i tried to edit the plugin of annual daylight of honebee and get my own sky matrix to be used instead of using gendaymtx. (just for research purpose).
i did this by adjusting command_util.py of honeybee radiance command to skip the use of of gendaymtx whenever command is passed to use it for sky mtx and instead pass on my own sky mtx file (obrtained for same location somehow differently).
doing this i noticed many negative lux levels of higher magnitude especially near glass skylights in geometry.
initially it looked it is beacuase of sundirect mtx mismatching with sky mtx.
what could be the possible reason
What’s happening here is that the value that is calculated under direct-only sky is much higher than the addition of direct-sun and total-sky. The final value math is something like this.
final_illum = total_sky - direct_sky + direct_sun
You should compare the values for each sensor is the 3 files and see what might be happening. If I remember correctly, we also have components that allow you to visualize these values in Grasshopper for debugging.