Hi All,

I have been playing around with a FEM implementation of a simple pressure-correction scheme for unsteady incompressible flow. Being a newbie to CFD, I noticed some things which I do not fully understand.

First of all, when using a triangular grid w/ linear shape function (both velocity and pressure), the scheme works when directly discretizing the (continuous) Laplacian by a simple Galerkin approach (using partial integration). On the other hand, when using the often-cited C^T M^-1 C matrix, the pressure field experiences pure oscillation ( there is no sign of a pressure distribution similar to what I get when directly discretizing the Laplacian). I tested this with T1P0 elements, but without any luck.

I created a small MatLab implementation of the scheme, which is accessible here:

http://www.geocities.com/secondresort/nsexample.zip

This code uses the Galerkin discretization of the continuous Laplacian and employs equal-order interpolation for velocity and pressure.

If you follow the instructions in the tunnel.m file and let the system compute the flow field (as given in the example parameters), you notice a build-up of "pressure" near the inlet at around timestep 1600 (yes, you have to wait a long time, set up some coffee inbetween). Do I take it right, that this is most probably caused by an insufficiently fine discretization near the boundary ?

With best regards, Markus

I found two serious bugs in the code (one should think that this is impossible in only a couple of lines) which are now corrected. I uploaded the new version, looks better now.

Regards, Markus

I selected the link you provided but was unable to get the download, both yesterday and today. No data is transferred and I get the the message "There is a temporary network error preventing download of your file." Has anyone else tried to get this download?

I too use Matlab for incompressible flow, but on rectangles with a pressureless, divergence-free method.

I've just downloaded the file without any problem.

Hi, It would be a great if anybody can help me in this regard. I have done enough of literature survey and at present having a 20+ papers on it. I have implemented the procedure but could not get the result.

if anyone implemented kindly contact me at yenyes_kumar@yahoo.co.in

Thanks in advance !

NSK