Hi folks,

I'd like to ask you any idea about generating an unstructured grid from a bathymetry data. The data is in the form of x, y regular coordinates and z (depth). I thought about turning the quadrangular regular data into triangles so as I could use any mesher to build the volume data (of course, I'd have to build the boundaries first). My main problem is that my bathymetry data does not represent a regular (square) bounded area.

Does anyone know any software that could help me in this task? I looked for mesh generators specifically developed for this kind of problem but I haven't found nothing yet.

Thanks for any help

Renato.

If you can get the triangles in STL (Points2Poly, for example) ICEM CFD would be a good choice. You can easily build the other boundaries as a box - then trim the STL triangles with the flat box boundaries.

Hi Myron, thanks for the hint... in fact I already tried to use points2poly to create a triangulated model but my version of the points2poly does not allow me to save the resulting STL file. I think it will not be hard to create the tetrahedra model after building the boundaries.

Regards

Renato.

I am not sure that do i understand your problem, but i give some general comments:

as i feel you have a surface point related to floor of ocean and objective is generation of water-tight 3d-manifold, e.g. as triangulated surface (is it?)

this subject was/are an extensive research area in computer graphics and computational geometries and is well-documented, there are several methods. But i think you look for crude cure and need a software to do this as black box, so:

- You can search about methods for triangulation of point cloud or ..., surly you find several work and their related opensource, e.g. see this: http://www.maths.dundee.ac.uk/~heikm...ngulation.html

main difficulty dealing with this problem is generation of health and water-tight 3d surface, so if you search about this subject they are connection with your case, i face this previously when look for method for handling dirty-deometries and now, mention that i see several methods/packages that can resolve your problem.

- note that it is better to add points related to other surfaces e.g. horzontal ocean surface and ... so your point cloud give implicitely 3d surface.

- VTK lib has a point cluod data structure support, i don't test but think that it has delaney tessellation of such data too.

- alos you can add other nodes and generate 3 volumetric point cloud and use delaney tetrahedrization of them.

Hope this help you.

I believe Rhino3d can do this job for you.

Hi folks,

sorry for my late response... (I was held in other urgent things).

The problem of building the bathymetry mesh was worked out. I asked the guy, who sent me the data, to put the grid in such a way that I could easily read and produce the triangular surface (of course he could do it much easier than me since he has the proper software). After this phase, I think, It will not be to hard to build a tetrahedra mesh with the meshers that I have here (IcemCFD, NetGen, TetGen, Ansys/Prep7).

anyway I'd like to thank all contributors.

Regards

Renato.

Out of curiosity,

what is application of such volumetric mesh?

Gravity currents like turbidity flows.

But, what is application of "Gravity currents" simulation?

any kind of horizontal flow driven by gravity influence like when you have two fluids with difference in their density. This kind of flow develops in several natural phenomena like avalanches, piroclastic plumes, turbidity reservoir formation, dense gases spreading, etc...

Thanks for your general info.

but my interest was to knowe what is application of simulation of gravity driven flow in sea, especially considering accurace geometry of sea floor, it seems to be a large scale simulation, i can't realize what can be its application and this is just my question.

We're trying to compute how far a load of sediments will reach after being released by a river. This process explain how some oil reservoir were formed in the past. Of course, there are much more details about the simulation to be considered, things like: is it feasible to consider a Boussinesq approach for the range of fluid densities considered? What kind of sedimentation model should be used? How the turbulence effects must be taken, etc..., etc..., etc... . In my case, I'm just starting to test our code regarding such concerns.

>> it seems to be a large scale simulation...

yes, can be... it'll depend on the desired accuracy and the level of details that you're describing, but this problem can be treated in several scales with different concerns, for example, in large scale it's important to determine the distance travelled by a load of sediments.

I hope it clarifies your doubts.

>We're trying to compute how far a load of sediments will reach after being released by a river. This process explain >how some oil reservoir were formed in the past

Informative material.

Thanks for your clarification.