How often to do a study

A study is only performed once for each piece of equipment generally using the IEEE 1584-2002 guideline for equipment operating below 15kV. Above 15kV OSHA recommends using ArcPro.

Once the initial (baseline) study is completed, NFPA 70E(R) - 2018 requires that it be updated when changes to the electrical network takes place. Practically, this would be done if a transformer is upgraded, protection settings changes, fuse ratings are changed, or new electrical equipment installed.

Article 130.5(G) states that the arc flash engineering study should be reviewed for accuracy at intervals not exceeding five years. This would typically allow the plant to catch changes in the utility supply. Also in this five year period, staff turnover may have left certain changes unnoticed.

Having an experienced engineer perform the five yearly review helps as they are able to focus on core installations that influence the calculated arc flash incident energies. This will reduce the cost of the review and offer the most bang for your buck. At the five yearly review, the entire study is not redone from scratch. It is a "review" and not a "redo".

Hugh Hoagland

Staff member
Dec 30, 2016
Louisville, KY
Some people get forced into a redo because they chose their arc flash study engineering provider poorly.

Many engineering firms do not provide the software digital files from the study.

We recommending putting that in the scope of work.

Require them to use a standard software when possible so all plants can share information.

We use SKM PowerTools, EasyPower, EDSA and ETAP depending on the site and ArcPro 3.0 for utility studies. When you have the digital files, including any spreadsheets they used, the original engineering assumptions are clear for the auditing engineer and updating the study is easier and MUCH more cost effective.

Just getting a PDF of the study makes the reproduction or updating much more costly. Most plants change over the course of the 5 years and sometimes even the utility feed/transformer will change, all of these factors affect the energy in the arc flash study. Adding motors, upgrading transformers, changing fuses or breaker settings can change the outcome of a study and require an update but if you don't have the digital files, the cost rises.

Additionally many plants gain value from a device coordination and arc flash mitigation study. Some of the "arc flash experts" don't provide these services and their studies don't gather adequate information to complete these types of value added services. Ask questions and gain expertise before putting out an RFP for an arc flash study.

Ask for one-line drawings. The study gives the company most of the information needed but often companies don't want to pay for this. It is required by NFPA 70E to have but not always provided in an arc flash study unless the customer asks for it.

If you need an arc flash study, protection coordination or a harmonics study of your power system, e-Hazard is always happy to assist.


New Member
Jan 30, 2018
Thank you for your answers. So as a follow-up, can one of our own on-site PE's perform the study? Or does it have to be an independent engineer who performs the study?
None of the consensus standards require third-party engineers for performing the study. The standard, however, is very clear in stating that:

"intended for use only by those experienced in power system studies".

If your site PE has experience on the software, power systems engineering, fault studies, and experience in IEEE 1584 and the NFPA 70E, then we would say that they may be able to perform the study. Some facilities have their own engineers, while many facilities have just outsourced this program. It is clearly more than just the arc flash study after all.

From a practical point of view, most engineers have other operations and projects related commitments and the study ends up being deferred. We have had to "fix" several studies where good plant engineers made worst case assumptions resulting in "false" lower arc flash energies.
Last edited by a moderator: