EP context doubts for energy simulation (customized shadings)

I was working on the custom shading for both daylight and energy simulation.
I have some doubts regarding the EP context surface component, when the shading is fully closed the solar gain is not reduced to 0. The following trials are fixing the hour at 21st Jun 0900, changing the opening angles of the shading from 0 degree to 60 and 90 degree.

Here are some results:

fully close: 0 degree opening

On the top left, I visualized the radiation result of the window and took the total radiation from Ladybug_Radiation Analysis, which seems correct to me.

The bottom left is the energy consumption on that date, the curve on the top is the solar radiation. And I took the solarGain for that hour from the result (on the right).

From OpenStudio, the solar gain could be slightly different from the radiation result by the Radiation Analysis. But I was expecting the solar gain would go to 0 since the geometry is fully close.

And to compare with the other 2 angles:

semi-open: 60 degree opening

fully open: 90 degree opening

I have tried like:

  • changed the mesh settings in EP context surface,
  • baked the whole facade in rhino and hooked up with EP context,

But nothing changes. And now I have no clues to solve this problem. Can somebody know what went wrong and how can I fix it in a proper way?
Here is the gh and rhino files.
help.rar (1.3 MB)

Thank you in advance.