2. • Brief background of Arc Flash Analysis
• How is a study done?
• How are the equations derived?
• Anomalies (and their consequences) inherent in the
equations
• Practical Considerations in Energy Calculations
• Efficacy of the conventional Arc Flash Hazard
Analyses
• New Developments
Outline
3. Simplified Procedure:
Acquire the Necessary Data
Calculate the Bolted Fault Current
Calculate the Estimated Arcing Current
Calculate Fault Clearing Time
Calculate the Incident Energy
How is an AFH analysis done?
img. from IEEE Std 141-1993
4. • Ralph Lee proposed in a
1982 paper, a method to
calculate arcing energy
• Assumes Pmax of ½ of
available MVA (max
power transfer theorem)
• Assumes Ia/Ibf = 0.7
𝐸 = 2.142 × 106 𝑉𝐼 𝐵𝐹
𝑡
𝐷2
How The Formulae Are Derived
– Theoretical Lee Equations
Img. from Practical Solution Guide to Arc Flash Hazards, ESA, Inc., 2003
5. • Approximately 300 individual
trials were done over the
following ranges:
• 13.8kV: 5.7kA – 40.8kA
• 4.16kA: 5.4kA – 40.4kA
• 2.3kV: 2.6kA – 16.6kA
• <1kV: 0.7kA – 106kA
• Statistical analysis via least
squares method
• Equation normalized for
• 0.2 seconds
• 610mm distance.
How The Formulae Are Derived
– Empirically Derived IEEE 1584-2002 Equations
Img. from Calculating Hazards, IEEE IAS, June 2005
6. Theoretical Lee Equations:
• Three Phase Fault
• Over 15kV
• No current constraints
• Open Air
• AC
• No gap constraints
IEEE 1584 Equations:
• Three Phase Fault (or
quickly escalate to three
phase)
• 208V – 15kV
• 700A – 106kA
• Any grounding method
• Many enclosure types
• AC
The Ranges of the Equations
There are currently no equations to
analyze single phase or DC arcing events
7. • Iarc > IBF
• E(air) > E(box)
• Unrealistic Clearing Times
• Induction and Synchronous Machine
Contribution
• High Arcing Current Tolerances
Some of the possible anomalies and
considerations
8. • Intuitively IAF < IBF
• IBF < 1.489kA effect
of gap length in
calculating IAF is
reversed
Result unrealistic
arcing current values
In practice max at 1.0
pu of IBF
Equation Anomalies
re Varying Conductor Gap
log 𝐼 𝑎𝑟𝑐 = 𝐾𝐴 + 0.662 log 𝐼 𝐵𝐹 0.662 + 0.5588𝑉 − 0.00304𝐺 + 0.0966V + 0.000526G
Img. from Calculating Hazards, IEEE IAS, June 2005
ArcingCurrent vs Bolted Fault Current for
varying gap size
9. • Intuitively
E(air) < E(box)
• Adjusting via “distance
exponent” produces an
anomaly.
• LV systems, d < 166mm
• HV systems, d < 358mm
Anomaly is less significant
Equation Anomalies re Varying Working Distance
Img. from Calculating Hazards, IEEE IAS, June 2005
Ratio of E(air)/E(box) vs. Working Distance
10. • Exceedingly long clearing
times can lead to unrealistic
results
• Arc sustainability?
• Personnel movement?
• Currently no standard to
address this
Result Widely accepted
maximum clearing time of 2
seconds.
Considerations for High Clearing Times
Max IAF
never clears
11. • Motor contribution is done by
discrete current steps, which
overstates the energy.
• The IEEE 1584 equations do
not take into account
asymmetrical (DC)
components
In Practice?
Considerations of Induction and Synchronous
Machine Contributions
Img. from Cahier Technique No. 158, “Calculation of Short-
Circuit Currents”, ECT 158, Sept. 2005.
Img. from J.C.Das, “Arc Flash Hazard Analysis and
Mitigation”, p. 194, 2012
0
𝑡
𝑖2
𝑡 𝑑𝑡
12. • For <1kV systems, at a 95%
confidence level, the error
can be substantial
Considerations of High Arcing Current Tolerances
Img. from Practical Solution Guide to Arc Flash Hazards, ESA, Inc., 2003
log 𝐼 𝑎𝑟𝑐 = 𝐾𝐴 + 0.662 log 𝐼 𝐵𝐹 0.662 + 0.5588𝑉 − 0.00304𝐺 + 0.0966V + 0.000526G
IE at Ia =
8.0 cal/cm2
IE at
77% Ia =
22.0 cal/cm2
13. • Given the above possibilities for
error, the formulas still produce
good results.
• An analysis done in [1] below
summarizes 22 arc flash events
involving 30 personnel. The article
concluded the following : 2
• PPE performs as expected as
long as its rating matched the
hazard calculation results
• When injuries occurred it was
because PPE was either used
improperly or not used at all
Efficacy of conventional arc flash hazard
analysis
[1] D.R. Doan andT.E. Neal, “Field analysis of arc-flash incidents,” IEEE IndustryApplications Magazine, vol. 16,
no. 3, 39-45, June 2010
2 J.C.Das, “Arc Flash Hazard Analysis and Mitigation”, p. 194, 2012
In conclusion:
Get an arc hazard risk assessment
done and wear the proper PPE!
14. • Testing for the new IEEE 1584 standard
will have more than 1000 tests conducted
• Differing bus orientations
• Vertical in cubic box
• Vertical in cubic box, terminated
• Vertical in open air
• Horizontal in cubic box
• Horizontal in open air
• New research and analysis will be done
• Time domain analysis (Ia Time-Variant)
• DC arc flash (may) be added
• Likely the “208V, 125kVA” rule will be
reduced, to examine AFH on smaller
systems
Result Modeling will become more
complex
New Developments in Arc Flash Hazard Analysis
Img. from Calculating Hazards, IEEE IAS, June 2005
The purpose of this presentation is to expose engineers and designers to just a few of the possible errors and inconsistencies one can encounter when conducting arc flash energy calculations.
If you ask five different people each their method and process for calculating arc flash energy, you may very well get five different answers.
For the shown low voltage bus, the following analysis is required. First the equipment parameters must be acquired, so in this case the fuse time current characteristics, the transformer data and the upstream source impedance. The modes of operation need to be considered such as variance in the utility contribution, as well as bus tie scenarios. A short circuit study is then conducted. Using the IEEE arcing fault equations, the estimated arcing current is found. From this arcing current, the upstream protective device clearing time is then found, which is a function of the calculated arcing current. Using additional info such as the grounding and equipment type, the overall worst case incident energy is then calculated.
In 1982, an engineer named Ralph Lee proposed a method to quantify the energy of an arcing fault. This was done theoretically, and leads to highly conservative results. It would be another 20 years until empirical formulas would become available.
-Arcing current of 0.7 bolted fault current is based on the maximum power transfer theorem. As maximum dissipation occurs when source and load impedances are equal
The empirically derived IEEE 1584 equations were done by laboratory tests. Over 300 individual trials were done. Three conductors would be placed in a vertical arrangement with calorimeters placed at varying distances away. An arc would be initiated vie fuse wires, and fault current, as shown in the specified ranges would flow for 0.2 seconds. Arc current would be measured, as well as the energy dissipated. Statistical analysis would then be done on the raw data to derive the formulas.
Here is a summary of the ranges of each of the equations. The equations are valid only for three phase faults, and while they statistically make up a small percentage of faults, it provides a conservative estimate. For instances where gap is out of the IEEE 1584 range, such as for substation IPS bus, the Lee equations are then used.
At this time there are also no equations which analyze single phase or DC arcing events.
Intuitively, arcing impedance will always cause the ratio of arcing current to bolted fault current to be less than one. However, at currents less than 1.5kA, the effect of the gap length is reversed, causing higher arcing currents than the possible bolted fault. As a result, it produces unrealistic arcing current values, which as shown in the last slide, drastically increase the overall error in the energy calculation. This can also occur at the opposite end of the equation range.
In practice, software packages will max out the arcing current at 1.0pu of the bolted fault.
-How the experiments are done both in open air and within enclosures
-Intuitively, energy in a box is greater than energy in air
-Anomaly shows that as you convert from the normalized to
While not necessarily an anomaly, this slide outlines the greatest source of error within an arc flash energy calculation, the arcing current estimate.
The incident energy formula gives a predicted value with a statistical 95% confidence level. This means that there is a numerical 95% confidence that the PPE level will be more than adequate.
However, to achieve this confidence level, the predicted arcing current can vary wildly (as shown in the histogram).
This is why the IEEE 1584 standard suggests using two energy values, one at the calculated value and one 15% lower.
These anomalies and errors shown previously may suggest that the formulas do not adequately represent reality.
However, the following should be noted:
It is the best we have as of now
It is the current standard
Field events conclude that arc flash analysis does prevent injury