Deterministic and Bayesian approach to u-value measurements

In the framework of reducing building carbon emissions, more efficient building systems are designed to reduce heating and cooling loads in domestic premises. Extensive measurements are essential to evaluate the thermal qualities of the single building systems as well as monitoring the performance “as built” pre and post retrofit. A series of experimental measurements have been carried out in a controlled environment with the aim of optimizing the analysis process in terms of time and reliability of the results. The data have been examined with standard and innovative techniques, based on Bayesian statistical analysis. Initial investigation shows that the combination of accurate data collection and the employment of mathematical statistical tools in the analysis can produce reliable results also under non-steady conditions. Further advantages are the simultaneous estimation of a statistical error on the values obtained.