You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Though not critical due to the boeing vertol model performing well enough for the fidelity of this model, the Lieshman Beddoes model is implemented, but not fully hooked up since it was not matching the experimental data (i.e. there was a bug), and because the state variables were much more complicated for book keeping.
The text was updated successfully, but these errors were encountered:
I have just implemented the BL model and validated against some examples in a text book. The basic goal is to find a good low order model to simulate the dynamic stall behaviour. BL is quite popular in the scientific community with its good accuracy and physical interpretation. Could I know more about how BL is not matching the experimental data (what is your airfoil profile)...
I haven't done much on this is a long time. Here is the original dynamic stall repo I was working with before I just integrated the boeing vertol model into the OWENSAero package. It has the BL model in it and some scripts that compare to data in papers that are also included (might be commented out in the test cases). Both the BV and BL models were scraped from CACTUS, and I have made some numerical improvements/changed to the BV model that is in OWENSAero. No guarantees that this attached old repo will run, but you're welcome to give it a shot.
Though not critical due to the boeing vertol model performing well enough for the fidelity of this model, the Lieshman Beddoes model is implemented, but not fully hooked up since it was not matching the experimental data (i.e. there was a bug), and because the state variables were much more complicated for book keeping.
The text was updated successfully, but these errors were encountered: