Showing results for 
Search instead for 
Do you mean 

103 Real Eigenvalue Solver - Harmonics Simulation Problem



I am trying to do a harmonic sim on a 2d mesh of aluminum sheet metal and the results are showing the first six modes being less than .0002 Hz. I meshed the assembly by doing a midplane surface between all of the components, and then used edge stitching to attach the pieces together. I also used Mesh Mating with the non-coincident glueing option for any pieces that required face to face stitching (which is not a thing).

I ran this same mesh with a linear buckling sim on the 105 Linear Buckling solver and it worked fine and gave accurate results, although that sim contained boundary conditions. I know that in NX 10, to do harmonic sims, our team would use the 103 Real Eigenvalues solver and got accurate results but there was a checkbox for inertial relief, which allowed us to use no boundary conditions. My question is, how should I be setting this up, because the current results I am getting do not seem to be too trustworthy, based on the first 6 modes being a little janky.


Here is a picture of the first mode and the chain of modes:



Re: 103 Real Eigenvalue Solver - Harmonics Simulation Problem

[ Edited ]

Modal solutions do not require the structure to be scatically determinant. Computing modes with no constraints is a free-free analysis. for a single body structure, the first 6 modes will be rigid body modes. Ideally, these will be "clean" modes with frequencies on the order of 1E-4 Hz or less. One or more rigid body modes with frequencies approaching 0.01 Hz or more can indicate grounding in the model.


As your rigid body modes are < 0.0002, the first 6 modes are clean rigid body modes and the remaining are the free-free flexible modes.

Re: 103 Real Eigenvalue Solver - Harmonics Simulation Problem

thank you so much for explaining that. You are the best!