Examination of Temperature-Dependent Iron Loss Models Using a Stator Core

Abstract

In order to examine the popular temperature-dependent iron loss models with unidirectional magnetization, we cut and measured a ring sample and stator core with several hundreds of combination of temperatures (RT-200 °C), frequencies (20–1 kHz), and magnetic inductions (0–1.9 T). Coefficients of various iron loss models are extracted based on the sample test, and then imported into the finite element method postprocessor to calculate the iron loss of the stator core. We found that constant coefficient loss separation temperature-dependent iron loss models can achieve a good fitting accuracy (about 5%) on ring sample in a wide range of 20–1 kHz, 0–1.9 T, RT-200 °C, while loss coefficients extracted on small ring sample could lead to 15%–20% error when predicting iron loss of larger sized stator stamp. The iron loss of silicon steel decreases with rising temperature, and thicker silicon steel is more influenced by temperature.