Learning Forum

[MOUNARAHAL]

hello, i'm simulating a transient case. does that mean that the solution converged.it is written that the solution is converged at the bottom but the residuals seem that they fluctuate.any help please

[MOUNARAHAL]

to be clearer, i just want to add that the residuals in the photo downloaded are not well appeared but they are not the same even though it is written converged.As you can see in the photo for example the residual of continuity was written 1.0200e-03 and then 7.2409e-04 and for x velocity was 1.3198e-07 and then 9.6826e-08. thanks

[aitor.amatriain]

Based on the information provided, it seems that the solution is converged at each time step. However, a steady solution has not been reached, as the difference bwtween the converged solution at time t_n and the initial guess at time t_n+1 is considerable.
In RANS, if the boundary conditions do not depend on time, except from some particular problems involving flow separation (flow around a cylinder, sudden expansions in pipes...), you should wait until the peaks are over (or almost over).

[Rob]

To add to 's comment. What transient are you modelling?

[MOUNARAHAL]

Thank you all. I' m modelling a first order implicit transient.

[Rob]

What are you modelling?

[MOUNARAHAL]

I'm modelling a 2D solidification of a phase change material set initially at 20 ┬░C and the two cylinders are set as walls at -39┬░C.the phase change temperature is -30┬░C.thanks

[Rob]

OK. You may need a smaller time step to start with as the near wall fluid freezes and the step can increase (very slowly) as the solution reaches an equilibrium. Plot a contour of solid fraction every some time steps and create a monitor of solid volume (or volume fraction) to see when you reach the equilibrium condition.