Shear Rupture of Uncoped Beam 1

[ChrisNYCEng]

I am trying to find an answer while reviewing calculations of shear connections. In my connections with bolts at webs of supported beams, the flanges in some cases are not coped. Do we need to check shear rupture for the net section of the web even though there are flanges present? I have always said yes, but am being told otherwise by the connection designer. Some other input, or a documented reference would help, I do not see anywhere that AISC specifically says where/where no to perform this check, so I interpret that as check anywhere you shear section is reduced.

Thanks.

 

[TehMightyEngineer]

Unless I'm grossly mistaken; no. Just think how could you shear the web off without shearing off the flange(s)?

Obviously check block shear though.

Maine EIT, Civil/Structural.

 

[KootK]

I vote yes. Imagine if the shear / tension planes in your shear rupture model were replaced by thin air gaps. Would you feel good about hanging the beam from the top flange then? I wouldn't. You would have to shear off the flange however:

1) There would be considerable deformation before you fully mobilized the flange in shear and;
2) The deformation mentioned in #1 might even be enough to initiate and unzipping kind of failure.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[BAretired]

Shear has to be checked whether or not the flanges are coped, but it will govern only on short, heavily loaded beams. The shear capacity Vr, listed in the steel handbook applies to an uncoped beam.

BA

 

[MacGruber22]

Absolutely check it. Flanges only are considered as bracing elements for webs - not load-resisting elements in combination. The elastic stress distribution on the I-section is parabolic through the web, right? So the flanges aren't resisting much shear. The only "resistance" the flanges could provide is that at real web rupture the flanges *might" keep the member from collapsing entirely, and the beam may hang on to the connection like a wet noodle. Interesting to think about, but not practical.

 

[ChrisNYCEng]

I agree with you all about checking it. I relayed this to the connection designer and he said he will get more info as to why it need not be checked, something regarding test info by AISC and prior communications with them. Fundamentally though, I believe it should be checked.

 

[KootK]

If the connection designer manages to send you something convincing, please share.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[nutte]

No, shear rupture of a section with at least one flange is not a valid limit state to be concerned with for connection design. So if the beam is uncoped, or even if it has a top cope but no bottom cope, shear rupture on the web need not be checked. (Of course, you still have to check shear yielding on the remaining portion, and block shear on the web, and all other regular checks.)

As for documentation, the AISC design examples that accompany the 14th edition Steel Construction Manual don’t check it. Neither do the 13th edition examples. Neither do the examples in the silver Volume II companion to the 2nd edition LRFD manual. The one exception is the Volume II companion to the 9th edition ASD and 1st edition LRFD manuals (the green and blue book). There is one example there where they check shear rupture on the web of an uncoped beam. But I’ve spoken with people involved with the creation of these manuals, and I believe this check is there because it was easy to calculate and include, not because it should have been checked. Regardless, we have several years’ worth of documents that have come out since the green and blue book that indicate it is not a valid limit state.

If you insist on checking it, you would have to include the flange area in your area calculation. If you were going to tear out the web, you’d have to take the flanges with you.

 

[RobertHale]

I don't think you need to check this limit state.
If you did try and check it what would you use for the Anv? Do you assume the whole flange, none of the flange, or just the k portion of the flange? In reading the commentary they suggest other examples of where block shear strength limit states apply, other than the coped beam end case (CJ4.3 paragraph 2) but none of those examples ever cross an out of plane element.
For further support Table 10-1 does not consider block shear of uncoped beams. The beam web available strength only considers bolt bearing for uncoped beams. If AISC intended this to be checked I would think they would have gone through the calculation like they did for coped beams. It is not explicit but the weight of evidence is toward not checking in my mind.

 

[KootK]

Nutte / Rob,

You guys have convinced me. I change my answer and no longer believe that shear rupture needs to be checked in this situation. Thanks for the education. Still, I would very much like to understand why shear rupture doesn't need to be checked. The arguments presented above to the contrary are pretty attractive.

I'm not at all convinced that the flange participates in providing shear resistance. Before the flange tips would be fully mobilized in shear, I feel that one of two things -- maybe both -- would occur:

1) You'd have to develop a ton of shear deformation in the web, perhaps enough to cause rupture and;

2) You'd have to yield the flanges locally in bending.

Check out the attached sketch for an illustration. At the very least, I would think that this would be a lousy serviceability condition. Maybe serviceability isn't pertinent here.

I did think of one significant benefit that would be provided by having the flange present above the tear out block. Without the flange, the tear out block requires bending stress on one or both of the failure planes in order for equilibrium to be satisfied. With the flange, the tear out block doesn't require these bending stresses. Rather, equilibrium can be satisfied purely through shear stresses. This is analogous to a building with bracing on three sides. Again, check out the attached sketch.



The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[nutte]

Theoretically, maybe you could experience some kind of shear rupture failure in an uncoped beam. But practically speaking, some other connection element would fail first. The bolts would fail, or you’d shear through the shear tab, or you’d exceed the shear yielding capacity of the beam.

Another item to consider: Have you ever heard of a beam failing in this way? If not, that might be a reason not to worry about this proposed limit state.

 

[KootK]

I agree on both counts. However, if you didn't already know AISC's feelings on the matter, you could apply the same arguments to a coped beam.

If a logical hierarchy of failure modes precludes shear rupture for an un-coped beam, that should able to be demonstrated by calculation somehow.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[MacGruber22]

I don't know if this makes sense, but I will try. After reading the AISC commentary (and 1978 reference) and reviewing the block shear equation, it seems that: The only way there is a chance to initiate block shear rupture, is when there is the possibility of tensile rupture (they are in tandem). If there is no chance of tensile rupture, then the shear rupture cannot mobilize. It seems that if there is a flange, it will prevent tension rupture from occurring, and then a precipitating block shear rupture.

 

[nutte]

If a logical hierarchy of failure modes precludes shear rupture for an un-coped beam, that should able to be demonstrated by calculation somehow.

So try it. Figure your net shear area, including the flanges, and see what you get.

 

[jayrod12]

If I'm understanding MacGruber correctly I agree that there can't be tension and shear block failure on a simple beam connection as I don't see the tension component of the loading.

 

[BAretired]

Live and learn.

BA

 

[MacGruber22]

Check this link out. It is the paper from the 2nd group of block shear tests performed to expand on the 1978 tests. These tests prompted AISC to add "Ubs" factor:

https://fsel.engr.utexas.edu/publications/docs/Ricles, James.pdf

 

[KootK]

Live and learn indeed.

@Nutte: You'll have to hold on to your gauntlet. Since I don't buy the flange participation bit yet, my numbers would be the same as for the coped case and the exercise would bear no fruit. That, and I'm way too lazy for a homework assignment

@Jayrod: if I understand correctly, the tension is an internal stress rather than an external loading.

@MacGrubber: your point ties in with what I was trying to get at with regard to the moment -- or lack off -- on the rupture block. The Ubs factor in J4.3 penalizes the capacity due to non-uniform stress on the tension rupture plane, at least for multiple row connections. I speculate that the presence of the flange minimizes that non-uniformity and might contribute to a higher capacity. I'll read the paper over the weekend. Thanks for digging it up.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[jayrod12]

I still fail to see how a pure shear load (simply supported beam) can create tension and shear block failure. The moment at the end of the beam is zero so there is no internal tensile stress?

 

[KootK]

See figure 1.5 of MacGrubber's paper. It's page six of the document, page 18 of the PDF. I can see how it might cause confusion. Without those flexural stresses that I mentioned above, equilibrium isn't satisfied.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.