Sensorless Control Strategy for Magnetic Drive-Trains Based on Adaptive Nonlinear State Observer

The magnetic drive-trains (MDTs) is widely employed in the various industrial fields because of its advantage of non-contact torque transfer. However, due to the non-linear torque transfer characteristics of permanent magnet couplings (PMC), the performance of the commonly used sensorless control strategy based on linear state observer dramatically decreases when it deviates from the linearization point. To solve this problem, a new sensorless control strategy based on adaptive nonlinear state observer is presented. By introducing an adjustable model, the system is transformed into a non-linear error feedback system consisting of linear forward and non-linear feedback paths. An asymptotically adaptive nonlinear stable observer is constructed by introducing an error feedback compensation matrix into the forward path and designing an appropriate adaptive law for the feedback path based on Popov stability criterion. The real-time observation of magnetic torque stiffness factor is achieved, and the estimation accuracy of electrical slip angle and load-side speed is improved. Finally, simulation experiments verify the validity and accuracy of the proposed scheme in load-side speed estimation.