[SamiBouterra]

Hello everyone,
Here are the details of my problem which I have been trying to deal with for months!


Transient simulation of S809 phase II using sliding mesh of rotational and static domain.


The steady state torque error in compraison to experimental result is 7 percent however, when switching to sliding mesh the torque values oscillate at a value of around 17 percent error to the experimental value which is not acceptable for validation. (comparison made with paper from Sorensen https://arc.aiaa.org/doi/abs/10.2514/6.2000-37)



The mesh is structured with y+ around 1 (first layer height 0.00002m) and the mesh export has no errors and max skewness is around 0.92 (the best I could obtain).


I had many fails until the sliding mesh solution succeeded (I fixed the error of cells handedness which I believe was related to skewness)


Here is a list of what I have tried to obtain close torque values:


*Used various Meshers : Gambit, Pointwise, Ansys Meshing (Structured and unstructured) and solvers Fluent and CFX as well as three versions of fluent 19.0 , R2 and 3.



*tried simulating whole three bladed turbine in a 2 cylinderical domains and one blade (120° domain)+ simple, coupled, piso and SIMPLEC.
* verified reference values influence to get correct value of Cm



I am trying fluent meshing now to see if results are different.


So if anyone has had a succesfull experience with sliding mesh technique get back to me, THANXXXX.


I have read here that the sliding and static domains interfaces should be separated by a certain minimal distance to avoid the sliding mesh errors but no information was provided on the extend of that distance?? any info on that?



****Problem details****
K-OMEGA SST, SIMPLE, sliding interfaces (Periodic repeats), time step 0.0006 according to (Sorensen 2000 paper).



Thank you