The simulation of annual daylight illuminance distributions — a state-of-the-art comparison of six RADIANCE-based methods

Abstract

The present work discusses simulation results of annual indoor illuminance distributions for two office geometries situated in Freiburg, Germany, calculated with six different RADIANCE-based daylight simulation methods. These methods are the ubiquitous daylight factor method [P.J. Littlefair, Predicting annual lighting use in daylit buildings, Building and Environment 25 (1990) 43–54.], ADELINE 2.0 [M. Szerman, J. Stoffel, ADELINE 2.0, Radlink Technical Manual, IEA Solar Heating and Cooling, Task 12.], the classified weather data according to Herkel and Pasquay [S. Herkel, T. Pasquay, Dynamic link of light and thermal simulation: on the way to integrated planing tools, 5th Int. IBPSA Conf., Prague, Sept., 8–10, 1997, Vol. II, pp. 307–312.] and two simulation procedures based on daylight coefficients according to Tregenza, namely ESP-r version 9 series [J.A. Clarke, M. Janak, Simulating the thermal effects of daylight-controlled lighting, Building Performance (BEPAC), (1) (Spring 1998).] and a new accelerated method developed by the authors. The new method calculates 145 diffuse and three ground daylight coefficients in a single raytracing run which considerably reduces the required calculation times for an annual daylight simulation. An explicit calculation of the indoor illuminances under all 4703 annual hourly mean sky luminance distributions from the Freiburg test reference year (TRY) serves as a reference case against which the other methods are tested. The simulation results reveal that the accuracy of an annual daylight simulation method is not necessarily coupled with the required simulation time. The quality of an annual simulation rather depends on the underlying sky luminous efficacy model and whether the method considers the hourly mean direct and diffuse illuminances for each time period explicitly. The two methods relying on daylight coefficients exhibit the lowest relative root mean square errors (RMSEs) for the straightforward office geometry. The results for the advanced office show that internal illuminance contributions due to external ground reflections are only considered by the new method.