Evidence for significant heat losses through party wall cavities in load-bearing masonry construction Short Title: Significant Heat Loss via Party Wall Cavities

This paper presents empirical evidence and analysis suggesting significant heat loss from air movement through cavities in party walls in masonry construction. The measurements have been made during the course of coheating tests undertaken as part of the Stamford Brook housing field trial. Direct measurements show the additional heat loss may amount to up to 30% of total heat loss in the dwellings tested, potentially making it the largest single contributor to heat loss in terraced dwellings built to the 2006 revision of the Building Regulations. The phenomenon of convective bypassing associated with masonry party walls was identified in the late 1970s in the course of the Twin Rivers Project, albeit in a somewhat different construction from that used at Stamford Brook. A similar phenomenon was reported by Siviour in a Address for correspondence: R Lowe, The Bartlett School of Graduate Studies, 1-19 Torrington Place, University College London, Gower Street, London, WC1E 6BT, UK. Telephone: +44(0)20-76795916, E-mail: [email protected] Pre-refereeing version July 2006. Revised version subsequently published in Building Services Engineering Research and Technology 28 2 (2007), 161-181. 2 the UK in the mid 1990s, but it appears that no action was taken either to confirm his results, to develop simple technical fixes to eliminate this large additional heat loss mechanism, or to amend standards for calculating heat losses from buildings. There are no references to heat loss associated with party walls in current conventions for heat loss calculation, and we have found no other recent literature on this subject. Practical Application The heat bypass mechanism described in this paper is believed by the authors to contribute to a significant proportion of heat loss from buildings in the UK constructed with clear cavities such as those found in separating walls between cavity masonry dwellings. It is proposed that relatively simple design changes could be undertaken to eliminate such heat loss pathways from new buildings. In addition, simple and cost effective measures are envisaged that could be used to minimise or eliminate the phenomenon from existing buildings. Such an approach could give rise to a significant reduction in carbon emissions from UK housing. Pre-refereeing version July 2006. Revised version subsequently published in Building Services Engineering Research and Technology 28 2 (2007), 161-181. 3 List of Symbols Aroof plan area of roof (m) cf dwelling fabric heat loss coefficient (WK) cp heat capacity air at constant pressure ≈1000 JkgK cv dwelling ventilation heat loss coefficient (WK) d dwelling plan depth (m) Dh hydraulic diameter (m) f friction factor g acceleration due to gravity (ms) H height of cavity (m) hloft floor thermal conductance of loft floor (WmK) hroof covering thermal conductance of roof covering (WmK) leave length of eaves (m) lparty wall length of party walls (m) Lbypass loft bypass heat transfer coefficient (WK) Lloft edges loft edge heat transfer coefficient (WK) Lloft floor loft floor heat transfer coefficient (WK) Lroof covering roof covering heat transfer coefficient (WK) n background ventilation rate (air changes h) Q daily mean heating power (W) Q’ corrected daily mean heating power (W) q50 dwelling air permeability (mh @ 50 Pa) R effective solar aperture (m) Re Reynolds number S solar insolation (Wm) Tcavity party wall cavity temperature (°C) Tin internal temperature (°C) Tloft loft temperature (°C) Tout external temperature (°C) Ucavity-loft conductance from party wall cavity into loft (WmK) Pre-refereeing version July 2006. Revised version subsequently published in Building Services Engineering Research and Technology 28 2 (2007), 161-181. 4 Uroof effective effective U-value of roof (WmK) Uroof notional notional U-value of roof (WmK) v upward speed of air in party wall cavity (ms) wcavity width of party wall cavity (m) ΔPf pressure drop due to friction in party wall cavity (Pa) ΔPstack stack pressure difference (Pa) ΔT inside-outside temperature difference (K) ε absolute roughness of party wall cavity (m) ρ density of air ≈ 1.2 kgm Ψeave linear thermal transmittance eaves (WmK) Ψparty wall linear thermal transmittance party wall (WmK) Pre-refereeing version July 2006. Revised version subsequently published in Building Services Engineering Research and Technology 28 2 (2007), 161-181. 5