new user - need help
 

hi, im an engineering student from ireland. i'm using fluent for my final year project. i cant find any books on it in the library, and i cant afford to buy one. i know basically how to use it, but i could really do with some information about what way to change the default settings to make the solution converge quicker (or converge at all), or to make the results more accurate. does anyone want to help me?

sort of :

if the velocity is above 200m/s, increase the third default number in the spalart-allmaras turbulence model. (or whatever)

or does anyone know of a website where i could find such information?

the model i'm working on is an axisymmetric, incompressible (more or less) jet.

one particular thing i cant figure out. my lecturer suggested i solve the problem using a coarse mesh first, and mess around with the controls to see how to make it converge. i actually managed to get the solution to converge for the coarse mesh, than used the exact same controls for the dense mesh, but it completely diverged! any idea why? the two meshes are as similar as i can make them, except for the density.

thanks, kirsten

Re: new user - need help
 

Sorry I can't help you more specifically, but assuming you have an academic licence for Fluent, you can give a support engineer an email with your queries. That, together with any replies you get here should see you right!

Re: new user - need help
 

Hi, i can recomend one website galaxy.sdsu.edu/fluent/fluent5/tg/html/main.htm but, unfortunately it almost duplicates the manual :( As to your problem with convergence using finer mesh, try to decrease time step.

Re: new user - need help
 

Hi kristen, I think you must use compressible flow for the jet flow. your assumption is totally wrong for incompressible jet flow.

Re: new user - need help
 

As jim said, your flow should be considered compressible as flows with mach number less than 0.3 are considered incompressible, whereas in your case it is 0.6 . You have to play with your mesh and turbulence models a little bit to see if you get convergence.