IEEE Power Electronics Magazine - September 2023 - 15

topologies, control schemes, and hardware fabrications
have greatly improved circuit efficiency and power density,
the bottleneck now lies with magnetic components [4],
with magnetics accounting for more than 30% of the cost
and more than 30% of the loss in almost all power converters
[5]. Magnetics design has become a critical issue for
power electronics as trends towards high efficiency and
high power-density.
The limits set by bulky and lossy magnetic components
must be broken by 1) radically new magnetic design techniques
and 2) novel magnetic materials with improved
properties (e.g., higher saturation limit, higher permeability,
and lower core loss).
Why is it Difficult to Deal With Magnetics?
The challenges of magnetics design and optimization are
attributed to two main aspects, both of which must be
addressed by magnetic material modeling capable of predicting
component behaviors:
First, deep understanding and accurate design tools
are required in magnetic components to comply with the
trend towards high frequency and high density. Instead of
being satisfied with the performance given by an off-theshelf
inductor, power electronics engineers nowadays are
obliged to design magnetic components from scratch with
highly customized cores and windings that yield better performance.
This requires advanced knowledge of magnetic
materials, loss analysis, high-frequency effects, and simulations
that aid the design. Advanced electromagnetic simulators
[6], such as 3D finite-element analysis (FEA) tools,
are able to simulate linear performance factors governed
by the Maxwell's equations, including winding loss, fringing
effect, geometry-based non-uniformity, and even numerical
optimizations. However, there is one exception, and it is the
major one that causes the discrepancies seen frequently
between simulations and real prototypes: the nonlinear
magnetic materials. Lack of accurate models and understanding
of the material's properties (such as dc bias- and
frequency-dependent loss and permeability) when used as
a power electronics component results in oversimplified
assumptions in magnetic design and then iterations of magnetics
prototyping based on trial and error.
Second, new magnetic materials are needed to break the
ceiling of magnetic component performance set by intrinsic
physical properties (e.g., saturation level, permeability, and
loss density) of existing materials. To this end, component
or even system level insights are essential to guide new
materials development. However, magnetic components
design is not a single-objective optimization process [7];
an optimal design balances several performance factors.
For example, a material that has an infinitely large permeability,
but low saturation level, doesn't necessarily lead to
improved component performance. An envisioned paradigm
of material-component-system co-design needs to
be facilitated by a simulation platform that links magnetic
material properties component prototype performance
metrics.
However, magnetic materials modeling is never an easy
job. The two most concerned performances of magnetic
materials in power electronics applications are permeability
and core loss, both of which vary significantly with
frequency and the strength of the magnetic field (excitation
signal) applied. Take the MnZn ferrite material N87 from
TDK [8] as an example (Figure 1). The nonlinear permeability
and core loss vary with frequency and ac excitation
amplitude. Besides the nonlinearity with ac excitation, the
effect of dc bias on the permeability and loss also strongly
impacts the core performance [9]. Such complexities,
rooted in the nonlinear dynamics of magnetic materials,
hinder good agreement between designs and prototypes of
magnetic components and lead to multiple trial and error in
engineering practice.
In practice, power electronics engineers tend to use
empirical approaches to characterize magnetic materials
based on large-signal measurements. Models to account
for the core loss and permeability, mostly empirical, are
extracted by curve-fitting experimental results measured
from specific prototypes [10], [11], with measurements
getting more extensive as more complex behaviors
of the core are observed. Even the physics-based models
FIG 1 Nonlinear behaviors of magnetic material N87: (a) permeability versus frequency, (b) permeability versus ac excitation
amplitude, (c) core loss density versus frequency and ac excitation amplitude [7], and (d) core loss versus dc bias [8].
September 2023 z IEEE POWER ELECTRONICS MAGAZINE 15

IEEE Power Electronics Magazine - September 2023

Table of Contents for the Digital Edition of IEEE Power Electronics Magazine - September 2023