Abstract The focus of this paper is to investigate the impact of using estimated solar irradiance data in building energy simulations. A series of annual weather files were developed with the estimates from two global and three diffuse/direct irradiance models for three locations in Australia. A weather file was also developed using the modelled values of global, diffuse and direct irradiance as estimated by EnergyPlus' (Version 7.0.0) auxiliary weather generator program. In addition weather files were generated with measured irradiance data and satellite derived irradiance data from the Australian Bureau of Meteorology. The influence of irradiance data was tested via the simulation of a simple 15 × 15 × 3.5 m building under two construction scenarios with varying window to wall ratios (WWR). This study indicates that the level of bias and uncertainty in the simulation results was low when global irradiance was measured and only diffuse and direct irradiance were estimated. However, when global irradiance was unknown the level of bias and uncertainty increased significantly and was shown to be highly dependent on the WWR, whilst only the level of bias was shown to be dependent on the building envelope construction. The results also indicated that the Skartveit diffuse/direct irradiance model used in combination with the BOM's satellite derived global irradiance data set achieved better results than the complete (global and direct) satellite derived irradiance data set from the BOM. The worst correlation in the simulation results occurred for the case where the irradiance data was estimated by EnergyPlus' auxiliary weather generator program.