There is a FiniteRateEulerEquations within amr_sol which was never quite
finished. I would like to use such a package and would like to know whether
anyone out there has made any headway in this direction or has any plans to
do it, whether from scratch or starting from what has been done in the
existing code.

Dear Hans,
I would also be interested in using such a package, although I have
not heard of it until your message.  What is the state of development
of the package?  Perhaps a collective effort in getting it up and
running would be desirable.
One of my students, Brian Maxwell, is looking into different options
for developing a compressible flow solver for multi-step chemistry.
Perhaps linking amr_sol with some of the cantera source could be a
desirable alternative.
matei

On Aug 20, 6:39 pm, Hans Hornung <hghorn...@gmail.com> wrote:

Matei,

The package that Hans is referring two is located at:

$AMRITA/plugin/amr_sol/BCG/equations/FiniteRateEulerEquations/chemlib

it hasn't been used since 1998 and should really be deprecated from
AMRITA's base installation kit.

The way to do it is to construct a patch-integrator which
is linked with Cantera. But rather than diving straight in,
Brian would be advised to look at:

    amrcp latex/shifter.1
    amrcp latex/shifter.2

which show the amr_sol to patch-integrator plumbing in its simplest
setting. The only problem I forsee is the fact that Cantera
is written in C++ and there can sometimes be problems with
the static initialization of objects, but we'll worry about
that when the time comes.

It's also worth noting that it would be straightforward to write
the patch-integrator in a way that holds Cantera at arm's length.
By that I mean the patch-integrator makes calls to EOS::P(W) and
EOS::C(W), EOS::S_UPDATE(W), W being the solution vector, and has
no knowledge of the back-end routines that work with the chemisty.

James