Use of a numerical weather prediction model as a meteorological source for building thermal simulations

Abstract

[Abstract]: Thermal simulations are a commonly used tool for energy efficiency analysis of buildings. Regional meteorological station networks are a prime source of weather data inputs, required for building thermal simulations. However, local measurements from weather stations are not always available, and when available, access to data may be expensive. This paper analysed a novel use of a numerical weather prediction mesoscale model, the Global Forecast System (GFS) sflux model, as a source of input data for transient thermal simulations. Two interpolation techniques (nearest neighbour and universal kriging) were used to generate local weather datasets from GFS outputs at 27 locations spread over an area of 29,574 km2 in Galicia (northwest Spain). The performance of the GFS estimations was tested against weather measurements obtained from a government weather agency. A representative building with the most common features was selected for running thermal simulations in the TRNSYS environment, focused on heating demands, with estimated weather data as the input. The results highlighted that GFS-interpolated datasets consistently performed better than using measured data from the nearest weather station. GFS was found to be an appropriate weather source for building simulations and was able to provide good-quality, free and global-scale local weather inputs.Xunta de Galicia; TOPACIO IN852A 2018/37