[tjwldhks79]

While turbulence itself occurs in three dimensions, the reason for conducting two-dimensional analysis when searching for data on drag coefficients in turbulent flow is intriguing.

Is there a specific rationale for employing two-dimensional analysis, and are the results obtained from such analyses meaningful?

It is commonly believed that two-dimensional analysis holds significance primarily in laminar flows, but why extend it to turbulent flows?

Could it be merely due to cost considerations in simulations?

I am curious to hear your opinions on this matter.

Thanks

[LuckyTran]

You are doing RANS. It doesn't matter if it's 2D or 3D. RANS is not a simulation of turbulence, it is simulation of a turbulence model.

Originally Posted by tjwldhks79

It is commonly believed that two-dimensional analysis holds significance primarily in laminar flows, but why extend it to turbulent flows?

Common to who!? That is fake news. We make spaceships using 2D simulations!

If you want to get drag coefficient of an infinite cylinder then the easiest and most logical way to do that is with a 2D case, instead of wasting time, money, resources, thinking, questions, on 3D. And now here you are questioning what even is CFD. I'm not saying there is a grand philosophical reason for doing 2D simulations but in your case, doing anything in 3D is clearly confusing you!


Now for the real question, so why aren't people doing 2D RANS simulations of cylinders? It's simple, they don't know how to mesh in 2D because all the software they use and youtube tutorials they watch only show them how to do it in 3D. Then they get assigned a very basic academic problem and they get absolutely stuck!

[tjwldhks79]

[QUOTE=LuckyTran;865332]You are doing RANS. It doesn't matter if it's 2D or 3D. RANS is not a simulation of turbulence, it is simulation of a turbulence model.



It seems that the information I've come across contradicts what I have known so far. Could you please explain this discrepancy in more detail?

Thanks

[FMDenaro]

Turbulence is unsteady and 3D. The statistical mean can become 2D and steady under suitable assumptions.
When you solve the RANS equation, you do not solve for the real turbulence details but only for getting the statistical means.
Just as an example, the turbulent boundary layer over a flat plate is unsteady and 3D. If you perform a DNS and save the 3d variables for a large period of time and the perform the mean in time and spanawise direction, you get a 2D steady field representing the mean velocity.

[askques123]

Originally Posted by LuckyTran

You are doing RANS. It doesn't matter if it's 2D or 3D. RANS is not a simulation of turbulence, it is simulation of a turbulence model.



Common to who!? That is fake news. We make spaceships using 2D simulations!

If you want to get drag coefficient of an infinite cylinder then the easiest and most logical way to do that is with a 2D case, instead of wasting time, money, resources, thinking, questions, on 3D. And now here you are questioning what even is CFD. I'm not saying there is a grand philosophical reason for doing 2D simulations but in your case, doing anything in 3D is clearly confusing you!


Now for the real question, so why aren't people doing 2D RANS simulations of cylinders? It's simple, they don't know how to mesh in 2D because all the software they use and youtube tutorials they watch only show them how to do it in 3D. Then they get assigned a very basic academic problem and they get absolutely stuck!

Hi,I am modelling multi mode Giesekus Model and it is turbulent case using k, omega SST. Simulation is running for few time steps and then I am getting this floating point error. Also I am using 3d Geometry. Can you please resolve this