[kh.aa]

What is the most common differencing scheme used for convective terms recently?

The cases are sudden expansion and a flow over cylinder. The flow is incompressible.

Using FV method and RANS
The grid is non-orthogonal non-staggered structured

[FMDenaro]

Originally Posted by kh.aa

What is the most common differencing scheme used for convective terms recently?

The cases are sudden expansion and a flow over cylinder.
The grid is non-orthogonal non-staggered structured

Considering only incompressible flows, I can suggest using FV formulation to fulfill conservation of mass and momentum. Often the type and order of the discretization depends on what formulation you use (RANS/LES/DNS).
You should provide more details about your interest

[kh.aa]

Originally Posted by FMDenaro

Considering only incompressible flows, I can suggest using FV formulation to fulfill conservation of mass and momentum. Often the type and order of the discretization depends on what formulation you use (RANS/LES/DNS).
You should provide more details about your interest

Thanks for your reply and sorry, I was not clear.

I am using FV with RANS formulation and the flow is incompressible.

[FMDenaro]

Originally Posted by kh.aa

Thanks for your reply and sorry, I was not clear.

I am using FV with RANS formulation and the flow is incompressible.

In this case you do not need high accuracy order provided by recent advanced discretizations and also upwind schemes can be used.

[kh.aa]

I see. I was intending to use the Upwind scheme.
I was afraid that this will affect the accuracy.

But can you please state the reasons that no need to use higher order schemes for such conditions?

Thanks

[FMDenaro]

Originally Posted by kh.aa

I see. I was intending to use the Upwind scheme.
I was afraid that this will affect the accuracy.

But can you please state the reasons that no need to use higher order schemes for such conditions?

Thanks

The contribution given by the RANS model has greater magnitude order than the local truncation error of the discretization. You will not get advantage in using high order schemes