[yaseen wsu]

hello
the following images are my model, after 1000 iteration running, the problem is the volume fraction for entire model is not appear (I used volume rendering in CFX post) what is reasons of this???
http://www.cfd-online.com/Forums/att...1&d=1455862066
http://www.cfd-online.com/Forums/att...1&d=1455862080
http://www.cfd-online.com/Forums/att...1&d=1455862099

[ghorrocks]

FAQ: http://www.cfd-online.com/Wiki/Ansys...gence_criteria

If those tips do not help then please post your CCL (not screen shots of CFX-Pre, please).

[yaseen wsu]

Originally Posted by ghorrocks

FAQ: http://www.cfd-online.com/Wiki/Ansys...gence_criteria

If those tips do not help then please post your CCL (not screen shots of CFX-Pre, please).

thanks those tips were usefull, I will try to apply all in my model. finally I will believe my mode should run in transient, the results is good but all equations in solver has no same ratio, only close to each other, what is the reason of this problem? I used 10 sec. as total time, 0.06 sec time step and the Courant Nomber at first iteration is 0.9 but at last iteration RMS Courant No. become 145 and Max. become 999.99 is it Courant No. related to fully convergence?

[ghorrocks]

If it is converging quickly and rapidly despite the high Courant number you should be fine.

[yaseen wsu]

no, I said does not converged, I selected time step 0.06 s according to mesh size 0.006 mm (min) and inflow velocity = 0.03m/s, but I dont know velocity in the model entirely, so how can reduced? Courant No, or Courant No. is a reason to the simulation become non converged? thanks

[ghorrocks]

I selected time step 0.06 s according to mesh size 0.006 mm (min)

And on what basis does that mean a 0.06s time step is appropriate?

My recommendation for time step size is to use adaptive time steps homing in on 3-5 coeff loops per iteration. Then the solver finds its own time step.

[yaseen wsu]

thanks for your recommend, based on Courant No.between (2 - 10)

[ghorrocks]

Unless you have shown that a Courant number of 2-10 is suitable for your simulation then it is meaningless. CFX is an implicit solver and does not have a easily definable stability limit like explicit solvers do. So I recommend you do the adaptive time steps homing in on 3-5 coeff loops per iteration instead.

[yaseen wsu]

Originally Posted by ghorrocks

Unless you have shown that a Courant number of 2-10 is suitable for your simulation then it is meaningless. CFX is an implicit solver and does not have a easily definable stability limit like explicit solvers do. So I recommend you do the adaptive time steps homing in on 3-5 coeff loops per iteration instead.

I used transient with adaptive time step but courant number at five time step is high (RMS = 40, MAX 999), ass follow
Time steps option - adaptive
first update time - 1s
update frequency - 1
initial time step -0.03 sec
min time step 0.00001 s
max time step 5 s
target min loops 3
target max loops 5
time step decrease factor 0.8
time step increase factor 1.06

[ghorrocks]

If that is converging every time step then it should be fine. Or is it not converging?

[yaseen wsu]

Originally Posted by ghorrocks

If that is converging every time step then it should be fine. Or is it not converging?

the rates are not same as for all equations only close to each other and RMS are near to the target (1e-5)
note: I choose the initial time step not randomly, but by a number trials and calculation like sensitivity analysis, in my view my mesh is also good
in most time steps equations of velocity (U, V, W) shows (ok not OK), I think this relates to inlet velocity B.C, as in image show i have velocity in (x,z) but for y axis (vertical) is zero. so I enter (U=0.03m/s, V=0, W=0.03m/s) is it x=U, y=V , z=W ? or I should enter value for (U & V) and W=0
thanks
http://www.cfd-online.com/Forums/att...1&d=1458571969

[ghorrocks]

You have done a sensitivity check on time step - good.

But you should also do one on convergence toleration and mesh size. Your convergence tolerance may be tighter (or looser) than is required, and your mesh may be too coarse or too fine. In this case, as you appear to be using a transient simulation to find the steady state answer you can probably loosen the convergence tolerance - this will help things a lot but you will need to check this before doing it.

[yaseen wsu]

thanks, I know how to done senstivity analysis for mesh size, but I dont know sensitivity analysis for convergence toleration, how to done?
thanks

[hamidciv]

Originally Posted by yaseen wsu

thanks, I know how to done senstivity analysis for mesh size, but I dont know sensitivity analysis for convergence toleration, how to done?
thanks

dear yaseen, Please refer to the link below.
http://www.cfd-online.com/Forums/cfx...imbalance.html
hope it helps.

[yaseen wsu]

thanks a lot