Analysis of Magnetic Properties of AlNiCo and Magnetization State Estimation in Variable-Flux PMSMs

Variable-flux permanent-magnet synchronous machines have gained a lot of attention for their ability to expand the speed range and reduce loss over an electric vehicle drive circle by manipulating magnetization state (MS) dynamically. In this paper, a combined-magnet-pole variable-flux machine (CMP-VFM) using both AlNiCo and NdFeB is analyzed. To obtain the variation laws of MS quickly and precisely, a 3-D neural-network hysteresis model of AlNiCo is proposed, and the accuracy and time consumption of the proposed neural-network hysteresis model are evaluated. In order to optimally utilize the speed expansion and loss minimization capability with precise MS control, a flux observer based on model reference adaptive theory is proposed for the closed-loop flux control of CMP-VFM and validated by simulations. The simulation results prove that the accuracy of the proposed adaptive flux observer is almost not affected by parameter disturbance.