A GRAPHICAL APPROACH TO INCIDENT ENERGY ANALYSIS

Copyright Material IEEE

Paper No. ESW2017-39

Rick Lutz, P.E. Maximilian Charbonneau Michael Garcia, P.E.

Senior Member, IEEE Member, IEEE Member, IEEE
Exponential Engineering Company Exponential Engineering Company Exponential Engineering Company
4251 Kipling Street 2950 East Harmony Road 4251 Kipling Street
Suite 425 Suite 265 Suite 425
Wheat Ridge, CO 80033 Fort Collins, CO 80528 Wheat Ridge, CO 80033
USA USA USA
[email protected] [email protected] [email protected]

Abstract – The goal of an incident energy analysis is to There are many sources of error when predicting the arcing
determine the largest incident energy that a worker could be fault current using the IEEE 1584 method. Modeling the type,
exposed to at a piece of electrical equipment. This is most length, and size of cables and conduits with one hundred percent
commonly done using the IEEE 1584 method. It is a common accuracy to determine bolted fault currents is virtually impossible
observation when performing this analysis that minor variations given typical economic constraints. Additionally, the fault
in fault current can result in large changes in a protective device’s contribution data obtained from utilities is subject to change over
operating time and the resulting incident energy. This would not time as the utilities system configuration changes. Even the
be a problem if it was possible to remove all errors from the IEEE 1584 arcing fault current equations have a margin of error
bolted fault calculation and the arcing fault calculation. associated with them because they are empirically derived.
Unfortunately, the arcing fault current calculation has some Because of this, IEEE 1584 recommends that a second fault
margin of error and the data used in arc flash models is often current that is 85% of the calculated value be used to help
estimated. mitigate the effects of this error in low voltage systems [1].
A graphical approach to determining the incident energy Taking a graphical approach to incident energy analysis, as
circumvents these problems and ensures that the worst-case presented in this paper, it is possible to reduce the effects of
incident energy is determined. This graphical approach is these errors on the final incident energy result.
derived from graphs of incident energy based on protective
device time current curves. An examination of the resultant II. ITERATIVE APPROACH
graphs reveals a few maximums in the incident energy curve.
These maximums are caused by the protective devices curve, Some power system analysis software allows the user to run
the minimum and maximum arcing fault currents, and any time an incident energy analysis using multiple study iterations to help
limitations that are applied. As a result, the locations of these account for the uncertainty inherent in system modeling, arcing
maximums are easily predicted allowing the maximum incident fault current predictions, and system configurations. Study
energy to be determined by performing the IEEE 1584 iterations are typically added for each source of error or for
calculation at a few key current values. situations that result in different levels of fault current. For
example, adding a tolerance to all cable lengths and to the
Index Terms — Arc Flash, Incident Energy, IEEE 1584 system’s fault current source. Using these tolerances, two
iterations representing the minimum and maximum fault current
I. INTRODUCTION can be created. The maximum fault current iteration uses the
minimum cable lengths, and the maximum utility contribution.
Calculations of incident energy directly impact the safety The minimum fault current iteration uses the maximum cable
decisions of workers who perform tasks on energized electrical lengths, and the minimum utility contribution. The ability to
equipment. In order to ensure safe work practices, knowing the automatically run a second calculation for each of these
worst-case hazard is of the utmost importance. Incident energy iterations, with the second being at a current value that is 85% of
calculations are commonly done using the IEEE 1584 the original, is often built into incident energy analysis software.
calculations. These calculations attempt to determine the The result is four study iterations and four incident energy values,
incident energy by predicting the amount of current that an arcing the highest of which will ideally represent the worst case incident
fault will have and then using that current to determine the energy.
upstream protective devices operating time. The incident energy Even with many study iterations to account for known sources
is then calculated using these time and current values. of error and different system configurations, it is difficult to know
Unfortunately, due to discontinuities in the Time Current Curves if the worst-case incident energy has been determined. The flaw
(TCCs) of Overcurrent Protective Devices (OCPDs), small in this iterative approach is more apparent when looking at a
changes in fault current can dramatically affect the OCPD’s graph of incident energy for an individual piece of equipment.
operating time and the calculated incident energy. This causes A graph of incident energy for all possible current values can
small amounts of error in the arcing fault current prediction to be created by determining operating times for the full range of
result in an incident energy result that is significantly different currents seen by the upstream OCPD. The operating times are
from the incident energy caused by a real-world arcing fault. determined by the OCPD’s TCC. The incident energy is then

1 978-1-5090-5099-4/17/$31.00 ©2017 IEEE

calculated as a function using these times and currents. An
example of an incident energy curve for one piece of low voltage
equipment and its OCPD can be seen in Fig. 1. Along the curve
four points are highlighted. These points come from running an
incident energy analysis for a system with its minimum and
maximum fault current. For each of these current levels the
incident energy was calculated at the arcing fault current and at
85% of the arcing fault current resulting in the four study
iterations shown.

Fig. 2 Incident Energy, TCC, and AFTL

Fig. 1 Comparison of Derived Incident Energy Curve to the

Iterative Method

This graph of incident energy allows one to see that the

iterative approach is calculating multiple incident energy values
along an incident energy curve. The use of multiple study
iterations increases the odds of finding the maximum incident
energy, but does not guarantee that it will be found. Performing
more study iterations between the minimum and maximum
currents further increases the odds of finding the maximum
incident energy, but this approach is only sampling more points
along the incident energy curve. With this method, it is difficult to Fig. 3 Incident Energy and ETCC
determine if the true maximum has been found or the required
number of study iterations that should be used to find it. Looking at the plot of incident energy alongside the plot of the
ETCC allows one to see how the incident energy curve is
III. GRAPHICAL APPROACH affected by the shape of the TCC. In Fig. 4, the different TCC
segments of the incident energy curve created by the ETCC are
An alternative to the iterative approach, a graphical approach, labeled. These segments are defined by the discontinuities in
comes from understanding how the incident energy curve is the ETCC. In the example used here, these discontinuities occur
shaped. This is made evident with a graph of the incident energy where the AFTL intersects the inverse time curve and the current
curve overlaid with the OCPD’s TCC (Fig. 2). Also, displayed on level at which the definite time setting picks up.
this graph is the Arcing Fault Time Limit (AFTL). The AFTL • Segment A: This segment is defined by the AFTL. Along
(recommended as two seconds by IEEE 1584 [2]) appears as a this segment, the current varies, but the operating time is
horizontal line. This time limit effectively becomes part of the constant. This causes the incident energy to increase with
TCC that takes over when the OCPD’s operating time exceeds the increasing current as it approaches the current level at
the AFTL. For the purpose of understanding the shape of the which the AFTL intersects the inverse time curve.
incident energy curve, the AFTL is a definite time segment of the
• Segment B: This segment is defined by the inverse time
TCC. The Effective TCC (ETCC) can be seen in Fig. 3. The
TCC shown in this graph is an extremely inverse curve with a curve. It begins where the AFTL and inverse time curve
definite time element. intersect. Along this segment, the incident energy begins to
change at a rate that is defined by the inverse time segment
of the OCPD’s curve. This segment ends where the definite
time part of the TCC picks up.

2 978-1-5090-5099-4/17/$31.00 ©2017 IEEE

 • Segment C: This segment is defined by the definite time Typical AFTL
section of the TCC. It begins at the current pickup level of
the definite time function. Along this segment, the operating Another possibility is for a local minimum to be created by the
time is constant and the incident energy increases with inverse time curve. In Fig. 6 the same moderately inverse curve
increasing current. from Fig. 5 is used, but in this graph the AFTL has been
increased so that a larger current range is affected by the
moderately inverse curve. This results in a local minimum due
to the effects of both the steep and flat parts of the inverse time
curve on the shape of the incident energy curve.

Fig. 4 ETCC and Incident Energy Separated by Segment

The incident energy curve in graphs one through four shows

the inverse time segment of the TCC causing the incident energy Fig. 6 Incident Energy with Moderately Inverse Curve and
curve to slope downwards with increasing current. This is the Increased AFTL
result of the slope of the inverse time curve and where the
definite time segment intersects the inverse time segment. It is Whether the slope of the resultant incident energy curve is
possible for the inverse time curve to cause the incident energy increasing or decreasing as the arcing current increases is
to increase with increasing current. This can be seen in Fig. 5 determined by the slope of the TCC. If the slope of a TCC is
which shows the incident energy curve for a moderately inverse greater than (more positive) the slope of a line of constant
curve. The moderate slope of the inverse time curve between incident energy, then the incident energy will increase with
the intersection of the AFTL and the definite time curve is not increasing arcing fault current. In Fig. 7 this is demonstrated
steep enough to reduce the incident energy with an increase in with a “TCC” that is a line whose slope is greater than that of a
current. Instead, the incident energy continues to increase with line of constant incident energy. A line of constant incident
increasing current. energy is shown for comparison along with the resulting
incident energy curve for the “TCC”. It should be noted that the
line of constant incident energy is plotted along the time and
current axis of the TCC and is unrelated to the incident energy
axis, which applies only to the resultant incident energy curve.

Fig. 5 Incident Energy with Moderately Inverse Curve and

3 978-1-5090-5099-4/17/$31.00 ©2017 IEEE

 of such a curve is shown in Fig. 9.

Fig. 7 TCC Segment with More Positive Slope than Line of

Constant Incident Energy Fig. 9 Concave Down TCC Segment
If the slope of a TCC is less than (more negative) that of a The local maximum created by this curve can be found at the
line of constant incident energy, then the incident energy will point at which the slope of the curve is equal to that of a line of
decrease with increasing fault current. In Fig. 8 this is constant incident energy. This point, where the line of constant
demonstrated with a “TCC” that is a line whose slope is less incident energy is tangent to the TCC, is plotted in Fig. 10. In this
than that of a line of constant incident energy. A line of figure, one can see that the rest of the TCC lies below the line of
constant incident energy is shown for comparison along with constant incident energy. The vertical line demonstrates that the
the resulting incident energy curve for the “TCC”. point on the TCC segment that is tangent to a line of constant
incident energy occurs at the same current value that results in
an incident energy maximum. This curve shape is relatively
uncommon in a typical power system, but it does exist.

Fig. 8 TCC Segment with More Negative Slope than Line of

Constant Incident Energy

The minimum seen in Fig. 6 is caused by the slope of the Fig. 10 Concave Down TCC Segment Indicating Maximum
TCC transitioning from less than to greater than that of a line of with Line of Constant Incident Energy Tangent to TCC
constant incident energy. This concave up shape is typical of
inverse time curves. Not all concave up curves will create a local minimum in the
Conversely it is also possible for a local maximum to be incident energy curve. In order for a concave up curve to create
created by a TCC. This is caused by a concave down shape. In a minimum in the incident energy curve, the slope of the curve
order for a local maximum to occur, the slope of a TCC has to has to transition from being less than (more negative) that of a
decrease with increasing current (concave down). An example line of constant incident energy, to greater than (more positive)

4 978-1-5090-5099-4/17/$31.00 ©2017 IEEE

that of a line of constant incident energy.
Similarly, not all concave down curves will create a local
maximum in the incident energy curves. In order for a concave
down curve to create a maximum in the incident energy curve,
the slope of the curve has to transition from being greater than
(more positive) that of a line of constant incident energy, to less
than (more negative) that of a line of constant incident energy.
A TCC can be thought of as separate, continuous segments
that are separated by discontinuities in the ETCC. There are four
basic types of curves that can make up a TCC: slope greater than
a line of constant incident energy, slope less than a line of
constant incident energy, concave up, and concave down.
Knowing how the incident energy curve is shaped by TCC
segments, it is clear that the behavior of the incident energy
curve is predictable for each TCC segment. With the exception
of TCC segments that are concave down, local maximums in the
incident energy curve are not created by an individual TCC
segment.
Therefore, the points of interest that could cause local
maximums in the incident energy curve are the discontinuities in
the ETCC that separate the TCC segments. In the case of the Fig. 11 Points of Interest on Incident Energy Curve
TCC shown in graphs two through six, they are where the AFTL
meets the inverse time curve and where the inverse time curve Once the minimum and maximum arcing fault currents have
is cut off by the definite time curve. It is possible for OCPDs to been calculated, the currents with discontinuities in the TCC
have TCCs with more discontinuities than are shown in these should be determined. Knowing these currents allows one to
examples. These other discontinuities have the potential to determine which discontinuities are within the region of interest
create local maximums in the incident energy curve and created by the minimum and maximum arcing fault current (Fig.
therefore need to be analyzed. 12). This can further narrow down the currents at which you
There are two other points that could create local maximums should calculate incident energy. Any discontinuities that are
in the incident energy curve. Those points are where the outside of the region of interest do not need to be analyzed. Also,
minimum and maximum arcing fault currents intersect the ETCC. for the constant operating time segment created by the AFTL, it
All of the points that could cause a local maximum in the incident is known that the highest incident energy will always be at the
energy curve are shown on the incident energy curve in Fig. 11. highest current value of the segment.
This figure, and the following figures, revisit the example first In the case of Fig. 12, we can see that the minimum current is
shown in Fig. 1. below the current at which the AFTL intersects the OCPD’s
In order to apply the graphical method, one needs to determine curve. Therefore, it is not necessary to calculate the incident
the minimum and maximum arcing fault currents that are energy at the minimum current; instead, it should be calculated
possible at the piece of equipment being studied. When doing at the point where the AFTL intersects the TCC.
this, reasonable engineering judgment should be applied to
ensure that unrealistic levels of fault current are not analyzed.
Possible causes of arcing fault current variance include: error in
the IEEE 1584 arcing fault current calculation, upstream system
changes, and estimated impedance data.

Fig. 12 Region of Interest

The maximum arcing fault current in this example is higher

5 978-1-5090-5099-4/17/$31.00 ©2017 IEEE

than the pickup current of the definite time part of the OCPD’s Currently, no software exists that will determine which currents
curve. Because we do not know which of these points will have need to be analyzed; the graphical approach has to be applied
the highest incident energy the energy needs to be calculated at manually. However, there are a variety of tools available to
the other discontinuities in the ETCC and at the maximum fault compute the incident energy for discrete values of current,
current. including the IEEE 1584 spreadsheet.
Going through this process allows one to reduce the number
of incident energy calculations and eliminates the need to run V. ACKNOWLEDGEMENTS
calculations across the entire current range. Once the incident
energy is calculated at these three current levels, the worst-case The authors would like to thank Dave Rightly, P.E. and their
incident energy rating possible at this equipment is the highest of other colleagues at Exponential Engineering Company for their
the calculated values. In Fig. 13 the results of this approach are support and assistance with reviewing this paper. We also like
highlighted. In this case, the highest incident energy occurs to thank the IEEE Technical Committee who also reviewed the
where the AFTL intersects the TCC. This approach allows one paper.
to find the maximum incident energy that was missed by the
approach shown in Fig. 1. VI. REFERENCES

[1] IEEE Std 1584-2002 Guide for Performing Arc Flash

Hazard Calculations - Section 9.10.4
[2] IEEE Std 1584-2002 Guide for Performing Arc Flash
Hazard Calculations - Annex B.1.2

VII. VITA

Rick Lutz, P.E. graduated from South Dakota School of Mines

and Technology with a BS (1992) and MS (1995) in electrical
engineering. Mr. Lutz is currently the Denver Area Engineering
Manager for Exponential Engineering Company in Wheat Ridge,
Colorado. He is a licensed Professional Engineer in multiple
states, a Colorado Master Electrician, and has been involved
with arc flash studies since 2002. He has also taught basic
electricity through the renewable energy program at a local
community college.

Max Charbonneau graduated from New Mexico Institute of

Mining and Technology with a BS (2007) and started working on
Fig. 13 Incident Energy at Important Points on Incident his MS at New Mexico State University in electrical engineering.
Energy Curve Mr. Charbonneau is currently an Electrical Engineer for
Exponential Engineering Company in Fort Collins, Colorado. He
IV. CONCLUSION has been involved with arc flash studies since he started with the
company in 2009.
Using the graphical approach enables the determination of the
maximum incident energy without the use of numerous study Michael Garcia graduated from Texas Christian University with
iterations, which are typically needed. This eliminates the need a BS (2009) in engineering with an electrical emphasis. Mr.
to study the effects of subtle differences in system modeling and Garcia is currently an Electrical Engineer with Exponential
the many possible variations in utility or generator fault current. Engineering Company in Wheat Ridge, Colorado. He has 6
It also avoids any of the error in attempting to calculate the exact years of experience performing power systems studies including
arcing fault current from the bolted fault current. arc flash studies. In 2015 he became a licensed Professional
This approach is particularly useful for studies with a wide Engineer in the state of Colorado.
range of available fault currents or when impedance data, such
as cables, are known to be estimated. In these situations, it
allows one to quickly determine the worst-case incident energy.

6 978-1-5090-5099-4/17/$31.00 ©2017 IEEE