|
[Sponsors] |
December 14, 2016, 18:07 |
Second order and PRESTO
|
#1 |
Senior Member
SinaJ
Join Date: Nov 2009
Posts: 136
Rep Power: 16 |
I get a strange result. When I use Second Order pressure scheme, I get a very good residual convergence, but Streamlines don't look fine. and when I use PRESTO! I have non-converged residuals (converged up to -3 or -4), but streamlines make much more sense. Does anybody have any explanation for this?
Thank you |
|
December 15, 2016, 02:49 |
|
#2 |
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,746
Rep Power: 66 |
Residuals are not a measure of accuracy nor are they measures of convergence.
PRESTO! solves for the pressure on cell faces whereas Standard / Second Order interpolates them. PRESTO! gives a more physically accurate representation of the pressure field. Interpolation naturally smooths the pressure field and can give better numerical stability and is probably the reason for better residuals. |
|
December 15, 2016, 18:59 |
|
#3 |
Senior Member
SinaJ
Join Date: Nov 2009
Posts: 136
Rep Power: 16 |
Thank you for your answer.
Then how should I get a more converged (based on residuals) solution with PRESTO!? In what I get, velocities, epsilon and k are fluctuating around <10e-3 , and continuity is around <1e-5. I have 4 points monitoring, and seems that velocities are not changing at those points. but residuals doesn't get smaller! Any idea? |
|
December 16, 2016, 07:00 |
|
#4 |
Senior Member
Kevin
Join Date: Dec 2016
Posts: 138
Rep Power: 9 |
1e-3 is actually pretty good as a residual. It means your solution improved by three orders of magnitude based on the initial conditions. As said earlier, don't base yourself on the residuals to judge proper convergence. Residuals only indicate how much your solution improved compared with the initial conditions. If your initial conditions are already really good, your residuals could be only 0.5. That doesn't mean your results are bad. Look at your monitor points, like the velocity points you have, and see if they still change or not. If not, it could indicate that the solution has converged (they could change if you simulate for longer). Now, whether the results are correct is a whole other story. That's up to you to determine, comparing with experimental/analytical results or other numerical results.
|
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
can you tell me best gradient, pressure & momentum order selection in fluent | sanjiiv | FLUENT | 6 | February 14, 2020 07:07 |
PRESTO! Order of Accuracy | dbsnu | FLUENT | 1 | December 4, 2013 15:24 |
2nd order schemes only 1st order accurate? | ThinkQuick | Main CFD Forum | 3 | September 11, 2012 06:16 |
2nd order and Turbulent-viscosity ratio error | freekeer | FLUENT | 3 | October 5, 2010 06:41 |
Hi, my name is Julie and need help with Fluent | Julie Blume | FLUENT | 9 | October 26, 2006 15:23 |