OPM
improving the SSI + Newton PTFlash method
with the master branch, the test introduced in https://github.com/OPM/opm-models/pull/845 can not run through with the SSI + Newton PTFlash method. The major reason is because it enters the Newton solution process with 0 or 1 valued L (single phase).
Through bounding the Lmin and Lmax between 0. and 1., it looks running fine now. Maybe more strict criteria need to be check whether we should jump into the Newton method.
Compared with the original version, it does not improve the SSI+Newton PTFlash solution anymore. (The first commit originally there were removed due to the merge of https://github.com/OPM/opm-common/pull/3956 )
It looks like rather cosmetic with its current form. But it formulate the SSI+Newton flash slightly more complete when coming to logic.
But SSI+Newton does not run through the case introduced from https://github.com/OPM/opm-models/pull/845 .
The PR introduces some nice reformulation. I hope it can get in.
most of the changes are cosmetic and small reformulation. Jenkins approves. I am self-merging it.