CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   pressure-velocity coupling scheme (https://www.cfd-online.com/Forums/fluent/169519-pressure-velocity-coupling-scheme.html)

johnsmyth757 April 12, 2016 10:32

pressure-velocity coupling scheme
 
Hi Folks,
I'm relatively new to CFD and was hoping someone could explain to me what the differences in the pressure velocity coupling schemes are and which one would be the correct one for me to use.

Im currently trying to set up a 2d transient simulation of a vertical axis tidal turbine to analysis the performance of the turbine, however I have seen several contradicting articles in publication on what scheme to use, where some use SIMPLE, others PISO and some more the coupled solver.

should all three give correct results? and which is best in this situation as Im aware each would have there strengths and weaknesses

Thanks,
John

LuckyTran April 13, 2016 02:15

I will not try to explain the differences in the schemes, it is much better to consult the Fluent manual or the original publications if you really want to know the details.

In short, there are many ways to solve the pressure-velocity coupling problem, either by brute force (the coupled scheme) or via some predictor-corrector techniques. These are algorithms to solve a certain problem. Think of it as different matrix inversion algorithms, or different root finding algorithms. The different literature do not necessarily contradict one another. They are all applicable.

You should also look into the density-based solver (and compare it to the pressure-based solver). SIMPLE/PISO/COUPLED schemes are only for the pressure-based solver, since that problem does not arise in the density-based solver.

All three (four if you count the density-based solver) should (in principle) give the same results. But for very obscure reasons, they often don't give exactly the same results. However, you should find that they are very close. In theory, the COUPLED scheme should give the most accurate results (since it basically is brute force). However, the COUPLED solver is by far the most expensive since it computationally inefficient.

PISO & COUPLED scheme generally is more aggressive and converge faster than SIMPLE per iteration cycle, but SIMPLE is computationally much quicker and the overall computational cost with SIMPLE can allow you to perform more iterations faster than you would be able to with PISO or COUPLED. PISO and COUPLED are generally preferred for transient simulations (because of their aggressive convergence behavior). Sometimes, in LES for example, you may switch back to SIMPLE because its faster and you make up for the slower convergence by calculating more iterations.

The choice between SIMPLE/PISO/COUPLED is really a choice of computational efficiency. One time where COUPLED can be vastly different than SIMPLE/PISO is with highly nonlinear (pressure & temperature dependent) property changes since the equations are fully coupled, there is no biasing of the solution.

johnsmyth757 April 13, 2016 07:11

2 Attachment(s)
Thanks Glenn for your detailed reply appreciated,

one more bit of advice would be really appreciated, I have attached a screenshot of my solution method setting and would appreciate your opinion on for my setup, all publish journal articles I have come across recommends using 2nd order and the manual recommends using the presto solver for swirling flows and fans which is applicable in my 2d turbine case.

Would you agree with these settings? I have also attached a snapshot of my mesh for reference and Im using the transition SST turbulence model.

Thanks in advance,
John

LuckyTran April 13, 2016 09:03

Yes, always use 2nd order unless you run into trouble and need to resort to 1st order.

I tend to prefer second order for pressure, but PRESTO is also a good choice.

sachin tom March 27, 2021 08:36

Which is better P-V coupling for transient flow boiling (time step size ~ 10^-4s) simulations?
Is it the phase coupled SIMPLE or coupled?

Regards,
Sachin


All times are GMT -4. The time now is 13:20.