Fixed or pinned connection 3

[adrianstructures]

I have a question.. why do many engineers model the main mast connection as well as the other connections as being in the structural calculation model as pinned and not fixed?

 

[rb1957]

The assumption is that the joint loses moment stiffness before it loses shear stiffness that a "fixed" joint will yield to a "pinned" joint and then fail.

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[SE2607]

They can be fixed or pinned... I'd usually use pinned to force the moment connection

This is the epitome of "calling a tail a leg" analogy. Just because you call it a pin to "force the moment connection" doesn't make it so.

What's it actually supported by? Before you assign a lot of moment at the base, you have to know how that will be developed. From the sketch, it looks like a few short pins connect it to a thin slab. If those drilled in anchors are going to pull out, it is foolish to depend on fixity. Too much modeling, too little common sense.

You don't "assign" moment at the base. Nature does that. Our modeling should represent nature, not to make up our own fairy tale.

The assumption is that the joint loses moment stiffness before it loses shear stiffness that a "fixed" joint will yield to a "pinned" joint and then fail.

This is the closest comment to reality I've seen on this and similar topics. You are describing progressive collapse. Wonderful! How do you know that assumption is valid? If/when those anchors pull out, what happens then? Do you still have shear capacity? Uplift capacity?

It really blows me away that there is such volume of comments supporting the position "I'm calling it a pinned connection because I am God!" justifying it as "engineering judgement" (with no data to back this up except for "it's accepted engineering practice"), when, in reality, it's a business decision illustrated by several comments such as

It may only add an hour to the project...but they add up. While not as impactful to me running things solo, my last firm could have employed an extra engineer for the time it would take to do that on every project.

as an example.

20 years ago, few of us had access to structural analysis software like we have today. A simple 2D frame analysis would put this issue to bed. I'm just amazed how lazy we've become intellectually.

 

[JStephen]

The original question was "why do many engineers model...as pinned and not fixed?", not "which is it, really".
Note that if you select either pinned or fixed, you are departing from reality already, as the connection is unlikely to fit either exactly. So it could be modeled as intermediate in various ways as well.
If you neglect deflections in the roof unrelated to this load, you are departing from reality.
If you neglect deflections due to weld distortions and fabrication and roof tolerances, you are departing from reality.
Anything that can be engineered can be engineered to any extent desired. Put NASA on the job, and a dozen engineers could spend a couple of years on it.

 

[SE2607]

JStephen - the debate strategy of extrapolating to infinity is weak.

 

[SE2607]

The original question was "why do many engineers model...as pinned and not fixed?", not "which is it, really".

You are correct!

I think phamENG answered the original question most accurately:

So we make up a plausible story, create a model around it, and then make sure the actual detailing makes room for what will really happen.

In other words, it's a business decision.

 

[BridgeSmith]

What I have been thinking is fixed connection transfers moment to the base hence more conservative than a pinned connection which doesn't transfer moment. Can you please explain what am I missing?

It's more conservative for the rest of the structure being designed to consider the connections to other structures/foundations pinned. All of the moment must be resisted internally.

Of course, it's not conservative for the connection to external structure. The external connections need to be designed with either enough strength or ductility/flexibility to resist or accommodate the rotation of the members coming into the external connection.

 

[dik]

I'm not sure of what you are saying. There is a lateral loading on the system due to the eccentric plan location. This will cause added lateral forces due to the eccentricity.

These lateral forces can be resolved by providing a moment capacity at the base, or alternatively by providing a moment capacity at the connection of the column and the beam over. I often use this latter approach because it's often a lot easier to make this moment connection.

I have no idea of what you are trying to stipulate. Sorry...

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[Howlyn2]

AISC DG #: 29 Chapter (1) outlines the applicability of the Lower Bound Theorem that can be used here. Pending satisfaction of all of the limits states an admissible internal force field - distribution of internal forces in equilibrium with the applied loads - the external load is in equilibrium with the internal force field and ideally less than or at most equal to the "connection" capacity. Therefore, pending the model is working whether it be pinned or fixed, your assumptions led to the result as allowed by the Lower Bound Theorem. Typically, it would be simplest to design this application as a pinned connection to keep stiffness in the frame rather than pushing it into the supporting structure as others above have outlined.

 

[rb1957]

"How do you know that assumption is valid? If/when those anchors pull out, what happens then? Do you still have shear capacity? Uplift capacity?"
you're using a bunch of terms I don't, but let's look at generally.

ok, take a joint fasteners, square pattern, reasonably stable flanges, loaded by direct load (tension), and two moments. So one fastener is getting hit by all three loads. He'll yield first and the moment stiffness will drop significantly. Assuming the fasteners are the weak link, maybe the flanges exhibit more deformation ... more flexibility, less stiffness. But it is reasonable to assume that all four fasteners will hold together and their final load state is tension.

Sure, you can say (correctly) that it is an assumption, that the fasteners could unzip (fail progressively). But a tonne of experience is telling you that this "doesn't" happen, that many other things will fail before you need to consider the moment stiffness of a joint. But work it out for yourself, build your own experience so you don't rely on "heard it on the internet". Personally I think it is a tonne of work to consider a joint fixed, to follow the loads through the structure, and to change is basic assumption and (pretty much) it all over again. Particularly if you have multiple joints and consider each fixed or pinned(neither of which is "true").

Clearly an anchor pulling out is a different mode of failure, presumably quite brittle IDK ... could one anchor of a pattern fail independently of the others ?


"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[TLHS]

You can treat something as a pinned connection if you can, by inspection, tell that the structure is not going to induce enough rotation at the restraint point to mobilize significant fixity out of a joint and that the rotation that is going to happen also isn't sufficient to negatively impact the strength of the joint.

Given any sort of moderately stiff alternate frame action load path, it's honestly surprising how little rotation at a joint is necessary to fully shed moment capacity. There are obviously situations where there are brittle and stiff connections that shouldn't be treated like that, but it's generally fairly obvious if you're in one of those situations.

The sanity check for me, if I'm concerned, is to look at the rotation of the connected elements at the pin joint and what that means in real terms (i.e. the amount of movement at the anchor points). If it's absolutely tiny, like it often is, then you're fine having treated it as a pin. If it's larger then you can start playing with springs to see how much fixity you need to drop that deflection down and if it's realistic.

In the case of this one, I'd need to take a look a the detailing, but if the steel to steel connection is tight then it's likely significantly stiffer in the frame direction than any connection to a roof would be and I'm fairly sure that the rotation at the base connections would be small compared to what a connection screwed into roof deck can handle.

 

[SE2607]

ok, take a joint fasteners,...

Instead of a hypothetical, let's use the example posited by the OP. I'm not that familiar with the metric system, but it appears the larger base plate is about 8"x12" and has 1/2" diameter anchors. It also appears to me that the roof deck is little over 4" thick. I'm not sure the material, but it is probably a concrete slab since it's so thin. I'm assuming the anchorage is by some sort of epoxy. I could be wrong on all of this, but play along if you can.

He'll yield first and the moment stiffness will drop significantly.

Perhaps in the aerospace industry, but not necessarily in the building industry.

Sure, you can say (correctly) that it is an assumption, that the fasteners could unzip (fail progressively).

In the submitted example, I see two modes of failure, the first one being the anchors fail in tension (that's a brittle failure) and the second is the base plate yielding. The base plate is about 5/8" thick and the anchors are about 1/2" in diameter with a pretty shallow embedment. I'm guessing here, but I believe the anchors would fail in tension before the base plate yields.

But a tonne of experience is telling you that this "doesn't" happen, that many other things will fail before you need to consider the moment stiffness of a joint. But work it out for yourself, build your own experience so you don't rely on "heard it on the internet".

Is that experience based on observed test data or that you haven't seen/heard of many failures? I know in the aerospace industry, everything is analyzed and tested before putting something out to the public. However, in the building industry, this seldom occurs. This "tonne of experience" or "engineering judgement" is based on conventional practice. Sure, there aren't many cases of failures. I contend that many of these conditions haven't failed (i.e., exceeded building code allowables) is mostly due to the fact that the connections haven't been subjected to the design loads. Roof live load only occurs when the roof is being re-roofed or when there is a fire and there are firemen and equipment on the roof. My particular concern is seismic loading. I'm in a high seismic region and, particularly with roof top equipment, this design load can easily exceed 1g which means there will be no gravity load to reduce the tension in the anchors. To me, when someone defends a pinned connection design with a 4 bolt anchor pattern with "engineering judgement", s/he really means s/he has made a business decision.

In performance based design (i.e., ASCE 41), there are two components of failure. They are force controlled components and there are displacement controlled components. Generally, force controlled components (mostly connections) are typically designed (based on ASCE 7) for lesser forces than the displacement controlled components (i.e., shear walls, moment frames, etc) can deliver. It is those components that will need to be retrofitted such that they remain elastic if subject to ASCE 41.

One might think that they don't have to worry about ASCE 41, and if one is designing residential or commercial structures or all structures in low seismic areas, they would be probably be correct. However, if the subject roof top equipment was installed at a hospital in a high seismic region, one would probably be wrong. ASCE 41 is being applied, or at least being considered, to all essential structures in high seismic zones.

Back in the "good old days", I used to design moment frames using the portal method with the point of inflection occurring somewhere between 0.4 and 0.6 x the height of the columns. I would design the moment connection for (F/2)x*0.6H and the base connection for also (F/2)x*0.6H.

Today, with the software we have at our disposal, it is very easy to change the boundary conditions from pinned in one model to fixed in the other model. On my models, that takes 10-15 minutes. I would design the moment connection based on the pinned based model and the anchorage for the fixed based model. That way, I don't have to worry about the rotational stiffness of the base connection. Easy peasy.

Those in the building industry who claim "engineering judgement" when modeling a frame with a pinned based connection are really making a business decision and ignoring the rotational stiffness at the base, particularly for a 4 bolt configuration. Most of the columns I design weigh less than 300 pounds, so I try to use a 2 bolt pattern (the line of bolts perpendicular to the frame) where ever I can.

build your own experience

I'm particularly annoyed by those who suggest I need more experience. I've been doing this a long time (those in California can tell that by my SE number). Yes, for much of that I made a business decision to use a pinned base connection. But, that was when the analytical tools were either not available or prohibitively expensive. Even then, however, I built Excel templates based on Kleinlogel's formulas. Limited geometry options and determining drift was tedious and thus, I was forced to make a business decision on those conditions for which I did not have a template. Today, I find it difficult to justify not having at least a 2D analysis program in one's tool box.

My four cents.

 

[rb1957]

ok, I didn't expect a "spanish inquisition" ...

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.