Thermal bridging in the new TGD L(2010)


The Ψ or Psi-value (W/mK) of a thermal bridge describes its extra over heat loss not accounted by the U-values of the adjoining walls, roofs or floors. It depends on several variables such as the U-value of the flanking elements, the geometry of the junction and the continuity of the insulation around.

The Y-factor (W/m²K) is the input for thermal bridging in DEAP. It’s a way of representing the total heat loss impact of all thermal bridges in a dwelling, expressed by square metre of envelope area. It has the same units as the U-value and therefore they may be directly compared: When the average U-value for a dwelling is ascertained the Y-factor can be directly added to get a sense of the full extent of fabric heat loss.

·         The Y-factor can be calculated from Ψ-values obtained by numerical modelling or Tables on Appendix D of TGD L

·         Y = 0.08 W/m²K can be applied for new dwellings that follow ACDs

·         Y = 0.15 W/m²K is the default value for all other cases

New build

The existing TGD L(2007) provides Table D1, apparently drawn up with an understanding that U-values don’t affect Ψ-values. The new TGD L(2010) proposes the more complete Tables D1-D5, based on the realisation that U-values do impact upon Ψ-values.

We have calculated the proposed Y-factors for the Example House on the Acceptable Construction Details following both 2007 and 2010 Tables. An extra column shows the results for a version with adjusted lengths, because we think the junction lengths used on the ACDs for windows are somewhat unrealistic.

Obviously different geometries & styles (e.g. modernist slit windows compared to historic tall, narrow windows) will result in different Y-factors but the variance shown here is already considerable. If someone were to use the ACD junction details on the basis that they could then simply input a Y-factor of 0.08 W/m²K in DEAP (without going through the process of calculating lengths as we have) they won’t actually meet that value in some cases and will be far exceeding it in other cases. We don’t think this is good guidance, and is unnecessary due to the proposed Tables D1-D5.

It’s interesting to note that a key element of the specification of the nine example dwellings used by the Building Regulations team in DoEHLG to show compliance is possible was  a Y factor of 0.05 W/m²K. A Y-factor of 0.08 is already not good enough.



A Y-factor of 0.15 W/m²K is arbitrarily used to describe the level of thermal bridging for all the following buildings:

·         A new-build meeting the Regs but not the ACDs

·         An old building, be it an oil crisis house, a 1950s mid-terrace, a Georgian house, or detached rubble cottage

·         A recent energy-focused retrofit where huge attention has been paid to U-values and no attention paid to thermal bridging


As recognised by the new Tables of TGD L(2010), plane element U-values directly impact upon the Y-factor. Perversely insulating the plane elements more without dealing with thermal bridging can lead to a significant increase of the Y-factor – the ‘Achilles heal’ of so many energy-focused retrofits.

We have calculated that the these could have a Y-factor more than 0.30 W/m²K (in one case we measured 0.58 W/m²K)! This is many times worse than the default value and leads to a serious disconnect between the DEAP estimate of heating load and the actual heating load, as well as risk of unforeseen surface condensation. In case anyone thinks this is unlikely, the 0.30 W/m²K Y-factor above is based on using the exact details advised by NSAI in one of their Agrément certificates (see ‘Breaking The Mould III‘).

Our proposal


We believe designers should calculate the Y-factor every time to know the impact of what they’re doing and input that exact figure in DEAP. As we now have the useful Tables D1-D5 in TGD L(2010) Appendix D listing the Psi-values for all ACD details, the generalized Y-factor of 0.08 W/m²K for ACDs is outdated and should no longer be permitted. In doing this we would be following England and Wales in a transition from SAP 2005 & Part L(2006) to SAP 2009 & Part L(2010). We should be changing with them, not 5 years behind them.

There is still a value in having a default Y-factor for other cases. The current value of 0.15 W/m²K is fine for existing uninsulated houses or those built under various Building Regulations with non-ACD details. However, based on our research we suggest that the default Y-factor for retrofits be raised to 0.25 W/m²K. Any improvement on this figure should be backed up by evidence or calculation. Some insulation manufacturers, such as CosiHomes, already have calculated Psi-values and Y-factors for their retrofit details.

Furthermore, as apparently Y-factors below 0.05 W/m²K might be necessary to comply with 2010 targets, an extended range of improved ACD junction details showing Psi-values would be of valuable help.



Friday, October 15, 2010