ANSYS modal analysis and normalization methods

[jgrzmmm]

Hello,

As we know, ansys when post-processing results, depending on analysis options, can show either "to mass matrix" or "to unity" normalized displacements.
One thing I don't understand then is why in the results of my analysis of simply supported cylindrical structure, the location of max displacement in "mass marix" normalized analysis is different than in the other one (please see the attachement)? I thought normalization is only about setting a relative value for modal shape display purposes, therefore the distribution of displacements shouldn't be affected. Or is it just an negligible numerical calculation error?

Other question: when I went through first few modes of this analysis with "to unity" normalization, I found that in most of the cases maximum value of X, Y or Z displacement is equal to 1. However in few cases neither of those displacements are equal to 1 .. Why is that?

Thank you,
Julian

 

[johnhors]

Your mode shape looks very symmetric. I suspect the values at top and bottom are the same to within round off error limits, thus both locations qualify for the maximum value and numerical round off determines which one gets it. It is not a calculation error.

You should ask ANSYS why some of the "to unity" normalizations are not to unity. This is very strange, since the normalisation routine is a trivial task and one wonders how this has gone astray.

http://www.Roshaz.com

quality, cost effective FEA solutions

 

[JoshPlumSE]

I wonder if the issue is an issue with the signs of the modes.

If joint A and B have the same displacement, but with different signs then which one is the "max". If normalizing to the mass matrix give A as positive and B as negative, then joint A may be said to control. If normalizing to unity gives A as negative and B as positive, then B may be said to control.

Regarding the reason why those are normalized "to unity". My guess is that they may be using a generalized algorithm for plotting contours geared more towards plotting internal stresses than joint deflections. If you look at the numerical (not plotted) results, they may be more exact.

 

[Drej]

The "greater than unity" numbers from unity normalisation in the contour plot could well be down to the nodal averaging that ANSYS performs. If you plot the element nodal results you should see that this is indeed unity. I agree also that the different locations for "max displacement" between unity/mass normalisation is probably down to round off - list and check the nodal displacements (the numerical values) and compare.


------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[johnhors]

Drej - nodal displacements are simply just that. There is no nodal averaging for displacement results. Elemental results are calculated using the nodal displacements as one of the inputs.
From "element nodal results" you can get stresses and perform averaging on the stresses of course, but not with displacements.

It would be nice to get an "official" explanation from ANSYS !
Can't the OP tackle them directly and post the explanation here?

http://www.Roshaz.com

quality, cost effective FEA solutions