Cap plate for Pipe - How can release moments? 18

[X4vier]

Cap plate for Pipe - How can release moments?
Typical cap plates for pipes will have bolts around the pipe, is there a way to release moments in that type of cap plate?

Thanks.

 

[Tomfh]

These threads are the best.

 

[KootK]

So, I'm sorry if my post came across that way.

I wasn't the least bit offended. On the spectrum of rigor here, I clearly fall on the low end of the spectrum. That, really, was why I felt compelled to chime in on a thread that was already deep into the weeds. I fear an outcome here that comes to pass from time to time:

1) Somebody answers OP with the practical answer that represents common practice.

2) Somebody responds with valid, if arcane, technical objection that does not reflect common practice.

3) Because a valid technical objection is tough to argue against, we wind up with a thread that does not accurately portray common practice.

So I wanted my vote counted, negligent as it may be.

I view threads like this as our community building what, oftentimes, winds up serving as definitive guidance for folks that make use of it later. So I like for there to be a lot of honesty in play. Yes to deep dives on technical nuances. Also yes to telling it like it is.

 

[KootK]

I assumed the OP was designing a beam over column and wanted to neglect the moment in the column, but that may not be what he/she was asking at all.

I'd bet a thumb that is what OP was asking. I wouldn't lose any sleep over that.

...valid, if arcane, technical objection...

One last salvo to make sure that this has gone all of the way arcane.

A column with axial load will be softer with respect to apparent bending stiffness than will the same column without axial load. As the axial load of the column approaches Pe, the flexural stiffness of the column approaches zero. This is not something that will show up in a first order model of course. This drop in column stiffness will mean that the column, and the connection, will tend to draw less moment that would be predicted by a first order elastic model.

The drop in column stiffness won't add up to to much until the axial load gets pretty high. That said:

1) Conditions of high load are precisely where one would expect real world problems to manifest themselves and;

2) Proportioning gravity posts to be slender with respect to the supported framing will tend to exacerbate the trend.

 

[SE2607]

I think we should separate a business decision ("I designed it as a pinned connection because this follows conventional practice and modeling it as fixed would take a lot of effort") from an engineering decision ("there is no moment in that connection").

 

[phamENG]

I disagree, SE2607. For the vast majority of us (probably all of us), engineering is our business. Any consideration of technical accuracy should be looked at through that lense.

 

[271828]

Since we're into the arcane, LOL... We should consider the accuracy of the entire process -- regardless of the level of complexity included. If we build one of these frames and test it with known loads, how accurate do we think the calculations will be? One of the guys in the mechanical vibration forum likes to say that every engineer should get the experience of comparing measurements and calculations. We'd probably be lucky to get deflections and moments right within 15% in this case. If this was a research project with very tightly controlled boundary conditions and everything else, we might get within 5%.

Not saying we should be careless, but that takes some of the intensity away when arguing over the details of the calculations.

Bigger Picture, IMO: About a billion lineal feet of such beams and columns have been designed for decades (a century?) without considering the moment going into the column. I don't think a guy in 1970 would typically use moment distribution on one of these, although someone can probably come up with an example. Ignoring the column moment will work out OK in the vast majority of cases. In the tiny percentage of weird cases, consider the moment.

 

[271828]

I just thought of a funny one. Maybe someone has mentioned this in the sea of responses above. These "exact" analyses are linearly elastic. Any time the moment is above 0.7FySx, there is yielding and nonlinear behavior, but that's never taken into account. Then we use Mp for the strength. Exact! LOL

I have 1965 Plastic Analysis class notes from a mentor. In the intro, they describe the deficiencies of elastic analyses, bsaying it's futile to accurately predict basically anything with it because it's so sensitive to small fabrication errors, support settlements, residual stresses, construction misfits, stress concentrations, deviations from perfect geometry, ideally modeled connections (hinge or rigid). [I would add to that list effect of nonstructural components. How accurate do you think that moment frame drift calc is with all those partitions and cladding? Also add unintentional composite behavior. Certainly many others.] The notes state that elastic analysis is "pure fiction." They then talk about how the plastic collapse mechanisms are much less sensitive to the stuff mentioned above, so if we want accuracy we need to move that direction.

Computerization in design offices, starting a little in the 80s, is why the industry didn't keep moving away from linearly elastic analysis. It's a lot easier to run a linearly elastic frame analysis program, get Mu, and compare to phiMn (= phiMp in some cases). That's very expedient for design with our current tools, and it will work. Thus, we do this for practicality, not because that's the gold standard of accuracy.

Again, not saying we should be careless, but let's get some perspective here.

 

[phamENG]

so if we want accuracy we need to move that direction.

And so the universities and academia have. My steel design course was focused exclusively on plastic design. I think dik has said that he learned it in university back in the 60s. But, as you said, the inertia of practice and the lack of widespread failures and structural issues from these design practices have made the switch almost impossible in practice (except for dik).

 

[milkshakelake]

@271828 I agree with you. But being a devil's advocate, we also have better tools than we did before. FEM makes easier work of moment distribution than what past engineers did, where it was more like a specialized skill. For example, we used to always assume that a grade beam gives linear pressure distribution under a footing. With FEM, we see that it's not the case.

 

[271828]

And so the universities and academia have. My steel design course was focused exclusively on plastic design. I think dik has said that he learned it in university back in the 60s. But, as you said, the inertia of practice and the lack of widespread failures and structural issues from these design practices have made the switch almost impossible in practice (except for dik).

It's awesome that you had that in school. I had a section on it also. Long ago, but not as long ago as the 60s. Early 90s.

AISC took a lot of the steam away from using plastic analysis. In every example I've done, Appendix 1.3.3a or 1.3.3c cause the resulting savings to be no more than a member size. That makes it not worth it for me, and I really like plastic analysis for sentimental reasons. It's just cool. LOL

 

[271828]

@271828 I agree with you. But being a devil's advocate, we also have better tools than we did before. FEM makes easier work of moment distribution than what past engineers did, where it was more like a specialized skill. For example, we used to always assume that a grade beam gives linear pressure distribution under a footing. With FEM, we see that it's not the case.

No argument, other than to point out "better" doesn't mean more accurate, as I argued a few minutes ago. The types of analysis we use in design are pretty coarse. They work well enough to keep structures from falling, and are practical, so we run with them. (Kinda reminds me of the approach of ignoring the column moment that is being condemned fairly harshly by some in the thread.)

IMO, the tools are better mostly because they help work flow. If I use manual calcs and there's a change, then it takes a long time to revise. If I'm designing one of the beams discussed in this thread, and a column moves by 2 ft, the change might take two minutes. And we all know how often things change.

 

[phamENG]

Yeah. I think one of the issues is that the vast majority of structures are still controlled by serviceability. Our primary concerns are usually within the elastic region. Do we care about the behavior into the plastic region? Yes. But it's so rare to experience that loading that we consider the structure to be sufficiently reliable using elastic approximations.

I think it's important, though, to not use that as an excuse to simplify everything away (not saying that's what you're doing, 271828). We need to be mindful of the difference between precision and accuracy - but also realize that reducing precision has a higher probability of also reducing accuracy than increasing it.

 

[271828]

Good points, phamENG. No argument there.

 

[haynewp]

I apologize if this was already mentioned, but almost all of the older single story buildings I’ve reviewed used this beam over column connection with a partial extension of the beam into the adjacent bay. The idea was that the moments equalize on each side of the column. But unbalanced roof live load was never considered nor the effects of that into the column below.

If there was a simple shear case as I see mentioned above(I was actually going to bring up the same thing), older engineers also ignored any simple shear eccentricity and just pulled out of the AISC axial tables. Actually, all Risa 3D models our new engineers produce still also ignore that eccentricity. I’ve seen that eccentricity fail the same perimeter column when modeled in RAM where the eccentricity is accounted for.

 

[phamENG]

haynewp - yep. You have to set your end offsets in RISA to capture those. The functionality is there, but most junior engineers are just given free reign to play with it and whatever comes out gets built. At least, that's how I was taught!

 

[SE2607]

Shall we discuss the rotational stiffness of a Simpson CCQ/ECCQ?

 

[SE2607]

I disagree, SE2607. For the vast majority of us (probably all of us), engineering is our business. Any consideration of technical accuracy should be looked at through that lense.

You may disagree, but business decisions are made all the time even in a structural engineering business. It has been demonstrated here that modeling a beam to column connection as pinned is unconservative, yet it is a common practice and no few structures have collapsed due to that assumption. Are the loads conservative (particularly roof live load)? Would significant yielding have to occur before a failure occurs and the analysis was based on ASD?

However, assuming pinned connections and designing to a CSR = 1.00 is a dangerous practice in my mind. A failure would be difficult to defend, and I'm not talking about in a court of law because few of these issues ever go to court. They end up in mediation where both sides have their expert witnesses and the insurance carrier will avoid going to court at (just about) all costs, so they would settle, particularly if the litigant's expert brings up this issue.

Making a business decision based on technical reasoning, as KootK and others have done, is valid to me. Saying "there is no moment transfer" is not.

 

[WinelandV]

Position:

If your column is failing when you account for the shear tab eccentricity, you're designing your columns too close to unity.

Please note that is a "v" (as in Violin) not a "y".

 

[271828]

Position:

If your column is failing when you account for the shear tab eccentricity, you're designing your columns too close to unity.

As with the question that's the subject of this thread, every firm I've worked for (big reputable firms) ignored the simple shear connection eccentricity. Just design the column for axial in 99% of the cases. If the case is weird enough, look harder.

There are lots of conservative assumptions. The column is continuous with the column above, which has lower load, and there will be some moment continuity even at simple shear connections, so K = 1.0 is conservative. At the bottom level, one could use Gbot = 10 and calculate an even smaller K. For W-shapes in typical buildings, KL/ry controls, and the eccentricity is small -- column tw/2. etc. etc.

 

[haynewp]

Whatever moment restraint offered from the beam to the column through the simple shear connection, which is also imparting the eccentricity, is what I have figured was helping to stabilize the column and why the many buildings designed before software ignored this eccentricity at the edge cols and didn't collapse. Agree on special cases, however, need to be considered.