unsteady corrections in free_flight

[tommy-engels]

The unsteady corrections in the free-flight solver need a thorough check.

For the falling cylinder, if we set
unst_corrections=0;
then we implicitly take those into account in the eqn, since it reads (rho_s-rho_f)u_dt = ...
so if you set
unst_corrections=1;
they are taken into account twice.
For the rotation, it needs to be checked as well