Demand

Status
Not open for further replies.
B

brian john

Senior Member
Location
Kilmarnock, Va
Occupation
Retired after 52 years in the trade.
I have a customer that wants KW Demand for a report that was already completed I have voltage Amperage, KW, KVA, PF, KWHR readings based on Min, max and average reading recorded every 5 minutes is there a method to give demand based on 15 minutes period.
 
Your 5 minute periods are more accurate than the 15 minute periods. If the customer wants it in 15 minute intervals to compare to a utility bill, then your average KW data is close enough. From what I have seen, utility 15 minute intervals are at :00, :15, :30, and :45 minutes past the hour. I suppose you could just start at a given time on the hour and take the average KW of three 5 minute average KW periods at a time. Seems like kind of a waste of time. I would just give them the average KW data over 5 minute intervals and be done with it.
 
Aren't you looking for the highest 15 min period? Doesn't the POCO factor your KWH charge for the maximum amount they will likely have to deliver? Why should the neighboring house have to pay for the large lines, transformers, capacitors etc. needed to run your factory?
 
Last edited:
Do you still have the raw 5 minute data?

If I understand '15 minute demand', that is simply the KW or KVA averaged over 15 minute periods, and then looking for the _highest_ 15 minute average.

If this is the case, then you can get a pretty good approximation by simply averaging your 5 minute windows in sets of 3. The averages of 3 consecutive 5 minutes windows, averaged together, is exactly the average of the composite 15 minute window. The only thing that you will miss are events that somehow happen to fall 'out of alignment' with your window periods.

For example, imagine that the customer has a load which doubles their KW, but only for 30 seconds, and irregularly. If one of these events happens, then it will increase the average of the 5 minute window that covers it, and be correctly represented in the 15 minute composite window.

But what happens if you have 2 such events, 13 minutes apart. Depending upon where these two events fall relative to the 5 minute windows, you may never catch that they were actually in the same 15 minute period.

If, as wirenut suggests, the utilities use _fixed_ 15 minute windows rather than a 15 minute rolling average, then the above is not a problem if the 5 minute windows are aligned with the utility windows.

-Jon
 
buckofdurham said:
IMO, Doesn't seem like it would be accurate. Which 15 min. would you use?
Click to expand...

MAX DEMAND for a thirty day period

Doesn't the POCO factor your KWH charge for the maximum amount
Click to expand...
Not on 10 different feeders from one switchboard.

Basically what they are trying to do is comply with 220.87 exception.

the calculated load shall be permitted based upon the maximum demand over a 15 minute period continously recorded over a 30 day period. using a recording amp meter or power meter.

Typically they would supply utility demand readings, I am trying to figure out how the utility meters arrive at this.
 
Last edited:
I am _guessing_ here based upon an understanding of physics and mathematics, rather than knowing the industry standard used, however:

"maximum demand over a 15 minute period continuously recorded over a 30 day period" means the peak value of the '15 minute rolling average' recorded over a 30 day period.

A 'rolling average' is a _continuous_ function; at each instant in time you have the power being used _right now_ (the instantaneous power) and the _average_ power being used over the last 15 minutes.

As a mathematical absolute truth, you _cannot_ generate a continuous rolling average from discrete data; you can only generate a discrete approximation to that rolling average. As I mentioned in my previous post, if you had some poorly timed peak power pulses, the pulses might show up in separate recorded windows, but the 'true' continuous rolling average might have a 15 minute window that contained both pulses.

As a practical matter, I would bet that the approximation to the 15 minute rolling average generated from the 5 minute discrete averages would be pretty good. Basically if you take your 5 minute data, and calculate 15 minute averages, and find the maximum, you know that it will be close to the true value that you want, but possibly low.

More importantly, the maximum of the 15 minute continuous rolling average (the true value that you want) _must_ be less than the maximum 5 minute discrete average. (I have this by inspection, not proof. I am quite certain that it could be mathematically proven if the proof does not already exist.)

Thus the maximum of the 5 minute average power use data that you already have (which is _not_ continuous, because you only record the value every 5 minutes) will approximate the true value that you want, but be slightly _high_.

You now have two numbers, which in any normal situation will be pretty similar, and you know that the _true_ number that you want to calculate is between them.

If these two numbers agree closely, then I would simply take the larger of the two and call it good.

-Jon
 
Last edited:
We used to have a demand meter. The peak needle would be pushed up by the KW usage, which would be reset every 20 min. The peak needle would stay at it's peak, but the KW usage that put it there would have to start all over again every 20 min.

When the guy read the meter he reset the peak needle.
 
Last edited:
The highest 5 minuite demand will always be larger than any 15 min. demand over the same 30 day period.

Given that, I think your 5 minute demands satisfy the code requirement, and I think you should just use your 5 min. data.

Are you saying your 5 min. demand is too high to let them add whatever load they want? I'm not sure i would want to add any more load to anything that has a 5 min. demand that says there isn't capacity for more load.

Steve
 
Status
Not open for further replies.
Top