moment in the end of brace pinned with link

[jackrussel]

Hi seismosoft,
According to your guidance in <fbd_inv error with pinned beams > topic, I used frame elements with links at the ends of my elements for releasing DOFs. my problem is that when I control element forces there is a moment around axis2 in the end of braces. I use 4 nodes for defining a link and the axis is coincident to element.

[huffte]

Hi jackrussel. Do you see the same moment in the link as you see in the brace? Did you set zero or very low rotational spring yield moments and stiffness values when you defined the link properties?

[jackrussel]

Hi huffte,
the link force is zero and is ok. but in end of braces I have moment.

[huffte]

Can you double-check the connectivity of the brace and link in your model jackrussel?

Of the two originally coincident nodes on your link - let's call them 1 and 2 - 1 should be connected to the node at the beam-column intersection node and 2 should be connected to the node at the end of the brace. I wonder if you might have mistakenly connected the end of the brace to the beam-column intersection node, thereby inadvertently bypassing the link?

I am just trying to come up with possible ways the problem could occur.

[jackrussel]

huffte, thank you. I have a link between node1 and node2. so it is not the problem. my beams have the same links at the ends and they work properly. I am suspicious about axes of links.

[huffte]

jackrussel,

Specifically, which link formulation are you using (i.e., plastic, bi-linear symmetric, bi-linear kinematic, . . .)?

[jackrussel]

huffte,
page 239 of manual explains about pinned joints. so I use Symmetric linear curve -­ lin_sym.

[huffte]

jackrussel, if you would like to e-mail your model to me, I will look at it as I have time over the next few days. Perhaps I can locate a modeling issue, or perhaps there is a bug which needs to be reported.

[jackrussel]

huffte,
I sent the model to Tim.Huff@tn.gov . M2 in braces is important for me to be about zero.

[huffte]

Hi jackrussel.

I notice 3 things about your model.

1. The braces in the lower 2 stories have significant out-of-plane (y-direction for this particular model) displacements (observe the final state of the deflected shape). This is unexpected as all loads are applied in the x-direction. However, the top floor braces exhibit no such behavior. I suspect that there is an inconsistency in the link definitions somewhere.

2. You applied load appears to be only a partial loading and the acceleration at the end of the record is far from zero (0.107g). The time span of the input accelerogram is only 9 seconds. I pasted the accelerogram into SeismoSignal and there are residual ground displacements at the end of the record. Most records from Northridge had accelerogram lengths of around 20 seconds, at least. This would not create the issues you are having with braces, but I believe it is worth mentioning and I strongly encourage you to be careful when reading accelerograms into SeismoStruct, making certain that the intended record has, in fact, been read in.

3. You have constraints in addition to links and frame elements applied to the nodes at which these three types of elements intersect. It is unclear why you have done this, but this is probably the most significant reason that the brace M2 moments are not zero. You also have constraints at the ground nodes in addition to restraints.

Perhaps a good strategy would be to:

1. Prohibit y-direction translation at the intersection of brace members in a panel. Of course, you must be the one who decides if this is valid for the structure you wish to model.

2. Review your input accelerogram to ensure that it has been correctly read into SeismoStruct.

3. Review thoroughly your use of constraints and see if there is some way you might eliminate most (or all) of them without losing modeling accuracy.

Best of luck jackrussel.