Modal Analysis: parameterizing certain mode shapes

[gman89]

Hi, I'm trying to verify mechanical models of shells with certain textbook data so I can then use the same parameters to make conclusions.

To simplify the question, I basically want to be able to to parameterize a specific mode shape (say n=1, m=1) in the modal analysis so I can get a response surface to see how the frequency changes when I change length, radius and thickness for example.

The problem is that as I change these parameters, the particular measurement number changes. For example, the n=1 m=1 mode might be the third measured node for a certain length, but when I change the length it might be the thirteenth and I therefore can't make the the frequency of mode number 3 an output variable. Hence, I can't track this mode shape and get a response surface.

Any ideas on how I can track a particular mode shape? Any information will be helpful.

 

[kan123]

Can you pick something about the mode of interest that is unique? For example, maybe your mode is characterized by having the displacement at one pair of nodes being equal while another pair is equal magnitude but opposite sign. Then you can loop through your modes until you find the first that satisfies your *IF statement.

 

[gman89]

Well I was trying to think of a way (say for m is 1 which is one half sine wave) that I could do a code that counts the number of peaks, but how do I even go about doing that, and making it a variable etc?

 

[kan123]

If you want to count local peaks you loop through each node one at a time. For each loop you have a parameter set to the magnitude of the node's displacement. Then select attached elements and the nodes of those elements (omitting the key node). Find the maximum magnitude of displacement among that set. If it is lower than the value for your key node, then you add one to a running counter.

 

[gman89]

Thanks for the help.

I have a few follow up questions to that, but for now where do I even write a code in ansys? Do I create a commands section in the solution part of the setup to perform what you described?

 

[kan123]

I don't use WB much, but you should be able to insert APDL commands in the post-processing section after the solution is obtained.

 

[flash3780]

I have a few follow up questions to that, but for now where do I even write a code in ansys?

Well, you can input command snippets in Workbench by adding them as you would add boundary conditions or results (Right Click>Insert APDL Command, I think). However, you'll find this very difficult and frustrating if you don't know much APDL.

Ansys Classic (now Ansys/APDL from v12 on) provides you with a command line which allows you to test your commands without running the entire model. This is a HUGE benefit.

<soapbox>
Building a model can be thought of as writing a program. Workbench is the equivalent of a "Visual" programming, with drag and drop type of options which do a lot of behind-the-scenes programming for you. Visual programming languages can be extremely powerful if you know the underlying language, but can be extremely frustrating if you need to do something that's not prepacked in the GUI.

Ansys Classic, on the other hand, is more like writing a program from scratch (not quite, but, go with it). It's extraordinarily flexible, allows you to make changes on the fly, and allows you to access functions and features that aren't available in the GUI.

As a new user, starting out with the "Visual" side of ANSYS seems great. It's very straightforward how to go through the steps of creating a model. However, as the modeling tasks become more difficult, the shortcomings of the GUI become painfully apparent. Outputting nodal forces, applying variable boundary conditions, reading results that aren't available in the GUI, and numerous other tasks that aren't supported by the GUI becomes extremely painful in Workbench. There is no command line, so there's no means of testing whether or not your commands did what you expected them to do without running the model. Even using command objects, Workbench isn't able to easily output boundary conditions which are applied outside of the GUI interface.

When I learned Ansys I started with Workbench, then moved to Classic. In hindsight, that was the wrong way to do it. It would have been better to learn the language in Classic, and then move to Workbench to take advantage of the modeling/meshing tools and the unified environment.

Ansys/APDL has some great tutorials in the help files. I'd recommend that you walk through them, the first with the GUI, and then with only APDL commands in the command line. It doesn't take terribly long and I guarantee that you'll find what you learn invaluable.
</soapbox>