Calculating weld stresses from FEA Shell model

[jimchest]

Hi All
I am reviewing some documentation from another engineer and am having a tough time finding references to back up the equations/Assumptions. Basically what was done was a shell model was created and von-mises stresses were captured at the plate interfaces. Then the procedure multiplies the stresses by the siesmic coefficient (which I understand) and also mulitiplies by the square of the plate thickness divided by the throat distance of the weld. I have no Idea how that relationship was developed. Thanks for any insight anyone can provide. SEE BELOW PROCEDURE;

The calculation of seismic stresses, if required, for the fillet and groove weld connections between different components of the SYSTEM Assembly, are based on the following:

1) The welded structural members are made of similar or different materials.

2) Consider, conservatively, that the maximum F.E. stresses presented in Table 4-1 at weld interfaces occur at the entire length of the welded connections.

3) Consider, also conservatively, that the Von Mises stress value reported at the weld interface (Table 4-1) is caused by combined shear, normal and bending stresses.
Note that:
For normal and shear stresses; fw proportional to (tc/twt)
Whereas,for bending stresses; fw proportional to (tc/twt)^2

4) Accordingly, the maximum weld stresses at the critical (i.e., throat) section of the welded connection can be calculated as follows:
fw = fc x (tc/ twt)^2

where: fw = maximum seismic stress in the throat section of the welded connection, psi

fc = maximum seismic stress in the base metal, psi, calculated from stresses presented in Table 4-1 and seismic stress formulas presented in section 6-3.

tc = base metal thickness of the structural component most stressed at the weld interface.

twt = Throat size of welded connection
For fillet weld;
twt = fillet weld size/ sqrt(2)
For groove weld;
twt = groove weld size

 

[dhengr]

Jimchest:
What you appear to be doing is certainly not welding engineering; rather it is dangerous ‘documentation reviewing,’ not to be confused with engineering of any sort, or a real fundamental understanding of the problem.

What kind of structure are you designing and welding? Who determines whether to use a fillet or groove weld at your company? Who selects the size of the fillet weld or groove weld and the configuration of that groove weld? Who selects process and filler metal for these ‘similar or different materials?’ If the materials at a joint are different, which material do you match with your filler metal? When do you apply the seismic coefficient? Are any of these joints subject to fatigue? I can see glimmers of meaning in your calculation procedure, but am not going to spend much time trying to make sense out of it. The whole thing seems to me to be an extreme example of arithmetic manipulation in an attempt to avoid any knowledge of what you are really doing, while calculating your way to a final number, maybe not a meaningful answer (solution) to a problem. You might want to get some good text books on Mechanics of Materials, and Theory of Elasticity; and learn about von Mises stresses, combined stresses, normal, shear and bending stresses and how they relate to von Mises stresses.

The James F. Lincoln Arc Welding Foundation (Lincoln Electric) has some very good books on the subjects of Welding and Weld Design, and they are very reasonably priced too. Maybe you should avail yourself of some of these books, so you can learn how to actually design a weld joint for a given set of stresses. And, keep in mind that the design of the welds is not all that complicated, if you know the basics. It’s our damn complex designs, shapes and joint configurations, which lead to multi-axial stress conditions, and highly restrained or hard spots, on which the welds are needed to work, which muck up the works.