[surajp92]

I was trying to determine the drag of underwater vehicle. I have also done CFD analysis of open water performance of propeller and I got good results by maintaining low Y+(nearly 1) over the blade surface.

But, when I run the simulation for determining drag of underwater vehicle I found that CFD model is under predicting the drag when I use low Y+ over the body surface. The CFD results gives good results when wall Y+ value is higher (around 30) over the body surface. I am comparing my results with drag obtained from ITTC 1957 line and using form factor.

My understanding was that when I have low Y+ value over the body surface, the velocity will get resolved till wall boundary by solving momentum equation. Also for first grid point from wall boundary u+=Y+ relation is used. And when Y+ is higher then log law is used for determining the velocity of first grid point from wall. (Please correct me if I am wrong). If my understanding is correct then CFD results with low Y+ should be more accurate because velocity near wall is determined by solving momentum equation instead of using log profile of velocity.

I am using K-w SST turbulent model with all Y+ treatment in StarCCM+.

Thank you in advance.

[piu58]

I am not familiar with StarCCM+.

If you run a RANS simulation the turbulence in the near of the wall is modeled. The model works fine if you have an y+ of around 30. If you have a too small y+ you don't have any turbulence model in the Prandtl layer.

To get valid results with low y+ you have to choose a solver which can model the turbulence effects in the boundary layer directly.

[tiam]

The k-omega sst model has been originally developed to work with meshes that resolve the velocity up to y+ around 1, so it should work fine. No idea how exactly the concrete implementation in StarCCM works though.


Sent from my iPhone using CFD Online Forum mobile app