Which method is the best for cfd applications finite volume,finite difference or spectral methods or finite element. Prasat

hi, According to my study so far, i think FV, method is better then all other for FEM, it is better for diffusion dominating problem, not that good for convecion. Although one can use FD, which give trunction error also, but it is not that flexible, and more over a mathematical nature. FV is better, most of the commercialy available software also use this method, also generaly we use same method for derivation of our flow equation..more close to fluid flow and heat transfer. So. I will suggest for FV. Look for other people what they say. bye

See previous discussion:

www.cfd-online.com/Forum/main_archive_2001.cgi?read=18678

It depends.

Each method has its own advantage. If high accuracy in simple geometry is required, Sp and FD may be the best choice. On the other hand, for complex geometry FV and FE are more convenient.

I will say definitely most popular is FVM.

Hi Friends, According to my view FV is the best method because of its basic principle of Conservation of flow properties.it is most accurate in terms of defining the physics of flow. Where as FEM fails in case of supersonic flows where there is discontinuity of the derivative(sudden change of flow properties). Also ,FD becomes very difficult when the geometry is complex due to lots of transformation involved from orthogonal to cartesian and again to orthogonal.Although programming for FD is the simplest.

Please read the book by T.J. Chung on CFD....He has elaborated the details, differences and similarities of all these methods in an understandable fashion. He also presented a lot of applications with results.

MT

hi sachu!!! nice to see you here its realy a cool place enjoy!!