Bent Steel design

[COEngineeer]

This is a residential project and I am using Risa 2D to analyze this component.

I have a bent steel beam spanning about 23' and the peak at the middle is about 6' high. I will use steel columns to support the beam. These columns will be inside 2x6 framed wall. Now, how would you analze it? Would you say the base are fixed X and Y? How much load at X direction can I have so that the wall can handle?

OR, shall i say one support is fixed at x y, and the other one is fixed only at Y. How much deflection can I have at the support so it is not too much for the framed wall to take?

I figure the steel columns will mostly take Y direction and I dont want to call it a moment connection. Suggestion? I dont do a lot of steel design in residential. Usually only flat beams.

http://www.krmconsultants.com

 

[ctcray]

Well, you could take a cue from wood scissor truss designs which I think TPI limits to .75" Live load and 1.25" total horizontal deflection. Modeling the beam as pin-roller supported. Maybe tighten things up to prevent problems with drywall cracks.

 

[csd72]

The top of the columns will be a pinned connection and there will be no lateral reaction from them.

For analysis you will need to provide horizontal restraint to one point (otherwise the analysis will think its like a car with the handbrake off and will come up with an error).

The lateral deflection at the support is actually your critical design criteria for this one. I would ensure that it is limited to 1/2" at most under typical vertical loads.

Make sure it is well braced to ensure that it is not prone to topple sidewards, and look closely at what happens to the stresses at the bend point.

 

[COEngineeer]

Yes I know I have to restrict x direction at at least one support. W10x36 is strong enough but if I analyze it as pin, roller at base, then one support will have 0.7" deflection at x direction. Is that small enough? How do we know a deflection is too much or not for a wall? What is the deflection criteria for a wall?

http://www.krmconsultants.com

 

[csd72]

A quick google gave me this link which agrees with ctcray:

http://www.mii.com/site/frameset.aspx?siteid=1&langid=1033&main=/page/open.asp?pid=1952

 

[COEngineeer]

Ok, I went trough IBC, for wall with brittle finish, you need L/240. So it really depens on the length of the wall.

http://www.krmconsultants.com

 

[jike]

With a model that has a pin at one support and a roller at the other, if the roller is experiencing 0.7" horizontal movement, then in reality you should see 0.35" movement at each support since there is really no restraining force at the top of the wall. In reality, both supports act like a spring support in the horizontal direction.

 

[COEngineeer]

yeah jike, that what i was thinking too. However, at one support the exterior wall jogs 90degree so I assume one side will not move anywhere so I assume it will move only one direction. Do you agree?

Now Jike, what happen if one support the wall is 10' long and the other support is 20' long, shall I assume one wall moves more than the other like different spring constants?

http://www.krmconsultants.com

 

[jike]

COengineeer:

yes and yes to your questions.

 

[twinnell]

Support the beam on top of the steel columns to provide a moment connection at that joint. Be sure to add a web stiffener at beam bearing. Also fix the base of the columns. This will reduce your outward deflection of the column enough to not have to worry about it. The footing can be sized to take the column moment. Shouldn't be too much to handle.