Dear Christoph,

As the structure is subjected to compressive stresses after a certain load it losses its stiffness, this particular load is critical load and ANSYS provides two ways (eigen value and nonlinear buckling analysis) to simulate this load….Structural instability causes a large deflection of structure for a small increment in load (due to loss of stiffness) and because of this classical nonlinear FE analysis (in ANSYS) aborts at the onset of buckling and you are not able to enter in post buckling regim.

ANSYS provides some methods to simulate the post buckling behavior

First one is arc length method (enough details are given how to do in ANSYS help)…..other method is nonlinear stabilization method…both these methods are very capable of capturing the zone when the structure snaps….I have personally used them for simulating the post buckling behavior of a thin cylindrical shell…only problem was after the onset of buckling the material ruptured in actual test (after a small load increment) as the deformations were very large and we stopped the test…. we simulated up to this point only….so I do not know how much these methods are capable to enter very deep in post buckling regime….

So in my opinion you would be able to simulate initial portion of the post buckling zone accurately with methods provided in ANSYS...

One more suggestion…instead of force control loading, try deformation control loading slowly…this does not allow structure to deform fast and you may be able to capture the behavior of structure..

Hope that will help.

Naveen

I know there are modified arc-length methods for that type of situation, you can check:

NON-LINEAR FINITE ELEMENT ANALYSIS OF SOLIDS AND STRUCTURES

Borst, Crisfield, Remmers, Verhoosel

Chapter 4.

An alternative is to use elastodynamics instead of trying to solve the static problem. At the very least, switching to a dynamics solver when buckling starts will give you an idea of how the system would evolve in time. I have done that for some problems in which I was not interested in the static equilibrium during buckling, but only the end result.

Dear Christoph,

As the structure is subjected to compressive stresses after a certain load it losses its stiffness, this particular load is critical load and ANSYS provides two ways (eigen value and nonlinear buckling analysis) to simulate this load….Structural instability causes a large deflection of structure for a small increment in load (due to loss of stiffness) and because of this classical nonlinear FE analysis (in ANSYS) aborts at the onset of buckling and you are not able to enter in post buckling regim.

ANSYS provides some methods to simulate the post buckling behavior

First one is arc length method (enough details are given how to do in ANSYS help)…..other method is nonlinear stabilization method…both these methods are very capable of capturing the zone when the structure snaps….I have personally used them for simulating the post buckling behavior of a thin cylindrical shell…only problem was after the onset of buckling the material ruptured in actual test (after a small load increment) as the deformations were very large and we stopped the test…. we simulated up to this point only….so I do not know how much these methods are capable to enter very deep in post buckling regime….

So in my opinion you would be able to simulate initial portion of the post buckling zone accurately with methods provided in ANSYS...

One more suggestion…instead of force control loading, try deformation control loading slowly…this does not allow structure to deform fast and you may be able to capture the behavior of structure..

Hope that will help.

Naveen

Thank you very much for your answers, especially Naveen. I appreciate it.

Dear Naveen,

I have done a non linear analysis in ANSYS using arc length method for a tubular joint. But after doing the analysis the load-deflection graph I plotted is a straight line. I also need to find the ultimate load for the joint. Can you please help me

Dear Prashob

From your description, problem (especially configuration of structure) is not very clear to me. FE simulations will be more helpful if you are interested in behavior of tubes but about  joint in my opinion better to trust on design calculations.

Please elaborate the configuration (if possible put a sketch) so that you may extract more useful information

Regards

The load is applied through the brace i.e the vertical member and both the ends of the chord i.e horizontal member is constrained. I need to determine the ultimate load for the joint using arc length algorithm

Dear Prashob,

What I am getting from the figure provided by you is – you have a higher diameter pipe on which a lower diameter tube is getting welded and both the components are perpendicular to each other. There is a compressive load applied on the free end of lower dia tube…You have made with shell elements…If I am getting correct you have coupled the nodes at the interface…..in my opinion you better to make it integral….

Let say vertical member is tube1 and horizontal is tube 2…

You are interested in the ‘load carrying capacity’ of this configuration…. I can imagine a few failure modes of it….first, tube one may buckle as a column….second, tube 2 may go to plastic collapse (both end fixed and lateral point load at centre)….third, at interface tube 2 may buckle locally if it is very thin….

Actually I am not getting the last part of your question “I need to determine the ultimate load for the joint using arc length algorithm”….As per my understanding arc length method is useful if you want to capture the post buckling behavior of a structure not for determining the ultimate load…

For you what I can suggest….You do a nonlinear buckling analysis of the system with nonlinear material properties of tube 2 (to capture its plastic collapse)….you may use elastic perfectly plastic material also…..give autots,off and keep refining you load near abort for this you may need to do multiple analysis….mean at the abort load you load step must be sufficiently refined….e.g. nsub, 200,300,100…or even more refined…see the deformed shape of unconverged and last converged step…you will know the reason of abort and will get the ultimate load also…

Hope that will help

Regards

ya thanks for your answer. I really appreciate your reply.  I need to find the post buckling behavior of this structure also.

Hi Everyone,

I am modeling thin steel plates under in plane quasi static loading. I have been reading and the arc length method will probably solve my non convergence issue. Any input on how to use tabular lading to define my loading sequence in parallel with using arc length method. Also any advice on the non convergence issue with my model.

Thanks