Measuring total load of a 3-phase panel (delta-wye)

[Im_Butters]

Hi all,

Need help measuring the total load of a 3 phase receptacle panel. The building operator told me the load on the panel is:

Iph(A) = 20A
Iph(B) = 22A
Iph(C) = 21A

Is the total load the summation of all three phase currents, or is it the average of the three phases times the square root of 3? Can someone explain the principles behind it?

Thanks

 

[kwired]

Is it all line to neutral loads or a mixture of line to line and line to neutral, mix of single and three phase but all line to line....?

And what is voltage?

 

[Im_Butters]

Forgot to mention that it's 120/208V system. I would say it's a mixture of line to neutral, and line to line loads.

 

[kwired]

Forgot to mention that it's 120/208V system. I would say it's a mixture of line to neutral, and line to line loads.

Quick and simple calculation then is 20+22+21=63x120=7560VA But remember that is total load at that the time those measurements were taken and is not evenly divided across all three phases, though is pretty close to evenly divided.

 

[Im_Butters]

Thanks, this helps!

 

[kwired]

Also keep in mind that if it were say 21 amps and totally balanced three phase load it comes out slightly different, but maybe the more load the more it will make a difference. 21x208x1.732 = 7565 VA

But that previous method gets you in right ballpark a majority of the time.

 

[Strathead]

So, first off, let's define terms. Load is amperage, so you have 20A on each phase. The 1.732 that kwired just mentioned is the square root of 3. when you have three phases, each phase partially cancels out the other two phases similar to the concept that 120/240 single phase is 180 degrees out from each other and cancel completely. Since three phase is only partial, the amount of cancellation is the square root of three or 1.732. So your total load is 20 + 20 + 20/1.732=51 amps What kwired forgot is the your totally va is not 7565, it is 7565, it is 21x120x1.732 I think

 

[winnie]

In general one should never talk about the 'total amp load' of a panel. Not because you can't define a precise meaning of the term, but because there are several different plausible definitions, and everyone in the room will pick a different understanding. Net result is confusion.

You are much better off saying what the total VA on the panel is, and if there is significant imbalance, the VA on each phase

Jon

 

[LarryFine]

You are much better off saying what the total VA on the panel is, and if there is significant imbalance, the VA on each phase

Absolutely. Using VA makes the math much simpler.

If one phase is more heavily loaded, multiply that VA by three.

 

[kwired]

So, first off, let's define terms. Load is amperage, so you have 20A on each phase. The 1.732 that kwired just mentioned is the square root of 3. when you have three phases, each phase partially cancels out the other two phases similar to the concept that 120/240 single phase is 180 degrees out from each other and cancel completely. Since three phase is only partial, the amount of cancellation is the square root of three or 1.732. So your total load is 20 + 20 + 20/1.732=51 amps What kwired forgot is the your totally va is not 7565, it is 7565, it is 21x120x1.732 I think

No.

If all the load is 120 volts it is simply V x A = VA. (my first calculation based on OP's numbers.

My balanced three phase load was 21 amps @ 208 volts (= 4368) and I did multiply that by 1.732 to get 7565. But what throws this off is I averaged his figures to a 21 amp three phase load, where he actually had 7565 of balanced three phase load the current would been just a little under 21.

If you have 100 amps on each of three lines supplying nothing but line to neutral loads you have 12000 VA on each line x 3 lines = 36000 VA

If you have a 100 amp balanced three phase load you have 208 x 100 x 1.732 = 36025 VA nearly the same thing, rounding is what makes it slightly different. 120 x 1.732 is a little less than 208, plus square root of three itself is a rounded figure also - it is one those fractional numbers that never comes out even at any particular decimal position.

 

[tortuga]

For sizing an existing panel load I have always taken the highest loaded leg and multiplied that by the L-L voltage X 1.732
I also round up to a even VA
So I'd do 22A X (208V X 1.732) = 7926 VA

 

[LarryFine]

For sizing an existing panel load I have always taken the highest loaded leg and multiplied that by the L-L voltage X 1.732
I also round up to a even VA
So I'd do 22A X (208V X 1.732) = 7926 VA

Which then is tripled for sizing a transformer, etc.

 

[tortuga]

Which then is tripled for sizing a transformer, etc.

Not really for a 220.87 Exception calc, just data logger for 30 days and that number times 1.25

 

[kwired]

Which then is tripled for sizing a transformer, etc.

?

7926VA would be all the transformer needs to handle.

If you were building a three phase bank out of individual single phase units, you would divide 7926 by three to get minimum necessary size of individual units for that load.

Unless you were talking about something else I wasn't seeing.

 

[tortuga]

In Canada NEC 220.87 would be CEC 8-106 (8):

Where additional loads are to be added to an existing service or feeder, the augmented load shall
be permitted to be calculated by adding the sum of the additional loads, with demand factors as
permitted by this Code, to the maximum demand load of the existing installation as measured over
the most recent 12-month period, but the new calculated load shall be subject to Rule 8-104 5)
and 6).

Thats from the 2018 CEC

 

[LarryFine]

?

7926VA would be all the transformer needs to handle.

If you were building a three phase bank out of individual single phase units, you would divide 7926 by three to get minimum necessary size of individual units for that load.

Unless you were talking about something else I wasn't seeing.

I thought that was just for the most heavily-loaded phase. My bad.

 

[tortuga]

I thought that was just for the most heavily-loaded phase. My bad.

It is. The way I was taught was to do the voltage by sqrt of 3 not the current.
I think either way is fine, but for code they want maximum demand.