[Greenbuilding] thermal bridging--letting it go

[Corwyn]

On Jul 31, 2006, at 13:20, Alan Abrams wrote:
>
> I understand all about stripes on vinyl siding, and linear patterns of 
> snow
> melt on roofs, and thermal imaging showing hot spots at framing 
> members--but
> where oh where is there some empirical evidence that this effect is not
> caused for the most part by convection at ill fitting fiberglass 
> batts?  As
> Holmes says, "never overlook the obvious."

Good point.  I haven't seen any.

> I also understand how an r value for wood can be substantially less 
> that
> that of insulation.  But to calculate heat loss through a 2x4 or 2x6 
> stud
> *presumes* that heat wants to travel in a straight line, 3 1/2 or 5 1/2
> inches from inside surface of stud to outside surface.

Well heat doesn't _want_ anything.  It just moves. You're right, the 
heat doesn't move in a straight line perpendicular to the wall.  In 
order to really see what it does, you could create a program which 
simulates a wall, where each infinitesimal volume is simulated and the 
total effect is taken by summing over all of them.  This would likely 
show that the thermal bridging effect is smaller than is approximated 
by the simple averaging of R-values.  This, is fact, what test appear 
to show.

> But in the case of tight fitting insulation--in particular, sprayed in 
> place
> foam--why doesn't at least some of the heat want to take a shortcut to 
> the
> left or right, a max distance of 3/4", into insulated void?  To a 
> certain
> extent, the insulation is buffering this tendency, but the effect 
> would be
> progressively less as the heat travels farther and farther away from 
> the
> interior.

It does, and some of the heat in the insulation takes the shortcut 
through the stud.  If they were both the same temperature, these would 
be equal, since the wood is warmer, it takes the shortcut more often, 
reducing the thermal bridge affect.  I am not sure what you mean by the 
insulation buffering this tendency.

Another interesting question is whether isulation's R-value is 
determined by putting it in a wall system and testing heat transfer.  
In which case, Icynene's R-value already includes its better sealing 
aspect.

> The point being, that there must be a certain point where the bridging
> effect is substantially nullified by lateral dissipation of heat.  And 
> I
> suspect that with the use of icynene or similar insulation, it happens 
> a lot
> closer to 5 1/2" away from the interior than the common wisdom has it.

It is still never going to equal a homogeneous wall of icynene.


Thank You Kindly,

Corwyn

-- 
Corwyn
Kermit didn't know the half of it...
http://www.greenfret.com/
corwyn at greenfret.com