부산대학교 도서관

Monitoring rotor temperature is a significant issue for safe operations in magnetically suspended turbo molecular pump (MSTMP). The increase in rotor temperature may cause the irreversible demagnetization of the permanent magnet (PM) and destroy the vacuum environment, eventually affecting the efficiency and security of the MSTMP system. Unfortunately, the direct measurement of the rotor temperature is complicated in practice due to the difficulty of accessing temperature sensors on a rotating shaft. This article proposes a novel rotor temperature estimation method based on the identification of the PM flux linkage. In order to address rank-deficiency in the identification of multiparameters, the resistance is calculated from the measured stator temperature, and two recursive least square (RLS) algorithm segments are combined to identify the inductance and flux linkage. Furthermore, the dynamic factor is introduced into the RLS algorithm to enhance the identification accuracy and convergence rate of the algorithm. The experiment results on the MSTMP test rig, equipped with the infrared temperature measuring module, demonstrate that the identified flux linkage has a good performance for the estimating rotor temperature under different working conditions.