Another Seismic Question

[Lion06]

I'm using an OMF in a high seismic zone (It's a one story home), and I'm looking at the connection design.

I'm using HSS round columns with WF beams and I'm looking at the connection design. I see that the hinge can be in the column or the beam for a OMF, and that the connection needs to accomodate 1.1 times the capacity of the member that the hinge will form in (including the ratio of Fy,avg/Fy,min).

It also looks like prequalified connections are not required for OMF, just that the connection capacity must be adequate per above.

So, here's my situation. My column and beam capacity are reasonably close, but the hinge forms in the column. I go through and determine what the connection capacity needs to be. All is good so far. I'm using HSS through-plate moment connections for the pipe column to WF beams. Now I design the plate thickness and bolts to transfer the momet - I picked larger bolts so I have a reasonable number. Next I check block shear of the flanges of the beam, and, of course, it fails. It's more than adequate for the demand, but the capacity is now less than the column.

Since block shear is a brittle failure, I'm assuming that this means that I need to either provide more bolts so that block shear doesn't reduce the beam capacity so much, but I can't find anything to confirm that.

Also, I'm unclear from AISC 341-05 if the welds from the pipe to the through-plates need to be full-pen welds for a OMF.

 

[msquared48]

That's why it is better to use WF sections rather than round or tubes - there are pre-qualified details to make the design simpler. I learned that the hard way and that's all I use now.

Good luck...

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[ishvaaag]

11.2a of 341-05 says

"... or the maximum moment that can be developed by the system, whichever is less."

In the same less conservative approach seemed wanted by the code, the connections then seemingly are permitted to be designed for 1.1 times demand. For the maximum moment that can be developed by the system, after an appropriate -office- analysis, and within the scope of what asked of the structure to resist by the code, is demand.

It is more or less the same interpretation than designing in general connections for actual demand, no provided member capacity.

Respect the second question, the paragraph 11.2a (3) says

"Single-sided partial-joint-penetration groove welds and single-sided fillet welds shall not be used to resist tensile forces in the connections" . So it depends on your detail, if it provides only single fillets, these are excluded and you are demanded to modify to either double fillet or some penetration weld.

 

[Lion06]

ishvaag-
The problem, as I see it, is that the failure mechanism of that joint (beam/connection/column) is a brittle failure, which just seems like a bad idea in a seismic zone.

 

[ishvaaag]

As you see by my previous post, I am generally interested in that the statements made by the codes be clear enough for everyone to understand. Certainly, fragile failures are not a good idea in seismic zones. But then, if such was the intent of the code, its contents should be entirely committed to it. Here in Spain we also have (for earthquakes of lower intensity) the ability to detail some earthquake resisting systems in more fragile way, as long as the forces are retained at the higher value concomitant with the lower reduction from elastic response then taken; and in fact I projected that way one of my latest works that subject to one earthquake of intensity close to that prescribed by the code has not shown any visible damage.

The force of the argument in my answer in my firt entry to this thread resides in that in NO case but when expressly exceptuated by the code, must anyone detail something for solicitations higher than those derived from a sound analysis following the guidelines of the code; and where exceptuated, the exceptions must be clear and clarified enough as to there not be any doubt of what is to be made. "That can be developed by the system" is no such kind of clear statement, and if understood within the usual frame of reading the codes, must be read "...whilst analyzing for the structural response to earthquake as prescribed by the present code". The structural response to earthquake as prescribed by the code is Demand. If not, they signal how, and not in smallcaps.

Neither I see reasonable to ask on qualification upon the works of connections. If asked, then as msquared48 one gets impelled to use qualified connections, that must not be in all cases the ideal solution; something like AISC's "or anything that works well" may not be as a bad idea as it may seem, as long the solution be overseen by some expert official reviewing party and projected by one qualified professional.

 

[JoshPlumSE]

Some comments on the code:

That 1.1 factor is an attempt to account for strain hardening. When you combine this with the Ry factor to account for actual yield, this should get you to approximately the "Maximum Moment that can be developed by the system". Make sense?

The prequalification document (AISC 358) uses a Cpr term instead of 1.1. That term is defined as (Fu+Fy)/(2*Fy) < 1.2

If you're using pipe or HSS material, the Cpr may end up being closer to 1.2.

If you design an OMF to the maximum probable moment that can be developed by the system, then do you allow the controlling failure method (block shear) to be non-ductile? I think it would be okay as long as your block shear failure occurs for a load above that maximum probable moment....

 

[ishvaaag]

Thanks, Josh. I am sure that the implied intent of the code must be of the kind of what you say or otherwise wouldn't be saying "whichever is less" for Demand for a member providing it is going to be always less than the capacity provided by design in the member, and so wouldn't be any doubt of which is less.

But my observation stands: what thing is that kind of statement? If you have been able to provide in scarce lines some evaluation, why the code does not? One could argue -as I have made- that in no case and by the technical state of the art evaluation of the response, the system can ever develop any maximum mommoment bigger than the demand. So the statement keeps being confuse, because in fact even by my stated interpretation, we would be providing some extra connection capactity above that required by demand, along what is expected for the condition.

 

[Lion06]

Josh, that makes perfect sense. The problem I'm coming across is that looking at gross sections, the column controls (smallest moment), but once I design the actual connection and the bolts, then the beam controls in block shear.

 

[ToadJones]

Lion-
Dumb question as I am not as well versed in seismic design as I wish I was...
Where is this block shear failure occurring?
I cant picture this connection.
Are there thru-plates that go though the HSS in the horizontal plane and attach to the top of the WF's?
Is the shear the result of the moment...tension or compression flange?

 

[Lion06]

Toad-
The block shear is failing the top flange. See the attached sketch.

 

[ToadJones]

Lion,
I guess that's about what I was picturing.

 

[sandman21]

AISC 341 Section 7.1

The design of connections for a member that is a part of the SLRS shall be configured
such that a ductile limit state in either the connection or the member
controls the design.


I would change the beam to a HSS and weld it but that is me.

 

[ishvaaag]

Lion06, by your sketch at the upper coverplate you would be having single-side fillet welds, that are not allowed. Hence there you need to bevel the pipe prior to weld for a complete penetration weld, or something so.

Respect the comment of Sandman21, once properly proportioned, say along Joshs's lines, the connection would be providing a ductile failure in the member, what would be enough, irrespective of whatever the fragile nature or not of the controlling aspect of design of the connection between coverplate and flange.

Respect internal consistence of the codes, I am very doubtful that they always have, at least in the pretended extent. Some days ago I was criticizing at this forum the use in 360-05 (or 10) of reduced stiffness to determine solicitations and then use unreduced stiffness to check the strength of the members. This would be in violation of B1 where it is mandated...

"The design of members and connections shall be consistent with the intended behavior of the framing system and the assumptions made in the structural analysis". Here -when we have required the reduction of stiffness to portrait the solicitations- the check is not consistent with the assumption made in the structural analysis; even if it is with the intended behavior -that would mean an effective OR in the clause opening the interpretation to some kind of "mistery of the law" where the implicit intent would be being -through not being explicit in the means- only known in practice by someone else -the codemakers- than the structural designer.

Now assume you make one change that makes analysis and intent consistent. I frankly doubt than after that the purported statistical analysis supporting the uniform reliability of the structures remains intact (respect that I doubt it has ever been for our quite mathematically intractable problem). Any change introduced in negotiations to adapt to international codes -common in Eurocode things- or new practices has the same effect.

Now undertake "calibration" to ASD "to produce the same effects or designs". That would be some kind of backwardness. Tantamount to say that what was being practiced 50 years ago should be kept untouched for practice today.

And then, and returning to the overstrength things, whomever has time for it please read something I was making around 2002 on moment connections in seismic setup; there I was clearly hinting that forces over material overstrength may be appearing on the seismic thrust (something on the line of the 1.1 factor but quite likely even bigger) and then that only what in current technical wording is called a mechanical event simulation with proper constitutive laws in place may identify what happening at the connection.

Again, with any reinforcement of the joint, you are violating whilst proportioning the assumptions made in the analysis, for some cases, attracting bigger forces out of rigidity. So either you go backwards and iterate or just subject the whole detailed structure to dynamical event simulation ... something not well amenable to response spectra procedures, precisely.

 

[Lion06]

Sandman-
The client is very sensitive to the actual process used. A lot of this is coming out prefabricated (fininshes and all), so minimizing the amount of finishes left off is a driving factor.

ishvaag-
When I think about a double sided fillet weld for this case, I think of outside the pipe and INSIDE the pipe. The inside the pipe fillet is impossible, so I was counting on needing full pen welds from pipe to plate at all connections. Additionally, if I don't match the capacity of the pipe then the pipe itself won't limit the design, which, as I read the provisions, is a requirement (not necessarily the pipe, but in this case the pipe because its plastic moment is the lower of the two).

 

[ishvaaag]

True, Lion. At all since you have to pass the forces to and from the pipe. If thw pipe would pass through you could count 2 fillets but then would have an issue with the bottom plate, hence you are entirely right.

 

[sandman21]

Lion: Understood I hate when we have to design around finishes.

Ishvaaag:

The net rupture lines drawn by Josh are NOT a ductile failure mode. Net area facture will result is a sudden loss of strength, with limited plastic deformation of the plate.

 

[ishvaaag]

Well, as you see, the gist of my intent is like in other things of these days to keep the procedures as consistent as feasible. I mean, if analysis gives some force, I am one of the less wanting to consider forces in excess over it. To me overstrength imposition = we have not determined the correct demand. I am not partisan of continuous corrections for details, but of comprehensive understanding if feasible; I can live with it but is not my preference.

Respect your comment on fragility of the detail as of now I don't know to which particular detail you are referring to but surely is (on the one you are just thinking of). However, again, if the detail is above any solicitation it has to see (by correct estimate of the demand) I wouldn't have particular worries (other than those the case would make arise in my mind). In the principle of standing earthquakes elastically there must be a number of instances of fragile failure yet all are expected to keep above demand.

And as a last observation, by your own quoted statement ductility can be provided in member OR connection, and a pipe member of thickness proper for seismic action has ample central area to plastify and then form the hinge; I mean, the code wouldn't be making much fuss for the connection in this case being fragile if strong enough.

 

[JoshPlumSE]

Sandman - FWIW: That was not my sketch.

My main point was related to the calculation of the plastic hinging moment in the column. This would likely be closer to 1.2*Ry*Z*Fy rather than the assumed 1.1 from AISC 341. This is based on using the Cpr term from AISC 358 rather than an assumed value of 1.1.

It represents the maximum moment that can be developed by the member. This includes probable yield strength (rather than minimum yield) and an allowance for strain hardening of the material. If you can design the connection to resist this force, then the connection should behave with decent ductility... because it should be difficult for the system to impart a force higher than this value into the connection.

 

[sandman21]

Josh sorry I saw the wrong post. I would agree that using the maximum probable loads would lead to a plastic hinge in the beam or column giving the system ductility. I personally try to have everything perform ductility, I feel that it simply allows another of ductility even if not needed. It’s been a long day and was a long weekend sorry for the misquote

Ishvaaag

Maybe things are getting lost in translation, however your main point is not since we don’t know the loads we should keep a consistent line of thinking and that overstrength is not really needed as a mean of increasing loads. If that is so, you sound like my boss. The issue with this thinking is that it does not correctly portray the behavior of a building during a seismic event. Take for example, Norhtridge, which showed that no matter what the strength of the beam column connection, non-ductile failure could still form. Do we ignore this because the moment frame buildings remained elastic during that one earthquake? We need to take a look at the system as a whole, simply providing ductile in one situation is not satisfactory. Earthquake takes a whole different thought process than just looking at the elastic behavior of a system.

 

[Lion06]

Ry is actually 1.4 for A53 pipe.
Thanks for the input guys.