Plasma-behavior-informed noise compensation and transient disturbance shielding quench detection in superconducting tokamaks

Abstract Reliable quench detection is essential for the safe operation of fully superconducting tokamaks. Conventional resistive voltage-based methods offer rapid response but are strongly affected by electromagnetic interference during plasma discharges, limiting their accuracy. A dynamic noise-compensated quench detection method has been developed and applied to EAST. The approach combines offline modeling with real-time inversion to update compensation coefficients according to plasma parameters, enabling effective suppression of plasma-induced coupling voltages. Additionally, a transient disturbance shielding scheme is incorporated to identify and filter short-lived perturbations arising from breakdowns, disruptions, and vertical displacement events. Experiments on EAST demonstrate that residual noise can be reduced to within ±50 mV for poloidal field coils and ±10 mV for toroidal field coils, corresponding to a signal-to-noise ratio greater than 10. A simulated quench test confirms accurate detection with a total delay of 1.74 s. The results indicate that dynamic noise-compensated quench detection can provide high-fidelity performance under plasma discharge conditions, offering a pathway for ITER and future fusion devices.