Analysis of a Welded Structure.

[LOKI1983]

Hello.
I would like to ask help in trying to evaluate a structure which is welded.I am no expert in Abaqus but I like to get to the bottom of it and with the help of Eng-Tips colleagues.
I had attached the drawings in a simplified form for my problem.Essentially I have the same principle for the scissors at which I apply a force at both ends with a hydraulic cylinder.
I have seen that the methodology has evolved for estimating the weld size in FEA, and for what I have seen I am very interested in the R1MS method approach.Now my question is how to I do this in Abaqus.I am modeling my structure in Solidworks because it is easier for me.For me there are 2 steps:
1) To model all the parts as separate parts.To model the weld chords.Then to import in Abaqus,create the assembly and contact conditions between the weld and parts.But from what I see I could not create the 1 mm radius required at the toe of the weld and at the end of penetration.
2) To create a entire solid in Solidworks with the weld geometry and the radius of 1 mm and then to import it in Abaqus and simply to mesh it, input the boundary conditions and loads and solve.
Could someone tell me if my approaches are correct?I don't imply that what I had written is a easy task.
Please Advise.
Loki.

 

[TerryR1]

We run all our weldments as one body. Gets a little tricky to put the gaps in behind fillet welds. But once modelled, FEA gives good results.

 

[LOKI1983]

Hello Terry.
Thank you for your answer.In your analysis are you considering the weld as a material with different properties? Are there any consideration to take in account except the weld geometry?
This weekend I shall finish the geometry of my model regarding with the R1MS method and I shall post in this thread and Abaqus Forum (where I had opened the same thread) the model with BC and LC so that should be clear the type of analsys is done. .
It will be welcomed if users from Eng-Tips could share their opinion regarding different methods of approach for welded structures.
Thank You.
Loki.

 

[rickfischer51]

They way I analyze welds with FEA is to use the FE solution to generate reactions at the weld, which I take to a spreadsheet and use with a traditional hand calculation ala Blodgett or Shigley. It's my understanding that it is impossible to get accurate stress predictions. A real weld has variable geometry, residual stresses, slag inclusions, stress concentrations, etc. Also, the weld allowables for your design code are intended to be used with hand calculations, and mixing allowables with an alternate method of calculating stresses is not a good idea.