Girder Anchorage Hook forces 10

[Quence]

Please see attached image. In airports and open hallways and malls.. you can often see long big secondary beam ends framing into girders. I'm concerned about the details of the anchorage and hooks. This is not often mentioned in structural books. Do you make the hook detail at the edge of the girder? I'd like to know the behavior of the vertical part of the hook.. would the forces be to the left or right? Won't it spall the concrete cover to the left?

 

[hokie66]

Those top bars are nominal, as the secondary beams are normally designed as pinned at that end. I like to place the hooks inside the main girder top bars, and that helps with the spalling you are concerned with. That spalling is generally due to inadequate cover rather than bending of the bar.

 

[Quence]

BAretired stated somewhere that "If the joint is pinned, none of the members can have any moment at that joint. If the joint is fixed, none of the members can have any rotation at that joint. If the joint is free to rotate, then all members meeting at that joint will rotate by the same amount." A secondary beam is moment connected to the girder owing to the secondary top and bottom bars so it can't be pinned. So why do you model it as pinned?

 

[hokie66]

Because it is safe to do so. We don't always match design and analysis. The top bars are necessary, as there is some moment, but standard practice can deal with that moment.

 

[Quence]

See attached illustration. If the ends are pinned.. how about the central girder.. do you model it as pinned too?

 

[hokie66]

Certainly not. The beam is continuous through the girder, so its maximum moment is there.

 

[Quence]

The reason the central girder with beam continuous through it has maximum moment is because it is much more rigid, right?

And the secondary beam framing into end girder has small moment because the edge girder can twist? If so, how do you prevent the twist in the edge girder to create more moments in the secondary beam framing into the end girder?

Just confirming because the books didn't state the above clearly. Thanks very much.

 

[hokie66]

Not necessary to prevent the edge girder from twisting, but you should have appropriate torsional reinforcement in all spandrel beams. The twist of the edge girder will be restrained by the stiffness of the supported beam. I don't know why you would want to "create more moment in the secondary beam" at that point.

The maximum moment I addressed was in the secondary beam, because it is continuous at that point, so tension in the top. The supporting girder also has its maximum moment there, but tension in the bottom. And depending on the stiffness of the two elements, there may also be deflection compatibility issues.

You can't learn everything from a book. You need an experienced mentor, which I hope you have.

 

[Quence]

Just verifying and double checking things the mentor says and things he didnt say. And he is not perfect... even mentors are learning. He is not so familiar with settlement behaviors in columns and beams. If the columns supporting the edge girder settles.. and there is big deflection... most of the loads can transfer to the central girder... how do you handle this? My mentor doesn't want to think of this as he has no experiences about settlements. Do you think beams directly connected to columns would have less drastic loading transfer to the unsettled columns than edge girders settlement issues? Any experiences?

 

[hokie66]

Lots of problems in the superstructure can arise from settlement of the footings. The key is to ensure uniform settlement, and that generally requires involvement of a geotechnical engineer.

 

[Quence]

My mentor was tied up designing a high rise now and has no time for theoretical discussions.. but I just wanna know.. see attached etabs output of the secondary beams on edge girders and central girders. The secondary beams on central girders have obviously huge moments compared to the edge secondary beams. Short of extending the beam at the secondary beam on edge girders. What kind of modification can you make to the edge girders to make the moment maximum too and equal to the central? Would a beam made of diamond do that? What kind of restrain (theroetical and hypothetical) should you do to create maximum moment at the edge secondary beams? This is just to learn the concept of controlling moments at will and not for any practical design.

 

[Quence]

The maximum moment I addressed was in the secondary beam, because it is continuous at that point, so tension in the top

This is the part I need clarification.. there is maximum moment at the secondary beam framing into central girder because it is continuous and tension in the top as you said. But at the edge beam. You also have tension at the top. So why is the moment almost minimum at edge. I reasoned it was because the edge girder twist but make the secondary beam almost pinned.. but you said the twist of the edge girder will be restrained by the stiffness of the supported beam. But why is the moment of the secondary beam at edge girder minimum? I tried asking this from my mentor and he said it is because the program outputs it that way. I want to know the theories behind it which most designers who relied on software just took for granted nowadays.

 

[hokie66]

If most designers just take computer outputs for granted nowadays, that is a sad commentary on our profession. I won't comment on your model, but follow the stiffness.

 

[Quence]

I just want to know why is the edge girder less stiff than the central girder? couldn't they be made similar in stiffness? if you were to make them identically stiff.. what would you do to the edge girder?

 

[Quence]

When we looked at structures.. they appeared very stiff to our flesh and blood perspective.. but by themselves I think structures can feel their own stiffness and where there are more stiffness there are more moments.. I think by making the edge girder bigger or brace it.. you attract more moments to the secondary beam framing it.. the advantage is to let it have more or less equal reactions in camparison to the central girder.

 

[TehMightyEngineer]

Hokie is spot on, loads will follow stiffness in indeterminate structures. Your computer model is very good at keeping track of stiffness and will distribute the loads accordingly. However, it might not know some things that make things stiffer than normal or otherwise will be very dumb about the actual construction. Garbage in = garbage out. We pin ends of beams in models to tell the program to "assume this connection has no flexural stiffness" and thus it will transfer loads to stiffer connections. In reality the connection is not pinned but if it sees some load then it reduces load elsewhere. If it sees too much load then (if your system is sufficiently ductile) it will begin to yield, reducing stiffness, and the load gets transferred to the stiffer elements as your idealized model originally predicted.

Thus, for ductile structures; as long as your load path is complete and all loads are accounted for, if all limit states have sufficient capacity then any redistribution of forces due to member or connection stiffness not accounted for will be acceptable and the structure safe as it can accommodate this redistribution by inspection.

For your setup, yes the edge girder can attract moment even though it's idealized as pinned. However, torsion of the edge girder is not as stiff as flexure in your center girder beam. Thus, the majority of the load will follow the stiffer path and go toward the center beam. We conservatively force more load toward this center beam by idealizing the connection of the edge girder and secondary beam as pinned.

Try this, take your pinned ends out of your computer model and compare the moments between the idealized model with pinned ends and the "realistic" model with fixed ends. I suspect you'll find the moment diagrams fairly similar. If you want, design the floor for both conditions (fixed ends and pinned ends); this is called "enveloping the design" and is used to ensure that your system works regardless of what actually happens in your structure.

Ian Riley, PE, SE
Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/

 

[TehMightyEngineer]

By the way; I applaud you for asking these questions. Your mentor sounds like a successful engineer and it's good to learn from him, but it's equally important to find your own answers and develop your own understanding of why your mentor does what he does.

My mentor doesn't want to think of this as he has no experiences about settlements.

I tried asking this from my mentor and he said it is because the program outputs it that way.

I'm sure you're paraphrasing but my original mentor was the same way. Smartest engineer I know but he had worked for long enough that he just stopped asking "why?" anymore. If presented with new methods, software, codes, etc. he just accepted things at face value. My mentor had the experience to backup their rote acceptance of whatever the black box spit out but, when I was in your shoes, I also asked "why?" a lot and had to crack a few books to get the answers. This has paid huge dividends in my career.

So, bravo for asking; knowing why the answer is correct is far more important than knowing the correct answer.

Ian Riley, PE, SE
Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/

 

[TehMightyEngineer]

but by themselves I think structures can feel their own stiffness and where there are more stiffness there are more moments

This is true. It generally would only not be true in uncommon cases like when small-deflection theory is invalid or you otherwise get non-linear or non-elastic effects (e.g. seismic).

I think by making the edge girder bigger or brace it.. you attract more moments to the secondary beam framing it.. the advantage is to let it have more or less equal reactions in camparison to the central girder.

Correct, making the edge girder bigger will attract more moments from the secondary beam. However, this isn't really an advantage. The flexure in the center girder is far stiffer than the torsion and flexure in the edge girder. Thus, it takes far more concrete and steel to "move" the flexure from the center beam to the edge. The end result is even though you distribute the forces more evenly the cost of your structure is greater than before.

Ian Riley, PE, SE
Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/

 

[BridgeSmith]

Excellent summary, TME. Enveloping some designs as you suggest, and closely scrutinizing the results, is a great way for someone less experienced to develop the judgment for when it is or isn't necessary to go through the process.

I don't know how well it would work for a structure like the one under consideration, but I have attempted to quantify where a beam falls between fully rigid and fully flexible by forcing the same deflection on the beam and the supports. It's akin to virtual work (it may be exactly that, but it's been so long since Structural Analysis class, I can't say for sure). I calculated the force to deflect a CIP abutment cap 1/8" with one of the supporting piles missing, and then forced a 1/8" axial deformation on the pile. I used the ratio of the forces required for each deformation to conclude the abutment cap was 95% rigid. Is that a valid way to determine rigidity and adequate justification to ignore the flexible abutment case?

 

[hokie66]

Thanks for the help, TME. I am just not patient enough to follow through as you have done.