Beam to column connection with eccentricity 4

[Mike1998]

Hi all,

I have a question regarding a beam to column (hinged) connection (see the attached file).

The beam transmits a reaction F to the flange of the column with an eccentricity e from the centerline of the column. Should I check the column with a bending moment M = F*e?

Thank you in advance.

 

[MegaStructures]

I say yes.

“The most successful people in life are the ones who ask questions. They’re always learning. They’re always growing. They’re always pushing.” Robert Kiyosaki

 

[KootK]

1) I do believe that the moment ought to be designed for, just as you've proposed. Whether or not that's common practice kind of depends on where you practice.

2) Through my participation on this forum, I've learned that some counties' standards explicitly direct designers to consider this moment as well as some moment from the eccentricity to the bolt line where the detailing includes that (yours doesn't). I believe that things work this way in AU and NZ. Not sure about Eurocode.

3) In my experience, it's pretty common to ignore these moments in North American practice. And there's some justification for that:

a) If you're not designing your columns to the nuts axially, there's a little margin there to cover this kind of stuff. Less so with automated design now.

b) The depth of the connection and the nominal fixity provided by it will tend to shorten the effective length of the column somewhat. There's truth in that but it's still a bit awkward in that you're weighing a benefit that you don't explicitly quantify in routine design (fixity) with a cost that is tangible (F*e).

The paper shown below supports this perspective to a degree although, in my opinion, it's kind of got the feel of "We like ignoring the axial load eccentricities and would like to continue with that".

 

[dik]

I'm of the 'other camp' that Koot refers to... I usually make sure the moment attachment of the bolts is OK to accommodate the shear load times the eccentricity to transfer the eccentricity of the load to the column centreline... specially if the beam stiffness is much greater than the column stiffness... forgot to add that for HSS sections, I use slip critical connections, usually...

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[MIStructE_IRE]

Yes, you should design for this moment.

I always wonder however that since the beam “applying” the moment is also the same beam effectively restraining the column at the point which the moment is applied, how much moment really materialises. But that’s yet another make & break thesis I’d try if I were back in college now! Regardless, to answer your question - yes.

 

[MIStructE_IRE]

If I were to justify omitting this, I could probably do some version of the attached sketch resolving forces around the bolts (so as you can see, your connection needs to be reasonably stiff, full depth end plate - not a pure hinge style pin!). I’ve taken an arbitrary diagonal line through the centroid of connection, you could probably refine this taking a number of diagonals to each bolt row.

But your code might still ask you to account for the moment anyway - mine does and I’ve always just accounted for it.

Edit - my T tension force might even be at the last bolt row... I dunno, I’d have to think about the plate stiffness..

 

[human909]

I get sick of agreeing with people especially Kootk. But I concur.

Should you check it? YES.
Do I always (and others) check it? NO. (despite code requirement) For pretty much the reasons Kootk has said. There is enough conservatism elsewhere to cover it in the vast majority of cases.

Likewise with beams. The reality is that shear plates (or angle clip plates) still can put a decent amount of eccentric load on beams. In most case this all washes out because any twist due to the eccentric load is counteracted by axial compression in the beam applying the load. Though this isn't always the case, in some cases if not properly accounted for can really cause severe torsion on a beam.

 

[KootK]

Nice. After a week of being pegged as wrong about everything short of my shoe size, I'm down with a little harmony.

 

[Agent666]

I believe that things work this way in AU and NZ. Not sure about Eurocode.

In Eurocode I believe it simply states something along the lines of designer must take account of all eccentricities. Which obviously means accounting for the moment induced in columns.

In NZ and AU codes for 'simple' construction (i.e. nominally pinned connections) connected to the face of columns, we are required to design the column for the shear acting at an eccentricity of half the column depth + 100mm or at the centre of bearing (whichever is greater). The moment is distributed based on the stiffness of column within the storey above and below, the design check is considered on a floor-by-floor basis in isolation treating the top and bottom of adjacent storeys being pinned to determine the moment diagram. The checks governs sometimes with higher axial load utilisation ratios, especially with heavy shear connections. I believe the additional eccentricity to consider is a nod towards pinned connections never truly being pinned, and the point of inflection is some distance past the line of bolts in reality. This additional eccentricity is only considered in the design of the column, not in the connections itself.

It seems like a logical way to check columns in simple construction if your code has no specific provisions.

https://engineervsheep.com

 

[HTURKAK]

The following excerpt from EC-3-8 ( Eurocode 3: Design of steel structures — Part 1-8: Design of joints )..




Apparently, designer's discretion is necessary and the eccentricity of the connection should be taken into account for global analysis..

The following doc. is useful to see how the axial load capacity is affected with eccentricity.



people.fsv.cvut.cz›…of…04…Structural_Modelling.pdf

 

[DrZoidberWoop]

For tallying purposes, company where I work doesn't include a calc for this, just a rotational ductility check with the assumption that the column has been properly sized by the EOR delegating the conx. design.

 

[IDS]

3) In my experience, it's pretty common to ignore these moments in North American practice. And there's some justification for that:

Not saying you are wrong, but what is the justification?


Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[JoshPlumSE]

Not saying you are wrong, but what is the justification?

Laziness based on experience?

My opinion on this is that this eccentric moment should be accounted for EITHER in the column or the connection, but that it's not necessary to design them both. That's my general view with pinned connections in general.

I think we often get away with not doing it for a few reasons:
a) The columns are frequently a little over designed anyway.
b) These gravity only columns are frequently laterally restrained by the slab or deck anyway, decreasing how much they can legitimately move laterally.
c) Usually the pure shear connections like this are a little over designed as well. And, it's really only the top bolt that would experience combined tension and shear (which is where we start to get nervous).
d) Even if you impart tension on the top bolt per your analysis, is it really enough to overcome the pre-tension force in a fully tightened bolt?

 

[human909]

Not saying you are wrong, but what is the justification?

Expediency. If I designed every member to include the eccentricity of the connection it would take me 3 times longer.

As structural engineers most of us make reasonable and often conservative assumptions for the sake of expediency. We don't micro design everything. Just like we make design decisions based on 'pinned' connections or 'rigid' connections neither of which are perfectly rigid or perfectly pinned.

Ignoring eccentricity is unconservative. But for several reasons already listed other conservative assumptions more than adequately cover you in the vast majority of circumstances. The trick is recognising when further analysis is required. And that is where good engineering judgment comes in.

For the sake of the excercise I just threw in the to code eccentricity requirement into the model currently in front of me. I put it onto the beams coming into the most critically loaded column. The eccentricity applied is greater than what it would be in reality. I lost about 2% of capacity. I already have what I consider as comfortable margins on these columns. I also for the sake of modelling convenience haven't explicitly included area reductions on live loads, so that gives me a hefty chunk of extra conservatism for my columns. (I did look at it but I considered the savings of half a tonne of steel in a two columns not worth the extra modelling and documentation effort.)

 

[Agent666]

I learnt a long time ago to always include eccentricities due to poorly noded connections, whether it's a simple shear from a beam or a more complex arrangement in a heavier braced connection. Relatively small moments especially about the minor axis of columns (narrower UB's sections especially) have this habit of using up a disproportionate amount of the available capacity as you're usually sailing quite close to the wind on buckling under axial load about the minor axis.

It's just my preference, I don't care if the design segments have a capacity/demand ratio of 1 at the end of the process, provided I know what loads when into the capacity checks and that I've done my best to account for everything that should be accounted for including eccentricities I'm happy. Could I tell you how many times I had to bump up the column weight or size because of the requirement, not really, because that would involve checking it without the eccentricity and it would take longer....

I've peer reviewed plenty of external projects where people are very blasé about intentionally introducing sometimes quite significant eccentricities into connections and never once checking what the effect is. Many times they based their ignorance of the effects by deeming that there engineering judgement is up to scratch.

Many times the designer is asked to assess further, resulting in increasing member size or modifications to the detailing to achieve sufficient strength for all the permutations of the loading.

Moral of the story as far as I'm concerned is unless you actually check, you really don't get a feel for the effect of the eccentric loads for your particular situation. You don't know what you don't know.

Check several of the more critical locations on a job, and see how it pans out and go from there, eventually getting to a point where your judgement might be applied with some guidance from prior results, rather than ignoring completely! Sometimes it's just a matter of the designers not really understanding what they are doing, others are deliberately ignoring and sticking their head in the sand.

https://engineervsheep.com

 

[Lomarandil]

Alright, help me out here.

My philosophy has always been like Josh expressed. If the connection (and local effects on the column, which I include in connection design) are designed to transmit the forces and moments to the column centerline, haven't I replaced my physical beam with an 'effective beam' spanning from node to node?

In this case, what is left for design of the column?

(I'm a fan of minimum eccentricities to simplify analysis, but in my mind they wouldn't be strictly necessary for a robust analysis).

----
just call me Lo.

 

[human909]

I learnt a long time ago to always include eccentricities due to poorly noded connections, whether it's a simple shear from a beam or a more complex arrangement in a heavier braced connection. Relatively small moments especially about the minor axis of columns (narrower UB's sections especially) have this habit of using up a disproportionate amount of the available capacity as you're usually sailing quite close to the wind on buckling under axial load about the minor axis.

Despite that being a fairly different approach to myself I value that imput. Eccentricities about minor axis of UB column DO make me look carefully.

It's just my preference, I don't care if the design segments have a capacity/demand ratio of 1 at the end of the process, provided I know what loads when into the capacity checks and that I've done my best to account for everything that should be accounted for including eccentricities I'm happy. Could I tell you how many times I had to bump up the column weight or size because of the requirement, not really, because that would involve checking it without the eccentricity and it would take longer....

My approach is pretty much the opposite though I would not argue it is superior. I would much prefer a healthy capacity/demand ratio of >1.5 in my columns and not include every little eccentricity.** Out of curiosity, how are you ensuring that your combinations include the worst possible states of eccentric loads as some eccentricities can counteract others. When dealing with multiple floors the combinations and permutations add up to many dozens.

**The size and type of structures obviously play a big role here. In some circumstances having a raw capacity/demand ratio of >1.5 might $1k difference, in other cases it might be $1mil difference. So naturally the time and effort suitable to eek out everything depends on the job.

I've peer reviewed plenty of external projects where people are very blasé about intentionally introducing sometimes quite significant eccentricities into connections and never once checking what the effect is.

I agree with this. I continually see under designed connections in hollow section struts.

EDIT:

I learnt a long time ago to always include eccentricities due to poorly noded connections

Going back to this. What exactly do you mean by always include eccentricities? Almost all connections have eccentricities from minor floor joists to secondary, primary beams and bracing, where are you drawing the line? (Not that my line is clear, currently the line I have expressed is "where my engineering judgment says this deserves extra checking".)

 

[human909]

You don't know what you don't know.

You are right about this. I know you were talking generally but this certainly applies to me and that is why I read these forums. It often raises pertinent items that once I start pulling on the thread I learn more. For my own benefit I'll chase down the capacity change on some minor axially loaded UB columns in another project and post the results. The % change is probably decently high but the demand/capacity ration is very healthy.

 

[Agent666]

Out of curiosity, how are you ensuring that your combinations include the worst possible states of eccentric loads as some eccentricities can counteract others. When dealing with multiple floors the combinations and permutations add up to many dozens.

Well if you're talking column design, don't rely on your analysis software to design in the first place. Some degree of post processing and hand checking is warranted. For example find worst unbalanced moment due to shear in combination with higher axial cases, check this column in isolation as I noted earlier (I'm guessing AS4100 requirements are identical to NZS3404 here). Keep going until you are happy that eccentricity is irrelevant. If you have a good method of importing design actions from an analysis, it becomes a click of a button to check every combination in post processing. Once you've setup a post processing process you are ok with you can obviously reuse it infinitely....

Almost all connections have eccentricities from minor floor joists to secondary, primary beams and bracing, where are you drawing the line?

In terms of poorly noded connections in steel like I noted, I almost always try account for eccentricities in the design process (whether its member design, weld design, bolt group design, etc), it's just another load to add into the design checks. While we are on welds, I always see people working out total load = total length of weld, therefore let's create the most eccentric weld group we can and as long as it has the bare minimum length of weld we calculated we're ok aren't we..... eccentricities can be important

This raises a question around what defines a poorly noded connection I guess....

I think as an arbitrary/definitive point at which to consider or not consider doesn't exist, but if you consider the requirements for not requiring to assess combined actions in a NZ/AU sense when less than 5% of the axial load capacity is present. Similarly, if you were to say you had bending in excess of ~5% of the member capacity then maybe you should be assessing. But like I said earlier, small eccentricities can cause disproportionate reductions in capacity (members about minor axes, weld group and bolt groups also are especially venerable), but I don't disagree some judgement needs to (and can be) be exercised.

Things like eccentric cleats can fail because of 5mm eccentricities in the load path for example. For years people ignored this because they didn't know any better.

https://engineervsheep.com

 

[human909]

Thanks for your input. Greatly appreciated.

Post processing of column loads in steel design isn't something I have been doing extensively for steel column design. But I would normally delve into the calculations inputs and outputs and check the impacts on some common assumptions which may not be completely conservative. But I'll take your comments on board and explore my assumptions. I find better wisdom on these forums than from my engineering peers.

For what it is worth lots of my designs are probably on the conservative side of things. Though I also see plenty of insanity from some engineers who design things ridiculously overcapacity such as a 2.5m deep truss for a small walkway load spanning 16m. Or another engineer who calculated his foundation based on seismic overturning moment using the mass of a full vessel but then ignored the stabilising effect of the mass of the full vessel. (To his credit the consultant quickly corrected things when I asked the question and the foundation more thank halved.)