Voronoi grid generation

[Groundwater-modeler]

Hi,

I have been following the examples in the flopy tutorial section for generating voronoi grids.
I have then cooked up some code to generate the grid based on shapefiles rather than coordinates.

I’m not quite satisfied with the end result and feel that its a bit coarse to assign a point inside a polygon to define the mesh resolution.
Are there any guides on how to construct a voronoi grid based on any number and type of shapefiles while also being able to refine them around the shapefiles?

best regards

[Groundwater-modeler]

in addition to my previous question. If you consider the example "Regions with Different Refinement and Hole", you notice that there is a part where refinement along a line is added. This part of the script seems to be rather case specific and not something that can handle a generic refinement of line objects. From what it seems it works only because the line is completely straight. should it deviate from this then the vertices will be interpreted as bounds of a polygon since the argument passed is "tri.add_polygon(line)".
This makes the refinement not very useful.

Could there perhaps be added some more functionality to the flopy.utils.voronoi module so that one have more options of refinement around specific objects. preferably with shapefile compatibility.

Best regards

[christianlangevin]

It's true that the Voronoi generation, and the triangle mesh generation upon which it is based, is pretty limited in Flopy. The capabilities available now allow us to create relatively simple meshes and use unstructured grids with MODFLOW-USG and MODFLOW 6. It would be great to have more robust mesh generation capabilities. In particular, I would like to see implementation of Lloyd's algorithm for improving Voronoi grids. We recognize that there is a need for additional work on this.

[Groundwater-modeler]

Right! Do you think this is something that we can expect any time som or is it for a more distant future?
Would be great with some more meshing capabilities!

[hatarilabs]

Have you tried mf6Voronoi that is a Python packages that couples with Flopy to create awesome models. It has templates as well and more capabilities to create 3D Vtk with even the water table surface. https://github.com/hatarilabs/mf6Voronoi And this is a tutorial: https://hatarilabs.com/ih-en/regional-groundwater-flow-modeling-with-voronoi-mesh-on-modflow6-disv-and-flopy-tutorial Christian, did you get a chance to look at the Fortran error on Modflow Disv? El dom, 13 abr 2025 a la(s) 11:04 a.m., David ***@***.***) escribió:

…

Right! Do you think this is something that we can expect any time som or is it for a more distant future? Would be great with some more meshing capabilities! — Reply to this email directly, view it on GitHub <#2487 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHEUBVFYZWUQBHU5VAX3OBL2ZKDHJAVCNFSM6AAAAAB24OTMAKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEOBSGA3DAOA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Groundwater-modeler]

Interesting, but does it stay true to the requirements for the cvdf formulation to be accurate?
is it based on Delauny triangulation?
Is mesh quality considered?

[hatarilabs]

Sure, We have worked intensively on the point cloud generation of the Voronoi grid. The process is fully geospatial and has options for overlapping and distance multiplier. This an example of the geospatial grid: [image: image.png] Viva el software libre! Saul El jue, 17 abr 2025 a la(s) 9:38 a.m., David ***@***.***) escribió:

…

Interesting, but does it stay true to voronoi conditions and keep angels lower than 30 etc? — Reply to this email directly, view it on GitHub <#2487 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHEUBVHZ5KRIDPWT4AEPRGT2Z64EXAVCNFSM6AAAAAB24OTMAKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEOBWHA2DAOI> . You are receiving this because you commented.Message ID: ***@***.***>