| Keywords |
Theoretical method ; Illumiance Sky conditions ; Photosensor ; Daylight ; Percent difference ; Correlation |
| Abstract |
A theoretical method for the prediction of illuminance under diverse sky conditions was proposed in order to examine the variation of daylight illuminance at photosensors. The method was used to predict the illuminance of photosensor with a full-shielding condition. The prediction results were validated using the results of field measurement under various daylight conditions. Results imply that the predicted values were lower than those of measurement under clear and partly-cloudy sky, but the predicted values were greater under overcast sky. Percent differences between predicted and measured values decreased as the distant between window and photosensor positions increased. The coefficient of determination between predicted and measured global illuminance was greater than 0.9912 under all sky conditions. The correlation between predicted and measured photosensor illuminance was also strong under clear and partly-cloudy sky, but the correlation was weak under overcast sky. Statistical test results showed that the linear regression models between predicted and measured values were acceptable with a significance level of 0.01. |