Fix oscilations

[totto82]

WIP

• ads iterations to the well/group constrain check. replaces the remaining parts of https://github.com/OPM/opm-simulators/pull/3410/files

• Keep well at individual controls when oscillating. See discussion in Set option maximum-number-of-well-switches default value to 10 #5599

Needs a lot more testing....

[totto82]

jenkins build this please

[totto82]

This PR with --shut-unsolvable-wells=false restore Norne results

[totto82]

jenkins build this please

[steink]

This PR with --shut-unsolvable-wells=false restore Norne results

Looking into this. Currently a well that has reached it's maximum switching limit is not only prevented from switching control, but also status (e.g., to stop). This leads to convergence failure for e.g., non-feasible bhp. This is a problem with master and not the PR, but probably good to fix in the current PR. Will do some experimentation and propose a fix.

[steink]

@totto82 could you make the following small update: aec5170 ? This will prevent forcing a non-operable well to be open in cases the control mode is forced (prevents well convergence failures).

[totto82]

@totto82 could you make the following small update: aec5170 ? This will prevent forcing a non-operable well to be open in cases the control mode is forced (prevents well convergence failures).

Thanks. I have updated accordingly

[totto82]

jenkins build this please

[steink]

PR fails to complete simulation for MOD4_UDQ_ACTIONX, however I'm not able to reproduce this behaviour locally. Will dig a bit more to try figure out what's going on.

[steink]

Found an issue which by coincidense resulted in failed simulation of MOD4_UDQ_ACTIONX for this PR. Will attempt to fix this separately.

[steink]

#5634 has been merged, so I think this should fix the trouble for MOD4_UDQ_ACTIONX for this PR. I think the changes here looks good except the comment above. In my opinion maximum-number-of-well-switches can safely get a default value of 10.

[atgeirr]

jenkins build this please

[steink]

See #5641