Experimental investigation of the impact of model complexity on indoor daylight spectral simulations

Daylight spectral simulation is crucial for designing functional, healthy spaces and predicting light interactions. It is essential for accurate non-image-forming effects of light calculations. This study addresses the knowledge gap in reproducing indoor daylight spectral conditions in the built environment. Using varying levels of geometry (LOG) and information (LOI), simulation accuracy was assessed by comparing it with experimental data from two offices over three days with cloudy and clear sky conditions. The lowest accuracy was found with high LOI and low LOG simulations. For the highest accuracy, specific material spectral properties are needed, while spectrally-neutral materials at low LOG produced comparable results. Simulations near and facing windows were the most accurate. The study concludes that to reproduce indoor daylight spectral conditions, modelling should use either the lowest or highest geometry and information complexity, depending on available modelling time and required accuracy.