on the second compartment from the right. The top row of meter
jaws and the bus bars above that had the most arcing damage. The maximum incident energy
measured was 43.6 cal/cm2 on the lower left calorimeter (as facing the calorimeters). Because the
calorimeter array was not exactly centered on the location with the most arcing, the incident
energy may have been higher in front of the second compartment from the right.
Figure 4-33
Fault Currents for Test 12
4-27
Figure 4-34
Test 12 Damage
Analysis of Meter Socket Test Results
Results from the 320-A meter socket and the multi-meter sockets showed that these styles of
meter bases can sustain faults considerably longer than the 200-A single-socket units previously
tested. Incident energies can be well above 20 cal/cm2, the 2012 NESC threshold for work on
480-V meters.
Table 4-1 summarizes the results of the testing on these meter sockets. Note again that in two of
these tests, the arcs did not self clearthe lab protection trips after two seconds to protect the
equipment. These meter socket tests were performed at a bolted fault current of 25.7 kA,
approximately the bolted fault current of a low-impedance 500-kVA transformer.
Figure 4-35 through Figure 4-37 compares this data to prior testing on meter sockets by PG&E
and EPRI. Note that several tests greatly exceed prior tests in terms of arc duration and incident
energies. Keep in mind that two of the incident energy measurements on the multi-socket units
were likely underestimated because the arcing moved to a different location within the unit. The
heat rate (Figure 4-37) for the 320-A and multi-socket units was comparable to the single-socket
units previously tested. This implies that the main difference is durationarcs did not self clear
or did not self clear as fast on the meter sockets with substantial bus bars.
4-28
Table 4-1
Summary of 2011 Testing of Meter Sockets at a 25.7-kA Bolted Fault Current
Average Average Maximum

Arc Arc Maximum
fault arc incident
Test Type duration, energy, heat rate,
current, power, energy,
msec Wh 2 cal/cm2/sec
kA kW cal/cm
2 200-A single 155 10.8 177 4143 5.4 34.8

3 200-A single* 106 13.3 197 6698 5.4 51.1
4 320-A single 454 17.1 830 6510 19.4 42.7
5 320-A single 109 17.1 194 6404 5.4 49.0
6 320-A single* 1,403** 16.1 3134 7511 86.9 62.0
7 200-A multi 2,007** 16.6 4311 7729 12.1 6.0
10 200-A multi 283 17.9 623 7953 8.9 31.6
11 200-A multi* 174 18.0 375 7847 8.4 48.4
12 200-A multi* 952 16.4 1716 6908 43.6 45.8
* Repeat test on the meter socket

** Test ended by opening the lab circuit breaker
2.0
1.5
Duration, sec
Type
200A multi
1.0
320A single
200A single
0.5
0.0
5 10 15 20 25
Current, kA
Figure 4-35
Arcing Durations by Meter Type
4-29
80
Peak incident energy, cal/cm 2

60
Type
200A multi
40 320A single
200A single
20
5 10 15 20 25
Current, kA
Figure 4-36
Maximum Incident Energy by Meter Type
80
Incident energy rate, cal/cm 2 sec
60
Type
200A multi
40 320A single
200A single
20
5 10 15 20 25
Current, kA
Figure 4-37
Maximum Incident Energy Rate by Meter Type
4-30
Because of the longer durations and much higher incident energies that are possible from these
events, utilities should consider options to reduce risks from these types of meter sockets. One
option is to de-energize this type of meter socket before work. Another option to consider is
training workers to identify these more hazardous meter sockets.
We only tested two models of multi-socket meters and one 320-A model. While arc flash is
equipment dependent, we have learned the main mechanisms that affect whether arcs sustain or
whether they self clear. These units can sustain arcing for a longer duration because they have
(1) large bus bars, (2) tight conductor-to-conductor spacings and conductor-to-enclosure
spacings, and (3) relatively tight confinement (the boxes are relatively small). These
characteristics can be found in other self-contained meter bases with higher ratings and/or have
multiple socket locations. Utilities may need to consider that faults wont self clear in multiple-
meter sockets and units rated above 200 A. Note that this applies only to 480-V equipment.
Meter sockets at 240 or 208 V are not expected to sustain arcing based on previous tests at 208 V
[EPRI 1022218, 2010; EPRI 1022697, 2011].
4-31
5
480-V OVERHEAD TESTS
Overhead secondary is another area where arc flash accidents have occurred, although risk and
incident energies were expected to be much lower than in metering or other more enclosed
equipment. On overhead conductors, there is nothing to confine arcs, making them more likely to
self extinguish. Because of this, the clothing threshold for overhead applications at 480 V is
4 cal/cm2 in the 2012 NESC. The tightest conductor spacings occur on bundled conductors, so
we arranged several tests to evaluate arc flash on overhead quadraplex conductors at 480 V. The
test results generally support the 4-cal/cm2 threshold in the 2012 NESC
Secondary Quadraplex Tests at 25 kA
Figure 5-1 shows the general test arrangement for the 4/0 Al quadraplex. The calorimeter array
was placed on one side of the conductors. As much as possible, faults were initiated to try to
aim the arcs towards the calorimeter array. Figure 5-2 and Figure 5-3 show the arrangement
for a phase-to-neutral fault initiation at a bolted fault current of 25.7 kA. We tested both phase-
to-neutral and phase-to-phase fault initiations at three bolted fault currents: 25.7 kA, 12 kA, and
6 kA. We sliced a wedge out of the conductor insulation, and bridged the gap between the phase
conductor and the neutral with a screw driver. The neutral was under tension, but not as much
tension as could typically be found in the field. Incident energies were measured at 15 inches
(38.1 cm) from the fault point, corresponding to the distance for glove work in the NESC [IEEE
C2-2012].
Note that the conductors were tied together with binding wires to limit conductor movement.
When a fault occurs between conductors, the magnetic forces will push the conductors apart. By
increasing the distance between conductors and allowing the arc more room to move, arcs are
much more likely to self extinguish. Because of this, the worst case is with the conductors held
closely together. Bundled secondaries can have a binding wire wrapped around the conductors.
Pre-spun secondary can have the tightest phase-to-phase and phase-to-neutral forces. As reported
by line workers [powerlineman.com, 2011], it can be difficult to separate pre-spun secondary
conductors. A plausible fault initiation mode occurs if line workers use conducting tools to try to
separate conductors with compromised insulation. These tests roughly match that scenario.
5-1
480-V Overhead Tests
Figure 5-1
Overhead Secondary Test Set-up for Test 13
5-2
Figure 5-2
Setup for Test 13
Figure 5-3
Line-to-Neutral Fault Initiation Prior to Test 13
5-3
Figure 5-4 shows damage from the line-to-ground fault in test 13. The fault stayed as a single
phase-to-ground fault and did not involve the other phases. Arcing lasted for about 3.5 cycles
before the fault self extinguished (Figure 5-5). Figure 5-6 shows video frames from the event.
The highest incident energy reading was 1.9 cal/cm2 (Figure 5-7).
Figure 5-4
Damage after Test 13
Figure 5-5
Test 13 Fault Current
5-4
0.2 1.2 1.6 cycles
2.2 3.2 3.6 cycles
4.8 5.2
Figure 5-6
Test 13
Figure 5-7
Test 13 Incident Energies
5-5
Figure 5-8 shows a phase-to-phase fault initiation on the 4/0 Al quadraplex, also at a bolted fault
current of 25.7 kA. This configuration is more severe than the phase-to-neutral fault because:
The voltage is higher, 480 V phase to phase instead of 277 V phase to neutral.
This fault configuration can support three arcing paths: phase to phase, phase A to neutral,
and phase B to neutral. More arcing path options make faults more likely to sustain.
We intentionally bound the conductors more tightly. The binding wires were placed closer to
the fault initiation point.
Figure 5-8
Phase-to-Phase Fault Initiation for Test 14
Figure 5-9 shows video frames from test 14. The arcing lasted for 3.5 cycles before self clearing
(see Figure 5-10 for the current waveforms). The neutral conductor completely separated, and
the whole bundle fell to the ground. Both faulted phase conductors burned through as well
(Figure 5-11). Figure 5-12 shows the calorimeter measurements; the peak incident energy was
3 cal/cm2.
5-6
0.2 1.2 2.2 cycles
3.4 4.6 8.2 cycles

Figure 5-9
Test 14
Figure 5-10
Fault Currents during Test 13
5-7
Figure 5-11
Figure 5-12
Test 14 Incident Energies
Figure 5-13 shows video frames through an infrared filter covering the last cycle of the event.
This shows that the calorimeter array approximately captured close to the worst of the arc flash.
In each half cycle, the arc expands out significantly and then collapses at the current zero
crossing.
5-8
2.2 2.3 2.4 cycles
2.5 2.6 2.7 cycles
2.8 2.9 3.0 cycles
3.1 3.2 3.3 cycles

Figure 5-13
Test 14 Infrared-Filtered Video Frames
Analysis of Overhead Secondary Tests
Table 5-1 summarizes the six overhead secondary tests. At lower fault currents (6 and 12 kA
bolted), faults extinguished very quickly. This may not be a general rule; with only two data
points, we cannot generalize this. The phase-to-phase tests were more consistent. In each case,
the peak incident energy was about 3 cal/cm2.
At lower currents, arcing lasted longer in the phase-to-phase tests. The product of current and
duration (It) stayed relatively constant for these events as shown in Figure 5-14. In these three
tests, the phase and neutral conductors burned completely through. The fault appears to arc until
the conductors are burned through, and then the arcs self clear.
5-9
Table 5-1
Summary of Overhead Quadraplex Tests at 480 V
Average Maximum
Bolted Arc Arc Maximum
Fault fault incident
Test Fault, duration, energy, heat rate
type current, 2 energy,
kA msec Wh cal/cm /s
kA cal/cm2
15 6 P-N ** 0.1
16 6 P-P 169.8 4.3 32.6 17.8 3.0
17 12 P-N 3.8 5.8 0.4 0.0 0.0
18 12 P-P 87.8 7.8 43.5 39.9 3.5
13 25.7 P-N 73.8 7.2 21.7 26.1 1.9
14 25.7 P-P 54.9 11.2 61.8 54.4 3.0
** = Duration and other parameters were too small to accurately represent.
3.5
3.0
Maximum incident energy, cal/cm 2
2.5
Bolted fault, kA
6
2.0 12
25
1.5
Fault type
PN
1.0 PP
0.5
0.0
100 200 300 400 500 600 700

I t (A sec)
Figure 5-14
Incident Energy and Current-Duration Product for 480-V Overhead Tests
5-10
These arc flash tests are similar to burndown testing on covered conductors that have been done
previously [EPRI 1017839, 2010; Goode and Gaertner, 1965]. Under the premise that arcing can
occur until the conductors burn clear, we can use the burndown results to extrapolate our arc
flash results from the 4/0-Al tests. Larger conductors should take longer to burn through. In
burndown tests on covered conductor [EPRI 1017839, 2010], tests showed that the duration to
burn-through is roughly proportional to the conductor cross sectional area. The average incident
energy for the phase-to-phase faults was 3.2 cal/cm2. As a first approximation, we could estimate
the incident energy on 350-kcmil Al. The ratio of cross sectional areas is 350/211.6 = 1.65 (4/0
has a cross-sectional area of 211.6 kcmil). Using that ratio, we can estimate the maximum
incident energy on the 350-kcmil Al as 1.653.2 = 5.1 cal/cm2. Note that this is an
approximation.
Overall, these test results support the 4 cal/cm2 threshold in the NESC. These overhead test
results generally support single-layer flame-retardant clothing. Consider also that the highest
energies occurred for phase-to-phase tests; these are less likely than phase-to-neutral contact. In
addition, our tests were worst-case in that we bound the conductors tightly for the phase-to-phase
contacts. If the conductors have more slack, we expect faster self clearing.
Note that burns can still happen on overhead secondary work even for exposures less than
4 cal/cm2, especially to exposed areas (the face for example) or if the worker is much closer to
the fault point than the 15 inches (38.1 cm) used for these tests.
5-11
6
480-V CT CABINET TESTS
CT-style meters have much less incident energy than self-contained meters [EPRI 1018693,
2009]. CT cabinets are of interest for arc flash exposure. In the 2012 NESC, the threshold for CT
cabinets is 4 cal/cm2. Figure 6-1 shows a CT cabinet used for testing. The incoming supply was
fed from the top. Cable stubs were added on the load side of the CTs. Based on previous testing
in the secondary compartment of padmounted transformers [EPRI 1020210, 2009], it is difficult
to sustain arcing with such wide conductor-to-conductor spacings at 480 V. Test parameters
included:
Calorimeter distance: 18 inches (45.7 cm) from the face of the CTs
Load side: 350-kcmil Al stubs
Fault initiation: vice grips or #12 Cu fusewire
Once faulted, the arcs will move away from the source. Because of the orientation of the CTs,
arcs will propagate to the cable stubs at the bottom of the enclosure. The key distance that
determines arc sustainability is the distance between exposed electrodes and the bottom of the
enclosure. Two different distances were tested as shown in Figure 6-2. The enclosure is
approximately 35 inches high. With the CTs exactly centered, there would have been a distance
from the bus bar to the bottom of the box of 12 inches (30.5 cm) and a distance of 9 inches
(22.9 cm) from the bottom of the exposed cable lug to the bottom of the box. In the first test on
this unit (test 19), the CTs were offset upward, leaving gaps of 13.75 and 10.75 inches (34.9 and
27.3 cm) at the bottom as shown in Figure 6-2. As we will see, the test at that distance did not
sustain long. A more severe arrangement was used for tests 20 and 21 where the CTs were
offset towards the bottom of the enclosure with bottom distances of 8.5 and 5.5 inches (21.6 and
14.0 cm).
6-1
480-V CT Cabinet Tests
Figure 6-1
CT Cabinet
6-2
Test 19: 13.75 in (34.9 cm) from bus bars to the bottom;
10.75 in (27.3 cm) from the bottom of the crimps
Tests 20 and 21: 8.5 in (21.6 cm) from bus bars to the bottom;
5.5 in (14.0 cm) from the bottom of the crimps
Figure 6-2
Bottom Distances in the CT Cabinet
Figure 6-3 shows typical arcing patterns as captured by 600 frame-per-second video through an
infrared-passing filter. As noted above, arcs tend to shoot out the bottom of the electrodes and
terminate on the enclosure.
6-3
Figure 6-3
Test 20 Video Frames through and Infrared-Passing Filter
Figure 6-4 shows fault initiations used for the three arc flash tests in the CT cabinet. The two
tests with vice grips were severethe vice grips were locked into place on one end and wedged
between bolts on the other end. The magnetic forces from the fault current will push the tool
away in this scenario.
Table 6-1 summarizes results from the three tests in this CT cabinet. Faults self cleared quickly
in tests 19 and 20 and incident energies at 18 inches (45.7 cm) were 1 cal/cm2 or less. Test 21
had the longest arcing and highest incident energy of 6.3 cal/cm2.
6-4
Test 19 Test 20
Test 21
Figure 6-4
Fault Initiations
Table 6-1
Summary of Overhead Quadraplex Tests at 480 V
Average Maximum
Bottom Arc Arc Maximum
fault incident
Test Fault type distance, duration, energy, heat rate
current, 2 energy,
inches msec Wh cal/cm /s
kA cal/cm2
19 Vice grips, P-P 10.75 75* 25* 118 13.3* 1.0

20 Wire, 3 phase 5.5 72 9.3 90 10.3 0.7
21 Vice grips, P-P 5.5 763 8.7 586 8.3 6.3
* = Approximate values due to intermittent restriking.
6-5
Figure 6-5 shows fault currents from the three tests. The first vice-grip test (test 19) only
involved a phase-to-phase fault, and arcing was intermittent before clearing completely in less
than 0.3 seconds. The longer distance between the bottom terminals and the bottom of the
enclosure probably explains why the fault did not become a three-phase fault. In test 20, a fault
initiated on all three phases, the arcing self cleared in less than five cycles, indicating that the
arcing pattern is not stable with this electrode geometry. In test 20, the phases with the closest
spacings had the highest fault currents (these are the two phases on the left which are white and
blue in Figure 6-5). Test 21 was the most severe with intermittent arcing that maintained a stable
pattern for almost 0.5 sec.
Test 19 Test 20
Test 21
Figure 6-5
Fault Currents
6-6
Figure 6-6 shows the arc flash from test 21, and Figure 6-7 shows damage after the event. One of
the cable stubs burned clear. As seen from the fault currents in Test 21, most of the arcing
occurred across the phases initially bridged by the vice grips. The third phase faulted, but that
current was more intermittent and of lower magnitude.
Figure 6-6
Test 21
Figure 6-7
6-7
Figure 6-8 shows infrared-filtered video frames from test 21. These show that the vice grip
shifted towards the end of the crimped lugs. This makes sustained arcing more likely because it
shortens the electrode to enclosure distance.
Figure 6-8
Infrared Filtered Video Frames from Test 21
6-8
Figure 6-9 shows incident energies measured on the calorimeters. The fireball was rather wide,
causing a relatively even distribution of incident energy measurements. The highest value
identified in the figure is an erroneous reading, probably caused by molten aluminum splatter on
that calorimeter.
Bad value
Figure 6-9
Calorimeter Values by Position from Test 21
Although incident energies were measured above 4 cal/cm2, this testing seems to validate the
range of incident energies used to arrive at the 4 cal/cm2 threshold given in the 2012 NESC. Test
21 was extreme for several reasons:
The vice grip fault initiation was severeit is difficult to imagine a real-world fault initiation
where a tool could be wedged that tightly.
During this test, the vice grip shifted to a position that seemed to enhance arc sustainability.
This may be unusual.
The bottom distances were intentionally shortened to evaluate worst-case conditions, and the
crimped connectors were left exposed. At shorter distances, arcs are more likely to sustain.
That said, this extreme case still led to incident energies in the range of single-layer FR clothing.
Although we think CT cabinets like this are relatively low risk, the test results highlight some
practices that can reduce risk:
Tape crimped connectors to reduce the phase-to-enclosure distances.
Keep CTs centered vertically to maximize phase-to-enclosure distances.
Evenly space the CTs horizontally to maintain the best phase-to-phase clearances.
6-9
7
FUTURE WORK
The testing, information sharing, and analysis work in 2011 has increased our knowledge on
distribution arc flash. The following highlights work still needed to fill in major gaps in
knowledge:
Low-voltage equipment
DC tests DC arc flash theoretically can have high incident energies. Tests would show if
that is realistic.
Network protectors Testing on single-phase external fuses and/or disconnects would
provide more confidence in using these solutions. Relay options could also be explored.
Medium-voltage equipment
Medium-voltage switchgear tests Arc flash is highly dependent on electrode arrangements
and box configuration. More tests on real switchgear or close mock-ups would better
determine incident energies. IEEE 1584 test scenarios do not replicate switchgear very well.
Other live-front gear: padmounted transformers tests Because of the surprising results on
PMH switches tested in prior EPRI research, it is worth exploring padmounted transformers
and other live-front equipment.
Underground vaults / cable splices Con Edison has tested a number of scenarios of faults in
vaults. A more detailed review of Con Edisons test data is warranted to determine the best
analysis approaches and assumptions for work in vaults. The Con Edison tests had durations
less than 0.2 sec with relatively currents above 20 kA. Additional tests may be needed at
other currents and durations.
Other issues
Clothing performance Fabrics generally underperform relative to ASTM ratings if exposed
to the fireball from an arc. Different fabrics also respond differently. Face shields tend to
perform better than their ASTM ratings if exposed to the fireball from an arc. More tests or a
better industry test would help provide workers with more effective clothing.
Better modeling of arc flash There is a need for a more universal modeling tool that can
account for electrode configurations, arc movement and directionality, and enclosure effects.
7-1
8
REFERENCES
ASTM E 457 08, Standard Test Method for Measuring Heat-Transfer Rate Using a Thermal
Capacitance (Slug) Calorimeter, ASTM International, 2008.
ASTM F1959, Standard Test Method for Determining the Arc Rating of Materials for Clothing,
ASTM International, 2006.
Cholewinski, K. and Davis, A., "May 13, 2010 Shubert Vault Fire Investigation Summary,"
EPRI North American Dense Urban Working Group Meeting, St. Louis, MO, 2011.
Eblen, M. L. and Short, T. A., "Arc Flash Testing of Typical 480-V Utility Equipment," IEEE
IAS Electrical Safety Workshop, Memphis, TN, 2010. Paper ESW2010-05.
EPRI 1017839, Distribution Conductor Burndown Test Results, Electric Power Research
Institute, Palo Alto, CA, 2010.
http://my.epri.com/portal/server.pt?Abstract_id=000000000001017839.
EPRI 1018693, Distribution Arc Flash: Analysis Methods and Arc Characteristics, Electric
Power Research Institute, Palo Alto, CA, 2009.
EPRI 1020210, Distribution Arc Flash: 480-V Padmounted Transformers and Network
Protectors, Electric Power Research Institute, Palo Alto, CA, 2009.
EPRI 1022218, 208-V Arc Flash Testing: Network Protectors and Meters, Electric Power
Research Institute, Palo Alto, CA, 2010.
EPRI 1022697, Distribution Arc Flash: Phase II Test Results and Analysis, Electric Power
Research Institute, Palo Alto, CA, 2011.
Goode, W. B. and Gaertner, G. H., "Burndown Tests and Their Effect on Distribution Design,"
EEI T&D Meeting, Clearwater, FL, Oct. 14-15, 1965.
IEEE C2-2012, National Electrical Safety Code.
IEEE Std 1584-2002, IEEE Guide for Performing Arc-Flash Hazard Calculations.
powerlineman.com, "Triplex spreader,"
http://www.powerlineman.com/lforum/showthread.php?7582-Triplex-spreader/page2,
2011. Accessed 2011-11-05.
8-1
Export Control Restrictions The Electric Power Research Institute Inc.,
Access to and use of EPRI Intellectual Property is granted (EPRI, www.epri.com) conducts research and
with the specific understanding and requirement that development relating to the generation, delivery
responsibility for ensuring full compliance with all applicable and use of electricity for the benefit of the public.
U.S. and foreign export laws and regulations is being An independent, nonprofit organization, EPRI
undertaken by you and your company. This includes an
brings together its scientists and engineers as well
obligation to ensure that any individual receiving access
hereunder who is not a U.S. citizen or permanent U.S. as experts from academia and industry to help
resident is permitted access under applicable U.S. and address challenges in electricity, including
foreign export laws and regulations. In the event you are reliability, efficiency, health, safety and the
uncertain whether you or your company may lawfully obtain environment. EPRI also provides technology, policy
access to this EPRI Intellectual Property, you acknowledge and economic analyses to drive long-range
that it is your obligation to consult with your companys legal
research and development planning, and supports
counsel to determine whether this access is lawful.
Although EPRI may make available on a case-by-case research in emerging technologies. EPRIs
basis an informal assessment of the applicable U.S. export members represent more than 90 percent of the
classification for specific EPRI Intellectual Property, you and electricity generated and delivered in the United
your company acknowledge that this assessment is solely States, and international participation extends to 40
for informational purposes and not for reliance purposes. countries. EPRIs principal offices and laboratories
You and your company acknowledge that it is still the
are located in Palo Alto, Calif.; Charlotte, N.C.;
obligation of you and your company to make your own
assessment of the applicable U.S. export classification and Knoxville, Tenn.; and Lenox, Mass.
ensure compliance accordingly. You and your company
TogetherShaping the Future of Electricity
understand and acknowledge your obligations to make a
prompt report to EPRI and the appropriate authorities
regarding any access to or use of EPRI Intellectual Property
hereunder that may be in violation of applicable U.S. or
foreign export laws or regulations.
2011 Electric Power Research Institute (EPRI), Inc. All rights reserved.
Electric Power Research Institute, EPRI, and TOGETHERSHAPING THE
FUTURE OF ELECTRICITY are registered service marks of the Electric
Power Research Institute, Inc.
1022002
Electric Power Research Institute

3420 Hillview Avenue, Palo Alto, California 94304-1338 PO Box 10412, Palo Alto, California 94303-0813 USA
800.313.3774 650.855.2121 [email protected] www.epri.com

480-V Distribution Arc Flash Updates PDF

Uploaded by

Copyright:

Available Formats

480-V Distribution Arc Flash Updates PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

480-V Distribution Arc Flash Updates PDF

Uploaded by

Copyright:

Available Formats

480-V Distribution Arc Flash Updates

Technical Update, December 2011

EPRI Project Manager

ELECTRIC POWER RESEARCH INSTITUTE

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER

REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY

THE FOLLOWING ORGANIZATIONS PREPARED THIS REPORT:

Electric Power Research Institute (EPRI)

Pacific Gas & Electric (PG&E)

The following organizations prepared this report:

Electric Power Research Institute (EPRI)

Pacific Gas & Electric (PG&E)

This report describes research sponsored by EPRI.

2 UPDATE ON 480-V SPOT NETWORK PRACTICES ............................................................2-1

3 FAULT CURRENTS IN 480-V NETWORK PROTECTORS...................................................3-1

4 480-V METER SOCKET TESTS ............................................................................................4-1

5 480-V OVERHEAD TESTS ....................................................................................................5-1

6 480-V CT CABINET TESTS ...................................................................................................6-1

7 FUTURE WORK .....................................................................................................................7-1

Figure ES-1 Meter Bases with Significant Busbar ........................................................................ x

Table 1-1 NESC 2012 Table......................................................................................................1-5

At 480 V, we identified several areas worth further testing:

Distribution Arc Flash: Analysis Methods and Arc Characteristics

Distribution Arc Flash: Industry Practices

Distribution Arc Flash: 480-V Padmounted Transformers and Network Protectors

Distribution Dispatch, Fourth Quarter, 2009

Arc Flash Phase II Work Practices Survey Report

Distribution Arc Flash: Phase II Test Results and Analysis

Arc Flash Testing of Clothing and 480-V Meter Sockets

Nominal Voltage Range and Calories/Cm2

CT meters and control wiring 42 45 68

Metal-clad switchgear / motor control

Pedestals, pull boxes, hand holes 42 87 128

Open air (includes lines) 42 47 68

Network Protectors 410 Note 11 Note 11

Panel boards - single phase (all) / three

Summary of Interest Group Discussions

A 24" Yes Yes 100 cal/cm2 CM-52, some with ARMs

100 or 140* Looking at smaller fuses; put in CM-52

External CLF fuses for large spots;

G 18" No 65 cal/cm2 Looking at smaller fuses

Looking at external disconnects for

External fuses for new units; de-

* = still under review

There were several themes common with several utilities:

Utilities raised several issues during discussions:

Analysis of Calculation Assumptions

Single-Phase, 277-V Scenarios

IEEE 1584 Fault Currents

log10 I a K 0.662 log10 I bf 0.0966V 0.000526G 0.5588V log10 I bf 0.00304G log10 I bf

Ia, arc current, kA

Ibf, bolted fault current, kA

K, 0.153 for open configurations

0.097 for box configurations (enclosed equipment)

G, distance between buses, mm

0.0 Square D 2002

Arcing current as a percentage of bolted current

Arcing current as a percentage of bolted current