Hello once again. I may have found the reason for the negative E_Grid values.
As the values occur during early morning and early evenings with low irradiance, could it be that PVsyst assumes that the site is still 'connected' to the grid at these times.
And as a result the transformer constant losses are applied, which are higher than the energy produced by the system for the given low irradiance, resulting in a negative E_Grid.
I ran the simulation again and got the following results with available energy at inverter output parameter - EOutInv (Sorted by lowest first) and MV transformer losses parameter - EMVTrfL, which supports the above explanation.
Please correct me if this explanation is incorrect.
Thank you.