Developing regulation of energy efficiency of private sector housing (REEPS): modelling improvements to the target stock - Main Research Report

6 Summary Of Technical Issues and Assumptions

6.1 This research project has attempted to be completely transparent in the use of the SHCS dataset in terms of its approach to developing the initial dwelling typology and the individual archetypes, its modelling of the base case archetypes, and the assumptions made in the modelling of measures to improve the least efficient dwellings in the private sector stock.

6.2 This section sets out the background to the SAP and RdSAP methodologies, their relationship with each other and with the SHCS data set that was used as the basis of the 355 archetypes, and how this affects the modelling of the archetypes. During this project there have been regular meetings with the REEPS RAG, and feedback from the REEPS Working Groups and subgroups that have raised various technical issues and concerns about aspects of SAP, RdSAP, the modelling assumptions and the improvement measures modelled that have informed the approach. A log of all of these queries, questions, and concerns were kept so they could be addressed in this report.

6.3 This area is one that is continually evolving, not just the knowledge about energy in dwellings, but also the SAP and RdSAP models themselves. This section summarises some of the key technical issues, the assumptions used, and how these relate to the outputs of the modelling.

6.4 These issues have been organised below into 5 sub-sections: SAP and RdSAP issues; specific technical issues; Improvement costs Issues; Occupant related issues; and Miscellaneous issues.

SAP and RdSAP issues

Evolution of SAP and RdSAP

6.5 Both the Standard Assessment Procedure (SAP) and the Reduced Data Standard Assessment Procedure (RdSAP) use the same calculation methodology and algorithms to calculate various energy performance indicators for a dwelling, e.g. the SAP Score, the EPC Band, the Environmental Impact Score, Environmental Impact Band, the SAP fuel costs, the SAP delivered energy, the SAP primary energy and the SAP carbon dioxide emissions.

6.6 That there are two approaches to calculate the same energy performance of a dwelling reflects the evolution of SAP and the implementation of the EU's Energy Performance Building Directive (EPBD). When SAP emerged in 1993 it was initially a methodology to compare two private sector energy rating schemes. SAP took on an independent life when it was then incorporated into the 1995 English and Welsh Building Regulations, and later into the 1997 Scottish Building Regulations, as one method to demonstrate compliance with the respective energy efficiency standards for new build dwellings. As demonstrating compliance for new build dwellings requires the calculation of individual U-values for, and the heat loss through, the different fabric components of the dwelling, the SAP methodology uses U-values and the area dimensions of the different fabric components as inputs into the calculation methodology. The presumption within SAP was that a person was doing the calculation using floor plans and elevations, and therefore this information would be available.

6.7 The original SAP methodology has evolved over time, so that SAP 1998, SAP 2001, SAP 2005, and SAP 2009 come and go. The current version is SAP 2012 version 9.92^[65].

6.8 In preparing for the introduction of the energy labelling of all dwellings at the point of sale or rent (energy rating of existing dwellings came into effect in Scotland on January 1st, 2009) a consensus emerged between the energy rating industry and software developers, the BRE (who developed the SAP methodology, and continue to maintain it) and the government, that a full SAP calculation was not feasible. For example, it would be very difficult to identify all the layers of an existing wall construction to allow a U-value to be calculated without intrusive surveys. Two important factors that influenced the thinking on extending energy rating to existing dwellings were the cost of producing the energy rating, and the length of time it would take to complete the survey. The more complicated the survey, the more information that would need to be collected This would lead to the survey being longer and more expensive to complete with possible consequences of market reluctance and householder hostility.

6.9 RdSAP evolved as a way of completing the SAP calculation for existing dwellings. Rather than entering the specific dimensions of all of the fabric components, a geometric model was developed to calculate the fabric components from a minimum data set. Rather than calculate the U-values of the individual fabric components, default U-values were adopted taking account of the known insulation levels, construction type of the component, and the age of the dwelling. RdSAP utilises other defaults for energy-related factors, such as the presence of flues and extract fans, and the air tightness of a dwelling. Initially, when RdSAP was introduced in 2009 these defaults could not be altered, but with the introduction of RdSAP 2009 version 9.91 (in October 2012) it became allowable for default U-values to be overwritten by assessors under specific, defined conditions. All of the default values used in RdSAP are set out in various tables within Appendix S of the SAP 2012 manual. When the REEPS modelling of the archetypes started, a working RdSAP 2012 (version 9.92) program was not available from any of the approved software developers, so an approved full SAP program was used (i.e. NES's Plan Assessor v6.1). RdSAP 2012 (also known as RdSAP version 9.92) became available when it came into effect across the UK on December 7th, 2014.

Comparing the SAP and RdSAP approaches

6.10 The two approaches are compared here by way of an example of a rectangular, 3-bedroomed 1930's detached bungalow with a cavity wall with a south facing front. The internal floor area of this bungalow is (9 x 8) 72 m2, with a floor to ceiling storey height of 2.5m, with a typical window area. There are 2 external doors: one in the hall and one in the kitchen. There is a window in each room spread across all 4 elevations. The resultant dimensional data for this dwelling for both SAP 2012 and RdSAP 2012 is set out in Table 6.1.

Figure 6.1: Floor plan of detached bungalow

Table 6.1: Comparing SAP and RdSAP entries for the same dwelling

	SAP	RdSAP
Total floor area (TFA)	9 x 8 = 72m2	9 x 8 =72m2
Heat loss floor area	9 x 8 = 72m2	assumed from floor area = 72m2
Heat loss roof area	9 x 8 = 72m2	assumed from floor area = 72m2
Gross heat loss wall area	((8 x 2.5) + (9 x 2.5)) x 2) = 85m2	exposed perimeter(9 + 8 + 9 + 8)x 2.5m height = 85m2
Net heat loss wall area	85m2 - window areas and door areas (see below) = 60.5m2	85m2 - window area and door area (see below) = 65.647m2
Window area	each window measured with details about frame, type of glazing, gap in multiple glazing, whether low -e coated, whether inert gas filled, and orientation - lounge = 3m2 to front and 2.07m2 to left side.	Calculated by formula in Appendix S
	Hall = 1.26m2 front.
	Bed1 = 3m2 to front and 2.1m2 to right.	(.122 x TFA + 6.875) = 15.659m2
	Bed 2 = 2.07m2 to rear.
	Bed 3 = 2.07m2.	Orientation - east / west.
	Bath = 0.864m2.
	kit = 2.07m2 to rear and 2.07m2 to left side.
	Total = 20.54m2
Door area	Front door (2.1 x 1) = 2.1m2	2 doors x RdSAP default of 1.85m2 = 3.70m2
	Rear door (2.1 x 0.9) = 1.89m2
	Total 3.99m2
Main living area %	Lounge area / TFA = 20 /72 x 100% = 27.78%	4 habitable rooms - 25% from Appendix S table S16

6.11 From Table 6.1, the SAP and RdSAP approaches produce different net wall areas (with implications for the total heat loss of the dwelling) as a result of using the typical window area approach in RdSAP compared to the full SAP calculation. RdSAP also assumes all windows are east-west facing, which is not the case here. These differences could be reduced by measuring the actual windows during the RdSAP survey, as could the impact of the very different solar gains from the different orientations, as RdSAP does have the facility to enter specific window details (although most RdSAP assessors probably do not use it). There would be some further difference in the calculated heat loss of the dwelling as a result of the difference in the total door area and the difference in the main living area percentage, neither of which can be adjusted in RdSAP currently.

6.12 Keeping the other data entry items the same as much as possible in the two programs, the overall difference in the SAP score produced by full SAP 2012 and RdSAP 2012 programs for this dwelling is small, but there is a difference. SAP 2012 produced a SAP score of 60, while RdSAP 2012 produced a SAP rating of 58. When the full window details were entered into RdSAP, the resultant difference is reduced as the RdSAP SAP score increases to 59.

6.13 More complicated dwellings are likely to give rise to much greater variation in the SAP scores produced by the two approaches where the actual dimensional details (and in particular, the windows areas) diverge significantly from the defaults. However, as seen above, it is possible to overwrite many of the details in RdSAP.

6.14 An important qualifier in this sample comparison was "keeping the other data entry items the same as much as possible". Both of these sample assessments used the RdSAP default U-values. This issue is discussed further in relation to the technical issues raised about the use of SAP and RdSAP.

SAP 2012 modelling of the SHCS dataset

6.15 The starting point for modelling the SAP scores for all of the 355 REEPS archetypes was the SHCS data collected during the 2010 - 2012 surveys. The SHCS was designed to collect the necessary data to allow the SAP scores and EPC bandings to be calculated as part of the outputs from the SHCS data analysis. A complication here is that the revisions to RdSAP data set are occurring faster than the revisions to the SHCS form. While many of the RdSAP revisions since 2012 have been to allow new technologies to be included within the RdSAP methodology, these technologies have not yet percolated into mainstream housing, and especially not amongst the dwellings with the poorer ratings. These are not the only changes. For instance, the SHCS does not collect the number of doors in a dwelling or whether the doors are insulated; it does not collect whether the windows are draught proofed; it does not collect any information about the length or construction of party walls; it does not collect whether solid walls or stone walls pre-1976 are lined or not; nor does the SHCS collect the age of the room of the roof (if one is present). To accommodate these data differences, assumptions are necessary.

6.16 Effectively a 2-page survey form was created to collate the necessary SAP data from the SHCS data set for each archetype (a sample is included in Appendix 11 of this report).

6.17 Then, the dataset for each archetype was to be converted into a full SAP data set. To do so required:

• the exposed perimeters for each storey of each dwelling to be converted to gross wall areas
• the areas and perimeters to be converted to internal areas and perimeters where the SHCS data set had measured externally (the formulae for the different dwelling types are set out in Table S2 in Appendix S of the SAP 2012 manual)
• the storey height of upper floors to be increased by 0.25m (in keeping with the procedure set out in Appendix S that RdSAP uses to work out the total volume of a dwelling and the total wall area)
• the RdSAP window areas to be calculated for typical, more than typical, and less than typical options as identified by the SHCS survey (again, the formulae for the different dwelling types and age bands are set out in Table S4 of Appendix S of the SAP 2012 manual)
• the U-values for the walls, roofs, and floors to be identified for the respective construction, age of the dwelling, and level of insulation present
• the U-value for the glazing to be identified
• the U-value for the doors to be identified.

6.18 As the SHCS data was entered into the full SAP 2012 program, it was also entered into a RdSAP 2009 (version 9.91) program to identify the RdSAP default U-values for the respective walls, roofs and floor constructions of each archetype. This tandem data entry approach was also used to identify the window areas for each archetype, and again for identifying the areas and default U-values of the different components of the room in the roof.

6.19 Assumptions used included:

• a house or a bungalow was assumed to have 2 doors; a flat was assumed to have 1 door, which was assumed to be located within the unheated corridor wall (where this was present) (these were the assumptions in RdSAP before v9.91 was introduced in October 2012)
• all doors were assumed to be uninsulated with a U-value of 3.0 W/m2K (the assumption in RdSAP v9.90, i.e. the version before v9.91 was introduced in October 2012)
• stone walls in pre-1930 dwellings where the wall thickness was greater than 450mm had an internal lining finish; solid brick walls in dwellings of the same age were assumed to be plastered on the hard (so no internal lining)
• the SHCS records wall thicknesses by one of 3 options: less than 450mm thick; between 450mm and 700mm thick; and more than 700mm thick: stone walls with a SHCS thickness of less than 450mm were assumed to be 400mm thick for the purposes of SAP (i.e. a midpoint between common stone wall thicknesses of 350mm and the 450mm SHCS cut-off); stone walls with a thickness of between 450 mm and 700mm were assumed to be 575mm thick for the purposes of SAP (the mean thickness); and stone walls with a thickness of greater than 700mm were assumed to be 800mm thick for the purposes of SAP (i.e. a mid point between the 700mm SHCS starting point and 900mm).
• the age of the room in the roof was the same as the age of the main dwelling
• the flat ceiling in the room in the roof was insulated to the main standard as the rest of the loft, but that the stud walls and slopes were 'as built'
• the number of habitable rooms was the number identified in the SHCS minus 1 if the kitchen was identified as a 'kitchen only' (as the SHCS habitable room count includes the kitchen); if the kitchen was identified as including another purpose (e.g. living or dining) then it was not subtracted from the habitable room count
• the number of heated habitable rooms was assessed from the room by room record data collected by the SHCS which includes assessing whether fixed heating is present
• there was draught lobby on a house or a bungalow if the SHCS indicated the presence of a porch; that there was a draught lobby on a flat or maisonette if the SHCS identified that there was a circulation area in the common block and that there were entrance doors on the front and back of the common block
• the windows were draught proofed if double glazed and not draught proofed if single glazed (these were the assumptions in RdSAP before v9.91 was introduced in October 2012)
• the doors were not draught proofed in pre-1991 dwellings unless they also had post 2002 double glazing
• the windows were orientated east - west (this is the assumptions in RdSAP unless the specific window details are entered)
• the increase of the typical window area by 25% where the SHCS identified the window area as 'more than typical', and the reduction of the typical window area by 25% where the SHCS identified the window area as 'less than typical'
• the pro rata split of the calculated window area between the main dwelling and any extensions based on the percentage of total floor area
• the length of the party wall, where the dwelling type would have a party wall, was estimated from the floor area and exposed perimeter. This was done on a case by case basis or on a trial and error basis, to obtain a fit. The construction of the wall type was based on the external wall of the dwelling (so if the external walls were cavity, then the party wall was assumed to be a cavity wall. All cavity walls and party walls were assumed to be unknown with regard to their insulation. The U-values ascribed to the different party walls were those set out in Table S8B of Appendix S.)

6.20 The SHCS collects data on the number of flues and number of extract fans within each dwelling, so there was no need for assumptions here within the full SAP data entry (although RdSAP makes assumptions on these items, and so there may be slight variations).

6.21 Before the full modelling exercise of the archetypes in SAP 2012 commenced, the data set for 12 dwellings was entered into both an RdSAP 2009 version 9.91 program and into an approved full SAP 2009 (NES's Plan Assessor 5.5.6 program) using the above assumptions and the results were compared. In all cases, the results were within ±2 SAP points (and 7 of the 12 achieved the same SAP score). The error bar used in assessing the accuracy of SAP programs as part of the approval process is ±4 SAP points.

6.22 One climatic assumption was made with the modelled archetypes. Both SAP and RdSAP use the post code to identify the degree day climatic zone, wind speed, and height above sea level used within the energy calculations. All of the archetypes were modelled on the Dalkeith post code of EH22. This post code has been identified previously by the SHCS as the median Scottish post code with regard to height above sea level, with a height of 61m above sea level. The post code and climatic zone has a variable role with the SAP / RdSAP process (see Appendix 6 where this is discussed in more detail).

U-value assumptions in RdSAP

6.23 The default U-values used in RdSAP are set out in various tables within Appendix S of the SAP 2012 manual:

• Table S7 - wall U-values (see Figure 6.2)
• Table S8B - party wall U-values (see Figure 6.3)
• Table S9 - roof U-values where thickness of insulation is known (see Figure 6.4)
• Table S10 - roof U-values where thickness of insulation is unknown (see Figure 6.5)
• Table S11 - thicknesses of insulation in ground floor (see Figure 6.6).

6.24 RdSAP includes a range of default U-values of common wall types as standard: granite or whinstone, sandstone, solid brick, cavity, timber frame, cob and system built (see Figure 6.2) and organised by age bands^[66]. Within Table S7, it can be seen that up to 1964, the important determinant for unimproved walls is the type of wall construction. From 1965 (and the arrival of the Scottish Building Regulations) the important determinant is actually the age band of the dwelling because of the increasingly higher insulation standards required to meet the respective Building Regulations. It is to be noted that for the U-values for pre-1965 stone walls, the Table refers to a footnote where the U-value is derived from formulae that varies by the type of stone and the thickness of the wall.

Figure 6.2: Table S7 Wall U-values in RdSAP 2012^[67]

6.25 RdSAP includes a range of default U-values of common wall types as standard either granite or whinstone. The degree of heat loss varies depending on the dwelling type and nature of the party wall (see Figure 6.3).

Figure 6.3: Table S8B Party Wall U-values in RdSAP 2012^[68]

Party wall type	Party wall U-value
Solid masonry / timber frame / system built	0.0
Cavity masonry unfilled	0.5
Cavity masonry filled	0.2
Unable to determine, house or bungalow	0.25
Unable to determine, flat or maisonette	0.0

6.26 RdSAP uses two approaches for establishing the default U-value, one where the thickness of the insulation is known for certain roof types (see Figure 6.4), and another where the thickness of insulation is unknown (see Figure 6.5).

Figure 6.4: Table S9 Roof U-values (known insulation) in RdSAP^[69]

Insulation thickness at joists (mm)	Assumed roof U-value (W/m²K)
	Slates or tiles	Thatched roof
None	2.3	0.35
12	1.5	0.32
25	1.0	0.30
50	0.68	0.25
75	0.50	0.22
100	0.40	0.20
150	0.30	0.17
200	0.21	0.14
250	0.17	0.12
270	0.16	0.12
300	0.14	0.11
350	0.12	0.10
>= 400	0.11	0.09

Figure 6.5: Table S10 Roof U-values (unknown insulation) in RdSAP^[70]

Age band	Assumed Roof U-value (W/m²K) (a)
	Pitched, slates or tiles, insulation between joists or unknown	Pitched, slates or tiles, insulation at rafters	Flat roof (b)	Room-in- roof, slates or tiles	Thatched roof (c)	Thatched roof, room-in- roof	Park home
A, B, C, D	2.3 (none)	2.3 (1)	2.3 (1)	2.3 (1)	0.35	0.25	-
E	1.5 (12 mm)	1.5 (1)	1.5 (1)	1.5 (1)	0.35	0.25	-
F	0.68 (50 mm)	0.68 (1)	0.68 (1)	0.80 (1)	0.35	0.25	1.7
G	0.40 (100 mm)	0.40 (1)	0.40 (1)	0.50 (1)	0.35	0.25	0.6
H	0.30 (150 mm)	0.35 (1)	0.35 (1)	0.35 (1)	0.35	0.25	-
I	0.26 (150 mm)	0.35 (1)	0.35 (1)	0.35 (1)	0.35	0.25	0.35
J	0.16 (270 mm)	0.20	0.25	0.30	0.30	0.25	-
K	0.16 (270 mm)	0.20	0.25 (2)	0.25 (2)	0.25 (2)	0.25 (2)	0.30
L	0.16 (3) (270 mm)	0.18	0.18	0.18	0.18	0.18	-

(a) If the roof insulation is “none” use U = 2.3 (all roof types).

6.27 RdSAP calculates the floor U-value from formulae taking account of the area to perimeter ratio, floor construction and the thickness of insulation. As floor insulation will not usually be known unless an intrusive survey is carried out, RdSAP has published default insulation thicknesses to use in the U-value calculations (see Figure 6.6). The other variables in the equation would be collected during a survey.

Figure 6.6: Table S11 Floor insulation thicknesses in RdSAP^[71]

Age band	Floor construction (1)	All-over floor insulation (2)
		England & Wales	Scotland	Northern Ireland	Park home (3)
A, B	suspended timber (4)	none	none	none	-
C to F	solid	none	none	none	none
G	solid	none	none	none	25 mm
H	solid	none	25 mm	25 mm	-
I	solid	25 mm	50 mm	50 mm	50 mm
J	solid	75 mm	75 mm	-	-
K	solid	100 mm	100 mm	100 mm	70 mm
L	solid	100 mm	120 mm	100 mm	-

(1) Where floor construction is unknown
(2) For floors which have retro-fitted insulation, use the greater of 50 mm and the thickness according to the age band.
(3) Suspended timber in all cases.
(4) Solid ground floor if underfloor heating.

6.28 These default U-values were used in the SAP 2012 modelling of the 355 archetypes. Other U-values could have been used. For the future of REEPS, it is important that it is possible to replicate the SAP results in RdSAP. So, while it is possible to overwrite U-values in RdSAP programs, the protocol conventions are quite explicit governing the over-writing of U-values and the documentary evidence that must be collected to support this action. These conventions limit most assessors to using the default U-values.

6.29 The default double glazing U-values for double and triple glazing were taken from Table 6a of the SAP 2012 manual rather than using the limited RdSAP defaults in Appendix S, to more appropriately reflect the common forms of double glazing found in Scottish dwellings. RdSAP does allow for the use of Table 6a glazing U-values to be used instead of the Appendix S defaults, so this is consistent with the RdSAP methodology.

Fuel costs used in SAP and RdSAP

6.30 Are there differences between real world fuel prices, and those used in SAP and RdSAP modelling? In particular, how do assumed and actual costs of biomass fuels compare?

6.31 SAP and RdSAP do not just use one set of fuel prices. They both use the same fuel cost basis, but within the modelling for EPCs, different reference fuel costs are used for specific purposes (see Appendix 7 for a fuller discussion on this).

6.32 SAP and RdSAP assume all fuels are bought in, that is, they do not take account of the possibility of 'free fuel', whether this comes from collecting sea coal from beaches, getting free coal because of past employment as a coal miner, collecting wood, or coppicing one's own wood lot.

Supplier switching / Tariff switching / Dual Fuel prices

6.33 How is switching supplier / switching tariffs / dual fuel tariffs accommodated for in the modelling of improvements?

6.34 Comparing the impact of switching suppliers, or buying into dual fuel discount tariffs, is outwith the ability of RdSAP or SAP to accommodate per se. There is no ability within the program to change unit rates to see how these would affect a household's fuel bills. The information from the energy consumption outputs from the two programs could be used and compared but would have to be done either by hand or in separate software (e.g. a spread sheet).

6.35 There is a limited ability to assess the impact of switching the type of electricity tariff a dwelling is on for given heating systems in SAP and RdSAP. Effectively, there are 4 electricity rates embedded within SAP and RdSAP - standard domestic rate, a traditional off peak rate, an 18-hour rate and a 24-hour rate. So it is possible to assess electric storage heaters on standard domestic rate, a traditional off peak rate, and a 24-hour rate; it is possible to assess electric room heaters on standard domestic rate, a traditional off peak rate and a 24-hour rate; it is possible to assess electric wet systems on standard domestic rate, a traditional off peak rate and an 18-hour rate. The results can then be evaluated to determine if there are any benefits for the householder to move them from one generic tariff to another. Several of the improvement measures included this comparison where electric heating was installed in the property (see the various electricity related improvement measures in Appendix 2 for more detail).

SAP or RdSAP?

6.36 Can full SAP be used (by assessors) instead of RdSAP? Can the EPC register 'take' full SAP EPCs?

6.37 Lodging EPCs for existing dwellings using SAP software has been possible in Scotland since 2013. However, for an assessor to lodge an EPC with SAP software they will need to have access to this software. This access will probably necessitate the RdSAP assessor acquiring additional competencies from a membership organisation to qualify as an assessor of new build dwellings from plans (known as an On Construction Dwelling Energy Assessor (OCDEA)).

Consistency of SAP / RdSAP ratings over time

6.38 As SAP and RdSAP evolve, e.g. as fuel prices, default 'u-values' change, will the impact of measures affect the resultant SAP rating differently?

6.39 The fuel prices that SAP and RdSAP reference within the PCDF are constantly evolving (see Appendix 7). The default U-values have had minor amendments applied to some levels of loft insulation in RdSAP for the first time, as opposed to additional defaults being added (which has happened at every revision of RdSAP). New tariffs governing certain types of wet central heating systems were added to RdSAP 2012. So yes, the SAP rating could vary between one version of RdSAP and the next.

6.40 The incorporation of party walls as a heat loss component within some dwelling types within RdSAP 2012 will have a major impact on some dwelling types, and no impact on others.

6.41 With each new version of SAP comes a basic relationship table to show the difference between one version and the other by fuel type. The relationship table for comparing SAP 2009 with SAP 2012 ratings is shown in Figure 6.6.

6.42 From the relationship table in Figure 6.7 it would appear that there is no difference between SAP 2009 and SAP 2012 ratings where gas heating is involved, and between 2 and 3 point decline where electric heating is involved. This table does not take account of changes such as the inclusion of party walls as a heat loss component that can reduce ratings by about 5 points (all other things being unchanged) or the introduction of the E18 hour tariff for electric wet systems that can increase the rating by over 20 points.

Figure 6.7: Table 15: Relationship between SAP 2009 and SAP 2012 ratings^[72]

	SAP 2012 for main heating fuel as:
SAP 2009	Mains gas	LPG	Oil	Electricity	Solid mineral	Biomass
1	1	(-9)	(-11)	-3	(-5)	3
10	10	1	(-1)	6	5	12
20	20	12	10	16	15	22
30	30	23	21	27	26	32
40	40	34	32	37	36	42
50	50	45	43	47	47	51
60	60	56	54	58	57	61
70	70	67	65	68	67	71
80	80	78	76	78	78	81
90	90	89	87	89	88	91
100	100	100	98	99	99	100

Accuracy within SAP / RdSAP

6.43 Are there issues with accuracy when using EPCs or the SAP methodology to assess the energy performance of dwellings?

6.44 Approved SAP and RdSAP programs are accurate within defined error bars when assessing the energy performance of the dwelling, assuming the correct information is entered into the programs. A number of government policies are already using EPC bandings or SAP scores. Notwithstanding RdSAP's limitations discussed in this section, the Scottish Government took the decision that this tool is currently the best one available for taking forward REEPS.

Replicability of the results

6.45 Can an EPC assessor replicate the modelling in the field?

6.46 The REEPS modelling was not a theoretical exercise divorced from real dwellings. As the bases of all 355 archetypes were actual surveys of actual dwellings, then an EPC assessor should be able to collect the same data. SAP and RdSAP fundamentally use the same algorithms to calculate the SAP rating - so that if the same data is entered into the two programs they should produce the same results. While the results were modelled in a full SAP program, which may not be available to all EPC assessors, the RdSAP default U-values were used in the SAP modelling, for just this reason. A sample of 12 dwellings was modelled in both SAP and RdSAP to compare the results from the two programs with very similar results.

Specific technical Issues

6.47 The approach to identifying the poorer performing dwellings within the Scottish dwelling stock, by developing the initial typology and then the further segmentation into the 355 archetypes, was intended first to group properties with similar characteristics (i.e. heating system and wall construction). Secondly, to capture the diversity of dwellings within a typology grouping by further segmentation to include dwellings from the different SAP rating bandings and size. The improvement measures modelled within each archetype were tailored to the specific heating and insulation characteristics of individual archetypes (see Appendix 2 for a complete description of the improvement measures and when these measures were not modelled). Critical considerations included ensuring that every set of improvement measures was bespoke to the archetype, that all improvements modelled were technically feasible, and that all of the improvements could be modelled in SAP and RdSAP.

6.48 A variety of technical issues and concerns were raised in the discussions with the REEPS RAG during the lifetime of this project over the methodology and assumptions used in the modelling of the archetypes and the assumptions used in modelling the improvement measures. These concerns and issues are detailed below.

Traditional Buildings

6.49 Are traditional buildings fundamentally different than other dwellings? Are there aspects of traditional dwellings which could be different?

6.50 At one level, using the label 'traditional' is not at all helpful. It implies that there is also a 'non-traditional' group of dwellings, but there is no agreement or consensus about what forms of dwelling constructions or age bandings comprise these two groups. Conceptually, it is too imprecise. The development of the typology focussed on breaking the stock down to a much finer resolution than 'traditional'.

6.51 Accepting that 'traditional' is short hand for certain older construction types, there is an emerging and growing literature and experience with improving older dwellings and particular types of older constructions. Organisations such as Historic Scotland and others are conducting research and producing technical papers on just these issues^[73]. What is emerging from this research is that it is possible to improve the energy efficiency of these dwellings and that, in principle, there is no need to preclude or exempt them from trying to achieve these higher standards for energy performance. That said, there are technical concerns with regards to the types of insulation materials being used and the techniques associated with their application that do need to be taken into consideration. These considerations may have cost implications for improvement measures. However, the improvement measure that gives rise to the biggest concerns - namely wall insulation in older stone and solid brick dwellings - was only needed in a small number of dwellings to achieve a higher EPC Band, and then usually only in achieving the biggest increase, namely improving dwellings in SAP banding G up to Band D.

Dealing with 'hard-to-treat' properties

6.52 Will the modelling identify appropriate measures to improve 'hard-to-treat' properties to enable them to achieve a higher EPC Band?

6.53 The report has already noted that this term is a misnomer, and one that is often wrongly applied, when the term that should be applied is 'expensive-to-treat'. Many of the wall types discussed previously under 'traditional' buildings (see section 6.49-6.51 above) would fall into this category, as would many non-traditional constructions, e.g. concrete constructions. These walls are being treated now, and have been treated for over 30 years. The insulation materials and the application techniques are known and available in the industry. Notwithstanding the discussion on technical issues regarding the suitability of certain insulation materials (see section 6.56-6.59 below), the issue is one of cost. They cost considerably more to insulate than filling a cavity wall. These higher costs were factored into the modelling of the archetype improvements. They are a major reason why many of the modelled improvement packages to meet the REEPS targets do not include solid wall insulation within them. Other measures or packages of measures to raise the energy rating standard within these properties were cheaper.

Dealing with Listed Buildings / Conservation Areas

6.54 How are listed buildings / conservation areas being dealt with in the REEPS modelling?

6.55 Listed buildings are not exempted from requiring an EPC^[74]. Nor are dwellings within Conservation areas. However, they were not a specific concern of the modelling with regard to their improvement as the SHCS does not separately identify whether a dwelling is a listed building or not, or whether it is in a conservation area, or not, so fell outside of the modelling of improvements. Where dwellings fall into these categories, their status may restrict the measures that could be undertaken to improve the property, and may require negotiation with the local planners.

Condensation in buildings

6.56 Will installing insulation measures cause problems with dampness or interstitial condensation in buildings?

6.57 First, assuming that there are no other technical issues relating to the application of particular insulation materials, even technically suitable insulation has to be installed correctly. There are BSI standards, industry guidance, and good practice guidance available. The BRE published the report, 'Thermal Insulation: Avoiding Risks' many years ago and have updated and revised it since that explicitly dealt with such issues^[75]. Over the last couple of years, the UK government has pushed the development of, and contractor compliance with, the PAS 2030 standards^[76] . Yet many contractors still make elementary mistakes by blocking up necessary eaves ventilation in roof spaces, or putting cavity wall insulation into walls that should not be done, or do not draughtproof loft hatches when insulating lofts. Poor installation can cause serious problems.

6.58 A different issue is whether certain insulation materials are suitable to be used in certain types of properties, e.g. whether 'vapour closed' materials are suitable for insulating pre-1919 stone and solid brick walls, or whether only 'vapour open' materials should be used. Historic Scotland is carrying out research on these issues.

6.59 The modelling of insulation improvements to solid walls within the archetypes assigned a default U-value to the improvement. RdSAP makes no specific determination of the type of material used in insulating a solid wall, other than set out that the basis of the U-value improvement for solid wall insulation is 100mm of a material with a thermal conductivity of 0.04 W/mK. So this default improvement could be achieved by installing 100mm of mineral wool, or 100mm of sheep's wool, or 100mm of cellulose fibre insulation. Historic Scotland's research has shown that U-values of 0.35 W/m2K can be achieved using 'vapour open' materials. The householder, designer or contractor prior to installation should carry out a technical assessment of the suitability of any insulation material for a given situation.

Default U-values and actual energy performance of dwellings

6.60 Are default U-values used in RdSAP markedly different from those measured 'in situ', which will have an impact on the assessment of the energy performance of a dwelling and the standards achieved through the improvements?

6.61 This issue goes well beyond the interests of REEPS, and affects all assessments of the energy performance of a dwelling. Again there is a growing literature arising from research projects suggesting that in-situ measurement of U-values are demonstrating differences with the defaults^[77] , but again that is nothing new. Thermal conductivity values are usually measured in the laboratory not 'in situ', and do not take account of the vagaries of the many things on site that could affect the actual U-value performance, therefore it should be expected that there will be differences.

6.62 RdSAP is measuring the notional asset performance across the whole of the UK, thus the publication and use of the standard defaults. Should these defaults be changed every time a research project identifies a difference, because different reports show different degrees of variation? Should the results of small samples be extrapolated and applied across all dwellings? Statistical testing of the validity of these research findings would require much larger sample sizes. As noted above however, RdSAP and the use of defaults evolved to deal in part with the issue of the cost of the survey. In-situ measurements of U-values are not cheap, but where they are undertaken, there is nothing in RdSAP to prevent the default U-values being over-written. Where measured U-values are better than the defaults, there would be an improvement in the energy performance of the dwelling, making it less onerous to meet any subsequently defined REEPS standard.

Indoor air quality

6.63 Will setting a standard in REEPS affect indoor air quality by increasing the air-tightness of the dwelling?

6.64 Air quality is likely to become an issue as properties become more air tight. The building regulations are effectively pushing dwellings towards requiring mechanical ventilation as they increase the degree of air tightness assumed in the compliance calculations.

6.65 Within the improvements modelled, the only ones that had explicit implications for increasing air tightness were the draught proofing of windows and doors, and the installation of double (including secondary) glazing or triple glazing where single glazing existed. To meet the modelled REEPS standards, changing from single to double was only included as a measure in a small number of dwellings. By contrast, the draught proofing of windows and doors (and this was usually just the door as the majority of dwellings were double glazed throughout) was a more common recommended improvement. Past research has shown that draught proofing is very effective at reducing draughts that are most noticeable by household occupants, but only affects about 10% of the overall air infiltration rate within existing dwellings, as much of the air infiltration is coming through sources other than the windows and doors. That said, there are many sources of good/recommended practice that note that draught proofing of windows in kitchens and bathrooms should take into consideration other forms of ventilation to avoid problems with condensation.

6.66 Equally, it may improve the air quality where properties were otherwise cold. Other insulation improvements could have an indirect consequence for the air change rate within the dwelling, e.g. floor insulation beneath a suspended timber floor will effectively reduce air infiltration through this pathway; external cladding of a dwelling may block up unintended air infiltration pathways. Insulation should not be applied in such a way as to restrict intended ventilation pathways, i.e. air bricks or extract fans.

Boiler Oversizing

6.67 Will improved energy efficiency lead to existing boilers now being over-sized?

6.68 In practice, possibly: in theory, no. Where significant amounts of insulation are added to the dwelling then the existing boiler may be oversized given the revised heat load from the dwelling. Historically, in older, pre-1998 boilers, this would have resulted in declining boiler efficiency. Since 1998, and the implementation of the EU's Boiler Efficiency Directive, the differences in boiler efficiencies between full load and 30% part load are limited to 3%, so increased insulation should only have a marginal effect in these situations.

6.69 However, within SAP and RdSAP this is not an issue, even when the boiler is specifically identified through the PCDF and therefore its output is known, SAP and RdSAP assume that that boiler is perfectly sized to meet the heat demand of the dwelling. SAP and RdSAP not only ignore oversizing, they also ignore under sizing (where the boiler is inadequate to meet the heat load of the dwelling).

6.70 SAP and RdSAP are not boiler sizing programs.

Biomass boiler efficiencies

6.71 Are higher biomass boiler efficiencies reflected in the modelling?

6.72 At one point in SAP and RdSAP the efficiencies that could be applied to biomass boilers were the default efficiencies from Table 4a of the SAP manual that limited them to between 63% and 65%. With the extension of the PCDF to include more than just gas, oil and LPG boilers, biomass boilers are included as an option. For biomass the efficiencies range between 78% and 87% and for wood log boilers, they range between 67% and 86% (see Figure 6.8a and 6.8b respectively).

Figure 6.8a: Sample of wood log boilers in PCDF with efficiencies

Figure 6.8b: Sample of wood log boilers in PCDF with efficiencies

Rooms in the roof

6.73 How does SAP and RdSAP account for rooms in the roof?

6.74 Rooms in the roof are particularly problematic. They come in many different shapes / sizes / configurations - with and without dormers. RdSAP allows for a simple way of dealing with rooms in the roof by simply measuring the floor area. If an assessor adopts this approach it calculates the overall surface area of the room in the roof components from equations (that are set out in the SAP manual) and allocates areas to different components of the room in the roof using more equations. Alternatively, since RdSAP v9.91, RdSAP allows for all the individual components to be measured separately, and all of the individual U-values to be calculated and entered into RdSAP both for an existing dwelling and for one with any improvements. RdSAP does allow for the different elements of dormers to be surveyed and included in the assessment if the assessor chooses to measure them and then model them. However, on this matter the choice is left to the assessor, who will probably make the decision based on how much s/he is being paid to do the survey.

6.75 The nature of the improvements that are recommended by RdSAP are prescribed by the rules set out in Appendix T of the SAP manual, and a minimum requirement of raising the SAP score by at least 1 point. So with an uninsulated room in the roof, RdSAP will more than likely recommend room in the roof insulation as an improvement because it is within the Appendix T prescriptions and will raise the SAP score by more than a point. The recommendation will appear on the EPC. This recommendation is regardless of whether there are technical concerns or otherwise about doing the suitability of undertaking this insulation work. That is, unless the assessor suppresses the recommendation, which is within the control of the assessor, and then must provide reasons why this was done.

6.76 There is also a note on the EPC that householders should seek technical advice on the suitability of any of the recommendations before being carried out. This is no different with rooms in the roof than it is with cavity wall insulation, fitting heating systems.

6.77 More on the modelling of rooms in the roof within the archetypes is discussed in M3 in Appendix 2 of this report.

Improvement cost issues

Costs of improvement measures

6.78 On what are the improvement costs based? If they are not reflective of real life they will over or under-estimate the actual costs.

6.79 The costs used are not based on real life but on the mean indicative costs drawn from the PCDF (see Chapter 4). They are intended to be indicative rather than actual costs. The median dwelling within each grouping was chosen as the archetype, and as such, the costs are as likely to over-estimate as they are to over-estimate the actual costs for a particular dwelling. As these costs are then grossed up to reflect the national stock then over-estimating and under-estimating the costs should cancel out in in the overall statistics when costing improvements for an archetype.

6.80 Attempts were made to get 'real costs' by interviewing contractors, but these too produced a range of costs. More discussion on the approach to identifying costs of measures is set out in Appendix 3.

Additional Costs, including maintenance costs

6.81 What costs should be counted in modelling impact of works?

6.82 In discussions within the REEPS RAG, whatever costs were used for modelling, there were concerns that households in more rural areas and the island groups would have to pay additional costs likely to be incurred while installing certain measures for ancillary services or transport costs. The basis of the costs ultimately decided upon were the mean indicative costs within the PCDF that SAP and RdSAP use. The background to the PCDF costs was explored, and alternative costs were sourced from several contractors, and another source (see Appendix 3 for a fuller discussion on this). The decision was taken that maintenance costs were an issue for the householder or owner, and acknowledged that rural areas and the islands will face higher costs because of the market, but these were not specifically quantified.

Occupant related issues

Occupant behaviour and energy savings

6.83 Will the occupants' use of a property following improvements affect the actual levels of savings achieved?

6.84 Undoubtedly, yes. SAP and RdSAP are asset based ratings, that is, they are about the energy performance of what is installed in the dwelling, and these are modelled using standard assumptions about occupant behaviour. Occupant behaviour does not affect the SAP or RdSAP assessments of the dwelling because it is effectively controlled in the modelling. In reality, the biggest difference between varying energy consumption in dwellings is often related to how occupants use their dwellings. To ensure that householders understand how to take advantage of the energy efficiency improvements, and maximise the benefits, energy advice should be an integral part of any heating and insulation improvement programme.

Occupant behaviour and condensation

6.85 Will the occupants' use of a property following improvements create problems of condensation? Occupants' use of a property can cause problems with condensation before or after insulation or heating related improvements.

6.86 In principle improving the heating and improving the insulation standards within dwellings should reduce the risk of condensation occurring. However, condensation occurs when there is an imbalance in the equilibrium between heating, insulation and ventilation, so the occupants' use of the heating or approach to ventilation can effect this equilibrium, so condensation can occur. To minimise the risks of condensation occurring and to ensure householders understand how take advantage of the energy efficiency improvements, and maximise the benefits, energy advice should be an integral part of any heating and insulation improvement programme.

Miscellaneous issues

Emerging Technologies

6.87 How are emerging technologies accounted for in SAP and RdSAP?

6.88 SAP and RdSAP are continuously evolving methodologies (as discussed above in Sections 6.5 - 6.9). One impetus for this evolution is that new technologies emerge. Simply because a new product appears on the market does not mean it will be incorporated immediately into the SAP methodology. There are published protocols for testing new products, and SAP has the facility and defined conventions to include the impact of new technologies on the energy performance of a dwelling. RdSAP is not as flexible, and there is usually a time lag between agreeing the impact of a new technology into the SAP methodology and including it in RdSAP. RdSAP v9.92 saw the inclusion of high heat retention storage heaters and new waste water heat recovery systems, and new control devices and systems.

Embodied energy

6.89 As one of the aims of REEPS is to reduce carbon emissions by improving the energy efficiency of properties, will the embodied carbon content in these measures be assessed?

6.90 No. Embodied energy was considered outwith scope of Scottish Government policy of seeking to reduce emissions so not included in the modelling of the archetypes or their improvements.

Feed-in-tariffs and Renewable Heat Incentive

6.91 Are payments for feed-in-tariffs and the Renewable Heat Incentive (RHI) taken into account in the calculations on cost effectiveness of measures?

6.92 No. For the measures governed by the feed-in-tariffs (FiTs) (i.e. PVs and the 2 wind turbine options) a value for the generation from the specific technology was available because the same size system was installed in every case where it was modelled. However, the rates being paid out through the FiTs is constantly being revised, and usually downwards, so any value may be quickly out of date. For RHI technologies (various heat pumps and biomass boilers), the payments are related specifically to the individual dwelling and would be specific to each of the applicable technologies, and have eligibility rules, so were deemed to be outwith the scope of this modelling. A note has been inserted into each archetype document where these technologies were assessed pointing out the availability of FiTs and RHI and that these payments may make these technologies much more cost effective for the householder to install.

In-use factors

6.93 Were in-use factors taken into account within the modelling?

6.94 The two initial sample dwellings took in-use factors into account because they examined lifetime energy reductions and lifetime CO_2e emission reductions. However, as the final decision was taken to model the packages on the basis of the cheapest cost to achieve the next EPC banding, it was not necessary to calculate the impact of the in-use factors.

Community heating

6.95 Were community heating and combined heat and power (CHP) taken into account in the improvement measures?

6.96 No. Community heating and CHP are valid interventions for certain properties in close proximity to one another. It was decided not to model them within the archetypes as an improvement measure because these options were deemed too site specific, and the actual systems are usually one-offs with individually calculated efficiencies.

Delivered Energy or Primary Energy

6.97 Did the modelling use delivered energy or primary energy in the modelling?

6.98 SAP and RdSAP calculate both delivered energy (i.e. amount of energy consumed in the home) and primary energy (i.e. the total amount of energy consumed by the nation, e.g. at the power station, to deliver the energy consumed in the home). It was decided to only present the delivered energy costs on the individual archetype documents as this related to what householders consumed in the dwelling.

Other Issues

6.99 Over and above all these issues, various individual factors were discussed that may have an impact on delivering on REEPS but fell outside the scope of this research project. These included the infestations of vermin and the presence of bats. These issues might affect the installation of improvement measures in individual situations, but were not accounted for in modelling archetypes. RdSAP and Green Deal assessors have the opportunity to 'turn off' recommendations arising from assessments of properties where local circumstances would prevent an improvement being carried out.