RowToTri wrote:
Can I take advantage of your presence here and derail this thread a tiny bit? What is your preferred turbulence model for external aerodynamics simulations - particularly when there is a significant turbulent wake? I just looked at this Daimer truck CFD analysis where they conclude the Spallart Allmaras DES model compared very favorably with wind tunnel tests at lower computational cost than an LES model, while RANS was off by 20%. It's 5 years old - would something like the k-omega-SST Improved Delayed Detached Eddy Simulation be better?
Thank you so much!!
My preferred model is no model. DNS is obviously the preferred choice in all cases. Now, once you have some kind of computational resource limitation, then you enter the realm of (in order of computational cost): (i) LES with some kind of subgrid scale model, (ii) DDES - implemented with k-w SST, (iii) Spalart Allmaras DDES, (iv) probably some others like SAS which is an improved URANS model, and (v) RANS.
The one other item in here that I did not mention in here is the concept of under-resolved DNS. Letting the length scale of the grid do the filtering for you and have no subgrid scale model.
You can hand pick papers that support anything here, e.g., some show RANS does well, some show LES does terribly, some show DES does well. You really need to first have a clearly defined objective for the project/simulation and then decide. (e.g., is the turbulent wake interacting with an object of interest? if so, you are probably better of with DDES or LES). Is the body that is generating the turbulent wake made of sharp corners with clearly defined separation points? if so, then DDES is fine since the simulation will have a clear location to transition between kw-SST and LES. If the body is smooth and you have complex free stream pressure gradients, then separation points may not be predicted adequately with kw-SST.
Edit: under-resolved DNS is also called Implicit LES depending on the paper.
Chris Morton, PhD
Associate Professor, Mechanical Engineering
co-Founder and inventor of AeroLab Tech
For updates see Instagram