Top of Wall Bracing in Tension or Compression?

[southard2]

OK so I have a light-gage steel wall that is about 18 feet tall. The top of the wall is braced by a horizontal stud that runs from metal building Frame to metal building frame. It is about 20 feet long. I'm using 600S200-54 stud as the horizontal brace. The weak axis is unbraced. If the stud is considered to be in tension it works just fine. However if I design it as a compression member it fails because of weak axis buckling. That and the load applied to the brace is not applied concentrically. The horizontal brace is actually part of a slip connection. So the vertical stud is not actually connected to the brace. See the attachment.

Normally I design conservatively and just pick a size that would also work in compression. But in this instance I'm helping the architect out of a jam the difference in cost is pretty significant. My gut tells me it should be fine. As soon as the member starts to buckle it would become a tension only member. Provided I design the connection at each end to take the full tension load it should be OK. Since shear shouldn't ever be a problem this seems to be reasonable but is it?

I'd just like to see what other engineer's opinions are of this situation because it really could act both ways to a point.

John Southard, M.S., P.E.

https://www.pdhlibrary.com/

 

[JAE]

southard2 - the attachment was missing.

 

[jayrod12]

I'm not sure how you expect it to act like either a tension or compression member. It acts as a bending member, which is it's own can of worms. If I'm understanding correctly, 18 foot steel stud wall, this is the top track (essentially speaking) that spans horizontally between frames. What is the spacing of the frames? If it's a standard 25 foot bay, I'm not exactly sure how you figure this stands a chance of working.

A sketch of the configuration would help.

 

[southard2]

uggh, here is the attachment.

So the vertical wall is restricted from moving laterally by the two rail studs. The rail studs are then both supported by the horizontal brace. So the vertical wall is locked into place laterally but vertical slip is allowed. The question is about the horizontal brace that currently is spaced at 6'-8" on center. Can it be considered as a tension only member if it fails in compression because of a lack of lateral bracking? That is the question.

John Southard, M.S., P.E.

https://www.pdhlibrary.com/

 

[human909]

Your first post doesn't make it clear but, assuming I understand it correctly, it seems like your 600S200-54 is continues in BOTH directions perpendicular to the plane of the wall. In which case well yeah you can design it as a tension only member for lateral loads and forget about the compression side of things.

No different to rod bracing or many EA bracing designs.

 

[southard2]

Yes the member labeled as 600S200-54 At 6'-8" on center is actually about 20 feet long. Each end of this horizontal stud will attach to side of 6 foot deep metal building wind frames with clip angles. So say there is a 10 psf internal wind pressure on the wall. This horizontal bracing stud would absorb and axial load. I just want to make sure I can assume it acts in tension.

John Southard, M.S., P.E.

https://www.pdhlibrary.com/

 

[KootK]

I'm sold. It's kind of clever really.

 

[r13]

However if I design it as a compression member it fails because of weak axis buckling.

It can be a tension member without doubt, but if it fails in compression, then you have a headache. A buckled member may not return to its original shape, then the next compressive force hits, it deforms more. A deformed member does not perform well in tension, I believe.

 

[human909]

It can be a tension member without doubt, but if it fails in compression, then you have a headache. A buckled member may not return to its original shape, then the next compressive force hits, it deforms more. A deformed member does not perform well in tension, I believe.

Well that depends on your approach to designing the member as a tension only member. Alongside that design choice you need to ensure it either won't receive compression or it won't be damaged from compression.

Just because a member 'fails' in compression and begins to buckle doesn't mean that it will be damaged. Appropriately designed with stiff alternative load paths damage won't occur the member when it buckles. Again plenty of cross bracing falls into this category.

 

[southard2]

When I get into the office tomorrow I might try looking at sort of a strain compatibility approach. The brace axial load isn't actually applied at the center of the brace. Instead it will be about 6 feet from one end where the load is applied. If I assume the compression settlement and the tension stretch are the same then the shorter side would have more strain and thus more stress and a higher share of the load. Then I can check in compression a 7 foot compressed column with a larger percent of load and the longer 13 foot compressed column with a smaller compression load (have to check both ways cause load is reversable). This assumes that the connection on each end of the brace member contrains movement in the direction along the axis. This might be a reasonable way to check to see if buckling in any meaningful way will actually happen. If both lengths of columns work for their share of the axial load than I could assume that the tension side reduces the load on the compression side enough so that it won't buckle.

Let me try to say this another way. If I assume an imaginary deflection. Say the load point upon being loaded moves 1 inch in the direction of the axis. Then the strain on the short side would be 1 in / 84 in. And the strain on the long end would be 1 in / 156 in. This would allow me to figure out what percent of the load is distributed to the tension side of the member and the compression side of the member.

I think this would work. It assumes of course that the load causes a deflection along the length of the axis. In real life I think when the compression side starts to buckle it will twist a bit and bow thereby relieving the strain along the axis in direction of compression. Hence the idea that this could be a tension member.

I think it is worth trying to check the member in this fashion cause retired_13 did a good job of expressing what could go wrong if the assumption that the other sides strain is not redistributed to the tension side before failure. I'll let you know how it turns out.

If the results are favorable I'm going to give the design a shot. If the results of the strain compatibility approach are at all scary then I might just punt (like I usually do) and go with that old expression "when in doubt make it stout".

John Southard, M.S., P.E.

https://www.pdhlibrary.com/

 

[Tomfh]

I agree that it's similar concept to tension X bracing. You only rely on the tension member.

If you're worried then do the analysis with imperfection one side and run a non linear analysis and see what the "buckled" side does under design loads.

 

[r13]

I need to adjust my view on "buckling" per reminder from published paper on continuous mechanics.

Buckling vs Yielding

As stated at the outset, classical buckling analysis is independent of a material's yield strength. This is evident in the above derivation because at no time was stress or strain discussed or compared to a material's strength.

But in fact, yielding considerations should never be totally ignored. Once one obtains an estimate of Pcr, one should always divide it by the column's cross-sectional area, A , to obtain a stress, σx=Pcr/A and compare this value to the material's yield strength to determine if yielding will occur before buckling.

[ADD] If in no time the member will be stressed to yield, then it will function under either type of load - tension, or compression.

 

[r13]

Once one obtains an estimate of Pcr, one should always divide it by the column's cross-sectional area, A , to obtain a stress, σx=Pcr/A and compare this value to the material's yield strength to determine if yielding will occur before buckling.

Actually I do have reservation on that statement, as buckling causes side sway (deflection), thus a moment is generated, and the section will have tension on the outer face, and compression at the inner face. To check the stress, I felt it is more appropriate to divide Pcr by the compression area only, then compare with the yield stress. The difference could be huge.

 

[r13]

I think if you can provide a connection with short slot hole, or oversized hole, along the line of action, it might be able to relieve some of the compressive force down to below P_cr, thus the member can safely work both ways. But your wall shall be able to tolerate the small displacement though.

 

[r13]

Another thinking is depends on how much vertical slip is required/permitted, how about connect the post/wall and the horizontal brace through slotted holes (two), so horizontal movement is prevented, but vertical movement is free. I will assume this is a laterally braced condition, for which you can calculate P_cr using the 13' length only. Will this case work?

 

[BAretired]

A horizontal force is applied to the brace at the third point of the 20' span. If nothing buckles, the short end takes 2/3 of the applied force and the long end takes 1/3. If the end in compression can't resist it's full share of load, it takes what it can and lets the tension end carry the remainder.

If the member is capable of carrying the full load in tension, there can be no failure.

BA

 

[southard2]

Baretired indeed that is the case. Just checked it with Risa3D and that is exactly what happens as I sort of thought it would as noted above as well.

retired13 I also love the idea of using short slotted holes. If placed on each end and then having the bolts positioned on the outside of each of the slotted holes than only the bolt on the tension side of the member would transfer load while the compression side would slip and thus have no load. I'm going to try to work this into my design as a fail safe. The member I've selected definitely works when taking the full tension load. That was a creative idea. Love it.

By adding a clip angle to one side where the rail beams connect I can say the horizontal stud is braced laterally at the load point and when using the actual compression in each segment with this reduced unbraced length it works for strength as well. So I'm good to go.

Just want to say I appreciate everyone's thoughts and ideas concerning this matter. Hopefully I'll never have to deal with the buckling Vs yeilding issue again but I'll know where to start if I have to now.

Thanks,

John Southard, M.S., P.E.

https://www.pdhlibrary.com/