- I'm trying to simulate a turbulent flow around an airfoil. I'm using Ansys software. The problem is my concern focus on the near wall region and I want to know which turbulence model is successful in such case?
Mohammed Alfahham : A couple of details are not clear from your question:
* Simulation in CFX or in Fluent or in Discovery AIM or in Discovery Live or in FENSAP-ICE - all 5 are fluid dynamics programs by ANSYS Inc.?
* Are you looking for (U)RANS, hybrid SRS models or LES?
* What is the purpose of this simulation - just C_D, C_L or more? At what angle of attack?
* Should the simulation be for any compromise at steady-state and rather quick/cheap, or can it be transient?
* At which Reynolds/Mach number are you investigating the flow conditions?
Are you looking into laminar-turbulent transition phenomena or not?
Despite all these more detailed considerations I would at least say, that the SST turbulence model, the new GECKO model or the SBES model might be suitable turbulence models to take into consideration, dependig on how much effort you are willing to spend.
Best regards,
Dr. Thomas Frank.
There is no answer in terms of the "best turbulence model". Ask yourself what is the goal of your simulation and you can start thinking about the proper formulation (of course that depends on your computational power). Then for each formulation you should think about the proper grid resolution and turbulence model.
More details are required about your project.
Dr. Thomas Frank
Thank you very much for these worthy point to look at. I'm trying to simulate the effect of span wise micro groove on lift and drag forces of NACA 0012. I'm using CFX and I'm not sure if the LES will be the best for my case. The flow velocity range between 20-80 m/s. The wind tunnel in my university is a sub- sonic and has maximum velocity around 40 m/s in which I hope to do some experimental tests to see the pressure distribution over the airfoil.
Thank you again for your time and valuable advise.
Mohammed Alfahham
the pressure distribution is a parameter that can be first analysed also in simple Euler model. Then a 2D RANS for low AOA can be a good choice.
The LES formulation makes sense if you are interested in drag and vortex dynamics at high lift, but that requires a 3D model and a very refined grid to solve the BL.
Professor Filippo Maria Denaro
Thank you very much for this short and valuable answer I will try your advice.
Best regards
You can use Ansys CFX for your case. for the turbulence model, it depends on your simulation, either you are going for steady or unsteady simulations. if you use steady simulation than simple RANS is good choice along with SST turbulence model
A good site to compare performance of the popular RANS models for some canonical test flows is Langley Research Center Turbulence Modeling Resource (https://turbmodels.larc.nasa.gov).
In particular, it contains information on workability of nine models for NACA 0012 airfoil, including the widely known SA, SST, EASM, LRR/SSG models. All models give Cp, Cd, CL close to experiment data for angles of attack from 0 to 16o. Indeed, SST and LRR/SSG are slightly better than other models for CL for higher angles, and SST computationally is less expensive because of extra differential equations for the Reynolds stress tensor components in LRR/SSG. All models fail for angles higher than 17o.
Mohammed Alfahham
There is no fit for all turbulence model suitable for any case.
If you happen to have enough time and resources to run DNS (Direct Navier Stokes), LES (Large Eddy Simulator) or at least decent hybrid model such as SST-SAS (Scale Adaptive Simulation); you will most likely get highly realistic results, close to experiments.
However, most of us do not need such a level of complexity and use RANS methods. In such cases, we need to rely on the 'artificial' eddy viscosity to simulate the turbulence.
In such case - here is the methodology I recommend to my students:
As previous speakers mentioned, the SST model might be an excellent first choice for your case but beware of the effect of high AOA (recirculation) and turbulence transition effect.
Bear in mind, that your case is not a new problem, and there might be somewhere on The internet very similar problem solved already.
As a good start, please refer to:
https://commons.erau.edu/cgi/viewcontent.cgi?article=1377&context=edt
The best turbulence model for simulate in ANSYS is k- w SST ( SST in CFX) which is give you a nears results to experiment especially for your case . where i was focused on same this topic in my research. you can read my papers for more details.
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators