Fix issue with stopped injectors in thermal runs

[totto82]

[totto82]

jenkins build this please

[totto82]

jenkins build this please

[totto82]

This is probably of given the current temperature model for wells, but there maybe issures:

1. For injecting perforations of injectors: l 2619 calculate enthalpy potentially with pressure and temperature from different plases i.e. temperature for injection (top well), but pressure from perforation.
2. For injecting perforations for produces (not(injectors) which is stopped) enthalpy is taken from reservoir at the cell... and derivatives is lost. This is probably a hack which is ok until temperature/enthalpy is added to wells. It also maybe "inconsistent" with output which set the "temperature" in the well to an average of producing perforations (but this would give dynamic large coplings in the model).

1. Do you suggest to use bhp?
2. I am open for using the average well temperature, but I am not sure how much it will impact the solution. The ideal solution is probably to add temperature as an variable to the MSW. With this we could actually model the heat transport in the well.

Do you want me to address some of these issues in this PR, or is that for another day. I think this PR fixes a actual bug as the well temperature from the input was used for stopped wells.

[totto82]

I think this should be merged as it fixes an issue with thermal runs and stopped injectors. As @hnil points out there are other issues that needs to be solved but they could be fixed in a different PR.