search

nikola-m / freeCappuccino

A three-dimensional unstructured finite volume code for fluid flow simulations.
GNU General Public License v3.0
51 stars 25 forks source link

missing files: RK4_FILES #2

Open Jasonzhaolinchen opened 6 years ago

Jasonzhaolinchen commented 6 years ago

are those files all missing?

nikola-m commented 6 years ago

Hi! I have removed those files since I was not able to finish the work on explicit solver with RK4 time stepping. It was just a sketch, and a reminder is left in the Makefile. So the functionality was not compete so I removed it. If you are interested, I am planning in creating explicit time-stepping solver for Euler and Navier-Stokes using existing infrastructure of the code. If you are interested let me know and we can join efforts.

Nikola

On Thu, Mar 15, 2018 at 9:49 PM Jasonzhaolinchen notifications@github.com wrote:

are those files all missing?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nikola-m/freeCappuccino/issues/2, or mute the thread https://github.com/notifications/unsubscribe-auth/AAf1jwHCamqtwEKs7yjMflPevbVaf3yKks5tetPkgaJpZM4Ss3In .

-- Nikola Mirkov - Web Studio Mirkov Principal Software Engineer

Jasonzhaolinchen commented 6 years ago

Thanks for the reply, yep, Let's put our efforts in this code. Will get couple of validation cases done later on

nikola-m commented 6 years ago

We may use this thread to report the progress. Are you more interested in compressible or incompressible explicit time-stepping solver?

My plan, at the moment, is to setup an Euler eq. solver and test it on 1D Sod's tube case and so on to 2D, and 3D geometries. Of course in Cappuccino meshes are always 3D, but if we align boxes in a line and do nothing with the lateral boundaries - it would be essentially a 1D case.

The infrastructure that is in place is mesh management, gradient reconstruction and limiting (via gradient limiters e.g. Bart-Jaspersen, Venkatakrishnana, and Wang modified Venkatakrishnan as in Fluent). Using cell centered gradients reconstruction of face values using extrapolation from two sides is possible. I am planning to use some open-sourced routine for inviscid fluxes (approximate Riemann solver), and the only thing left after flux evaluation is evolution, which is easily done with popular solutions like third order TVD Runge-Kutta of Gotlieb and Shu.

Later viscous fluxes ca be added to form Navier-Stokes eqns.

Also I can send a link to paper which I wanted to use in implementation of RK4 projection method - if you find it interesting.

Please let me know what do you think about the code so far, and what could be possible improvements.

Nikola

On Fri, Mar 16, 2018 at 11:16 AM Jasonzhaolinchen notifications@github.com wrote:

Thanks for the reply, yep, Let's put our efforts in this code. Will get couple of validation cases done later on

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nikola-m/freeCappuccino/issues/2#issuecomment-373667800, or mute the thread https://github.com/notifications/unsubscribe-auth/AAf1j86UQ2HfouJcuiCgabdcLE9XkAq1ks5te5DrgaJpZM4Ss3In .

-- Nikola Mirkov - Web Studio Mirkov Principal Software Engineer