Problem with refraction in Dielectric material

radiance-general
1

Dear all,
I am running a simulation test to asses if Radiance calculates correctly the refraction inside a dielectric material. (Snell law n1*sin(A1) = n2sin(A2) ).

I build up two identical geometry one with a dielectric pane and the second without. The direct light can pass through a gap between two plastic surfaces.

Using just direct calculation (sun) I should have two identical bright stripes in dimension but the one passing through the dielectric panel should be shifted due to the refraction in the dielectric pane.

I copy and paste down the geometry and the material definition.

I am running the following command:

cat sensor.grd |rtrace -h -I -ab 0 test-1.oct | rcalc -e '$1=47.4*$1+120*$2+11.6*$3' > test-1-sensor1.out

the file sensor.grd is the list of my sensor:

X Y Z 0 0 1

where:

"X" is 2500 (for the case of the dielectric panel) or 7500 (for the case without the dielectric panel)

"Y" is changing between 10 and 5000 (with a step of 0.1)

"Z" is 0.01

the simulation results are not clear to me.

···

--------------

!gensky -ang 45 0 +s

void plastic ConcFlr_Tiles_Suspended

5 0.576 0.604 0.396 0.0 0.0

void plastic BrickPlaster

5 0.898 0.898 0.898 0.0 0.0

void dielectric Glass

5 0.90 0.9 .900 1.520 0.0

Glass polygon zone01.up

15
    0.0 0.0 5000.0
    5000.0 0.0 5000.0
    5000.0 8.0 5000.0
    0.0 8.0 5000.0
    0.0 0.0 5000.0

Glass polygon zone01.west

15
    0.0 0.0 0.0
    0.0 8.0 0.0
    5000.0 8.0 0.0
    5000.0 0.0 0.0
    0.0 0.0 0.0

Glass polygon zone01.nord

12
    5000.0 8.0 5000.0
    5000.0 8.0 0.0
    0.0 8.0 0.0
    0.0 8.0 5000.0

Glass polygon zone01.south

12
    0.0 0.0 5000.0
    0.0 0.0 0.0
    5000.0 0.0 0.0
    5000.0 0.0 5000.0

Glass polygon zone01.est

12
    5000.0 0.0 5000.0
    5000.0 0.0 0.0
    5000.0 8.0 0.0
    5000.0 8.0 5000.0

Glass polygon zone01.botton

12
    0.0 8.0 5000.0
    0.0 8.0 0.0
    0.0 0.0 0.0
    0.0 0.0 5000.0

BrickPlaster polygon panel-1a

12
    0.0 4.0 0.0
    5000.0 4.0 0.0
    5000.0 4.0 2450.0
  0.0 4.0 2450.0

BrickPlaster polygon panel-1b

12
    0.0 4.0 2550.0
    5000.0 4.0 2550.0
    5000.0 4.0 5000.0
    0.0 4.0 5000.0

ConcFlr_Tiles_Suspended polygon zone02.rad00008

15
    5000.0 8.0 0.0
    5000.0 10008.0 0.0
    10000.0 10008.0 0.0
    10000.0 8.0 0.0
    5000.0 8.0 0.0

ConcFlr_Tiles_Suspended polygon ceiling

15
    0.0 8.0 5000.0
    0.0 10008.0 5000.0
    10000.0 10008.0 5000.0
    10000.0 8.0 5000.0
    0.0 8.0 5000.0

BrickPlaster polygon panel-2a

12
    5000.0 4.0 0.0
    10000.0 4.0 0.0
  10000.0 4.0 2450.0
    5000.0 4.0 2450.0

BrickPlaster polygon panel-2b

12
    5000.0 4.0 2550.0
    10000.0 4.0 2550.0
    10000.0 4.0 5000.0
    5000.0 4.0 5000.0

ConcFlr_Tiles_Suspended polygon zone02.rad00011

15
    0.0 8.0 0.0
    0.0 10008.0 0.0
    5000.0 10008.0 0.0
    5000.0 8.0 0.0
    0.0 8.0 0.0

-----------

Any comment or idea are appreciated

Regards
Francesco

2

Hi Francesco.

A couple of comments:

If I understand correctly you are trying to render an image of a beam of
light passing through
a dielectric material. If you try to create an image you should use the
rpict program and not
the 'rtrace' command line you used.

You mention that you don't understand the results but you don't show us what
they are.
The output of your command line should give you lux values but I don't know
how the
input relates to the geometry (haven't looked at it close enough).

Your scene description below is incomplete. There are some '0's missing in
the descriptions.
I hope that is an effect of copy-and-paste. Otherwise I doubt that the
simulation would run
at all.

Do you know about the docs in

http://radsite.lbl.gov/radiance/refer

in particular

Regards
Thomas

···

On Thu, Aug 20, 2009 at 1:58 PM, <[email protected]> wrote:

Dear all,
I am running a simulation test to asses if Radiance calculates correctly
the refraction inside a dielectric material. (Snell law n1*sin(A1) =
n2sin(A2) ).

I build up two identical geometry one with a dielectric pane and the second
without. The direct light can pass through a gap between two plastic
surfaces.

Using just direct calculation (sun) I should have two identical bright
stripes in dimension but the one passing through the dielectric panel should
be shifted due to the refraction in the dielectric pane.

I copy and paste down the geometry and the material definition.

I am running the following command:

cat sensor.grd |rtrace -h -I -ab 0 test-1.oct | rcalc -e
'$1=47.4*$1+120*$2+11.6*$3' > test-1-sensor1.out

the file sensor.grd is the list of my sensor:

X Y Z 0 0 1

where:

"X" is 2500 (for the case of the dielectric panel) or 7500 (for the case
without the dielectric panel)

"Y" is changing between 10 and 5000 (with a step of 0.1)

"Z" is 0.01

the simulation results are not clear to me.

3

Hi Francesco,

Radiance doesn't compute "caustics" directly -- refractions of beam radiation through dielectric materials. You can visualize the displacement of light rays by looking through a dielectric solid, but the beams of light will not be translated by default. (It is possible to get such behavior using the prism types, but it's not provided as a default behavior because such modeling is expensive and not required by most architectural environments.)

Best,
-Greg

···

From: [email protected]
Date: August 20, 2009 5:58:39 AM PDT

Dear all,
I am running a simulation test to asses if Radiance calculates correctly the refraction inside a dielectric material. (Snell law n1*sin(A1) = n2sin(A2) ).

I build up two identical geometry one with a dielectric pane and the second without. The direct light can pass through a gap between two plastic surfaces.

Using just direct calculation (sun) I should have two identical bright stripes in dimension but the one passing through the dielectric panel should be shifted due to the refraction in the dielectric pane.

....

Any comment or idea are appreciated

Regards
Francesco

4

Dear thomas,
I am not rendering a picture. I would like to calculate the illuminance value on
different sensors using rtrace to calculate the real position of the bright stripes on the "floor" and the dimensions.
I attache now part of the result (result-ff.txt) that I obtained with my simulations. If you look into the file you can understand that radiance seems to calculate correctley the start point of both stripe (last two columns): the bright stripe for the case without the dilectric panel (4th column) starts at point 2454mm (Y value) far from the origin and the bright stripe for the case of the dielectric panel (5th column) starts at point 2454.9mm (a bit further due to the ray refraction inside the dielectric material). But both the two stripes end at point 2553.9mm and there is no more shifting.

I also attach the rad files (geometry-rad and sky.rad) and my script (calc-refr.run) and the two sensors files (points1 and points2.grd).

Thank you very much in advance.
Francesco

Quoting Thomas Bleicher <[email protected]>:

calc-refr.run (572 Bytes)

geometry.rad (1.95 KB)

points1.grd (32.6 KB)

points2.grd (32.6 KB)

result-ff.txt (26.5 KB)

sky.rad (69 Bytes)

···

Hi Francesco.

A couple of comments:

If I understand correctly you are trying to render an image of a beam of
light passing through
a dielectric material. If you try to create an image you should use the
rpict program and not
the 'rtrace' command line you used.

You mention that you don't understand the results but you don't show us what
they are.
The output of your command line should give you lux values but I don't know
how the
input relates to the geometry (haven't looked at it close enough).

Your scene description below is incomplete. There are some '0's missing in
the descriptions.
I hope that is an effect of copy-and-paste. Otherwise I doubt that the
simulation would run
at all.

Do you know about the docs in

RADIANCE Reference Materials

in particular

http://radsite.lbl.gov/radiance/refer/materials.pdf

Regards
Thomas

On Thu, Aug 20, 2009 at 1:58 PM, <[email protected]> wrote:

Dear all,
I am running a simulation test to asses if Radiance calculates correctly
the refraction inside a dielectric material. (Snell law n1*sin(A1) =
n2sin(A2) ).

I build up two identical geometry one with a dielectric pane and the second
without. The direct light can pass through a gap between two plastic
surfaces.

Using just direct calculation (sun) I should have two identical bright
stripes in dimension but the one passing through the dielectric panel should
be shifted due to the refraction in the dielectric pane.

I copy and paste down the geometry and the material definition.

I am running the following command:

cat sensor.grd |rtrace -h -I -ab 0 test-1.oct | rcalc -e
'$1=47.4*$1+120*$2+11.6*$3' > test-1-sensor1.out

the file sensor.grd is the list of my sensor:

X Y Z 0 0 1

where:

"X" is 2500 (for the case of the dielectric panel) or 7500 (for the case
without the dielectric panel)

"Y" is changing between 10 and 5000 (with a step of 0.1)

"Z" is 0.01

the simulation results are not clear to me.

5

Francesco.

As Greg already wrote, Radiance is not suited for your type of calculation.

The difference in your results files might be caused by interpolation of
cached
values. You are calculating tiny steps (0.1units) in a large geometry (10000
units).
I wouldn't trust Radiance to produce absolutely accurate results on that
scale and
the differce you see is less than 2 units.

You can try to disable the ambient cache completely by setting -aa to 0. No
need
for an ambient cache file, then.

Thomas