Laminated Material¶
The Lamination module ships two parallel material classes — one per magnetic model — that share the same parameter set and the same underlying physics. See the overview for the laminate geometry, the two eddy-current channels and the relevant references.
| Magnetic model | Material class |
|---|---|
| Time-Domain Magnetic | TimeDomainMagneticLaminationMaterial |
| Time-Harmonic Magnetic | TimeHarmonicMagneticLaminationMaterial |
The two classes implement the same constitutive law. They differ only in how the in-sheet eddy-current correction is wired into the formulation. The time-domain class uses a separate weak form; the frequency-domain class uses an imaginary part of the complex reluctivity.
Effective constitutive tensors¶
In the local frame both tensors are diagonal with two equal in-plane components and a distinct out-of-plane component.
Magnetic permeability.
The in-plane direction is an arithmetic mean (continuity of \(H_\parallel\) across iron/air interfaces); the out-of-plane direction is a harmonic mean (continuity of \(B_\perp\)):
So the model uses this permeability tensor:
Electric conductivity.
Electric current cannot flow across the insulating interlayers, so \(\sigma_\perp = 0\). The volume fraction of conducting material reduces the in-plane conductivity:
This \(\bar{\boldsymbol{\sigma}}\) accounts only for the homogenised \(\vec{j}_{\alpha\beta}\) — the macroscopic eddy-current loops in the lamination plane.
In-sheet eddy-current correction¶
The component \(\vec{j}_{\beta\gamma}\) — the in-sheet loop driven by the parallel flux variation \(\partial_t \vec{B}_\alpha\) — has a known analytical structure1. It comes from the 1-D Faraday equation across the sheet thickness. In the thin-sheet limit the in-sheet current density varies linearly with the through-thickness coordinate. The correction augments the magnetic-field strength2: $$ \vec{H} = \bar{\boldsymbol{\mu}}^{-1}\,\vec{B} + \underbrace{\frac{\sigma\,d^2}{12}\,\partial_t \vec{B}}{\displaystyle \vec{H}\text{LS}}, $$
where \(\vec{H}_\text{LS}\) is the lamination-stack augmentation that accounts for the in-sheet eddy currents.
Validity range¶
The first-order augmentation \(\vec{H}_\text{LS}\) neglects the in-sheet skin effect. So it is valid when the in-sheet skin depth \(\delta_{\alpha\beta} = \sqrt{2 / (\omega\,\mu_\perp\,\sigma_\parallel)}\) exceeds about half the sheet thickness:
For 0.5 mm silicon-steel (\(\sigma = 5 \times 10^6\,\mathrm{S/m}\), \(\mu_r = 1000\)) at 50 Hz, the in-sheet skin depth is \(\delta_{\alpha\beta} \approx 1.0\,\mathrm{mm}\). The ratio is then \(\sim 0.5\) — well inside the validity range. For thinner sheets or lower frequencies the approximation is better; for thicker sheets or frequencies above a few kHz at these material parameters the linear \(\gamma\)-expansion breaks down.
A second length scale, the penetration depth into the stack \(\delta_\gamma = \sqrt{2 / (\omega\,\mu_\parallel\,\sigma_\parallel)}\), governs the mesh size near the stack ends. Because \(\mu_\parallel \gg \mu_\perp\) for typical iron laminates, \(\delta_\gamma \ll \delta_{\alpha\beta}\). The perpendicular flux decays much faster from the stack ends than the in-plane skin depth would suggest. You must resolve the boundary layer at the stack ends accordingly.
Parameters¶
| Argument | Units | Description |
|---|---|---|
name |
– | Material name. |
marker |
– | Volume marker of the laminated region. |
sheet_permeability |
N/A² | Magnetic permeability \(\mu\) of a single ferromagnetic sheet. May depend on \(\vert\mathbf{B}\vert\) for nonlinear iron. |
sheet_conductivity |
S/m | Electric conductivity \(\sigma\) of a single ferromagnetic sheet. |
stacking_direction |
– | Unit vector \(\vec{e}_\gamma\) normal to the sheets (the lamination plane is perpendicular to it). |
stacking_factor |
– | Volume fraction \(F \in (0,\,1]\) of the magnetic phase. \(F = 1\) recovers a pure-iron stack with no insulating interlayer. |
lamination_thickness |
m | Thickness \(d\) of a single sheet (used only by the in-sheet eddy-current correction). |
has_eddy_currents |
– | If True (default), the homogenised eddy-current channel \(\bar{\boldsymbol{\sigma}}\) is added to the formulation. |
with_homogenization_correction |
– | If True (default), the in-sheet eddy-current correction \(\vec{H}_\text{LS} = (\sigma d^2 / 12)\,\partial_t \vec{B}\) is wired in. Disable to study the bare anisotropic constitutive law in isolation. |
All scalar / vector parameters accept the standard mufem coefficient
inputs (CffConstantScalar, CffConstantVector, field-dependent
coefficients, …).
Time-Domain Magnetic¶
import mufem
from mufem import Vol
from mufem.electromagnetics.module.lamination import (
TimeDomainMagneticLaminationMaterial,
)
iron_material = TimeDomainMagneticLaminationMaterial(
name="Core",
marker="Core" @ Vol,
sheet_permeability=mufem.CffConstantScalar(1000.0),
sheet_conductivity=mufem.CffConstantScalar(5.0e6),
stacking_direction=mufem.CffConstantVector(0.0, 0.0, 1.0),
stacking_factor=mufem.CffConstantScalar(0.95),
lamination_thickness=mufem.CffConstantScalar(0.5e-3),
)
time_domain_magnetic_model.add_materials([iron_material])
Time-Harmonic Magnetic¶
import mufem
from mufem import Vol
from mufem.electromagnetics.module.lamination import (
TimeHarmonicMagneticLaminationMaterial,
)
iron_material = TimeHarmonicMagneticLaminationMaterial(
name="Core",
marker="Core" @ Vol,
sheet_permeability=mufem.CffConstantScalar(1000.0),
sheet_conductivity=mufem.CffConstantScalar(5.0e6),
stacking_direction=mufem.CffConstantVector(0.0, 0.0, 1.0),
stacking_factor=mufem.CffConstantScalar(0.95),
lamination_thickness=mufem.CffConstantScalar(0.5e-3),
)
time_harmonic_magnetic_model.add_materials([iron_material])
References¶
-
P. Dular, J. Gyselinck, C. Geuzaine, N. Sadowski and J. P. A. Bastos, A 3-D Magnetic Vector Potential Formulation Taking Eddy Currents in Lamination Stacks Into Account, IEEE Trans. Magn., 39(3), 1424–1427 (May 2003). doi:10.1109/TMAG.2003.810386. ↩
-
H. De Gersem, S. Vanaverbeke and G. Samaey, Three-Dimensional–Two-Dimensional Coupled Model for Eddy Currents in Laminated Iron Cores, IEEE Trans. Magn., 48(2), 815–818 (February 2012). doi:10.1109/TMAG.2011.2172924. ↩