Bracing Basics / Question

[sshields]

Hello everyone!

I'm an entry level structural engineer, working on my first few tasks. Everything has been going great and I'm learning alot. Here is a lingering question that I feel was never really explained to me in school, and I can't seem to find an answer, or maybe I'm just not understanding the answers given to me. But here goes...

If you took a simple two dimensional bay, and loaded it laterally at the top with a point load, it seems to me (assuming braced frame design), that bracing would be needed. Now in this discrete case we would only need one diagonal member to act in tension as our brace. But when we flip the direction of the lateral load, we find ourselves having to put a diagonal in the other direction, that will act in tension. My question is under this circumstance, when you either have X bracing or alternating diagonal bracings along a stretch of bays, do you need to make sure the diagonals that are in compression, satisfy the same criteria as any other memeber that would be designed for compression, say like a column?? I've read several posts here trying to find a real simple explanation and all I find are people saying things like, "tension only member this" and "ignore compression that"...but how can we design a member to act exclusively in would type of loading? And ignore another (which in most cases its the compression loading that is controling the member size)? Along the same line (but perhaps not relevant), when does a knee brace become and better option then a diagonal (excluding the added space / wide openness point), are knee braces functionally the same as diagonals and x-braces?? Some of the other posts I've read skip ahead to address the unbraced length question, which is a very intriguing and interesting disscussion by the way.

To me it seems like a no brainer, compression is compression and euler doesn't discriminate which members he decides to buckle. LTB is another failure mode (mentioned above), I understand this concept is more nebulous and even my co-workers have differing opinions on what is considered 'stiff enough' (regarding the intersection at a x-brace, and does that provide in and out of plane bracing)....but thats another post for another day...

I'm sorry if all these seem basic, but I would love to hear thoughts and explanations. I really love reading the posts here, structural engineering is such a great field and just by being a 'passive' member to this site I've learned alot...I know this is alot of stuff, and I'm grateful for any explanations on any topic I may recive. Thanks in advance!

Sincerely,
Scott

 

[MiketheEngineer]

Usually - "X" bracing is used. And the compression member is basically ignored because as you say - load it from the other side - it goes into tension.

SO design the members for maximum tension load and when the other one goes into compression and maybe buckles - so what.

Thats why you will see threaded rods w/ turnbuckles on steel framed buildings. They don't care if one goes into compression.

 

[Ussuri]

X bracing is typically designed for the tension only member. You are correct that one member will be in tension and the other in compression. However for the compression member to fail (ie buckle and get shorter in length) the tension member will also have had to fail and get longer. The tension bracing prevents the compression member collapsing. The tension capacity will be higher than the compression capacity.

When you have braced systems using Chevron or 'K' bracing though you must check both.

 

[sshields]

Ok, so the compression member in an x-brace system won't physically buckle until the tension member fails? (assuming the LTB is not an issue). Also, what about bays with a single diagonal?

 

[csd72]

Scott25,

basically it works like this:

The tension member has elongation by stress only.

The compression member has shortening by compression as well as by buckling.

This means that the compression member will shorten more per unit force than the tension member and will effectively take less load. This occurs more and more as the member buckles.

It is therefore conservative to ignore the compression member and design both for the full force in tension.

 

[sshields]

csd72:

Thank you for your response, but now I have 2 more questions :/

1.) Does the ignored member need to be designed to any criteria or just made the same as the tension member? Does it need to carry x% of load or should it just be checked for LTB?

2.) Equation E3-4 in the AISC 13th steel code has 'L' for compression members in its denominator, and so why does the force in the member go to zero as it gets shorter? Shouldn't it get larger?

I'm sorry if I'm missing your point.

 

[csd72]

Perhaps I will explain it another way.

If the compression member fails by buckling then all the force will go into the tension member, when the force acts the other way then the compression member becomes the tension member and any residual buckling will be straightened out by the tension in the member.

So as long as you design each diagonal as if it was a tension member acting on its own you will be fine.

I forgot to mention that if this is in a significant seismic area then there may be more stringent criteria that you have to follow.

 

[JaredS]

Scott

The stiffest parts of the structure will carry the greater majority of the load. If they didn't then the structure would likely explode apart as different elements deflected incompatible amounts. As a great majority of bracing tends to be very slender, read very small I, equation 3-4 of AISC 13 tends towards zero and the compressive stiffness of each brace is much much smaller then the tensile stiffness. Furthermore because of this effect it is not necessary to design the bracing member for compression, but only for the tensile load.

Alternatively there are braces that can be used in compression. However, these braces will primarily be designed with the manufacturer's help based upon the lateral force demand. They are called Buckling Resistant Braced Frames, BRBF, and are featured in chapter 17 of AISC 341-05, the seismic provisions. To construct these braces the manufacturer will encase the bracing member in a grout filled sleeve. The brace will act independently of the sleeve allowing the sleeve to restrict the buckling action of the member and allowing the member to develop its full yield strength while in compression.

A BRBF in action

http://youtube.com/watch?v=ewkCvK3_f0E

 

[JAE]

csd72 is correct. I would only add the following:

1. Even if the compression member does buckle...so what? The reverse loading will straighten it out when it goes into tension and the possibility that part of the brace's cross section goes inelastic really doesn't matter. (except in seismic and fatigue cases).

2. For a tension only design, you normally would not even bother to check for compression buckling - just design it for tension.

3. For bays with a single diagonal (where there are no other braced lines beyond it in line) and the diagonal is required to resist compression to stabilize the building, then yes, you must design it as a compression brace and the compression check will normally control.

4. Equation E3-4: The L in the denominator is the unbraced length of the compression member. As L gets larger, the value of Fe gets smaller. Fe is the stress at which critical buckling occurs so this makes sense. A large L means less stress in the member is required to buckle it. Conversely, a small L means that Fe gets larger. A short stubby column must be stressed to a very high Fe to buckle it....so much so that the material will yield before it buckles (i.e. Fcr = Fy)

 

[hokie66]

Scott25,

Now that you have been exhaustively instructed in cross bracing, I will try to answer your question about knee bracing. The term "knee bracing" is a bit confusing, becausing knee bracing is used really the same way as a haunch, that is to increase the bending capacity at a node. So while a knee brace normally takes only axial load itself, it contributes to frame stiffness rather than acting as tension or compression members do in a braced building.

One point on the tension braces buckling: when a tension only brace buckles, it has not failed. Buckling is an elastic phenomenon, so when it is again asked to take tension load, its properties are unchanged.

 

[sshields]

Thank you everyone for you help!

 

[JJMT]

How much consideration does everyone give for providing physical room for the compression brace to buckle? Off hand i can't think of a situation were allowing the compression member to develop some loads but. . .