[Back]

C. Tsiopoulou:
"Calibrated Sky Luminance Maps for Daylight Simulation";
Supervisor: A. Mahdavi; Continuing Education Center, 2006; final examination: 2006-06-19.

Building design and control applications can benefit from daylight simulation. Sky models help to model the sky conditions and predict the availability of daylight in indoor environments. Sky luminance is changing according to the weather, the season of the year and the time of the day, therefore it is difficult to create an accurate sky model. The simplified models that are currently used for computational simulation do not take into account these constant changes. It is important to test if there is the possibility of creating a sky model that approaches the characteristics of real sky and provides the architects with more precise daylight predictions. As past research has demonstrated, relatively low-cost sky luminance mapping via digital imaging can provide an alternative to highly sophisticated sky scanners and support the provision of information on sky luminance distribution patterns on a more pervasive basis. The aim of this research is first to explore the potential of deriving sky luminance distribution maps based on digital imaging and then to test their efficiency for the prediction of indoor daylight. A comparison is made between the predictions based on existing sky models (CIE Standard Skies and Perez All-weather sky) and the camera-based sky model. Thus, the effects of the selection of the sky model on indoor daylighting prediction are explored. A set of measurements were performed at the roof of the TU Vienna in order to obtain the necessary data. The horizontal illuminance levels due to 12 sky sectors were measured with the help of a sky monitoring device. A scale (1:5) model of an architectural space was used to measure the indoor illuminance values with the help of three sensors. At the same time, images of the sky were obtained with the help of a digital camera with a fish-eye converter. Luminance values were derived from the images and four calibration methods were used to generate accurate sky luminance distribution maps. These variously calibrated luminance values were then compared with the corresponding photometric measurements. Finally, the application of a digitally derived sky model based on the best calibration method was compared with the other two sky models toward the prediction of indoor illuminance levels using the case of the scale model. The results demonstrated that the camera-based sky model was more reliable than the other two sky models. It was concluded that digital imaging combined with parallel photometric calibration can provide a valuable means for a real-time generation of sky luminance maps. Detailed sky luminance models can be generated and their application can increase the predictive accuracy of the computational daylight prediction tools. Moreover, the reliability of daylight simulation can be increased toward supporting the design process and the operation of daylighting systems in buildings.

Daylight simulation; sky luminance mapping; digital imaging; sky models.

http://publik.tuwien.ac.at/files/pub-ar_7966.pdf