curved steel beam design 1

[keyPitsimplE]

I am looking for some guidance on designing a steel beam that is "S-Curved" in plan view. The beam is approx. 30' long, the max. offset from a straight line connecting each end is approx. 3'. This beam supports one edge of a residential floor that is open to below (joists hanging flush to curved beam, railing over curved beam). The floor trib. is approx. 5' at one end, increasing to 12' near the middle, then back to 5' at the other end. 40 psf LL and 15 psf DL. I propose a rectangular HSS for torsional stability.

I am aware this is complex and am looking for other ways to create what this designer wants. I would like to know what it would take if he insists on this design. Online research proved fruitless.

 

[BAretired]

Could you provide a sketch of the beam in plan view and the applied load? It need not be anything more than a freehand sketch.

 

[msquared48]

Analyze it using RISA 3D.

Mike McCann
MMC Engineering

 

[BAretired]

If you do not have RISA 3D (and I do not), analyze it using established structural principals. If you do not know how to analyze it precisely using established structural principals, try to approximate the answer using the principals you do know.

If, in the end, you do not know how to approximate the answer you are seeking with a reasonable degree of confidence, you must advise your client that you are unable to continue the project using his criteria.

Either he must vary his criteria or you should abandon the project. To use a piece of software which produces results which you do not have a "feel" for is not professional. You must decline to continue with the project.

 

[HamishMcTavish]

Hi keyPitsimpLE,

The Steel Construction Institute (SCI) are UK based but they have a publication "Design of Curved Steel" - it may get you started, all you'd have to do is try to match up code requirements if you're based in the states.

Link is:

http://shop.steelbiz.org/Product.pa...ory=Specialist Design Guides&txtProductID=243

Spiel is:
"Covers the design of common types of curved steel members used in building structures, and demonstrates how to take account of the curvature so that member checks can be made in accordance with BS 5950-1:2000 or suitable modifications of that Standard. The curving process is described, together with the effect that this has on the material properties. Guidance is offered on when the changes in material properties are small enough to be neglected, and additional checks are described where necessary. Where the behaviour of curved members differs from that of straight members, this is described. Where necessary, modified resistance checks are prescribed for curved members, together with straightforward methods to take account of additional effects, such as transverse bending stresses. The design of steel members curved in elevation, such as arches, and portal frames with curved rafters, is described, with guidance on the choice of effective lengths, modelling for computer analysis and simple approaches to determine approximate maximum forces and moments for initial sizing of members. Members curved on plan experience torsion in addition to bending effects. The publication provides two methods to design such members, allowing for the torsional effects by either a simplified approach or by using a detailed analysis model of the member. The design of both open (I and H) and hollow sections is covered. Includes six worked examples demonstrating the design of curved steel members used in different applications, curved in elevation, and curved on plan."

Hope this helps, HM


No more things should be presumed to exist than are absolutely necessary - William of Occam

 

[DWHA]

msquared48

I have RISA 3D and I was not aware that you can analyze a curved member in the program. How is this done??

 

[ChipB]

DWHA:
thread1419-229571

 

[DWHA]

Thanks chip. I have done that before with an arch. I was hoping that there was a way to do it as an actual curved member to see the torsional forces being induced.

 

[msquared48]

Feed your coordinates into the program in the XZ plane segmentally, the same as an arch, then load the member on the joints in the Y axis direction, placing the proper supports and restraints as the conditions direct.

No worries...

Mike McCann
MMC Engineering

 

[keyPitsimplE]

Attached is a sketch, if anyone is curious what this designer is trying to put me through. Due to cost, I have decided to push for cantilevering the joists over a straight beam, then trimming the joist ends into the curve he wants. Simple, as long as he can work around his height restrictions and aesthetic desires.

I agree, of course, with BAretired. I would never dream of continuing on a project that I was not comfortable with. That would clearly be unethical. I'm just trying to learn here.

Also, there are a couple of older references available to AISC members online, but reproduction is prohibited, so I will not post here. "Analysis of curved girder bridges" and "curved beam analysis", both .pdf's available through AISC. Good resource for comparison to a computer model.

 

[jheidt2543]

The applied load is variable along the length of the beam too isn't it? Do you design for the maximum and size the beam so that you can stuff all the reinforcing into it OR, do you make use of the varing load along the length to adjust the amount of reinforcing along the length? Perhaps in a 30' long beam it's not worth the extra work or field inspection.

 

[keyPitsimplE]

Yes, the load is variable, as seen in the sketch. It is a steel beam, so I would design the section for max. moment and torsion. I would hope to keep the beam from rotating by a rigid flush connection of the joists to the beam, and avoid major fixed connections at each end of the beam.

 

[BAretired]

keyPitsimplE,

The framers won't like the variable direction of the joists...too much cutting of deck.

I don't know the width of the rectangular area, but would it be acceptable architecturally to provide a beam from east to west (assuming north on top) at the north end of the stair opening? North of that beam you could frame all of the joists north-south or east-west, as you prefer.

South of that beam, you could place a straight beam angled so that it approaches the curve as closely as possible. Joists could frame east-west with a five or six inch deep joist extension on the east end to cantilever out to the required perimeter.

You really need to draw the plan to scale to see what would work best.

 

[keyPitsimplE]

The designer knows it will not be easy. The joists are exposed from under, so the skewed design is the visual intent. Using 2x decking anyway, not plywood, so no big deal.

See my post at 16:19. Cant. joists trimmed to the curve is what I am pushing for.

The freehand sketch was only because someone asked for one to get an idea of what was going on.

Thank you all for your comments. I'm done, unless anyone has any other wisdom to share on curved beam design.