How to calculate weld stress for a skewed pipe joint

[manwich]

Can some one help me find the calculation to calculate the weld stress for a pipe attached to a plate at an angle. I cant seem to find a sheer calc for an ellipse. Please see the attached diagram

 

[kingnero]

If I understand your drawing correctly, you've got a moment you need to react with your weld, not a shear force?

 

[manwich]

Basically yes. for some reason I am having a lot of trouble trying to find an equation to calculate the moment of inertia for an ellipse. I do need to take sheer into consideration as well.

Thanks

 

[kingnero]

first hit on google gives you the formula. my mechanics of materials handbooks have it as well.
I fail to see how shear is an issue here.

 

[dhengr]

Manwich:
You can FEA the hell out of this thing, and you might well want to, to check pipe stresses and buckling. That weld is a pretty odd shape and you may be able to find a moment of inertia for it too. But, then what about the section modulus, and then what do you do with that? Every problem does not have a simple formula or simple closed form solution which you can pick out of a handbook someplace. Why not be just a little conservative and think about how this connection really works and how it must be made and welded. Let’s simplify it so we can get our head around it. Here’s the way I see it.

I’d know what to do with a rectangular tube section and a 90̊ connection. To simplify it and be conservative, I might say; let the top and bot. sides (flanges) take the moment as a force couple, T & C and a lever arm; and neglect the contribution of the two side walls (webs) w.r.t. the moment calc. Now, we actually have t&b flanges which are elliptical in shape, a geometric complication; and we know that the max. tension stress (the critical stress) is going to be at the tip of the toe, from beam theory. Just by eye balling, for the time being, I’m going to assume that about 2-2.5" of the elliptical shape, measured along the central axis, is going to be my flange and the lever arm is about 7-7.5". What are T & C, and can I develop them with reasonable welds? And, what are reasonable welds, for that matter? This is a pretty funky weld to be spending a lot of time GD&T’ing and worrying about exact AWS symbols, etc. Through the little cross section of the pipe shape at the plate (your sketch, lower left) I’ve draw the central axis (horiz. center line), and I’ve drawn vert. lines 2 & 2.5" from the heel and toe, to imagine my flanges. At the toe, the elliptical arc of the flg. would have a vert. chamfer cut on it; zero chamfer at the 2.5" line and max. at the central axis, size to be determined. This weld is essentially a fillet which decreases in size and finally blends into the fillet welds along the two webs, which finally blend into a groove weld at the heel. The tough weld is the weld at the heel for welding access reasons and for access to the root right near the central axis. The saving grace is that the loading/stresses here will almost take care of themselves, in bearing, if we do a good job of fit-up.

Try thinking of your problem this way, and see if you can come to a solution. Some details to be worked out yet, and this does not lend itself well to one simple AWS weld symbol. But, we fool ourselves if we think every engineering or welding problem has a perfect or absolutely correct solution.

 

[manwich]

dhengr,
FEA was performed and submitted the along with a similar hand calc. to the one that you described, only I used a circle rather than a rectangle for the assumption. The customer wanted to see the hand-calc for the ellipse specifically. Performed a buckling analysis using solidworks simulation and found that the pipe was not going to work anyways. Switched the configuration to square tube and bumped up the wall thickness and made things much easier for all involved (not to mention increasing our safety factor margins). Fingers crossed that it gets approved this time around.

I would however still be interested if anyone out there has a way to calculate the moment of inertia for an ellipse.

Cheers,

 

[SnTMan]

From my handbook: Ix = 1/4 * pi * a * b^3, with b dimension perpendicular to x axis.

Regards,

Mike

 

[Ron]

Your pipe is rotating around the heel of the weld at the inside. That puts the entire weld in tension. You need only resolve the moment into a force couple, averaged over the weld area. You can distribute the concentrated force of the couple in tension with an equivalent distributed load for a bit more distributional accuracy.

 

[chicopee]

Ron, while the weld will be in tension on the flat plate, wouldn't the weld also be in shear with the pipe?.

 

[kingnero]

Your pipe is rotating around the heel of the weld at the inside. That puts the entire weld in tension.

I beg to differ: I'd say the smaller part of the weld is in tension.
see sketch attached: when combining both the applied moment and the force, you get the idea as per the diagram on the right.