[Bdew8556]

Hi guys,

Just a general question. I've heard CFX can't handle a true 2D mesh. Anyone know the technical reason?

Brett

[aero_head]

I believe it is just because of the coding of the solver; it is not coded to handle a 2D mesh, it is naturally a 3D solver. The fact that Fluent can do it might be why they have not implemented it.

[Bdew8556]

I know that FLuent is a centred based code, whereas CFX is node based. Could that have something to do with it?

[flotus1]

I don't think it's a technical reason. Developing a true 2D solver that 99% of your paying customers don't need -alongside the 3D solver- just isn't worth it. Especially when you have a similar solver that can handle true 2D.

[LuckyTran]

CFX (Flow3D in the UK) has been 3D only since the 1980s. Hardly any of the commercial CFD software developed since has supported a truly 2D environment (Star-CCM doesn't, OpenFOAM doesn't). These software straight cannot import a 2D mesh. And they have no reason to.

Fluent was 2D and then added 3D whereas most of thee other software skipped that development phase entirely. Fluent continues to support the 2D environment, I don't really see why. It's really hard to even find a 2D mesher nowadays. Gambit is defunct. GridPro is honestly the only 2D mesher I use.

It's like asking why you cannot put a 12VDC lead-acid battery in a Tesla or a Boeing 787, they were never designed for it.

[arjun]

Originally Posted by LuckyTran

CFX (Flow3D in the UK) has been 3D only since the 1980s. Hardly any of the commercial CFD software developed since has supported a truly 2D environment (Star-CCM doesn't, OpenFOAM doesn't). These software straight cannot import a 2D mesh. And they have no reason to.

Fluent was 2D and then added 3D whereas most of thee other software skipped that development phase entirely. Fluent continues to support the 2D environment, I don't really see why. It's really hard to even find a 2D mesher nowadays. Gambit is defunct. GridPro is honestly the only 2D mesher I use.

It's like asking why you cannot put a 12VDC lead-acid battery in a Tesla or a Boeing 787, they were never designed for it.

Very much true. Its not worth the efforts.

Wildkatze like fluent also support true 2d (that is only x y mesh and not cell cell width mesh) but we did not put efforts to support unmatched interface with it because usage is so low.

(Can add but effort is not worth) .


Not many solvers support true 2d nowadays.

[naffrancois]

On a user standpoint, most of our customers (but not all of them) are indeed in the end solely interested in fully 3D simulations.

On our side however, in the early stages of development of a new model equation or new Riemann solver starting directly on 3D grids is simply not an option. This is why, while our code was still in the range of 30k lines we decided to make it truly multi-D (1D/2D/2.5D/3D) in an almost fully transparent way. After all a base solver mainly sees vertex-face-cell connectivities. Our code has since grown a lot and this just makes things way much easier having it all in the same environment.