Future Weather Data for Dynamic Building Energy Simulations: Overview of Available Data and Presentation of Newly Derived Data for Belgium

Abstract

As buildings have a relatively long life span, it is important to consider climate change in energy performance modelling. Good quality weather data are needed to obtain accurate results. This chapter discusses widely used methods to predict future weather data (dynamical downscaling, stochastic weather generators and morphing) and provides an overview of available weather datasets (multi-year, typical years, extreme years and representative years) for building simulations. A Flemish office building is used for a comparative analysis of the estimated heating and cooling load making use of 1-year weather files (typical and extreme future climate conditions) derived from a recently developed convection-permitting climate model for Belgium. Climate models and weather generators are identified as the most preferred for the estimation of the average energy consumption and thermal comfort in average and extreme situations. Climate models have the advantage to better represent extreme weather events and climate differences due to territorial settings, while weather generators can generate multiple climate realizations. A combination of a typical year with an extreme cold and extreme warm year was found to result in an overall good representation of the energy need for heating and cooling in average and extreme weather conditions. Further, the influence of the methodological choices to extract 1-year weather files (typical or extreme years) from the 30-year climate data is highlighted as different results were obtained when different meteorological variables were considered for the creation of the 1-year files.