Thermal bridging

[retnuhnibor]

While SAP10.2 allows manufacturer data and fallback to default values that are 'tolerable', SAP10.3 and it's follow on Home Energy Model (HEM) essentially force the calculation of thermal bridge PSI values. Using the defaults may fail your whole property and manufacturer data is no longer allowed. Failing SAP10.3 would mean you do not get you completion certificate and not be able to occupy your house.

 

I have deliberately chosen SAP10.3 as it certifies a higher thermal performance which to me translates to increased value in the house. However, getting the psi-values done has been a cost and professional involvement nightmare with one wanting between £5-8k for 12 junctions.

 

Recently I came across a real alternative that would give you back control. It's easy to use and does not cost a fortune. The site is https://psiclops.io with plenty of verification links, the ability to use DXF drawings as well as layer definitions and it validates against BR497 and gives you a full report.

Edited 5 hours ago by retnuhnibor

[-rick-]

That looks cool. Just need to find a suitable competent person willing to use it and sign which might be a problem. Having said that, in time this looks like a solution that will work once the competent people get familiar with it.

[ADLIan]

Do you already have Building Reg approval for this house? SAP 10.3 is in support of the next Appr Doc L1 (2026) effective from March next year. Using both SAP 10.2 and SAP 10.3 is just extra work for no real benefit. As I understand it both versions deal with linear thermal bridges in the same manner and manufacturers of insulation/blocks/TF etc (as competent persons) the psi-values are acceptable. Not much background info on who or what is behind psiclops - tread carefully!

[retnuhnibor]

SAP10.3 is very different to SAP10.2 with regard to thermal bridges and is a stepping stone to HEM. You cannot use manufacturer tables or values. With regard to the competent person requirement; there is no register for such people and the main concern that the SAP assessor will have is that the calculations have been done to ISO standards and BR497. BR497 requires certain boundary conditions for the report and getting these right can be horrible. I tried using THERM (which is free) but came badly unstuck with respect to BR497. Also psiclops is fully ISO verified and BR497 compliant, which is what is required for SAP and BR. It's low cost so if it's not producing acceptable output then I will not have lost much compared to going to a thermal engineer. Also, the SAP assessor has to tell me why the output is not acceptable.

 

I do not yet have building control. Getting all of the paperwork together that I need has been a big task and I decided to wait until February to cut ground. While I could have done my work under SAP10.2 the design of the house is such that my total heat loss is circa 2.5kW and SAP10.2 doesn't do this justice and I have to use defaults or dig through manufacturer data to get a psi value. Using software like http://psiclops.io just made this all easy. It also has an architect verification piece if you think you need that reassurance. 

 

Make no mistake - SAP10.3 and HEM are really forcing calculations for pretty much all of your junctions. The defaults will potentially fail your building as you can see from table K1 in the SAP10.3 documentation.

Edited 4 hours ago by retnuhnibor

[retnuhnibor]

Paddocks532 story is hard to understand. The 'As Designed' SAP assessment should have picked this up way before you cut ground. For sure the assessor should be able to say what failed not just 'it failed' and if they can't give you more then it may be worth talking to Stroma or whoever that are 'accredited' by.

 

One thought I did have is that timber frames under BR497 have to take into account the thermal path though the insulation as well as the thermal path through the timber frame. You must use the combined method or a finite element model (FEM).  Both of these will more accurately calculate the wall true U-value, with FEM giving the true value. The defaults SAP uses for combined method are not very helpful to getting the U-value lowered.

 

I am using I-Joists in my wall. I used a FEM to calculate the U-value and it was 97% better than the combined method as I-Joists are hard to model without FEM. 

[ADLIan]

Agree with comments re the OP.

 

If the FEM Uvalue of an I-joist wall is so much better than that from the BS EN 6946 method I would suggest one result is incorrect. The combined method gives accurate results provided the I joists are modelled correctly. Therm output is not very intuitive!! I’ve used both methods on a wide range of wall details and they tend to agree.

 

A quick read of SAP 10.2 and 10.3 appears to have identical guidance on dealing with linear thermal bridges. Where do you see any differences?

[retnuhnibor]

I'm sorry if I misread your comment where you say 'manufacturers of insulation/blocks/TF etc (as competent persons) the psi-values are acceptable'.  I read this as comparing SAP10.2 and 10.3 as following the same path with respect to what source of values are acceptable. For sure the ethos is the same as building physics is building physics. However, SAP10.3 does not allow the use of manufacturer details. This was a specific change as is was found the manufacturers were not necessarily giving accurate psi-values  as it could give a competitive disadvantage if the real values are poor but the stated values are not. 

 

With respect to I-Joists BR443 recognises that the modelling of these is poor which is why  Section 4.4.2 three layer method exists for I-beams (I-Joists). Even using this method the result is still higher than can be achieved with a FEM, although close enough for most. However not applying section 4.4.2 at all really does negatively impact the U-value.

 

For the property that started this thread I just wondered if the timber bridging had not been applied originally but had then been applied by the SAP assessor . If so it could also use the wrong ratio (I think 11% or 15% is used as a default) while a typical 600mm insulated bay using 45mm timber would actually be about 7.5%.