welding retangular tube

[ball61]

My question pertains to welding of a rectangular (or square)
tube that is in double bending (M1 and M2). At the corners,
the moment is additive, creating substantial peaking at a
weld (of any type). The reason for this is that the actual
throat or weld stress area is theoretically a singularity at
this corner. 1) can I just call out a "pullback" (no weld) at the corners?--has anyone seen this done? 2) does anyone know of quick analytical techniques to quantify the throat
area??--I know that a rect/sqr. tube really has rounded corners, so there is a finite gradient around the corner

Thanks for any help you can provide

 

[desertfox]

Hi ball61

Can you upload a sketch then it will be clearer to see what your talking about.

desertfox

 

[ball61]

Sure--I'll try and put something together this aft.

Thanks

 

[ball61]

DF-

here's a sketch of what I'm talking about--apprently it's
common practive to "pullback" a weld from a corner

Rob

 

[desertfox]

Hi ball61

I have never heard of pulling back a weld,I might be completely wrong but I would of thought that leaving an area without weld would be like having a crack in a welded joint.
I think I see your situation ie:- you have bending in two directions and therefore your stress distribution for both loads are at 90 degrees to each other, however at the corners you have two tensile and compressive stress distributions due to bending not to mention shear distrbution.
I am truefully not sure how you analyise those corners other than FE, if it were me doing it I would calculate the stresses seperately for both bending planes,sum the stresses and use a large safety factor.
this site might help:-

http://www.roymech.co.uk/Useful_Tables/Form/Weld_strength.html#Weld_lines

desertfox

 

[Ron]

ball61...if you carried that logic through, you would eliminate the weld and therefore have no stress in the weld....and no connection.

DF has it right...a pull-back is equivalent to a crack and therefore a loss in capacity. You still have to resist the same moment...why not get as much weld in there as possible...not less weld.

 

[tr1ntx]

I assume you are talking about standard mill-run tube, so I will say this: the corners are substantially rounded, with the outside radius greater than the wall thickness (unless you're talking about an extruded tube of some sort).

Perhaps your sketch, with its sharp corners, has caused you to think along those lines and not allow for the substantial radius. I do that kind of thing all the time, i.e., jot something down as a "general idea" sketch and then catch myself letting some ill-defined feature steer me down an incorrect path. What size tube are you talking about?

Yes, I have seen the "pullback" that you are talking about, but not with rectangular tubing.

 

[ball61]

Hi All--

Yes, i think that my original assumption of compound moments (bi-planar) summing to a "fictitious" corner was
overly conservative--taking a rather generous radius in account gives me positive margins--also, there is a redundant load path in this weldment that I'm not counting on (more conservatism)-
I really appreciate everybodys timely and helpful inputs

Rob