New typical meteorological year generation method based on long-term building energy simulations

Outdoor climate data are needed to predict building energy use. Since the 1980s, Typical meteorological Years (TMY) have been developed based on actual observed, typically 30-year-long, meteorological data. EN 15927-4 uses Finkelstein-Schafer statistics to construct a TMY. To address varying climate parameters' impact on energy use, EN 15927-4 method requires supplementary weighting factors that correlate with the climate parameters' impact. However, there is no straightforward method to select these weighting factors, and often, building performance simulations with reference buildings are conducted. To eliminate weighting factor determination uncertainty, this study proposes a TMY-generating method fully based on long-term energy simulations. As the main finding, it was shown that using building energy simulation results instead of climate parameters to determine the best months for TMY corresponded better with the average annual energy needs of the 31-year period. In principle, the developed method has no geographical/climatological limits, but it was tested with the Estonian cold and dry climate and typical reference buildings. The best statistical method for TMY generation was mean square error normalized by average annual heating/cooling need. It reflected long-term global warming and has low and consistent deviation values from 31-years results averaging less than 1% for heating and 4% for cooling compared to EN 15927-4 with 5% for heating and 7% for cooling and EN 15927-4 with weighting factors with 4% for heating and cooling. The developed TMY generation method has wide application potential, and it can be recommended to be validated in other climates in the future.