Modeling shear critical column hysteresis

[fakharifar.mostafa]

Hello,

While I have been successful simulating accurate hysteresis behavior of FLEXURAL dominant RC columns with pretty much any type of confining device (e.g. stirrups, steel jacket, FRP jacket, welded stirrups, FRP+steel jacket) in SStruct, I have difficulty capturing the pinched post-peak load carrying capacity and hysteresis behavior of SHEAR critical columns. Examples of the hysteresis behavior for a shear critical and flexural dominant columns could be seen below.

Simulated behavior of flexural dominant column:
https://www.dropbox.com/s/sy6lvtec4wi53 ... s.jpg?dl=0

Shear critical column hysteresis:
https://www.dropbox.com/s/svvr9ej082dl2 ... s.jpg?dl=0

As can be seen, simulated flexural column hysteresis is accurate, however the shear critical column was not the same. Any help and hint on accurate modeling (strength, stiffness and residual displacement) shear critical columns is highly appreciated. Should I use link elements at base, if yes, what link to capture the significant degradation in strength with the negative post-peak stiffness?

Thanks.

[ruipinho]

Hi fakharifar.mostafa,

Yes, no surprise in what you are reporting. Current fibre element formulations ignore shear deformations, and hence the strength degradation and hysteresis pinching that are typical in the response of shear-dominated elements cannot be readily reproduced.

Over the past 10 years or so, a number of researchers have proposed advanced fibre element formulations that were indeed capable of capturing the response of elements dominated by shear deformations (see e.g. papers listed after one googles the expression "shear fibre column elements"), however, and most regrettably, none of such innovative formulations is robust enough for general application (they normally require ad-hoc calibration against experimental results).

I also searched this Forum for the expression "shear deformations" (asking for an 'exact matching' in the search) and found posts that go as back as 2008 (e.g. http://www.seismosoft.com/forum/topic.asp?TOPIC_ID=54), and they all say the same thing; in order to be able to reproduce the cyclic response of shear-dominated elements, it is indeed likely that you will need to introduce a link element at the base (or at the middle) of your element.

Of course, then a second difficulty will be the necessary (empirical) calibration of such link elements, which, once again, would require the availability of experimental results (which, as usual, will have the drawback of being element-specific, rather than of general application).

You can again see what others have done in the past (just add the word "spring" to the Google search string that I mentioned above, and you will find many references), but the difficulty will always be that none of such previously proposed shear springs will be applicable to all and every case.

Best,

Rui

[fakharifar.mostafa]

Thank you really so much Rui for your discussion on this topic.

Does anybody have any recommendation/hint regarding the modeling strategy and link element type to be used to model shear-dominant columns? I have been able to reproduce the pinched response, but attainment of the significant strength deterioration at peak is somewhat challenging, and I don't know what link element could be calibrated accordingly for such purpose.

Many thanks in advance.

[ruipinho]

Hi again fakharifar.mostafa,

According to the Help/Manual description of the 'stl_mp' material model, setting a negative value for the 'a3' parameter should help reproducing post-peak softening.

Regards,

Rui

[rashid]

Dear fakharifar.mostafa

The multi-linear hysteretic model by Prof. Reinhorn might prove of some good use to you. Though the backbone curve does not have an EXPLICIT (by specifying input values for the negative branch) post-peak negative stiffness branch, but this model is capable of simulating such behaviour (Cyclic & Monotonic) by adjusting the input parameters related to stiffness and strength degradation.
You can find some calibration parameters to start with in the following document (Page no. 76);

http://mceer.buffalo.edu/pdf/report/09-0006.pdf

[fakharifar.mostafa]

Dr. Pinho, thank you really so much for your input on this topic. I had forgotten such parameter. That helped. THANK YOU!

Dear Rashid, many thanks for your contribution. Indeed I had the older version of IDARC Manual and had tried to tweak such behavior through multi_lin element. Though that link element when calibrated still simulates mostly flexural response for plastic hinge model. Thank you for your help!

SStruct may include a beam-column element with shear-flexure interaction in their future versions.