Welcome to ideas.sasaki.com! This is where Sasaki Associates shares ideas and inspirations.
Visit www.sasaki.com to learn more about our company and our projects.
explore by topic:
Get a Weekly email with a list of new posts!
-
Anonymous
Ecotect Daylight Reality Check
By Meredith Elbaum posted on Fri, 2009-08-07 09:38 , 608 reads, 1 commentComputer simulation models can be very valuable tools but how accurate are they? Sasaki uses Ecotect and Radiance for daylight analysis and wanted to make sure the data we were looking at was accurate. We took light level readings in an enclosed room in our office and then modeled the same space in Ecotect and compared the two. We concluded that the Ecotect model seems to provide an acceptable first order estimate of interior light levels due to daylight. The full report is included here.
About ideas.sasaki.com
Sasaki Associates is an interdisciplinary firm of 300 professionals. Our founder, Hideo Sasaki, instilled the firm with a passion for critical thinking and open inquiry. In this spirit we created ideas.sasaki.com, to share our ideas and process with our colleagues, clients, and anyone else interested in our work. We hope you are inspired to engage us in further conversations about the topics you see here. Email us at ideas @ sasaki.com for more info
Tags
Hi Meredith,
I agree with your conclusion that the sky model is likely to be a major factor in the discrepancy. The CIE Overcast sky model used in Ecotect assumes that the sky is about three times as bright at the zenith as at the horizon, no matter what time of day it is, when in reality an overcast sky is brighter near where the sun is at that moment. This leads to various inaccuracies in luminance from different parts of the sky model. The intermediate sky is basically a mathematical average of the sunny and overcast skies, so it still contains this overcast sky inaccuracy. I believe that the CIE clear sky is fairly close to a real clear sky, so this would have been a better sky model to use for the experiment, but you'd have to wait for a perfectly clear day to compare to.
Did you take into account any obstacles outside the window, the ground reflectance, or any other reflective surfaces? Another possible source of error is material reflectances, if these were based on guesses and not measurements. There are a lot of papers out there validating Radiance as the most physically accurate lighting simulation engine, with very good correlation to measured values, but the problem is that there are many complex inputs that have to be correct in order to achieve the high accuracy that is possible.
Rosie Osser
Post new comment