I built a (12 * 8) room with a window on the east wall and generated some point in the room with height(0.8). A sun_sky with time closed to noon was generated. The following command was used to predict indoor illuminance.
However, for the points close to the window, there is a significant jump in illuminance value – nearly 10 times. For instance, for points with an X value of 11.125, the illuminance value is 4300 or so, but for points with an X value of 11.375, the illuminance value is more than 50000 (X-axis points east). I referred to the manpage of rtrace and used different settings of -ar and -ad, but the issue persisted.
So, I changed the grid to avoid points being too close to the window and it works. However, I would like to understand how to properly address this issue in a general sense. Is there a better approach to handle such cases and ensure accurate results?
I would appreciate any advice or insights to help me stay on the right track.
Have you generated an image of your room? Are you sure the points near the window are not in direct sunlight? This would be the simplest explanation of the jump in illuminance values.
I have generated two images of the room. It seems like the points near the window are in direct sunlight, but the false-color image shows that the direct sunlight area doesn’t have very high illuminance.
Unless you created the image with rpict -i+, the image shows luminance rather than illuminance. Illuminance is luminance times PI divided by the reflectance of the surface, which makes sense if your floor reflectance is somewhere between 20 and 25%.
You are absolutely right. I apologize for the confusion in my earlier reply, I mixed up the units. The new false-color image is right and aligned with the calculation.
Back to the original question, the illuminance values jump because these points are in direct light area. Beyond adjusting the calculation grid, is there any other way you could recommend to minimize the illuminance value jumps?
I am confused by your question. Do you not wish to measure the illuminance at the workplane? The solar patch is what would happen in the real world, which Radiance simulates. You can model a clear sky with no sun using the gensky -s option if you want to see that, but it won’t correspond to the real world at that point.
Sorry for this ‘weird’ question. I think you are right. The solar patch would happen in practical cases. My question arose because I built the same case in DIALux Evo, and the direct light area didn’t exhibit very high values. This made me wonder if there might be specific rules or adjustments in place to handle such areas and avoid extreme values.
I can’t really read the numbers in the false color image from DIALux Evo, and I’m not familiar with this software. It seems from the legend that there are illuminance values on the floor at least exceeding 10,000 lux near the window. I am not sure if these are reflected in the workplane illuminance values or not.
In any case, you need to specify the direct solar radiance or irradiance in both programs if you hope to have any kind of agreement between them. Again, I don’t know the software you are comparing to or its calculation methods, so there’s little I could tell you regardless.
