Parameterize a Magnetic Core Block Using B-H Curve Data

Since R2024b

This example shows how to parameterize a magnetic core block using B-H curve data. This block represents a magnetic core with magnetic hysteresis implementing a Jiles-Atherton model.

Model Overview

Open the parameterizeIronCoreWithBHCurve model.

myModel = 'parameterizeIronCoreWithBHCurve';
open_system(myModel);

In this example, you model a core that an electromagnetic converter magnetizes. To parameterize the model with B-H curve data, you set the Parameterization option parameter of the Magnetic Core block to Saturated BH Curve. The Magnetic flux density, B(H) and Magnetic field intensity, H parameters represent the ascendent or descendent half of the B-H curve.

Import B-H curve

Select a saturated B-H curve chart and use the graphImporter to capture either the ascending or descending curve. This example uses Figure 5 of a research paper.

Open the graph importer and open the session parameterizeIronCoreWithBHCurveGraphImporterSession.mat.

Click Export and save the data with the name BHCurve.mat. Load the mat file assigning the data to the corresponding parameters in the workspace.

load parameterizeIronCoreWithBHCurve.mat
Hdescendent = tableData(:,1);
Bdescendent = tableData(:,2);

Alternatively, you can follow the examples from Graph Importer to extract the B-H curve data from parameterizeIronCoreWithBHCurveFig5BH.png.

The journal article does not specify values for the Effective length and Effective cross-sectional area parameters of the Magnetic Core block. Use the default values. If these values are incompatible with your B-H curves, you can change them later.

You can estimate an appropriate value for the Magnetic flux density at saturation point using the curve you imported earlier. Plot the data to visualize the saturation point.

parameterizeIronCoreWithBHCurvePlotImportedCurve

The maximum value of the imported curve is 1.38 T. However, the curve has not yet reached a plateau, so the actual flux density at the saturation point is higher than this value. Set the Magnetic flux density at saturation point parameter to 7.5% above the maximum value.

Bsaturation = max(Bdescendent)*1.075;

To define the input signal to the Controlled Voltage Source block, choose a sensible value for the Frequency (rad/sec) parameter of the Sine Wave block. In this example, you do not model eddy current losses, so you can choose any value that the solver can handle.

f = 1e3;

Choose a value for the Amplitude parameter of the Sine Wave block such that the electrical network provides sufficient voltage to saturate the B-H curve, but it does not saturate it excessively. The resistance is equal to 1$\Omega$, so the voltage is equal to the current. From the previous figure, you can estimate that the saturation value for the magnetic field intensity is around $$500A/m$$. Set the Amplitude parameter of the Sine Wave block to 500 multiplied by the value of the Effective length parameter of the Magnetic Core block.

mmfSat = 500*meanMagneticLength;

Plot Simulation Results

Simulate the model and plot the simulated and the input B-H curves.

parameterizeIronCoreWithBHCurvePlotCurve

The simulated curve overlaps well with the input data.

Investigate Sensitivity to Magnetic Flux Density at Saturation Point Parameter

To see how important it is to parameterize your model using an appropriate saturation point, set the Magnetic flux density at saturation point parameter to the maximum value from the input data plot the B-H curve.

Bsaturation = max(Bdescendent);
parameterizeIronCoreWithBHCurvePlotCurve

The simulated B-H curve no longer overlaps with the input data at high flux density magnitudes.

Set the Magnetic flux density at saturation point parameter back to 7.5% above the maximum value from the input data.

Bsaturation = max(Bdescendent)*1.075;

References

Sadowski, N., N.J. Batistela, J.P.A. Bastos, and M. Lajoie-Mazenc. “An Inverse Jiles-Atherton Model to Take into Account Hysteresis in Time-Stepping Finite-Element Calculations.” IEEE Transactions on Magnetics 38, no. 2 (March 2002): 797–800. https://doi.org/10.1109/20.996206.