modeling of knee braced frame

[abazvand]

hi all;
i am verifying siesmostruct model with an exprimental work called"Seismic Design and Behavior of Ductile Knee-Braced Moment Frames",i want to produce the knee brace buckling and plastic hinge at the first and end of the link beam as the original exprimental work, how can i model them.
thanks.

[seismosoft]

Dear abazvand,

We suggest you to start by searching this Forum for "buckling" (complemented with an always useful searching and reading of literature on the subject of buckling modelling).

You should then also "play"/experiment with different modelling assumptions, and withdraw the necessary conclusions.

Seismosoft Support

[Behnam]

Hi
Many thanks for the supportive forum.
I would like to model an Experimental Knee-brace frame with resulting hysteresis loop.(attached pdf )
All i get is a linear hysteresis curve,(which i searched in the forum but no results).(pls find attached the .spf model)
1. i have modeled brace with initial imperfection to consider buckling.
2.Also,i wish to model a plastic Hinge in the Beam-Brace connection ,but i do not know if i should model it with Link element(e,g:st-bl) or not?

Any help is highly appreciated.

http://wikisend.com/download/137052/Model.rar

[huffte]

Hi Behnam. I took a quick look at your model and notice some things:

1. The PDF file shows pinned supports, but you have modeled fixed supports. I removed the ry restraint at nodes 1 and 2.

2. Your braces have rigid connections at the column and the beam. You may need to use releases (which can be a problem with inelastic elements) or a truss element for the knee braces. Or it may be ok as is - you will have to see if you can tell from the report . I added the releases at the brace ends which attach to the beams and columns, not at the internal brace node.

3. Your static time history loading is 32 seconds long, yet you specified a single step in your loading phase definition. I used 1000 steps.

4. I changed the settings to indicate that loads are not to be derived from masses.

5. I added a small amount (1%) of damping.

6. I decreased the load factor on the applied loading. This may not be necessary, but I wanted to see if something might be failing with the full load.

7. I get an error message when I try to run. The message states the node N1' has no member connected to it, which it does not. I had to change node N1' to a non-structural node to get the model to run.

After making these changes the model ran fine and the global hysteresis looked similar to that shown in your PDF file. You may still have to tweak some things in your model to match the test, but this should get you started.

Best of luck, Behnam.

[Behnam]

Dear Tim
Thank you very much.
I did make the changes you kindly said.but the thing is i get 1.8m displacement during plotting which seems irrational.
Also you said"I added the releases at the brace ends which attach to the beams and columns, not at the internal brace node", how can i not add the releases at the internal brace node.would you please email me ur modified model?(behnam.atd1368@gmail.com)
Here is the short report to the experiment.
At the 3% story drift, the mechanism, consisting of "yielding and buckling of knee braces" and "flexural hinging of the beam", formed as intended in the design.
PS:would u pls tell me if i have done the knee-brace model, considering buckling, correctly?

http://wikisend.com/download/123540/Model.rar

Kindly Yours
Behnam

[huffte]

Behnam. I have sent the revised model. I would note that the 1.8 meter displacement should not be surprising to you since your loading history attempts to impose a 5 meter displacement. This is why I previously mentioned that you may need to reduce the load factor. Watch what you're doing and know what you're doing - advice from one who has made plenty of mistakes. Best of luck Behnam.

[Behnam]

Dear Tim
Thanks a lot for your support.
i did make some changes to the model but could not reach to the experimental model hysteresis curve.
Based on the test report ," localized yielding was first detected at the drift of approximately 1.0% in one of the knee braces" but in my model happens sooner.
At the 3% story drift, the mechanism, consisting of yielding and buckling of knee braces and flexural hinging of the beam, formed as intended in the design. but In mine no flexural hinging of beam was seen.
The maximum load at the story drift of 3% was approximately280 kN, but in mine it is about 120 kN.
i guess my frame stiffness needs to be more,but how ?

http://wikisend.com/download/460168/Verifing Model.rar

P.S:Dear,I am really pressed for time and unless i don't get answer from the verifying model i can't go through my main modeling.

Thanks in advance.
Kind regards.
Behnam

[huffte]

Hi Behnam.

1. Your column depth is 260 in the model, but 250mm in the report.
2. Your model brace is a rectangular tube but the report braces are circular tubes.
3. Yield is 350 MPa in the model, but not in the report.
4. The end releases in the braces may need to be tweaked - try releasing M2a only at the lower braces and M2b only at the upper braces.
5. You could apply rigid zone offsets equal to say 0.0625 meters - this is 1/4 of the member depths, and only a suggestion, not a rule - to the members where they frame into other members. Rigid zone offsets are defined at the same location as member releases.

Lots of options, Behnam. You will have to think them through, test your logic, and tweak the model. The discrepancy in max load is the most glaring. Read that full report and make sure you aren't missing something.

Best of luck with your interesting work, Behnam. - Tim

[Behnam]

Dear Tim
I really appreciate your help.your ideas made me closer to my verifying model.
1.The verifying Model Max Drift(%) & Force(kN) are 4 & 300 but Mine are 4.7 & 200.Also the model says Fy=340MPa
2.The knee brace is supposed not to continue after 26s but mine does.(pls find the Gauge.xlsx ,3rd sheet)
3.In the .8m beam(b3) 110kN moment has happened and plastic hinge happened which is not intended.(pls find attached .jpg)
3.Do you think definfing a Panel Zone might help?(since in the verifying model article in an example application,it has been done)
"Simulating the shear resistance of a joint by two rotational springs corresponding to the contributions of the column web and the column flanges surrounding the PZ. and use Rigid eccentricities to specify the finite dimension of the elements"
What kind of link element i should define?(

[url]http://wikisend.com/download/430380/Request(II).rar[url]

Kindly Yours
Behnam

[huffte]

Behnam, I believe that your loading still does not match the experiment. The experiment imposed a displacement which was maximum of 5% drift.

With a frame height of 1.95 meters, 5% drift would be 0.05 X 1.95 = 0.0975 meters, while your displacement history has a maximum of about 0.05 meters.

Note as well that the test you are modeling is the first test, for which the experiment was stopped at 4% drift. Only when the RBS sections were constructed did the frame permit 5% drift.

So this is one major reason that your results are not matching the experiment. You are comparing displacements from your model to drift percentages from the report.

Your imperfections in the braces via nodal coordinates at the center of each brace could certainly change the model results from the experiment.

Your panel zone members may not be necessary once you get the rigid end offsets correct. Your rigid offsets should only be applied to beam ends framing into a column, column tops where they frame into beams, and braces where they frame into both beams and columns. And the rigid offsets should have the correct sign - some should be negative - the ones which are toward the negative direction from the centerline of the element. Beam rigid offsets will be in the x-direction. Column rigid zones in the z-direction. Brace rigid offsets in both the x and z directions. When you get them right, they will make sense in the plot of the structure.

Best of luck.