Ferromagnetic Modelling for In-Vessel Components

[DanShort12]

In GitLab by @NathanUKAEA on Jul 26, 2021, 11:41

Some in-vessel components, such as the breeder blankets, may be made of magnetically permeable materials which could have a sizable effect on the plasma as they're in close proximity and similarly may effect the performance of nearby coils.

It's likely that the material will be magnetically saturated, so the magnitude of the magnetisation is set, but the direction of the magnetisation will be given by the sum-total of all fields present from the plasma and coils and its own field. This requires a self-consistent non-linear solver.

This is quite a development, and for the non-FEM demagnetisation tensor approach would require the following:

• Discretization of the material into elements of constant magnetisation (in a Cartesian sense, fixed radial magnetisation is still varying)
• Create demagnetisation tensor from analytic formula
• Calculate all fields inside material from external sources
• Non-linear solver to find self magnetisation
• Calculate stray field outside material from analytic formula to analyse effect

The stray field outside of a toroidally (near) continuous magnet is very small (micro-tesla) but, numerically speaking, this is a case of Catastrophic Cancellation, in that any one element actually creates a large field but the sum-total is near zero. This means that accuracy in the geometry may be vital - in the simulation, design and real-world construction. So a sensitivity analysis may be very useful.

[DanShort12]

In GitLab by @NathanUKAEA on Jul 26, 2021, 11:44

changed the description

[DanShort12]

In GitLab by @sebkahn on Jul 26, 2021, 15:09

Thanks for that issue @NathanUKAEA