Whiteout/Remove some part of the contour plot

I use the built-in MATLAB function "scatterInterpolant" to plot a contour. The plot is shown below.

It's s stress contour and the ellipse is an empty space and there are no nodes inside the ellipse. I want the ellipse to be whiteout. Somehow I managed to set the values inside the ellipse to be zero (because I can't delete them) and I got the following result

I want the ellipse to be whiteout. Any sort of help is highly appreciated. Thanks in advance.

P.S. I use the fill command to fill the ellipse with white color but that's not a proper way to do it.

Answer 1

[Robert U]

Hi Muhammad Usman,

instead of assigning Zero to the nodes you don't want to use, apply "nan"-Values. That will be treated as if there is no value assigned and will be plotted with white color.

x = -100:100;
y = -100:100;

inputData = rand(201);

inputData(abs(x)<=10,abs(y)<=10) = nan;

contourf(x,y,inputData,'LineStyle','none');

Kind regards,

Robert

[darova]

I like that solution. It works

Answer 2

[Muhammad Usman]

Thank you very much sir for your kind response, but there is another problem I am encountering. As you can see in the figure below that it also whiteout some of the nodes outside the ellipse.

What do you think about that?

[Robert U]

contourf fills the space between gridpoints. I assume the space between a grid point with value and gridpoint without value will not be filled as well.

You can observe in your initially posted figures, that the boundary between results and zero-assigned values is not correct as well, since colormap interpolates from gridpoint value to zero which leaves you a rainbow-like contour-line outside of your ellipse.

In order to correct the issue you could snap the displayed ellipse to valid gridpoints (which will probably increase the ellipse area).

If it's just for optical reason: you can either reduce the size of your ellipse for the logical expressions by one gridpoint in each direction or assign synthetic values. Before applying any measure like this, make sure they make sense on your data.

[Muhammad Usman]

SIr again thank you for your kind response. But I cannot "snap the ellipse" because the area very close to ellipse boundary plays a very vital role in the results.

[Robert U]

If so, you might consider to provide a finer mesh.

[Muhammad Usman]

The mesh is very fine near the ellipse boundary as you can see below and refining it more will increase the computational cost abruptly.

[Robert U]

As far as I understood, you are exporting your FEM node results to Matlab. There, you apply scatteredInterpolant in order to map your original data on a (equidistant) grid that is easy to plot.

You should have a look whether your ellipse is matching the used grid for plotting. I am quite sure that this is not the case since you used an analytical description of the ellipse and plotted it overlaying the contourf plot.

The result grid for plotting could be chosen finer in order to have smaller deviation from FEM grid near the ellipse.

[Muhammad Usman]

You are right that I use "scatteredInterpolant" to map original data on a equidistant grid but I am not importing anything to Matlab. The whole code, from Pre-processing to Post-processing, is written in Matlab.

I use " DistMesh - A Simple Mesh Generator in MATLAB" to generate mesh.

You can find it here: http://persson.berkeley.edu/distmesh/

[Tamas Lanci]

Is there any chance you can share your code Muhammad?

[Muhammad Usman]

F = scatteredInterpolant(xgrid, ygrid, plotcomp, 'nearest') ;
% plotcomp contains the values of stress etc at xgrid and ygrid i.e. mesh

[X,Y] = meshgrid(); % creat equally spaced rectangular mesh

Z = F(X,Y);

% then access the nodes of [X,Y] mesh grid which are inside the hole\ellipse
Z( " nodes inside the hole " ) = nan ;