Uncertainty of the predictions of different programs and modelling teams based on a detailed empirical validation dataset

Abstract

The results obtained from programs predicting the thermal-energetic performance of buildings depend on the assumptions made when parameterizing the constructed models. This is because some parameters are not known or are subject to uncertainty. Common simulation programs are required to undergo rigorous quality control and validation. These involve analysis of predictions against analytical test cases, inter-model comparisons such as BESTEST (ASHRAE. 2020. ASHRAE 140-2020 Standard Method of Test for the Evaluation of Building Energy Analysis Computer Programs) and empirical validations against measurement data. This paper uses an empirical validation data set and corresponding modelling predictions from IEA EBC Annex 71 to analyse the uncertainty of different programs by means of a one-step-at-a-time sensitivity analysis. Since the results show a strong influence from the various sub-model calculations of the air flows, an additional sensitivity analysis is added regarding the air exchange through the envelope and between the zones .