Re:

[Daniel Wheeler]

On Mar 3, 2006, at 12:57 PM, ASSERIN Olivier 202603 wrote: I’m working in a lab which deals with numerical simulation of the welding processes (Thermal metallurgical residual stresses distortion, Navier-Stokes in 2D). Up to now we have used finite element method. We are now very interested in evaluating the capabilities of finite volume methods to simulate the weld pool (NS model coupled with thermics). I’ve downloaded FiPy and read your powerpoint presentation (in attachment) giving few examples of simulation in which you announce future work concerning fluid flow examples. Could you please confirm that FiPy is suitable for solving Navier-Stokes equation and further if you have already any example available. Thanks for you interest in FiPy. We have not as yet implemented any examples that solve the Navier-Stokes equations. It is not the main interest of the group we work with. We mainly use FiPy to solve phase field and non-fluid level set problems. I have in the past worked with and developed CFD codes and am aware of the complexities involved. FiPy has the basic underlying infastructure to solve Navier-Stokes. It has the traditional FV convection schemes implemented such as (Power Law, Upwind etc). It also uses a cell-centred discretization scheme developed mainly for expediency and efficiency which is common in CFD software. The missing part of FiPy as far as the Navier-Stokes is concerned is a nonlinear solver for pressure/continuity and momentum coupling. FiPy is set up to easily state and solve equations in a high level language. We pointedly do not write the non-linear solver algorithms that deal with convergence issues mainly because these algorithms are problem oriented and hard to abstract. For example CFD codes use all sorts of coupling schemes such as SIMPLE, SIMPLER, SIMPLEC, PISO etc. The point being (hopefully) is that the user can write their own script for the non-linear solution using the basic tools that FiPy provides.  Having said all that, we currently do not have many examples of this coupling in the manual. Since you emailed I worked on a rudimentary script to solve Navier-Stokes for the moving lid cavity problem. It isn't finished yet. I'll post it to the list when I have something that works properly. Cheers Daniel Wheeler ------------------------------------- Daniel Wheeler Telephone: (301) 975-8358