New techniques for validating building energy simulation programs

An Anglo-French collaborative project is presently under way with the dual purpose : a) to collect high quality data sets for the empirical validation of thermal simulation programs of buildings and b) to develop and test advanced data analysis techniques. The aim is to identify and remedy the causes of poor predictive -performance in thermal programs of buildings. Two data analyses are presented in this paper: one is a time-domain technique developed by the Energy Monitoring Company in UK, and the other is a frequency-domain technique developed by the University Paris 12, France. Both techniques have been applied to empirical validation results from four state-of-the-art thermal programs, and the derived error diagnoses are discussed.