Tangential Load for a Cylinder

johnkent

Here is a Mecway model for the torsion-tube tester in _Mechanics of Composite Materials_. Using node values to simulate strain gages, the "measured" shear modulus at various angles of uniaxial fibers agree nicely with theory.

However, to model a uniform torque around the end of the tube, I didn't see a direct way to apply a tangential load. Finally I resorted to adding a circle of truss elements around the tube end and applying the line pressure load along the local "u" coordinate. this works, but is there a more direct approach?

Victor

There's no more direct approach yet. One suggestion though - I would use beam elements that are much softer than the tube's material, otherwise they introduce a little bit of artificial stiffness when they have similar Young's modulus and thickness.

I'd like to have a more general moment load that's distributed over a surface or edge, which would work here too. There's a philosophical problem with these types of loads though - they improve the accuracy of modelling something that's already an idealized model. Text books use them and they make hand calculations easier but I don't know anything a uniformly distributed tangential load or moment corresponds to in real life.

If you're willing to use such an idealization, then you might as well choose whatever is easiest - in this case it might be two point loads, one on each side of the cylinder. Of course, the results won't match the text book but the text book won't match a real cylinder either because whatever applies the load will probably also add some stiffness.