A method to quantify uncertainties in airtightness measurements: Zero-flow and envelope pressure

•Multi-level model should be used to estimate uncertainties in zero-flow pressure.•Zero-flow pressure approximation leads to substantial uncertainties.•Uncertainties can be reduced by increasing the measurement period for zero-flow pressure.•Uncertainties in zero-flow pressure lead to a change in regression method. In airtightness measurements using the fan pressurization method, the calculation of uncertainties does not take into account the zero-flow pressure approximation. It was recently suggested that neglecting the uncertainties related to zero-flow pressure approximation could be an unrealistic hypothesis. In this study, a method for quantifying this source of uncertainty is proposed, illustrated and discussed. The method is applied to a series of 31 zero-flow pressure tests performed on a newly-constructed apartment within a period of 15 days in Brussels, Belgium. For each test, 32 different zero-flow pressure approximations were compared. Since the data had a nested structure, multiple multi-level models were used for their analysis. The results show that, for the tested building, the zero-flow pressure approximation uncertainty is 0.45, 0.91 and 1.52 Pa respectively under low-, medium- and large-wind conditions. These uncertainties can be reduced to 0.42, 0.80 and 1.39 Pa when using longer zero-flow pressure measurement periods. As a comparison, uncertainty in pressure measurement at 50 Pa due to the equipment is 0.25 Pa. The uncertainty of zero-flow approximation makes the envelope pressure uncertainty non-negligible, therefore having an impact on the regression technique used to determine the building airtightness.