[Bldg-sim] Radiance Materials database

Lars O. Grobe grobe at gmx.net
Thu Sep 24 02:30:02 PDT 2015


Hi Christopher,

one rather simple way to estimate reflectance of mostly diffusely reflecting surfaces is to use a gray-chart. The ColorChecker by Gretag/MacBeth is rather popular for such applications. Often you do not need color, and I would not worry too much if you are off by few percents here.

Specular reflection is a game changer, it is rather hard to estimate and introduced the question about the exact characteristics of the specular component. Radiance uses a Gaussian representation for the specular reflection, with an additional roughness parameter to describe the "sharpness" of the reflection. Imagine the difference between a polished aluminum surface and one with some roughness, they will both mostly reflect to the specular direction but highlights will be spread by the roughness of the latter.

Glass obviously is crucial for your simulations, and before thinking about measurements I would have a close look at the International Glazing Database. I would not guess when it comes to glazing.

The case of translucent panels is probably one of the most difficult ones. For a meaningful analysis, you need to know not only the typically given direct/specular transmission/reflection, but the angular dependence of both (the scientific term would be the Bidirectional Scattering Distribution Function BSDF). There are very different "white translucent" materials, which may share the same direct/specular transmission and reflection (as found by direct-hemispherical measurements using e.g. an integrating sphere), but lead to very different effects and require different models. E.g. it makes a difference wether scatter is caused by surface roughness, or by inclusions in the volume of a transparent solid. The latter may have perfectly smooth surfaces and show sharp, mirror-like reflections, while the surface roughness would effect both reflection and transmission. Radiance provides the trans material class, but this comes with a lot of assumptions...

Besides it may also have the most significant impact, depending on the height of the partitions. On the one hand, the scattering properties affect how much light is transmitted to a direction where it is accounted for in work-desk illuminance. If the scattering effect is very strong, only a small fraction will pass through towards a desk behind such a partition, and quite a lot of light will be scattered towards directions where it is kind of lost for your "useful" illuminance on the work-desks behind (a perfect diffusor would equally scatter light to all directions, so to the ceiling, floor, ..). What may even be more crucial is the fact that the scattering properties may lead to glare, if the partition is hit by direct sun-light. Light that would otherwise reach a work-desk or floor may be scattered in a way that a significant portions goes straight towards the eyes of occupants - that is why materials with a strong scattering effect are rather difficult to handle if they may be exposed to direct sun-light.

For translucent panels I would thus recommend going with measured BSDF data. Recent versions of Radiance allow you to use the data-driven BSDF material class, which is directly based on measurements and does not rely on all the assumptions inherent in the trans model. There are laboratories offering such measurements dedicated to applications in daylight simulation (LBL in the US, Pab-opto in Germany, my employer Lucerne University of Applied Sciences and Arts in Switzerland), and databases are created to make more data accessible. You may contact the supplier of the product to provide you with data, and if not available, hint them at those labs. I think that it is important to rise awareness for the importance of realistic models. Too often a lot of effort is spend on running complex simulations and analysing results derived from models based on nothing but guessing.

Do not spend time on textures if you are looking for illuminance, they have virtually no impact and take your valuable time. And do not estimate from photographs. Cameras have a lot of image-processing built in, the pixel values have little to do with reflectance.

Cheers, Lars.

> I am looking for some example materials I can use in a Radiance simulation for daylighting calculations for LEED v4.  The project is your typical open office concept with half-height partitions between work stations.
>  
> The partitions are a white, translucent material.  The workstations are a wood grain laminate, the walls are flat off-white, the carpet tiles are a blue-gray color.  I can get the colors from photographs of the materials but I am not sure how to estimate the reflectance.  Also, I am not sure of how to handle the texture of the carpet tiles. 
>  
> Thanks for any assistance!
>  
> <image001.png>
> Christopher Jones, P.Eng. 
> Senior Engineer

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.onebuilding.org/pipermail/bldg-sim-onebuilding.org/attachments/20150924/0a7ce6df/attachment-0001.html>


More information about the Bldg-sim mailing list