* 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.

* update ref sol

* Adjust salt water intrusion test that was meant to use mass fractions
but actually used mole fractions before this bugfix

Fix #579 where is was noted that 1.0 as default is more a disadvantage
than an advantage. Just specify 1.0 in the input file explicitly
instead.

Fix #608

Katharina Heck authored 6 years ago and Kilian Weishaupt committed 6 years ago

[test][multidomain] add 1p2c_1p2c test to compare diffusion laws. MaxwellStefan and Fick's Law produce the same solution for 2 components

Kilian Weishaupt authored 6 years ago and Timo Koch committed 6 years ago

* test failed on master with old 360 deg 3d ref solution