Have you looked at placing windows at different locations within the thickness of the wall? One of the Stamford Brook .pdf files suggested that the psi value diminishes the closer the window (or what ever) element is located towards the center of the insulation plane.
How do you get THERM to give you the U-value? Its a while since I used the software and I don't recall the feature. Is it just on a simple pull down menu?
I looked at a couple of different reveal depths but didn't graph the results. I was interested in this because with a deep reveal the thermal bridging results seemed fairly high which appeared at odds with the surface temperature staying above 18.9 degrees all the way to the frame I didn't look into it in detail but guess that this was due to deep reveal adding approx 230mm of additional surface area. If someone else wants to model some of the window details from European inward opening windows, with additional insulation on the outside of the frame, I would be interested in seeing the results.
I have already ordered windows from Thermotech with an integrated fitting flange, which dictates the position of the window in the wall so I focussed on how we could get the details with the lowest bridging results. The airtightness approach my builder uses is to externally wrap the building and then connect onto the door/window frames. The fitting flanges are designed for this purpose.
Therm calculates the U values automatically but you need to create multiple models in order to find the difference caused by introducing the thermal bridge. This is the reason I created baseline models for each junction. To find the U value for a model you need to:
* create the model,
* click on the BC button to assign boundary conditions to the model
* the first time you need to select the exterior boundary, by clicking on the beginning and ending edge, press enter and assign the exterior boundary condition
* do the same for the interior but as well as selecting the boundary condition also set the U factor surface, I typically used frame but I don't believe this is critical
* run the simulation
* click on the show U factors button
* then plug the U values from both models and the junction length into the first page of the spreadsheet which will give the thermal bridge [It took me a while to realise that Therm was working out the energy lost though the model and then scaling this up to a U-value. Ie if the model was only 500mm long it would multiply the results by 2. But as the thermal bridges are quite localised this process gives a higher answer than it should. I confirmed this by extending one of the junction models to 1m long and the U-value was lower even though the junction remained the same. ]
I'm going to try to get the ISO document from the library and see how closely I can follow its approach using Therm.
Chris: I'd skip the calculations and just have a look at the report, there are two pages of text and 8 pages of detail drawings with overlaid isotherms.
http://www.rjaowilliams.plus.com/tmp/thermal-bridge-calcs/thermal%20bridge-report-03-04-07.docRod
I am unable to access the link above, it is asking for a username and password? Any help would be appreciated.
Author