Service Calculation - NEC 220 or not

[elec_eng]

For design engineers out there, how do you perform service calculation in early stage of the design? Especially new construction with big square footage such as, 200K+ SF. This is very important calculation because this is necessary for the space planning in early stage.

I can’t really use NEC 220 in SD calcs because 220 is based on the connected loads and they are not available in that early stage.

So when I do the load calcs in SD or/and early DD, I use the historic data from the pervious similar projects or other technical references, such as IEEE, ASHRAE depending on the type of the projects and that gives me the pretty good watt/sf value per each type of building. Of course, I always give myself some extra capacity and accounts for the future capacity as well.

Is this code violation since I did not use the 220 method?

I have run into situations where the contractors adding all the panel schedule loads (from the construction set) and my service size doesn’t meet the NEC 220.
For example, for a typical hospital, our historic data (and IEEE references) shows that the building loads are avg. some where around 8 – 10 watts/sf, most times less than 10. So my service size is some where around 13 – 15 watt/sf depending on the location (heating fuel type) and the chiller efficiency (water vs air) etc. We have actually reviewed the utility used of several hospital loads after completion and their peak loads have been well under the 8 watt/sf, most of times. This is an example of the hospitals but other type of building loads are well below that.

But when you add up the panel schedule loads (connected) at the end of the design (CD), it always comes out to be higher than the first estimated watt/sf load, even after the demand factor allowed in NEC 220. Then the contractors claims the building service is under sized per the NEC 220 and we have to upsize the service. But we know it is already oversized and the utility info can support that.

So far, I have talked to couple of AHJs and they accepted my calculations but this question keeps coming up so I just want to see what you guys are doing out there in case there are some AHJs won't accept my approach.

I know there is some leeway in NEC 517.30(D) for the generator sizing but I don’t think there are any exceptions on the service calculation.

Any thoughts?

 

[ron]

I have a similar method to you, although I don't have good historical or standardized w/sf.

Often if there will be lots of electrical infrastructure, like heat or HWH, etc, and tight layout like office space, I use 20 W/sf.
For general warehouse or storage, I use 3-6 W/sf.

The AHJ has't questioned me in the past because a core and shell just doesn't have 220 info.

What type of IEEE or ASHRAE references have you seen that give guidance?

 

[charlie b]

We use something similar. The SD load calculation is on a W/SF basis. But we have not had the experience you describe, of having the final panel schedule loads significantly higher than the initial W/SF numbers. Perhaps we are shooting a bit higher than you are. Our calculation is also more granular, in that we separately calculate lighting, mechanical, and plug loads on a W/SF basis, and we separately calculate by the occupancy of each floor (e.g., retail versus office versus residential). Please note that you can't use the utility numbers as the basis for your calculation. As design engineers, we need to use 220, even though we know that the real load will be much lower.

 

[elec_eng]

I have a similar method to you, although I don't have good historical or standardized w/sf.

Often if there will be lots of electrical infrastructure, like heat or HWH, etc, and tight layout like office space, I use 20 W/sf.
For general warehouse or storage, I use 3-6 W/sf.

The AHJ has't questioned me in the past because a core and shell just doesn't have 220 info.

What type of IEEE or ASHRAE references have you seen that give guidance?

IEEE White Book for hospital and Gray book for the commercial buildings..i found their numbers are very close to our historic data. you will be surprised that most of hospital utility data we collected, never exceeded 10w/sf. Most of time less than 8.... IEEE color books provide w/sf value (low/avg/high) per various building type.


I use NEC for lighting worst case and ASHRAE 90.1 or IECC for probable lighting load (I think this is allowed in newer version of NEC). I tried to pick somewhere between depending on the project. ASHRAE 90.1 also gives you the receptacle w/sf value per building type. I think it was in the ASHRAE handbook.

 

[ron]

IEEE White Book for hospital and Gray book for the commercial buildings..

Just looked through Chapter 2 of the Gray Book. I didn't realize what I had available (in reality, I haven't looked at it in years).
Thanks

That's better than my WAG of 3 - 20 W/sf

 

[elec_eng]

We use something similar. The SD load calculation is on a W/SF basis. But we have not had the experience you describe, of having the final panel schedule loads significantly higher than the initial W/SF numbers. Perhaps we are shooting a bit higher than you are. Our calculation is also more granular, in that we separately calculate lighting, mechanical, and plug loads on a W/SF basis, and we separately calculate by the occupancy of each floor (e.g., retail versus office versus residential). Please note that you can't use the utility numbers as the basis for your calculation. As design engineers, we need to use 220, even though we know that the real load will be much lower.

I found that panel schedule loads are way high than initial value partially because some non-concurrent loads (heating and cooling) are shown as connected... and excessive receptacle loads, especially for the hospitals. Also, the dedicated appliance and equipment loads are always connected with high load.

I know we can't use the utility numbers but do we really have to design for 25-30W/sf when we know they will only need 8-10W? So if I find the middle ground and design around 13-15W, it won't meet the code then? Should we tell the owner that they have to pay the twice of the service they will never use? (since the utility won't provide that size service/transformer anyway)

 

[jrohe]

I found that panel schedule loads are way high than initial value partially because some non-concurrent loads (heating and cooling) are shown as connected... and excessive receptacle loads, especially for the hospitals. Also, the dedicated appliance and equipment loads are always connected with high load.

I know we can't use the utility numbers but do we really have to design for 25-30W/sf when we know they will only need 8-10W? So if I find the middle ground and design around 13-15W, it won't meet the code then? Should we tell the owner that they have to pay the twice of the service they will never use? (since the utility won't provide that size service/transformer anyway)

The calculations in NEC 220 are so conservative that almost any large building service is oversized by a magnitude of 2, if not even more. I just looked at an industrial facility today that has a 1200 amp service that was designed per NEC Article 220 and their peak demand load over the past 2 years was only 572 amps. Unfortunately, the NEC does not allow engineers to apply that knowledge.

There are really two things at play here. You have to design a service that will work in reality. You also have to design a service that will be compliant with NEC Article 220 load calculations, no matter how obscenely conservative they are. The NEC Article 220 compliance aspect means that you have to design a service much larger than is really necessary.

When we are in SD and DD phases, we take a similar approach as Charlie. We look at each occupancy type and use a specific power density (VA/sf) for that specific occupancy. Those power densities are broken down into lighting, HVAC, receptacles, and miscellaneous. In general, these power densities are likely close to 2 times what you will see in the real world because we know that compliance with NEC Article 220 is going to come into play during the CD phase. These power densities are also on a sliding linear scale so that the power densities decrease the larger the building gets. Once those power densities are finalized, we add in large loads such as elevators, escalators, fire pumps (if they're not being fed directly off a utility transformer), etc. and we use this final value for space allocation for electrical distribution equipment. Using this method, we have been very successful in coming up with accurate estimates even during the SD and DD phases. And if we oversized a bit, the clients never balk at us telling them that we can back the service sizes off and maybe even give them some room back.