Noninvasive Kalman Filter Based Permanent Magnet Temperature Estimation for Permanent Magnet Synchronous Machines

Abstract

Permanent magnet temperature (PMT) is crucial to high-performance control and condition monitoring of permanent magnet synchronous machines (PMSMs). This paper proposes a noninvasive PMT estimation approach based on the PMSM steady-state equation. First, a linear temperature model, dependent solely on the PMT, is derived from the steady-state equation and the PM thermal model. Thus, the PMT can be directly estimated from the measurements using the derived linear model. In order to improve the estimation performance, a linear state-space model is developed based on the derived model, and the Kalman filter is employed for PMT estimation. The inverter nonlinearity is considered and compensated in the proposed model to improve the estimation performance. Compared with the existing methods, the proposed approach is noninvasive and computationally efficient. More importantly, the derived model does not involve machine parameters such as winding resistance and self and mutual inductances, and thus, the proposed approach is independent of winding temperature rise, magnetic saturation, and cross-coupling effect. The proposed approach is evaluated with extensive experiments under various speed and load conditions.