Speed controller for a drive with complex mechanical structure and variable parameters

The paper presents the problem of obtaining a good dynamics of the PMSM drive with complex mechanical structure and variable moment of inertia. Proposed approach bases on proper design of feedback filter to reduce influence of mechanical resonance's and applying a neural speed controller trained online, which can adapt itself to variable inertia. As an alternative, robust PI controller with set-point weighting is presented. The results of simulation and experimental investigations show good properties of designed control system.