OSHA asks for a written certification with four named elements. That is what we sign.
Most of what is written about arc flash in this country explains the hazard. Very little of it explains the document a compliance officer actually asks to see, or when in your maintenance year the study has to land to be worth anything. This page covers both, for the two US markets we deliver into: the Texas Gulf Coast, and the Northern Virginia data centre corridor.
Nearly every conversation about arc flash compliance in the United States is about labels and clothing. The regulation is about paperwork first.
29 CFR 1910.132(d)(1) requires the employer to assess the workplace to determine whether hazards are present that make personal protective equipment necessary. Where an arc flash hazard exists, that assessment has to address it.
1910.132(d)(2) then requires the employer to verify that the assessment was performed, through a written certification. The regulation names what the certification must identify, and the second element is the one most people get wrong · it is the person certifying, not necessarily the person who held the clipboard:
The absence of a documented assessment is itself the finding. Not the missing label. The missing document.
A spreadsheet of incident energy values does not satisfy 1910.132(d)(2). Neither does a modelling file, or a set of labels with no traceable author. The certification has to identify a person who certifies the assessment was performed, carry the date it was done, and identify itself as a certification.
Every VB Engineering deliverable is drafted by a Chartered Engineer and signed by an independent Principal Reviewer, with the site, the certifying engineer and the date on the face of the document. The study is the engineering. The certification is what you hand to the compliance officer. We issue both, and the second one is the reason the first one counts.
Tell us what your last hazard assessment certification says, and who signed it. If it names nobody, that is the finding waiting to happen.
This is where a great deal of published material goes wrong, including material written by people selling studies.
| Standard | What it covers | Who it binds |
|---|---|---|
| 29 CFR 1910.132(d) General PPE |
Hazard assessment and the written certification of it | Almost every industrial employer in general industry. A refinery, a data centre, a pharmaceutical plant, a steel mill, a food line. If you have workers and energised equipment, this one is yours |
| 29 CFR 1910 Subpart S Electrical · 1910.301 to 1910.399 |
1910.302 to 1910.308 design safety for electric utilisation systems · how the installation is built. 1910.331 to 1910.335 electrical safety related work practices · how people work on or near it, including 1910.332 training and 1910.335 safeguards for personnel protection. 1910.309 to 1910.330 are reserved | General industry, alongside 1910.132 |
| 29 CFR 1910.269 Electric power |
Operation and maintenance of installations existing for the sole purpose of generating, transmitting or distributing electric power | Electric utilities, and equivalent installations inside an industrial establishment. It does not govern your utilisation system · the switchgear and motor control centres feeding process load · and supplementary standby generation sits under Subpart S. Widely quoted at plants where only part of it applies |
NFPA 70E is not law. It is a consensus standard, and its 2027 edition took effect on 6 May 2026. What makes it enforceable in practice is that it is the recognised method for doing what OSHA requires, so a study performed to NFPA 70E and IEEE 1584-2018 is how an employer demonstrates the assessment was competent. That is the relationship. Anyone who tells you NFPA 70E is federal law has not read either document.
We reference the standard that governs your facility, name it on the certification, and do not pad the deliverable with standards that do not apply to you. If you operate generation or transmission assets as well as a plant, both frames appear, separately. How to compare arc flash study quotes covers what else belongs in a scope of work.
A Gulf Coast refinery and a Northern Virginia data centre both need the same calculation. They need it on completely different clocks.
The refining and petrochemical belt runs on turnarounds. Major units come down on a planned window, and that window is the only time much of the electrical system can be verified without an energised work permit.
If you are scoping the study when the turnaround starts, you have already missed it.
The corridor holds the densest concentration of data centre capacity anywhere. Qualified staff work 480 V bus duct, large-format UPS, generator transfer switches and medium voltage distribution around the clock.
A boundary that lives only inside a report cannot survive a decision made twice a week at two in the morning.
Tell us which of these you are, and when your next outage window or maintenance review falls. That date decides the schedule, not ours.
On the property and casualty side, loss prevention engineering drives the cadence as firmly as the regulator does, and it asks a question the regulator does not.
| Loss prevention data sheet | What it governs |
|---|---|
| DS 3-10 | Electrical safety related work practices, including arc flash and protective equipment. This is the arc flash one |
| DS 5-20 | Electrical testing, and the intervals at which equipment is tested |
| DS 5-19 | Switchgear and circuit breakers |
| DS 5-32 | Data centres and related facilities |
The testing sheet is the one that quietly decides whether your labels are true. Every incident energy figure assumes the protective device upstream clears in the time its curve says it will. A breaker that has not been exercised or tested can clear considerably slower. When that happens the real energy at the working distance is higher than the number printed on the label, and the clothing selected against that label is under-specified.
NFPA 70E 2027 now points to a defined reference for assessing condition of maintenance, which makes this explicit rather than assumed. So the insurer's testing programme and the validity of your arc flash study are not two separate files. They are the same question asked by two different parties. We say so in the report, and we tell you which buses the assumption is carrying the most weight on.
A fair question, asked more often than it is answered honestly.
Engagements are contracted and delivered through our United States presence in the Greater Houston area, with the analytical bench and the Chartered Engineer review behind it.
The walkdown is where a study is won or lost. Our engineers do it, photograph every data point, and put the evidence in front of your team for approval before a calculation runs.
Where your specification or your authority requires a Professional Engineer stamp on the deliverable, we arrange it. Tell us at scoping, because it changes the schedule.
A study changes behaviour only if the people who work the equipment have been taught against their drawings and their boundaries. We bring the training to the plant. The classroom session runs against your single line diagrams, and the immersive session runs in the room the study flagged, so a worker practises standing in the right place before the night it matters. Where the programme is run on a recurring cycle rather than as a one-off, that is VB Arc360.
Labels are issued in English, to the NFPA 70E 2027 format. Where a site has a substantial Spanish-speaking workforce we supply a Spanish-language safe work practice insert alongside the labels, rather than translating the label itself, so the label on the equipment stays in the format an inspector and a visiting contractor both expect to read.
Send us your single line diagrams, whatever state they are in. The condition of the drawing set is diagnostic, and we would rather see it early.
Anonymised, as every engagement on this site is.
An IEEE 1584-2018 incident energy baseline across the medium voltage and 480 V distribution, NFPA 70E format labels installed, and a five year recertification retainer signed before the concurrent maintainability proof was completed. The parties reviewing the outcome were the operator, the insurance carrier and the facility certification body, each asking for the same underlying evidence in a different format.
That engagement is also why the certification wording on our deliverables reads the way it does. Three reviewers, one document, no ambiguity about who evaluated what and when.
This page covers delivery in the United States. Each of these answers a different question, and none of them repeats this one.
Send us the last hazard assessment certification for the site and the date on it. If it does not name an evaluator, or if the electrical system has changed since, we will tell you what that means before we quote anything.
A Chartered Engineer responds inside one business day.
Gulf Coast turnaround or a data centre review coming up? The date drives the schedule.
Tell us the site, the date on your last hazard assessment certification, and your next outage window.