Abstract
The Experimental Advanced Superconducting Tokamak (EAST) has achieved thousand-second plasma discharges, placing unprecedented demands on its Magnetic Diagnostic System (MDS). Conventional integrators, limited by saturation, excessive drift (>1% Full-Scale Range, FSR), and nonlinearity, fall short of the requirements for prolonged discharges. This paper presents a comprehensively optimized alternating integration system designed to overcome these limitations. Through rigorous numerically-driven analysis and quantification of error sources, tailored compensation strategies were derived and key operational parameters optimized, promoting error convergence and ensuring precision consistency for distributed deployment. Bench tests over 4000 seconds show that 99.04% of integration drift samples remain within 0.1% FSR, while plasma experiments confirm a maximum drift of only 9 mV during the 1,066-second discharge. The system operates autonomously, performing self-correction of drift induced by sensor load variations and self-calibration of drift rates during alternation cycles, thereby eliminating the need for manual intervention. Excellent reproducibility has been confirmed through repeated testing. This system delivers the precision performance surpassing the all-digital integrator developed for the International Thermonuclear Experimental Reactor (ITER) at a fraction of the cost. Suitable for deployment across EAST’s full magnetic diagnostic array of over 600 sensors, it not only meets the stringent requirements for long-pulse operation but also establishes a reliable foundation for the upcoming Burning Plasma Experiment Superconducting Tokamak (BEST) program.
Funding Information
  • National Key Research and Development Program of China (2023YFF0719700)
  • National Magnetic Confinement Fusion Program of China (2015GB103000)
  • Key Research and Development Plan Project of Anhui Province (202304a05020012)
  • Anhui Province Natural Science Foundation (2408085ME145)

This publication has 29 references indexed in Scilit: