•GHI: if you set tilt angle to 0 (a horizontal panel) and select raw data, the sum of the “direct” and “indirect” columns will correspond to GHI.
Just for a clarification, this would also require treating the direct radiation with (before adding it to diffuse component):
((60 * df['direct'] * np.cos(df_b['sun_zenith'])) / (np.cos(np.arccos(np.sin(df_b['sun_alt'])) * df_b['duration']))
Right? I have few pages of proofs leading to this solution (mainly because I'm a bad mathematician, could have done it in one).
At least this allowed me to change the angle/azimuth freely for gsee-tool from just one downloaded dataset and still receive +-0,01% same values for direct and diffuse as files manually downloaded from ReNinja for any combination of angle/azimuth. Not having this meant that the direct component was off for other angles/azimuths, and thus the GHI was off.