If you wanted to model it with perfect accuracy, that would require a DNS model with a high order of accuracy in terms of your numerical grid resolution, a lot of computational power and time at your disposal.

To get an some understanding at least (not perfect accuracy) of the eddies and drag associated with something like a car wheel LES would describe the effect of the larger eddies, as the name suggests, whereas RANS is averaged, again as the name suggests, I recommend reading this thread on stack exchange for some more understanding of the differences "What are the differences between an LES-SGS model and a RANS based turbulence model?"

good luck

To describe the vortical structures up to a certain characteristic lenght you need to RESOLVE them NOT to mode them. The SGS model accounts for turbulent structures below this lenght. This is in brief the meaning of LES and the reason for which RANS cannot describe these structures as the model accounts for the whole range of structures.

You will get large eddies with RANS simulation, but probably they won't be perfectly well resolved because you won't simulate correctly smaller structures that at the end have a certain impact on larger eddies (to a more or less extent depending on your configuration AND the associated mesh). The size of the structures you can resolve correctly is also certainly a matter of meshing, as well as of turbulence model.

The SST model (whether in transition formulation or not) is an eddy viscosity model combining advantages of a k-epsilon and k-omega model formulation. All common turbulence models aim to represent the transfer of momentum caused by turbulent eddies as an artificial turbulent viscosity. In such, none of RANS turbulence models actually resolve any turbulent eddies. They, however, aim to represent the effect turbulent eddies have on the flow solution by using the turbulent viscosity concept. I think you are struggeling with the definition of what an eddy actually is. And that is often not very clear. Some people define an eddy as a flow structure whose flow direction differs from the main one in the problem. This also applies to large scale vortices which you can very well resolve with RANS CFD. So one could also call these vortices eddies.

So in compact: RANS turbulence models do not resolve turbulent eddies, they model their effect. If you define eddies also as large-scale vortices, RANS can represent this.

In LES simulations, the larger turbulent eddies are really resolved and the smaller ones (below the grid resolution) are modelled with a subgrid model.

If you wanted to model it with perfect accuracy, that would require a DNS model with a high order of accuracy in terms of your numerical grid resolution, a lot of computational power and time at your disposal.

To get an some understanding at least (not perfect accuracy) of the eddies and drag associated with something like a car wheel LES would describe the effect of the larger eddies, as the name suggests, whereas RANS is averaged, again as the name suggests, I recommend reading this thread on stack exchange for some more understanding of the differences "What are the differences between an LES-SGS model and a RANS based turbulence model?"

good luck

The difference between using RANS and LES depends on the required accuracy you want. Both of them may be accepted however using DNS with suitable fine mesh may be used.