Discussion Forum

I posted a question recently about trying to apply a distributed torque density to a model, but to no avail. Since then, I've come up with a way to potentially apply the load, but I'm not sure if it's possible to use the inputs I want.

I'm able to convert my torque load into an applied stress tensor which includes only shear stress terms. Since there are no boundary conditions on my stress element, it's not in static equilibrium and so has imaginary eigenvectors and values. If I could calculate them, I would just find the appropriate body load to simulate the tensor.

I was wondering if it's possible to input a stress tensor as an boundary/initial condition. The torque is always present, so COMSOL would need to calculate the stress produced because of the boundary conditions.

If my description above was not clear, here's the idea:
1. Make a shaft that's fixed somewhere.
2. Apply a torque about an arbitrary axis using the stress tensor (which may be spatially or time dependent). For a non-fixed geometry, the stress tensor has components sigma_xy=-sigma_yx=(torque about z-axis), -sigma_xy=sigma_zx=(torque about y-axis), sigma_yz=-sigma_zy=(torque about x-axis)
3. Calculate the steady-state stress/strain/deflection of the model.