Structural repair or replacement?

[prblmfxr]

Analyzing an existing 30 year old grandstand bleacher structure that seats 2000 people (20 rows of bleachers total). The tallest columns in the structure (17.5 ft long, angle 3 1/2 x 2 1/2 x 1/4) show signs of buckling or deflection of the tops of the columns (about 1 1/2 in. from plumb line at top of column vs 0 in. from plumb line at midspan) about the strong geometric axis (the column is unbraced in this direction, lb=17.5 ft, max vertical load = 4.5 kips). Bracing is provided in the weak geometric axis direction at 5.5 ft intervals, no noticeable deflection in this direction. Computer analysis shows that these columns are undersized due to the very large slenderness ratio about the "weak" principal axis. However, I'm not exactly sure how the slenderness ratio should be calculated, since this angle column potetially can buckle about 4 axes: both principal axes and both geometric axes. The bracing in the weak geometric axis direction would provide some support for buckling about the principal axes, but the question is, how much, and what is the true maximum slenderness ratio?

In addition, any experience out there for welding another section to this angle to adequately increase the section properties. I'm thinking of welding another angle to this angle, so the final shape will look something like a channel section. Any problems with welding to this already curved angle?

 

[ishvaaag]

Old and flawed since inception (column angles?).

Respect analyisis, I would do a 3D analysis with P-Delta, at least elastic. You might decide to also account for material nonlinearity by reducing bending and axial stiffnesses, but since you are going to revamp the thing and so slender may be it is not even worth.

After that you have only members or segments of with known forces at ends (P-Delta accounted for) amd only in-segmentr additional P-delta to be considered. If the in-segment out-of-straightedness is less than 1/1500 the equations in LRFD (and I assume ASD) will directly account for all the necessary effects.

But since some parts of memebers are already curved, you may elect to directly represent the members by a greater number of segments. When the initial geometrical imperfection is under L/1500 you will be in the conditions of prceeding along the code checks without any need to do more. Otherwise you should account the in-segment P-delta buckling effects separately. Since curved, arch segments, with all the arch buckling implications...and surely lack of engineering data, for angles. Hence it may be better to assume a kink in a way the worst effect is covered, and let the automatical analyisis of P-Delta cover for the effects. It is obvious that the simplifications won't be able to cover all the feasible failures, form the required inferences.

How to forfeit this? You don't want curved columns, do you?
Then engage in straightening/substituting the columns whilst supported, then weld to form preferably boxes, these are columns. Then you will have straight elements and whenever the initial imperfection for your segments is under L/1500, LRFD checks for the results given by an analysis with P-Delta will be OK. If bracing opposes your new setup, make temporary substitution of it as well and replace in way in accord with your new design.

 

[redhead]

Single angle struts require a special analysis because the load is usually not applied through the shear centre of the section. See AISC-ASD Manual 5-309 thru 5-323.

 

[prblmfxr]

Thanks for your response. I plan to weld another angle to this angle so that the final section will look somewhat like a channel section. We cannot remove the existing curvature in the angle.

Any thoughts on whether welding to this already curved angle will weaken and/or make the column more curved?

 

[ishvaaag]

Welding a loaded column is always risky to some extent. I understand you pretend more or less make your weld a longitudinal seam (how to in the curved part -intercalation of an orthogonal tab?- yet to wonder). My main and best advice is that you discharge to soil itself the load at the column you are making welds. If single story and at so moderate loads this may be made very economically by ordinary quickly removable scaffold column items and will forfeit crushing hazards by buckling or accidental shearing of the being welded column.

 

[Zulak]

If you orientate the angles to form a "new" channel section, try connecting the two angles by welding plates across the back faces every 18" or so. This will effectively give you a tied column section, and the edges of the angles don't necessarily have to touch each other (to accomodate the curve). Of course, surfaces should be properly prepared prior to welding.

 

[evelrod]

"OLD AND FLAWED SINCE INCEPTION (COLUMN ANGLES?)"---In Spades !!!!!!!!

Very nice reaserch in the posts above, but that doesn't change the facts. I will agree that you can weld another angle, toe to toe and probably make it work (with tabs where the bend is). BUT WHY? I have had the unfortunate duty to pull workmen from under a collapsed platform made of, as it happened, 3 X 3 X1/4 angle, and it was only a temporary structure not nearly as loaded as your grandstands. Take the time NOW to replace these weakened columns with tubesteel and save yourself a LOT of grief later. It is just NOT that difficult to substitute square tube columns or indeed, to add them to the existing angle (assuming you are not concerned with how it looks and only in how it performs). The tubesteel is approximately the same cost and much less difficult to weld in place!
My personal opinion---anything less is at best, MICKEY MOUSE, and at worst, CRIMINAL! 2000 libel suits, thats a lot of 'PROBLEMFIXING', eh?


Just my opinions ,with respect to you all, who are probably much smarter engineers than myself, anyway.



Rod