IN-24-C043 - Predicting Discomfort Glare Using Simple Illuminance Measurements – An Experimental Evaluation

Predicting visual comfort has always been a challenging task, and it typically requires a calibrated High Dynamic Range (HDR) imaging process for a precise assessment. Visual comfort assessment is particularly needed in the case of smart window shading design to be able to optimize the comfort perception of occupants. This study proposes an experimental evaluation of a concept that relies only on convenient, non-intrusive indoor and outdoor illuminance measurements, along with sun positions (azimuth angle and elevation angle), to predict discomfort glare. The experiment was carried out by recording the aforementioned illuminance measurements along with HDR images using a calibrated camera to derive Daylight Glare Probability (DGP), as the metric of discomfort glare, over a five-week period in an unoccupied experimental laboratory. Then, justified by its interoperability and relative simplicity, a Random Forest model was developed to predict the DGP from the illuminance measurements with an accuracy of 0.92. Overall, this study demonstrates that glare assessment is possible without a convoluted network of sensors or equipment, and thus the techniques or similar techniques used in this study can potentially improve the practicality of relevant applications such as smart window shading control for visual comfort.