This reverts merge request !1615

Gabi Seitz authored 5 years ago and Johannes Hommel committed 5 years ago

instead of the water vapour pressure.

One reference had to be updated because the computations of mole and
mass fractions is not exactly the same.

Katharina Heck authored 5 years ago and Timo Koch committed 5 years ago

This fixes a bug introduced with !1517 and thus fixes the
test_mpnc_kinetic test. However the newly introduced 2p2c test
references were created with the buggy code and thus have to updated (in
this patch). We make it more robust by increasing the Newton accuracy,
introducing checkpoints and incresing the initial time step size, which
results in many less time steps.

Katharina Heck authored 5 years ago and Bernd Flemisch committed 5 years ago

[test][nonequilibrium][2p2c] make test nonisothermal and set better boundary conditions. add reference solution for box and tpfa

* Use boundary velocity v_j in term (v_i*v_j) both for upwinding and
  central average
* do not use upwinding for kovasznay test
* update reference solutions for kovasznay and angeli test

Beatrix Becker authored 6 years ago and Bernd Flemisch committed 6 years ago

[sequential 2p2c flash] revert change in Rachford-Rice equation (both quantities are mass-based), add literature, change references

Dennis Gläser authored 6 years ago and Timo Koch committed 6 years ago

The reference solution of this test needed to be updated anyway now that
the derivative computation for solution-dependent parameters is fixed.
We now use a more robust setup of the test with a fixed time step size
as well as a shorter end time to speedup the test.

* Always compare the last vtu file
* Set maximum time step size lower
* Set Newton and linear solver tolerance lower

Dennis Gläser authored 6 years ago and Timo Koch committed 6 years ago

this jump should be independent of the choice for xi, which is why
here we test if for different values of xi the same reference solution
is obtained.