Gusset Connection Eccentricity Problem 9

[dik]

I have a new project to design bracing for an existing building. The force in the bracing is Cf = Tf = 103K... big forces!

The problem is that the gusset will be secured at the centroid of the resisting beams and columns because there is an existing masonry wall located to the centerline of the connection. The EOR want me to design the connection for an HSS 10x4x0.5 (located at the face of the wall), welding the HSS to the face of the gusset, creating an eccentricity of 2" (from the HSS) + half the thickness of the gusset.

I've designed it using elastic combined stresses, based on a Whitmore width for compression. If it weren't for the compressive load, I'd have used the plastic section. I'm looking for some comments how others would design this for the eccentricity. I've suggested going to plate that is roughly 1"x12", in lieu of the HSS, to minimise eccentricity or possibly using the existing masonry wall for lateral forces.

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

-Dik

 

[JLNJ]

I see it as the simple axial force in the gusset. Sort of like when you squeeze a toothpick between your fingers and make it bend out of plane. The connection at the fingers is pinned, for sure.

Whether or not the HSS10x4x.5 is appropriately sized for the 103k is a separate question. If the brace is long and flexible, the deflection of the member itself will need to converge under P-delta analysis.

I just ran some numbers in RISA. 103k x 2.5" = 21.63'-k. With NO P-delta this causes a 4" (!) deflection in a 25' long 10x4. Yike. If you subdivide the member so that P-little-delta can run then the model doesn't converge. No wonder, the 4" of deflection causes another 43'-k of moment which causes more deflection, and on and on. P-little-delta is often not significant, but this is a prime example of how it works. Just like the toothpick between your fingers.

I just don't see how you can make the gusset stiff enough to suck the moment out of the brace and drive it into the connection.

 

[dik]

Thanks, that was one of my first concerns about the use of this eccentric bracing (making it more flexible in the plane of bending). The heavy loaded bay, the one with the 103K is not cross braced. The lighter loaded one is... with the cross-bracing it is somewhat restaind, and I suspect only a small part of the load will go into it. The loading is +/- 103K. I've already noted noted my concerns.


Ya, 285"-K. I didn't run a deflection check on it, but length may only be 15'-20'. My initial gusset thickness was 1-1/2" and the EOR was concerned that I was attracting too much moment into the connection.

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

-Dik

 

[cliff234]

In addition to the challenges of dealing with design of the brace for the eccentric loading, you’ll also need to look at the secondary stresses and torsion induced on the beam and column due to the brace eccentricity.

 

[dik]

The EOR does, not my circus, not my monkey...

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

-Dik

 

[WinelandV]

dik,

This question isn't germane to your brace eccentricity question, but how did you get involved in trying to sort this new brace issue out? That is, why are you looking in to this brace, and not the EOR?

 

[dik]

My current employment is designing connections for steel fabricators. the original connection had a concentric HSS 8x8 with a factored design load of 103K. Because of an existing brick wall, the new bracing was changed to HSS 10x4, offset, to clear the wall. I designed the gusset plates and the original offset gusset was 1-1/2" thick due to the eccentricity caused by 1/2*(tg+bHSS). As a result of the discussion here, I feel comfortable using a 3/4" thick gusset. The EOR was concerned that the thicker gusset would transfer too much moment to the existing connection. As it stands the EOR will be responsible for the existing connection plus any p-delta effects.

I had used Deker's Option 3 and have revised it to Option 1 with the HSS moment to include a moment equal to 1/2 the thickness of the gusset.

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

-Dik

 

[RangeRock]

It's unfortunate the EOR didn't design the frame as a tension-only system... Maybe a little extra steel tonnage but saves a lot of heartbreak as far as connection design is concerned, particularly given the inevitable eccentricity.

 

[dik]

That was one of the first questions I asked... normally if tension and compression are used at the same time, the compression is about 40% of the tension value. The one with the highest load is chevron braced which may be more problematic.

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

-Dik

 

[rowingengineer]

We had an article in our local steel mag about the effect in she roof bracing and option 3 is most likely. an ideastatica is a great program for this, it is how I do my more difficult connections.

 

[dik]

Do you have any reference to this? rowing...

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

-Dik

 

[CANPRO]

dik, I scaled one of your previous sketches to get rough beam and column size, made some assumptions on the existing loads/connections and ran the connection through Idea Statica - report is attached. I used the 3/4" gusset you noted above.

I ran (4) load cases. The first two load cases have 100% eccentricity resolved within the brace (tension/compression case), the second two load cases have 100% of the eccentricity resolved in the beam/column (tension/compression case). NOTE - I updated the gusset thickness, which changed the eccentricity, but I didn't go through and change the balancing moments...so the results you're seeing are based on loads slightly out of equilibrium (I got lazy, general results still the same).

The overall results are only shown for the worst case. Everything passes but the existing beam connection in the first two cases.

You can see based on the deflected shape that it isn't realistic to assume the beam/column will resist the eccentricity. Far too much movement that will likely be picked up by the brace first.

Goes without saying, but please take this report with a gigantic grain of salt.

 

[dik]

Thanks so much... it appears to be a really neat program. I will revisit the calculations and the approach... never with a grain of salt...

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

-Dik

 

[dik]

canpro... can I trouble you to run this again using only load case 2, which is the one I'm using. I cannot distinguish in the attached output, the results for the individual load cases. Does it output the calcs at the end; I'd like to be able to compare my calcs with a 'real output'. or is this a compendium of all load cases? Also if you can use 300MPa for the gusset strength... sorry to be such a pain, and thanks if you can.

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

-Dik

 

[Deker]

Enable, human909: No problem, glad the sketch was helpful.

Dik: Seems odd that the EOR reduced the brace size from an HSS8x8 to an HSS10x4 given that the 10x4 not only has to resist the original axial load that the 8x8 was designed for, but an additional weak-axis moment as well. Any indication that the EOR considered the moment in the brace? And how exactly do you coordinate your design with the EOR on something like this? Do you provide a FBD of your gusset to ensure that a consistent set of forces is used in the brace design? Situations like this make me grateful that we design our own connections where I practice. Recently had my first experience doing delegated connection design on a local project designed by an out-of-state EOR...they forced me to use the max UDL reactions even after I went through the trouble of determining the real beam reactions using the load criteria listed on their drawings. What a headache...

rowingengineer: Since you mentioned the article was for roof bracing, any chance it was based on tension-only rod bracing rather than the stockier HSS that dik is dealing with?

 

[CANPRO]

hey dik, updated report is attached. Load case 2 only - brace in compression, eccentricity resolved in the brace. I adjusted the bending moment in the HSS to get equilibrium.

The calculations only cover the governing case - i.e. if bolt #7 is governed by load case 5, that is the calculation you'll see. This program does not handle flare bevel welds at all - it is treated as a fillet weld...so for the HSS to gusset weld, you need to take the demand on the fillet weld from the program and evaluated manually.

No trouble at all on my end, its easy to tweak this and re-do the output. Once you get over the learning curve in this program you can model a connection like this in minutes.


EDIT: Deker, I had to circle back and give you start for the FBD's. I wish I could sketch/print that neatly. There was a thread a while back discussing how useful a simple FBD can be, this is the perfect example....you've taken a problem that is not easy for everyone to visualize and made it clear as day with a few simple sketches.

 

[dik]

Deker... my error, the original was HSS 6x6x0.5 and the loading was concentric. I checked the HSS 10x4x0.5 and it's OK for P/A and M/S; I didn't check the original HSS 6d6. The HSS 10x4 is loaded eccentrically. The weak axis moment is actually fairly high. Checking for member strength is outside... If I see something that is wildly out of whack, I note it. My SMath program checks for elastic strength (not buckling, though). Even with the HSS 10x4, I note that any bracing required to the existing wall for buckling or any P-delta moments is by the EOR.

I simply stipulate on the drawings that the EOR is to confirm existing structure is capable of resisting loads and, in this case, the gusset will be 3/4"? thick and connected concentrically at the connection, for the EOR to confirm, the connection is the case.

Regarding the other issue, for UDLs I have an SMath program (also an Excel one) that calculates the max reactions based on UDL and spits out the forces. If there are a lot, I use the Excel one since I cannot get SMath to load from my AISC/CISC database.

As far as redesign, there is some interesting case law (in Canada, anyway) regarding stipulating design loads and using some other criteria, not stipulated, or not clear. Generally with contract law, if there is an ambiguity, then the person that wrote the contract is deemed at fault.

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

-Dik

 

[Deker]

Thanks for the compliment, CANPRO. Glad it was helpful. And thanks for modeling the connection and confirming our intuition. If only my brain didn't turn to mush once I start reading the metric units.

 

[dik]

Yes gentlemen (Dik's binary term for political correctness ), it's been a neat learning experience.

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

-Dik

 

[N G P]

Referring to CANPRO's idea statica report, I happened to have a doubt about the bending moment you have additionally applied.
As I understand idea statica considers the eccentricity ( (B+t)/2 ) by itself. Even if we do not apply the additional moment, idea statica evaluates the bending moment and verifies the gusset plate, bolts and welds with the moment similar to option 3 mentioned by Deker above.

I just want to ask you whether you applied the additional moment to remove the bending moment calculated by idea statica to make it look like option 2 in deker' FBD diagram.

I am also checking a similar case in idea statica and I just to avoid any moment in the gusset plate resulting from the eccentricity to the member centroid.

 

[dik]

I went with a modified Option 2; I think that most closely represents the ocnnection.

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

-Dik