[rsunil]

Hello everyone,



For my CFD project, I have been working on an Unsteady analysis over a 2D simplified dragonfly wing (corrugated wing) using DNS in StarCCM+ (laminar model).



Initially, I worked on a structured mesh for this profile and performed a mesh independence test at a Reynolds No. of 6000 and an angle of attack (AoA) of 9 degrees.


In this process, I obtained results where the flow behaviour changed from periodic to chaotic for a refined mesh. To be precise, I used a mesh of 450,000 cells for my study at various Reynolds numbers and AoAs and the results obtained were well within 3% deviation in comparison to a research article's results. I increased the cell count by a factor of 2, obtaining a mesh cell count of 1.75 million (for testing mesh independence). I tested the mesh at the same Re= 6000 and AoA= 9deg, the Coefficient of lift (Cl) plot was shown to be chaotic for the refined mesh (1.75 million cells) but periodic for the mesh of 450,000 cells. Meshes close to a cell count of 900,000 showed a periodic nature as well.


The flow was determined to be chaotic or periodic by taking the Fast Fourier Transform (FFT) of the Cl plots.


I was not able to pin-point where the problem could be as the quality of the Mesh seemed to be good as the cell qualities were >0.6 in ICEM CFD. So, I made a hybrid grid and tested for Mesh Independence at the same Re and AoA mentioned above. But for this mesh, there was no change from Periodic to Chaotic. It was periodic throughout.


The research article had obtained a Periodic result for the same conditions.



So, it would be really helpful to know what could be causing such changes for the structured grid and in such scenarios, how do I rely on which grid to choose ?


Thank you in advance for the help.


PS: Please let me know if further information is required