Triangluar shaped tube structure

[jheidt2543]

I am starting a project that involves three tube structures and I'm looking for some helpful hints to get started. The cross sections are approximately an equilateral triangle with a base width of (wb1) and tapers to a height of (h) at the apex. One side of the tube is a plate of thickness (tp) to be determined (assumed at the start as ¼” thick stainless steel) and it is vertical, 90 degrees to the base. The two other sides will be sheet metal, thickness (tsm) to be determined (assumed at the start as 14 ga. stainless steel).

There will be a stiffener plate that, in plan view, is from the apex of the triangular tube to the midpoint of the back plate. The stiffener plate will have dimensions (wb2) x (h). The thickness of the stiffener plate will, for simplicity, be the same as the thickness of the back plate (tp). There will also be triangular stiffeners in the cross section at approximately six (6’) foot intervals along the vertical height of the tube. These stiffeners will be either angles or plates and be used for attaching the eight sections together as well as the sides.

The dimensions for the three tubes will be approximately wb x h: 3’ x 48’; 2.5’ x 36’ and 2’ x 28’ and the three tubes will be set on a common base plate, anchored to a concrete foundation. The tubes will be oriented in such a way that none of the sides will be in the same local X-Z plane. The assembly will be outdoors so, it will be subject to wind loading and gravity, no seismic.

I am looking for some pointers on how to approach analyzing these structures and here are my questions:

1. This structure is modeled as a shell, but it takes four points to define the shell elements, these are triangles. Would I be better off looking at the tube sections as truncated prisms?

2. Do I compute the center of gravity of each tube and the combined center of gravity by hand or is there a way I get that information from my FEM program (RAM Advanse)?

3. I am thinking of calling base plate bolt locations supports to get reactions for bolt design, it that the proper way? Or, should I just take the full base plate as a fixed surface and analyze the bolts separately?

4. I am thinking of building a model of one tube, analyzing that, then combining the three tubes and analyzing them as a unit with the base plate.

This isn't the "normal" square box building, so I'm just trying to think of a few different ways of looking at it. Any helpful comments would be appreciated.

Thanks in advance for your comments.

 

[corus]

Perhaps I'm reading this wrong but why use triangular elements because the cross-section is triangular? Surely you'd use 4 or 8 noded shell elements to model the side walls of the tubes that form the cross-section?

Normally FE programs will give you the C of G but you'd have to check with your software.

For bolted connections restrain the model at the bolt positions. The reaction forces will be given in the results.

You could always model the tubes as beam elements with appropriate area and Moment of Inertias for the section properties but it's easier and less complicated just to use shell elements.

corus