Determining effective force and moment through solid section

[bfillery]

Hi all,

I was wondering if anybody had any idea how to determine the effective axial force and bending moment that results from a solid cross sections stress distribtution. That is, say I have a cross section through either a p.strain p.stress or 3D simulation, and I would like to resolve the stresses acting on that cross section into effective forces and bending moments for a subsequent beam type analysis.

Can this be done by writing a python script and utilising NFORC in some way, or is there an alternative approach? The cross section is not nessecarily associated with boundary condtions so I don't believe I can sum the reaction forces etc.

Any suggestions.

Thank you

bfillery

 

[corus]

See the FAQ in the boiler and pressure vessel forum. Basically the axial force is simply the average of the axial stress times the area. The equivalent linearized bending moment is calculated from the axial stresses less this average axial stress. such that the moment is the sum of the equivalnet forces times the distance from the neutral axis. Personally I'd write the stresses to a file together with the co-ordinates and write some code or use a spreadsheet to do the calculation rather than use puthon.

corus

 

[brep]

It would be pretty simple to code up a Python plugin (with GUI!) to do this, and could be made very general in terms of picking groups of elements etc etc (the Abaqus scripting and GUI programming APIs have all you need).

On the other hand, people who have access to V6.8 preleases (beta) might find something interesting in store...

 

[bfillery]

Dear corus and brep

Thanks for your replies.

Corus, in response to your reply, I have a further couple of questions.

Firstly, when calculating the axial force I understand that it is the average of the axial stress times the area, but when working in regards to a finite element simulation, what axial stress is used (i.e. stress reported at the elements or nodes) and what is the associative area?

Secondly, with regard to the moment, again, I understand that it is the equivalent forces times the distance from the neutral axis. How does one attain what these equivalent forces from the finite element simualtion?

I pictured that it would be most benefical to be able to first determine the effective axial forces acting on the nodes along the required section, then average these etc to determine the axial force etc. If this seems an appropriate apporach how does one either calculate or report these effective nodal axial forces.

Thanks

bfillery

 

[corus]

If you're using linear elements then the element stress would be the one to use in calculating the axial force as there would be only one value acting over the element area, though this wouldn't be the surface stress. You could however 'coat' the solid elements in thin shell elements so that you get a better value of the surface stresses and use those values. As another approximation though you could use the average of the nodal surface stresses for each element, or assume an associated area around each node, say a quarter of the surface area for each linear element, for which the nodal stress applies. The values calculated would be the equivalent force at each node.

I've not seen 6.8 but I do know that some codes provide an output of the equivalent linearized bending stress across a section, though I believe that this is for 2D axisymmetric models where there is a line of nodes across the section. This enables the stresses to be assessed against primary/ secondary stress limits for pressure vessels, without the addition of peak stresses at the surfaces. I've not seen anything as yet for stresses acting across a 2D area in commercial codes.

corus

 

[bfillery]

Thanks corus

I have discovered that a cross section can be defined in ABAQUS using *SECTION PRINT. The effective forces and moments across this section can then be output but not in an .odb, just in the .fil or .dat.

A work around is to define a sub-model with a submodel boundary at the cross section of interest. The reaction forces on these submodel boundary nodes can then summed appropriately to extract the effective force and bending moments.

It will be interesting to see what 6-8 has in store.

Thanks for your help.

bfillery

 

[corus]

Interesting, and a good work round. The sub-model will of course obtain the displacements from the main model from which you can obtain the reaction forces at these 'constraints'.

Personally I think what ever 6.8 offers will no doubt be fixed in the next version, and so on..

corus