[BLDG-SIM] Stat Est. response to Larry Degelman

BKoran at aol.com BKoran at aol.com
Thu Jun 3 16:21:38 PDT 1999

In a message dated 6/3/99 12:27:27 AM Pacific Daylight Time, 
Larry at taz.tamu.edu writes:

>  If you look at this purely as a statistics problem, it's interesting to 
> that the larger the number of units operating with independent 
probabilities of 
> being on at any one time, the lower will be the probability of all of them 
> on at any one time.

Yes, that is partly why I brought this issue up.

>  What I'm skeptical of is applying statistics to
>  building cooling loads.   I don't agree that this is a statistics problem. 
> It has a bit of probability in it, but the air conditioner's operation is a 
> reaction to a fairly predictable pattern of thermal loads.

Absolutely, for an individual A/C unit.  But I am concerned with the case for 
which the impact of the load and ambient T on equipment cycling, efficiency, 
and capacity have already been accounted.

> To back this up 
> (admittedly not by a scientific method), just look at the electric codes 
again and you 
> will see that even though diversity is allowed in plug loads, you are 
required to 
> sum 100% of all air conditioner loads when sizing the service panel.  

That makes sense, because there is (nearly) always the possibility that all 
units could be on at once.  (Conversely the probability that the all of the 
units will NEVER be on at the same time is < 1.)

However, for estimating peak demand by month, (on average, over a several 
year period) one might look at, for Pr =0.5, how many units would be on 
simultaneously for at least 15 minutes, or perhaps more units on for less 
than 15 minutes if that results in a greater 15-minute integrated demand.

So, I think I agree with you, except I think it is at least partly a 
statistics problem.  The biggest uncertainty I have is whether zonal load 
diversity has such a big impact on the peak kW (few of the units ever see 
peak demand at the same time, most months) that this type of statistical 
diversity becomes insignificant.  The zone load diversity impact is certainly 
dependent upon lots of factors, but my guess is that it does not always 
dominate the statistical diversity.  For example, many schools have little 
diversity between zones, especially if there are not east- or west-oriented 

Thanks for your response!  This is an interesting subject for me.


You received this e-mail because you are subscribed 
to the BLDG-SIM at GARD.COM mailing list.  To unsubscribe 
from this mailing list send a blank message to 

More information about the Bldg-sim mailing list