|
[Sponsors] |
Difference between solve() and solve(...finalInnerIter())? |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
June 8, 2016, 06:01 |
Difference between solve() and solve(...finalInnerIter())?
|
#1 |
Member
Lennart
Join Date: Feb 2016
Posts: 46
Rep Power: 10 |
Hello,
can you help me understand the difference between a simple call to Code:
p_rghEqn.solve(); Code:
p_rghEqn.solve(mesh.solver(p_rgh.select(pimple.finalInnerIter()))); Your help is much appreciated! |
|
June 8, 2016, 17:59 |
|
#2 |
Senior Member
Joachim Herb
Join Date: Sep 2010
Posts: 650
Rep Power: 21 |
The second version uses the settings for p_rghFinal in the system/fvSolutions file (for solver settings) for the last iteration in the pimple outer loop (so for the last iteration step per time step) whereas the first version *always* use the p_rgh settings. So the p_rghFinal settings are ignore in the first case.
|
|
June 9, 2016, 03:56 |
|
#3 |
Member
Lennart
Join Date: Feb 2016
Posts: 46
Rep Power: 10 |
Thanks for the quick and clear answer!
My settings for p_rgh and p_rghFinal are the same, except for the relTol, which is 0.05 and 0, respectively. So naturally, the solver needs more iterations in the last outer pimple loop because it can't stop at any relative tolerance. I was wondering if it's normal that THAT much more iterations are required: Code:
DICPCG: Solving for p_rgh, Initial residual = 0.0509331, Final residual = 0.00187709, No Iterations 3 time step continuity errors : sum local = 9.11852e-06, global = 7.5083e-06, cumulative = -0.00101879 DICPCG: Solving for p_rgh, Initial residual = 0.00738722, Final residual = 0.000365238, No Iterations 80 time step continuity errors : sum local = 7.85558e-06, global = 7.48801e-06, cumulative = -0.0010113 DICPCG: Solving for p_rgh, Initial residual = 0.00345672, Final residual = 9.48528e-08, No Iterations 250 time step continuity errors : sum local = 7.50304e-06, global = 7.50295e-06, cumulative = -0.0010038 |
|
June 9, 2016, 04:30 |
|
#4 |
Senior Member
Joachim Herb
Join Date: Sep 2010
Posts: 650
Rep Power: 21 |
It would say it is normal.
|
|
June 9, 2016, 04:40 |
|
#5 | |
Senior Member
Anton Kidess
Join Date: May 2009
Location: Germany
Posts: 1,377
Rep Power: 29 |
Just look at your residuals. At first you are reducing them by 10, then in the final iteration 10'000x. Of course that will require a lot more iterations. By the way, GAMG typically shows much better performance for the solution of the pressure equation.
Quote:
__________________
*On twitter @akidTwit *Spend as much time formulating your questions as you expect people to spend on their answer. |
||
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
SIMPLE Algorithm Finite Difference Equations: how to discretize and solve? | DA6righthand | Main CFD Forum | 0 | August 3, 2015 12:12 |
Can OpenFoam solve this problem? | salazardetroya | OpenFOAM Running, Solving & CFD | 1 | July 29, 2015 22:34 |
Additional 2D solve at an inlet patch during each iteration of a usual 3D solve | incompressible | OpenFOAM Running, Solving & CFD | 1 | July 5, 2015 11:09 |
solve the scalar gradient | yp5547 | FLUENT | 0 | December 24, 2014 05:05 |
Difference between codes and representations | titio | OpenFOAM Running, Solving & CFD | 3 | July 3, 2009 07:58 |