Welding 2

[Pingen]

Hi,

I want to determin the minimum safety factor of a statically loaded weld. Does this task require any weld-specific considerations, or can I simply approximate the case by connecting the welded parts with a separate weld-like part?

Pingen

 

[JStephen]

I'm not sure how to approach it, but do know of some things to watch for. First off, with fillet welds, it wouldn't be too surprising if you calculated local yielding in places even at otherwise acceptable stress levels. Also, welding in general involves a fair bit of residual stress of unknown levels, so the stresses you might calculate aren't necessarily going to be the actual stresses. And finally, the weld metal is normally a higher alloy than the base metal, and therefore often stronger than the base metal, which would complicate the modeling. Welded test specimens are expected to fail in the parent metal, not in the weld metal. Also, welds are more susceptible to fatique than the base metal.

 

[GBor]

You have to consider the "heat affected zone", which is dependent upon several factors associated with the welding process. The reduction in material strength due to this heat effect is significant for most metals. You also have weld porosity, which is difficult if not impossible to model with any confidence, so you should probably reduce the cross-section of your "weld-like part" to give yourself some comfort. And, of course, you have to know the penetration of the weld...full pen? partial? No pen? It isn't an easy situation.

I think most analysts simply ignore the weld bead on large, complicated models and we look at the areas that would be welded for stress beyond the "heat effected zone" knock-down value. If it appears to be a problem, we model local areas and add the weld beads.

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Lower Alabama SolidWorks Users Group

 

[Pingen]

Thanks a lot for your replies.

You guys, describe weldments as complex structures that can be analysed with different levels of detail. I have a feeling that the relation between the amount of time spend on the analysis and the accuracy of the obtained results is quite non-linear, and that it sometimes takes much additional time to achieve slightly better results. Usually, I don't need to obtain super accurate results, but they must be on the safe side and reasonable. I need a happy-medium-method that I can trust.

Furthermore, I'm not very experienced in FEA and I think it would be interessting to know how other professionals analyse weldments.

Thanks again,
Pingen

 

[JStephen]

Hopefully, you realized that asking this in a FEA forum points your answers in a certain direction. People don't generally use FEA to analyze welds, if that helps- much more mundane techniques are available. For certain geometries and codes, the allowable loads are simply tabulated- see AISC books, for example.

 

[torpen]

Hello,

This article can be helpful:

http://www.aws.org/wj/supplement/Weaver/ARTICLE2.pdf

For a fillet weld, the required conditions are as follows:

1) equivalent stress < tensile ultimate strength / (beta_w * gama_M2)

2) normal stress prependicular to the weld throat < 0.9 * tensile ultimate strength / gama_M2

where,
beta_w : correlation factor function of the grade of the steel
gama_M2 (= 1.25) : safety factor for welded assemblies

In spot weld analysis, no spot tensile shearing may be lower than a given shear force based on the thickness of a sheet e, the diameter of the weld nugget, the material and the tensile shear ultimate strength.
Tables give this type of data.
If two sheets of different thicknesses are welded together, e is the thickness of the thinner of the two.
In this case you can use a coarse FE mesh to transfer loads only.

Regards,

Torpen

 

[dulmant]

Hi Pingen,

All the above are valid answers and, as you said it yourself, cover the subject way deeper than you need.

Going back to the core of your question: for simple applications it is reasonable to consider that you don't need to model the weld, as long as you probe the results in the areas where the weld is and compare these results with allowable values for welds. As JStephen said, there are "documents" where you can find these comparison values globally accepted as reference documents, namely the AISC, DIN or FEM norms and codes (ex. FEM 1.001) - which actually take into consideration the type of weld, loading direction, etc.

Just make sure your application can be considered a "simple" one, with no asymmetric or particular stress concentrating welds.

dulmant