Truss modeling and connection question..another one

[faromic80]

I have some more questions about trusses and decided to start another thread because the one I started the other day is a little off topic. I have attached a scan of some questions I have on a truss I am analyzing. I got these shop drawings and have to verify everything works.
From what I understand, truss members have only axial loads, but this is not the case here. In the attachment I show the model i used for hand calcs: simply supported truss. Then I show the model with the existing columns coming up the sides. The top and bottom chords are attached to the existing columns at the sides. I also attached the RISA output for the top and bottom chords. The bottom chord has moment and NO axial loads?? The (2) C10x15.3 channels can't handle that moment. I'm thinking this is because of the connection to the existing columns? the top chord has both moment and large axial forces in it and I can't understand why? because of the distributed load on the top chord?
Also, the two center diagonal M15 and 16 have no axial force in them?
I don't exactly understand the splice connection at the top chord. I don't understand why the the two plates are welded on the the larger splice plate...is this for out of plane buckling reinforcement?

You can also see in the RISA model at the intersection of M8 and M9 the members don't coincide with the center of the top chord (see truss shop dwg) how can I model this in RISA?

As for the model I think it's pretty accurate because my hand calcs for the joint forces is within +-10 kips (attached)

 

[faromic80]

as of right now, i assumed vertical load only and no moment transfer in the column. The connection is a shear tab with 3 bolts. the gusset will be added when the truss is fit up in the field. i don't think this will induce moment at the connection so it should act as a pinned connection. If I fix the end connections, I will get a moment which I don't want and is not how I modeled it. The top chord is connected to an existing metal deck roof which acts as a diaphram. However, the lateral load is dumped into masonry shear walls, so the truss frames aren't assumed to resist any lateral loads. I'm going to reinforce the existing columns with 2 MC7 channels for the increased vertical loads. I'll run them up to the top to of the existing column and cut them off as necessary to clear the existing W14 beam (top chord).

 

[rb1957]

but aren't the upper and lower chords connected to the column ? ... that'd provide moment stiffness to the end of the beam. if one chord isn't connected to the column, this'd reduce the moment stiffness, but i think this'd drive the chord load into the nearest load path, presumably not really designed for this load.

 

[faromic80]

The bottom chord is connected to the column with L3x3 at top and bottom of the channel. it's a pretty flexible connection just for the sake of continuing the bottom chord. it should act as a pinned connection? what do you think?

 

[rb1957]

the top chord isn't connected to the column ?

lower chord only would be a pinned joint

 

[lkjh345]

My 2 cents:

Am I reading something wrong: Your taking the reaction into an existing 25' tall TS 4x4 column? Is the column braced in both axis somewhere along the 25' height?

FWIW: I disagree with connecting the bottom chord at all. This connection will attempt to 'fix' the truss end until it fails. With a very small column size (TS4x4) any moment introduced into the column could be too much. JMHO, but I would not attach the bottom chord at all.

 

[faromic80]

I am reinforcing the TS4x4 with 2 MC7x19.1 at it's sides. It is unbraced along it's height. the top chord is currently connected to the existing TS4x4 by a 3 bolt shear tab connection. So it's pinned I am reinforcing the TS col below the existing top chord for the increased vertical load. so I'm assuming a pin connection because of the concentric connection and no moment at the connection....

 

[OzEng80]

I really think it is worth while modeling the frame in its entirety to get insight into the moment the column will attract. Two 'pinned' connections still achieve fixity to the column due to the distance/lever arm between the connections - the connection does not permit rotation.

It is my interpretation that a 'roller' assumption should not be assumed in the frame design unless the horizontal deflection determined in the analysis can actually be achieved. Unless slotted holes are detailed, the truss's rotation will be prevented by the column fixings inducing moment into the column - all a function of the relative stiffnesses of the elements.

Note that even in a true 'pinned' connection, the load is still eccentric to the column CL and some moment will result - P delta effects need to be considered with any horizontal deflections.

Cheers

 

[csd72]

At a minimum you should apply the load at the face of the column.

 

[rb1957]

if both chords are pinned, then the overall truss beam is effectively fixed at the ends. if the top chord is already attached to the column, then leaving the bttm chord unattached (as ljkh posted) would make the truss pinned.

as the other posters are noting, there are many "issues" to be addressed with the columns.

 

[faromic80]

I agree. It may be beneficial for me to model the columns to see the moment transfer I get at the column. I knew that some moment transfer had to occur just because of the shear tab connection at the ends. By the way, this connection is transferring the vertical reaction at the ends, right? I decided to not take into account the existing connection and see if the 7.5" vertical length of the gusset at the existing end columns could resist the 50 kip reaction. Just for curiosity, I found that the 3 bolted connection can handle much more than 50 kips from the AISC manual tables. Do you think it's sufficient to check the existing column for the moment due to the reaction at the face of the column?

 

[csd72]

Look up the design rules for shear tabs, taking the reaction at the face is only applicable for a beam bearing on a cap plate.

 

[faromic80]

My biggest concern is the end connections at the top chord. I am not going to connect the bottom of the truss to the column. I will extended it past the last diagonal but stop it 2" short of the column. Then I'll weld some plates to the sides of the existing column for torsional considerations since it's going to be field built.
I just keep asking myself if the top chord connection to the column is fixed or pinned after I add the gusset plate at the intersection of the existing column and existing top chord. When you introduce that brace and gusset, it should still act as a pinned connection since the member centroids are concentric, right? I guess my biggest concern is my unfamiliarity of the detailing of the connections to ensure they behave like "pinned" connections and not fixed connections, especially at the ends.

 

[IFRs]

You are to be congratulated for being concerned about the connections. This is too often overlooked! There will be some degree of local fixity. Analyzing the truss with pinned ends will be conservative for the truss members. Some moment will be transferred into the column. You could try using the vertical end reaction of the truss times the width of the gusset to get a moment into the column. There will also be some moment due to the tendency of the top chord and/or truss to deflect under load and transfer that deflection into the column via the "fixed" gusset. So, you have an induced moment from deflection and a moment from the gusset connection. Add them together and see what that does to your column. Good work!!

 

[faromic80]

How would I determine the approximate load transferred into the column due to the deflection of the top chord imparting load to the column through the "fixed" gusset plate. I could figure out the horizontal movement due to vertical deflection, but how to translate that to a force?

 

[faromic80]

One more question:
I was asked to take a look at the end connection as well as the splice connection at the top chord to make sure the load could be transferred. I've attached a picture of my detail of the end condition, showing the MC7x22.7 reinforcing members. Is the vertical reaction of 49 kips transferred to the column through the gusset plate. I've designed the weld to transfer this load. Should I add a stiffener? It's a 3/4" plate! The reason for the 3/4" plate is that it's being built entirely in the field and I wanted to account for construction tolerances and any eccentricities being introduced.

 

[OzEng80]

faromic

Your analysis software should provide the horizontal deflection and associated column moments. If you want to determine the force from a pre determined horizontal movement simply work backwards by applying a deflection instead of a force to the appropriate deflection formula. ie the stiffer the member the larger force.

Can't say i'm a fan of not connecting the bottom chord. I think fixing it and slotting the holes horizontally will provide a more stable arrangment. I assume there can be no lateral loading on the bottom chord and that it doesn't go into compression (no upward loading possible)?

 

[faromic80]

The load is not reversible. I was not a fan of it either, but did not want a moment induced. I modeled the truss with fixed connections at the top and bottom chords at the ends and still didn't get a moment in the column. There were only vertical loads acting on the column, but still, a moment should be present. What I did was calculate the moment by multiplying the vertical reaction by 4" eccentricity.

 

[OzEng80]

The moment that is transferred into the column is a function of the stiffness of the connecting elements. So if you have a stiff beam (little rotation) and an unstiff column (pin based?) you will get little to no moment transfer through the connection and into the column. ie the column offers little resistance to the rotation of the beam.

Regarding the load eccentricity you have to run a second order analysis to account for p-delta effects. What you have done with the nominal offset seem adequate though.

Cheers

 

[kslee1000]

Your model is correct. But have you removed the rotational constraint at the ends of the truss members?

 

[faromic80]

I have. I'm getting an instability error in RISA, but the results are reasonable so I can ignore the error. This is according to the help file. The results are close to what I get by hand.