Future PI 110.15

[tortuga]

Open-delta configurations are not more likely to have voltage issues than any other undersized transformer.

In my experience open delta is the most commonly overloaded transformer setup here, part of the reason the local POCO wanted to quit offering them.
3 out of 4 of the transformer failures I have dealt with in recent memory were from projects exactly like the one felon describes.
Old industrial loft building with a open delta bank probably sized for whatever 240V motors were in there back in the day, the stinger pot usually much smaller than the lighting pot (1/3 the size), sparky just sees 400A 240V service.
You just dont ever see different transformer sizes in a wye.
Difficult to replace transformer bank, new loads mostly 240V single phase, improperly balanced (due to not possible to know open phase), a few months go by and boom.

 

[jim dungar]

In my experience open delta is the most commonly overloaded transformer setup here, part of the reason the local POCO wanted to quit offering them.
3 out of 4 of the transformer failures I have dealt with in recent memory were from projects exactly like the one felon describes.
Old industrial loft building with a open delta bank probably sized for whatever 240V motors were in there back in the day, the stinger pot usually much smaller than the lighting pot (1/3 the size), sparky just sees 400A 240V service.
You just dont ever see different transformer sizes in a wye.
Difficult to replace transformer bank, new loads mostly 240V single phase, improperly balanced (due to not possible to know open phase), a few months go by and boom.

It sounds like people are not following current code or notifying the utility when adding loads, additional code language is not likely to cause this problem to disappear.

 

[electrofelon]

It sounds like people are not following current code or notifying the utility when adding loads, additional code language is not likely to cause this problem to disappear.

Also I assume many utilities have demand meters on these and should be able to easily automatically flag an excessively loaded transformer bank.

 

[kwired]

Also I assume many utilities have demand meters on these and should be able to easily automatically flag an excessively loaded transformer bank.

They would need the software developed to notify them if/when it a particular metered application is excessively overloaded, which if they have demand based energy rates maybe they do already have this. Though is also possible they only monitor net VA and not VA per each phase conductor.

Open delta services around here is mostly fixed or limited load being supplied, or larger capacity single phase load but limited three phase loads being supplied. Farms with significant three phase motor loads being supplied are often supplied with a full delta with all three transformers of equal size, though the total demand and size or quantity of equipment has been increasing over the years and there are now many with 480/277 for service voltage anymore, particularly for facilities for grain storage and handling. Farming more acres leads to larger equipment like combines which leads to larger grain wagons needed to efficiently haul from the combine to the bin, larger wagon means larger transfer equipment or else you will be waiting too long to unload kind of lessening the reasons you got the bigger combines and wagons. Which also leads to needing 40 or 50 thousand bushel bins instead of 10K to be more efficient - and those need more HP for aeration fans to get the job done right... one thing just keeps leading to another.

 

[tortuga]

It sounds like people are not following current code

Perhaps you could give an example how to solve this problem without knowing the open phase set:

A customer needs to add a new 6kVA single phase 240V load.
Service is 240V open delta, with a 37.5 kVA power pot and a 100kVA lighting pot.
Utility provides load numbers for each Xfomer.
Going by article 220 calc for existing service 125% of existing load is 30kva and 97kva on each transformer respectively.
A-N and C-N has 120V to ground, B-N is 208V and marked orange.

Due to the nature of the panel all the A-C breaker spots are full so the only open spot are B-C or B-A, also load calc confirms A-C is 97kva so 6kva cant go there.
There are also two existing '240V' 9.6kW resistive heat strip loads, one is B-C connected and one is A-B
Which set of phases do you choose for the new load?

 

[jim dungar]

Perhaps you could give an example how to solve this problem without knowing the open phase set:

A customer needs to add a new 6kVA single phase 240V load.
Service is 240V open delta, with a 37.5 kVA power pot and a 100kVA lighting pot.
Utility provides load numbers for each Xfomer.
Going by article 220 calc for existing service 125% of existing load is 30kva and 97kva on each transformer respectively.
A-N and C-N has 120V to ground, B-N is 208V and marked orange.

Due to the nature of the panel all the A-C breaker spots are full so the only open spot are B-C or B-A, also load calc confirms A-C is 97kva so 6kva cant go there.
There are also two existing '240V' 9.6kW resistive heat strip loads, one is B-C connected and one is A-B
Which set of phases do you choose for the new load?

I would flip a coin. Or, take some field current measurements at the service.
Odds are the NEC loads calls are conservative and your loading is not close to the utility transformer capacity.
Odds are the NEC calculated new 6kVA load will not overload the utility transformer sufficiently enough to cause problems.
I would notify the utility of the additional load and remind them it is their job to provide the proper voltage regulation.

 

[tortuga]

If I am understanding the math right technically you cant add the 6kva load, With 9.6's kw single phase loads on A-B and B-C then that tells us there has to be a load on open set.
Therefore the Lighting pot load (A-C) can only be 95.2kva and the power pot can only be 32.7kva,
so the new load cannot be added to the power pot unless the 9.6kva load is moved off the open phase to the power pot (since its an existing load its already accounted for on there).
Only then could the 6kva load be added to the power pot.

 

[wwhitney]

If I am understanding the math right technically you cant add the 6kva load, With 9.6's kw single phase loads on A-B and B-C then that tells us there has to be a load on open set.
Therefore the Lighting pot load (A-C) can only be 95.2kva and the power pot can only be 32.7kva,

Not following your line of thought or how you came up with those numbers.

I took your statement that "Utility provides load numbers for each Xfomer" to mean that they have meters on each transformer. And that "125% of existing load is 30kva and 97kva on each transformer respectively" means the utility's measured peak current is 30 kVA/240V/125% = 100A on the power transformer, and 97 kVA/240V/125% = 323A on the lighting transformer.

In which case there would be no further reduction of the transformer capacities attributable to the single phase load on the open phase, as that load's current is already reflected in both the measured 100A and 323A figures.

Cheers, Wayne

 

[jim dungar]

The formulas I have used for open delta transformer sizing are:

kVA (lighting) = .58T + S
kVA (power) = .58T

where T = the 3-phase balanced load and S = 1-phase load

So your example of "37.5 kVA power pot and a 100kVA lighting pot" would tell me this bank can supply 64.6kVA three phase balanced 240V load and 62.5kVA 120/240V single phase load.

 

[wwhitney]

So your example of "37.5 kVA power pot and a 100kVA lighting pot" would tell me this bank can supply 64.6kVA three phase balanced 240V load and 62.5kVA 120/240V single phase load.

That's one combination of 3 phase and single phase (on the lighting pot only) loads that would fully utilize both transformers. And the formulas you gave (where 0.58 = 1/sqrt(3)) are good for sizing transformers when you treat the loading as fixed and put all the single phase loads on the lighting pot.

But if you treat the transformers as fixed, then there are other loading scenarios that would fully utilize both transformers. For example, if you have half as much 3 phase load (32.3 kVA instead of 64.6 kVA), you could put an additional 37.5/2 kVA of single phase loads on both the lighting pot and the power pot, for 37.5 kVA total of additional single phase loads. Or, if you don't know which phase of the delta is the open phase and which phase has the power pot, then you'd be limited to adding 32.3/3 kVA to each of the 3 phases, effectively making a 32.3 kVA 3-phase load.

And that's the point of the PI as I understand it, 37.5 > 32.3.

Cheers, Wayne

 

[jim dungar]

That's one combination of 3 phase and single phase (on the lighting pot only) loads that would fully utilize both transformers. And the formulas you gave (where 0.58 = 1/sqrt(3)) are good for sizing transformers when you treat the loading as fixed and put all the single phase loads on the lighting pot.

But if you treat the transformers as fixed, then there are other loading scenarios that would fully utilize both transformers. For example, if you have half as much 3 phase load (32.3 kVA instead of 64.6 kVA), you could put an additional 37.5/2 kVA of single phase loads on both the lighting pot and the power pot, for 37.5 kVA total of additional single phase loads. Or, if you don't know which phase of the delta is the open phase and which phase has the power pot, then you'd be limited to adding 32.3/3 kVA to each of the 3 phases, effectively making a 32.3 kVA 3-phase load.

And that's the point of the PI as I understand it, 37.5 > 32.3.

Cheers, Wayne

These are available source design issues, already covered by code articles.
The NEC makes no mention of transformer loading.
The utility would need to provide you with the proper voltage regulation based on their contract.
Single phase 240V loads may be distributed/balanced around the delta and count against the 3-phase loading rather than the 1-phase loading. Knowing the power pot is not important.

I see no need to add further code language for these circuits.

 

[tortuga]

Not following your line of thought or how you came up with those numbers.

I am using syncro's formula he posted here:

Trouble finding straight 240v breaker vs 120/240 breaker

There is no real derating necessary for using the 'open' phase. Typical formula for sizing the unequal transformers are: Power pot = .58T Lighting pot = .58T + S Where T is the total three phase load and S is the total single phase load. Note how the transformer sizing takes into account the...

forums.mikeholt.com

If you want to fully load the AB winding to its rated current IBmax when there's already a load current across the open jaw of IBC, then allowable load current across AB would be:
IAB = 0.5 x [ √ ( 4 I_Bmax² - 3 I_BC²) - I_BC ]

Similary, the allowable load current placed across AC when the AC winding is rated at ICmax and with a load current across the open jaw of IBC is:
IAC = 0.5 x [ √ ( 4 I_Cmax² - 3 I_BC²) - I_BC ]

Since the L-L voltages are the same, you can substitute kVA for all currents in these equations.

And yeah I might not be doing it right, but I put it in a spreadsheet and was able to reproduce his results.
Where
A-C is the 100kva lighting pot
Cmax² = 10000 X 4 = 40000.
A-B is the 37.5 power pot
Bmax² = 1406.25 X 4 = 5625.
and the B-C load is 9.6 kva
BC = 9.6 and 9.6² = 92.16.
92.6 X 3 = 276.48.
AC = 0.5 [sqrt (40000 - 276.48 ) - 9.6]
AC = 94.9 kva

then the same for the power pot

The utility would need to provide you with the proper voltage regulation based on their contract.

Well in the 3 cases of open deltas that went boom here the utility contract had the customer paying the 10-30k for transformer replacement, and put the loading issue right back in their lap.

 

[jim dungar]

Well in the 3 cases of open deltas that went boom here the utility contract had the customer paying the 10-30k for transformer replacement, and put the loading issue right back in their lap.

Did they advise the utility they were adding loads?

 

[tortuga]

Did they advise the utility they were adding loads?

Yes

 

[wwhitney]

I am using syncro's formula he posted here:

OK, but those formulas are for calculating the A-B and A-C coil currents on an open delta given known loading A-B, B-C, and A-C.

Whereas in your example I don't see what "Utility provides load numbers for each Xfomer" could mean other than that the utility has measured the actual peak coil currents A-B and A-C.

So those measured coil currents already include the effect of any load B-C across the open coil, there's no need to do further calculation because of the existing B-C load.

Cheers,
Wayne

 

[Hv&Lv]

Whereas in your example I don't see what "Utility provides load numbers for each Xfomer" could mean other than that the utility has measured the actual peak coil currents A-B and A-C.
Cheers,
Wayne

Yeah, we don’t (and won’t) do that.
Electrician gives us single phase loading, the three phase loading, we will build the bank by kVA with a diversity factor built in..

 

[jim dungar]

Yeah, we don’t (and won’t) do that.
Electrician gives us single phase loading, the three phase loading, we will build the bank by kVA with a diversity factor built in..

And if an existing bank is not sufficient, do you roll the dice or upgrade the bank? Who pays for changing the bank?

 

[wwhitney]

Yeah, we don’t (and won’t) do that.

Do you report, or at least capture, peak current on lines A, B, and C separately? For an open delta, that's all the info you need to determine peak coil loading, at least for L-L single phase and L-L-L three phase loads.

Cheers, Wayne

 

[Hv&Lv]

And if an existing bank is not sufficient, do you roll the dice or upgrade the bank? Who pays for changing the bank?

We keep up with all our transformer loading. Residential included.
It’s really not hard to do.
Take the demand readings on the transformers (from the meters), add diversity, plug it all in a data base and an overloaded transformer report spits out.

We upgrade on our dime. With increased loading you’re gonna pay anyway with increased kWh use.
We win by not blowing one up when a lineman re-fuses it a little higher because it blew on load and he thinks he knows things..

 

[Hv&Lv]

Do you report, or at least capture, peak current on lines A, B, and C separately? For an open delta, that's all the info you need to determine peak coil loading, at least for L-L single phase and L-L-L three phase loads.

Cheers, Wayne

Nope. Only at the substation level

Doing that at each bank would require expensive CTs at each bank. Then someone has to install them, get the readings or purchase expensive equipment to send the readings into our office through SCADA.
It would be cheaper to let it blow up and then replace it.