Pre-stressed modal analysis

hooshsim

Hi,

Is it possible to perform a pre-stressed modal analysis? If so, please provide some guidance. A simple test case will be greatly appreciated.

Hooshsim

prop_design

yes it's possible. just setup a regular modal analysis and add the 'new stress stiffening' load. obviously, fix the model and apply whatever loads you want.

i'm looking for some old models to post. be a few minutes.

prop_design

ok, here are the examples i have. due to the large size of the download, i had to clear the solutions from the examples. if you want to see the results, you will have to re-solve the files.

update; added one additional example

https://www.dropbox.com/s/1gxdl9ugjf3fvmg/Modal Analysis Examples.zip?dl=0

if you have any questions about the files, let me know.

hooshsim

Prop design,
Thanks for the effort. How would you approach this? Do you run static first or what? Could you explain the procedure step by step please. I have a static and modal analysis results of a stationary shroud. What should I do next?
Thanks

prop_design

when you add the 'New Stress Stiffening' it runs the static analysis then the modal analysis. so it does it for you. what i usually do is just save the static analysis as another file with a different name. then switch the solve to modal analysis 2d or 3d. delete all the solution results. only have displacement as the solution. then run the file.

hooshsim

Thanks again

prop_design

yw. i added one new example to the download. when making it, it reminded me of something. if you have custom scripts/cards etc for CCX you will most likely need to delete those too. the modal analysis doesn't like/need them. all of the examples started as nonlinear stress solutions that needed custom cards. however, the modal analysis doesn't use/need them. so i delete those too. i don't know a lot about the cards. victor helped me with them to get the nonlinear solution to work. so if you start doing stuff like that and run into problems, ask victor for help.

i should also mention, you can see that the model setup for some of the examples is very time consuming. that is the reason for starting the modal analysis off with a saved copy of the stress analysis. it can save hours of setup time. it only takes about a minute to convert a stress analysis over to a modal analysis, regardless of how complex the stress model was. it also makes it easy to ensure both models are using the same mesh. so you're not having to try and duplicate a mesh you already did.

hooshsim

I followed your procedure but did not get any where. Can you please setup the attached static run for a pre-stressed modal?

prop_design

sure, here you go

i updated the download again; see post below

JohnM

hooshsim

Thank you sir. Ran it and results look good. One more question:
Is there any option to display von-mises stresses form the pre-stressed solution?

prop_design

hi,

i'm not 100% sure, but I don't think so. i do the stress first then save the file with a different name and do the modal analysis with stress stiffening.

the stress i have been doing is nonlinear. the modal analysis uses a linear stress. there is an equation that calculates the stress stiffening effect from the linear stress analysis. this was common back in the day. however, a lot of programs have switched to just doing a nonlinear prestress and thus not needed the stress stiffening effect equation. both provide similar results and each has it's pros and cons.

others may no more.

oops; i should also say that the mecway and ccx solvers seem to use the same method and both also include the spin softening effect. however, neither includes the gyroscopic effect. so you need to have a rigid shaft. that's not really an issue most of the time. just about everything uses a rigid shaft. the only real devices i can think of that could use the gyroscopic effect are MEMS sensors.

lastly, the example you provided already had the linear stress setup. so that should be the same that the modal analysis is using. so you can get what you want, just not in the same file, as far as i know.

hooshsim

Thanks. Much appreciated

prop_design

hi hooshsim,

i have been trying to find the formulas i was referring to. unfortunately, the stress stiffening effect is hard to find good info on. the ansys manual has a good explanation of the spin softening effect.

i updated your example to include a run using the ccx nonlinear stress solver. i believe that would match best to the prestress modal analysis. linear stress wouldn't have the stress stiffening and spin softening aspects. those are geometric non linearities that should only be captured with a non-linear stress analysis.

also, your original static stress file is using the mecway linear stress solver. so, for comparison purposes, i added a run using the ccx linear stress solver.

anyways, if you look into geometric non linearities, stress stiffening, and spin softening, those would be the topics to study.

here is the link to your updated example:

https://www.dropbox.com/s/vr5iw74aji7q6ky/rod-update.zip?dl=0

ps; a few versions ago, ccx added complex frequency analysis. that includes the coriolis effect as well. however, just for modal analysis. i don't think any fea offers the coriolis effect in a stress analysis. that is something i have tried for a long time to work with victor on. unfortunately, it gets very confusing via email. mecway doesn't offer the ccx complex frequency analysis though. so you need to stick with rigid shaft type equipment. mainly MEMS devices are the only I can think of that could utilize these features. all rotating equipment, that I know of, uses a rigid shaft.

victor was able to include the gyroscopic effect in the mecway linear stress solver. that is the same effect as the coriolis effect. so right now, in mecway linear stress, you can apply two angular velocities. that is something i worked with victor on. however, you can't apply one linear velocity and one angular velocity. it would be great to be able to do that, but it's not available right now. both abilities would be especially useful in a ccx nonlinear solve.

one other limitation with the two angular velocities is that they are oriented about the global origin. so you can't have a propeller or spinning wheel gyroscopic off axis. this is very limiting too.

so we only got so far with it. it's still better than any other fea software i had used up to that point. not sure if any other fea software now has these abilities or not.