Simulating Daylight Propagation through Complex Fenestration Systems in a Urban Context Using Variable Sampling Subdivision Scheme

The use of Complex Fenestration Systems (CFS) within buildings can contribute to a significant reduction of cooling loads by the way of two assets: the redirection of daylight and the shading of sunrays. In order to back-up these features, daylight performance evaluations need to be performed individually in each case according to the geographical location, outdoor conditions, building orientation and design as well as indoor material properties. The use of computer simulations can make this assessment easier than on-site CFS monitoring requiring the availability of testing facilities, materials and equipment transportation as well as the physical installation of a CFS. Computer simulations can be carried-out using the RADIANCE lighting software, which allows the estimation of the daylight propagation through the CFS using monitored light transmission properties. Those features named BTDF data (Bi-directional Transmission Distribution Function) are assessed using a bidirectional gonio-photometer and stored in an internationally standardized format. In order to perform such simulations, BTDF data are assigned to a planar polygon that models the CFS and/or a window becoming a secondary light source in the virtual model. The accuracy of the simulation results is generally relying on the computer simulation parameters that drive the lighting calculation in the virtual model, as well as the BTDF data resolution assigned to the polygon mimicking the CFS. In an urban context moreover the shadowing effects due to adjacent buildings have a significant impact on the incoming daylight flux transmitted by the CFS. It is important accordingly, to subdivide such a polygon - which may be a large glazed area - in an optimal way in order to enhance the simulation accuracy. Computer simulations were carried-out for that purpose using a virtual model of an office room placed in an urban context. The BTDF data of a Laser-Cut Panel (LCP), made of a 6 mm thick acrylic panel with 4 mm spaced parallel laser cuts, were applied for that purpose to: i) a single pane representing a full-size office window and ii) subdivided polygons of the size of the LCP sample benefitting from BTDF monitoring. All results are compared in order to determine the influence of the polygon subdivisions referenced to the BTDF data on the final computer simulation accuracy.