I have a buckling problem where a cylinder is lying down and has a load applied on one side. I was asked to include body loads in the analysis as well as the applied force. Do all applied loads scale for the eigenvalue result, or is there a way to set some loads to be constant and have the analysis only only output an eigenvalue for some of the loads?
I have not found anything about being able to do that so I was thinking I simply have to iterate the body load - start at 1g, then vary it so that at the end of the last analysis the buckling load results include a 1g body load. To properly get the reaction forces from the applied force + body load on the curved surface I was thinking I would copy all nodes on the contacted surface, displace them a small amount, and add stiff compressive gap elements. Then there would be no constraint preventing that section from crumpling in, but it would not be able to move radially out, though the ground, and the reaction forces on that side would increase as the applied force increases.
It is a rather large model, so I plan on making a smaller, simple, model to check things before going all out on it, but I figured I would check in here as well to see what issues I will have with that setup and if there is a better way to go about it. I do plan on checking the final results with a non-linear analysis, though I have not really started thinking about that one yet, since there is also contact to deal with (which I may just turn off and consider the results conservative). We do not have the advanced non-linear solver or DYNA for non-linear contact.
Solved! Go to Solution.
This is prestressed buckling problem. Nastran can keep gravity as constant while calculating critical applied load.
Perfect! Thank you. I figured there would be some way to do it.
So I just get the load into the input file as a 'preload' in FEMAP to make sure it is in the file properly, then edit the case control and use the edited file to run run. That is hopefully fairly straight forward get done.
I made simple example.
Here is simple rod, L = 1 m, D = 5 mm, E = 2e11 Pa. Critical force according to theory is 15,14 N. If i use force of 10 N. I got critical load coefficient of 1,514 in Nastran. If I apply 10 N as preload and 5 N as buckling load, then load coef. is 1,02796 and critical load is 10 + 5 * 1,02796 ~ 15,14 N.
Thank you Karachun. I did something similar with a simple setup that was similar to what I will be working with to make sure I was able to set it up properly on a small model. I hope to have the real one running in a few days.