To compute phenomena like the eddy currents induced in the rotor of an electric machine at steady state (or an induction motor at steady-state), a full transient simulation is normally needed. Because the simulation must start from a known initial state (typically rest) many electrical cycles are usually required to reach steady state conditions, making it VERY time consuming to do (especially in 3D).
I know there's been a lot research into methods that will directly solve for the time-periodic steady-state solution (without the need for a full, time-stepped, transient) but to my knowledge none of the reported methods have made their way into commercial FEA software. However, I did notice that in the latest COMSOL release (4.3b) there was blurb about a tutorial example that uses the Optimization Module to find the "long-term periodic steady-state solution to a nonlinear time-dependent simulation." Does that mean COMSOL can now solve these types of problems directly? That would be great, if so. And if so, does anyone know how that works? As indicated above, I'm interested in applying this to certain rotating electric machinery analyses.
Thanks for any feedback on this!
Pat
I know there's been a lot research into methods that will directly solve for the time-periodic steady-state solution (without the need for a full, time-stepped, transient) but to my knowledge none of the reported methods have made their way into commercial FEA software. However, I did notice that in the latest COMSOL release (4.3b) there was blurb about a tutorial example that uses the Optimization Module to find the "long-term periodic steady-state solution to a nonlinear time-dependent simulation." Does that mean COMSOL can now solve these types of problems directly? That would be great, if so. And if so, does anyone know how that works? As indicated above, I'm interested in applying this to certain rotating electric machinery analyses.
Thanks for any feedback on this!
Pat