Extract Stress at Node from Raw Output

[ESPcomposites]

By default, most solvers seem to output element stresses and not stress at a node. In general, if I want to get a stress at a node, I would do it in the post processor.

But for what I am doing, I will not be using a post processor, but rather just reading the output into Excel. What would be some methods to extract the stress at a given point? Since it is a high stress gradient area (near a cutout), it has to be very accurate to be meaningful. In other words, I can't just use the direct element stresses.

Brian

http://www.espcomposites.com

 

[TGS4]

You do appreciate that the "nodal" stresses are simply the extrapolation (using the element shape function) from the gauss points? So, you need to know the element shape function a priori before doing what you're proposing. By the time you go ahead and do that, you might as well be routing things through your post processor.

Of course the question of whether or not the nodal stresses are accurate is entirely dependent on whether or not your solution is grid converged...

 

[ESPcomposites]

Yes, I do understand all of this. But for what I am proposing, I cannot use the post processor.

For this project, I will only be running Calculix, a solver similar to ABAQUS. The pre processing will be done in Excel and the raw data is read back into Excel. Unfortunately, the post processor is not an option.

I am looking for some ideas I suppose. Something I thought might work would be to make a small element in the area of interest such that it could be approximated as a point. But that could get tricky as well for several reasons. Any other ideas though?


Brian

http://www.espcomposites.com

 

[TGS4]

Do you know the element shape functions? If you do, then the math is rather straight forward - sort of like generating your own post-processor.

 

[ESPcomposites]

Calculix is open source so I should be able to get the shape functions. But I think that is going to be more complicated than what I really want to get into.

I will try to exhaust all of the simple approaches and see what I can come up with. If nothing works, then my backup plan would be to request a feature like this from the developer. He has said he will work with me on it, but I don't want to get into that just yet.

Brian

http://www.espcomposites.com

 

[rb1957]

you should be able to get nodal stresses ... sure the default fringe stress is probably element CG stress but (in FeMap at least) you can query the element and get the nodal stresses, whihc could be element averaged nodal (ie the stress at the node is the average of the nodal stresses from the various elements (that share the node).

or you could look into the F06 output file ...

this assumes you're running a reasonable FEA code (NASTRAN, ANSYS, ...) and not something "dumbed down" for designers

 

[johnhors]

Calculix outputs average nodal stress values in the frd results file.

http://www.Roshaz.com

 

[ESPcomposites]

Johnhors, thanks. I was looking into last night and thought that might be the case, but did not have time to run a test model. Now that you have confirmed that, I can move on to other parts of the project that might trip me up. So thanks for the help there.

Brian

http://www.espcomposites.com