Using Von Mises to size welds 2

[JAG-E]

I came across a paper (Weaver Engineering Seattle WA, 1997)that indicates how to use FEA derived Von Mises to size welds. The examples are simple like a Tee bracket, but the method loses me. The part I am working with is much more complicated in shape and loading. So if I can get to the point of understanding the published method I don't know if it could be applied to my case.

The analyst running the FEA is providing VM and Max principal stresses. Prior to the above paper I thought the VM stress could be used in conjunction shear strength of the filler to size the throat and length of the weld. Now I am thinking it is more complicated. I have done many weld hand calculations using methods from AWS and Roymech

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

The component is part of an an axle for a large coal haul truck. The part is a thick wall tube in bending, as in a beam, and also torsion at the ends. There are internal ribs that will be fillet welds. The main welds will be full penetration so according to the paper the VM can be used directly. I would likely used a higher factor of safety at the weld to account for weld defects and stress concentrations, not likely found in the base metal.

Any advice?

Joe

 

[AWDMIKE]

I am very interested to see where this post goes.

Recently I have had to size several welds and found myself asking some of the same questions. Even referencing AISC 360-16 if you have a fillet weld the specification only has criteria for shear in the weld, and states that "tension of compression in parts joined parallel to a weld is permitted to be neglected in design of welds joining the parts." While this may not be directly applicable to your situation, it's something to think about.

I have recently seen a paper which discussed the complexities of stress distribution within a weld, basically suggesting that you could make a research project out of it.

 

[scheah]

Hi JAG-E,

I assume you are referring to this paper: [URL unfurl="true"]http://www.weavereng.com/feweld/dload/FEWeld_WhitePaper.pdf[/url]. On page 5, he used total weld load to calculate required weld size. This is not dissimilar to your method of hand calculation.


Kind regards,
Jason

 

[JAG-E]

In simple brackets as shown in the paper (you are correct as to which paper) I can do the weld calcs without the FEA.

I have a 1.25" thick wall steel tube, 65" in diameter, 95" long, two vertical loads evenly spaced from the center. Each concentrated load 396 Tons. At the ends of the tube, bolt circles with (60) 1.25" dia bolts each side. Each bolt circle flange has an applied torque 807,700 ft-lb (yes Ft not in).

The VM stress does not indicate orientation to weld. If the steel part are being pushed together the weld may be doing little. If the force is 90 degrees to weld we have a different situation.

As I type this I think the easiest way will be to estimate the stresses by hand cal's and proceed with weld cal's.

Joe

 

[dik]

Sk_cheah: Good paper...

Dik

 

[Tmoose]

Will the parts be stress relieved after welding?

I bet As-deposited welds in restrained joints have significant regions with residual stresses right at the yield strength of the material.
Some sources elaborate that even the residual stress distribution near the weld start and end points can show extraordinary patterns since the heat sequence at the weld start and end points is so complicated.

https://onlinelibrary.wiley.com/doi/pdf/10.1002/srin.200200008

http://publications.iowa.gov/22401/1/IADOT_HR_94_Residual_Stresses_Fatigue_Behavior_Welded_1967.pdf

https://onlinelibrary.wiley.com/doi/pdf/10.1002/srin.200200008