|
[Sponsors] |
March 3, 2017, 06:31 |
Simple question : Simple Vs Coupled solver
|
#1 |
New Member
Maria Angela - Roma
Join Date: Mar 2017
Location: italia - Roma
Posts: 2
Rep Power: 0 |
Hello,
i am working on a simulation around an airfoil at very high Rey, the Simple solver doesn't converge even i decreased the under-relaxation factors but the coupled V-P works. my question is why ? Thank you for your help & support Maria |
|
March 4, 2017, 14:10 |
|
#2 |
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,764
Rep Power: 66 |
Is it diverging or what do you mean by converged?
|
|
March 6, 2017, 12:44 |
Response
|
#3 |
New Member
Maria Angela - Roma
Join Date: Mar 2017
Location: italia - Roma
Posts: 2
Rep Power: 0 |
Hey first of all thank you,
Well the idea is that i am runing a same mesh with SIMPLE & COUPLED solvers : - The simple solver diverged. - The coupled solver converged. i just wanted to know why ? |
|
March 6, 2020, 08:05 |
|
#4 |
New Member
Elif Gündoğdu
Join Date: Mar 2020
Posts: 2
Rep Power: 0 |
Hi, have you figured it out yet? My solutions were converged too when I used coupled solver. I just can't understand the reason
|
|
March 6, 2020, 08:29 |
Convergence
|
#5 |
Senior Member
|
SIMPLE and Coupled are just two numerical approaches. Whichever works is good. The reason why one works and other not is more of physical nature. If pressure and velocity are very tightly coupled in a scenario, such as in rotating machines, then Coupled works better. If that is not the case, SIMPLE is alright. Coupled is almost always better than SIMPLE but there are a few scenarios where SIMPLE works better, such as for low Gr number Natural convection flows.
__________________
Regards, Vinerm PM to be used if and only if you do not want something to be shared publicly. PM is considered to be of the least priority. |
|
March 6, 2020, 16:38 |
|
#6 |
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,764
Rep Power: 66 |
Well there are non-physical reasons as well.
Always first you need to state your entire model. It's hard to say why your case diverged when you don't state what you're doing. The P-V coupling solver is merely a means of solving one part of the entire problem, namely how to solve the momentum equation for pressure and velocity. Your solution can fail to converge due to problems with coupling from other equations, i.e. the energy equation. If you are not running any energy equation at all, which means you're doing an incompressible simulation, then we can actually blame it from the P-V algorithm. SIMPLE solves a pressure correction problem and uses this to update the velocity field. Once it does this, the pressure field is in an inconsistent state with the velocity field. You are guaranteed to have the wrong pressure field, every time until your solution converges. If you carry this wrong pressure field into another equation (e.g. the energy equation) then you're done-zo. In SIMPLE you rely on iterating many times to and pray that these inconsistencies get smaller. Next, SIMPLE doesn't use the non-orthogonal correctors and this tends to make it even more unstable for non perfect meshes. That's why the default urf's for SIMPLE are as low as they are. The bottomline is, SIMPLE is not designed to be stable, it's designed to be fast by exploiting weak coupling. For strong coupling, SIMPLE will be smelly. For bad meshes, SIMPLE will stink. For bad initial guesses, SIMPLE will really really stink. |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Question about matching of solver and turbulence model | louistse | OpenFOAM Running, Solving & CFD | 1 | February 1, 2017 22:36 |
pressure-based coupled solver for compressible NS equation | wangmianzhi | Main CFD Forum | 19 | July 29, 2016 04:37 |
Question about Implicit Unsteady Solver and Maximum Inner Iterations | Awesomo | STAR-CCM+ | 4 | June 28, 2016 09:57 |
Convergence with coupled implicit solver | Henrik Ström | FLUENT | 1 | October 29, 2005 04:57 |
Segregated or coupled solver? | Nori | FLUENT | 0 | September 29, 2005 15:24 |