“Truncated octree” in run Daylight Simulation under electric lighting

Hi everyone,

I’m a new user of Ladybug and Honeybee, and I’m trying to simulate illuminance and glare under electric lighting. I’ve built a simple geometric model, which includes a room with a desk and chairs, and I’ve assigned materials to the geometric models accordingly. However, when I run the model, an error occurred in the “run daylight simulation” component. The error message says, “Solution exception: Failed to read the results! rtrace: fatal - (unnamed_RAD.oct): truncated octree.”

I’ve tried deleting parts of the geometric model separately to identify the cause, and I’ve also searched for similar issues on forums, but I haven’t been able to resolve the problem, and the error persists. I have previously run similar examples on my computer, and they worked fine. However, I keep encountering this error when I create the geometric model myself. Does anyone know how to solve this issue? I’d appreciate any help.

By the way, since I’m doing an electric lighting simulation, I’m using older versions of Ladybug and Honeybee—Ladybug version 0.069 and Honeybee version 0.066.

Because new users cannot upload attachments, please find my model and error file at the link below.

https://1drv.ms/f/s!AjSBZTqX5vjGhRf01RTy-ASWoUJ1?e=Es2rR4

I think removing the space in the customLumName input will likely fix this issue. You can call it Middlelight or something else (with no spaces in between)

Regards,
Sarith

1 Like

Hi @sarith

After following your suggestion to modify the names, my problem was indeed solved! Thank you so much! However, I have encountered another issue. Although the model did not show any errors, the simulated illuminance values are all 0, which is unreasonable. The IES file I used works correctly in other examples, so I believe the IES file itself is not the problem. I found an error file that shows the following:

*** PID 20448: rtrace -I -h -dp 64 -ds 0.5 -dt 0.5 -dc 0.25 -dr 0 -st 0.85 -lr 4 -lw 0.05 -ab 2 -ad 1000 -as 128 -ar 300 -aa 0.1 -af unnamed.amb -e error.log unnamed_RAD.oct

rtrace: warning - non-planar vertex for polygon “object_6.641”
rtrace: warning - non-planar vertex for polygon “object_46.5605”
rtrace: warning - non-planar vertex for polygon “object_44.5396”
rtrace: warning - non-planar vertex for polygon “object_6.748”
rtrace: warning - non-planar vertex for polygon “object_21.2629”

Could this mean there is an issue with my geometric model?

Looking forward to your reply.

This likely a units issue. The ies2rad program from Radiance, which converts the IES to a Radiance format, assumes the units of the scene to be in meters. From the documentation…

-d units

Output dimensions are in units, which is one of the letters respectively. The letter specification may be followed by a slash (’/’) and an optional divisor. For example, -dm/1000 would be millimeters. The default output is in meters, regardless of the original units in the IES input file. Note that there is no space in this option.

Although other options are possible (as the documentation indicates), the meter option has been hardcoded in the Honeybee source code.. I think you will get proper results if you work with meters instead of mm (and scale everything appropriately). Right now in your simulation a single light is trying to illuminate a room which it assumes is 7.5 km x 5.5 km big.

Regards,
Sarith

Hi @sarith,

After adjusting the geometric model units to meters, the illuminance calculations are now functioning as expected. Additionally, I incorporated the “Image Based Simulation” component for DGP calculations, which successfully computes and generates visual outputs. However, the warnings noted in the error file from the previous issue persist, although they do not appear to impact the computation process. Thank you again for your assistance!

Hi, @sarith

About three months ago, based on your suggestion, I resolved the issue with my model that prevented proper simulation of illuminance distribution under electric lighting. However, I have now encountered a new problem. I simultaneously used Climate Studio and Honeybee to simulate the illuminance values in a rectangular room. The dimensions and materials of the room, as well as the number, positions of the lights, and IES files, are all the same, and the number of grid points for the calculations is also identical.

Notably, there is a significant difference in the illuminance values obtained from the two plugins. The average illuminance calculated by Climate Studio is about 7% higher than Honeybee, and the coefficient of variation of illuminance is about 15% lower in Climate Studio than in Honeybee. I believe this difference is not negligible.

I am actively looking for the cause of this discrepancy. I found an error log in Honeybee’s calculation file, which is as follows:

*** PID 1676: rtrace -I -h -dp 64 -ds 0.5 -dt 0.5 -dc 0.25 -dr 0 -st 0.85 -lr 4 -lw 0.05 -ab 2 -ad 1000 -as 128 -ar 300 -aa 0.1 -af unnamed.amb -e error.log unnamed_RAD.oct

Could this error be the reason for the differences in the results between the two plugins? How can I resolve this error?

Looking forward to your reply.

Hi @Congcong , your parameters are way too low. I think setting something like -ab 5 -ad 4096 -as 2048 -aa 0.05 -lw 0.005 -ar 4096 might improve the results (as in calculate the illuminance properly).

Climate Studio algorithms for handling electric lighting have several improvements over Radiance ies2rad (which is what Honeybee legacy uses). So, I cannot guarantee that the parameters that I suggested above will necessarily bring the results closer to Climate Studio.

Regards,
Sarith

Hi, @sarith

I’m sorry for my delayed response. In the honeybee_radparameters component, I set the quality to 2, which corresponds to high quality. The associated Radiance parameters are as follows:
-dp 512 -ds 0.05 -dt 0.15 -dc 0.75 -dr 3 -st 0.15 -lr 8 -lw 0.005 -ab 6 -ad 4096 -as 4096 -ar 300 -aa 0.1.
These default parameters are very similar to the ones you recommended. Under these settings, compared to Climate Studio, the difference in the mean illuminance dropped from 7% to 1%, while the difference in the coefficient of variation dropped from 15% to 7%. I believe the difference in the mean illuminance is negligible, but the difference in the coefficient of variation is still relatively significant.

What caught my attention even more is the speed of calculation. Climate Studio completes the calculation in just 5 seconds, whereas Honeybee takes 270 seconds. This is a huge gap. Even though I set the number of CPUs in Honeybee to 6, its speed is still much slower than Climate Studio.

I’m quite surprised by this. Considering that Honeybee is widely used for building energy simulations, why is its calculation speed so much slower than Climate Studio? Is there any good way to resolve this issue? Otherwise, this would severely limit my ability to use genetic algorithms for lighting optimization, as the calculation speed for each individual is just too slow.

They are not. You might want to check up on what the parameters mean and control, especially ambient accuracy and ambient resolution.