by Trevor Rushton, Director
Article published in RICS Journal. Click here to read the full article.
The risk of condensation occurring in flat, unventilated roof constructions is nothing new. Early attempts at insulating flat roofs brought about significant problems in terms of decay and failure, particularly during the early 1970s as efforts were made to improve the thermal performance of buildings. The advent of warm and inverted roof construction – in which, respectively, insulation is placed directly below the roof membrane or on top of it – improved performance significantly, although with the latter there was much early debate on the wisdom of concealing the membrane below hard finishes as this made repairs or diagnosis of leaks difficult.
Provided that a sound vapour barrier is included at deck level, the risk of interstitial condensation in a warm roof is much reduced. But the roof covering will absorb heat as a result of infrared radiation, and because the insulation prevents or reduces the dissipation of that heat back to the structure, the roofing membrane will have to work hard to accommodate the potential range of temperatures involved. Indeed, this was the downfall of many asphalt warm roofs until matters were improved using polymer-modified asphalt.
Today, warm and inverted construction is commonplace; the problems of condensation are understood and, rather like measles, they can be eliminated. Unfortunately, also like measles, cold roof construction still crops up from time to time, and when it does the results can be catastrophic.
A cold roof is one in which the insulation is placed between the roof joists or below the deck. Air and water vapour will enter the roof construction even if you try to prevent this, and as the air cools the amount of moisture that it can support is reduced, potentially leading to 100 per cent relative humidity – or saturation point – and the resultant condensation.
By ventilating the roof space, moisture can be removed before it condenses, or be allowed to dry before harm is caused. A good air- and vapour-control layer will reduce the passage of vapour into the roof; but its effectiveness will depend entirely on good-quality work and meticulous attention to detail.
Although a ventilated cold roof can theoretically be designed to avoid condensation this is a high-risk strategy, and such a design ideally needs to be avoided. Equally, an unventilated cold roof is simply asking for trouble. BS 5250: 2002 Code of practice for control of condensation in buildings provides guidance and recommendations; although a later version of this code has been produced, this version is referred to in Approved Document B to the Building Regulations.
The code itself refers to BS EN ISO 13788, which sets out design calculations for the avoidance of condensation that are known as the Glaser method. Such calculations tend to be performed by software, whether used by the roofing component manufacturer, or by the designer through engineering design programs such as Bentley Hevacomp.
Earlier versions of BS 5250 suggested that, provided the condensation dried out from one year to the next and did not exceed 350g/m2, the roof would perform satisfactorily – a somewhat rash statement that was removed from the 2011 version. Nevertheless, the code was very clear that cold deck construction should be avoided. Where this was unavoidable, there should be a 50mm ventilation gap, with cross-ventilation provided continuously. The vapour control level should be at least 250MN·s/g, which is a hopelessly low threshold. However, effective cross-ventilation is unlikely to occur with spans greater than 5m, so unless alternative measures can be contrived such a roof is likely to be intrinsically unreliable.
Bear in mind that Glaser calculations are suitable for comparing different constructions but are not an accurate prediction tool. They do not model moisture in the structure under service conditions, and are unsuitable for calculation or drying out of built-in moisture. Alternative programs are available; one, WUFI , created by the Fraunhofer Institute, enables modelling of moisture conditions in building envelopes. The software performs hygrothermal calculations on building component cross-sections considering, where appropriate, built-in moisture, solar radiation, long-wave radiation, capillary transport and summer condensation.
This program can help to consider component performance under actual climate conditions. However, the analysis is affected by many variables, and given that accurate, common and consistent data on specific building materials can be difficult to come by, one must exercise a degree of caution. It is rather like playing a violin: harmony if you hit the right notes, but cacophony if you don’t.
While the effects of condensation on non-organic materials may be marginal, its implications for timber and other hygroscopic materials can be profound. During construction one can assume that the moisture content of construction timber may be in the region of 15–18 per cent; it may be a bit higher if the building was exposed to rainfall during construction, but if constructed properly will probably settle down to an equilibrium within a few months.
The relationship between moisture content and relative humidity is important. Timber held in 80 per cent relative humidity or more for an extended period will, depending on species, reach a moisture content of 20 per cent or more. If this is allowed to continue for several months then the risk of timber decay becomes acute.
Furthermore, if the roof is unventilated the construction will take a correspondingly longer time to dry out and, in this case, the vapour resistance of the decking and the roof membrane become very relevant. Some materials such as PVC roofing are reasonably permeable, but some thermoplastic polyolefin materials and ethylene propylene diene monomer membranes have a much higher resistance – meaning that the risk of condensation becomes higher. Seemingly simple substitution of materials during construction can therefore have unexpected consequences during the service life of a building.
One would have expected that, faced with the above difficulties, a reasonably competent designer would avoid this form of construction. Unfortunately that is not necessarily the case, and over the past two to three years several contemporary buildings have failed due to extensive timber decay within five years of construction, and there are probably many more unreported cases.
In one instance, the designer was encouraged to depart from his original warm roof construction and use a prefabricated approach involving very large unventilated cold roof cassettes. Although a condensation analysis suggested the construction would perform satisfactorily, the roof structure failed within a short time, and repair costs ran to millions. Another case involved a highly insulated eco-house, where again the entire roof deck and most of the structure failed within about 24 months.
Both examples may have been compounded by exposure to rainfall during construction. It is imperative that any such excess moisture is removed because it will tend to redistribute itself around the roof owing to evaporation and the subsequent movement of vapour from high- to low-pressure areas, usually condensing on the underside of the roof deck and the top portion of the joists or engineered beams. This, coupled with a highly impermeable deck and or roof membrane, will almost certainly result in failure.
The message is simple – avoid cold flat roof construction, and if you come across it treat it with the utmost suspicion unless you have good reason to think otherwise. If the roof is unventilated, proceed with even more caution.