Shock reflections over a ramp

8 views
Skip to first unread message

Andrea

unread,
Jun 10, 2008, 3:16:17 PM6/10/08
to amrita-ebook
Hello,
I'm studying the reflection of a shock wave over a ramp. I first
studied the problem using the body_roe solver. To evaluate the
precision of the solution I tried to use other solvers, like the
godunov_km for example, but I obtained only unrealistic solutions.
Is the body_roe solver the only one able to treat the problem? Or are
there also some versions of the other solvers for the ramp problem?\
Thanks
Andrea

James Quirk

unread,
Jun 10, 2008, 3:36:01 PM6/10/08
to amrita-ebook
Andrea,

Without seeing your script it's hard to determine where the problem is.
But if you send me a mailit, I'll take a look at it.

James


> Thanks
> Andrea
>

James Quirk

unread,
Jun 10, 2008, 11:38:00 PM6/10/08
to Andrea, amrita-ebook
Andrea,

If you're using the ramp.mailit from the VKI notes then
please be aware that it is showing its age i.e. it's
more klunky than it needs to be. A better place to
start would be:

amrcp latex/method.3

which itself can be polished to give the attached.
Note to run these examples you will need to fix
a bug that I recently posted to the newsgroup.

Once you can run the attached, you can convert
the CylinderProblem into a WedgeProblem by
changing the level-set passed to the routine:

AXS::InternalBoundary

It defines a distance function from a cell to the
body of interest.

James

run_smokefoil_polished

Andrea Papi

unread,
Jun 11, 2008, 11:36:38 AM6/11/08
to James Quirk, amrita-ebook
James,

thank you very much for your advice. I'll start to look at the script right away.
However, there's the script I was working with.

Andrea
nonreactiveramp.mailit

James Quirk

unread,
Jun 11, 2008, 12:35:30 PM6/11/08
to Andrea Papi, amrita-ebook

Andrea,

The mailit you sent me is incomplete. At least the top-level scripts
are missing the procedure BasicRampGrid2. Anyhow it looks like
you've modified the ramp.mailit from the VKI notes, and as I noted
earlier that particular script is a bit grungy. The smokefoil
example I sent you is much more in line with how AMRITA would
be programmed today. It does, however, require you to view
the source with a folding editor and from your mailit
I'm guessing that you're using emacs and so you won't
see the source in the intended fashion.

James

Andrea

unread,
Jun 19, 2008, 4:10:02 PM6/19/08
to amrita-ebook
James,

thank you for your help. Using a level-set method the script seems to
be more accurate.
However, I'm facing another problem. I'd like to print the values of
the thermodynamic parameters along the ramp, but it seems that the
"along" command works only with a constant value of y so every time I
try to write y as a function of x I have an error. Is there another
command that work with a variable y? Or, is it possible to use the
"along" command with other frames of reference?(ie: a rotated and
translated one with the X' axis parallel to the ramp)
Thanks,

Andrea

James Quirk

unread,
Jun 19, 2008, 7:34:08 PM6/19/08
to Andrea, amrita-ebook
Andrea,

On Thu, 19 Jun 2008, Andrea wrote:

>
> James,
>
> thank you for your help. Using a level-set method the script seems to
> be more accurate.
> However, I'm facing another problem. I'd like to print the values of
> the thermodynamic parameters along the ramp, but it seems that the
> "along" command works only with a constant value of y so every time I
> try to write y as a function of x I have an error. Is there another
> command that work with a variable y? Or, is it possible to use the
> "along" command with other frames of reference?(ie: a rotated and
> translated one with the X' axis parallel to the ramp)

As it stands, the along command is limited to lines of x or y = constant.
I notionally have the machinery to be able to cope with arbitrary
curves, but I havn't found the time to put it all together. There is,
however, a probe command that can be used to interrogate the flow
solution at arbitrary points. Therefore there is nothing to stop you
doing something along the lines of:

do t=0,1,$dt
set x #= f($t)
set y #= g($t)
probe Q[] at $x,$y -> Q
end do

For details on the probe command, see amrhelp.

James

> Thanks,
>
> Andrea

Reply all
Reply to author
Forward
0 new messages