PLOT OF LOAD AND DEFLECTION 1

[Showkat Ahmad]

HELLO,
I HAVE DONE THE NON LINEAR BUCKLING ANALYSIS OF A BEAM. WHEN I PLOT LOAD - DEFLECTION CURVE ITS SHOWING A LINEAR GRAPH WHICH SHOULD NOT BE SO.
PLEASE SUGGEST ME HOW CAN I GET THE NON LINEAR PLOT.

 

[FEA way]

Was the analysis completed or it stopped before reaching full load ? What is the maximum registered load value ? It seems that only linear response before buckling was captured and it stopped when critical load was reached. Maybe more stabilization will help. Or you can increase imperfection scale.

 

[Erik Panos Kostson]

In the upgeom command make sure that your factor creates the desired imperfection (you should be able to look a the nodes to see that they have the imperfection). Secondly I would use displacement control to drive the analysis as it is easier to get convergence on. So instead of a force use an enforced displacement if possible.

As per previous use stabilisation (STABILIZE is the command can be set up also from /SOLU in Ansys) in ansys it is much better than ARCLEN.

Finally make sure you push past its limit load to see instability/collapse and make sure that you included BISO for plasticity for say steel.

Obviously the results are greatly influenced by the imperfection which say in steel design of buildings is given by the code (e.g., EC3). So make sure which imperfection to use and why it needs to be applied (say due to manufacturing)

 

[Showkat Ahmad]

ITS SHOWING BI-LINEAR CURVE BUT I THINK IT SHOULD HAVE A DESCENDING LIMB. THE ULTIMATE LOAD IS AS EXPECTED.
THE CURVE IS ATTACHED BELOW

 

[FEA way]

Is the curve obtained for one node (which one) ? How does the global deformation look like ?

This horizontal part of the curve means that displacements increase without change in load (neutral equilibrium path). For me it indicates that even though this approach shows correct buckling load, it can’t predict postbuckling behavior correctly. To obtain better results I advise you to follow Erik’s suggestions. Especially in terms of the use of displacement instead of load control and adding plasticity and stabilization.

 

[Showkat Ahmad]

HOW TO PLOT LOAD DEFLECTION CURVE.

 

[rb1957]

please learn to turn "CAPS LOCK" off … we don't like being shouted at.

What code are you using ? ANSYS ??

another day in paradise, or is paradise one day closer ?

 

[Showkat Ahmad]

sorry rb1957
I am using ANSYS APDL, From post-processing I have the options of plotting the curve between node displacement and time. But I need a curve between displacement and load.
By the way who told u I was shouting at you.

 

[rb1957]

"By the way who told u I was shouting at you." … that's the commonly understood interpretation of all caps text.

if you can only get displacement vs time, do you have a fixed relationship between load vs time ? then you can plot displacement vs load.

but then that may be a different interpretation of your graph. if your loading is a step function (100% load applied at time = 0), then the graph shows that the deflection increases to some steady state ?

another day in paradise, or is paradise one day closer ?

 

[Showkat Ahmad]

Ok thank you

 

[Showkat Ahmad]

I have the load vs time relation I want to plot the graph between the two. the number of nodes is around 1700 I don't know which node to select for the curve.
your advice will be highly appreciated.
thank you

 

[rb1957]

boy, long time between posts !

"I have the load vs time relation I want to plot the graph between the two." … sorry, what do you mean ?
you have load vs time, and displacement vs time, and you want load vs displacement ? … at some time, what is the load and the displacement ?

"which node to select" … can you plot the deformed shape ? then see which nodes look most interesting … the one with the largest displacement, the one with the largest change in displacement (per unit time), ...

another day in paradise, or is paradise one day closer ?

 

[Showkat Ahmad]

I need the load deflection curve, the time steps itself represent load.
I am using time history plot for the load deflection curve

 

[rb1957]

I still don't understand if you have a problem ?

If you're saying that you're using load vs time instead of load vs deflection,
then you can find the deflection at whatever time from the model output (yes?, do you understand this)
then you have the load and the deflection at whatever time (time from zero to the end, yes?).

Then decide which node to plot.

another day in paradise, or is paradise one day closer ?

 

[Showkat Ahmad]

OK I got your point thank you for the reply