Uncertainty in daylight calculations

Abstract

There is uncertainty in every daylighting calculation, an inherent and unavoidable imprecision which exists in addition to any error due to inaccuracy in a numerical model. It occurs because neither the future state of a building nor the future luminance pattern of the sky can be predicted with certainty. This paper examines the propagation of error in a daylight coefficient calculation and concludes that the overall effect of uncertainty in the various parameters is significant and must be taken into account in practice. The theory of daylight coefficients is extended. The daylight factor under a uniform sky is shown to be equal to twice the mean of the daylight coefficients, making it possible to compare criteria used in climate-based simulation with older daylighting standards. Simple examples are used as illustrations, with theoretical working and numerical experimentation. Guidelines are given for good practice.