features
* tests for zeroeqni and zeroeq2cni
* split update() in "classical" update() and updateRANSProperties()
* do not use asImp_() but virtual inheritance
* cleanup

* ransnc uses now nPhaseIdx = 1

navierstokes
* make vtkoutputfields more modular

rans
* move eddyViscosity functions to the specific rans implementation

* better distinguish between total mass balance and component mass balances

* do NOT use updateStaticWallProperties() in problem constructor
* correct calculation of maximum velocity
* add Laufer velocity profile
* update wall variable names

Simon Emmert authored 7 years ago and Kilian Weishaupt committed 7 years ago

richardswelltracerproblem

The box-reference solution is adapted because it fails in the context of
very small values. Additionally the box-solution's max value for
x^SimpleH2O_D2O is now closer to the tpfa-solution's max value.

This commit introduce a model traits class that encapsulates a few properties,
such as NumEq, NumComponents as well as the specifications on active processes
(EnableAdvection ... ). The respective property tags become obsolete and can be
deleted.

Differences to the release are sufficiently small.

Synchronizing the boundary condition evaluation points makes the
test pass. Keep the setting on the master, update the reference
solution, simplify the Neumann boundary condition.

Changes to the release are negligibly small.

Manually picked from 8aae0b9f.

A lot has been changed in the test setting from the 2.12 release:
boundary condition types, evaluation points for boundary condition
values and spatial parameters, domain heterogeneity, complexity of
the fluid system relations, Newton and linear solver tolerances,
maximum time step size. Synchronizing the settings between 2.12 and
master renders the differences negligibly small. Keep the new
setting, uncomment the domain heterogeneity, simplify the Neumann
boundary condition, adapt the reference solution.