Generation of typical weather datasets for daylight simulations using the enhanced simplified Sandia method

Existing generation methods of Typical Weather Datasets (TWDs) have mainly been established for energy-related applications, overlooking the accurate representation of luminous conditions. This limitation is accompanied by the reliance on statistically aggregated weather elements, coarse temporal resolutions and subjective weighting schemes. Building upon related studies in the context of daylight simulations, this work aims at addressing these discrepancies by systematically investigating the influence of four critical factors on the representativeness of TWDs: weather elements, typical periods, aggregation methods and weighting schemes. The objective is to determine the optimal combination of factors that yields a TWD, whose sky luminance distribution closely aligns with the corresponding long-term data for a given location. The enhanced simplified Sandia method is introduced as an alternative procedure to generate TWDs, prioritizing the establishment of reference luminous conditions through improved element selection, refined temporal resolutions, robust statistical references and objective weighting schemes. Three climatically diverse locations, Cairo (Egypt), London (UK) and Key West, Florida (US), were selected, from which 36 TWDs were generated using different combinations of key factors. Besides, three readily accessible TMYx files, corresponding to the selected locations, were included for further comparative analysis. The systematic evaluation revealed that the optimal combination of factors includes the use of hourly weather data, daily temporal resolution, average aggregation method and metaheuristically optimized weighting scheme. These findings underscore the potential of advanced generation methods to enhance the accuracy of TWDs in climate-sensitive applications, thereby contributing to more reliable daylighting analyses and better-informed building performance simulations.