Truss / Joist Girder Stability Bracing 4

[TTUengr51]

I realize there are other past threads that have attempted to address this topic, but I wanted to take the time to get some feedback on an item I am currently contemplating. I have attached a sketch to follow along with.

I currently have designed a couple of 120' clear span, 10' deep steel trusses for a heavily loaded roof. Typically I would assume the supported bar joists/metal roof deck would provide ample stiffness to laterally brace the truss. However, since this truss is 40' longer than any I have designed in the past, and the diaphragm depth is 1/3 of the truss span, I thought it would be worth while to look into this issue further to make myself more comfortable.

I originally assumed the bar joists at 5' o.c. braced the top chord of the truss, thus making my Ly = 5'. While summing the brace forces (0.004*chord force at each joist - treated as column per App. 6) required along the truss at the bar joist locations (see 2nd page of attachment), I came up with a total lateral brace force on the diaphragm of 76.6k, which puts 38.3k in my end walls. With only a 40' diaphragm depth, this requires the decking to resist 958 plf shear for D+L stability bracing. Obviously, this seems ridiculous.

So my first red flag was is the 0.004*Pr required to be calculated at each brace point? Or is this force taken at the maximum chord force and then distributed amongst the brace points? The second item involves the LFRS. Is the diaphragm required to be designed to resist these cumulative loads and therefore distribute to the LFRS? Or is this the purpose of the required brace stiffness criteria listed in Appendix 6 and no further assessment of the load transfer is required?

In moving forward, I decided to be conservative and utilize brace trusses at the quarter and mid-span locations of the truss (Ly now is 30' - see 1st page of attachment) in order to significantly reduce the lateral stability forces transmitted to the diaphragm. While this makes the D+L forces in the diaphragm fairly negligible (150 plf), it obviously makes the compression chord of the truss much larger in size. While this is probably viewed as very conservative, I felt comfortable and moved on.

To complicate matters, during the bidding phase, I had steel fabricators request that a joist girder be explored as an alternate to the custom truss I had designed to try to cut cost. I had considered this early on, but thought the span/depth and loads would not make this a viable option. After discussions with Vulcraft, I had asked about their assumptions regarding compression chord bracing. As I suspected, they assume each bar joist braces the truss. When asked how they account for the brace forces transmitted to the bar joists and beyond, they were not able to really give me an answer other than it's just assumed to work, which did not make me feel very comfortable.

In conclusion, I'm hoping to get feedback for:

1) How do you interpret the requirements listed in AISC Spec Appendix 6 in regards to placement/distribution of the brace loads, and cumulative effects on the diaphragm?

2) Dealing with joist girders/open web joist bracing?

Nick Deal, PE, SE
Michael Brady Inc.

http://www.michaelbradyinc.com

 

[WillisV]

See the attached paper which addresses this issue.

Basically, brace forces are not cumulative (they do not need to be resisted by the LFRS) because the truss buckles in a sine curve between brace points so that the brace forces are opposite of each other at each point. The induced diaphragm shear is also not cumulative, it is a panel by panel phenomenon as addressed in the paper. You also need to check the stiffness provided by the diaphragm - again as addressed in the paper.

 

[a2mfk]

Thanks WillisV, this article was very timely for me..

I am looking at a couple of roof beams that I will have to design as drag struts and I was just thinking about how to design it for compression and what kind of bracing I could assume from the joists/deck. This is perfect. Of course, I also have to consider bending stresses, but for a built up truss you have to consider that too in the top chord- bending between the webs plus axial, correct? (I have never designed a large truss like this before..)

 

[WillisV]

a2mfk - you can design the top chords of trusses either for axial load only (plus bending induced by intermediate point loads between panel points as if it was simply supported) such as in traditional truss analysis, or as if the top chord is a continuous beams where you get flexure throughout the length of the chord. Either option is valid. Designing as a true truss simplifies the design and can produce smaller members but results in higher deflections. Designing as a continuous beam might require larger members due to the moments but will be stiffer in the calcs and you can produce a shallower truss design.

 

[BAretired]

I don't know what your code says about bracing trusses of this span, but in my opinion, bracing the top chord with open web steel joists is not adequate. The bottom chord needs to be braced too and it makes sense to provide braced frames between trusses to prevent truss rotation and to laterally brace the bottom chord at L/300 (or whatever your code requires for a tension member).

It also assists in erecting the trusses without disastrous results. Trusses are very unstable until all the bracing has been installed. Collapse during erection is not uncommon with trusses of this span.

While erection may be the responsibility of the erector, I would encourage him to connect the trusses to the horizontal braced frames on the ground, then lift the assembly into place and connect to the columns before attaching open web steel joists.

BA

 

[hokie66]

I support BA's approach, and would not like to depend on a metal deck diaphragm to brace these trusses.

 

[TTUengr51]

WillisV,
I understand your point concerning the sine wave mode of failure, which would indeed cancel out any cumulative forces at the diaphragm edge. However, for this to occur, wouldn't the brace points have to be assumed to be rigid out of plane? It seems to me you would still have to design the diaphragm/horizontal truss to account for these internal shear forces.

I keep going back to pre-engineered wood/metal truss bracing scenarios. For bottom chord uplift bracing, it is common to run rows of lateral bracing when no ceiling diaphragm is present. The brace forces that accumulate in the lateral braces are distributed to the supporting walls with a bay of horizontal diagonal bracing every 4th to 5th truss. The forces developed in the horizontal diagonal bracing are typically cumulative.

I think you would also want to consider a failure mode in which all the trusses would buckle in the same direction at mid-span. It seems this would put a lot of force into the diaphragm that would not cancel out.

BARetired,
I plan to keep the brace trusses at the quarter and mid-span to provide torsional restraint and to brace the bottom chord to meet tension slenderness requirements and against compression buckling under uplift loads.

In regards to the metal deck diaphragm, I don't have a problem using it in conjunction with the brace trusses/bar joists. I just want to be clear on how much load to induce into it. Using an 18 gage deck with decreased spacings on the puddle welds and sidelap fasteners, you can get shear values that exceed 1000 plf.

Nick Deal, PE, SE
Michael Brady Inc.

http://www.michaelbradyinc.com

 

[hokie66]

My comment about depending on the deck to brace the trusses was not for the built case, but for erection. If the contractor does not adequately brace the trusses laterally before he loads the decking onto the joists, it could all be unstable. I know...this is means and methods, but it is also consideration of constructability.

 

[DaveAtkins]

I agree with WillisV. Many, many buildings have been built with steel deck diaphragm, joists, and joist girders, and no additional in plane roof bracing. The lateral force resisting system is designed for wind or seismic, whichever controls, but additional buckling forces due to vertical loads are not added to the wind or seismic forces.

Otherwise, where do you stop? Joist buckling forces? Joist girder buckling forces? Top of column buckling forces?

Don't overthink this one.

DaveAtkins

 

[RFreund]

There was a helpful post that included notes from Yura a few days ago.

Here is the thread:
thread507-286765

I will attach the notes that were posted. Hopefully they will help.

EIT

 

[TTUengr51]

I tend to agree that I am probably over thinking the situation and I would have never really even considered this with smaller scale members such as joists or with a roof system that contained a more robust diaphragm. However, due to the magnitude of the brace forces and the depth to length ratio of the diaphragm, I had to make some more sense of this issue.

One way I thought about it to make myself more comfortable about not applying the brace forces to the diaphragm was to think about the chord bracing in-plane with the truss. In this axis, you utilize the vertical/diagonal web members as braces, but you do not add additional loads into these elements, thus increasing the load on the truss.

On the flip side, I had posed this question to AISC a few days ago and finally got a response today. I have included a snippet of the response below.

"...something ultimately needs to resist the brace force and this would most likely be the LFRS."

The rest of the response was pretty vague and really didn't lead me either way on whether or not to accumulate these loads into the diaphragm.

Thanks everyone for your responses so far on an issue that leaves a lot to judgment and interpretation.

Nick Deal, PE, SE
Michael Brady Inc.

http://www.michaelbradyinc.com

 

[TTUengr51]

RFreund,
Good link...Thanks!

Nick Deal, PE, SE
Michael Brady Inc.

http://www.michaelbradyinc.com

 

[RFreund]

This question got me thinking and it seems that the braces (bar joists) should be designed for Pbr1=0.004*chord force as you did and appropriate stiffness. Then the deck should be checked to make sure it can brace the bar joists from buckling or moving relative to each other (using Pbr2=Pbr1*0.004). I don't believe that the deck needs to resist the cumulative force of the braces because if the structure were to buckle in a single sine wave this would be similar to global lateral buckling failure of the system. However I would think that the deck should be design to resist the Pbr1 force between bar joists. This leaves me with two questions.

How would you calculate the global lateral buckling resistance?
I've seen this article referenced before:

http://scitation.aip.org/getabs/ser...0009001487000001&idtype=cvips&gifs=yes&ref=no

How did you decide to design the system?

EIT