[Greenbuilding] thermal bridging--letting it go

[George J. Nesbitt]

Don't let go untill you understand! and there are no stupid questions, 
just answers.

Lets see if I muddy the waters on the subject.

Lets start with some review of the basics.
Heat flows from hot to cold (don't ask why, just except it).
Heat flows through several methods, Conduction, Convection, and Radiant.
Conduction in heat transfer through solids, Convection is by fluids, and 
Radiant is between objects through air.
Different materials conduct heat at different rates. K(conductivity) is 
how much heat flows through a 1" thick 1 sqft slab of material per hour 
at a 1 deg F temperature difference. K=Btu's/hr/sf. Aluminum K-117, 
Steel K=26, Glass K-0.65, Concrete K=0.48, Brick K=0.36, Wood K=0.13, 
Glass Wool K=.0.22, Air K=0.015. As you can see wood is 6 times more 
conductive than glass wool. Steel is 200 times more conductive than wood.

Frame Construction.
Of course most of our walls are built out of a variety of materials and 
layers with different conductivities. Most of the time we use wood, and 
it is continuous from the inside to the outside. Because wood has a high 
(relative to insulation) conductivity we call it a "thermal bridge". How 
much of a thermal bridge will depend on what % of  your assembly is 
wood, 20-25% is common and on multifamily higher. I have herd of as low 
as 6% for a well designed and built "advanced framing" house.
Because of  the thermal bridging the "whole wall R-Value" is something 
less than the R-Value of the insulation in the cavity. It will also 
depend on whether the thermal bridging is wood or steel, and and 
thickness of the wall, and the type of insulation. With a steel framed 
wall (2x4 or 2x6) with spray foam cavities, you have to add R-4,5 
continuous external insulation to get it to work as well as a 2x4 wood 
wall with R-13 cavity insulation, steel sucks! Joe Lstiburek recomends 
not insulating the cavities of steel walls, just the exterior.
We can reduce the thermal bridging by two ways, "advance framing" or by 
using a "thermal break". We can create a thermal break by wrapping the 
assembly with a continuous layer of insulation, or by separating framing 
(double wall or larson truss). We get a lot more benefit by having a 
thermal break than by just making an assembly thicker.

So now lets consider a 2x4 assembly with spray foam insulation, or best 
case framed assembly with no thermal break. If we took a picture with an 
Infrared camera in the winter when it is hotter inside that out, we 
would see the framing because it is warmer due to its higher 
conductivity. Because the framing is warmer some heat would transfer to 
the insulation because it is colder, but it still wants to go out 
because it wants to go from warmer to cooler. Assuming it is perfectly 
airtight we would see no effect from convection or infiltration. If we 
added a continuous layer of insulation on the outside of the assembly we 
would see spots from the nails that attach it too the framing.

Now lets consider a 2x4 assembly with batt or blow/spray in fiberglass, 
cellulose, or cotton insulation. When we look at it with an Infrared 
camera we will probably see a much lager effect from the framing. This 
is due to gaps in the insulation that allow convection to occur, and due 
to infiltration through the assembly. If you use a blower door you can 
see the effect from the infiltration better, which can be quite dramatic.

If we build assembles with a thermal break, or just thicker assemblies 
without a thermal break but don't also address infiltration we will not 
get the benefit we are hoping for.

Solid Walls
Since the materials most solid walls are made of are highly conductive, 
you could say the whole wall is a thermal bridge. Their performance 
would be greatly implroved by adding continouse exterior insulation.

Build Tight, Superinsulate, Ventilate Right




Keith Winston wrote:

>It occurs to me, finally, that we don't have to do the "thought 
>experiment" I mentioned before. Just consider this:
>
>According to California Energy Commission calculations, metal-framed 
>wall assemblies with R-30 insulation have an effective R-value of less 
>than 12.  
>(http://www.homeenergy.org/archive/hem.dis.anl.gov/eehem/94/940707.html)
>
>Now, all that heat loss is occurring through a metal stud, who's 
>thickness in cross-section is perhaps 22 gage (damn thin!), though with 
>very good conduction! We've got all that insulation-edge-effect we're 
>talking about still going on, and yet a huge reduction in effective R-value.
>
>Of course, (sigh) those are probably batts, so there's still probably 
>convection and gaps. If only things could be simple...
>
>Keith
>
>
>Corwyn wrote:
>  
>
>>On Aug 1, 2006, at 00:34, Keith Winston wrote:
>>    
>>
>>>Heat conduction is from hotter to cooler points. You could perhaps more
>>>accurately imagine that heat (like water) is "trying" to go in all
>>>directions, but can only actually succeed going in the direction of
>>>cooler areas. The stud is the coolest point because the heat is
>>>transfered through and out of the stud fastest (into the cold outside
>>>air), compared to the Icynene. Heat will try to transfer through the
>>>Icynene, but due to the high R-value, will not move well through the
>>>foam, and the foam will, in essence, not be cooler.
>>>      
>>>
>>Looking from the outside, the studs will be the warmest part.   Thus, 
>>for any given slice of wall (crossways to the heat flow, parallel with 
>>wall surface) the stud will be the warmest section and heat will move 
>>from it into the cooler insulation.  This is why thermal bridging 
>>isn't as bad a simple computation by R-values would indicate.
>>
>>Thank You Kindly,
>>
>>Corwyn
>>
>>    
>>
>
>  
>