Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page
Ask Your Question

Daylighting results inconsistencies

asked 2016-05-13 17:55:18 -0600

Determinant gravatar image

updated 2016-11-12 09:05:21 -0600

I have set up 6 different simulations using Radiance. See the pic.

image description

image description

Cases 2 & 3 have the same width-to-spacing ratio and the same angle and the geometry is such that it blocks direct beam between 8am and 4pm year round (including DST). Same for cases 4 & 5. Yet, the results between 2 &3 differ significantly. Same between 4 & 5 . I can't understand from a physics standpoint why they are so different (even accounting for shadowing the ceiling and workplane). Why are shorter slats resulting in less daylighting?

I consulted a daylighting guy and he said they should all be equivalent (within the accuracy of the model). He suggested changing some Radiance settings and I did but the trend didn't change. Smaller, more numerous louvers still results in less daylighting.

for Case 6, I used the OpenStudio Blinds BSDF and these results are more consistent with what I'd expect. So, not only do the results seem self-INconsistent, but they seem to be 2-4 times less than what the BSDF predicts.

Any ideas of what physical concepts I might be misunderstanding or what inputs I might need to change here?

edit retag flag offensive close merge delete

2 Answers

Sort by ยป oldest newest most voted

answered 2016-05-14 02:57:26 -0600

Diffuse transmission through small and crowded geometrical structures is one of the fundamental weaknesses of a backwards raytracer. You would need to massively crank up -ad, -as, and probably a few other parameters to see more light coming through those narrow and deep gaps in your model. The results might still not really be satisfactory, though.

This methodical difficulty is the reason why the genBSDF program was created, which I assume is the tool behind what you call "OpenStudio Blinds BSDF". So it's no surprise that you're getting more reasonable results that way.

The "internal inconsistency" is easy to explain once you recognize the problem with the narrow and deep gaps. The two scenarios would probably converge a lot better (ie. at similarly bad values) if you were to add another slat at the top of the windows. Right now, both models have a comparatively unobstructed zone there, where daylight will reach the ceiling after a single diffuse bounce off the top slat. Just that in one case, that zone is three times as large as in the other.

edit flag offensive delete link more

answered 2016-05-16 09:43:46 -0600

updated 2016-05-16 10:45:42 -0600

What Schorsch said. Just to add some OpenStudio-specific responses, we added a few BSDFs to OpenStudio to model blinds (and shades, and daylight redirecting devices) better for annual simulations than you could get in a reasonable timeframe using actual geometry. You're stuck with the ones we provide if you stick to the OpenStudio measure workflow, however. See this post and the ones it references for more details on the OpenStudio-Radiance-BSDF connections.

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Training Workshops

Question Tools

1 follower


Asked: 2016-05-13 17:55:18 -0600

Seen: 298 times

Last updated: May 16 '16