Arc Flash Survey

[mivey]

My State disagrees with you. My OP is about learning the "Process" of
doing a Arc Flash Survey. Nothing more nothing less. I've already compiled
29 hours of time and material in the understanding of the "Process."
I was simply looking for "Time-Saving" tips or words of wisdom from those that
do it professionally. Other than Stephen. I've only read posts that second guessed my education and suggested I'm not smart enough to learn such
a task. I'm really cranky from being up all night reviewing material on studies
and formulas. I will tolerate you much better after a nap. :grin: Just
kidding.

Well, you did say you were going to perform arc flash studies. I have no doubt you can learn to do it.

I also have no doubt that your boss will not appreciate what you will have to do once you learn how. A lot of bosses think most complicated tasks are less involved than what they really are. They might take your word for it but unless they get involved, will be impatient for you to finish, and will be appalled at how much it costs.

 

[Karl H]

Well, you did say you were going to perform arc flash studies. I have no doubt you can learn to do it.

I also have no doubt that your boss will not appreciate what you will have to do once you learn how. A lot of bosses think most complicated tasks are less involved than what they really are. They might take your word for it but unless they get involved, will be impatient for you to finish, and will be appalled at how much it costs.

Yes Your right, I went thru the" Impatience" thing once I spent two years
learning AutoCAD then demonstrating the step by step process of
doing a simple electrical drawing. It looks like Drag-in-drop but, you don't
see the 50 steps before you can actually drag -n-drop. :grin: Mivey,
have you done a Arc Flash Survey?

 

[TxEngr]

Karl,

I applaud your efforts to learn the process of an arc flash study. Even if you were to outside contract the work, your knowledge of the process would be an asset in evaluating the work of the contractor and extremely helpful to the contractor. I agree with you about doing work in-house not requiring a P.E. In the states I have worked in for the last 20 years, if you performed the work for internal use, then you didn't have to be a P.E. I also think our P.E. system here is screwed up and a model like Canada uses would be much better. Then every engineer with 5 years experience would have a P.E. or pretty much be out of the business. But that's another topic and I don't want to hijack the thread.

The articles by Brainfiller are great articles and give you a good overview. Learning to do the calcs by hand is a good thing, but if a study is of any size, then the computer software is the way to go. Having the software and doing the study yourself is also helpful when you make changes to the system (or are planning changes) and you can rerun the model with the changes to see the impact - all without having to pay an outside consultant again who may or may not still be in the business.

The gathering of the data for the study is the most time consuming part of the work. Cable sizes and lengths must be verified, breakers and relay settings confirmed and much of this data is buried in files (hopefully) that is many years old. It is tedious work and eats large amounts of time. It seems the one-lines are never up to date to the extent needed for a study.

Building the model in the software can be a bit tedious, but if you have all of the data in front of you, it really helps. Running the numbers in the software is easy - almost too easy. I typically run small sections at a time and make sure the numbers make sense. That part is what comes with experience and what many of the posters are concerned about. It comes from doing it a few times and maybe having a guiding hand to help you with questions. Perhaps a consultant could be used at this point to help out. But even they had to start somewhere so I encourage you to persevere. Recoginze your limitations and ask questions when necessary.

Good luck with your efforts and keep us posted.

TxEngr

 

[jghrist]

You should get a copy of the IEEE Red Book, Std 141, IEEE Recommended Practices for Electric Power Distribution for Industrial Plants and IEEE Brown Book Std 399, IEEE Recommended Practices for Industrial and Commercial Power Systems Analysis to help with the fault current calculations. Then get IEEE Std 1584, IEEE Guide for Performing Arc-Flash Hazard Calculations to explain the incident energy and flash boundary calculations.

 

[jim dungar]

When performing SC calculations, it is not uncommon to take short cuts, such as ignoring cable lengths. Most of these are intended to generate a high (safer?) SC value. However, when performing arc flash calculations a artificially high SC value may result in an unrealistically fast clearing time and a lower than actual incident energy.

Because of the uncertainties with available utility fault currents, a good practice is to run several arc fault scenarios using a range of SC values. The report can be issued showing only the worst case incident energy value for each bus analyzed.

 

[Karl H]

When performing SC calculations, it is not uncommon to take short cuts, such as ignoring cable lengths. Most of these are intended to generate a high (safer?) SC value. However, when performing arc flash calculations a artificially high SC value may result in an unrealistically fast clearing time and a lower than actual incident energy.

I ran into this issue when I was learning how to do a SC. I had a desk full of
SC and coordination studies for existing plants that I manage. Since I had
access to the actual plant site I would do my own survey and compare
mine to the EE's results. On a few instances my numbers would not add
up to the original study. The mistake I found most frequently
were the length of the conductors which lead to a few trips being around
15 to 30% delay compared to the original calcs. That's why I find it important
to learn to do calcs by hand before you trust a computer program.

 

[jim dungar]

That's why I find it important to learn to do calcs by hand before you trust a computer program.

Doing calcs by hand vs. by computer would not flag cable length approximation errors created during the 'data collection' process.

It is not possible to create a good report from bad data.

What I was trying to point out is, experience in SC does not make one experienced in arc flash, in many cases the required system assumptions are polar opposites. Look for a support group of some type (like the members here) for assumption making. But remember, if you ask 3 engineers a question you may get 4 answers.

 

[Karl H]

Doing calcs by hand vs. by computer would not flag cable length approximation errors created during the 'data collection' process.

It is not possible to create a good report from bad data.

What I was trying to point out is, experience in SC does not make one experienced in arc flash, in many cases the required system assumptions are polar opposites. Look for a support group of some type (like the members here) for assumption making. But remember, if you ask 3 engineers a question you may get 4 answers.

I completely understood what you were saying. I didn't explain what I was
saying. I tested my knowledge by trying to re-create or come to the same
conclusion of other EE's calcs on a certain system, by taking the real-world
data and doing the calcs on my own. I found in some cases the conductor
length was not "Real" so my calc's did not match the original calcs,as my lengths and sizes were "Real-world". So my numbers would never match
the EE's. So that taught me in "Some" cases the available fault current
would have been way much higher than the report suggested.

So when calculating Incident energy at a point I need to realize to look
at all impedance at All points and in order to be accurate it Has to "Real-
world" and not an assumption. The SC education taught me how important
length is , so with my experience with length and SC, I can apply that
knowledge to Arc-Flash.

 

[StephenSDH]

The most critical part of arc flash studies are transformers, OCPD, and generators. You will find that wire lengths do not vary the results much. Once you cross over a transformer the primaries wire length becomes even less critical. Ultimately you have to approximate the cable length. As long as you get out there with you wheel and follow the conduits you can get within +/- 20'.

If you are competent enough to perform a coordination study you almost there complete an arc flash study.

 

[Karl H]

The most critical part of arc flash studies are transformers, OCPD, and generators. You will find that wire lengths do not vary the results much. Once you cross over a transformer the primaries wire length becomes even less critical. Ultimately you have to approximate the cable length. As long as you get out there with you wheel and follow the conduits you can get within +/- 20'.

If you are competent enough to perform a coordination study you almost there complete an arc flash study.

You made a real good point in this post. Transformers ,and OCPD's are a given. You know, I didn't think about the generators. It's so obvious yet I didn't even consider the generators. Thank You.

 

[ptrip]

The most critical part of arc flash studies are transformers, OCPD, and generators. You will find that wire lengths do not vary the results much. Once you cross over a transformer the primaries wire length becomes even less critical. Ultimately you have to approximate the cable length. As long as you get out there with you wheel and follow the conduits you can get within +/- 20'.

If you are competent enough to perform a coordination study you almost there complete an arc flash study.

Conductor lengths don't have to be perfect and +/- 20' is a good estimation ... but saying it's not critical reduces the impact that conductor length can have on an arc flash study.

I spent last summer performing arc flash studies for a company (and I have to say that I was terribly unprepared and uneducated) and found in a couple of cases that a simple 208V 3ph panel ended up having a higher than expected incident energy because of the distance from its source (a "mere" 100'). "They" (whoever they really are) say you can ignore 120/208V panels after a 125kva or smaller transformer ... but if there's signficant distance involved ... include it in the study!

 

[zog]

"They" (whoever they really are)

In this case they would be the IEEE 1584 Commitee

say you can ignore 120/208V panels after a 125kva or smaller transformer ... but if there's signficant distance involved ... include it in the study!

You are assuming the Ei is based on an arcing fault that requires an OCPD to clear it, the basis of that cutoff pont is the arc will not self sustain so the calulated values do nto apply.

Of course, that is a big assumption and further testing will give us more data on this, I have doubts it is correct.

 

[steve66]

Here is a ton of information on arc flash:

http://www.easypower.com/arc_flash/arc_flash_resource_center.php

Steve