Solar Footprint

Hi group,

Just curious if anyone has tried this before.

I need to make a diagram showing where direct sunlight hits the floor
in a room with an odd skylight. I have a basic idea just from using a
solar chart, but I would like to accurately illustrate every possible
sun patch location.

My thought was to make a loop where I render all possible sun positions
for half the year (renderings with sun only, no sky, and with -av 0 0 0
-ab 0) and add all the images together as I go. The resulting image
would show where the sunlight is falling for the entire year. Depending
on how fine I make my sky changes, this might result in quite a few
calculations. (6 months x 16 hrs per day = 96 calculations.. not too
bad)

Has anyone tried to do this before?

Mark

Mark de la Fuente wrote:

Hi group,
Just curious if anyone has tried this before.
I need to make a diagram showing where direct sunlight hits the floor in a room with an odd skylight. I have a basic idea just from using a solar chart, but I would like to accurately illustrate every possible sun patch location.
My thought was to make a loop where I render all possible sun positions for half the year (renderings with sun only, no sky, and with -av 0 0 0 -ab 0) and add all the images together as I go. The resulting image would show where the sunlight is falling for the entire year. Depending on how fine I make my sky changes, this might result in quite a few calculations. (6 months x 16 hrs per day = 96 calculations.. not too bad)
Has anyone tried to do this before?

I've always used animations for that sort of thing, but your idea sounds cool. One thing that you may want to do is actually give your ambient value a little boost, so that the other surfaces are slightly illuminated. Otherwise, it's tough to tell what you're looking at. The direct sun patch against an all-black background could be tough to place in context, depending on what is illuminated by the sun. John M. shared a tip at last year's workshop too, he used a green-tinted ambient value (-av 0 .5 0) so that it's clear what illumination is "fake" and what is direct sun. Giving your model a little "ambient glow" with -av will make for a clearer rendering, I think.

I've never merged images together like that and would be interested to hear how you go about it, and what issues there are to consider, such as exposure & filtering. I'd love to see what you come up with. This could be a useful tool.

P.S.
Since you're doing -ab 0, the number of calculations shouldn't be a cause for alarm, the direct calculation is really fast.

···

----

      Rob Guglielmetti

e. [email protected]
w. www.rumblestrip.org

Mark,

I suggest you to generate a single radiance scene file containing as many suns as you need. Then using a single radiance run you will get the expected result.

I is a small program I have used to generate multiple gensky calls for a whole year:

#include <stdio.h>
#include <stdlib.h>
void main() {
         int mois,jour,heure;
         int nj[12] = {31,28,31,30,31,30,31,31,30,31,30,31};
         double hh;
         for (mois=0;mois<12;mois++) {
                 for (jour=0;jour<nj[mois];jour++) {
                         for (heure=0;heure<24;heure++) {
                                 hh=heure + (double) rand()/RAND_MAX;
                                 /* Adjust latitude, longitude and timezone according to your location hereafter */
                                 printf("gensky %d %d %.4lf -a 46.4824 -o -7.0894 -m -15\n",mois+1,jour+1,hh);
                                 }
                         }
                 }
         exit(0);
}

You will note that I have added a random perturbation on the hour parameter in order to avoid having all the suns aligned on meridian lines.

Raphael Compagnon

···

At 28.10.2004 15:40, you wrote:

Hi group,

Just curious if anyone has tried this before.

I need to make a diagram showing where direct sunlight hits the floor in a room with an odd skylight. I have a basic idea just from using a solar chart, but I would like to accurately illustrate every possible sun patch location.

My thought was to make a loop where I render all possible sun positions for half the year (renderings with sun only, no sky, and with -av 0 0 0 -ab 0) and add all the images together as I go. The resulting image would show where the sunlight is falling for the entire year. Depending on how fine I make my sky changes, this might result in quite a few calculations. (6 months x 16 hrs per day = 96 calculations.. not too bad)

Has anyone tried to do this before?

Mark

Mark,

This paper (below) describes one way to accelerate the computation of
annual irradiation images.

Mardaljevic, J. & Rylatt M. (2003). Irradiation mapping of complex urban environments: an image-based approach. Energy and Buildings, Vol. 35, pp. 27-35

I can send you a PDF if you have trouble finding it. Unfortunately,
this appeared just before Elsevier decided to leave the colour in the
PDFs.

The presentation "Precision modelling of parametrically defined solar
shading systems: Pseudo-Changi" gives some more examples (in colour).
It's in the list here:
http://www.iesd.dmu.ac.uk/~jm/publications.html

-John

···

-----------------------------------------------
Dr. John Mardaljevic
Senior Research Fellow
Institute of Energy and Sustainable Development
De Montfort University
The Gateway
Leicester
LE1 9BH, UK
+44 (0) 116 257 7972
+44 (0) 116 257 7981 (fax)

[email protected]
http://www.iesd.dmu.ac.uk/~jm

Rob, I like the off-color idea.. this is after all just a diagram. To
combine images all you need to do is type pcomb pic1.pic pic2.pic >
newpicture.pic. It just adds all the values together. This has been a
very useful for me in the past. Keep in mind you are adding light, so
you want to try and keep from "double dipping" sort of speak. So you
would not have light A and B in one picture and B and C in the other and
add them up. You would have A, B, and, C all in separate pictures.
This can be a big time saver if you are trying to develop images for
various "scenes".

Mark

Mark de la Fuente wrote:

Hi group,

Just curious if anyone has tried this before.

I need to make a diagram showing where direct sunlight hits the floor

in

a room with an odd skylight. I have a basic idea just from using a
solar chart, but I would like to accurately illustrate every possible

sun patch location.

My thought was to make a loop where I render all possible sun

positions

for half the year (renderings with sun only, no sky, and with -av 0 0

0 -ab 0) and add all the images together as I go. The resulting

image

would show where the sunlight is falling for the entire year.

Depending

on how fine I make my sky changes, this might result in quite a
few calculations. (6 months x 16 hrs per day = 96 calculations..

not

too bad)

Has anyone tried to do this before?

I've always used animations for that sort of thing, but your idea
sounds
cool. One thing that you may want to do is actually give your ambient

value a little boost, so that the other surfaces are slightly
illuminated. Otherwise, it's tough to tell what you're looking at. The

direct sun patch against an all-black background could be tough to
place
in context, depending on what is illuminated by the sun. John M. shared

a tip at last year's workshop too, he used a green-tinted ambient value

(-av 0 .5 0) so that it's clear what illumination is "fake" and what is

direct sun. Giving your model a little "ambient glow" with -av will
make for a clearer rendering, I think.

I've never merged images together like that and would be interested to

hear how you go about it, and what issues there are to consider, such
as
exposure & filtering. I'd love to see what you come up with. This
could be a useful tool.

P.S.
Since you're doing -ab 0, the number of calculations shouldn't be a
cause for alarm, the direct calculation is really fast.

···

----

      Rob Guglielmetti

e. [email protected]
w. www.rumblestrip.org

Raphael, thank you for your response. Random perturbation noted. You
guys think of everything! :slight_smile:

I was under the impression that if you call out the longitude and the
standard meridian that Radiance will figure out where the sun needs to
be in relation to your exact location.

Making one sky with all possible sun conditions was option B. But I
had not been able to get it to work. I believe the problem lies in the
fact that when gensky creates the sky, it always uses the same naming
convention. I have been creating csh scrips (running Cygwin) and have
been unhappy about how fickle they seem to be. What is this written in?
Unfortunately I don't understand what the script is doing so I could
not really use it.

Anyway, I wrote a script and wound up with the following sky.

http://images.ofoto.com/photos887/5/91/16/85/95/0/95851691506_0_ALB.jpg

It is very strange that the suns don't line up like they do on a sun
chart. Any idea why this is so?

Start of file.....

#!/bin/csh -fv

···

#
### This script generates multiple skies and then creates one multiple
sky octree.
#

### Daylighting parameters

set day = "21"

rm ./scene/xskies.rad

foreach month (012 001 002 003 004 005 006)
foreach hr (005 006 007 008 009 010 011 012 013 014 015 016 017 018 019
020)

gensky $month $day $hr +s -a 38.7 -o 90.5 -m 90 >
./skies/'sky'$month$day$hr'.rad'
echo "\!xform -n sky ./skies/sky$month$day$hr.rad" >>
./scene/xskies.rad

end
end

oconv -f ./scene/xskies.rad ./skies/sky.rad > ./octree/skies.oct

... end of file

Mark

Message: 3
Date: Thu, 28 Oct 2004 17:10:56 +0200
From: Raphael Compagnon <[email protected]>
Subject: Re: [Radiance-general] Solar Footprint
To: Radiance general discussion <[email protected]>
Message-ID: <[email protected]>
Content-Type: text/plain; charset="us-ascii"

Mark,

I suggest you to generate a single radiance scene file containing as
many
suns as you need. Then using a single radiance run you will get the
expected result.

I is a small program I have used to generate multiple gensky calls
for a
whole year:

#include <stdio.h>
#include <stdlib.h>
void main() {
         int mois,jour,heure;
         int nj[12] = {31,28,31,30,31,30,31,31,30,31,30,31};
         double hh;
         for (mois=0;mois<12;mois++) {
                 for (jour=0;jour<nj[mois];jour++) {
                         for (heure=0;heure<24;heure++) {
                                 hh=heure + (double) rand()/RAND_MAX;
                                 /* Adjust latitude, longitude and
timezone
according to your location hereafter */
                                 printf("gensky %d %d %.4lf -a 46.4824
-o
-7.0894 -m -15\n",mois+1,jour+1,hh);
                                 }
                         }
                 }
         exit(0);
}

You will note that I have added a random perturbation on the hour
parameter
in order to avoid having all the suns aligned on meridian lines.

Raphael Compagnon

Do a search for "analemma" (or "analema") and see why what you've found is to be expected.

I especially like the image on:
http://vrum.chat.ru/Photo/Astro/analema.htm

Mark

···

On Fri, 29 Oct 2004, Mark de la Fuente wrote:

Anyway, I wrote a script and wound up with the following sky.

http://images.ofoto.com/photos887/5/91/16/85/95/0/95851691506_0_ALB.jpg

It is very strange that the suns don't line up like they do on a sun
chart. Any idea why this is so?