Using Abaqus, I am trying to generate symmetric lamb wave mode in a simple Aluminum plate by applying symmetric point force parallel to the plate. The extra wave followed by the S0 mode appears all the time. This is definitely not the A0 mode because the geometry is symmetric and hence no mode conversion would take place, these are not the reflections from the boundary, I am new in this topic and I need some guidance,
You have a wave packet due to the the impact on the plate. We do not know where is the environment of the plate. The trailing packets can be excited by a far reflection from the boundary the enclosure of the plate.
I proposed this hypnosis because as you see the second wave packet is attenuated relative to the first wave packet. So, i assume trailing pulses are echos of the first pulse. They may come from a far boundary to the plate.
Best wishes
Hi Haider,
If you plot the dispersion curve, and check what modes are getting excited in the frequencies that you excite the plate, you might narrow down the search about where that unknown burst is coming.
I suspect you will find that one group may be for a torsional vibration mode and the other is for a bending mode of vibration.
Could you clarify the geometry and boundary conditions? How long is the plate, where is the excitation, where do you extract results? By "symmetric point force", do you mean that you excite on both upper and lower plate surface? If the geometry as well as the boundary conditions are perfectly symmetric, I agree that this should not be the A0 mode. If you are also sure that it's not a reflection from the boundary (how did you check?) it must be a higher symmetric mode. In that case, you may look at the dispersion curves to see which modes could exist and check whether the group velocity matches the results.
Hi Hauke, It'd help if you provide the thickness of the modelled plate and the frequency contents (FFT) of your pulse (five sine periods in a Hanning fenster?). Right now I think of the following possibilities:
1. S1 Mode:
With the the thickness, the frequencies and the dispersion curves of the Symmetric S-Modes you have everything you need to find out if you are exciting an S1 (S2, S3, ...) mode.
2. Non symmetrical FEM-mesh:
If your FEM-mesh is non symmetric you'll get some artificial A-Modes, if your mesh is not fine enough.
3. Numerical dispersion:
If you are using an explicit time integration method and your mesh although symmetric is not fine enough, you'll get some artifacts due to the instability of the explicit time integration method.
It doesn't seem you have a convergence problem but the looks of your plots. I'd check Nr. 1, in case you are sure you are not getting any reflections.
As a recommendation I'd like to add, that if you are only interested in symmetric modes you can save a lot of time and memory by modeling only half of a plate. Use then symmetric boundary conditions for the mid plane of the plate and impose shear forces on the surface.
I hope it could help.
Best regards
Miguel
- Badges
- Science topic