Utility arc-flash calculations at higher voltages were identified as an unresolved engineering area because available methods did not consistently agree with test data or with one another. The reported EPRI work divided the problem by system voltage and sought utility equipment test data rather than assuming that one calculation method covered every application.
Research Scope and Voltage Boundary
EPRI organized two separate projects: transmission and substation applications above 69 kV, and distribution systems at 69 kV and below. This boundary determines which research project applies; it does not establish that a particular calculation method is valid for every installation within either group.
The work was initiated amid concern over the NESC arc-flash compliance deadline of January 1, 2009. Treat that date as historical context, not as evidence of present requirements. Verify current obligations against the governing documents applicable to the utility and installation.
Calculation Methods and Reported Uncertainty
Methods then used for higher-voltage evaluations included NESC tables, ArcPro, and the Duke Power Heat Flux program. An unpublished analysis also compared Lee's formula, IEEE 1584 calculations, and ArcPro results. Because the analysis was not published in the supplied evidence, its findings are research observations rather than validated selection rules.
| Method or issue | Reported observation | Engineering implication |
|---|---|---|
| Lee's formula | Tended to produce the lowest values relative to test data in some cases. | Do not interpret the lowest result as inherently correct or conservative. |
| IEEE 1584 and ArcPro | Produced overestimates relative to test data in some cases. | Compare calculated results with applicable equipment test evidence where available. |
| IEEE 1584 versus ArcPro | Tracked each other within 30% to 50%, with a reported crossover point. | Do not apply a fixed ranking between the two methods across all conditions. |
| Test datasets | Some sets did not agree with other sets. | Record the equipment configuration and test basis before transferring results to another installation. |
| Arc topology | Whether arcs should be modeled as single-phase, three-phase, or phase-to-phase remained unresolved. | State the assumed topology explicitly and evaluate alternatives when the governing method permits them. |
480 V Utility Meter Testing
PG&E reported tests on self-contained and current-transformer meters at 480 V. Incident energy was measured at four current levels spanning 6.6 kA through 44 kA. The results were shared with NESC SC8 and were used in developing new low-voltage tables for utility equipment.
This evidence is equipment-specific. It supports evaluation of the tested meter configurations but does not establish that the same incident-energy behavior applies to unrelated 480 V equipment, different enclosures, or untested current levels.
Method Selection and Verification
- Classify the application as above 69 kV or at 69 kV and below, then identify the research and calculation scope relevant to that voltage class.
- Confirm whether the selected method covers the actual utility equipment and configuration. Do not infer coverage solely because the nominal voltage matches.
- Document the calculation method, arc-topology assumption, equipment configuration, and available test-data basis.
- Where multiple applicable methods exist, compare their results without assuming that IEEE 1584, ArcPro, or Lee's formula will always be highest or lowest.
- Investigate material differences against applicable test evidence and current governing requirements before using a result for a safety decision.
FAQ
How were the EPRI utility arc-flash projects divided by voltage?
One project addressed transmission and substation applications above 69 kV. The other addressed distribution systems at 69 kV and below.
How closely did IEEE 1584 and ArcPro results agree?
The reported unpublished analysis said their calculated values tracked within 30% to 50%, with a crossover point. That observation does not establish a fixed correction factor or a consistently more conservative method.
What current range was tested on 480 V utility meters?
PG&E reported incident-energy testing at four current levels spanning 6.6 kA to 44 kA on self-contained and current-transformer meters. Apply those findings only after confirming that the evaluated equipment matches the tested configuration.