Rapid Thermal Analysis of Nanocrystalline Inductors for Converter Optimization

Abstract

To capitalize fully on modern component technologies such as nanocrystalline cores and wide-bandgap devices, multi-objective converter design optimization is essential, requiring simple, accurate component models. In this work, a lumped parameter thermal model is presented for nanocrystalline inductors with ceramic heat spreaders. The key challenge is the non-uniform loss distribution in gapped, tape-wound cores, particularly the high loss densities adjacent to the gaps. However, uneven loss distributions are not handled easily by lumped-parameter techniques. It is shown that by treating the ceramic heat spreaders as 'passive' heat sources, a simple thermal model of the inductor can be derived to estimate the hot spot temperature of the core. The model is validated through comparison with 3-D finite element analysis (FEA) and experimental measurements on a 60 kW DC-DC converter. The proposed model offers a comparable level of accuracy to FEA with a fraction of the running time, executing in 99 µs in MATLAB.