Total value on meshSelector component is not the same as the total radiation analysis result

It seems the total value reported on the meshSelector component is not the same total radiation result reported by the radiation analysis component: (using Singapore’s epw file)

According to the annotation for the “total value” output node, it should report the total radiation value for the input analysis mesh if it is a radiation analysis mesh:

Considering that the difference is quite huge, may I ask if Mostapha or Chris can kindly help to check if there is any mismatch between the “radiation analysis” component and the meshSelector component regarding the way of mass addition of the “irradiation X area” values for all the faces in a mesh?

Thank you very much!

meshSelector issue.gh (563 KB)

Someone needs to have a closer look to this but just a quick check shows there is a difference between the mesh face area if I use the Area component or calculate it for each face and add them up together. Since your geometry doesn’t have non-planar faces, or that’s what I assumed, we shouldn’t have this difference. This can by itself be the source of the error.

Thank you, Mostapha!

Adding to the previous demo file, it seems that three different total radiation values can be calculated depending on the component used, and the difference can be as high as 3%:

meshSelectorissue _v02.gh (569 KB)

Grasshope,

Thank you for pointing this out. Because of this issue, I realized that Rhino’s DeleteFaces() method for meshes completely destroys the original order of the faces in the mesh, which was causing a mis-calculation of the area that meets the threshold. I have fixed this now by running the area calculation before the deletion of faces and this give perfect agreement between the mesh threshold selector and the totalRadiation:

https://github.com/mostaphaRoudsari/ladybug/commit/3c75d204afbcfb53…

The reason why the Grasshopper manual calculation result is a little different is because of tolerance issues (Grasshopper is rounding off the number of decimal places for each mesh face area). The results output from the Ladybug components are the more accurate results.

-Chris

meshSelectorissue_v04.gh (575 KB)

Dear Chris, thank you very much for fixing the issue!

If I may, it seems that the total value derived from the manual calculation method is still not the same:

Does this mean that the manual calculation workflow as shown above somehow scrambled the order of the mesh?

The difference can also be seen when a threshold value is applied, when the total areas of the mesh meeting threshold are the same for both methods, but the total radiation values are not:

This is strange because both the cullFace component and the dispatch component take the same pattern derived from the flattened results.

Appreciate if you can kindly advise.

meshSelectorissue_v05.gh (585 KB)