[@E18]

Hi,

I am simulation flow past a cylinder under controlled vibration with DES in Star-CCM+ and the result is obviously not what I was expecting and the experiment shows. The Reynolds number is in a range that I have a laminar boundary layer and turbulent wake.

I am just wondering whether or not DES is a good option for situations where there is separations right at the wall? The original DES uses a distance function to switch between RANS and LES and RANS modeling is not the best option to deal with separations and reattachments.

Anyone here with a successful DES simulations for separations at the wall? I am not sure if my mesh is refined enough but I am pretty sure it is as good as many other studies for a similar case. I am using a radial mesh with max(y+) = 2 at the wall, the circumferential and spanwise direction discretizations are in an acceptable range.

Thanks,
A E

[FMDenaro]

Have a look to this page:
https://cfdisraelblog.wordpress.com/...he-aid-of-les/

[@E18]

Originally Posted by FMDenaro

Have a look to this page:
https://cfdisraelblog.wordpress.com/...he-aid-of-les/

Thanks for the link Filippo. It addresses many features of DES but I am still confused.

I have used IDDES with SA, SST-KW, and KE-EB RANS based formulations for my simulations. The model works fine for flow past a stationary cylinder in terms of force time history and statistical values. However when the cylinder is moving, it seems like the lift and drag time histories are damped and do not modulate as much as the experiment shows.

The results show that the near wall region is entirely treated by the RANS part even though there is separations and reattachments right at the wall as well.

Thanks,
A E

[FMDenaro]

Is the cylinder oscillating in the direction normal to the streamwise direction?

[@E18]

Originally Posted by FMDenaro

Is the cylinder oscillating in the direction normal to the streamwise direction?

Yes, to preserve the mesh quality, instead of moving the cylinder, the mesh domain is moving with a prescribed sinusoidal motion (This is a controlled VIV)

A E

[FMDenaro]

I cannot say more, I have no experience about DES in such flow conditions but I have doubt you can get an accurate solution from the RANS formulation on an oscillating body...

[@E18]

Originally Posted by FMDenaro

I cannot say more, I have no experience about DES in such flow conditions but I have doubt you can get an accurate solution from the RANS formulation on an oscillating body...

That could be the reason. Do you have any references for deficiency of RANS in moving boundary cases?

Thanks,