Torque-speed Adaptive Observer and Inertia Identification without Current Transducers for Control of Electric Drives

This paper develops a high dynamic and robust speed-torque control system for servo drives based on adaptive observer and self-tuning techniques with on-line inertia identification. Only the rotor position is measured even for a low precision shaft encoder. The instantaneous speed and the disturbance torque are accurately estimated by an extended Luenberger observer inertia adaptive. A current to voltage decoupling and disturbance torque feedforward compensation realize a fast current sensorless torque control with dynamic torque limitation. The mechanical inertia, used for the selftuning of both speed controller and Luenberger observer, is on-line identified by a discrete time recursive gradient algorithm. Extensive simulation results with the proposed control structure, applied to a permanent magnet synchronous motor (PMSM) drive, prove high-dynamic performances in wide speed range. A good robustness to 1:10 inertia variation at rated load torque is obtained.