220.87 (2) - Determining Existing Loads

[parsonspd]

Does anyone know reasoning behind the 1.25x factor on top of max recorded demand discussed in the this section? I Have a customer I'm tryinig to explain this to; I can explain it's there, but why it's there, I can't. I'm wondering if this is an extra arbitrary margin, or is there some historical reasoning behind it?

 

[iwire]

I would say it is because the recorded values do not let you know if it is a continuous load or not so they are covering that possibility.

Of course that is just my guess.

 

[G._S._Ohm]

Does anyone know reasoning behind the 1.25x factor on top of max recorded demand discussed in the this section? I Have a customer I'm tryinig to explain this to; I can explain it's there, but why it's there, I can't. I'm wondering if this is an extra arbitrary margin, or is there some historical reasoning behind it?

Without seeing the exact text it looks like a safety factor.
Sizing it at 1.00x may cause an overload half the time.
Sizing it at 1.25x causes an overload 5% (or 1%?) of the time.
Sizing it at 2.00x causes, dunno', an overload 0.1% of the time.

I'd hope that this 1.25x factor is based on historical data and statistical analysis rather than just a WAG, because the customer's money is riding on this gamble.

E.g., HVAC equip. is sized on the Outside Design Temperature, which is exceeded only 1% of the time, depending on who you talk to.
It's based on historical data.
You can have 0.1% (99.9% certainty) too, but it will cost you.

While I have your attention, what's your WAG on the cost difference in your case between sizing for 1.00x and sizing for 1.25x?

 

[iwire]

Without seeing the exact text it looks like a safety factor.
Sizing it at 1.00x may cause an overload half the time.
Sizing it at 1.25x causes an overload 5% (or 1%?) of the time.
Sizing it at 2.00x causes, dunno', an overload 0.1% of the time.

I'd hope that this 1.25x factor is based on historical data and statistical analysis rather than just a WAG because the customer's money is riding on this gamble.

E.g., HVAC equip. is sized on the Outside Design Temperature, which is exceeded only 1% of the time, depending on who you talk to.
It's based on historical data.
You can have 0.1% (99.9% certainty) too, but it will cost you.

Did you even look at the code section?

The load is determined by recording the actual demand data for at least a 1 year period. Once you have that you add the 25%.

The NEC has some faults but under-sizing services, feeders, or branch circuits is not one of them. Typically the NEC load calculations run much higher than real life. The power company knows this and sizes their supplies accordingly.

 

[charlie b]

I don?t know the answer. But I suspect it is to provide a safety margin. You don?t know the actual loads, so you are making an educated guess on the basis of a period of load measurements. That won?t always give you the actual, maximum load that the building would ever experience. So they make you ?pretend? that the load measurement was 25% higher that you actually read, in order to allow for the possibility that if you had measured during some other time period you might have gotten a higher reading.

 

[G._S._Ohm]

Did you even look at the code section?

I don't have a copy of the current code, but. . .this is a statistical calculation. Once you have the average demand and the Standard Deviation of that demand you can adjust how much uncertainty you're willing to tolerate.

This 1.25x gives you 98% certainty that you won't be overloaded. . .provided that the SD of the demand is +/- 12%.
If the SD is +/- 1% the customer is paying for certainty that he doesn't need.

 

[hmspe]

What I've read is that this is a diversity factor. The basic calculation without this factorassumes that the phases are equally loaded. For panels with single phase loads it unlikely that the phases will have equal loads. The choice of 125% may be the result of a statistical analysis, but I'm more inclined to believe it was a number the CMP thought would be reasonable.

 

[iwire]

I don't have a copy of the current code, but. . .this is a statistical calculation. Once you have the average demand and the Standard Deviation of that demand you can adjust how much uncertainty you're willing to tolerate.

You can't make a calculation without the facts and expect it to be right. We don't use the average demand, we must use the maximum demand.

 

[parsonspd]

Without seeing the exact text it looks like a safety factor.
Sizing it at 1.00x may cause an overload half the time.
Sizing it at 1.25x causes an overload 5% (or 1%?) of the time.
Sizing it at 2.00x causes, dunno', an overload 0.1% of the time.

I'd hope that this 1.25x factor is based on historical data and statistical analysis rather than just a WAG, because the customer's money is riding on this gamble.

E.g., HVAC equip. is sized on the Outside Design Temperature, which is exceeded only 1% of the time, depending on who you talk to.
It's based on historical data.
You can have 0.1% (99.9% certainty) too, but it will cost you.

While I have your attention, what's your WAG on the cost difference in your case between sizing for 1.00x and sizing for 1.25x?

Thanks for the response. The WAG could be significant in this case; the customer will have to spend anywhere from $100 to $150k on new electrical equipment in the form of a new service...based on 1.25x. On 1.0x, he can use the service he has.

 

[parsonspd]

I would say it is because the recorded values do not let you know if it is a continuous load or not so they are covering that possibility.

Of course that is just my guess.

I was wondering too if that was the true, as services are sized 1.25x the max continuous load as calculated. However, in 230, there's an allowance to size services based on 1.0x if the service entrance overcurrent device is a 100% rated device. Since this provision didn't appear in 220.87, It made me wonder if there was a different reason for the 1.25x.

In my case, I have max demands over a year, reported max demands are at 15 min intervals. I would therefore think that the max demand is a non continuous load, and not subject to 1.25x. Again, this pointed to a different reason for the 1.25x.

 

[parsonspd]

What I've read is that this is a diversity factor. The basic calculation without this factorassumes that the phases are equally loaded. For panels with single phase loads it unlikely that the phases will have equal loads. The choice of 125% may be the result of a statistical analysis, but I'm more inclined to believe it was a number the CMP thought would be reasonable.

I don?t know the answer. But I suspect it is to provide a safety margin. You don?t know the actual loads, so you are making an educated guess on the basis of a period of load measurements. That won?t always give you the actual, maximum load that the building would ever experience. So they make you ?pretend? that the load measurement was 25% higher that you actually read, in order to allow for the possibility that if you had measured during some other time period you might have gotten a higher reading.

Did you even look at the code section?

The load is determined by recording the actual demand data for at least a 1 year period. Once you have that you add the 25%.

The NEC has some faults but under-sizing services, feeders, or branch circuits is not one of them. Typically the NEC load calculations run much higher than real life. The power company knows this and sizes their supplies accordingly.

Thanks for all the replies folks. This was my first post on the forum; I realized perhaps I should have been a little more detailed in my question....
My customer is an industrial setting; much of the equipment is pumps, fans; nearly all dynamic load. The existing service transformer is 2500kVA (utility owned), with a 480V, 3000A main breaker and 8-500's per phase underground. I haven't done the cable sizing calcs yet to get an idea on what my ampacity is, but I'm guessing right now about 3000A is a good number to start with. (My experience is that banked conduit runs underground can result in lower ampacities, so that is something I?ll be checking!) Proposed modifications will add roughly 260kVA in load.
Utility max demands, over a year, are 1719kW. Assuming 0.88pf and using 1.25x factor as per 220.87, I get 1719 x 1.25 / 0.88 = 2441kVA; or 2937A. More or less, this service is done per 220.87.
However, if I don?t use 1.25x, the results are significantly different; i.e., 1719kW/0.88 = 1953kVA, or 2349A.
A large outlay in capital has to happen in order to accommodate the additional 260kVA, as such, my customer was a little surprised to see 1.25x applied, with him thinking he had the max demands. I am looking for some good verbiage to explain why the 1.25 is needed?to go along with ?It's what the Code says to do?. Based on responses, seems like the consensus is leaning towards the 1.25x is a measure of margin to cover the fact that a years worth of utility data may not cover all possible scenarios.

 

[jrohe]

I suspect it has something to do with power factor. NEC 220.87 does not specify whether the units of the maximum demand are required to be in kW (real power) or kVA (apparent power). I think that with the CMP likely knowing that most utility companies record peak demand in kW, they decided that a 1.25 factor would be appropriate to account for power factors as low as 0.8.

 

[raberding]

1.25 AND .8 pf

1.25 AND .8 pf

I recently had a plan reviewer require the use of both the 125% demand factor AND a .8 PF.

and I could disagree with that...

 

[jrohe]

I recently had a plan reviewer require the use of both the 125% demand factor AND a .8 PF.

and I could disagree with that...

Ouch. I would have argued that one and asked where the NEC requires the kW demand load to be divided by the power factor and then that result multiplied by 125%. I am confident they would not be able to provide a code reference.

 

[parsonspd]

Makes the most sense

Makes the most sense

I suspect it has something to do with power factor. NEC 220.87 does not specify whether the units of the maximum demand are required to be in kW (real power) or kVA (apparent power). I think that with the CMP likely knowing that most utility companies record peak demand in kW, they decided that a 1.25 factor would be appropriate to account for power factors as low as 0.8.

Jason: That makes the most sense to me. Demand to me has always meant real power; i.e., kW....the load calcs have to be done in kVA; multiplying by 1.25x is a conservative way to get there. It just would have been nice if the paragraph clarified that; it would have saved all this time!!

 

[jrohe]

It just would have been nice if the paragraph clarified that; it would have saved all this time!!

I agree with you 100%. Or is that 125%?... Horrible joke. At any rate, nearly every other NFPA publication has an annex with explanatory material which clarifies the reasoning behind specific sections. It would be great if the NEC also had this annex, since it is, after all, an NFPA document. It would help everyone at least understand the intent behind some of the more ambiguous requirements and may prove to be especially valuable to those forum members who have had plan reviewers require peak demand data to account for both power factor and the 125% required by 220.87(2).

 

[charlie b]

I don't think that is the answer. I copied a form from the Washington State web site, the form that the inspectors want us to submit when we do a load addition using a 30 day measurement of existing load. The form starts with the maximum peak demand from the 30 days (in KW), then has us assign a power factor and divide the peak KW by that value (thus converting to KVA), and THEN we multiply by 1.25.


I am sticking with my original guess. But let me explain it a different way.

• Suppose we wanted to add a 50 amp, 480V, 3 phase, resistive heating load to an existing facility that has an 800 amp, 480/277 volt main service panel.
• The load is rated at about 42 KVA, and since pf =1 it is also 42 KW.
• The panel's rating is about 665 KVA.
• At an assumed 0.8 power factor, the panel's rating is about 532 KW.
• Suppose the maximum peak demand that we measured during a 30 day period was 490KW.
• We could say that it is acceptable to add the load, because the new 42 KW added to the measured maximum of 490 KW would give 532 KW, and that would not exceed the panel's rating.

I think that is simply cutting it too close. I think we are required to multiply the measured value by 1.25 in order to provide some margin for error.