Comparison of methodologies for Test Reference Year (TRY) generation for Mediterranean sites

The improvement of the energy efficiency of buildings calls for the availability of tools for evaluating and simulating the thermal behaviour of buildings. These simulation tools generally need a complete input data set and, in particular, information referring to the climatic conditions of the site where buildings are built-up. Among climatic issues, particularly important for the thermal energy balance is the solar radiation. However, this huge mass of data is not very useful in this form, so simple aggregation methods have been developed in order to run these simulation computer programs: the test reference year (TRY) is one of the most used methods. In this paper the comparison of two methodologies for compiling TRY of the solar radiation is presented. The two methodologies, elaborated respectively by Dogniaux and Finkelstein-Schafer, have both a statistical approach to the problem. The first one is based on the normal distribution of the monthly average of each climatic parameter and his variance, while the second one is a more powerful version of the Kolmogorov-Smirnov test, based on the cumulative probability function. The two methods give quite different results, suggesting a particular caution in the choice of the method for generating such reduced sets of data.