Hi everyone.
I'm currently working on replicating the load-displacement backbone curve result from an experiment of a simple circular column subjected to slow-cyclic loading. The specimen is subjected to a constant axial load of 890 KN and a lateral load with a 978 KN servo-controlled hydraulic actuator. The column has couplers placed near the support.
I modeled the column with couplers using SAP2000 with auto hinge property ( P-M-M). The axial load is assigned as nonlinear static ( zero initial conditions ) while a displacement controlled static nonlinear load case for the lateral load ( continue from state at end of NL case "axial" ). My column eventually failed at step 27 of the displacement-controlled load case.
The results between the 2 are quite the same for the yielding load and displacement, however, drastic changes can be seen in terms of ultimate load and strain. Kindly see photos attached for comparison.
What seems to be the problem? Any suggestions Re: use user-defined hinge properties generated from other software.
Thanks!
I think in experimental work you measured displacement by millimeters while in program it's too small (may be in meter).
Furthermore, in experiment you had data only for the end of plastic stage while program provides post failure behavior. You need to match every units
Muhammad Tayyab Naqash
Yes the experiment was done using cyclic loading. However, I read an extension of that paper where the author modeled this experiment thru opensees using pushover analysis. But he used force-based distributed plasticity.I attached a snapshot of his statement where he subjected the frame to monotonic lateral displacement twice. But I don't think It will change my result since my curve will only be mirrored on the third quadrant ( of the curve ).
I'm wondering if doing a force-based analysis would fix my problem.
Also, can you elaborate further why you said I have convergence issue on SAp2000? I attached hinge result from my analysis, and I think it is erroneous.
P.S. Unfortunately, I cannot use Abaqus or other Fem programs of such since part of my study is to conduct global performance of structure, and modeling a 10 storey structure would be difficult using this software.
Thank u so much for the effort of responding. :)
Abdullah Sikar Al-Issawi Apologies, I forgot to mention that the curves are actually load-drift curve so the displacement is actually the drift. I divided the displacement by the total length of column.
Yes, I plotted the data until failure. How should I know the end of plastic stage? However, my problem still is that the load is continuously fluctuating as the drift increases, unlike in the experiment.
Muhammad Tayyab Naqash
Here is my model. I hope you can take a look at it.
Muhammad Tayyab Naqash
I guess that will work. i have SAP2000 v20. Thank you so much!- Badges
- Science topic