AI Arc Flash Study Review

No. This review reads the study you upload and checks its model, inputs, assumptions, and internal consistency against a 64-point checklist. It never produces its own incident energy value, corrected number, or PPE category. If a finding suggests a value may be wrong, it tells you what to go verify and why — the recalculation, if one is needed, is the Engineer of Record’s.

Jay R. Prigmore II, Ph.D., P.E. — a Ph.D. electrical power engineer, a voting member of the IEEE 1584 working group that writes the arc flash calculation standard itself, and a founding member of the NFPA technical committee on electrical inspection practices. The checklist is original prose describing what to look for; no IEEE, NFPA, or NEC text is reproduced in it or bundled into this platform.

No, and it is important to be exact about this. The automated lane is performed by a four-pass software pipeline running the checklist a licensed P.E. wrote. Findings then pass through a human queue where each one is approved, edited, or removed before release. The engineering judgment in the review is the checklist author’s; the reading of your specific document is the software’s.

Yes. Intake asks which software produced the study, but the review reads the PDF you upload rather than a native project file, so any study exported to PDF works. Submit the incident-energy study output — the one-line plus the incident energy and PPE table — not a project archive. The checks are on the model, inputs, and assumptions the study reports, which are the same questions whichever of the three produced it; the software only changes where on the page the answers are printed.

That is a determination for your Engineer of Record and your AHJ, and this product does not make it for you — but it is the most common reason a study gets a second look, and the review is built to surface what the decision turns on. The 2018 edition replaced the 2002 empirical model with one derived from a much larger test program, made electrode configuration an explicit input rather than an assumption, and changed the range of voltages, gaps, and currents the model is stated to apply within. A study built on the older model is not automatically wrong, but it is frequently silent about inputs the newer model treats as decisive. The review reports which edition and model your study says it used, whether the equipment falls inside that model’s stated applicability limits, and which inputs the study leaves undocumented — which is the evidence the redo-or-not conversation actually needs.

NFPA 70E requires the arc flash risk assessment to be reviewed at intervals not to exceed five years, and also when a change in the electrical distribution system could affect the results — which in practice is the one that catches people, because the system changed and nobody re-ran anything. Before that review, the useful questions are whether the study still reflects the system as built, whether the protective device settings it assumed are the settings actually in the devices today, whether utility or source data has changed, and whether the study documents its own assumptions well enough that someone can tell. This review reads the existing study and reports what it does and does not establish. It does not inspect your facility, and it cannot tell you whether your system has changed — only your records and your people can.

No. This is an automated preliminary screening. It is not a stamped or sealed engineering deliverable, not a professional engineering opinion, and not a code-compliance determination by any Authority Having Jurisdiction. Absence of a finding does not mean the study is compliant. Every finding must be independently verified by the Engineer of Record before it is relied on.

Your uploaded files are scanned for malware before anything reads them, and are deleted automatically 30 days after the job completes — or sooner, on request, from your portal.