Heritage Windows / Stained Glass in Radiance/Honeybee

Hello Fellow Modelers,

I was just curious if anyone has experience with producing heritage stained glass / mosaic glass representations in a Honeybee daylight analysis. I am looking to get hourly illuminance data within a space that only receives natural light through stained glass.

I am starting my research by looking at Mostapha’s paper on Modeling Frit Patterns for Daylight
Simulation
but I am not entirely sure if a stained glass window can be represented the way a fritting system would be. I have rectified images of the windows to try mapping with.

I’ll update if I run into any revelations. Any insights from others would be awesome!

Thank you for your time,
Adrian

Hi @Elliot_Glassman, @drulff, @daniel.knapp I’ll share with you whatever I run into!

(In my opinion,) It does not make much sense to draw any conclusions from numerical values of illuminance for light filtered through stained glass. Illuminance, as measured empirically with an illuminance meter or simulated through Radiance, is defined for full spectrum (white or in the vicinity of white) light sources. When light passes through a stained glass it will most likely get filtered to some other color.

Anyway, Matiu Carr’s write-up from 1993 is still arguably the definitive resource for incorporating patterns into Radiance simulations. I am not sure if @mostapha got around to incorporating image mapping into Honeybee, but assuming this functionality does not exist, you might have to do some manual calculations to get the images to map to a glass polygon. I tried a random image to see if the process described in Carr’s write-up still works or not.


(Sunny sky)

(Overcast sky with a bunch of ambient calc artifacts )

The problem with the process used to create above images, and also the methodology described in Mostapha’s SIMAUD paper, is that none of it has been empirically validated. These are definitely more photorealistic than physically-based. The closest that I know of a “scientific” approach used by anyone to do this with Radiance is the work that Lars Grobe from HSLU did recently.

Regards,
Sarith