Reinforced Tower Leg

[CBlaker114]

So I've got an interesting problem. I've got a telecommunications self support tower where the legs have been reinforced by u-bolting a channel onto the tower legs.(I've attached 2 photos for reference) So from what I've seen, there's nothing that will allow the reinforcing channel to develop any of the compressive load that the tower leg is taking. I'm assuming the u-bolts are snug tight and as there would be no shear transfer there, the leg still takes all of the load. However, the with channel the overall shape is more rigid. So I was wondering if the following approach would be applicable of if I'm just a little outside the box on this one..

Pn= FcrAg

Fcr- I would use the composite member to calculate this... reductions as required for fastener spacing and all
Ag - Use the original 1 3/4" SR leg

So I'd be using the rigidity of the composite section to get my flexural buckling stress but I would only use the area of the leg as if unreinforced.

I'm thinking... either A) I'm oversimplifying.... B)This approach is already used and I've just never seen it... C) I'm a genius (this one is the least likely)

Thanks All

 

[KootK]

Genius for sure. I'd make one change. Base Fcr on the non-composite combined section. You know, because they're not shear connected for composit behavior.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[oldestguy]

Shortening the unsuppoted length seems to be all that this does. L/r is less.

 

[msquared48]

It .is just kl/r related. Nothing more.

Mike McCann, PE, SE (WA)

 

[Teguci]

Reminds me of page 23 of the this PDF

http://www.unistrut.biz/Literature/SeismicBracing/Seismic_R0120.pdf

Need to check:
Capacity of the original member with braced length = U-bolt connection spacing
Capacity of the combined members for the full unbraced length. I would use Ag for only the original member as you noted.

Definitely oversimplifying because:
The bracing member is eccentric to the buckling member - not sure how to consider this. The buckling moment from the original member will place a small torsional load on the members (buckling reaction is eccentric to action).
Are the U-bolts tight? Will they remain tight?

 

[CBlaker114]

oldestguy/msquared,

That's correct. With a reduced kl/r... my original legs will pass... essentially the rigidity of the channel will brace the SR leg. I just didn't know if this approach was appropriate.

 

[SAIL3]

as Koot pointed out, if there is no capacity for shear transfer between the reinforcement and the original, then Fcr should be based on the original member and it's unbraced length.....IMO the channels are not adding any extra strength, stiffeness, etc to the overall structure....

 

[CBlaker114]

Thanks All. I've designed an extension and welded termination to develop the required loads. I do think in reality there's some bracing capacity from channel to original leg which would increase the compressive capacity.... but I agree my previous approach could be an over-simplified approach as there's no shear transfer.

 

[msquared48]

CBlaker114:

Check in TIA222G, but I think there is a provision to relax the KL/r ratio under certain conditions.

Been a couple of years since I investigated one of these towers, and I cannot remember the application clearly, but I know it does exist.

Mike McCann, PE, SE (WA)

 

[TheDW]

TIA 222-G 4.5.3(a) lets you reduce the Kl/r with bolted connectors provided the connectors are close enough.

 

[msquared48]

If you do not use this reduction in the program (TNX Tower, or whatever), the program will tell you that the web members fail. The allowance of this trick is in some obscure place in the program...

Big help, huh...

Mike McCann, PE, SE (WA)

 

[TheDW]

msquared: From what I remember (its been ~6months since I've used tnxTower), you would have had to calculate the effective kl/r yourself outside of the program, and then it was simple enough to apply a factor to the kl/r of individual members in the program.

If you were considering it a built up member we would also do a number of checks outside of tnx on the member as well, since tnx would only do a built-up check and not check each component of the member or anything regarding the connections between them.

 

[KootK]

To clarify, I think that it's as simple as this:

1) Radius of gyration = SQRT((I_leg + Iy_channel) / (A_leg))

2) Everything else the same as usual using the properties of the original leg alone.

Since the centroids of the leg and channel are located fairly close to one another, it's probably not much of a loss to ignore any potential for composite behavior.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[rb1957]

I think all those channels do is limit the deflection of the leg ... if the leg tried to bow towards the channel, that'd produce some bending in the channel. maybe the leg was originally bowed and the channel added to straighten it ... in which case there's a whole pile of unknown stress in the background ...

another day in paradise, or is paradise one day closer ?