Summary
This paper first deals with the sensorless estimation of shaft torsional torque, and turbine and generator speeds in permanent magnet synchronous generator (PMSG) based wind turbines (WTs). In this way, cascade observers, comprising the inner and outer ones, are introduced for estimation of the turbine and generator speeds and shaft torsional torque. Then, by using the estimated speeds of the turbine and generator, a new sensorless active damping compensator is proposed that is efficient for suppression of WT torsional vibrations. The proposed sensorless compensator operates based on the difference of the turbine and generator speeds and provides sufficient damping even under the drive train model uncertainty. The conventional approaches, used for damping of WT torsional oscillations, mainly do based on the generator speed feedback and with using the band pass or high pass filters. However, the performance of the conventional approaches, because of existence of the band pass/high pass filters, may be failed under the drive train model uncertainty because the drive train resonant frequency changes under the drive train model uncertainty. At the end, the performance of the proposed sensorless compensator is compared with that of the conventional damping approach with/without considering the drive train uncertainties.