Application of Adaptive Neural Controller and Filter Tune for Multi-Mass Drive System

The article presents a comparison of dynamic properties of adaptive neural controller for the system with a complex mechanical structure and variable moment of inertia. Due to presence of several mechanical resonances (since it is a multi-mass system) at higher frequencies two filters were implemented: a simple LPF (Low Past Filter) and Fixed Filter, taking into account the changing parameters. A suitably tuned speed filter is applied to the feedback loop in order to damp the selected resonance frequencies In the paper was placed a results of simulations and laboratory tests with PMSM motor for the multi-mass system of variable parameters.