The Magnetic Skeleton

Abstract

The magnet system gives a tokamak its plasma geometry, but it also determines structural force, stored energy, cryogenic demand, access, assembly sequence, and much of the plant's capital mass. I develop the magnetic skeleton as one integrated system: toroidal-field coils confine particles, poloidal-field coils shape and position the plasma, a central solenoid supports startup, correction coils manage error fields, structures close the Lorentz-force paths, and cryogenics preserve superconductor margin. I compare low-temperature and high-temperature superconductors, estimate magnetic pressure and stored energy, and define quench protection as a plant safety function. The design objective is not peak field alone. It is useful field at the plasma after shielding, structure, joints, tolerances, radiation, and maintainability are included.

Field is purchased with structure

A charged particle with perpendicular momentum $p_\perp$ follows a Larmor orbit whose radius is

$$r_L=\frac{p_\perp}{|q|B}.$$

Stronger field reduces orbit size and generally supports higher pressure or a more compact plasma. This is why high field is so attractive. The magnet does not experience field as an abstract parameter. It experiences force.

Magnetic energy density and pressure scale as

$$u_B=p_B=\frac{B^2}{2\mu_0}.$$

At $12\ \mathrm{T}$, $p_B$ is about $57\ \mathrm{MPa}$. Local stresses in winding packs and cases can be much higher because geometry concentrates loads. Doubling field quadruples magnetic pressure. A compact high-field plant may save blanket circumference while demanding stronger conductor, support, and protection.

The coil family performs different work

The toroidal-field coils create the dominant field around the torus. Their in-plane forces tend to straighten the curved coils, while out-of-plane loads emerge from interactions with other coils and plasma transients. Cases, wedges, shear keys, gravity supports, and the surrounding structure must provide a continuous load path.

Poloidal-field coils shape the plasma and control its radial and vertical position. The central solenoid acts as a transformer primary during plasma-current initiation and ramp. Fast control coils correct instabilities; smaller correction coils compensate manufacturing and assembly errors.

ITER’s system illustrates the scale of integration: eighteen toroidal-field coils, a central solenoid, poloidal-field coils, correction coils, feeders, current leads, cryogenic distribution, and power conversion all operate as one magnetic machine. A commercial plant may choose different field and current strategies, but it cannot remove the system relationships.

Coil function	Timescale	Design consequence
Toroidal confinement	Steady or long pulse	High stored energy and continuous structural load
Plasma shaping	Seconds to hours	Large controlled currents and coupled forces
Central induction	Startup and pulse evolution	Fatigue, volt-seconds, and central-space competition
Vertical stabilization	Milliseconds	Fast sensors, supplies, and conducting structures
Error-field correction	Commissioning and operation	Tight metrology and distributed trim authority

Superconductor choice changes the plant

Low-temperature superconductors such as niobium-titanium and niobium-tin have extensive fusion and accelerator experience. They operate near liquid-helium temperature and have known cabling and qualification methods. Niobium-tin supports higher field than niobium-titanium but becomes brittle after reaction heat treatment.

Rare-earth barium copper oxide and other high-temperature superconductors can carry current at higher field and temperature. Their name can mislead: a fusion magnet may still operate at cryogenic temperature to gain current density and margin. Higher allowable temperature can reduce refrigeration penalty and provide more thermal headroom, while tape geometry, joints, quench propagation, mechanical strain, manufacturing scale, and radiation response introduce new constraints.

I will not select a conductor by its short-sample critical current. I need an operating surface

$$I_{\mathrm{op}}<I_c(B,T,\varepsilon,D),$$

where $\varepsilon$ is strain and $D$ represents accumulated radiation damage. Margin must include spatial nonuniformity, joints, disturbances, fabrication variation, sensor uncertainty, and credible faults.

Cryogenic margin is a heat-flow problem

A magnet remains superconducting only while local heat generation and incoming heat stay below what conduction and coolant can remove. Sources include nuclear heating, alternating-current losses, joint resistance, current leads, supports, instrumentation wires, thermal radiation, eddy currents, and mechanical motion.

The refrigeration wall-plug demand greatly exceeds the heat removed at cryogenic temperature. A simple thermodynamic lower bound is the reversed Carnot coefficient of performance:

$$\mathrm{COP}_{\mathrm{ideal}}=\frac{T_c}{T_h-T_c}.$$

At $T_c=4.5\ \mathrm{K}$ and $T_h=300\ \mathrm{K}$, the ideal value is about $0.015$. Real systems achieve only a fraction of the ideal. One watt intercepted at room temperature is therefore not equivalent to one watt deposited in a cold winding.

Thermal shields at intermediate temperature, multilayer insulation, low-conductivity supports, optimized leads, and neutron shielding protect both current margin and net plant power. ITER uses helium systems near $4\ \mathrm{K}$ for magnets and cryopumps, demonstrating the industrial scale of this support plant.

The neutron shield protects an irreplaceable asset

The breeding blanket and shield sit between plasma and magnets. They absorb neutrons, reduce nuclear heating, and limit displacement and dose in conductor and insulation. Increasing shield thickness protects magnet life but enlarges the machine because every radial centimeter repeats around the torus.

This creates a radial build negotiation:

$$R_{\mathrm{coil}}=a+\Delta_{\mathrm{scrape}}+t_{\mathrm{first\ wall}}+t_{\mathrm{blanket}}+t_{\mathrm{shield}}+t_{\mathrm{vessel}}+t_{\mathrm{gap}}.$$

Here $a$ is plasma minor radius. Compactness depends on all layers, not only field strength. A credible high-field design must show that blanket breeding, shielding, vessel structure, cooling, and maintenance clearances still fit.

A quench is an energy-management event

When a region becomes resistive, local Joule heating can damage conductor before the normal zone spreads far enough to produce an obvious voltage. High-temperature-superconductor magnets can be especially difficult because normal-zone propagation may be slow.

Stored magnetic energy is approximately

$$E_m=\frac{1}{2}LI^2.$$

For $L=10\ \mathrm{H}$ and $I=60\ \mathrm{kA}$, $E_m=18\ \mathrm{GJ}$. That energy must remain magnetically stored, transfer into other coupled coils, or dissipate as heat and voltage. Detection, energy extraction, dump resistors, quench heaters or alternative spreading systems, insulation, grounding, and structural response belong to one protection case.

The protection architecture uses diverse measurements: differential voltage, temperature, acoustic events, fiber sensing, current, field, strain, helium behavior, and model residuals. It must reject inductive voltage during routine ramps while detecting a small resistive signal quickly.

ConductorCarry current with margin

Control field, temperature, strain, joints, and radiation over life.

StructureClose every force path

Transfer steady, cyclic, gravity, seismic, and fault loads predictably.

CryogenicsRemove heat at the right level

Intercept room-temperature loads before they reach cold mass.

ProtectionDetect and dissipate abnormal energy

Limit conductor temperature, voltage, pressure, and coupled force.

Field quality begins in the factory

Small geometric errors produce field errors that can seed locked modes or reduce confinement. Each conductor length, winding pack, case, splice, support, and assembled coil needs a metrology record tied to the plasma model. Correction coils provide finite authority; they should not become permission for uncontrolled manufacture.

The plant needs magnetic commissioning before nuclear operation. Individual circuits undergo insulation, pressure, flow, sensor, and low-current checks. Selected assemblies receive cold and full-current tests where practical. Integrated energization proceeds through bounded plateaus while field maps, strain, motion, cryogenic load, and protection channels are compared with predictions.

Maintainability reaches inward

Large superconducting coils may be lifetime components, yet feeders, valves, sensors, cryogenic equipment, power converters, switches, and joints require access. The machine arrangement should permit their isolation and replacement without disturbing the activated vacuum vessel.

QualifyConductor, joints, and insulation AssembleMeasure geometry and preload EnergizeValidate field, force, and protection OperateTrend margin through plant life

I regard the magnet system as a permanent capital asset protected by replaceable inner systems. That choice makes shielding, vacuum cleanliness, and assembly access part of magnet design. It also leads directly to the next subject: the vacuum environments that separate plasma, cryogenic mass, atmosphere, and fuel processing.