During my thesis I developped a way to make Finite Differences on Unstructured grid. The problem is that I found very few references on that subject. If anyone has some information on that topic, it would be very helpfull to me. THANKS IN ADVANCE.

Can you describe your Finite Differences on Unstructure grids. I have experience about Finite Volume Method on Unstructured grids. Most used Finite Volume Methods on Unstructured grids are two types: 1. time marching( e.g. Jameson) 2. pressure based method(e.g. Patankar,Baliga)

What is the difference of your method from these?

Contrary to Finite Volume Methods UFD doesn't use control volume nor integration of the equations. Like in structured grid, the starting point is the Taylor series expansions ; but here, the Taylor series are made over the space on points that may not be on line coordinates. Thus, any kind of partial differential equations can be discretized, like Navier-Stokes equations, heat transfert or Maxwell 's equations. Then, UDF are not particularly time marching related or pressure based.

you conducted an interested subject, I didn't know any refernces in this subject, hopefully after you finish your thesis, I can obtain one copy of it.

wang

Hi,

You definitely have an interesting academic topic to study on. As you know, there aren't many people (if any) doing finite difference on unstructured grid.

There is, however, one very siminar idea in a paper published by Batina of NASA Ames several years back (92-93) on a gridless method. The method used a cloud of points to form the partial differential, which is idential to your idea in a sense. The paper was published as an AIAA paper. You probably could find it by searching the name of the author.

good luck

ZJ Wang CFD Research Corp.