[Diptaru]

Hello everyone,

I'm currently conducting a 2D steady-state simulation on a sharp trailing-edge airfoil.

I've set up a C-type domain with a structured mesh in ICEM CFD, using the k-w SST turbulence model. The Reynolds number is 1,000,000, the chord length is 1 meter, and I'm aiming for a Y+ < 1.

To maintain this Y+, the first cell height needs to be very thin (around 0.012 mm in my case). However, this creates a problem: as I reduce the first cell height to 0.012 mm, the height of adjacent cells in the downstream blocks also becomes 0.012 mm, leading to cells with high aspect ratios downstream.

Additionally, my solution becomes very unstable and tends to diverge. The lift and drag coefficients are significantly higher than expected. I'm also encountering the error message: “Turbulent viscosity limited to viscosity ratio of 1.00e+05 in xxxx cells.”

One approach I've considered is increasing the number of divisions stream-wise, but this also increases the aspect ratio of the normal-wise cells at the top and bottom, which are further away from the airfoil. Eventually, this leads to a significant increase in the total cell count (around 450,000 in my case).

However, I've found in the literature that some authors have conducted similar simulations with much lower cell counts, around 60,000-160,000.

Is there a more convenient solution to this issue? if it is then what is the meshing approach?

[shereez234]

Originally Posted by Diptaru

Hello everyone,

I'm currently conducting a 2D steady-state simulation on a sharp trailing-edge airfoil.

I've set up a C-type domain with a structured mesh in ICEM CFD, using the k-w SST turbulence model. The Reynolds number is 1,000,000, the chord length is 1 meter, and I'm aiming for a Y+ < 1.

To maintain this Y+, the first cell height needs to be very thin (around 0.012 mm in my case). However, this creates a problem: as I reduce the first cell height to 0.012 mm, the height of adjacent cells in the downstream blocks also becomes 0.012 mm, leading to cells with high aspect ratios downstream.

Additionally, my solution becomes very unstable and tends to diverge. The lift and drag coefficients are significantly higher than expected. I'm also encountering the error message: “Turbulent viscosity limited to viscosity ratio of 1.00e+05 in xxxx cells.”

One approach I've considered is increasing the number of divisions stream-wise, but this also increases the aspect ratio of the normal-wise cells at the top and bottom, which are further away from the airfoil. Eventually, this leads to a significant increase in the total cell count (around 450,000 in my case).

However, I've found in the literature that some authors have conducted similar simulations with much lower cell counts, around 60,000-160,000.

Is there a more convenient solution to this issue? if it is then what is the meshing approach?

Your approach is correct. At the end of meshing, go to edges on the outer side (far field), unselect copy parameters to parallel edges and use uniform spacing there, (you need to be careful not to copy these parameters to parallel edges again as this can destroy your boundary layer).. Takes a bit of practise but you will get there..



Regardless, also, reduce your relaxation factors in ANSYS fluent/CFX etc to stabilize your solution