Domestic and non-domestic energy performance certificates review: supplementary notes

2. RdSAP and U-values

The default U-values defined with RdSAP have changed since RdSAP's introduction in Scotland in 2009. Up to October 2012, RdSAP in Scotland effectively had single, age band determined-default U-values for each of the 'as built' wall constructions (see Table UV1).

Table UV1 – Default U-Values for Scottish 'as built' wall constructions – RdSAP v9.81,v9.83 and v9.90

Wall Type	RdSAP Age Band
	A	B	C	D	E	F	G	H	I	J	K
Stone: granite or whinstone (as built)	1.7	1.7	1.7	1.7	1.7	1.0	0.6	0.45	0.45	0.3	0.25
Stone: sandstone (as built)	1.5	1.5	1.5	1.5	1.5	1.0	0.6	0.45	0.45	0.3	0.25
Solid brick (as built)	2.1	2.1	2.1	2.1	1.7	1.0	0.6	0.45	0.45	0.3	0.25
Cob (as built)	0.8	0.8	0.8	0.8	0.8	0.8	0.6	0.45	0.45	0.3	0.25
Cavity (as built)	2.1	1.6	1.6	1.6	1.6	1.0	0.6	0.45	0.45	0.3	0.25
Timber frame (as built)	2.5	1.9	1.9	1.0	0.8	0.5	0.4	0.4	0.4	0.3	0.25
System build (as built)	2.0	2.0	2.0	2.0	1.7	1.0	0.6	0.45	0.45	0.3	0.25

With the changeover to RdSAP v9.91 in October 2012, while many of the 'as built' age-related U-values remained unchanged, the U-values for sandstone and granite or whinstone wall types were replaced by a footnote which referred the reader to equations in S5.1.1 of the SAP manual (see Table UV2 below). 'Limestone' was added to the sandstone wall type category with RdSAP v9.92 in December 2014, but there were no other changes to any of the 'as built' U-values.

Table UV2 – Default U-Values for Scottish 'as built' wall constructions – RdSAP V9.91 and v9.92

Wall Type	RdSAP Age Band
	A	B	C	D	E	F	G	H	I	J	K
Stone: granite or whinstone (as built)	a	a	a	a	1.7	1.0	0.6	0.45	0.45	0.3	0.25
Stone: sandstone or limestone (as built)	a	a	a	a	1.5	1.0	0.6	0.45	0.45	0.3	0.25
Solid brick (as built)	2.1	2.1	2.1	2.1	1.7	1.0	0.6	0.45	0.45	0.3	0.25
Cob (as built)	0.8	0.8	0.8	0.8	0.8	0.8	0.6	0.45	0.45	0.3	0.25
Cavity (as built)	2.1	1.6	1.6	1.6	1.6	1.0	0.6	0.45	0.45	0.3	0.25
Timber frame (as built)	2.5	1.9	1.9	1.0	0.8	0.5	0.4	0.4	0.4	0.3	0.25
System build (as built)	2.0	2.0	2.0	2.0	1.7	1.0	0.6	0.45	0.45	0.3	0.25

The most recent manifestation of the default RdSAP U-values table for 'as built' has made changes to the U-values of both solid brick and cavity brick walls. These changes were introduced after on-site in-situ testing of walls found that solid brick walls and cavity walls performed better than the theoretical defaults^[1], and as part of the UK Government's changes to SAP in the wake of the SAP 2016 consultation^[2]. These changes to the default U-values were incorporated into RdSAP with v9.93 in November 2017 (see Table UV3).

Table UV3 – Default U-Values for Scottish 'as built' wall constructions – RdSAP v9.93

Wall Type	RdSAP Age Band
	A	B	C	D	E	F	G	H	I	J	K
Stone: granite or whinstone (as built)	a	a	a	a	1.7	1.0	0.6	0.45	0.45	0.3	0.25
Stone: sandstone (as built)	a	a	a	a	1.5	1.0	0.6	0.45	0.45	0.3	0.25
Solid brick (as built)	1.7	1.7	1.7	1.7	1.7	1.0	0.6	0.45	0.45	0.3	0.25
Cob (as built)	0.8	0.8	0.8	0.8	0.8	0.8	0.6	0.45	0.45	0.3	0.25
Cavity (as built)	1.5	1.5	1.5	1.5	1.5	1.0	0.6	0.45	0.45	0.3	0.25
Timber frame (as built)	2.5	1.9	1.9	1.0	0.8	0.5	0.4	0.4	0.4	0.3	0.25
System build (as built)	2.0	2.0	2.0	2.0	1.7	1.0	0.6	0.45	0.45	0.3	0.25

2.1 Calculating Stone Wall U-values

With the changes to the default U-values introduced in RdSAP in October 2012, the respective single as-built U-values for the sandstone / limestone and the granite / whinstone wall types were replaced with equations. For these two stone wall type categories, these changes created infinitely variable U-values, as the thickness of the stone wall became a major determinant of the calculated U-value. The two equations were, respectively:

• for granite or whinstone, the U-value = 3.3 - 0.002 * thickness of wall in mm
• for sandstone or limestone, the U-value = 3.0 - 0.002 * thickness of wall in mm

Another variable was also introduced in the process: an additional shelter factor was applied to the calculated U-value if the stone walls in age bands A to E^[3] had internal dry-lining or lath and plaster finish (that is, an airspace):

where R_dl is the additional thermal resistance introduced by the internal finish.

Use R_dl = 0.17 m²K/W.

So three factors now determine the stone wall U-value within RdSAP: the type of stone, the thickness of the stone wall, and the presence (or not) of an air space. A sample of the resultant U-values for stone walls is sent out in Table UV4.

Table UV4: Sample RdSAP U-values for sandstone or limestone and granite or whinstone walls

	sandstone or limestone		granite or whinstone
wall thickness in mm	plastered on the hard	with drylining or lath and plaster	plastered on the hard	with drylining or lath and plaster
200	2.60	1.80	2.90	1.94
300	2.40	1.70	2.70	1.85
400	2.20	1.60	2.50	1.75
500	2.00	1.49	2.30	1.65
600	1.80	1.38	2.10	1.55
700	1.60	1.26	1.90	1.44
900	1.20	1.00	1.50	1.20

From Table UV4, the presence of the drylining or lath and plaster layer on the stone wall has a marked impact on the resultant U-value. In effect, if the stone wall was thick enough it would be give a U-value of 0.00 in RdSAP. For a drylined sandstone or limestone wall it would need to be 1498mm thick, and for a granite or whinstone wall it would have to be 1648mm thick.

In its current formulation, RdSAP does take into account the varying thickness of stone walls, and walls of different stone constructions, but does it go far enough?

Dr. Moses Jenkins of Historic Environment Scotland suggested in a paper presented at a recent seminar^[4] that a minor adjustment to the RdSAP stone wall equations might produce a better alignment between the RdSAP U-values and the evidence collected through various in-situ testing projects. He suggested that rather than use a value of 0.002 in the equations set out above, that a value 0.0025 be used. The U-values in Table UV4 were recalculated using this modification and the results are set out in Table UV5 below.

Table UV5: Sample RdSAP U-values for sandstone or limestone and granite or whinstone walls with adjusted to RdSAP calculation equation (0.0025 rather than 0.002)

	sandstone or limestone		granite or whinstone
wall thickness in mm	plastered on the hard	with drylining or lath and plaster	plastered on the hard	with drylining or lath and plaster
200	2.500	1.754	2.800	1.897
300	2.250	1.627	2.550	1.779
400	2.000	1.493	2.300	1.653
500	1.750	1.349	2.050	1.520
600	1.500	1.195	1.800	1.378
700	1.250	1.031	1.550	1.227
900	0.750	0.665	1.050	0.891

This suggested adjustment does seem to align the 600mm and 700mm thick sandstone walls with his suggested actual performance of sandstone wall U-values of between 1.0 and 1.2 W/m²K, but only when internal drylining or lath and plaster is included. Dr Jenkins's initial calculations presented at the seminar did not include for an internal lining. The resultant U-values for 900mm thick stone walls also appear to be 'too good'. Certainly, further empirical research work is required here to confirm and validate these possible changes to the RdSAP stone wall U-value equation.

2.2 Other Types of Stone Walls in Scotland

RdSAP is currently set up for four different stone wall types, grouped in two pairs: 'sandstone or limestone' and 'granite or whinstone'. Scotland has a rich diversity of stone used in the construction of its dwelling stock, though sandstone and granite are likely to be the most common. The RdSAP methodology advises assessors that if they come across other types of stone to categorise them as 'sandstone or limestone' if they are a sedimentary type of stone, or as 'granite or whinstone' if they are an igneous or metamorphic type of stone. The assessor is then to select Addenda note 1 from the software which states that the assessor has chosen the closest match to the actual wall construction. Thus, mudstone from the Orkneys would be recorded as 'sandstone or limestone', while basalt around Bathgate would go through as 'granite or whinstone'. Some work could go into collating more information on the energy performance of other types of Scottish stone, or undertaking in-situ testing, to broaden the range of stone wall U-values that are available in RdSAP^[5].

2.3 Solid Brick Walls

It was noted above that with the introduction of RdSAP v9.93 in November 2017, that the RdSAP default U-values for solid brick walls up to Age Band E were revised downwards (that is, effectively improved). The basis of this change was an empirical data exercise whereby in-situ monitoring of wall U-values was completed on a significant number of dwellings (i.e. 300 dwellings)^[6]. The results found that the actual U-values measured on site were lower (that is, losing heat slower) than calculated theoretically, or assumed by the RdSAP default U-value of 2.1 W/m²K for solid brick walls. The range of measured results is shown in the graph in Figure UV1. The RdSAP default of 2.1 W/m²K is shown at edge of high side of the distribution curve.

Figure UV1 – Distribution of measure U-values in standard (i.e. <330mm) solid brick walls (source: BRE (2014) In-situ measurements of wall U-values in English housing, BRE, Garston)

While the empirical evidence presented certainly supports such a change, the technical background paper to the SAP 2016 consultation on this topic noted "There are also likely to be other factors, such as air voids in nominally solid walls leading to (fortuitously) better U-values."^[7] What the supporting research did not identify was the impact of such factors, aggregating all standard solid walls together and noting that there could be other factors in play. This research would have been more robust if it had identified these various factors, and displayed the distribution of the measured U-values for different factors rather than amalgamating them all into one graph.

The result of this research was to reduce the RdSAP default U-value for solid brick walls from 2.1 to1.7 W/m²K. However, in RdSAP, one of the wall construction items to be identified as part of the assessment survey is whether there is an internal drylining or lath and plaster internal finish because of the airspace (a fortuitous air void?). If present, indicating this in the previous version of the RdSAP software changed the default U-value from 2.1 to 1.55, that is, better than the revised 1.7 W/m²K value.

The same original research^[8] also identified that 'non-standard solid walls' which were defined as solid walls greater than 330mm in thickness, were also performing significantly better than predicted or accounted for by the RdSAP default (see Table UV6 below). For the non-standard solid walls, the mean and median values were 1.28 W/m²K compared to the RdSAP default of 2.1. Yet, there was no proposal in the SAP 2016 changes to revise RdSAP to include dimensional components into the solid wall U-values.

Table UV6: Summary of Results: Classification of wall and calculation of U-values^[9]

Table UV4: Sample RdSAP U-values for sandstone or limestone and granite or whinstone walls

Wall type	Number of Cases	Measured U-values: mean (standard deviation) W/m2K*	Measured U-values: median W/m2K	Calculated U-values: mean (standard deviation) W/m2K	Calculated U-values: median W/m2K	Typical RdSAP U-values W/m2K
Solid wall, standard b)	85	1.57	1.59	1.9	1.92	2.1
		-0.32		-0.2
Solid wall, non-standard b)	33	1.28	1.28	1.91	1.68	2.1
		-0.42		-0.49

While a brick-thick solid wall is the norm for many houses, in the pre-1919 tenements in Scotland, the close walls are often only a half-brick thick (i.e. 150-180mm thick), while the external walls of the lower levels of the tenement can be 1.5 or 2 bricks thick to support the weight of the upper levels of the tenement. The result is that the close walls are given too much benefit by the RdSAP defaults; the external walls not given enough.

Introducing a dimensional component for solid brick walls would differentiate solid brick walls by the wall thickness and assign default U-values: for example:

• a half-brick thick (approximately 150-170mm) (so anything less than 200mm)
• a brick thick wall (usually between 220-230mm) (so anything more than 200mm up to 300mm)
• a 1.5-brick thick wall (usually between 330-350mm (so anything more than 300mm up to 400mm)
• a two-brick thick wall (usually between 450-500mm) (so anything more than 400mm)

This task would not require any more effort from the assessor: RdSAP assessors already measure the thickness of the wall as part of the survey. Identifying solid brick walls is also part of the survey. All of this change could be managed as a software issue.

2.4 Non-traditional and System-Built dwellings

Scotland has a large legacy on non-traditional and system-built dwellings. Much research has been completed on these dwellings to catalogue and describe these constructions in great detail. The Guide to Non-traditional housing in Scotland^[10] gives descriptions of the constructions, the numbers of dwelling units built, the locations, as well as floor plans and section drawings. Behind this summary book were a series of detailed BRE reports on many of the individual non-traditional dwelling types. These dwellings encompass a wide diversity of metal, concrete and timber constructions. Yet, despite the availability of all this information, RdSAP aggregates them under a single, age-related, default U-value. Then, because RdSAP treats "system-built" as a catch-all categorisation, it does not recommend improvements to such wall types under Appendix T. They are not included in the type of walls that can be insulated under RdSAP conventions.

A major effort should be carried out to produce different defaults for known 'system built' wall constructions, and Appendix T be opened-up to recommend external or internal wall insulation to defined non-traditionally constructed walls.

There is much published research on different types of 'non-traditional' and 'system built' walls so that default U-values could be calculated and built into a database within RdSAP so that they could be selected where the 'system built' type can be identified. Databases and records still exist indicating where many of these dwellings were built in Scotland. This information could be utilised to a produce a guide for RdSAP assessors. If the 'non-traditional' and 'system built' wall type could not be identified then the catch-all default would still prevail.

Where RdSAP assessors / DEA's would be required to positively identify stone wall types or 'non-traditional' building elements, this change may require additional certification, and training provided by professional bodies e.g. HES or CIBSE.

2.5 Accounting for Insulation that you can measure

The SAP 2016 consultation document noted that "the energy consumption of a dwelling predicted using SAP is very sensitive to a wall U-value". With cavity wall insulation, the current assumption is that the cavity walls are being insulated with blown fibre or bead with a thermal conductivity of 0.04 W/mK. A number of companies however are using types of bead insulation with a thermal conductivity of 0.032 (i.e. about 25% better than blown fibre or traditional bead insulation) but get no added benefit to carbon savings or improved SAP scores, even if this can be identified or certified.

Similarly, this approach should be extended to internal and external wall insulation so that better modelling of the impact of the actual thickness of the insulation and the material being used is include within the RdSAP assessment. Currently RdSAP offers default insulation thicknesses of 50, 100, 150, and 200mm thicknesses of insulation (and again, this insulation is assumed to be the equivalent of Expanded Polystyrene (EPS)) with a thermal conductivity of 0.04 W/mK. However, many properties were retro-insulated in the past with 25 to 40mm of EPS or similar insulation materials – currently they would get no benefit for this insulation as the actual thickness is less than the minimum 50mm thickness, and assessors are advised by their membership schemes to ignore it^[11]. Currently, many properties are receiving 70 – 90 mm of external or internal wall insulation (which under RdSAP convention should be rounded down to 50mm).

If the concern is getting to the right U-value, then we need to move RdSAP away from rigid standard thicknesses of insulation to allowing actual materials and thicknesses to be specified where these can be identified.