The paper presents a formula useful for prediction of loss density in soft magnetic materials, which takes into account multi-scale energy dissipation. A universal phenomenological P(Bm, f) relationship is used for loss prediction in chosen soft magnetic materials. A bootstrap method is used to generate additional data points, what makes it possible to increase the prediction accuracy. A substantial accuracy improvement for estimated model parameters is obtained in the case, when additional data points are taken into account. The proposed description could be useful both for device designers and researchers involved in computational electromagnetism.

This paper presents a method for estimation of core losses in banks of single phase power transformers that are subjected to an injected DC current such as geomagnetically induced currents (GIC). The main procedure of the core loss calculation is to obtain a magnetic flux density waveform in both time and location by using a novel algorithm based on 3D FEM inside the core and then to calculate the loss distribution based on loss separation theory. Also, a simple and effective method is proposed for estimation of losses of asymmetric minor loops by using combination of symmetric loops. The effect of DC biasing on core losses in single phase power transformers is investigated and the sensitivity of core type and material is evaluated. the results shows that DC current biasing could increase core losses up to 40 percent or even more.