Clustering of European climates and representative climate identification for building energy simulation analyses

This research presents a methodology to cluster weather climates with the purpose of generating homogeneous and distinct groups suitable for the selection of representative cities, to use as reference for robust generalization of building simulation findings. The proposed methodology is based on Hierarchical Clustering and Kolmogorov-Smirnov tests, applied to dry bulb temperature, water vapour partial pressure and solar irradiation data. As an example, a dataset with more than 300 European climates is analysed, a new classification is prepared and compared with the well-known Köppen-Geiger system. Results on this paper provide preliminary indications of the large potential of this approach to identify representative cities for Building Energy Simulation analyses.