pipe elbow stress modelling 2

[yoojeen]

I am trying to develop a model for an elbow in thick-walled power piping where the wall thickness is non-semitrical due to errosion / corrosion. This model is not one of the standard models in ANSYS or Abaqus which are the two FEA programs that are available to me. Thanks for your help.

 

[corus]

Is there a question here?

 

[loki3000]

the question is, how does one model a non-circular form, as the original circle is eroded and is not an exact circle.

answer: do it with splines/ curves, but it will be somewhat difficult.
if you have autocad (or similar free program), you can cut the pipe, photograph its profile, insert into drawing and draw the shape based on it.
then you can import it into abaqus somehow (dxf, dwg) or make a 3d model in autocad.

http://cad-notes.com/2011/06/how-to-embed-image-to-autocad-drawing/

 

[yoojeen]

I am sorry that I was not clear. I have ultrasonic inspection data that indicates thinning of a thick walled elbow and this thinning is non-symmetrical. Based on this data I can develop a reasonable guess as to the wall thickness at various locations. I have access to ansys and abaqus be neither one has a model in the example data that I can build upon. Does any one know where I could obtain a model for either of these programs or is there another program that I should look into. Thanks for your help.

 

[TGS4]

When I have been faced with a similar problem, I developed a python (for ABAQUS) or APDL (for ANSYS) script that modifies the mesh of an ideal elbow, based on the UT data. The toroidal geometry of the elbow makes the math rather complicated (have you ever done geometry in toroidal coordinate systems before??).

ANSYS has the advantage of being able to use a local toroidal coordinate system built-in. ABAQUS has no such thing. If I had the choice, I know which program I would use...