Truss Deflections 1

[medeek]

I'm sure you are all familiar with the standard output from the truss manufacturers that shows the loads, geometry etc... According to the code they are required to provided deflection (vertical) values for the DL and TL. I've got a copy of ANSI/TPI 1-2007 here and I can't seem to find any equations, algorithms or specification of method used to arrive at those deflection numbers. I've done quite a bit of reading on trusses recently and I've seen matrix method, energy method, castiglianlo's method thrown around. Being that the truss manufacturer's use software that is proprietary and they probably protect their methods quite carefully its not like I can email them and ask what is their standard method of calculating deflections. Just wondering if anyone has any experience or insight about real world truss deflections and could send me searching in the right direction.

 

[paddingtongreen]

I'm not clear on what you want, they publish the deflections in their tables. The only difference in the methods that I can think of is whether the chords are modeled as continuous or not. Manual calculations usually ignore the continuity. The calculations tend to overestimate the deflections, probably because there is some frictional fixity at the internal member connections providing some additional load paths.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[focuseng]

Medeek, I think the most responsible approach is FEA. Each member has an associated stiffness/length/loading condition. The nodes would have to be modeled too as paddingtongreen suggested, there is some inherent fixity at ea plate and is dependent on the plate size/thickness and location. The nodal fixity is the real "magic" black box part of the truss manufacturer calcs which I can never seem to replicate 100% when do truss repair.

______________
MAP

 

[XR250]

Just get yourself a 2-D frame program and model it yourself using pinned connections.
That will get you close. Most of the deflection comes from axial shortening or elongation of the members so modeling as pinned versus fixed does not make much difference. There is nothing proprietary about the way they calc deflections - it is just simple matrix analysis - just like you can do with a 2-d frame program - unless they are taking into account plate creep - which would be proprietary.

 

[rb1957]

solving a determinate truss is just an excel s/sheet.

solving an indeterminate truss involves more work ... you can do it on a s/sheet or you can use FEA.

as michael says, typically hand calcs neglect continous chords, pinning the elements at each joint; which is probably conservative for the truss element loads, but possibly unconservative for the local stresses of the continuous chords.

Quando Omni Flunkus Moritati

 

[medeek]

I'm just solving for a determinate truss so this shouldn't be too hard. After further searching I noticed I had managed to overlook a couple of paragraphs in the commentary to the TPI 1-2007 which does briefly talk about pin jointing the truss and using the energy method, deflection being a result of axial tension and compression and the sliding of the reaction points. I will give this method a shot and see if I can get results that compare with their output. Thank-you for everyone's timely input.

 

[rb1957]

i'm sorry, i was thinking more about getting the loads in the truss elements.

getting the defelction of the truss will be somewhat of a pain ... element load > element stress > element strain > element elongation (ok that was the easy part) now put all the element elongations together ... what assemble a bunch of arcs ?? that's got to be only a short step away from insanity !

if you need deflections, i think you need FEA.

Quando Omni Flunkus Moritati

 

[a7x1984]

medeek: If you don't have an FEA program, calculations by the "virtual work method" work very well, and can be completed in an excel spreadsheet, if desired.

I can send you the procedure for trusses in pdf if you like.

In Russia building design you!

 

[medeek]

I have some examples I have been looking at but do send me your procedure (pdf) A7x1984, maybe it is a little less confusing than what I am looking at right now.

 

[paddingtongreen]

Castigliano should be relatively easy on a spreadsheet.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[renderu]

medeek, are you required to double check the values provided by the truss manufacturer? I buy Crosby shackles but I don't feel compelled to check that they are correct in their provided 4 or 5:1 safety factor. Just wondering if this is more of personal interest than a requirement of your work.

Thanks.

 

[medeek]

It's a little bit of both actually. I'm working on a custom app. that helps me at work with some of my preliminary design work. I don't need results that are 100% accurate but the closer I can get to a standard output from a truss manufacturers software then the better.

 

[a7x1984]

medeek: Here are the procedures. I recommend using RC Hibbeler's structural analysis textbook - he seems to be clear about procedures.

In Russia building design you!

 

[rb1957]

"apply the principle of virtual work to determine the deflection ..."

i assume that means multiply the n and N and L/EA values for each element of the truss and sum ?

Quando Omni Flunkus Moritati

 

[a7x1984]

rb: Yup.

In Russia building design you!