[yaseen wsu]

my project does not reach the convergence, why? how can help me?
(my project is free surface flow in spillway)

http://www.cfd-online.com/Forums/att...1&d=1445016778

[Opaque]

Your convergence looks fine for the standard defaults, RMS = 1.E-4, correct ?

What is the problem ?

[yaseen wsu]

no, it does not show me any results.

[hamidciv]

i dont underatsnd your mean, please explain more for solve problem.

[yaseen wsu]

dear hamidciv the above attachment is the first iteration, which is finished run said (run completely normally ), after I did second iteration with finer mesh also the results is not acceptable (please see the attachments for second iteration) , thanks
http://www.cfd-online.com/Forums/att...1&d=1445078680

http://www.cfd-online.com/Forums/att...1&d=1445078688

http://www.cfd-online.com/Forums/att...1&d=1445078698

http://www.cfd-online.com/Forums/att...1&d=1445078707

http://www.cfd-online.com/Forums/att...1&d=1445078715

[hamidciv]

dear yeaseen wsu
if results is not acceptable, probably problem is induce of your mesh, i think your mesh is coarse.
whether you control mesh charactristics?
whether you doing verify in your model?
whether your time is Enough for To reach steady state flow?
are your boundary condition is suitable?

[yaseen wsu]

thanks a lot hamidciv
yes I have problem in mesh generation, but I sure my boundary condition is ok, what is time? time scale control, I choice physical time scale (1 second)

[hamidciv]

dear yaseen , I thought you do transient simulation, my mean of time was enough time for reach to steday state in transient simulation.

[yaseen wsu]

no I did stead state simulation, but I have problem how to select appropriate time is there any law?

[Antanas]

You geometry is fairly simple, hexa mesh aligned with the flow will give you more accurate results.

[hamidciv]

dear yaseen
as antanas mentioned, i also suggested doing mesh with hexa mesh, becasue it decrease number of cells and give same accurate in compared tetra mesh.
about physical time scale , it better you read cfx help, in my order your time is large.
good luck

[Antanas]

Originally Posted by hamidciv

dear yaseen
as antanas mentioned, i also suggested doing mesh with hexa mesh, becasue it decrease number of cells and give same accurate in compared tetra mesh.
about physical time scale , it better you read cfx help, in my order your time is large.
good luck

I think you're wrong. In fact in Steady State simulation you should use the largest time step size which still gives you convergence (in terms of residuals, imbalances and quantities of interest), because converged steady solution doesn't depend on time step size. If converged solution doesn't satisfy your expectations, then you should check other things, like physical models that you use, boundary conditions, mesh independence and so on.

[yaseen wsu]

thanks a lot Antanas and hamidciv for your suggestion

[Bazinga]

If you are worried about the convergence of your simulation you can additionally add monitor points in the flow and observe during calculation if the flow variables converge with the rms residuals at 1e-04. If not, lower the residual target.

[yaseen wsu]

dear Bazinga how to add monitor points in the flow?
thanks

[Bazinga]

Check Pre-Processor, Simulation-Flow Analysis-Solver-Output Control-Monitor.

There you are able to add Monitor Points at the different positions in the flow to observe for example the velocity and check convergence.