Two experimental options for compositional flow solver

[paveltomin]

Two experimental options:

• targetRelativeCompDensChange: use relative change of component density to select time step size
• maxRelativeCompDensChange: chop component density Newton update exciding maximum relative change Both are inactive by default

Should go after https://github.com/GEOS-DEV/GEOS/pull/3138 (merged) and https://github.com/GEOS-DEV/GEOS/pull/3152 (merged)

Was hoping it might help with SPE11a but no, still worth to keep for testing imho.

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 46.66667% with 48 lines in your changes missing coverage. Please review.

Project coverage is 56.41%. Comparing base (02f04e3) to head (db80fad).

Files	Patch %	Lines
...nents/physicsSolvers/NonlinearSolverParameters.cpp	0.00%	30 Missing :warning:
...sSolvers/fluidFlow/CompositionalMultiphaseBase.cpp	72.97%	10 Missing :warning:
...w/IsothermalCompositionalMultiphaseBaseKernels.hpp	50.00%	6 Missing :warning:
...Flow/ThermalCompositionalMultiphaseBaseKernels.hpp	0.00%	1 Missing :warning:
...dFlow/wells/CompositionalMultiphaseWellKernels.hpp	0.00%	1 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## develop #3160 +/- ## ======================================== Coverage 56.41% 56.41% ======================================== Files 1023 1023 Lines 86543 86596 +53 ======================================== + Hits 48822 48855 +33 - Misses 37721 37741 +20 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[dkachuma]

We already have the composition change maxCompFractionChange. Doesn't this indirectly cover the functionality provided by maxRelativeCompDensChange?

[paveltomin]

We already have the composition change maxCompFractionChange. Doesn't this indirectly cover the functionality provided by maxRelativeCompDensChange?

If one density value is much smaller than another, maxCompFractionChange is not going to capture the changes i think, that's why i tried to introduce that relative scaling