Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

problems with the linear buckling analysis

Please login with a confirmed email address before reporting spam

hei,
I am trying to calculate the critical loads using the linear buckling analysis. I have two problems:
1. when the model is not to include geometric nonlinearity, the critical factor computed is negative value, is it valid? What does it mean?
2. when the model is to include geometric nonliearity, do I need to select it in both study's steps? There are different values of the critical factors if I select it only one of the study's steps.

thanks!
Dazhuo

3 Replies Last Post 2012年5月9日 GMT-4 09:18
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年5月9日 GMT-4 02:16
Hi

I have noticed too that the non linear tick, tends to lower the load factor by 1 unit about, and that you can get negative "security factor".

As it's called "linear buckling" so turning on non-linear geoemtry is perhaps an assumption violation


I find this value of "1" quite puzzeling, why tend assymptotically to "1" ? this could also explain the negative values, when you go above the buckling limit, at least it makes some sens for a simple case of a vertical fixed free rod, comparing to standard buckling theory

I would expect a non-linear geometry simulation to give a lower security factor, and say that anything below +1 is critical and will self buckle


It's perhaps worth a question to support to get it explained once.
There is a similar question when adding some damping, how to consider the compelx results, take the abs() values ?
Sorry no reply here, just more questions ;)

--
Good luck
Ivar
Hi I have noticed too that the non linear tick, tends to lower the load factor by 1 unit about, and that you can get negative "security factor". As it's called "linear buckling" so turning on non-linear geoemtry is perhaps an assumption violation I find this value of "1" quite puzzeling, why tend assymptotically to "1" ? this could also explain the negative values, when you go above the buckling limit, at least it makes some sens for a simple case of a vertical fixed free rod, comparing to standard buckling theory I would expect a non-linear geometry simulation to give a lower security factor, and say that anything below +1 is critical and will self buckle It's perhaps worth a question to support to get it explained once. There is a similar question when adding some damping, how to consider the compelx results, take the abs() values ? Sorry no reply here, just more questions ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年5月9日 GMT-4 05:16
Dear Ivar

Do you mean "security factor" is the same concept as "the critical load factor"?

I turn on nonlinear geoemtry in step 1 and step 2 (study 1).
I got the critical load factors in three load cases:
140 for load case 1(=Gd+Qd);
-98 for load case 2 (= load case 1+Fd1);
44 for load case 3 (= load case 2+Fd2).
The result (negative critical load factor) shows that the buckling occurs in the load case 2, right?
But when I continue to add a new load in the neighbour position with the same direction, the critical load factor becomes positive value, the structure "becomes" safe status!
Is anything below +1 critical and will self buckle?

The model is attached.
Thank for your reply!

-dazhuo
Dear Ivar Do you mean "security factor" is the same concept as "the critical load factor"? I turn on nonlinear geoemtry in step 1 and step 2 (study 1). I got the critical load factors in three load cases: 140 for load case 1(=Gd+Qd); -98 for load case 2 (= load case 1+Fd1); 44 for load case 3 (= load case 2+Fd2). The result (negative critical load factor) shows that the buckling occurs in the load case 2, right? But when I continue to add a new load in the neighbour position with the same direction, the critical load factor becomes positive value, the structure "becomes" safe status! Is anything below +1 critical and will self buckle? The model is attached. Thank for your reply! -dazhuo


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年5月9日 GMT-4 09:18
Hi

Nice model, but I cannot really give you "the response" because I see we are stuck with the same issue.

What I can tell you and that I verified with the simple model hereby (its a 4.2a I do not have any older version up running), is that when you add self weight AND a force load, your "critical load factor" (= my security factor, sorry for the confusion) corresponds to a mix of BOTH loads, as these are cummulative.
So I would say start with only its proper weight under 1*G, get the load factor, which then corresponds to N*G. Then remove the body load and have a single force load and get the load factor per Newton (perhaps for forces applied at different places).

Then you need to do a good engineering guess how to combine both results into a fixed self weight part and an additional extra load part. These do not really add linearly, as the self weight is a globally distributed force, while the force load is generally a locally distributed load

And I would run it without the "non-linear geometry" to start with, then the behaviour at least is linear (doubling force means half the critical load factor

By the way a body load scales qith a (SF)^(1/3) law, a force with a linear law

--
Good luck
Ivar
Hi Nice model, but I cannot really give you "the response" because I see we are stuck with the same issue. What I can tell you and that I verified with the simple model hereby (its a 4.2a I do not have any older version up running), is that when you add self weight AND a force load, your "critical load factor" (= my security factor, sorry for the confusion) corresponds to a mix of BOTH loads, as these are cummulative. So I would say start with only its proper weight under 1*G, get the load factor, which then corresponds to N*G. Then remove the body load and have a single force load and get the load factor per Newton (perhaps for forces applied at different places). Then you need to do a good engineering guess how to combine both results into a fixed self weight part and an additional extra load part. These do not really add linearly, as the self weight is a globally distributed force, while the force load is generally a locally distributed load And I would run it without the "non-linear geometry" to start with, then the behaviour at least is linear (doubling force means half the critical load factor By the way a body load scales qith a (SF)^(1/3) law, a force with a linear law -- Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.