Fusion of a lumped-parameter thermal network and speed-dependent flux observer for PM temperature estimation in synchronous machines

High utilization of permanent magnet machines without monitoring their internal temperatures has negative impact on windings and permanent magnets. Thus, different temperature estimation techniques such as lumped-parameter thermal networks (LPTNs) and permanent magnet temperature observers (PMTOs) have been developed to estimated magnet and winding temperatures. LPTN's identification is an intricate process and, therefore, its accuracy varies from model to model. PMTOs, on the other hand, show disturbance rejection problems which yield large temperature estimation errors at low speeds. Since a high accuracy and robust disturbance rejection are desirable, fusion of different techniques is currently under research. In this paper, a new observer topology is presented which makes use of the speed-dependent Gopinath-style flux observer. Thus, this topology combines an LPTN model, operating at low speeds, and a voltage-model-based observer operating at medium and high speeds. Simulation and experimental results on a 50 kW IPMSM show a performance enhancement over the standard topology; particularly, a superior disturbance rejection to voltage estimation errors over the entire operating range.