And I am very surprised with the Ladybug difference here and started wondering what I am missing.
The scene has 5 faces (cube 10 x 10 x 10 without one) and a ground face. I applied 50% reflective material with 0% specular and 0% roughness to all faces. The grid is every 2m with 1m offset from walls (25 points in total). I used weather data from Copenhagen (DNK_Copenhagen.061800_IWEC.wea) and Tregenza sky (can’t find the option to specify it in ClimateStudio). And the schedule for the simulations is full year.
I attach a script (including Ladybug and ClimateStudio), a model and the files I used for Radiance. To create the .rad files I used su2rad for SketchUp.
I would really appreciate any comments on what might be causing these inaccuracies.
Thank you for the suggestion. I ran the test and it increased the average illuminance to around 600 lux. This is closer to the results from Radiance and ClimateStudio but still significantly lower.
Applying the same parameter (-lw 0.0002) in ClimateStudio gives almost identical average illuminance results as for the previous simulation.
it was recently also talked about in a post, that for an annual daylight analysis is recommended to use an -ad of 25.0000 for a high quality.
Another little advice is to cretae your room a little more realistic for this example and create the surfaces with a closed thickness, avoiding surfaces with only one face and to also insert the glazing, because perhaps the 3 methods you used need a different orientation of the surface to recognize that this is an obstruction.
It would be important if after your last analysis, you would post the result again as you did in the first message.
Unfortunately, the difference still seems significant. For the Ladybug with -ad 4048, the average illuminance is about 400 lux higher.
I couldn’t find a component to create simple glass in ClimateStudio, but I just copied the material .rad file from LB into the CS library, so it should be the same.
If you want to run a fair comparison between LBT and native Radiance with “-lw 0.01 -ab 6 -ad 2048”, you should not pass the parameters through HB Radiance Parameters. This component includes various other parameters that will override the default parameters in rcontrib. You should just add the panel with “-lw 0.01 -ab 6 -ad 2048” to the radiance_par recipe input. And although the enhanced version is modeling the direct sun more accurately, you should set it to False for this comparison (though setting it to True might show the same results).
Also note that the default -lw value that you get from HB Radiance Parameters is 1/(ad*100), i.e., lower than the rule of thumb that @LaFleur mentioned. So that would be -lw 0.000002.
If you are using HB Radiance Parameters, it might be the -lr value in combination with the -lw value that are causing issues. If you set -lw to 1/ad and change the -lr value to -10, which is the default in rcontrib, you might get results closer to native Radiance.
