Coordination study dependent upon exact feeder cable length?

[Isaiah]

I was told today that the coordination study had not yet been done because they on didn’t have the exact cable length of 480V feeders.
Yet the one lines have all been completed - from switchgear to MCCs, Panelboards and Transformers with CB sizes and trip settings assigned for protection.
It seems odd that they were unable to perform the study.
In the past I’ve seen estimated feeder lengths utilized to determine short circuit and CB/Fuse sizes and the studies were done well in advance


Enviado desde mi iPhone utilizando Tapatalk

 

[jim dungar]

If you estimate the feeder lengths you are estimating the short circuit currents.
If you over estimate the current, your coordination may show protective devices opening early, which can make coordination difficult. If you under estimate short circuit current it may appear you have coordination when it doesn't exist.

So, how close are your cable length estimates? How close does your coordination need to be. Can you check results using high and low estimates.

I have performed many coordination studies using estimated cable lengths.

 

[Isaiah]

If you estimate the feeder lengths you are estimating the short circuit currents.
If you over estimate the current, your coordination may show protective devices opening early, which can make coordination difficult. If you under estimate short circuit current it may appear you have coordination when it doesn't exist.

So, how close are your cable length estimates? How close does your coordination need to be. Can you check results using high and low estimates.

I have performed many coordination studies using estimated cable lengths.

If I said the lengths were +/- 20’; would this be good enough for an accurate study?


Enviado desde mi iPhone utilizando Tapatalk

 

[jim dungar]

If I said the lengths were +/- 20’; would this be good enough for an accurate study?


Enviado desde mi iPhone utilizando Tapatalk

If the runs were 1000' of #500kcmil, probably not.
If the runs were 50' of #12 definitely yes.

Typically I would ignore fault currents for branches (these OCPD are usually not adjustable, so their curves are what they are) and assume high fault currents for feeders. For the majority of circuits, if you coordinate at high fault levels you will also coordinate at low level.
If you have coordination issues on a few circuits, trying refining your estimated length. On more than one study I was able to say something like, coordination exists as long as the circuit length is a minimum of XX'.

 

[Isaiah]

If the runs were 1000' of #500kcmil, probably not.
If the runs were 50' of #12 definitely yes.

Typically I would ignore fault currents for branches (these OCPD are usually not adjustable, so their curves are what they are) and assume high fault currents for feeders. For the majority of circuits, if you coordinate at high fault levels you will also coordinate at low level.
If you have coordination issues on a few circuits, trying refining your estimated length. On more than one study I was able to say something like, coordination exists as long as the circuit length is a minimum of XX'.

The runs are all 500kcmil parallel feeders but closer to 500 feet vs 1000’


Enviado desde mi iPhone utilizando Tapatalk

 

[Deltaforce]

The runs are all 500kcmil parallel feeders but closer to 500 feet vs 1000’


Enviado desde mi iPhone utilizando Tapatalk

Check with 5 ft difference.
Include any motor contribution to fault current
Account impedance nonlinearity for large size cables

 

[Isaiah]

If the runs were 1000' of #500kcmil, probably not.
If the runs were 50' of #12 definitely yes.

Typically I would ignore fault currents for branches (these OCPD are usually not adjustable, so their curves are what they are) and assume high fault currents for feeders. For the majority of circuits, if you coordinate at high fault levels you will also coordinate at low level.
If you have coordination issues on a few circuits, trying refining your estimated length. On more than one study I was able to say something like, coordination exists as long as the circuit length is a minimum of XX'.

I’m trying to make the argument we don’t need to wait until we have the exact feeder lengths to do the study. I’d like to have Arc-Flash labeling done in advance rather than wait another year


Enviado desde mi iPhone utilizando Tapatalk

 

[jim dungar]

I’m trying to make the argument we don’t need to wait until we have the exact feeder lengths to do the study. I’d like to have Arc-Flash labeling done in advance rather than wait another year


Enviado desde mi iPhone utilizando Tapatalk

Coordination is easy, unless you need to meet 100% level. It needs high fault current values. Arc flash becomes harder, it needs low fault current values.
So you basically need to perform your studies using two different sets of estimates. Now you are performing 4 or 5 studies instead of just one. Who is paying for this extra engineering?
BTW, I think my record was 18 different scenarios before we felt comfortable with the results.

If you want arc flash values you must have most of the system already installed. How many cable lengths are you missing?
Have you thought about not putting exact incident energy values on your labels? By rounding them up, to say 4, 8, or25 cal/cm², you might not need to change labels unless the system had significant changes.

I agree with wanting to have labels on as soon as stuff is energized.

 

[Deltaforce]

Application of a known load and determining fault current from voltage drop at required points, the standards permit(IEEE 141 Annex D), the uncertainty in conductor lengths be thus avoided

 

[jim dungar]

Application of a known load and determining fault current from voltage drop at required points, the standards permit(IEEE 141 Annex D), the uncertainty in conductor lengths be thus avoided

This is a design level analysis. If the cabe, feeding the load, was installed they would know its length.

 

[D. Castor]

For a typical system study of an installed system (short circuit, coordination, arc-flash), we generally advise that cable length to within 10% is accurate enough. Cable lengths are generally NOT documented on one-lines or anywhere else, in my experience.

The lower the system voltage, the more important cable lengths become. When we used to do short circuit calcs by hand, medium-voltage cable lengths were routinely ignored.