220.87

Status
Not open for further replies.
ron said:
I think the language in 220.87 is chosen (and possibly poorly) to allow use of the 15 minute measurement made by the utility
Click to expand...
I think that is the intent of 220.87(1) with the exception permitting field installed measuring devices for 30 days.
 
It seems in the two examples that have been worked in this thread, the 15 minute demand number comes out to be equal to the 15 minute usage number multiplied by 4.
 
don_resqcapt19 said:
It seems in the two examples that have been worked in this thread, the 15 minute demand number comes out to be equal to the 15 minute usage number multiplied by 4.
Click to expand...
Yes, simply because one hour divided by 4 is 15 minutes, and the kWh unit is based on a hour. (If the unit were kW15mins then you wouldn't need to multiply by 4.)
 
I don't know what 'and maintained' is supposed to mean in the exception language. I can't figure out how it makes any difference. Average kilowatts over 15mins is exactly that, and once it's 'reached' it can't be 'maintained' beyond the 15 minutes inquestion.

That said, the exception doesn’t state we're allowed to use an interval that begins exactly on the hour, 15 mins after the hour, etc. What if the highest average kW over a 15min period began 4 mins after the hour and ended 19mins after the hour.

A conservative approach is to take the highest kWh sum of two adjacent intervals from the on-the-hour based data, assume all the consumption occured in a 15min period that overlapped the boundary between those intervals, and then multiply the kWh sum by four (1hr/15min) to get an upper possible bound on kW under the broadest interpretation of the exception.

This is overly conservative, i.e. likely massively overestimates the actual peak demand, by as much as twice. But it unassailably meets the language of the exception. For commercial, the overestimate may be less likely to give you the answer you want but for residential it will often confirm the service is much bigger than it needs to be.

For me this is not theory, I use kWh data to satisfying myself on 220.87 frequently. I also take hourly data and multiply by four (although doing the sum exercise above seems excessive). I've not had to justify these methods to an AHJ yet, but I consider them completely justified. All the more so if utility provided kW data isn't doing anything more precise.
 
As far as I know, demand is just a 15 minute average, and the highest demand is the highest one of those 15 minute averages. I don't know if it's based on set 15 minute intervals, or the highest of a "sliding window".

I remember having some discussion in the past, in a somewhat recent thread, where I was complaining about that word "maintained" in the definition of demand, and it didn't really make sense. Maybe there is more to the calculation of a demand figure than I am aware of? Like I said, I thought it was just a simple average.
 
electrofelon said:
As far as I know, demand is just a 15 minute average, and the highest demand is the highest one of those 15 minute averages. I don't know if it's based on set 15 minute intervals, or the highest of a "sliding window".

I remember having some discussion in the past, in a somewhat recent thread, where I was complaining about that word "maintained" in the definition of demand, and it didn't really make sense. Maybe there is more to the calculation of a demand figure than I am aware of? Like I said, I thought it was just a simple average.
Click to expand...

Back in the 1980s (last time I dealt with this) our utility used 5 minute sliding windows. The demand charge was based on the the 3 highest consecutive 5 minute windows.
 
retirede said:
Back in the 1980s (last time I dealt with this) our utility used 5 minute sliding windows. The demand charge was based on the the 3 highest consecutive 5 minute windows.
Click to expand...
I would call that a 15 minute sliding window. 5 minutes would be the granularity or sampling period or something.

jaggedben said:
For what's its worth this utility's page shows the same thing.
Click to expand...
That doesn't have a sliding window. In the example it happens to break up each 15 minute interval into 5 minute intervals, but it only looks at consecutive 15 minute intervals, there's no sliding. And the description refers to "over 2,800" data points per month, which is just 4 per hour, not the 12 per hour you'd need to implement a 15 minute sliding window with 5 minute sampling period.

Cheers, Wayne
 
wwhitney said:
I think you are misinterpreting the word "highest". It is the same as the word "maximum" in "maximum demand". I.e. the sentence tells you to take create one average number for each 15 minute interval, for at least 30 days (so at least 4 * 24* 30 numbers), and take the highest of those numbers.

Cheers, Wayne
Click to expand...
Perhaps you could give us your interpretation of the phrase,". . . and maintained for a 15-minute interval.”
 
jaggedben said:
I don't know what 'and maintained' is supposed to mean in the exception language. I can't figure out how it makes any difference. Average kilowatts over 15mins is exactly that, and once it's 'reached' it can't be 'maintained' beyond the 15 minutes inquestion.

That said, the exception doesn’t state we're allowed to use an interval that begins exactly on the hour, 15 mins after the hour, etc. What if the highest average kW over a 15min period began 4 mins after the hour and ended 19mins after the hour.

A conservative approach is to take the highest kWh sum of two adjacent intervals from the on-the-hour based data, assume all the consumption occured in a 15min period that overlapped the boundary between those intervals, and then multiply the kWh sum by four (1hr/15min) to get an upper possible bound on kW under the broadest interpretation of the exception.

This is overly conservative, i.e. likely massively overestimates the actual peak demand, by as much as twice. But it unassailably meets the language of the exception. For commercial, the overestimate may be less likely to give you the answer you want but for residential it will often confirm the service is much bigger than it needs to be.

For me this is not theory, I use kWh data to satisfying myself on 220.87 frequently. I also take hourly data and multiply by four (although doing the sum exercise above seems excessive). I've not had to justify these methods to an AHJ yet, but I consider them completely justified. All the more so if utility provided kW data isn't doing anything more precise.
Click to expand...


Theory as refers to projections and future incidence (15- minute sampling) and final result is still a theory.

It is speculation, presumption , a guess.

This follows the Mathematical concept of Probability Theory , thus--clearly a theory
 
charlie b said:
Perhaps you could give us your interpretation of the phrase,". . . and maintained for a 15-minute interval.”
Click to expand...
Sure. In the 2014 NEC, the language is "maximum demand (measure of average power demand over a 15 minute period)." Here the parenthetical comment appears to be a definition of demand for this section. And the phrase pretty clearly just means take a 15 minute period and compute the average power over that period. There's no discussion of whether we should use a sliding window, or just divide a day up into 96 intervals of length 15 minutes.

In the 2017 NEC, the phrase changed to "maximum demand (the highest average kilowatts reached and maintained for a 15-minute interval)" That change was introduced by CMP 2 under FR 344-NFPA 70-2015 with the only statement "The added text will improve clarity and understanding of the term “maximum demand”." Which, by the way, indicates that this definition in the exception is intended as a general clarification of "maximum demand" and so would apply to 220.87(1) as well.

That statement certainly doesn't sound like CMP 2 intended a change in meaning. So I believe the phrase "reached and maintained" is just some redundant language emphasizing that you take the average over the whole 15 minute period, not over some shorter period.

Cheers, Wayne
 
Status
Not open for further replies.
Top