Arc Flash and Electrical Safety

W-J Lee (Univ. of Texas at Arlington), T. Gammon (John Matthews and Associates, Cookeville, TN),
Z. Zhang (Univ. of Texas at Arlington), B. Johnson (Thermon, San Marcos, TX), and J. Beyreis
(Retired)

Abstract
According to NFPA 70E, more than 2000 workers are admitted to hospital burn centers per year
for extensive injuries caused by arc flash accidents. Arc flash incidents occur when unintended
electric current flows through air, superheating the air and causes an explosion. Recognizing the
significant threat posed by arc flash hazards, IEEE and NFPA have joined forces on an initiative to
support research and additional testing to increase the understanding of the arc flash phenomena.

Several areas of the arc flash phenomena need further research and testing validation to provide
relevant information that can be used for developing safety strategies to protect workers. The
identified areas include but are not limited to: (a) Heat and Thermal Effects, (b) Blast Pressure, (c)
Sound and (d) Light Hazards. The test results of this project will provide information to help more
accurately predict the hazards associated with high energy arcing faults, thereby improving electrical
safety standards and providing practical safeguards for employees in the work place. This report
highlights the activities of this collaborative research project.

Keywords: Arc Flash, Incident Energy, Non-Thermal Hazards, and Vertical and Horizontal
Electrodes.

Introduction
Shortly after Thomas Edison’s Pearl Street Station started generating electricity, people
quickly learned that, in addition to the benefits of electrical power, direct contact with electricity
caused personal shock injury and improperly installed or malfunctioning electrical systems
initiated fires. By the 1950s, it was widely understood that serious injury and death could result
from an electric shock; however, there was little research and data to define exactly what
threshold of energy damages human tissue. In the late 1950s and early 1960s, Alice Stoll and
Maria Chianta studied the effect of heat related to burn injury. Their research determined the
level of heat energy that would produce the onset of a second-degree burn, where skin blisters as
the outer layer (epidermis) separates from the inner layer (dermis). Their research led to the
development of the “Stoll Curve,” which is used to predict burn injury. The Stoll Curve is also
the benchmark for determining the ATPV (arc thermal performance value) of flame-resistant and
flame-retardant materials used in clothing designed to protect against the thermal effects of an
arc flash.

In 1976 the National Fire Protection Association (NFPA) formed a new electrical standards
development committee at the request of the Occupational Safety and Health Administration
(OSHA) to develop an electrical safety standard. Three years later, NFPA 70E, Standard for
Electrical Safety Requirements for Employee Workplaces, was first published; it serves as the
foundation for electrical safety practices in the United States today. In 1982 Ralph Lee presented
“The Other Electrical Hazard: Electrical Arc Blast Burns,” the first paper to quantify the
potential burn hazards from arc flash incidents [1]. Lee’s work made the engineering community

‹,(((  3UR5HOD\
aware of the occurrence of arc flash incidents in the workplace and their associated thermal and
pressure hazards [2]. Although thermal burns usually inflict the greatest harm, arcs can also
injure or kill from blast-pressure related injuries.

Much of the earlier arcing-fault testing was performed single phase. Research focused on
quantifying electrical parameters, assessing system damage or protecting the electrical system from
damage. Due to the growing awareness in the 1990s that electrical arc flashes account for a
significant number of electrical injuries, the arc testing focus expanded to quantifying incident (heat)
energy, fabric testing, and determining the personal protective equipment (PPE) needed to protect
workers from arc-flash hazards. Testing was done to quantify the reduction in incident energy
associated with current-limiting, overcurrent protective devices. The recognition of the arc flash
hazard in the 1995 edition of NFPA 70E was the beginning of standards formally addressing this
additional electrical hazard. The 2000 edition of NFPA 70E introduced the Hazard Risk
Category classification system. Tables were developed for FR fabric and PPE selection to protect
against the thermal effects of arc flash.

In 2000 Richard L. Doughty, Thomas E. Neal and H. Landis Floyd II published “Predicting
Incident Energy to Better Manage the Electric Arc Hazard on 600V Power Distribution Systems”
[3]. This paper included the results of 25 three-phase arc tests conducted at 600V and presented
separate open-air and enclosure incident energy equations based on bolted-fault current, arc
duration and distance to the arc. These equations have been included in NFPA 70E as an
acceptable method for calculating incident energy in an arc-flash hazard analysis. The test data
from this paper became part of the IEEE 1584 data set. The distance exponents formulated for
the open-air and cubic enclosure tests are almost identical to the distance exponents presented in
the IEEE Standard 1584.

First published in 2002, IEEE 1584, IEEE Guide for Performing Arc Flash Calculations, has
become the predominant method in the industry for performing arc flash calculation studies. This
guide presented detailed incident energy and arc current calculation methods and included an
extensive data set consisting of more than 300 entries. The 1584 data was recorded during three-
phase arc testing performed over a wide range of test conditions: supply voltages from 208 V to 13.8
kV, bolted-fault currents from 700 A to 106 kA, and gap widths from 7 mm (0.28 in) to 152 mm (6
in). In addition, tests were run with grounded, ungrounded or high-resistance grounded setups. Low-
and medium- voltage tests were conducted in open air and in metal enclosures. The enclosure sizes
were: 12”x14”x7.5” (LV, shallow), 20”x20”x20” (LV, cubic) and 45”x30”x30” (MV). The scheduled
test duration and the calorimeter distance to the arc also varied. The 1584 arc current and incident
energy equations were formulated from statistical analysis. The calculation factor was introduced to
reach a 95% confidence factor for predicting a PPE category to meet or exceed the incident energy
thresholds of 1.2, 8, 25, 40, and 100 cal/cm2. These thresholds combined the NFPA defined
categories of 1 (1.2 – 4 cal/cm2) and 2 (>4 – 8 cal/cm2); combining them is a common industry
practice.

Although the 1584 model was shown to fit the data well, concern has arisen on how well the
1584 Standard can predict the arc current and the incident energy associated with “real-world”
arcing faults. A statistical analysis is not based on physical observation; it may indicate bogus
relationships which do not really exist or obscure fundamental relationships. The 1584 equations


have been shown to counter known arc behavior when the variables or variable relationships are
outside the normal test range. Anomalies and discontinuities have also been shown to exist in the
IEEE 1584 equations [4, 5].

After the publication of the 1584 Standard, Stokes and Sweeting reported that most of the arc
power is stored in the plasma cloud as high temperature enthalpy, and that the convective heating due
to the plasma cloud is three times higher than the heating due to radiation alone [6]. When the
calorimeters are placed directly in front of horizontal electrodes in open air as shown in Figure 1(a),
the arc plasma is driven toward the calorimeters, which results in significantly higher calorimeter
measurements than 1584-type setups with vertical electrodes shown in Figure 1 (b) and (c). The
open-air vertical electrodes drive the plasma cloud toward the ground, not directly toward the
calorimeters placed in front of (and 90 degrees off-axis) the electrodes.

(a) Horizontal Electrodes in the Air (b) Vertical Electrodes in the Air

(c) Vertical Electrodes in the Metal Enclosure

Figure 1 Projection Trajectory of Arc Plasma


 Several areas of the arc flash phenomena need further research and testing validation to develop
safety strategies for protecting workers. The identified areas include but are not limited to: (a) Heat
and Thermal Effects, (b) Blast Pressure, (c) Sound and (d) Light Hazards. The test results of this
project will provide information that is expected to help more accurately predict the hazards
associated with high energy arcing, thereby improving electrical safety standards and providing
practical safeguards for employees in the work place. The proposed research and testing plan will
focus on, but will not be limited to (a) The Development of Physics and Engineering-Based
Modeling and (b) Testing to Validate Theory related to Heat Transfer and Thermal Effects, Arc Blast
Pressure, Sound, and Light Hazards. This effort will include arcs initiated in both open-air and in
enclosures. Enclosures tend to contain and direct the arc cloud and to deliver higher heat levels to
unfortunate persons standing in front of enclosure openings. Arc flash incidents are known to occur
while working on switchgear, motor control centers and power panels. This report highlights the
activities of this collaborative research project.

Test Program
The project is currently in Phase II of the test program. Series of arc flash tests have been
conducted for 600 V, 2700V, and 13.8 kV. Additional tests at other voltage and current levels are
planned. The medium-voltage enclosure sizes and electrode gap widths and spacing dimensions are
based on current engineering practices. The low-voltage test enclosures and dimensions were
selected to conform to UL 1558, UL 67 and NEMA MCC Standards.
The original test plan for the 4.16 kV tests, consisting of 270 tests is listed in Table 1. The
original number of tests was expanded by adding tests conducted at 2.7 kV; reducing the test voltage
from 4.16 kV to 2.7 kV was found to have minimal impact on the test results.

Table 1 4.16-kV Test Protocol*

Fault Duration Gap Distance to Measure VCB VCB- HCB Measure VOA HOA
Current (cycles) (inch) Back Panel Distance Barr Distance
(kA) (inch) (inch)a (inch)b

20/40/63 6 1.5/3/4.5 4 24 1 1 1 18 1 1

20/40/63 12 1.5/3/4.5 4 24 1 1 1 18 1 1
20/40/63 6 1.5/3/4.5 4 33 1 1 1 27 1 1
20/40/63 12 1.5/3/4.5 4 33 1 1 1 27 1 1
20/40/63 6 1.5/3/4.5 4 42 1 1 1 36 1 1
20/40/63 12 1.5/3/4.5 4 42 1 1 1 36 1 1
*VCB or HCB: Vertical or horizontal electrodes in cubic box (26”x26”x26” metal enclosure)
VCB-Barr: Vertical electrodes, terminated in barrier, inside cubic box, “Barrier test”
VOA and HOA: Vertical or horizontal electrodes in open air
a and b: Distance from calorimeters to electrodes a) located inside box or b) open air


Sample Thermal Related Testing Results
Voltage, Current and Power
Voltage and current were measured near the electrodes with the sampling rate of 10 or 20 k
samples per second. Arc power and energy were calculated from the sampled voltage and current
signals. A sample, 480-V three-phase arc current is displayed in Figure 2 for a bolted-fault current of
5 kA; the vertical electrodes were spaced 10 mm apart in a cubic box enclosure. The phase arc
currents decrease after the first cycle in Figure 2, and then increase as the arc stabilizes; such
waveforms are common for lower bolted-fault current levels. The arc’s ability to stabilize and to
sustain is primarily an issue for lower voltages (480 V and less) and depends on several factors
including: the bolted-fault current, electrode gap width and configuration (presence of an enclosure,
as well as dimensions and interior spacings). Table 2 summarizes the ranges of arc power associated
with the 480-V testing.

ArcCurrent(Ibf=5kA)
15000

10000

5000

Ia
0
A

Ib
61

1021
1081
1141
1201
1261
1321
1381
1441
1

121
181
241
301
361
421
481
541
601
661
721
781
841
901
961

Ic
Ͳ5000

Ͳ10000

Ͳ15000
Sample

Figure 2 Example of a Three-Phase Arc Current

Table 2 Arc Power (MW) for 480-V Testing

Ibf (kA) VCB HCB/HOA VOA VFC*
Lab 1 5.9 – 6.3 1.7 – 2.0 0.4 – 0.6
Lab 2 5.2 1.3 – 1.5 1.8 – 1.9
Lab 3 5.0 – 5.1 2.0 1.8 – 1.9 0.4 – 0.5

Lab 1 17.2 – 17.6 4.0 – 4.8 4.4 – 5.4 5.1 – 5.6 1.8
Lab 2 19.8 – 20.1 5.2 – 6.4 6.9
Lab 3 20.6 – 21.7 3.6 – 6.9 4.8 – 7.3 4.9 – 6.7 2.6
* Vertical electrodes inside Faraday cage, single-phase 480 V


Incident Energy Comparison
Copper slug calorimeters, painted black, are used to measure incident energies at several
distances from the electrodes. The spatial arrangement of seven slug calorimeters is shown in Figure
3. The middle row, center calorimeter is placed in-line with the tip of the center electrode. During an
arc test, each calorimeter experiences a temperature rise, which is converted to incident energy.
Although equation development for incident energy will be based on the maximum incident energy
associated with the highest temperature rise experienced by any single calorimeter during an arc test,
the overall and row average incident energies provide much insight into understanding the heat flow
and the heat levels experienced. Figure 4 displays the row-average incident energy measurements for
a series of twelve-cycle, 480-V tests. The x-axis labels designate the electrode orientation (V-vertical
or H-horizontal), configuration (CBu-unbonded cubic box or OA-open air), bolted-fault current (5 or
20 kA) and the gap width (10 or 25 mm). The results presented in Figure 4 suggest that the bolted-
fault current, electrode orientation and presence of an enclosure strongly impact incident energy.

6"

6"
6" 6"

6" 5 feet above ground

6"
Figure 3 Arrangement of Slug Calorimeters

Figure 4 Sample Results for Average Incident Energy Measurements


 Based on the Phase I and Phase II test results, the following factors have been found to
impact the level of incident energy (IE):
x Bolted-fault current level
x Duration of the arc
x Electrode Orientation/Presence of an Enclosure
x Calorimeter arrangement, height and measurement distance
x Voltage level
x Gap width between electrodes
x Distance between electrode and back panel
x Dimensions of the metal enclosure

Non-Thermal Hazards
Blast pressure, toxic gases and shrapnel can seriously injure or kill anyone in the vicinity of an
arcing fault. Blast pressures may cause permanent hearing loss and intense light may cause blindness.
Published information on non-thermal arc injuries is not abundant. Attachment 5 of the 2005
NFPA/IEEE Research and Testing Planning Committee (RTPC) Report provides one of the best
assessments of non-thermal arc injuries [7].

In the 2005 RTPC report, David Wallis analyzed 454 public OSHA records on investigations
involving electric arcs; thirty records which involved arc welders, arc furnaces and electric shock
were eliminated. Non-thermal injuries were reported in 7.3% of the remaining 424 records involving
electrical faults. He noted, “The risk of burn injuries from this hazard can be very severe. However,
an electric arc poses a substantial risk of non-burn injuries…less well known…frequently less severe
than the potentially threatening burn injuries.” David Wallis provided the Project Team with an
additional 100 relevant public OSHA records created after the RTPC report was written [8]. Five of
the 100 records were removed because they involved arc welders, an arc furnace or electric shock. Of
the 95 remaining records, 14 or 14.7% documented non-burn injuries. Table 3 provides a summary of
the non-burn injuries recorded in the additional OSHA records. The four records of eye injuries are:
two flash burns to the eyes (with no face injury specified), momentary blindness, and required eye
flushing from a fault causing a battery to blow up in a person’s face. An additional record, not
included in the table, included both face burn and eye injury. It is quite possible that many of the
OSHA records did not document additional non-thermal burn injuries.

Table 3 Non-Burn Injuries listed in OSHA Records [7, 8]

Injury Type Wallis’ 2005 Analysis Subsequent Records
Number of records considered 424 95
Smoke inhalation and asphyxia 13 (3.1%) 3 (3.2%) – 2 smoke inhalation
Thrown, knocked down, fall, loss 13 (3.1%) 6 (6.3%) – 2 falls
of consciousness, and fracture – 4 thrown
Eye injury 4 (0.94%) 4 (4.2%)
Laceration 1 (0.24%) 1 (1.1%) shrapnel
Hearing loss 1 (0.24%) 0
Total 31 (7.3%) 14 (14.7%)


Pressure Measurement
When an arcing fault is initiated, the gases expand rapidly in the vicinity of the arc. A high-
pressure front is created as the expanding gases compress the surrounding air. “A phenomenon called
‘blast overpressure’ forms from the compression of air in front of a blast wave which heats and
accelerates the movement of air molecules [9].” The severity of the blast pressure depends on the
initial peak pressure, the duration of the overpressure, the distance of individuals from the incident
location, and “the degree of focusing due to a confined area or walls [10].” Blast pressures are greater
when the explosion occurs indoors, particularly in small enclosed rooms, and the pressure wave
reflects from the walls [11].

Individuals may be injured or killed by blast pressures through three mechanisms. Injuries which
directly result from the pressure wave striking the body are known as primary blast injuries. Air- and
fluid-filled organs, such as the lungs, gastrointestinal tract and middle ear, are susceptible to primary
injuries. Primary blast injuries can cause concussions or mild traumatic brain injury without a direct
blow to the head [9]. Secondary injuries result from flying debris propelled by the blast wind.
Shrapnel wounds can occur anywhere, including the eye and head. Tertiary injuries result from the
individual being thrown by the blast wind [12]. Individuals may be injured by a fall or being
propelled into a wall or equipment. More than one blast injury may be sustained, and damage to one
organ often affects other organs [9].

Arc blast pressures have been measured or estimated using pressure sensors, a pendulum, and
high-speed video. Recording accurate measurements using traditional direct pressure sensors is
challenging due to the high magnetic flux and high temperature plasma gas during the arc event.
Based on the two consecutive high-speed video frames (1000 frames per second) showing the
movement of the arc cloud and air in Figure 5 during a 2.7-kV, VCB, 30-kA, 6-cycle, 3”-gap arc
flash test, the estimated pressure reached 1.7 psi (245 lb/ft2) at the opening of the enclosure.
Actual pressure is affected by air temperature and composition. Furthermore, the calorimeters
block some air flow and influence measurement. The most accurate techniques for arc blast
pressure measurement continue to be researched and tested.

Figure 5 Estimating Blast Pressure from Air Movement


 Lab testing has also provided evidence that secondary injuries in the form of shrapnel
wounds can occur from flying debris. As illustrated in Figure 6, molten copper deposited and re-
solidified on the surface of the camera filter lens ten feet away from the electrodes during one arc
test.

Figure 6 Resolidifed Molten Copper, Propelled on Camera Lens from Blast Force

Sound Level Measurement

Blast pressure frequently injures the ear. The initial positive air pressure may cause lesions on
the eardrum and internal ear; it may also dislocate or interrupt the chain of auditory ossicles or
rupture fenestrae (membranes). According to one study on the effects of blast lesions on the eardrum,
conduction hearing loss occurred most of the time [11].

In the Code of Federal Regulations 1910.95(b)(2), it is stated, “Exposure to impulsive or

impact noise should not exceed 140 dB peak sound pressure level.” When the potential peak sound
pressure is 140 dB or greater, individuals should wear personal hearing protection devices (PHPDs)
to reduce the exposure level within OSHA limits. Using a protective hood may also attenuate the
sound pressure level. Personal Protective Equipment (PPE) categories are based on incident energy, a
summation of heat flux over time. Since peak sound pressure is linked to the initial formation of the
arc, the PPE categories are not an effective method for assessing the sound hazard. Figure 7 shows
the peak sound pressures from 16 representative, medium-voltage arc tests. The peak sound
pressures, measured at a distance of three meters from the electrodes, were in the range of 150 to 170
dB. The x-axis label in Figure 7 specifies the configuration, bolted-fault current (kA), test duration
(12, 60-Hz cycles) and gap width (in), respectively. The C-weighting scale, indicated by LCpeak is
used to measure sound pressure levels above 85 dB.


 SoundLevelat3Meters
175

170

165

160

155
LCpeak(dB)
150

145

140

135
1 11 21 31 41 51 61 71 81 91

Figure 7 Summary of Peak Sound Pressures for Medium-Voltage Testing Series

Light Measurement
The wavelengths of visible light lie between 400 and 700 nm. It has been reported that the
light radiated by an arc flash covers part of the ultraviolet region, and is predominately in the range of
200 to 600 nm [13]. A number of research projects were conducted by the U.S. armed forces on flash
blindness in the 1960s. The research was fueled by the need to know how well service personnel,
especially pilots, could perform after a nuclear blast. The recovery time from flash blindness depends
on the light flash (size, intensity, direction, duration and spectrum), as well as the individual (pupil
size, age and individual variations). Flash blindness is the temporary loss of vision when the retina
receives an excess of thermal energy, but less energy than required to cause a burn. A reduction in
visual acuity can last a few minutes or a few days. Contributing factors are glare, afterimage and the
bleaching of the photochemical substances within the rods and cones of the retina. Glare is an excess
of light which hinders vision; even after the light source is no longer present, scotomatic glare from
intense light causes a reduction in the sensitivity of the retina [14].

The Project Team has used both a spectrometer and a light sensor to measure illumination
levels in lux (lumen/m2). Neutral density filters were used to attenuate the light to levels falling
within the measurable ranges of the device ratings. The frequency response of the neutral density
filters were characterized in the laboratory. The light measured through the neutral density filters
were calibrated to their actual values. Sample illumination measurements, taken during six recent
medium-voltage arc tests, at distances of 3, 4.5 and 6 meters from the electrodes are displayed in
Figure 8. The x-axis labels specify the HOA (horizontal electrodes, open air) configuration, the
bolted-fault currents (10, 20 or 32 kA) and the gap width (3 or 4.5 in). The results in Figure 8 show
that the light intensity increases at closer measurement distances and for larger bolted-fault currents.
Additionally, for the test configuration shown in Figure 8, the larger electrode-gap width also
significantly increases the illumination levels. It is worth mentioning that a bright summer day will
have a midday ground level illumination in the order of 100,000 lux. Since some light may be


blocked by the calorimeters, the Project Team will obtain more accurate measurements when the
calorimeters are removed to perform special arc blast pressure measurements.

IlluminationlevelsatDifferentMeasurement
Points
14
12
10
MillionLux

8
6 6.0M
4 4.5M
2 3.0M
0

Figure 8 Sample Illumination Levels

Conclusion
On average, approximately five to ten arc flash explosions occur daily in the United States.
Personal protective equipment can prevent or reduce casualties during these accidents. The IEEE and
the NFPA have joined forces on an initiative to fund and support research and testing to improve the
understanding of the arc flash phenomena and to better quantify arc current, incident energy and
several non-thermal hazards. The results of this collaborative project will be used to improve
electrical safety standards, predict the hazards associated with arcing faults, and provide practical
safeguards for employees in the workplace. Please visit the following website for additional
information on this IEEE/NFPA Collaborative Research Project:
http://standards.ieee.org/esrc/arcflash/index.html.

References
[1] R. H. Lee, "The other electrical hazard: electrical arc blast burns,” IEEE Transactions on
Industry Applications, vol. IA-18, no. 3, pp. 246-251, May/June 1982.
[2] R. H. Lee, “Pressures developed by arcs,” IEEE Transactions on Industry Applications, vol. IA-
23, no. 4, pp. 760-764, July/Aug. 1987.
[3] R. L. Doughty, T. E. Neal, and H. L. Floyd, II, “Predicting incident energy to better manage
the electric arc hazard on 600 v power distribution systems,” IEEE Transactions on Industry
Applications, vol. 36, no. 1, pp. 257-269, Jan./Feb. 2000.
[4] R.Wilkins, M. Allison, and M. Lang, “Calculating Hazards,” IEEE Industry Applications
Magazine, vol. 11, no. 3, pp. 40-48, May/June 2005.


[5] C. St. Pierre, “Putting arc-flash calculations in perspective,” Online, http://www.epc-
website.com.
[6] A. D. Stokes, and D. K. Sweeting, “Electric arcing burn hazards,” IEEE Transactions on Industry
Applications, vol. 42, no. 1, pp. 134-141, Jan./Feb. 2006.
[7] NFPA / IEEE Research and Testing Planning Committee, Final Report, July 28, 2005.
[8] OSHA records with “electric arc” keyterm, released between March 2005 and September
2008.
[9] G. L. Wallace, “Blast injury basics: A guide for the medical speech-language pathologist,”
The ASHA Leader, vol. 11, no. 9, pp. 26-28.
[10] “Blast injury,” Wikipedia: The Free Encyclopedia, modified on Oct. 3, 2008
http://en.wikipedia.org/wiki/Blast_injury.
[11] N. Sperm, S. Branica, K. Dawidowky, “Tympanoplasty after War Blast Lesions of the
Eardrum: Retrospective Study,” Croatian Medical Journal, vol. 42, no. 6, pp. 642-645,
2001.
[12] “Explosions and blast injuries: A primer for clinicians,” Center for Disease Control, May 9,
2003.
[13] R. A. Wilson, R. Harju, J. Keisala, S. Ganesan, “Tripping with the speed of light: arc flash
protection,” Proceedings of the 60th Annual Conference for Protective Relay Engineers, pp.
226-238, College Station, Texas, Mar. 27-29, 2007.
[14] J. D. Teresi, “A review of research on flash blindness,” USNRDL-TR-68-76, Radiological
Defense Laboratory, San Francisco, California, July 1, 1968.