Three different Sun Vectors: Geco, Heliotrope, Diva

Replies to This Discussion

Permalink Reply by Andrew Heumann on March 29, 2012 at 10:09pm

I can't vouch for the others, but I have found DIVA's solar vector to be consistent with the NOAA algorithm as implemented by Ted Ngai. 

Permalink Reply by Michael Thomsen on September 7, 2012 at 1:10pm

Hi Arthur

I have choosen DIVA. 

Specialist in energy-buildings from my university use Radiance, which DIVA is based upon. They advise against Ecotect, which GECO is based on. I don't know about Heliotrope. Ecotect is preferred by many architects students because they get free license, and because it's compatible with lots of other free software from Autodesk. 

Read about Ecotect vs radiance from MIT here --> http://web.mit.edu/tito_/www/Publications/BS09_DaylaightingNovices.pdf

Permalink Reply by Ángel Linares on September 7, 2012 at 1:37pm

Hi michael...the link you posted is broken :(

Interesting conversation :) Interesting to see the same error (or success) in Heliotrope and Geco...mmm...

Permalink Reply by Sameer Kumar on September 7, 2012 at 1:43pm

There is a typo in Michael's link. This one works:

http://web.mit.edu/tito_/www/Publications/BS09_DaylightingNovices.pdf

Permalink Reply by Ángel Linares on September 7, 2012 at 2:24pm

Thanks Sameer :)

Now I've read the paper I could say that this error exposed by Arthur is not about "split-flux method" used in Ecotect (it's used for daylight factor for example). To calculate sun position Ecotect uses other algorithms...

However, I extract two conclusions reading this paper.

1.- Harvard students don't know how to modelling proper analysis 3d models :P

2.- "Furthermore individual ECOTECT models both grossly over- and under predicted daylight factor levels according to RADIANCE. This finding suggests that ECOTECT-based daylight factor predictions  cannot be considered  to be  worst case assumptions and that RADIANCE should always be used instead of the build-in ECOTECT daylighting engine."

Permalink Reply by Piotr on September 7, 2012 at 1:45pm

Maybe the difference is in the Daylight Daving Time?

Permalink Reply by Luis Fraguada on September 7, 2012 at 3:09pm

Arthur,

That is both fascinating and worrisome.  Thanks for sharing.

-As a sidenote...I do enjoy the way Heliotrope works.  Looking forward to trying out the new version.

Permalink Reply by to] on September 14, 2012 at 8:40am

Hi,

very intresting discussion and great comparison. Thanks to Arthur to coming up with that. So coming back to Geco it is based on the implemented calculation within ecotect. The algorithm used in Ecotect to determine the position of the Sun from any location on Earth is outlined below. The azimuth and altitude of the Sun are calculated using formulae first proposed by Spencer (Spencer, 1965), then refined by Szokolay (Szokolay, 1996). Values for solar declination and the equation of time are determined using formulae proposed by Carruthers, et al [1990]. For mor inforamtion on that topic you can check http://wiki.naturalfrequency.com/wiki/Solar_Position_Calculator.

hope we can get on with this topic and we will add addiotional things in the future.

 

Permalink Reply by to] on September 14, 2012 at 10:22am

After comparing Heliotrope, Ted Ngai's definition with the implementation of Noaa algorithm and Geco(Ecotect) i come to following result which should be further evaluated. Ted's result and Geco are absolute equal but Heliotrope has a slightly variation. So according to Michael Thompson this has nothing todo with Radiance which is a Raytracing-Software and of course is indeed much more accurate in case of Light analysis but does not deliver Sun position data. So it seems that the two algorithms come to the same result which is promising.

 

 

Permalink Reply by to] on September 14, 2012 at 12:15pm

Hi Brian,

thanks for the reply.... and contribuitng to this discusson

Geco(gets information from ecotect) and Noaa algorithm are different but deliver the same result. I am curious about the Diva algorithm and what they use to calculate the position....