FACTORS INFLUENCING AIRTIGHTNESS OF DWELLINGS   

YEAR OF CONSTRUCTION

Airtightness is influenced, to a certain degree, by when the dwelling was built. In those countries that have well established whole building airtightness requirements, such as Canada, Sweden and Switzerland, new dwellings tend to be more airtight that older dwellings.

In countries with less well established airtightness requirements, such as the UK, the trend for new dwellings to be more airtight than older dwellings is almost non-existent. This was highlighted in work undertaken in the late 1990’s by the BRE (see Stephen, 1998 & 2000), which suggested that dwellings constructed at the beginning of the 20th century are as airtight as those that have been constructed since the 1980’s.

Effect of age on airtightness. Adapted from Stephen (2000) and Olivier (1999)

Recent tests undertaken by Leeds Metropolitan University on a small number of dwellings constructed to comply with the 2002 Edition of Approved Document Part L1 found that the air leakage of the dwellings was broadly in line with the air leakage of the existing UK housing stock as a whole (Johnston, Miles-Shenton, Bell & Wingfield, 2006).

Given the qualitative nature of the work, it was not possible to extrapolate the results to the post 2002 new build stock with any degree of confidence, however the results obtained were broadly in line with the results from pressurisation tests on 99 post-2002 dwellings (also not a random sample) reported by Grigg (2004). These results suggest that, in the UK, there has been no real improvement in the airtightness of new dwellings following the introduction of the 2002 Edition of Approved Document Part L1.

CONSTRUCTION TYPE

Airtightness is strongly influenced by the type of construction. Theoretically, certain types of construction are intrinsically more airtight than other methods of construction. For instance, wet plastered masonry and concrete overseas dwellings tend to be more airtight than comparable timber or steel-frame dwellings (Olivier, 1999). Evidence of this has been gained from pressurisation tests undertaken on a large sample of dwellings in Norway and Sweden (Olivier, 1999). These tests indicated that the dwellings constructed from concrete were tighter than plastered brick masonry dwellings, which in turn were tighter than the timber-frame dwellings.

Conversely, in the UK, cavity masonry construction tends to be inherently leaky. Work undertaken by the BRE (Stephen, 2000) on different types of wall construction found that cavity masonry walls were considerably leakier than solid masonry, timber-framed and large panel system (LPS) walls (see graph below). The reasons for this appear to be twofold:

1. The use of plasterboard dry lining which can be a significant source of air leakage if it is not properly edge sealed (see Stephen, 1998 & 2000 and Lowe, Johnston & Bell, 1997), as the air gap behind the plasterboard effectively interconnects all the air leakage paths within the dwelling, producing a complex network of inter-linked voids.
2. The use of timber intermediate floors that are supported using built-in joists. Built-in joists are also known to be a significant source of air leakage (Stephen, 1998 & 2000).

Influence of wall type on mean air leakage rate in the UK.
After Stephen (2000)

Both of these construction practices contrast with the rest of Europe, where wet plaster is used as an internal finish to masonry walls and in-situ concrete upper floors were adopted many decades ago (Olivier, 1999). Further information on the issues associated with using plasterboard dry lining and built-in joists can be found within the Construction Observations section.

One of the most important influences on the eventual airtightness of a dwelling is the location and the continuity of the primary air barrier. In the UK, it has not been common practice to identify the primary air barrier on the design drawings, or to identify those areas of the design where attention to detail is required to ensure airtightness and provide the required detailed design guidance. Instead, it is common for the design drawings to contain little or no information on the location of the primary air barrier or airtightness issues. This results in those involved in constructing the dwelling not being aware of the location of the primary air barrier, its purpose, the importance of maintaining continuity of the air barrier, nor areas of the construction where particular attention to detail is required to ensure airtightness. Consequently, dwellings are constructed with no explicit primary air barrier. Instead, different elements within the dwelling default to form part of the primary air barrier. One such element that commonly defaults to form part of the primary air barrier is the plasterboard dry lining in masonry cavity construction, and it is known that plasterboard dry lining can be a significant source of air leakage.

The lack of detailed design relating to specific details and airtightness issues also results in the construction teams having to make decisions on critical details based upon the limited design information that is available and the experience gained from constructing similar details on other dwellings. This inevitably can have an adverse effect on the airtightness of the dwelling.

There is some evidence to suggest that the number of storeys has an influence on the air leakage of dwellings. Work undertaken in the United States by Sherman & Dickerhoff (undated) on a database of almost 13000 air leakage measurements suggests that multi-storey dwellings are leakier than single-storey dwellings. Similar findings have also been found in Canada (see Allen, 1985 after Sulatisky, 1984). This is probably a result of the greater influence of wind pressure effects and the stack effect on taller buildings.

SIZE & COMPLEXITY

Other things being equal, the larger and more complex the floor plan and the more complex the construction techniques used, the greater the number of junctions between the elements of the thermal envelope. This increases the potential for air leakage.

Section through a dwelling illustrating the details
requiring information on air barrier continuity

Recent work undertaken on a small number of dwellings built to conform to ADL1 2002 illustrated that significant variations in air permeability (up to 4 m³/(h.m²) @ 50Pa) can be observed in dwellings of similar size, construction and form that had been constructed with comparable levels of workmanship and site supervision and by the same site team. The only observed difference between the dwellings was the complexity of the detailing. Higher levels of air permeability were consistently observed in those dwellings that contained the most complex detailing. The disparities in detailing were most common where certain design features required the primary air barrier to cope with complex changes in plane, negotiate structural members and accommodate changes in material. Such details included ground floor projections, rooms adjacent to semi-exposed areas, timber bays in masonry construction and complex junctions with ventilated cold roof loft-spaces. A more detailed discussion of these results can be found in Johnston, Miles-Shenton & Bell (2006).

This does not mean that complexity should be avoided as a matter of principle. Clearly where complexity serves no purpose there are benefits to be gained in all aspects of design by simplification, but where there are clear aesthetic or other reasons complex detailing need not be avoided. However designers and constructors need to understand the airtightness problems that may be introduced by adopting complex detailing and devise appropriate and robust solutions.

SEASONAL VARIATION

There is some evidence to suggest that air leakage is seasonal. Work undertaken on an unoccupied heated test house by Warren & Webb (1980) found a substantial seasonal change in total air leakage (about 25%), with the maximum occurring in winter and the minimum occurring during the summer. This difference was attributed to seasonal changes in moisture content of the timber. It could also be argued that the differences observed by the authors might also be due to thermal expansion and contraction effects.

More recently, a seasonal change in air leakage of around 20% was observed to occur in an unoccupied masonry cavity showhome (see Miles-Shenton, Wingfield & Bell, 2007). This difference in air leakage is assumed to have occurred due to thermal expansion and contraction effects.

The air leakage of a dwelling tends to increase over time. Increases in air leakage ranging from around 25% to more than 80% have been observed in a small number of dwellings during the first year of occupation (see Miles-Shenton, Wingfield & Bell (2007), Elmroth & Logdeberg (1980) and Warren & Webb (1980)). This suggests that the reasons for the increase in air leakage of these dwellings was probably shrinkage cracks caused by drying out and settlement of the foundations. The following photographs illustrate common areas of shrinkage in new build housing.

Shrinkage at [1] wall string on stairs, [2] intermediate floor/wall junction, [3] window sill and [4] junction between timber window frame sections

A number of other factors are known to contribute to increased air leakage over time. These include:

• Wear-and-tear of construction materials, particularly window and door seals.
• Changes to the building fabric carried out by the occupants. For instance, poor sealing of penetrations through the air barrier that have been made once the dwelling is occupied.

However, it is unclear how much additional air leakage is likely to be attributable to these different factors.

Sequencing can have an important impact on the airtightness of a dwelling. Work undertaken at Stamford Brook (Miles-Shenton, Wingfield & Bell, 2007) found that the build sequence adopted can often make it difficult to gain access to and maintain continuity of the primary air barrier. For instance, two separate approaches to constructing the top floor ceilings, which formed the air barrier to the loft, were adopted at Stamford Brook. These separate approaches were as follows:

• Approach 1 - The other developer installed the top floor ceiling prior to the installation of the metal stud partitioning.
• Approach 2 - One developer installed the metal studwork partitioning first, and then inserted a timber head plate over the top of the head channel in an attempt to reduce air movement through the head channel.

Plasterboard ceiling erected prior to partitioning

Partitioning erected prior to plasterboard ceiling

This variation in build sequence had important consequences for the continuity of the air barrier at the ceiling/studwork partition junction, particularly at the junctions between studwork partition walls. In those dwellings where a continuous plasterboard ceiling was installed, no gaps in the ceiling existed, apart from at the ceiling edge/wall junction. This contrasted with the dwellings where the ceiling had been installed after the installation of the studwork, where numerous gaps existed between the studwork and timber head plates. These separate approaches are illustrated in the photographs below.

Another common example relating to the build sequence adopted concerned a number of the service penetrations made through the external walls. Penetrations for the kitchen sink wastes in a number of dwellings were made directly through the external walls after the kitchen units had been fitted.

This made it very difficult, if not impossible, to seal the waste to the primary air barrier after the penetration had been made. In addition, it was also common for the boiler flue to be sealed internally to the plasterboard dry lining, but not with the primary air barrier due to the limited access available once the flue had been installed.

Installation of kitchen waste pipe

Sealing of the boiler flue to plaster board dry lining internally

A general lack of sequencing of the various construction processes can also result in a situation where a detail is constructed then damaged or dismantled for a subsequent installation before being repaired or reconstructed. This “build – damage – install – repair” approach very often results in damage to the primary air barrier, damage that is often not adequately repaired.

SITE SUPERVISION & WORKMANSHIP

The level and quality of site supervision and workmanship during the construction of a dwelling can influence its overall air leakage. Experience has shown that nominally identical dwellings on the same site can have very different air leakage rates and leakage distributions (BRECSU, 2000 and Allen, 1985). Consequently, workmanship is often cited as being one of the main reasons why airtightness standards are not consistently achieved in house building in the UK.

At Stamford Brook, the focus for those dwellings that were included in the detailed airtightness study was on workmanship, rather than making changes to the design of the dwellings (see Miles-Shenton, Wingfield & Bell, 2007). This focus resulted in all but one of the tested dwellings in this study achieving an air permeability of less than 5 m³/(h.m²) @ 50 Pa. Despite this result, experience suggests that focusing on workmanship per se is unlikely to lead to a consistently high (over 95%) “pass” rate at anything much below 5 or 6 m³/(h.m²) @ 50 Pa. Of course, workmanship is important, but very often it is the context in which trades have to work, the lack of specific training, the buildability of designs, the lack of detailed design and the lack of a general quality control process that underlie many workmanship problems. If such issues are not addressed, workmanship will always appear to be poor.

One of the best examples of a perceived workmanship problem at Stamford Brook was the maintenance of a continuous ribbon of plaster adhesive around the perimeter of the plasterboard dry lining. Observations of one gang indicated that with very careful attention to detail and enough time, airtightness in the region of 2 m³/(h.m²) @ 50 Pa was possible using plasterboard on dabs. However, this was the exception. Under normal subcontracting arrangements the use of plasterboard dry lining is not consistently buildable, yet site managers and developers continue to focus on the workmanship aspects rather than demand, from designers, a more robust solution.

It is also impossible to divorce workmanship, not only from design but also from other issues of construction management such as training, communication and quality control. It was clear that many operatives were keen to do a good job but that, as far as airtightness was concerned, it was difficult for them to be clear about what they had to do or who was responsible for achieving an airtight envelope. This tended to manifest itself in inconsistencies in airtightness, such as between dwellings plastered by the same team. The general picture suggests that not all support processes were in place nor was it possible, particularly at times of rapid increases in production, to maintain the necessary level of training.

QUALITY OF CONSTRUCTION

The quality of the construction can have a significant impact on airtightness. The overwhelming conclusion from the Stamford Brook Field Trial (see Miles-Shenton, Wingfield & Bell, 2007) and from more general observations of the UK housebuilding industry as a whole, is that quality control processes are extremely diffuse with a number of actors playing similar but different roles which are almost always carried out in isolation. It is perhaps not surprising that with no clear airtightness quality control process in place, sequencing is often out of phase and known errors tend to be repeated time and time again.

At Stamford Brook, a reoccurring fault was observed at the ground floor slab extensions at the thresholds. This is illustrated in the photographs below.

Misalignment of ground floor slab thresholds

In one of the earliest dwellings constructed on-site, the slab extension at the threshold was often misaligned with the opening [1]. In some cases, the misalignment was as much as 450mm and it was common occurrence for air leakage to be observed at these thresholds when the dwellings were tested. Although this information was feed back to the site management teams, the same errors were still observed in dwellings constructed later on in the development [2] & [3]. The fact that the same error was still being made later on in the construction phases shows that the feedback loop to the existing current quality control system has not been set up to deal with such issues adequately.

Another quality control issue observed at Stamford Brook that can lead to airtightness problems concerned the different tolerances that were worked to by different trades. Items and components that are manufactured off site to high tolerances are often fitted into structures built by trades that are not operating to the same precision. Consequently, workmanship and performance issues can arise when construction is outside these tolerances. However, even when trades work within their expected tolerances, the discrepancies between the different elements and materials can result in larger than expected gaps. Subsequently, adjustments often have to be made to the construction which can have an adverse affect on Airtightness. At Stamford Brook, variations in the cavity width and the offset between the outer brick leaf and the inner blockwork at openings often resulted in poor fit of the propriety cavity closers. This often gave rise to some physical damage to the closer. In these particular cases, the closers were replaced, but in other cases such small gaps can easily be overlooked and remain as hidden air leakage paths on dwelling completion.

Variations in cavity width resulted in the origially specified cavity closers having to be replaced

COMMUNICATION

Studies at Stamford Brook have highlighted the critical nature of communication and the potential impact that it can have on airtightness (see Miles-Shenton, Wingfield & Bell, 2007). It is clear that there is considerable scope for improvement in flows of information both upwards and downwards throughout the organisations involved whether developer, designer, subcontractor or individual trade.

Very often, in common with many other sites observed nationwide, design information was not available, not at the right level of detail, confusing or just not referred to by operatives. This led to a rather diffuse process as operatives followed their instincts rather than using detailed design information.

At a more general level, there did not appear to be any particularly well developed mechanism for feeding back information on airtightness performance, nor was it clear how the design and construction lessons were being absorbed for use in making improvements to processes or actual designs. To a large extent this is linked with the need for a clearly defined quality control process, for without such a process there can be no definition of problems, identification of their causes or framing of solutions.