I have got a small project to model something with CFX, however, i do have to find the validity of my model...some sort proves for the final result of the model i will be done. again, any one of you could help me to give any idea for a new comer to become my school project? only a simple but can get a pretty good mark.

thanks herry

The first question I would ask is: what is your particular field of interest? Testing the validity can be interpreted as comparing with something known. A good "known" is a closed form solution. Of course, closed form solutions are available for relatively simple cases. However, there are multitudes of examples out there that may be appropriate for the complexity you seek.

One particular suggestion that I would make – if you are familiar with compressible flow, look up the area Mach number relation in your text. From this relation, you can construct a convergent-divergent nozzle (a.k.a. de Laval nozzle). Once created, you can adjust back pressure for your solution to obtain subsonic and/or supersonic flow. Note the pressure ratio that you get a normal shock in the throat (a.k.a. choked flow). Then note what happens in the divergent part of the nozzle when you further reduce the back pressure. All of this can be compared to your text book discussion to give you an idea how this solver (an incompressible formulation, pressure correction method I believe) behaves with for a highly compressible, supersonic flow situations.

Hi THere,

i'll try to work it out. and i think it is quite chalanging ( although not really simple for me ). However, do you have any suggestion for the incompressible flow? thanks

herry

Again, there are many options. One would be classical Couette shear flow between parallel plates. The upper plate is moving at a specified velocity, and the lower plate is stationary. This would be a very simple case to mesh, and simple to set the boundary conditions as well. You can look up the closed form solution in a viscous flow text for the velocity (and temperature if so inclined) distribution between the plates for a specified pressure gradient. Comparing with the closed form solution would give you a good idea of the affects the turbulence model selection and near wall grid resolution have on the numerical results.

HI beachcomber, i think i'd take ur second suggestion, but i will probably try to model ur 1 suggestion. anyway thanks 2 you...big thanks

How about flow around a perpendicular flat plate ? you can get results, compare against theory, WT data ? and also there is great scope for showing how CFD can struggle to get the correct result ie Turbulence/unsteady flow ? just a suggestion ?