[BLDG-SIM] Stat Est response to John P. Karasaki

BKoran at aol.com BKoran at aol.com
Thu Jun 3 16:47:24 PDT 1999

In a message dated 6/3/99 8:40:35 AM Pacific Daylight Time, 
jpkarasa at cbg-engrs.com writes:

> I don't think you can estimate AC equipment demand with the statistics
>  variables you mention.  The number of AC units on at any time is a function
>  of the internal and external building loads, the thermal inertia of the
>  building and contents, and the equipment operating schedule set by the
>  building operator.
And, I believe, statistical variability in things such as the duty cycle of 
the equipment, which is dependent upon equipment variability, T-stat 
deadband, etc.

>  We had a project in which we had the good fortune of over 2 years of 
>  data and coincident weather data.  The goal was to come up with a series (a
>  minimal amount!) of equations by which we could predict a building's 
>  load based on outdoor air parameters.
>  We were not able to do it as there were too many inconsistencies and too
>  many variables.  e.g. so what if it is 100 F out if that occurs on a Sunday
>  when the building isn't occupied; or worse, a Monday that also just happens
>  to be a holiday.

I had a retrofit project in which I also had "2 years of chiller data and 
coincident weather data"  This was a large hospital, and I WAS able to come 
up with a series of equations to model the cooling load based on outdoor air 
parameters.  My spreadsheet model was evaluated by the P.E. for the local 
utility, and we obtained a $180,000 rebate based solely on my model.  

So now we have 2 anecdotes.  Neither of us is right or wrong in the 
reasonabless of simulating a facility based mostly on OSAT.  I happened to 
have a particularly consistent building.   I'm guessing you didn't.  Also, 
the building I was looking was relatively unaffected by day-of-the-week, 
holidays, etc.
>  It seems the more one tries to simplify a complex process, the more the
>  accuracy of the results are compromised.

I couldn't agree more.  But my concern is that our *best* sim programs 
already include big simplifications in this regard.

Thanks for responding to my post...this is most educational, and making 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