[zyzz]

The default Fluent solver solves the Navier-Stokes equations for incompressible flow. It solves the momentum equations first then solves the pressure equation separately.

If you have compressible flow (e.g., modelling the flow of a gas that experiences a significant temperature change), you need to use the coupled solver, which solves the momentum and pressure equations together.

I had some problems in my project in the last few days due to this, so I hope this helps someone someday.

Further reading:

https://www.sharcnet.ca/Software/Ans...eOverview.html

https://www.sharcnet.ca/Software/Ans...ns_scheme.html

https://www.sharcnet.ca/Software/Ans..._solve_pv.html

[LuckyTran]

No. You are right on one point but for completely wrong reasons. Please don't go around spewing such nonsense because you only contribute to people's further misunderstanding.

The default Fluent, you really mean the pressure based solver which is a segregated solver, still solves the compressible Navier-Stokes. It solves them using some algorithms which I guess you do not at all understand. There is no such thing as a pressure equation. There is no independent equation that describes the time evolution of pressure, this is the well-known pressure-velocity coupling problem.

However, the default SIMPLE algorithm uses a predictor-corrector approach to solve the momentum equation for pressure & velocity. It utilizes the continuity constraint to formulate the pressure correction equation, which is not an equation for pressure. The choice of solving momentum and a pressure correction is the algorithm itself.

But you have a choice of using the pressure-based solver with a COUPLED algorithm that solves simultaneously the pressure & velocity. However, this approach is not the same as the density-base solver which solves continutiy, momentum, and the energy equation simultaneously whereas the pressure-based solver only solves the continuity & momentum equations simultaneously but segregates the energy equation.

For steady flows, the SIMPLE algorithm is naturally unstable, and that is why it needs under-relaxation. People mistake this desirable property of SIMPLE and interpret it as "SIMPLE only works for incompressible flows and cannot be used for compressible flows". It is straightforward to use SIMPLE in a compressible flow, you just update the cell density before entering the SIMPLE part of the algorithm or update the density after you are done with the SIMPLE part. Updating the cell density really has nothing to do with SIMPLE, but it is so easy to do that people often include it in their diagrams for how their code looks. If you are able to follow up to here on how this updating the cell densities works, you should understand why the pressure-based solver is finicky.

Bottom line is, the pressure-based solver is extremely capable of solving compressible flows. At the end of the day, the solution satisfies the compressible Navier-Stokes, regardless of how it got there. However, user mileage varies. One thing that really helps, for both pressure-based and density-based solver, is to always have a good initial guess.

You are right to suggest using the density-based solver, but just because you don't know how to fly a plane does not mean that a plane cannot take you to your destination.

[zyzz]

Originally Posted by LuckyTran

No. You are right on one point but for completely wrong reasons. Please don't go around spewing such nonsense because you only contribute to people's further misunderstanding.

The default Fluent, you really mean the pressure based solver which is a segregated solver, still solves the compressible Navier-Stokes. It solves them using some algorithms which I guess you do not at all understand. There is no such thing as a pressure equation. There is no independent equation that describes the time evolution of pressure, this is the well-known pressure-velocity coupling problem.

However, the default SIMPLE algorithm uses a predictor-corrector approach to solve the momentum equation for pressure & velocity. It utilizes the continuity constraint to formulate the pressure correction equation, which is not an equation for pressure. The choice of solving momentum and a pressure correction is the algorithm itself.

But you have a choice of using the pressure-based solver with a COUPLED algorithm that solves simultaneously the pressure & velocity. However, this approach is not the same as the density-base solver which solves continutiy, momentum, and the energy equation simultaneously whereas the pressure-based solver only solves the continuity & momentum equations simultaneously but segregates the energy equation.

For steady flows, the SIMPLE algorithm is naturally unstable, and that is why it needs under-relaxation. People mistake this desirable property of SIMPLE and interpret it as "SIMPLE only works for incompressible flows and cannot be used for compressible flows". It is straightforward to use SIMPLE in a compressible flow, you just update the cell density before entering the SIMPLE part of the algorithm or update the density after you are done with the SIMPLE part. Updating the cell density really has nothing to do with SIMPLE, but it is so easy to do that people often include it in their diagrams for how their code looks. If you are able to follow up to here on how this updating the cell densities works, you should understand why the pressure-based solver is finicky.

Bottom line is, the pressure-based solver is extremely capable of solving compressible flows. At the end of the day, the solution satisfies the compressible Navier-Stokes, regardless of how it got there. However, user mileage varies. One thing that really helps, for both pressure-based and density-based solver, is to always have a good initial guess.

You are right to suggest using the density-based solver, but just because you don't know how to fly a plane does not mean that a plane cannot take you to your destination.

I see. Thank you very much for the correction LuckyTran. I apologize if I've misguided anyone.