Mesh generation and manipulation for finite difference method usage

Resumo

This work proposes a rectangular mesh generator software which discretize complex geometries, al-
lowing computational fluids dynamics (CFD) simulations. To this porpuse, the software uses different image pro-
cessing and data analysis techniques to extract a finite set of points that describe the image contour. The resulting

coordinates represents the maximum refinement, describing each pixel from the image contour. Hence, procedures
with meshs would require high cost of memory. For this reason, the software approximate the coordinates to mesh
nodes, compliant with the mesh size selected by the user, allowing a efficient contour description with low memory
cost. Then, the approximate contour mesh is obtained, which can be used as parameters to numerical simulations
of partial differential equations using the finite difference method. Also, the software allows selection of regions

from the geometry to contain more nodes than the rest of the mesh, creating sparse meshs, resulting in better re-
finement using less memory. Finally, the obtained meshs are compared with the original coordinates by their area,

ensuring the mesh generation efficiency.