Damping of resonance peaks using adaptive notch filters in gearless servo drives

The performance of a gearless servo system is significantly degraded by mechanical resonances present in the system. For applications on a machine the parameters of the mechanical resonances differ from workpiece to workpiece and thus cannot be calculated a priori. Adaptive notch filters can dampen these unknown resonances online. In this paper an online mechanical resonance frequency identification algorithm and online notch filter depth tuning algorithm are presented which will be processed in a closed loop control system with a control frequency of 32 kHz in software. The performance of these algorithms is shown on a test axis and on a round table with a direct torque motor on a machine. It will furthermore be shown that multiple resonances present on the machine can be damped subsequently and the controller gains can be increased to improve the overall tracking performance.