Force Calculation #111
-
|
I've done the forced oscillation of the cylinder by modifying the BC-Cylinder.f90 file and ibm.f90 file. The changes that I've made are shown below and it worked successfully. The control volume for the forces is taken as follows **"&ForceCVs xld(1) = 3.0 !X left for volume control /End"** And the Coefficient of lift is plotted with time and is as shown The lift coefficient seems not reasonable. The animation of the cylinder oscillation is shown below oscillating_cylinder.mp4Thanks in advance.. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
The oscillations which occur are to be expected (they are commonly known as spurious force oscillations (SFOs)). The oscillations will decrease in amplitude with the increase of the spatial resolution and the increase of the time step. The simplest solution is to filter the hydrodynamic coefficients in a post-processing step which should work just fine for both cases mentioned. With a simple filtering you should be able to get accurate predictions for the hydrodynamic coefficients. You can take a look at our recent paper where we have performed simulations of the VIVs of a cylinder at various Reynolds numbers here: https://trebuchet.public.springernature.app/get_content/f00e5a2b-c2b8-437e-9bfd-078dd012ff44 For more information on the filtering you can also look at the following paper: https://www.sciencedirect.com/science/article/pii/S0307904X21003103 From the sort video it looks like the domain may be too small in the Y-direction and perhaps too wide in the X-direction. A final comment, please ensure that the surrounding box which is used for the calculation of the hydrodynamic forces is large enough to capture the amplitude of the cylinder oscillations (the coordinates of the box are printed by the forces.f90 script during the first iterations of the solver.) Hope this helps! |
Beta Was this translation helpful? Give feedback.
-
|
Got it, Thank You @nasos94 |
Beta Was this translation helpful? Give feedback.
The oscillations which occur are to be expected (they are commonly known as spurious force oscillations (SFOs)). The oscillations will decrease in amplitude with the increase of the spatial resolution and the increase of the time step. The simplest solution is to filter the hydrodynamic coefficients in a post-processing step which should work just fine for both cases mentioned.
With a simple filtering you should be able to get accurate predictions for the hydrodynamic coefficients. You can take a look at our recent paper where we have performed simulations of the VIVs of a cylinder at various Reynolds numbers here: https://trebuchet.public.springernature.app/get_content/f00e5a2b-c2b8-437e…