opm-simulators Assessing Convergence Problems

Hi, Is there a way of either exporting the number of newton iterations in the restart file or reporting the cells with the highest number of iterations to the PRT file?. I'm trying to identify grid portions that may be causing convergence problems. Thanks a lot in advance

Dec 06 '23 23:12 mbmfer

With the Newton method, the number of iterations is the same for all cells in the grid. I think we support some output that can help with identifying the hardest regions though, @bska or @totto82 maybe you have some suggestions?

In the meantime: cells with small pore volume can give you a hard time, as a workaround try using MINPV to remove the smallest, possibly combined with PINCH to reconnect cells across the removed cells.

Dec 07 '23 07:12 atgeirr

I think we support some output that can help with identifying the hardest regions though,

You may be thinking of the CONV=N result arrays (CNV_OIL, CNV_GAS, and CNV_WAT). At the moment Flow does not support those arrays. For what it's worth, we do plan to add support for them but it likely won't happen until the new year.

Dec 07 '23 08:12 bska

Thanks a lot for your replies. In my original message I was referring to:

RPTRST
  CONV=n

and

RPTSCHED
  NEWTON=m

Dec 12 '23 12:12 mbmfer