At the best of my knowledge, the extrapolation that is implemented in fluent should be already congruent to the convective conditions as it assumes zero diffusion. Have a look here https://www.cfd-online.com/Forums/fluent/206317-convective-outflow-bc.html

Dear Ravi,

check this link

https://www.researchgate.net/post/How_to_write_and_insert_boundary_conditions_as_UDF_in_ANSYSY_Fluent

Dear Ravi,

You can use the link to get the initial inputs and knowledge about macros in Fluent.

https://www.afs.enea.it/project/neptunius/docs/fluent/html/udf/main_pre.htm

Thanks Filippo Maria Denaro, Syed Mughees, Arun K Raj Ali for giving your valuable time to answer my question.

I am simulation flow over a Low pressure turbine linear cascade. I had applied outflow boundary condition at outlet and run 2D Simulation, but result are matching with experiments. Actually in experiment a separation bubble is formed at the suction surface around 70% of chord but in simulation its not there Due to which there is a discrepancy in the result. People have done LES on the same problem and found the Separation bubble using convective BC.

I tried with URANS and v2f Model but both are giving the same result.

So I think Boundary condition needs to changed and for that i have to write UDF. I have never written any UDF. It would be great if someone can help me out.

LES would require a full 3D unsteady sumulation. Do you have enough computational power?

Yes, I know that Filippo Maria Denaro, But my question is that my BC at outlet is correct? Or there is problem somewhere else. Thats why i want to try with convective BC to see whether the result would match or not.

I don't think at all that the problem is the BC in outlet, just set the outlet at a correct distance from the trailing edge. Setting no diffusion at the outlet is congruent to the convective BC.

The results in the experiment is very different from what you are simulating. There are real 3D effects and the statistics can be affected. Your 2D statistical solution assumes the flow is not confined. Thus, check the condition of the experiment and how the results are measured. The statistical solution can be too dissipative due to the turbulence model or a not adequate grid resolution of the BL.

Then, you see numerical wiggles in the solution, those are not physically correct.

Thanks evey1

the problem was in the turbulence model initially I was using v2f now I changed it to transition SST thts sove the problem