[casper]

Hello all,

I am using immersed boundary method (IBM) with finite difference discretization to simulate viscous flow around a circular cylinder (there are grid points and internal flow as well inside the immersed body). As IBM uses cartesian grid, I am finding it difficult to extract pressure values at the surface of the cylinder. Also for streamline and contour plots, using some post processing tool like paraview doesn't seem possible since it uses points inside the cylinder to interpolate at the surface. I must somehow define the presence of the immersed boundary during post processing, but I am unable to figure out how. Would be grateful if someone could give suggestions on this.

Thanks!

[FMDenaro]

Originally Posted by casper

Hello all,

I am using immersed boundary method (IBM) with finite difference discretization to simulate viscous flow around a circular cylinder (there are grid points and internal flow as well inside the immersed body). As IBM uses cartesian grid, I am finding it difficult to extract pressure values at the surface of the cylinder. Also for streamline and contour plots, using some post processing tool like paraview doesn't seem possible since it uses points inside the cylinder to interpolate at the surface. I must somehow define the presence of the immersed boundary during post processing, but I am unable to figure out how. Would be grateful if someone could give suggestions on this.

Thanks!

could you post your figures?

[casper]

I have attached the streamline plot plotted in paraview as an example. Near the cylinder, it can be seen that the stream lines aren't smooth. As we know that they must smooth in all cases except when there is shock, I can infer that paraview is using points insider the cylinder as well.
finer grid streamlines at x=6.7.jpg

[FMDenaro]

Originally Posted by casper

I have attached the streamline plot plotted in paraview as an example. Near the cylinder, it can be seen that the stream lines aren't smooth. As we know that they must smooth in all cases except when there is shock, I can infer that paraview is using points insider the cylinder as well.
Attachment 41849

the streamlines seems not to follow the geometry of a cylinder but feel as a rough approximation...however, I strongly suggest to see the field you really solve (u,v,p), not the post-processing with streamlines. Check first if your solution is correct

[casper]

To check the solution I need to plot, say for instance, pressure coefficient values on the surface of the cylinder. This creates a problem in Paraview since it interpolates values near the surface using values inside the cylinder.
Is there any way to remove the points inside the cylinder? Or any other post processing tool for that matter?

[kaya]

I am sure there is a workaround in paraview. But wouldn't it be easy if you open your data in matlab

[FMDenaro]

Originally Posted by casper

To check the solution I need to plot, say for instance, pressure coefficient values on the surface of the cylinder. This creates a problem in Paraview since it interpolates values near the surface using values inside the cylinder.
Is there any way to remove the points inside the cylinder? Or any other post processing tool for that matter?

I do not agree that a check of the pressure coefficient is equivalent to a check on the solution....
however, is better you extctract point-vales on the body and save in a vector to be further analysed