A Data-Driven Method for Iron Loss Estimation in Bearingless Permanent Magnet Synchronous Motors

This paper proposes a data-driven iron loss estimation method to reduce the calculation time of iron loss in bearingless permanent magnet synchronous motors (BPMSMs). The iron loss calculation dataset is obtained by design of experiments, and the iron loss prediction model is established under different input parameters and working conditions. Firstly, the stator core flux density changing rules at selected points are analyzed, and the flux density components of different parts of iron loss are studied. Secondly, the effects of temperature and frequency on iron loss are studied. Finally, a data-driven iron loss estimation method is established based on the nonlinear autoregressive exogenous (NARX) model using input current, frequency and temperature data. In the proposed method, the iron loss of BPMSM under all working conditions is considered, which significantly shortens the calculation time compared to the finite element analysis. This method can be used to quickly obtain the iron loss variation range of each speed and different working conditions and provide a basis for calculating motor efficiency and design optimization.