Neutral Axis in Torsion

[sapayne]

Hi

I am trying to find the torsional neutral exis of a long beam-like part with a complicated but constant cross section (composite airfoil).

I am applying a surface torque at one end and fixing the other end. Then I am looking for the minimum defelection on the end plate to estimate the neutral axis. 

The issue I am having is that I calculate a slightly different neutral axis depending on if I look at the end of the beam or in the center.

Any suggestions for determining the neutral axis?

[AstroJohnPE]

If the mesh was created automatically, then "a slightly different neutral axis" could be a result of the randomness of the mesh. I would use the average.

I am not an expert with torsion, but my guess is that the neutral axis does not remain constant depending on whether the shape is open or closed. The analogy is a beam in bending. The formulas that calculate the bending and deflections are based on the assumption that plane cross-sections remain plane. So there must be some cases where that is not true. The same may be the case for torsion.

Finally, if you fully fixed one end, then you are forcing that cross section to remain the same size, planar, and perpendicular to the axis. Some of those assumptions may not be true in reality. A better solution may be to apply a torque to each end of the model, and use 3D Springs with low stiffness, attached to 3 points not in a line. The springs will provide the stability needed when the two torques are not mathematically equal through the entire model.