arc flash study based on assumed information

billk554

Member
our company recently had an arc flash study done. we have the results back and in the beginning of the report is a paragraph that say something like " arc flash studies are done with the most complete data available" in more words than that but that's the idea. two paragraphs below it states that the data used in our study is assumed based on the overcurrent device rating should have the proper sized conductors for what ever the device is rated at. the person doing the study would not open one panel cover. he told me if I opened a panel cover (for inspection purposes) he would have to leave the building. this was done by a highly recognized engineering firm associated with our electrical supply company. I am thinking I may have to throw this study out and start all over again. What if any opinion does this page have on this?
 

Hv&Lv

Senior Member
Location
-
Occupation
Engineer/Technician
ASSumptions...
especially when someone’s life may depend on these numbers.
 

MyCleveland

Senior Member
Location
Cleveland, Ohio
I was given a similar report on a building renovation project.
This firm did not even look at the SCCR of ANY panel either.
Informed customer of about six panels where SCCR values were less than required.
They ended up replacing panels, but was never told if they were refunded any monies for the report or if it was ever redone.
 

jim dungar

Moderator
Staff member
Location
Wisconsin
Occupation
Retired Electrical Engineer - Power Systems
Do you believe the circuit conductors were not sized and installed per the NEC?

At least one software program allows you to run the analysis using different conductor sizes. Maybe the engineering firm also considered conductors one size smaller as well as one size larger?

Have you talked to the engineer?

I know of several firms that consider it 'live work' to remove a panel cover simply to look at the conductors.
 

MyCleveland

Senior Member
Location
Cleveland, Ohio
well the engineering firm is Schnieder and we used them through a supply house that is a square d distributor.
I am sorry, I thought you were the supply house.

I called the EE that generated report, and was told they had no one-line diagram and the owner did not want to pay for a SCC study because the building was only a few years old assumption being everything was done correctly.

Result…very nice arcflash-labels on panels it was not safe to be in front of to read.

Hopefully others can provide advice to help negotiate with Sqd.
 

jim dungar

Moderator
Staff member
Location
Wisconsin
Occupation
Retired Electrical Engineer - Power Systems
I am sorry, I thought you were the supply house.

I called the EE that generated report, and was told they had no one-line diagram and the owner did not want to pay for a SCC study because the building was only a few years old assumption being everything was done correctly.

Result…very nice arcflash-labels on panels it was not safe to be in front of to read.

Hopefully others can provide advice to help negotiate with Sqd.
What do you mean it is not safe to be in front of the panel?

Schneider has been following the policy of not opening panels for at least 4 years. So what should they have done if the customer was not willing to provide the data or paid to have it collected?

I usually say if you guess at the input data, you might as well just guess at the results. But, I also know Schneider's methodology was not arrived at lightly or quickly.
 

MyCleveland

Senior Member
Location
Cleveland, Ohio
What do you mean it is not safe to be in front of the panel?

Schneider has been following the policy of not opening panels for at least 4 years. So what should they have done if the customer was not willing to provide the data or paid to have it collected?

I usually say if you guess at the input data, you might as well just guess at the results. But, I also know Schneider's methodology was not arrived at lightly or quickly.
Sorry Jim & all
I meant to say work on or in. I was discussing the project I was involved with.
If a panel is rated say 14k and AFC available at that location was 28k, no series ratings involved, and panel has an arc flash label with any values you want is it not unsafe for an EC to work on live.

I agree with EEs not opening anything, I always insist on owners maintenance or EC assist in data gathering. Don,t have the data the report in worthless.

I agree with last statement first sentence. I am just sharing a similar event.
 

paulengr

Senior Member
our company recently had an arc flash study done. we have the results back and in the beginning of the report is a paragraph that say something like " arc flash studies are done with the most complete data available" in more words than that but that's the idea. two paragraphs below it states that the data used in our study is assumed based on the overcurrent device rating should have the proper sized conductors for what ever the device is rated at. the person doing the study would not open one panel cover. he told me if I opened a panel cover (for inspection purposes) he would have to leave the building. this was done by a highly recognized engineering firm associated with our electrical supply company. I am thinking I may have to throw this study out and start all over again. What if any opinion does this page have on this?

You should report this guy to the state board and gave their license revoked. I would not pay for it. IEEE 1584 has a second part detailing what is required. Electrical equipment is not inherently dangerous. It’s only when the activities are such that they can cause an arc. As per 70E. That guy should leave as he recommended. Highly suggest meeting with the principals at the firm. If they can’t do it right they don’t do it. Any idiot can punch wrong data in a computer.
 

jim dungar

Moderator
Staff member
Location
Wisconsin
Occupation
Retired Electrical Engineer - Power Systems
You should report this guy to the state board and gave their license revoked. I would not pay for it. IEEE 1584 has a second part detailing what is required. Electrical equipment is not inherently dangerous. It’s only when the activities are such that they can cause an arc. As per 70E. That guy should leave as he recommended. Highly suggest meeting with the principals at the firm. If they can’t do it right they don’t do it. Any idiot can punch wrong data in a computer.
Where do you find that opening an energized panel is allowed per NFPA70E, as it is not part of troubleshooting or testing?

If the data collection is live work then a work permit would need to be signed by all levels of management.

Yes, meet with the firm. Their written quotation, for at least the past four years, has specified this methodology would be used.
 

paulengr

Senior Member
Where do you find that opening an energized panel is allowed per NFPA70E, as it is not part of troubleshooting or testing?

If the data collection is live work then a work permit would need to be signed by all levels of management.

Yes, meet with the firm. Their written quotation, for at least the past four years, has specified this methodology would be used.

Visual inspection is testing. And it’s a whole lot safer than taking measurements since it is noncontact. Visual inspection of molded case circuit breakers is required as part of annual testing per NEMA AB-4 annually for example. It is required as part of NFPA 70B which NFPA 70E requires. Visual inspection is inherently one of the most critical aspects of troubleshooting as well as routine testing and maintenance. Most insurance companies are requiring infrared inspections and NFPA 70B specifically mentions that too as does NETA MTS.

Ignorance of basic inspection and troubleshooting procedures and methods demonstrates lack of qualification, both to meet 70E requirements as well as OSHA 1919.269, and NFPA 70B.

A fundamental lack of understanding that overestimating available short circuit current results in significantly underestimating arc flash hazards clearly demonstrates lack of qualification to perform an arc flash study. In the case of a so-called professional engineer delivering a report to someone who may not have such a lofty title by hiding behind an extremely unsafe assumption that can potentially drastically underestimate the risk is putting the safety of others at risk. That’s a clear violation of NCEES Code of Ethics. I don’t know how much more clear I can be. As I said, I would never pay for such garbage and at a bare minimum the principal engineer and I would have a heated discussion about the next step. It would get reported to the state board and no matter what the contract terms say I’m not paying for it unless some idiot agreed to those terms and even then I don’t see where I should pay for something that is nothing but a huge liability issue.

I can understand where compromises sometimes must be made in light of a lack of information or the ability to get it. Cable sizes especially with fine strand do vary a lot more than people realize and trying to get accurate measurements with a plastic caliper is sometimes challenging but not checking cable sizes is just as bad as using the infinite bus assumption on transformers, or using the table method instead. They all have their weaknesses. This is where the engineer and customer have to work together to evaluate the alternatives and make a mutually acceptable decision,
 

Strathead

Senior Member
Do you believe the circuit conductors were not sized and installed per the NEC?

At least one software program allows you to run the analysis using different conductor sizes. Maybe the engineering firm also considered conductors one size smaller as well as one size larger?

Have you talked to the engineer?

I know of several firms that consider it 'live work' to remove a panel cover simply to look at the conductors.
Correct me if I am wrong, but say, the conductors were 600KCMIL for a 400 amp panel, and the assumption is that 500 was used, there would be a marginal difference. So CYA. Also an assumption always made is wire length. I would assume they usually underestimate the length of the wire. Both of these assumptions would cause the calculated AIC to be higher than the actual, a potential inconvenience, but not a safety issue. Still CYA
 

jim dungar

Moderator
Staff member
Location
Wisconsin
Occupation
Retired Electrical Engineer - Power Systems
Correct me if I am wrong, but say, the conductors were 600KCMIL for a 400 amp panel, and the assumption is that 500 was used, there would be a marginal difference. So CYA. Also an assumption always made is wire length. I would assume they usually underestimate the length of the wire. Both of these assumptions would cause the calculated AIC to be higher than the actual, a potential inconvenience, but not a safety issue. Still CYA
But what if the study was done using 350, 500, and 600kcmil then the worst case value was used. Cable side can kind of be checked via the conduit, 500 won't fit in a single 2inch EMT.
 

jim dungar

Moderator
Staff member
Location
Wisconsin
Occupation
Retired Electrical Engineer - Power Systems
Visual inspection is testing. And it’s a whole lot safer than taking measurements since it is noncontact. Visual inspection of molded case circuit breakers is required as part of annual testing per NEMA AB-4 annually for example. It is required as part of NFPA 70B which NFPA 70E requires. Visual inspection is inherently one of the most critical aspects of troubleshooting as well as routine testing and maintenance. Most insurance companies are requiring infrared inspections and NFPA 70B specifically mentions that too as does NETA MTS.
If your company legal department says they view data collection as energized work and not testing, are you going to do it anyway just because other people do?

How many near misses have you heard about simply from trying to collect data? Ever had cover fasteners fall off into a panel, pinched a conductor with a clamp, poked into insulation with a screw, had a bypass switch fail causing an product line to shut down?

Schedule an outage and create a complete one-line. Future maintenance and expansion will likely be easier.

Don't accept any report results without first understanding the process used to produce it. Talk to the engineering firm don't simply say they are incapable or ignorant until you understand their procedure. Maybe they ran a scenario with 10 different wires and reported the worst case, but simply used the NEC sized conductors on the one-line. Maybe they round all incident energies up to the values of 1.2, 8.0, or, 40.0 cal/cm^2 for simplified PPE selection per NFPA70E suggestions?
 

paulengr

Senior Member
Correct me if I am wrong, but say, the conductors were 600KCMIL for a 400 amp panel, and the assumption is that 500 was used, there would be a marginal difference. So CYA. Also an assumption always made is wire length. I would assume they usually underestimate the length of the wire. Both of these assumptions would cause the calculated AIC to be higher than the actual, a potential inconvenience, but not a safety issue. Still CYA

Incorrect.


As fault current is reduced, trip times increase. Although it is certain not linear incident power (ignoring time for a second) is roughly linear with current. However trip time is inversely proportional to current so it increases exponentially. So on an inverse time curve the net result is that lower fault current results in increased incident energy, and the increase is substantial.

Try it yourself. Increase a cable length by say 10% when within the inverse time region and watch your incident energy. It is NOT nitpicking. It makes a huge difference.

Also cable inductance varies a lot with wire gauge. This should not be taken lightly either.

It’s one thing to mistake 500 for 400 let alone 535 it quite another to use say 500 vs. double 350s because of a chart. And aluminum vs steel vs PVC conduit makes a substantial difference too. And it’s a big deal if you are say looking at the feeder from a 1500 kVA, 480 V secondary where every foot drops incident energy by quite a bit compared to a couple hundred feet away. In fact it’s not the high fault current that matters. It is going to be above the instantaneous settings or in the short term tripping. Usually arc flash is relatively harmless because trip times are short. It’s when trip times get out to 100+ ms that these minor details matter.
 

Strathead

Senior Member
Incorrect.


As fault current is reduced, trip times increase. Although it is certain not linear incident power (ignoring time for a second) is roughly linear with current. However trip time is inversely proportional to current so it increases exponentially. So on an inverse time curve the net result is that lower fault current results in increased incident energy, and the increase is substantial.

Try it yourself. Increase a cable length by say 10% when within the inverse time region and watch your incident energy. It is NOT nitpicking. It makes a huge difference.

Also cable inductance varies a lot with wire gauge. This should not be taken lightly either.

It’s one thing to mistake 500 for 400 let alone 535 it quite another to use say 500 vs. double 350s because of a chart. And aluminum vs steel vs PVC conduit makes a substantial difference too. And it’s a big deal if you are say looking at the feeder from a 1500 kVA, 480 V secondary where every foot drops incident energy by quite a bit compared to a couple hundred feet away. In fact it’s not the high fault current that matters. It is going to be above the instantaneous settings or in the short term tripping. Usually arc flash is relatively harmless because trip times are short. It’s when trip times get out to 100+ ms that these minor details matter.
One shouldn't be in the inverse time range of a circuit breaker when one is approaching the arc flash rating of the circuit. Instantaneous trip is generally about 10 times the breaker rating, so even a 10KAIC breaker rated 1000 amps (not likely) will not have the problem you are expressing.
 

paulengr

Senior Member
One shouldn't be in the inverse time range of a circuit breaker when one is approaching the arc flash rating of the circuit. Instantaneous trip is generally about 10 times the breaker rating, so even a 10KAIC breaker rated 1000 amps (not likely) will not have the problem you are expressing.


A common example is this. Say you have several distribution panels, panel boards or switchboards. Those in turn are fed by switchgear feeder breakers, which are in turn fed by at least one main breaker off the plant main transformer. So at the switchboard level we must use UL 489 breakers. Coordination is due to series rating and everything can or preferably should trip “instantaneously”. Now the feeder breakers have to coordinate too so either we have to delay the instantaneous by 3-4 cycles to allow for breaker tolerances or use short or long term trip settings (inverse time). By far most do not use delayed instantaneous. The same trip mechanism is used at the main(s) and tie(s) so we need to add another 3-4 cycles. At this point we are looking easily at 10-12 cycles or around 160-200 milliseconds trip time at a minimum. So at a point where short circuit current is at the highest for the whole plant, trip times are the longest.

Now your response will probably be using a maintenance switch. This sounds nice in theory but you are also disabling coordination. That can be banned by regulation (hospital Code for instance). And it may not be possible on some equipment like old breakers with integral trip units. Plus do you really want to disable coordination on the breaker that feeds the entire plant?

Add in long cable lengths and large motors, fault currents can be very low and trip times excessively long by nature.
 
Top