[Radiance-general] Current practice for LEED sky modelling?

[Guglielmetti, Robert]

Juicy subject matter. It's be cool if Zacks work could get incorporated
into Radiance. I also agree though that the Perex model is preferred, and
Ian Askdown has written his own version of gendaylight in C++ and
validated it. Only issue there is that his tool puts out data in the way
he wishes to use for his work. There's beed to be some effort writing a
wrapper for the Ashdown_gensky output to be able to use his source in
Radiance, but it's totally open source. Food for thought.
 

Rob Guglielmetti  IESNA, LEED AP
Commercial Buildings Research Group
National Renewable Energy Laboratory
15013 Denver West Parkway MS:RSF202
Golden, CO 80401
303.275.4319
robert.guglielmetti at nrel.gov





On 6/19/12 12:20 PM, "Greg Ward" <gregoryjward at gmail.com> wrote:

>I wish I had more to offer to this conversation.  Gensky was written some
>aeons ago before IESNA updated their standards.  I'm not familiar with
>the later standards, though I probably have them in my files somewhere...
>
>The bottom line is that the zenith luminance calculation in gensky is
>based on a turbidity & luminance study published back in the 70's (I
>think) and hasn't been changed.  The general assumption was that anyone
>who cared about absolute levels would input their own zenith value, which
>is still the recommended practice when working from a weather tape or the
>the like.  Honestly, I didn't realize people were relying on the
>turbidity-estimated luminance anymore.
>
>I would have no objection to a volunteer updating the zenith luminance
>calculation in gensky if that's what people want.  Adding the new sky
>types would probably be of even greater interest, though that might
>exacerbate the already sticky issue of choosing which sky model is
>appropriate.  This is why a general sky model like Perez is usually
>preferred.  I don't really have time to dig into the new standard and
>sort it all out.
>
>Cheers,
>-Greg
>
>> From: "Guglielmetti, Robert" <Robert.Guglielmetti at nrel.gov>
>> Date: June 19, 2012 10:33:09 AM PDT
>> 
>> On 6/19/12 11:23 AM, "Zack Rogers"
>><zrogers at daylightinginnovations.com<mailto:zrogers at daylightinginnovations
>>.com>> wrote:
>> 
>> Hi Lars, Rob,
>> 
>> From what I understand, gensky produces accurate CIE sky descriptions
>>according to the older definitions of 3 different sky types (clear,
>>partly cloudy, and cloudy) - but does not include definitions for the
>>newer CIE standard that has 15 different sky types derived from 5
>>different parameters (A-E).  For the older CIE sky types, I am not clear
>>on what gensky uses to predict the magnitude (zenith luminance, Lz) but
>>from what I have seen it does not match the methodology that the IESNA
>>lays out for predicting Lz.  With gensky alone I might get around 70,000
>>lux on a sunny summer day where as the IESNA guidelines would predict
>>around 100,000lux.  This is the reason I wrote a IES_gensky.py python
>>script, it produces the same CIE sky distribution functions (adopted by
>>IESNA) but also uses IESNA guidelines to determine Lz based on some
>>lookup tables.  I recall some babbling from me on this topic around when
>>I developed this:
>> 
>> 
>>http://www.radiance-online.org/pipermail/radiance-general/2003-October/00
>>1074.html
>> 
>>http://www.radiance-online.org/pipermail/radiance-general/2003-October/00
>>1090.html
>> 
>> 
>> Hi Zack,
>> 
>> It's my understanding that gensky produces sky type 1, which is the
>>standard overcast sky, and sky type 12 for the clear sky from the 15 in
>>the CIE table from 2003. I'm not sure which one it uses for the
>>intermediate sky, nor do I know if it's using the 5 parameter method or
>>something else.
>> 
>> - Rob
>
>_______________________________________________
>Radiance-general mailing list
>Radiance-general at radiance-online.org
>http://www.radiance-online.org/mailman/listinfo/radiance-general