Allowable compression of unstiffened elements 1

[haynewp]

Does anyone know where to find in ASD an equation to find the compressive stress(in bending) of a TEE section that is being bent with the outstanding unstiffened leg in compression, with no lateral restraint? Or strong axis bending of a rectangular plate without lateral bracing?

 

[par060]

I've battled this problem before with the strong axis bending of a plate....the only way I found to get an allowable was to use the procedure in the LRFD manual which gives you a moment that you can convert into a "allowable stress"..by dividing the moment resistance by the load factors....it may not be a prefered procedure but for lack of anything better

 

[austim]

Haynewp,

Tha Australian ASD steel code AS 3990 (which can still be used legitimately for mechanical equipment design), has clauses limiting the unsupported widths of plates in compression.

For flanges in compression the maximum outstand is specified to be 256*T/sqr(Fy) where Fy is in MPa (N/mm^2 or MegaPascals). Anything outside this width is to be ignored when calculating the design properties of the section.

Although this is strictly intended for elements in pure compression, it would at least give a conservative result if used for the web of a Tee in bending compression.

 

[haynewp]

Austim, is there anything in the code that limits lateral torsional capacity. AISC's ASD gives local buckling information but I can't find anything to adjust allowable stress for unbraced length. The built up "TEE" section I have is 12 feet long.

I had also saw the LRFD method, maybe I should just use it.

 

[austim]

Haynewp,

That is a very good question - sorry my response is so long.

You had me scurrying back to the code and to a very good 1975 reference "Source Book for the Australian Steel Structures Code AS 1250" by Dr. M.G. Lay. (Our ASD code has remained essentially unaltered since then). This was not intended as a design guide or commentary, but to provide "background data, explanations and reference sources for those rules...whose origins and reasons for inclusion are not immediately apparent" [If only there was a similar document for every code that has been written since then ]

The Code includes :
Rule 6.1.3 Slender-leg struts When the value B/T for any angle exceeds 208/sqr(Fy) [Austim comment - still metric - Fy in MPa], the calculated stress shall not exceed the lesser of :
(a) 0.50 FY, and
(b) the value of Fac calculated from Clause 6.1.1 [which is the standard equation to allow for flexural slendness effects].

Dr. Lay's booklet has the following to say :

"Rule 6.1.3 Slender-leg Columns
Column buckling is usually associated with flexural Euler-type buckling of the type covered by Rule 6.1.1. However, stocky members may also buckle in a torsional mode by twistng about their longitudinal axis. This mode is actually closely related to the local plate buckling failure mode discussed in Comments 32-35 and 45 [which relate to the effective width formulae I mentioned in my earlier post - Austim]. It is not very common in structural components as the torsional buckling load is much higher than the squash load for practically proportioned columns. In addition, because of its relation to plate buckling it is usually effectively restricted by the Rules of Section 4 [again, the effective plate width rules].

However, one case which can be of practical concern is associated with the use of high strength (Fy>=350 MPa) angles with slender legs. Whereas I and channel sections cannot usually be made with critically high B/T ratios, the simple angle is available with B/T ratios of up to 16. To avoid the possibility of torsional buckling mode (ie local twisting of the angle legs) the Rule limits the maximum permissible stress in angles with B/T over 208/sqr(Fy) to 0.50Fy. It effectively downgrades the yield strength of slender legged angles when used with short slenderness ratios. Many angles are used at high slenderness ratios, and so are unaffected.

[Haynewp - Dr. Lay must have known you would ask about this ] Other cases of outstands on compression members could also be covered by this Rule. However, it would be too severe for a case such the "web" of a T section made by splitting an I beam. In such a case the slender web would be attached to a much stiffer flange. Control of such cases is adequately supplied by Ruyle 4.3 [effective width rules again] which provides an upper "effective' B/T limit of 256/sqr(Fy). The more conservative 208/sqr(Fy) in Rule 6.1.3 is restricted to poorly supported outstands in pure compression.

The theoretical background to this Rule is given in Reference 84 - (Lay, M.G. Width-thickness Limits for steel angles in compression, BHP Tech Bull, 14(2), Aug 70, 24-28)"

Getting hold of a copy of that reference now would be nigh impossible, but Max Lay's commentary would generally be taken as "Gospel" in these parts without worrying too much about going any further back.

If you have read all this - well done. Time for a coffee.

 

[redhead]

haynewp:

Lateral buckling of a narrow rectangular beam is covered in "Steel Structures" by William McGuire - Prentice-Hall.

 

[trainguy]

This might be considered an unhelpful post by some, but in my 15 yrs of experience as a structural engineer, I've come to appreciate the reason why some info. is absent from design codes. It's the code writer's way of suggesting the following:

Don't use a tee upside down...

You can discuss allowables all you want, but it just doesn't seem like a good idea. Whenever I spend a day and a half checking allowables (using the Canadian code S136 for cold formed structural members, I come to the same conclusion.

When accounting for marginal lateral stability locally, lateral-torsional buckling, column buckling, warping, and possible out-of straightness of your ideal compression flange, you suddenly realize that you just spent 12 hours of your client's money proving that this inferior design should have been flagged sooner.

Sorry to be such a downer...

 

[teng22]

I agree with trainguy. I spend almost half day to derive the formulae for design upside-down T beam and another half day to run full finite element analysis to check my formulae. And,it is found that using upside-down T beam is not a good design at all...soory to be another downer

 

[haynewp]

I don't really like upside down tee's either, but we have a client who has been using them for years as lintels for small openings in cmu walls. They have been getting a plate welded to the flanges of the tee. This plate then sticks out and catches the brick. The stem part of the tee just sticks up into the bottom knock out course of the wall (directly above the opening) so you can't actually see the stem. I am trying to get them to use castcrete lintels instead, with an angle expansion bolted to the side to catch the brick.

For smaller openings (less than 8ft) in cmu walls with brick, what are others using for lintels?

 

[Bannu]

haynewp,

For smaller openings (<8 ft) we use trough shaped reinforced grouted bond beams as lintels.

You can see these kind of lintels in most of the CMU load bearing constructions in US.

 

[haynewp]

That's what I have used also. Recently, I started using Cast-Crete. I think the Cast-Crete lintels are a little better in terms of quality control over a standard u block bond beam. They have the punch outs already manufactured in for vertical rebar to pass at the bearing ends, and horizontal bars already formed in the lintel and more can be added if needed. They aren't very expensive either, but I think they mainly distribute to southeastern US states. So far, the upside down tees just don't seem to be worth it in terms of engineering time.

http://www.castcrete.com/products.shtml

 

[redhead]

Some further thoughts on this thread.
1. Tees are typically used as support beams for precast plank across corridors in wall bearing hi-rise construction.
The grouted joint assures lateral stability of the compression stem.
2. The same logic should apply to short lintels - say up to 8 feet masonry opening. The masonry is quite stiff and will provide adequate lateral support by the 2 percent rule.
3. If you use tee sections for truss chords, you cannot avoid having the stem in compression due to bending if the load is applied through decking rather than purlins at the panel points, since the chords are continuous members.

 

[jheidt2543]

Excuse me for stepping in here, but why use Cast-Crete (I am assuming Cast-Crete means precast concrete lintel) lintels with angles bolted on the side? It seems to me that those lintels would be much heavier, need added labor operations to install the angles, be suseptable to fastener pull out and COST much more than what you are trying to replace.

In my experiance here are what I see used for lintels in the Midwest of the U.S.:

1. By far the most used and least expensive, depending on loading, is an angle or two back-to-back on smaller openings. The up turned leg fits in the masonry cavity.

2. Next is an angle, a channel or a wide flange with a plate welded on the bottom to pick up the brick or block above. Since these are larger lintels, the upturned leg of the angle or channel doesn't fit in the wall cavity, hence the plate on the bottom.

3. Next is a masonry bond beam. The only problem here is the the bond beam must be shored in place untill the concrete fill reaches strength. It does solve the potential problem of steel lintels rusting in an exterior wall when they are not properly flashed and you don't have to paint the exposed steel plate of #2.

4. Precast lintels. Personally, I would never bolt an angle onto precast lintel to carry the brick. Remember, the idea is to provide the needed support at the least cost. Why add another costly, labor intensive operation? Also, you run the risk of drilling into the rebar while tring to install the expansion bolts.

 

[trainguy]

Most of the lintel applications I've designed involved the (usually correct) assumption that the lintel's compression flange (for back to back angles, etc) was fully laterally supported, thereby signifying a marked reduction of the doom and gloom of my earlier response.

 

[JAE]

Well, with a Tee - there is a subtle, but hard to calculate, amount of lateral bracing when using it for a lintel. The block/brick itself has some level of stiffness, both horizontally and vertically. For a typical lintel, the triangular load applied over the opening (due to arching) doesn't amount to a lot of bending for most moderately sized openings (24&quot; to 60&quot. Beyond that, I would be a little queezy about using an upturned tee.

We typically use block lintels (or just cast-in-place or precast) for most of our openings - these can be built on the ground and lifted into place - you end up with an opening that is clean - uniform, and doesn't include a different material in the system. If the wall is hidden, the wide flange with plate option is also utilized.

But ---- if there is suspended brick veneer hanging off of a ledge angle attached to the lintel, then the WF is out as it has no torsional rigidity - the block or concrete lintel then works best. This occurs with larger openings. For openings smaller than about 6 feet, a loose brick veneer lintel angle works fine.