Skip to content

SC 002 - Twisted superconductor AC Loss

In this tutorial, AC loss in a high-temperature superconducting (HTS) wire is simulated using the H-φ formulation.

The wire consists of twisted superconducting filaments embedded into a copper matrix. The whole modelling domain with an air cylinder around the wire is illustrated below.

Tutorial image

Element Dimension
Air cylinder diameter 10 mm
Copper cylinder diameter 535 μm
Filament diameter 350 μm
Domain length 10 mm
  • Joule losses as a function of time in the copper and the superconducting filament regions. The losses over the volume of interest can be computed as
P(t)=E(t)J(t) dV.P(t)=\int\boldsymbol{E}(t)\cdot \boldsymbol{J}(t)~\rm{d} V.

Magnetic permeability, μ\mu:

  • all domains: μ0\mu_0

Electric resistivity, ρ\rho:

  • Copper: 1010 Ωm10^{-10} ~ \Omega \rm m
  • Superconducting filaments:
    • ρ=EcJc(JJc)n1\rho = \frac{E_{\rm c}}{J_{\rm c}} \left( \frac{||J||}{J_{\rm c}} \right) ^{n-1}
      • Critical electric field strength, Ec=100 μV/mE_{\rm c} = 100 ~ \rm μV/m
      • Exponent, n=30n = 30
      • Critical electric current, Ic=100 AI_{\rm c} = 100 ~ \rm A
      • Total cross-section are of superconducting filaments, Asc=3.4541107 m2A_{\rm sc} = 3.4541 \cdot 10^{-7} ~ \rm m^2
      • Critical electric current density, Jc=Ic/AscJ_{\rm c} = I_{\rm c} / A_{\rm sc}

The problem is sourced by applying a total current of

I(t)=0.8 Ic sin(2πft),I(t) = 0.8 ~ I_{\rm c} ~ \sin(2 \pi f t),

where the frequency ff is 50 Hz50 ~ \rm Hz.

Here you’ll find a detailed step-by-step tutorial on how to simulate AC loss in a twisted filament HTS wire Quanscient Allsolve.

  1. Start with a new project and name it as

    SC twisted filament AC loss
  2. Import the geometry as a .step file with default import options.

    File download link: twisted-superconductor.step

    Tutorial image

  3. Confirm model changes.

  1. Go to the Common sidebar.

  2. Define a region for air:

    Region name Region type Target
    air Volume Air cylinder

    Tutorial image

  3. Define a region for copper:

    Region name Region type Target
    copper Volume Copper matrix

    Tutorial image

  4. Define a region for the superconducting filaments:

    Region name Region type Target
    sc Volume SC filaments

    Tutorial image

  1. Assign the Air material to the air region:

    Tutorial image

  2. Assign the Copper material to the copper region:

    Tutorial image

  3. Assign the YBCO superconductor material to the sc region:

    Tutorial image

  1. Define new variables:

    Name Description Expression
    f Frequency [Hz] 50
    YBCO_Ic Critical current [A] 100
    YBCO_Asc Filament cross-section area [m^2] 3.4541e-7
    Iop Operating current [A] 0.8 * YBCO_Ic * sin(2 * pi * f * t)
  2. Modify existing variables:

    Name Modified expression
    YBCO_Jc YBCO_Ic / YBCO_Asc
    YBCO_n 30

Step 5 - Define physics and apply the current source

Section titled “Step 5 - Define physics and apply the current source”
  1. Go to the Physics section.

  2. Add the Magnetism H physics:

    Physics Target
    Magnetism H Copper matrix and SC filaments

    Tutorial image

  3. Add the Magnetism φ physics:

    Physics Target
    Magnetism φ Air cylinder (air region)

    Tutorial image

  4. Add a Lump I/V cut interaction to Magnetism φ.

    Interaction name Interaction type Target Value
    Current source Lump I/V cut a counter-clockwise loop at the top edge of the copper matrix Iop

    Tutorial image

  5. Add a Constraint interaction to Magnetism φ:

    Interaction name Interaction type Target Value
    Gauge Constraint a point at the external boundary of the air domain 0

    Any point on the external boundary of the air domain can be selected. This constraint is the gauge condition for the scalar potential φ.

    Tutorial image

  6. Add the H-φ coupling interaction to Magnetism H.

  7. Before moving on, check that your physics tree matches the one below:

    Tutorial image

  1. Go to the Simulations section.

  2. Add a new mesh.

  3. Set Autorefine to Disabled.

  4. Open the collapsible menu for Mesh element size.

  5. Set Curvature enhancement to 25.

  6. Generate the mesh and check the preview.

    Tutorial image

    Tutorial image

  1. Add a new simulation.

  2. Set Analysis type to Transient.

  3. Select Transient settings:

    Timestep algorithm Start time [s] End time [s] Timestep size [s]
    Implicit Euler 0 0.01 0.0001
  4. Set Solver mode to Iterative solver.

  5. Set Node count to 50.

  6. Select Mesh 1 as the mesh for your simulation.

  7. Define custom value outputs for computing Joule losses in the filaments and copper:

    Output name Output type Output expression
    SC loss Custom value output integrate(reg.sc, transpose(E) * j, 4)
    Cu loss Custom value output integrate(reg.copper, transpose(E) * j, 4)

    In the expressions above, E and j are 3 x 1 column vectors. transpose(E) creates a 1 x 3 row vector from E so that the multiplication with j produces a dot product EJ\boldsymbol{E} \cdot \boldsymbol{J}. The last argument 4 is the integration order.

  8. Open the Script for your simulation.

  9. Enable Scripting mode.

  10. Replace the first autogenerated line under # Magnetism H with the following Newton-linearization [4]:

    rho = 1/par.sigma(df.j)
    dedj = rho*qs.eye(3) + (expr.YBCO_n-1.0)*rho/qs.max(df.j*df.j, 1e-40) * df.j * qs.transpose(df.j)
    dofe = rho*df.j + dedj * (qs.curl(qs.dof(fld.H))+var.curl_dof_Hs - qs.curl(fld.H)-var.curl_Hs)
    form += qs.integral(reg.sc, dofe * (qs.curl(qs.tf(fld.H)) - var.curl_tf_Hs))
    form += qs.integral(reg.copper, qs.inverse(par.sigma(df.j)) * (qs.curl(qs.dof(fld.H)) + var.curl_dof_Hs) * (qs.curl(qs.tf(fld.H)) - var.curl_tf_Hs))

    Tutorial image

Step 8 - Run the simulation and see results

Section titled “Step 8 - Run the simulation and see results”
  1. Run the simulation.

  2. To follow the simulation progress, open Logs.

  3. The SC and Cu loss results can be seen in Plotting, even while the simulation is running:

    Tutorial image

[1] H-φ Formulation in Sparselizard Combined With Domain Decomposition Methods for Modeling Superconducting Tapes, Stacks, and Twisted Wires. https://doi.org/10.1109/TASC.2023.3240389

[2] Allsolve demo project of Superconductor AC losses. https://allsolve.quanscient.com/#/projects/demo/8fed82d1-5bf0-4c02-835b-94e65a60f847

[3] Youtube tutorial of Superconductor AC losses. https://youtu.be/B9QZZ5y7RpQ

[4] Newton Linearization. https://en.wikiversity.org/wiki/Nonlinear_finite_elements/Newton_method_for_finite_elements