Surface heat transfer coefficients in building envelopes: Uncertainty levels in experimental methods

Abstract

There are several research methods for the on-site assessment of U-values that aim to avoid the use of surface heat flux measurements and rely on tabulated or empirically developed correlations to define this parameter. This works performs a detailed process to estimate indoor surface heat transfer coefficients based on several experimental campaigns over building walls. Data is filtered out to remove periods with large temperature variations and/or unstable convective conditions due to HVAC. A statistical analysis is conducted, and the outcomes used to test the validity of U-value estimation approaches. The outcomes show that the actual surface heat transfer coefficients are in the range of reference works, but variations in the range of up to 2 W/m² K are found. Uncertainty levels associated to the estimation of surface heat transfer coefficient are in the range 60% for instantaneous values while this is reduced down to 12%–20% for 8-h averages. Variations and uncertainty levels are higher for low temperature gradient situations, which are considered to be very likely for modern insulation levels. It is concluded that methods seeking to avoid the use of surface heat flux measurements need to develop much deeper knowledge in this field to gain accuracy and reliability.