Transverse Stiffeners and Diagonal Stiffeners

[Buzzbromp]

I have a long deep beam, which has a slender web. The loads acting on this beam are a combination of concentrated and distributed loadings. Flanges are non-compact, but the bending stress induced from these loadings is less than the reduced allowable bending stress. The beam has one-sided transverse plate stiffeners which are more than adequately spaced. The web of this beam is buckled in many of the panels along its length, however, none of the stiffeners have buckled. These buckles occur all along the height of the web. Also, these buckles do not occur in areas of concentrated loadings but of the distributed loadings. Therefore, I do not think the buckling is due to web yielding, web crippling, or the compression buckling of the web as described in section K of the AISC 9th edition. I believe this buckling must be due to the internal shear stress.

If there were no stiffeners, then the allowable shear would be drastically reduced according to equation F4-2 of the AISC 9th edition. With the stiffeners spaced as they are, my Cv value is greater than one, and my value of allowable shear to use according to that equation is the 0.4Fy, which is more than enough to handle the internal shear stress. Is it possible to get inner panel plastic failure, because without the stiffeners your stress is too high? Then the beam is still able to handle the load, because although the web has buckled in numerous locations, the shear allowable is the same because it is transferring to the stiffeners utilizing tension field action?

I made a similar post recently, and i'm just still trying to clarify this. I thought that the intermediate stiffeners stiffened the beam, so that when the load would transfer to the stiffeners not allowing the web to buckle at all. Tension field action (diagonal tension) is the only thing I can think of that would explain why the stiffeners have not buckled, but the web has. I just would think that the web would not buckle in the first place, if the stiffeners were taken the load.

Is it possible that the stiffeners are not sized correctly? I would think that would cause the stiffeners to buckle. Is it possible that Creep or Fatigue could cause this?

Thanks for any input, sorry for the long post, hope it makes sense.

 

[csd72]

Are the concentrated loads over stiffeners or between - if so this could be the cause.

Are the web stiffeners spaced too far apart?

Just a few thoughts

csd

 

[Buzzbromp]

The stiffeners are not spaced too far apart, and the concentrated loads do occur on the stiffeners.

Again, the web is slender, and the stiffeners although their spacing is closer than necessary, I do not think they meet the code requirements for area and inertia. Although, I'm not sure if since they are closer than necessary, you can combine the area and inertias.

Still, I think the stiffeners would be buckled also if they could not take the extra shear.

 

[miecz]

It would be good to put some numbers on this. Sizes of web and flange plates, stiffener size spacing at the buckled location and the minimum spacing, moment and shear at the location, material strengths...What kind of loading does this girder carry? Is it subject to impact? What is the Concentrated load over the stiffener?

Are the stiffeners acting as bearing stiffeners and web stiffeners, or are they only needed for shear?

 

[HgTX]

How buckled is it? Was it distorted even before it was erected? How deep is the beam?

It could be welding distortion from stiffener attachment.

Hg

Eng-Tips policies: faq731-376

 

[csd72]

Could the stiffeners have been put on because of the buckling (i.e. years after the beam was originally installed)?

 

[40818]

Dont know any of the codes etc, but the panels can buckle between stiffeners quite easily, depending upon the modeshape. Shear does sound like the problem, and the allowable shear stress is heavily dependant upon the minimum panel dimension. If a stiffener is not adequately sized it will not offer simple support to the panels and th buckle will not be "held" by it. So it sounds that they are doing their designated purpose by not carrying the buckle. However, as the panels are buckling it seems that there is more likely not enough stiffeners in place (or in the right places), the web is to thin or the shear loads are to high for the structure.

I would rule out diagonal tension, because if it were, then you would know about it, as it doesn't look anything like a typical buckle. Do some basic handcalcs and freebody the beam shearflows and check each section out individually.

There is no such thing as "web yielding buckling", the "web crippling" is confused, as cripling is a shortwave problem, the web will not see this problem before another critical criteria has long given up the ghost.
"Inner plastic failure", only generally seen in composites, and even then very rare. "creep and fatigue" are other non-starters, creep by its nature will not occur in a beam, and fatigue wont cause buckles. I would advise reading up on the subject outside the scope of the AISC to find out how instability works.

 

[40818]

Just had a thought after i posted,

Have a look where your panels are buckling in relation to the beams bending moment diagram and panel shear flows. It is probable that your panels are buckling under combined compression, bending and shear.

 

[Buzzbromp]

Sizes:

The beam is a built up section, with an 84" (clear distance between flanges) deep web with a 0.4375" thickness. This is welded to plates that extend very far in either direction, with other webs down the line, so when analyzing these as a flange width, i used the formula

2*t(flange)*(E/Fy)^(0.5) to calculate the effective width.

the thickness of the top and bottom plates are 0.5 inches, which yields an effective width of about 29.5 inches. Material is steel, and since it is older, I used an Fy of 33 ksi.

So,

t(web) = 0.4375", t(flange) = 0.5", b(flange) = 29.5"
h(web) = 84", d(web) = 85"

Stiffeners are one sided plates, which are 0.5" thick and only 4" wide. They extend top to bottom and have a center to center distance of 27".

As loads are concerned, the bottom of the beam is all distributed loading, and their are distributed and concentrated loads at the top. Where the main concentrated loads act, a very large stiffener is in place, and the web is not buckled in this vicinity.

Max shear on beam is about 150 kips
Max moment on beam is about 3200 ft-kips
Also, there is an axial compressive load of about 200 kips (The beam web is braced every 15 feet to another beam with the same loads and geometry, and also the top and bottom flanges are stiffened continuous plates exhibiting no deformation)

The stiffeners are transverse stiffeners, not bearing stiffeners.

The beam is not subject to impact.

Buckles are in many of the pannels, extend almost the entire height of the web, and are as large as 3".

There were stiffeners in the original design, and i don't think the welding caused these buckles. Also, i do not believe this was buckled prior to being erected.

 

[miecz]

With that flange width to thickness ratio, your allowable stress should be cut in half. Are you sure this girder will hold 3200 ft-kips?

 

[Lion06]

Isn't this an excellent reason to never use a section (built-up or otherwise) that is slender or requires stiffeners for shear stresses?

 

[Tomfh]

Which panels do the buckles occur in?

What shape are the buckles?

 

[miecz]

To get back to your original post:

I believe this buckling must be due to the internal shear stress.

I don't see this buckling is shear. The shear stresses are low (4 ksi) and the stiffener spacing is small enough to allow 14 ksi.

Is it possible to get inner panel plastic failure, because without the stiffeners your stress is too high? [\quote]

If this were possible, then a lot of girders would fail in this manner.

Is it possible that the stiffeners are not sized correctly? [\quote]

It appears to me that the stiffeners are adequately sized. That is, they meet the formula for moment of inertia.

 

[Buzzbromp]

The buckles extend almost the entire height of the web, and buckle in plane (no twisting). They occur all along the length of the beam.

Yes, shear load seems fine, but is it possible to get this kind of buckled shape from bending? The flanges exibit no deformation.

 

[miecz]

If the buckling were related to beam bending, I would expect the distortion to occur only on the compression side of the beam in the high moment areas. Would have to agree with HgTX then, the distortion seems to be related to welding. Are the flange to web welds continuous? How about the stiffener welds? As an aside, did you check your bending stresses for local buckling of the flange?

 

[Tomfh]

I'm still confused about the buckling pattern. How many half waves, and what orientation?