The localtoglobalmap is there to be a cheap way of obtaining a c0 continuous field but is simply doing a local degrees of freedom to global degrees of freedom copy and then a copy back. So at the interface between two elements it will just take the element information that is copied in last. So this is not a formal way of obtaining a more accurate C0 field.

Without this option the module assembles and inverts a global mass matrix and then uses this to project the local field to a C0 global continuous finite element field. My understanding is that this is therefore a formally more accurate process to get a C0 field but obvious is more expensive. In taking this approach you cal also decide whether to impose boundary conditions when inverting this matrix. The default does not do this but if you select the usexmlbcs it will enforce (in a strong sense) the boundary conditions in this inverse from the input xml file.

Underresolved data will always likely lead to high values at the wall so smoothing this data is then a challenge. Another option is to consider using the helmsmoothing option, however this its mainly possible for 2D fields rather than 3D fields as it has to invert a non-positive definite matrix. Details of this projection are discussed in the user-guide as I believe are the other two options.

Spencer Sherwin FREng, FRAeS
Head of Aerodynamics Section,
Director of Research Computing Service,
Professor of Computational Fluid Mechanics,
Department of Aeronautics,
South Kensington Campus,

Imperial College London, SW7 2AZ, UK
Phone: +44 (0)20 7594 5052

http://www.imperial.ac.uk/people/s.sherwin/

On 9 Oct 2020, at 02:03, 王昭予 <clemden@sjtu.edu.cn> wrote:

*******************
This email originates from outside Imperial. Do not click on links and attachments unless you recognise the sender.
If you trust the sender, add them to your safe senders list https://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address.
*******************
Dear all,

     I wonder the options of the C0Projection module. Could you please help me with it?

     First, the option 'localtoglobalmap' is cheap to perform. However, sometimes it may cause some interpolation problems, leading to a different result compared with the expansive default one. I want to know which

option is better for the precision of results?

     Second, since the value of Q-Criterion and vorticity is very high near the wall region, which may cause inconvenience when post-processing, will the option 'usexmlbcs' improve it?

     I will be very grateful with your kind reply.

Many thanks,
Zhaoyu Wang
Shanghai Jiao Tong University

_______________________________________________
Nektar-users mailing list
Nektar-users@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/nektar-users