Posted By rvalue on 11 Feb 2010 08:15 AM
All;
I have heard that insulations perform differently at greater Delta T's. I have heard that fiberglass "loses" r value at lower temps. Is there a study of this? Are there any published specifications I can reference? I don't want to go off hearsay.
Thanks;
Jake Vierzen
R-Value Concrete Structures
This issue has been studied for 25 years or more, and the mechanics are now well understood. The Oak Ridge National Labs climate simulator studies in the early '90s was sort of a mind-blower for some, but the worst-case scenario was apen blown low density fiberglass in a cold-side up configuration, where warm air convected easily up through several inches of fiberglass from the warmer ceiling into the colder attic in volume that increased with delat-T. An outline of the study is online here.
In walls the problem isn't as pronounced, since the convective forces are vertical, and the delta-T is across a horizontal axis. In warm-side up configurations (under floors) the problem is all but non-existent, since the convective forces stratify the air within the fiber in the same orientation of the delta-T (warmer air wants to rise, and the warm side is up.)
But not all fibers are created equal: Finer spun blowing wools and higher density batts ("cathedral ceiling" variants) allow less convection than the standard low-density stuff that was around 20 years ago. But cellulose at 2lb density inhibits convection to only ~10% of what standard density fiberglass allows. Reference
Fiberglass also has interesting non-linearites under high radiant flux that other fibers don't have. This was studied extensively at Texas A & M in the 1980, and at the Florida Solar Energy Center in the '80s & '90s in conjunction with radiant barrier. Fiberglass is somewhat translucent to infra-red, and heating of the fiber within the top few inches of the material due to the radiation causes a small amount of convection similar to the cold-side up attic in winter issue (the insulation was hotter than the attic air 1.5-2" below the surface!) The effect is to pretty much neutralize the insulating value of the top inches of the material when the roof deck is the hottest. This nonlinearity is absent in denser and more IR opaque fiber. Reference.
You can dig a lot more detail than you'll ever have time to read out of the ORNL and Texas A & M websites in an hour of search engine use if you like, but the principle is clear- convective forces will drive air movement in all fiber insulations with increasing rates at higher delta-T. Radiant heat penetrates fiberglass more than some other fiber insulations (but is only a problem at high delta-Ts between the radiating element and the air on the hot side of the structure.) All fiber insulations benefit from air-barriers on all sides (unfaced batts in an unfinished room need an air barrier over it to perform.) Foams don't have convective losses with delta-T. (They have dimensional change issues creating a less pronounced shift in R value with delta-T however). Cellulose has much lower convective losses than fiberglass even at 2lb density, and performs similar to foam if "dense packed" to 3lb+ density.
But nothing is perfect- they all have their issues. Use them all with some forethought to their weight, hygric capacity, thermal mass, vapor retardency & other properties in mind. Applied appropriately they're ALL good.