How to use contourf to plot a mesh not generated in matlab, i.e. imported mesh coordinates from abaqus. The issue is that coordinates are not sequential.

I'm trying to use contourf with x and y coordinates being random or non sequential. I have z value for each X y coordinate. Normaly we would create a rectangular mesh and then fill up the z values. Here the issue is that the geometry is meshed in abaqus and the mesh nodes and element related information is fed to my code. Now the nodes are not sequential matrix.

Answer 1

[Star Strider]

It would help tto have your data.

Since there are (x,y,z) coordinates (there can be duplicates), I would use the scatteredInterpolant function to create an interpolant, and then use either ndgrid or meshgrid tto create the relevant argument matrices using the ‘x’and ‘y’.vectors to define them.

There are several examples on MATLAB Answers using that approach, that create column vectors for ‘x’, ‘y’ and ‘z’ to use with scatteredInterpolant,.and then generate the ‘z’ matrix from the argument matrices. Then use surf or surfc to plot them.

.

[Subham Prasad]

I have attached the data points. First two colums are x and y coordinates. Third column represents z values for each coordinate. I have also attached the meshed geometry that was imported from Abaqus.

[Subham Prasad]

I am looking for a similar function like trisurf like function where i can give the nodal coordinates and node element connectivity along with the z values for colourmap. Because the meshing is 4 noded quad element.

[Star Strider]

This sort of works, however it takes 605.3 seconds to complete and about as much time to render the surface plots. It fails, because it fills (interpolates) the voids areas. I do not see any way to correct for that.

imshow(imread('boundary conditions.jpg'))

LD = load('plotdata.mat');
plot = LD.plot;
x = plot(:,1);
y = plot(:,2);
z = plot(:,3);

figure
scatter3(x, y, z, 5, z, 'filled')
colormap(turbo)

% return
tic
xv = linspace(min(x), max(x), size(plot,1));
yv = linspace(min(y), max(y), size(plot,1));
F = scatteredInterpolant(x, y, z);
[X,Y] = ndgrid(xv, yv);
Z = F(X,Y);

figure
surfc(X, Y, Z, 'EdgeColor','none')

T = delaunay(X, Y);

figure
trisurf(T, X, Y, Z, 'EdgeColor','k', 'LineWidth',0.2)

toc

This also ended up crashing my computer.

.

[Subham Prasad]

Thank you for the answer. The scatter plot (2D) will do for now. I think contourf kind of plot is not possible in this case due to the randomness of data. Had it been a sequential set of data points it would have worked. Do let me know if you find any way to get contour plot kind of image. I am attaching an example of how the plot should have looked and also one image to show how it looks with scatter plot (2D).

Answer 2

[Walter Roberson]

Perhaps try

https://www.mathworks.com/matlabcentral/fileexchange/10408-contours-for-triangular-grids

https://www.mathworks.com/matlabcentral/fileexchange/38858-contour-plot-for-scattered-data

https://www.mathworks.com/matlabcentral/fileexchange/40847-tricontf

https://www.mathworks.com/matlabcentral/fileexchange/72090-tricont

https://www.mathworks.com/matlabcentral/fileexchange/78901-discrete-contour-mesh-patch-2d-3d

Also, possibly it would help to use https://www.mathworks.com/matlabcentral/fileexchange/27830-mesh2tetra

(List order is numeric)

[Subham Prasad]

I have used 4 noded quad element for meshing. The nodal coordinates are in that way.

[Subham Prasad]

These are good for triangular elements, is there any similar funtion for quad element?