The geometry looks like this:
 

The geometry looks like this: There is a throttle plate in a pipe. Upstream the throttle plate the pipe bends.

The aim is to compute the pressure distribution of the throttle plate. My idea is the following: First I use a large domain covering the bend of the pipe upstream the throttle plate. During this stage I use a RANS model to compute the flow.

Then I put the inlet closer to the throttle plate and use the RANS result of this position as an inlet velocity profile for a LES computation.

1) I will use ANSA to create my mesh. I would like to create a plane inside my RANS domain where the LES inlet will be positioned. Is OpenFOAM able to handle a transparent patch, i.e. consisting of internal faces, only used for postprocessing?

2) How do you suggest me to map the RANS velocities of the position of the LES inlet onto the LES case? Please consider the two alternatives of question 1:
a) I have no transparent surface in the RANS case at the position of the inlet of the LES case.
b) I have a transparent surface...

3) Now, assume I have the velocities from the RANS case set at the inlet of the LES case. My question concerns the turbulent boundary condition. I've understood that "fluctuationScale" of the b.c. "turbulentInlet" simply means white noise.
a) What is the freqency of this white noise?
b) Assume I have a flow of 10 m/s ONLY in the x direction. Now I set the fluctuation scale to be (0.1 0.1 0.1), what does this mean for the y and z velocities (0.1 times 0?)?

4) Any general advices on solving this task? Especially concerning a proper inlet boundary condition.

Hope to hear from someone soon.

Best regards,
Christian Svensson