ANALYSIS METHOD OF A ROTATED AT SPLICE BEAM

[flyingdebris]

Hi Guys,

Imagine a simply supported rectangular hollow steel beam which at mid span has been cut in half. One half of the cut beam is sloped down at 15 degrees and would then be full strength butt welded to rejoin the two beams ie the two supports are at different heights (though its not welded yet, read on). Now the beam that is sloping down is twisted around its longitudinal axis by 18 degrees and then rejoined with FSBW. I hope I am clear at this stage. How would one analyse this beam in bending? I see a whole issue of biaxial bending, torsion issues that are not conventional. Has any one come across this? I realise each end of beam would require end plates FSBW together to form a splice as such, but struggling to find an analysis method.

Thanks.

 

[CELinOttawa]

Your biggest problem will be in manufacturing... Box type HSS do not like to be twisted about their long axis. The walls are going to bow and your beam is going to look pretty wrong, to say nothing of analysis.

Now, in terms of the analysis, you aren't so very bad off. If you have a section that is torsionally stable, you do not need to consider secondary moments within the section unless the geometry also causes the input loads to be arranged in such a manner as to cause incidental torsion.

Think about it this way: If you have a box on an angle, will this necessitate a secondary moment in the design? No, but if you are applying load to a corner which is no coincident along the line of gravity with the centre of stiffness, YES, now you have a secondary moment. The easiest design method is going to be to break this into 250mm long segments and solve for all the secondary effects through the section, applying this as a torsion within the shape. As for your primary bending case, the classic tension and compression bending formulae for a beam will still apply; It does not matter that the section is changing as you go along the axis, only that there is material far enough away from the neutral axis to handle the stress induced by the loading. Just solve for the weakest (Iyy and Syy in this case), although this would be conservative and you could break the section down into segments, solving for each rotation through that segment, solving for the worst section (which WILL be the lowest value of the section applied in your problem) compiling all the rotations, and solving for the vertical shift to sum the set for the solution of the deflection part of the problem.

The bigger issue is that you're likely going to have to make the twisted HSS out of plate, unless your client is willing to accept a very odd looking final result from the twisting.

 

[hokie66]

CEL,
I think the OP is rotating the section, not twisting it.

Providing a nice thick plate at the splice, along with backing bars for the full strength weld, should be adequate for the connection. I would just figure the section properties of the weaker part, probably the rotated section, and use that to analyse the beam.

 

[CELinOttawa]

Okay, if Hokie is right, why was this question asked? This is very basic structural analysis and design...

 

[JStephen]

How it would be done would depend a lot on the details.
Depending on the end conditions, vertical loading or loading normal to each beam or loading in the vertical plane of the combined beams will also produce axial compression, so it's not pure bending.
I'm thinking bending in one beam will produce torsion in the other, would have to give that some more thought.
Depending on the lateral restraint, vertical loading will produce horizontal deflections.
Depending on the lateral restraint and end conditions, the whole thing could be unstable and simply flop over.

How to analyze it: Frame software would be quickest. Watch the angles, releases, end conditions.
It may be a situation where allowable stresses are not obvious from the codes.

 

[dhengr]

Flyingdebris:
I wouldn’t seem that you have expressed/explained your problem/structure very clearly, since you have three of the smartest people here on E-Tips debating about what you really mean. Sketches, with some sizes, dimensions, loads, reasonably proportioned, are truly worth a thousand words; plan, side view and end view. I would also be interested in why you need such a crazy beam, is there a simpler way to skin this cat? I think you have a cranked beam when viewed from the side. It starts at a lower bearing point on the left, slopes up at 15̊ to midspan and then turns (cranks) to a horiz. member at a higher elev. to the other bearing point on the right. Now, looking at an end view, we see the horiz. portion straight on in cross section, moving away from us, and do you want to move the lower bearing point left or right so the sloped member also slopes 18̊ from vert. in this view? Or, are you really rotating this sloped portion by 18̊ about its longitudinal axis, and why? This third manipulation makes the analysis and the joinery very difficult. The first two manipulations just involve a compound miter cut on the ends of both halves, then beveling, then a CJP weld all around. But, this is still a very difficult joint and weld to do properly at a max. stress location.

 

[flyingdebris]

Sorry I thought the idea of a forum was indiscriminate no matter the question? But you contradict yourself by saying "IF he is right"? So in fact by stating that, it sounds like you don't know either. If it's as easy as that, that's great, I'm happy. It didn't need a dickhead response from someone who obviously needs to re-evaluate life and get out more. Go take a walk to clear the mind.

 

[flyingdebris]

Thank you to everyone else for giving the question some time for a response. Much appreciated.

 

[CELinOttawa]

Well, which is it? There are three differing interpretations, and we still don't know which is correct.

You are right, however, that I was being judgemental; I simply thought the question more complicated than Hokie believes. How 'bout you post a sketch so we know which is correct, and can help you. If I was wrong and the question is more simple, I do apologise. We all start somewhere and I should not have been so quick to quip a kurt response. Where needed, you have my apology.

 

[flyingdebris]

Apology accepted. Please accept mine. Have talked the client out of this anyway. For it's intended purpose it was a silly design. Another win over the architect!!

 

[CELinOttawa]

Well that's good... In for the win over another shmoo. I do hope to see you around the board!

 

[JStephen]

The sketch attached is how I interpreted it. It could be really awkward to work by hand. Good to know you dodged it.

 

[hokie66]

Me too, JStephen, except I thought it was just a simple beam, and braced adequately so that torsion etc. was not an issue.