Nodal Bracing for Columns

[slickdeals]

Folks,
I am having a discussion with one of my senior engineers regarding the attachment of a nodal brace for a column.

The nodal brace is designed for 1% of the compression load on the column and meets the axial stiffness requirements.

My question is thus:
Does the connection from the column web to the bracing beam need to be a slip critical connection or can it be a bearing type connection with fully tightened bolts?

Even though the standard holes only have a hole diameter of 1/16 more than the bolt diameter, I am questioning whether things will be fine in the period between when the column goes into compression and then eventually into bearing at the brace?

Thanks

 

[mudflaps]

Use the same bolt design as the rest of the structure. If you start mixing connection theory on the job something is bound to get screwed up.

Old CA SE

 

[slickdeals]

Are you then suggesting that it does not make any difference on whether there is movement in the connection prior to it engaging in compression?

I am using bearing type bolts A325 N.

In principle, I totally agree with your KISS concept. []

 

[WillisV]

If you are concerned, just take the (1/16") slip into account when determining the provided stiffness to check against the required stiffness.

 

[271828]

I'm thinking that the additional 1/16" wouldn't change the physical reality enough to worry about considering the coarseness of the overall method.

It just changes *when* the brace starts picking up load--when the column is bent to its initial out-of-plumbness (OOP) or its OOP+1/16". The bigger its initial OOP, the bigger the brace force and stiffness required.

I don't have my seminar notes here, but I think those equations are based on an initial OOP = L/500. I guess it depends on how 1/16" compares to L/500 for your case. If the column is 30' long, the 1/16" wouldn't be adding much.

The way I look at calcs like this, right or wrong, is in the context of how closely we can actually predict what's going to happen. In a lot of cases, you could predict behavior one way versus another, go run a series of tests and who knows which method would correlate better with reality. Splitting gnat hairs?