About high order non-traditional Lagrange Finite Element

[cfdbooks]

Does paraview support high order non-traditional Lagrange Finite Element? For example, no solution are saved on vertices, or edge points, or face points, solutions are saved on volume Lagrange points only(All control points are in the volume). Thanks in advance.

[vrkssai]

@cfdbooks Welcome to the community. It supports lagrange type of dataset at lobatto points.

For example, no solution are saved on vertices, or edge points, or face points, solutions are saved on volume Lagrange points only(All control points are in the volume)

You would have to interpolate your data to these points and then plot in Lagrange Hexahedron cell type. I’m assuming you have some data at like Legendre points, make it Lobatto by interpolating from those Legendre points and then write your data.

[cfdbooks]

Thanks. Is there any reference on high order Lagrange Hexahedron cell, especially the points order?

[vrkssai]

It was actually really difficult to get the point order, but you could check out this post below in the comment section.

Arbitary Order Lagrange Hexahedron Element Connectivity

The new versions of ParaView(5.5 and above) supports arbitary order lagrange hexahedron elements. The ordering of the points (lagrange points facepoints+interior+corners)associated with the elements hasn’t been provided either in the ParaView reference guide or in VTK file format specification in VTK guide. (For simple hex the ordering of nodes is clockwise from inside the cell for one face and similar for the nodes in the opp face ). How do we figure such ordering or pattern for the Arbitary Or…

Basically from the comments there is a link to blog post, in the blog post’s comments section you get an idea where the ordering, you’ll have to go through the C++ code for that at this link:
https://gitlab.kitware.com/nick.laurenson/vtk/blob/d8aa5b89c622cf04b3322112fda420afc9d2a16d/Common/DataModel/vtkLagrangeHexahedron.cxx

search for int vtkLagrangeHexahedron::PointIndexFromIJK(int i, int j, int k, const int* order)

Giving you a headsup, there are a few bugs in the viewer like below:

Segmentation fault while Visualizing/Creating contours,slices.

ParaView Support

Segmentation fault while Visualizing/Creating contours,slices.

Loading a 2x2x2 dimensions lagrange hexahedron element type of order 1 and trying to visualize the slices, clipping or contour leads to segmentation fault. I suspected corrupted vtk file so I created a sample element from , sources>alphabetical>unstructured grid types; Properties: dimensions- 2x2x2, lagrange hexahedron,cell order-1. I chose the default location of clipping or slicing and the Paraview closes with segmentation fault. Anyone able to reproduce this and check? This seems to be a ma…

Vorticity and Gradient python calculator calculation(gives wrong result-Lagrange hexahedron)

Vorticity calculated for a smooth vortex function using python calculator as curl(Velocity) gives wrong result using the Legrange Hexahedron element. I tried to plot the same in tecplot using a point data and traingulation, the result looks fas it should on tecplot. I used the same number of levels,range,etc. [tecplotvortex] [paraviewvortex] As you can see from the figure for paraview plot the vorticity doesn’t vary in the cell for the paraview plot as it should(since its lagrange hexahedron), …

Just out of curiosity may I know what you are working with or on?

[cfdbooks]

Thanks. I am working on Flux Reconstruction method of CFD.

[R-e-M]

Hello !
I’m also working on high order element, only testing with quad for the moment.
When I try to create Lagrange Quadrilaterals from the GUI with the “Unstructured Cell Types” tool in the v5.6, I only get high order cells with uniform point spacing.
@vrkssai : You talk about Lobatto points, is there something I missed in order to get this spacing ?
Thank you !