Composite Truss 1

[slickdeals]

Guys,
Have anyone of you modeled a composite truss where the top chord is composite? Do you have any suggestions on how to model it in say ETABS or SAP?

The truss is 100' long and carries a roof top garden about 300 psf SDL.

Thanks

 

[civeng80]

Hi slick

I note that you have used the term composite truss before and wasn't sure exactly what that was. Is it like a composite beam i.e. a beam to concrete slab with composite action? except that its the top chord of the truss in composite action with the concrete?

Just want to be sure we are talking the same lingo !

Cheers!

 

[slickdeals]

Yes, the top chord of the truss is composite with the concrete. I am trying to understand behavior and modeling to reach a level of comfort where I can understand if an analysis program is throwing out crappy results.

Any references, reading materials in this regard is very much appreciated.

 

[kslee1000]

How the composite action on top chord is developed?
1. A continuous member with top flange encased.
2. A continuous member with metal sheeting and shear studs.
3. A typical truss with metal sheeting and shear studs (unusual case).

Also, how is the composite truss supported.
1. At ends of the top chord.
2. At ends of the bottom chord.
3. both

For a continuous top chord supported on its ends, I think the top chord (with the concrete)essentially is a beam for strength and flexural considerations. The remaining part (the truss members below) should be utilized for deflection calculation only.

 

[slickdeals]

kslee100:
A continuous top chord with studs + metal deck + concrete
Supported on the top chord at the ends, bottom chord braced at ends.

 

[civilperson]

Just a definition usage, chords of trusses usually are non-flexural members, just axial forces with loading applied at panel points. The description seems to be a frame with continuous loads along the top member of the frame, (which is composite). What is the tributary with of the frame, (spacing between two frames)?

 

[slickdeals]

The trusses are spaced 20' on center spanning 100'. The loads are of the order of 350 psf superimposed. The panel points on the 100' truss are at 10' on center. In this case, the top chord is not a flexural member as it has no significant loads between panel points.

 

[dik]

Yes... about 25 years ago... the Canadian NRC has some publications/papers on the design of composite trusses. They should be dated about 1970?. I'll rummage through my files tonight and see if I still have the papers.

Dik

 

[slickdeals]

I found a couple of papers based on full scale testing dated 1984 at

http://www.thestructuralengineer.info/library/composite_structures.html

Also I downloaded the ASCE task committee document from 1996.

But I am looking at real world advice and tips from people who have actually designed these trusses.

Thanks

 

[WillisV]

Real world advice: design it as non-composite and then put a shear stud ever foot as well - move on to next project.

 

[slickdeals]

WillisV, you sounded just like my boss. I get the point. However, this is the time I learn. You probably have enough grey hair from designing these things.

 

[civilperson]

How do you have no loads between panel points if the top member is composite with concrete deck?

 

[WillisV]

Sorry, you asked for real world advice and I couldn't help myself =).

Technically, a starting point might be to go about it in a similiar manner to a composite steel joist (see

http://www.steeljoist.org/graphics/... Engineer/SE1207_MaterialMattersReprint.pdf).

Calculate a moment capacity for the truss by summing moments of the forces in the tension member (bottom chord) and compression concrete block combined with some portion of the top chord (depending on the location of the PNA, shear transfer capacity of studs, etc).

Then design the truss web members for whatever shear you have just as you would a normal truss (the composite action is going to help flexure only i.e. top and bottom chords, it will not do anything to aid in the shear strength and therefore the web members can be designed using typical non-composite assumptions).

 

[slickdeals]

The reason I said I don't have significant loads between panel points is because I have beams that frame into these panel points. These beams are composite also and the deck spans between these beams. I am assuming that the load between panel points is not very significant.

WillisV, thanks for your advice. This is my weekend exercise to figure this thing out.

 

[hokie66]

WillisV has given you some good advice. The web members should be the same as if the truss were non-composite. The bottom chord can be smaller, just because your neutral axis will be higher. If you calculated the forces in the bottom chord neglecting composite action, then factored them down based on the neutral axis shift, you should get a good answer on the forces in the composite truss chord. I think I would then use the same member for the top chord, provided it is OK for the construction condition.

 

[kslee1000]

suggested thinking for your weekend practice:

1. The top chord is a relatively rigid beam (composit) supported on ends. Its deform-ability is much less than the free hanging bottom chord (no support at ends but braced for sway only).
2. The diagonals are pinned to the top and bottom chords, and follow the deformed shape of the top chord, however, the diagonals will not shorten or elongate.
3. Due to compability of displacement (the deformed shape of the bottom chord has to agree with the top chord), now the bottom chord(s) will experience forces carried through the diagonals. (Please note the bottom chord can be a continuous member, or formed by segments)

The analysis and design by hand will be quite tedious. The best and quick way to gain understanding on the behavior of this composite beam/truss is to model it on a PC with proper joint releases. Good luck, and have fun.

 

[slickdeals]

kslee1000,
Thanks for your suggestions. I tried calculating the effective moment of inertia by using transformed area for the top chord and area of bottom chord. I reduced this by 15% to account for web axial deformations.

However, when i run the same in an analysis program, I am getting different results. Is there a suggested way to model the composite top chord? Once thats done, I guess the problem of modeling is solved. Any tips?

Thanks

 

[kslee1000]

Sounds you are in the right track. Question is did you run the same model in two different programs to compare the results? If this was the case, make sure all inputs and the design parameters are identical on both runs. Check the outputs carefully, you may have bad member somewhere.

 

[slickdeals]

No, I ran the different models in the same program, in fact in the same file. I am getting a bigger MOI with hand calc than with the program.

Thanks for taking the time to help fellow engineers.

 

[kslee1000]

Two different models should generally yield two different set of results, don't you think so? Try to understand what had caused the differences, and adopt/modify the model per your conviction (that is best representing the structure to be built). Also, discuss your results with your colleagues at work will do great help.