Hi,

I'm working with a model of a two-story residential building with an attic
being modeled as an unconditioned space in three climates: Phoenix, San
Francisco, and Chicago. I've set up parametric runs varying the R-values of
the exterior walls and the ceiling (attic floor) in order to determine the
optimal values for efficiency's sake, but the results of the parametric runs
don't make much sense. Specifically, while varying the wall insulation has a
notable impact on the heating and cooling loads, there is only slight
variation when the ceiling insulation is varied (between R20 and R60--would
expect a fairly significant difference). The large savings that one would
expect from such a significant increase in ceiling insulation are not
apparent here in any of the climates. The roof itself is not well insulated
(R2) but is highly reflective.

The problem seems to be rooted in some issue modeling temperature in the
unconditioned zone. Hourly reports show a much narrower range of
temperatures annually in the unconditioned zone (about 20F) than outdoors
(100F)--also somewhat counterintuitive. Somehow, I'd like to find a way to
simulate the larger expected temperature swings in an attic in order to see
the full benefits of increasing ceiling insulation.

I've made a few attempts to resolve the issue. As per an old post from the
BLDG-SIM list
(http://lists.onebuilding.org/htdig.cgi/bldg-sim-onebuilding.org/2007-July/0
05513.html), I tried adjusting the heating and cooling design temperatures
to better reflect the range of outdoor temperatures, but this had no effect
upon household energy consumption.

For curiosity's sake, I also switched the attic to a plenum to see how this
might affect the calculation of loads but got some pretty nonsensical
results in that case as well.

I did also come across another posting
(http://lists.onebuilding.org/htdig.cgi/bldg-sim-onebuilding.org/2005-Septem
ber/002502.html) that suggested integrating glass into the roof deck in
order to allow for the simulation of radiative heat transfer, though this
was in the case of a commercial building. I'm not sure if this is the
solution I'm looking for, but any insight would be appreciated.

Thoughts? Thanks for your help!

Nick Schlag