Which result is important after modal analysis; displacement or mode frequency convergence

[Hkn88]

Hi there,

I have run several modal analyzes on the same model with different mesh sizes. Eveery time, i change the mesh size (gradual decrease) i get different mode frequency and displacement. My questions to experienced users:

1. First of all, why do i get different frequency and displacement results for different mesh sizes although it is the same model(mass is the same)?
2. Another question, after executing the analysis acording to which criteria i can assume the model converged to the correct values. During mesh refinement do i need to watch displacement to converge or first mode frequency to converge or what?

Many thanks.

 

[L_K]

If you plot frequency (y axis) and number of elements (x axis) you should see the
frequency converging to a value.

 

[rickfischer51]

What elements are you using to mesh your model? Bricks, Tets, linear, quadratic, etc. Is the mode shape changing, or just the frequency? Frequency is proportional to sqrt(k/m), so if the mass is fixed and the mode shape is the same, the stiffness must be changing. Linear tets are notoriously stiff, so refining the mesh should reduce the frequency. bricks and quadratic elements are less prone to this. As LK said above, it should converge on a value. Try this on a simple model like a cantilever beam where you can compute the frequency by hand. Also, the displacement plot you are looking at is not a real displacement. It is a shape derived from the eigenvalues of the modal run. You should be converging on frequency.





Rick Fischer
Principal Engineer
Argonne National Laboratory