How to calculate Mises stress from complex stress results

[nicolasb65]

Hello,
from a frequency response analysis have complex stress tensors.
How do I calculate the magnitude of the mises stress?

Be aware that the mises stress itself is not harmonic!! So i have to find the maximum over the phase angle. Theirfor a formula should exist.

Do you know this formula?
Thanks for you help

 

[prost]

Maybe I don't understand what you analysis you are doing. Typically "frequency analysis" is just you trying to find the eigenvalues, the natural frequencies, of the structure. These are a function of the stiffness matrix, not the loads. There are no stresses to compute in this case.

 

[johnhors]

Are you a student?

 

[GBor]

Prost,

This is a frequency response analysis from which you do get displacement and stress, not just the eigenvalues; however, I don't fully understand the rest of the question either.

nicolasb65,

If you have the "complex stress tensors", you should be able to calculate the Mises stress from well-published equations unless I'm missing something.

GBor

 

[nicolasb65]

The problem is, that the Mises stress is a function of the phase angle. Usally you need the maximum of the Mises stress over the phase angle.
You can not solve this analytically - heavily nonlinear. And if you have a hughe FE-Model with several 100.000 elements you need a fast approximation.
This approximation exsit - several FE-solvers and postprocesors use it. But I need to know it, because my combination of tools did not calculate it.
nicolas

 

[GBor]

What software are you using and can you export the data to Excel? Or can you import the ASCII text version in to Excel? What information is your processor outputting?

 

[spongebob007]

ANSYS published an article about this very subject in their magazine several years back. VonMises stresses are not exactly accurate for a frequency response analysis since the stress components may be out of phase with one another. I have a hard copy of the article somewhere, I will see if I can find it.

 

[Bausk]

Nicolas,

I'm not that into dynamics but I think you will find this Sandia labs report useful. It's dated 1998 and is freely downloadable:

http://infoserve.sandia.gov/sand_doc/1998/980260.pdf

 

[nicolasb65]

It is right, their is common formula to calculate the maximum Mises in this way, that the maximum of all stress components (independing of their phase angele is used). That leads to wrong result if they are great differences in the phase for stress components.
But this is not my problem know.

And thanks for the artikle - I'll go throught and tell you if that solves my problem.
Thanks for your help!!
Nicolas

 

[hacksaw]

nicolasb65,

Best to transform the stress components into time domain. That is a little easier to comprehend and visualize how the dynamic amplitudes fall into place.