Accurate demagnetization modelling is mandatory for a reliable design of rare-earth permanent magnet applications, such as e.g. synchronous machines. The magnetization of rare-earth permanent magnets requires high magnetizing fields. For technical reasons, it is not always possible to completely and homogeneously achieve the required field strength during a pulse magnetization, due to stray fields or eddy currents. Not sufficiently magnetized magnets lose remanence as well as coercivity and the demagnetization characteristic becomes strongly nonlinear. It is state of the art to treat demagnetization curves as linear. This paper presents an approach to model the nonlinear demagnetization in dependence on the magnetization field strength. Measurements of magnetization dependent demagnetization characteristics of rare-earth permanent magnets are compared to an analytical model description. The physical meaning of the model parameters and the influence on them by incomplete magnetization are discussed for different rare-earth permanent magnet materials. Basically, the analytic function is able to map the occurring magnetization dependent demagnetization behavior. However, if the magnetization is incomplete, the model parameters have a strong nonlinear behavior and can only be partially attributed to physical effects. As a benefit the model can represent nonlinear demagnetization using a few parameters only. The original analytical model is from literature but has been adapted for the incomplete magnetization. The discussed effect is not sufficiently accurate modelled in literature. The sparse data in literature has been supplemented with additional pulsed-field magnetometer measurements.
The accurate prediction of iron losses has become a prominent problem in electromagnetic machine design. The basis of all iron loss models is found in the spatial field-locus of the magnetic flux density (B) and magnetic field (H). In this paper the behavior of the measured BH-field-loci is considered in FEM simulation. For this purpose, a vector hysteresis model is parameterized based on the global measurements, which then can be used to reproduce the measurement system and obtain more detailed insights on the device and its local field distribution. The IEM has designed a rotary loss tester for electrical steel, which can apply arbitrary BH-field-loci occurring during electrical machine operation. Despite its simplicity, the proposed pragmatic analytical model for vector hysteresis provides very promising results.
Some physical concepts important for a hysteresis model (effective field, anhysteretic magnetization) are discussed on the example of Jiles-Atherton model. The Jiles-Atherton model reveals some drawbacks, which make this model more difficult to be applied in electrical engineering. In particular, it does not describe accurately the magnetization curves after a reversal, moreover complex magnetization cycles are poorly represented. On the other hand, the phenomenological description proposed by Takács seems to be a valuable alternative to the Jiles-Atherton formalism. The concept of effective field may be easily incorporated in the description.
A temperature dependent model is necessary for the generation of hysteresis loops of ferromagnetic materials. In this study, a physical model based on the Jiles-Atherton model has been developed to study the effect of temperature on the magnetic hysteresis loop. The thermal effects were included through a model of behavior depending on the temperature parameters Ms and k of the Jiles-Atherton model. The temperaturedependent Jiles-Atherton model was validated through measurements made on ferrite material (3F3). The results have been found to be in good agreement with the model.
An extension of the modified Jiles-Atherton description to include the effect of anisotropy is presented. Anisotropy is related to the value of the angular momentum quantum number J, which affects the form of the Brillouin function used to describe the anhysteretic magnetization. Moreover the shape of magnetization dependent R(m) function is influenced by the choice of the J value.