magnetic_components​_toolbox_matlab

MATLAB Toolbox for Power Magnetics: Model and Optimization

MATLAB Toolbox for Power Magnetics: Model and Optimization

This MATLAB toolbox allows for the modeling and optimization of power magnetic components:

• medium-frequency inductors
• medium-frequency transformers
• computation of the mass and volume
• extraction of the equivalent circuit
• computation of the core and winding losses
• fast and accurate semi-numerical methods
• plotting of the winding and core geometries
• brute-force optimization (parallel code)
• flexible object-oriented design

The following methods/functionalities are provided for the core modeling:

• iGSE for the core losses (with locally fitted parameters from a loss map)
• linear reluctance solver with 3D air gap models
• multiple air gaps are allowed
• multiphases components are allowed

The following methods/functionalities are provided for the winding modeling:

• mirroring method with inductance matrix and field evaluation (with/without air gaps)
• solid wire windings (including skin and proximity losses)
• stranded (Litz) wire windings (including skin and proximity losses)
• multiple air gaps are allowed
• multiphases components are allowed
• model of the winding heads

Currently, the following components are implemented:

• inductors and two-winding transformers with shell-type windings
• U-core, C-core, and E-core

However, additional components can be added by implementing abtract classes. More specifically, the code could handle the following cases (without modifying the core classes):

• multiphase components (transformers or chokes)
• other winding geometries (core-type, matrix, etc.)
• other core geometries (ELP, RM, etc.)
• distributed airgaps

Getting Started

Two DC-DC converters are considered as examples:

• a resonant converter (SRC-DCX) with a MF transformer
• a bidirectional Buck converter (Buck DC-DC) with a MF inductor

Both converteres are operating between 400V and 100V buses with a rated power of 5kW. The component geometry (core and windings) and the operating frequency are optimized.

The example consists of the following files:

• resonant converter (SRC-DCX) with a MF transformer
  • run_src_dcx_1_single.m - modelization of a single design
  • run_src_dcx_2_combine.m - brute-force optimization of the component
  • run_src_dcx_3_plot.m - optimization results (Pareo fronts
• bidirectional Buck converter (Buck DC-DC) with a MF inductor
  • run_buck_dcdc_1_single.m - modelization of a single design
  • run_buck_dcdc_2_combine.m - brute-force optimization of the component
  • run_buck_dcdc_3_plot.m - optimization results (Pareo fronts)
• example_files - definition of the parameters, materials, and waveforms

Gallery

Buck DC-DC Inductor

SRC-DCX Transformer

Pareto Fronts

Toolbox Organization

The power magnetic toolbox contains the following packages:

• add_path_mag_tb.m - add the toolbox to the MATLAB path
• core - core reluctance and core losses
  • core/README.txt - package documentation
  • core/DATA_STRUCT.txt - data format documentation
  • core/core_class.m - main class
  • core/core_lib - package internal classes
  • core/core_example - example/test files
• window - winding window stray field and winding losses
  • window/README.txt - package documentation
  • window/DATA_STRUCT.txt - data format documentation
  • window/window_class.m - main class
  • window/window_lib - package internal classes
  • window/window_example - example/test files
• component - simulation of complete components (inductor or transformer)
  • component/README.txt - package documentation
  • component/DATA_STRUCT.txt - data format documentation
  • component/component_class.m - main class
  • component/component_lib - package internal classes
  • component/component_example - example/test files
• sweep - simulation of design sweeps (brute-force optimization)
  • sweep/README.txt - package documentation
  • sweep/get_sweep_single.m - simulating a single parameter combination
  • sweep/get_sweep_combine.m - simulating a many parameter combinations
  • sweep/sweep_lib - package internal functions
  • sweep/sweep_example - example/test files

Compatibility

• Tested with MATLAB R2015b and R2021a.
• Parallel Computing Toolbox.
• Compatibility with GNU Octave not tested but probably problematic.

Author

Thomas Guillod - GitHub Profile

This toolbox shares some files/ideas with the following repositories:

• mirroring_method_matlab - ETH Zurich, Power Electronic Systems Laboratory, T. Guillod, BSD License
• litz_wire_losses_fem_matlab - ETH Zurich, Power Electronic Systems Laboratory, T. Guillod, BSD License

License

This project is licensed under the BSD License, see LICENSE.md.