Lateral Torsional Buckling of Truss Beams

[Robs1121]

Good Afternoon, I have a truss tower made up of rectangular/square sections and usually I analyze the tower using a Finite Element program. However, I have to analyze the same tower (cross section wise for the full length of the tower) for different heights (tower 1 is 20 ft high while tower two could be 50 ft high)for wind, ice and earthquake loads many many times as this tower is located in different states and jurisdictions.

My question is could you just calculate the section properties of the tower legs assuming the legs act as a group and obtain a graph similar to the moment/unbraced length graph. This way I can calculate the load demand on the tower and compare that to my graph? And are there any references that cover this type of analysis? Modeling the tower could be very cumbersome for each height and location. I can not seem to find any information that says you could use this type of analysis.

Thank you in advance for your responses.​

 

[Lomarandil]

Short answer: no, not really.

Truss structures don't experience LTB in the same manner that a single beam does.

You can use beam analogies to start your process of designing/selecting a tower section -- compute "cantilever beam" moments to estimate tower width and leg area, etc. But most of the nuance of truss design is in the individual members and connections.. ensuring that assumption of all of the legs acting as a group is valid.

Thankfully, you live in the era of really cheap and easy computer modeling. So running the analysis for different heights, on the surface, is not very difficult.

----
just call me Lo.

 

[KootK]

1) I'm a bit more hopeful that this could work.

2) I do not know of any applicable references specific to your situation.

3) I have a few questions:

- This is a 3D space truss, right?

- What are the overall proportions of the truss cross section?

- Does the truss taper towards the top?

 

[MotorCity]

I think your method would be a quick and dirty gut check method. Not sure what your tower supports but in my experience, communication/cell towers can be very sensitive to added load (i.e. adding a 300 lb antenna can require reinforcing).

Your proposed method analysis would not account for local effects (buckling, bracing, connections, etc.)

Instead of designing a unique tower for every height, can you group your tower designs, say one design for towers up to 20', another design for towers up to 30', etc?

 

[Craig_H]

I used to work for an employer where we routinely designed slender steel towers for hoisting loads (usually tailing pressure vessels in refineries and plants). Because of the repetitive use of the same equipment (constant cross-section), we did create an analogous beam model for first-pass bidding checks. The beam's properties were chosen slightly conservatively, and models were run on past jobs to validate the beam properties. I was wary of it when I first saw it being used, but found it to be surprisingly accurate. To be clear, this was used when a job was being bid, and we wanted to ensure that the in-house equipment could feasibly perform the work. A detailed FEM model was always used after the award to completely evaluate the structure.

Keep in mind that such a model will not be able to predict web lacing failures (picture large lateral loads at the base connection), and may become inaccurate outside of the range of lengths for which it was validated. I would question the analogous model's ability to accurately predict horizontal deflections under lateral loads, and would carefully evaluate the ability to predict second order effects. Very slender structures often have a tendency to fail in a buckling mode other than simple cantilever bending, and a graph's ability to predict this may be limited.

You may be able to create a model that enables you to give quick & dirty "it will probably work" answers, prior to performing a full analysis.

 

[Robs1121]

I greatly appreciate everyone's response, yes this tower is of uniform cross section for the full length and it is a 3D space truss. We do have different tower sections based on height, so we do have a wider section for heights up to 100'-0". So the take away is, I could use group section property analysis but it yield preliminary results and there really is no way to obtain a detailed analysis assuming group section properties.