Airtightness predictive model from measured data of residential buildings in Spain

The need for airtightness control is a reality given its impact on buildings’ energy use and IAQ. For the past few years, this fact has resulted in energy performance regulations being established in many countries in Europe and North America. However, compliance proof is not always required, and on-site testing is often avoided. In this sense, predictive models have become useful in the decision-making process and to estimate input values in energy performance simulation tools. In Spain, maximum envelope permeability values were introduced recently, but pressurization tests rarely undergo. The most common approach to prove compliance is by means of reference values, which were proved to be inaccurate. This paper presents a predictive model for airtightness, which offers an alternative procedure for airtightness estimation. The model was developed from an airtightness database which included a representative sample of the residential building stock in Spain. A General Linear Model was considered to assess significant variables related to the climate zone, the age of the building, typology, building state, construction system, and dimensions. As a result, a predictive model that explains 42.9% of the variability of the response is presented, containing 12 main effects and 2 interactions. Overall, even if some limitations were identified, the relevance of the model proposed is warranted from the statistical point of view by the significance of the coefficients and the validity of its residual analysis.