Computing Library › 3D Model & Digital Twin
3D Model & Digital Twin

Plasma State Reconstruction

Plasma state reconstruction infers the internal condition of a confined plasma from external diagnostics through a physics model.

Seeing inside the plasma

The interior of a confined fusion plasma is largely inaccessible to direct measurement; the most important quantities, the current, pressure, and magnetic-field profiles, cannot be sampled point by point. Plasma state reconstruction infers these internal fields from external and line-integrated diagnostics, constrained by the equations that govern the plasma. It is the fusion-specific heart of a plasma twin.

The reconstruction problem

Kronos motion — plasma state

The task is an inverse problem: find the internal state that is consistent with both the measurements and the physics. It is typically ill-posed, meaning many internal states could roughly fit the data, so the physics constraints and prior information are essential to pick the physically valid one. This is why reconstruction is done through a model rather than by direct inversion.

Diagnostics used

From static to dynamic

Traditional reconstruction solves for a snapshot. A twin turns this into a continuous, dynamic estimate by embedding reconstruction inside a data-assimilation loop, so the plasma state is tracked in time with uncertainty, and can be predicted a short way ahead for control. This is the bridge from classical equilibrium fitting to a live plasma twin. See equilibrium reconstruction.

In the Kronos machines

For the Hyperion breeder, a spherical tokamak with negative triangularity of -0.30 and a plasma current of 9.86 mega-amperes, reconstruction supplies the shape and internal profiles that control depends on, since no diagnostic measures the internal current directly. For the burner, a tandem-mirror configuration, the reconstructed state describes plug and throat conditions. Both are developed today on design models and validated against related experiments, pending real data after first tritium near 2030. See disruption prediction.