Unbraced Lengths

[slickdeals]

Just trying to get a collection of thoughts on what unbraced length you would use for the member circled in the pdf? The member is in compression and is W14x550 (web horizontal). The truss has a span of about 500'. The gusset plates 1.5" thick and are flange bolted.

Using a node to node is certainly conservative, but is using a unbraced length from edge of gusset to edge of gusset (15'-8") appropriate considering that there will be rotations at the column due to truss deflection?

 

[connectegr]

In your sketch it appears there are also truss to truss connections orthogonal to the chord in question. Does that change the un-braced length.


Doug

http://www.FerrellEngineering.com

 

[slickdeals]

The truss in the orthogonal direction does does not affect the member capacity really because the member is web horizontal.

 

[connectegr]

How do you know the specific connection lengths, prior to designing the truss members? Therefore the lengths would be node to node.

A few years ago we consulted on the design of a 600' span arena roof. In order to decrease the unbraced lengths, a "kicker" was added vertically connecting to the diagonals and chord at mid point of the diagonal. Effectively cutting the lengths in half. But we did use workpoint to workpoint for the lengths.

http://www.FerrellEngineering.com

 

[paddingtongreen]

This is first impression, bear that in mind when you read below:
All of the members at the joint are weak direction so it looks as though you have to use node to node and fake up your own alignment chart, substituting diagonals for perpendiculars.

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

 

[BAretired]

I would use node to node, i.e. 25'. I would not argue with k = 0.8 for a length of 20'. It may be a bit conservative, but certainly 15'-8" is not justifiable by any logic.

BA

 

[ash060]

paddingtongreen

How can you use alignment charts when you are assuming pin ended conditons?

Slick

When modeling the structure do you get any moment in the member, if you did I might consider going from inflection point to inflection point. I assume that this is only happening in the uplift condition.

In a AISC presentation I have they use the real length of the member for effective length. This was for braced frames not trusses though

 

[ToadJones]

how did you size the member sans the unbraced length?
are you refining the design at this point?

Might be a question of relative stiffness between the connections and members which would certainly be hairy.
Are the wide flange chord members sandwiched between the gusset plates?

How deep is the truss in the middle.
A 500' span truss is long enough to scare me into using K > 1.0 !!!!!!

 

[slickdeals]

All:
I cannot divulge information on the project. The design was done by a fabricator's engineer and we are peer reviewing the design (obviously). The design is lean and mean, with absolutely no extras.

The WF is sandwiched between gussets. The truss is 26' deep in the middle. They are relying on continuity of the truss over the first support. As a result the bottom chord close to the supports are in compression.

The truss is braced out-of-plane by sway trusses framing at every panel point.

I am only trying to prove that the design may not be optimal. There are many ways to skin a cat, but this one does not seem right. I don't think I would have designed it this way (with no extras), but that does not mean it is wrong.

Hence, I was trying to get some opinions from you "been there-done that" type folks to see if an unbraced length of 0.63 * center-center distance was too unconservative.

 

[ToadJones]

Spanepth of close to 20 seems to be pushing the envelope a little ...just my opinion.

 

[paddingtongreen]

ash060, If pinned connections are assumed, the node to node distance is correct. If it is continuous, then a reduced length may be acceptable.

Slickdeals, I'm a little leery of using the reduced length, I can't see right off, which of the standard cases this most represents. There will be translation and there will be rotation of both ends in space, but what will they be for one joint relative to the other? Will there be an arc or an "S" bend.

I haven't sorted this out in my mind but I offer it as a consideration.

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

 

[slickdeals]

To make matters worse, the connection between the W14 and the gusset has only been designed for axial forces. In order to use a reduced length, wouldn't the bolt group have to resist this rotation?

Assuming the member buckles at mid-span and assuming a 2% brace force would be required to prevent buckling, then wouldn’t the joint (bolts) have to be designed for say a moment = P L / 8, where L could be started off with a 15’-8” initial value.

The calculations for the gusset plate bots are purely designed for axial forces and no reserve capacity.

 

[BAretired]

I think the fabricator's engineer is taking an unwarranted risk with that design. I don't think it can be justified by analysis. Not only would the bolts have to take a moment, but the other members meeting at each joint would have to be capable of resisting rigid body rotation of the gusset plates.

BA

 

[slickdeals]

Thanks guys, all of you seem to unequivocally agree that the design is not really adequate.

What is also worrying is that all the members in this joint are in compression. It is bound to rotate in space as paddington says.