Currently, the parallel velocity (UA in SOLPS notation) is misinterpreted as the projection of velocity onto the poloidal direction. This is not so, parallel velocity is the projection of velocity onto the direction of the magnetic field: V|| = (VB)/|B|.
To correctly calculate the velocities we must use particle fluxes at poloidal and radial directions defined on cell faces. This will give us 4 velocity vectors for each cell at the cell faces. By averaging them we'll get the projection of velocity on the poloidal plane for each cell, Vplane. Then using Vplane, V|| and B, we can obtain Vtor (many thank to Andrey Pshenov for clarifying all this).
Note that if the flux from cell A to cell B is positive, than the velocity at the face between A and B is obtained by dividing this flux by the density at cell A, but if it's negative, than the flux must be divided by the density at cell B.
I'll add this to the code.
Currently, the parallel velocity (
UAin SOLPS notation) is misinterpreted as the projection of velocity onto the poloidal direction. This is not so, parallel velocity is the projection of velocity onto the direction of the magnetic field: V|| = (VB)/|B|.To correctly calculate the velocities we must use particle fluxes at poloidal and radial directions defined on cell faces. This will give us 4 velocity vectors for each cell at the cell faces. By averaging them we'll get the projection of velocity on the poloidal plane for each cell, Vplane. Then using Vplane, V|| and B, we can obtain Vtor (many thank to Andrey Pshenov for clarifying all this).
Note that if the flux from cell A to cell B is positive, than the velocity at the face between A and B is obtained by dividing this flux by the density at cell A, but if it's negative, than the flux must be divided by the density at cell B.
I'll add this to the code.