U-value of a warm flat roof

Condensation risk in a warm flat roof

A warm deck puts vapour-tight waterproofing on the cold side of the insulation, so moisture that reaches it has nowhere to go — which is why the vapour control layer under the insulation matters. As pre-filled, U-Monkey’s free BS EN ISO 13788 check fails it. With the roof surface taken 2 K below the air every month, the standard’s simplified method for roofs, a little condensation builds up under the waterproofing through the winter, peaking at about 7 g/m² in April, and about 3.5 g/m² of it is still there at the end of the year, so it accumulates. The waterproofing is entered at BS 5250’s figure for roofing felt laid in bitumen, sd 90 m, and a 500-gauge polythene VCL, sd 50 m, is simply less vapour-resistant than the roof above it.

The fix is the one warm-deck specifications use: a high-performance vapour control layer, such as a foil-cored bitumen VCL. Left on the generic felt, the check asks for a VCL rated at least sd 2,400 m, because the result has to hold with the waterproofing anywhere up to BS 5250’s sd 200 m for a reinforced bitumen membrane; enter your membrane’s declared value and that falls, to about 950 m for a declared 90 m. Swap the PIR for mineral wool slabs and it fails by far more with the polythene VCL. Model a cold deck instead — insulation between the joists, deck on top — and it fails outright, with or without a VCL at the ceiling; even a cold deck that clears the calculation is reported as outside its scope, because BS 5250:2021 relies on prescriptive rules for cold roofs. BS 5250 counts a warm deck with a sealed vapour control layer among the arrangements experience shows are safe, and the calculation is harder on it than experience is: its simplified roof method leaves out the sun, which dries a roof in summer. At colder or wetter stations than Manchester it asks for more, and at some none of the standard fixes clears it, so run it at your own.

It is a screening to BS EN ISO 13788 (the Glaser method), honestly labelled as one: it does not model air leakage, rain or moisture stored in materials, so for high-risk or heritage constructions a full hygrothermal simulation (BS EN 15026) is the right next step. Figures above use the app’s defaults — Manchester (Ringway) climate, humidity class 3 (dwellings with high or unknown occupancy), a 1-in-10 design year and, on a roof, the surface taken 2 K below the air for night-sky cooling — so pick your nearest weather station for your own build-up: the result can change from one station to another.