CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   Mesh Refinment (https://www.cfd-online.com/Forums/cfx/25296-mesh-refinment.html)

Steven February 25, 2008 06:32

Mesh Refinment
 
Hi

I was having convergence problems on my 3d airfoil analysis but seemed to have solved them through changing inflation layer expansion factor and number of inflaion layers. I now have a simulation that does converge with a y+ of 1 (SST) and 9 layers of inflation with expansion of 1.05.

I wanted to study how refining the mesh ( reducing the face spacing on the airfoil so it was finer at the leading and trailing edges where pressure gradients are highest) would effect important parameters, namely lift and drag. As expected as i increase the value form 0.1mm to around 1.5mm the value decreases from level showing 'less accuracy' but after this point the value increases again! Can some one explain this? The airfoil has a thickness fo 3mm if that helps.

Thanks Steve

Rogerio Fernandes Brito February 25, 2008 06:53

Re: Mesh Refinment
 
Tell us about the boundary conditions (flow rate or velocities at the inlet and static pressure zero at the outlet?). How much elements are you using? Maybe 1,500,000 is a good number. I suggest that you must use 1 mm of thickness and choose "simmetry condition" on z direction. What´s the level of turbulence at inlet?

Steven February 25, 2008 08:28

Re: Mesh Refinment
 
I have a flow rate of 3ms at inlet and static pressure at out let. I am only modeling half the model as you suggested (ie sym in z direction) with a spacing of 0.3 i have 750000 elements, im reluctant to go to finer due to cpu time.

I have max volume size at 200mm Max surface of 200mm and minimum 0.3mm On airfoil max of 200mm and min 0.3mm

Inflation applied to teh sureface of airfoil Ive gone back to using 8 inflation layers with y+ of 2 and expansion factor of 1.2

When I do go finer i get aspect ratio errors aswell.

Steve

Glenn Horrocks February 25, 2008 17:40

Re: Mesh Refinment
 
Hi,

It is common to not get monotonic convergence in mesh refinement studies, particularly as you move away from mesh convergence.

A good discussion of the issues of mesh refinement can be found in "Computational Fluid Dynamics" by Roache.

A good summary of the issues is discussed in the Journal of Fluids Engineering Numerical Accuracy Policy: journaltool.asme.org/Templates/JFENumAccuracy.pdf

Glenn Horrocks


All times are GMT -4. The time now is 08:59.