Development of a Monitoring System for Forced-Flow-Cooled Superconducting Coils With Principal Component Analysis

To enhance the efficiency of system operation and reduce the manpower required for monitoring a superconducting magnet system in nuclear fusion experiments, we have incorporated principal component analysis (PCA) into the monitoring system. This monitoring system uses PCA to transform high-dimensional data, comprising temperature, pressure, and mass flow rate of Large Helical Device (LHD) poloidal coils into lower-dimensional data for efficient coil status monitoring. The LHD poloidal coils are forced-flow-cooled superconducting coils wound with cable-in-conduit conductors. In the system with PCA, the coil's state is assessed visually and qualitatively. Consequently, the monitoring with PCA of a superconducting magnet system demonstrates practical applicability.