Speaker
Description
During the acceptance phase of the China Fusion Engineering Testing Reactor (CFETR) Central Solenoid Model Coil (CSMC), Paschen testing was conducted to verify the reliability of the insulation structure. To locate insulation faults during magnet testing, this study developed an Insulation Condition Monitoring System (ICMS) based on visible light imaging technology. The ICMS software, developed using C++ and OpenCV, integrates camera control, image acquisition, discharge image detection, and fault area localization. To address large-scale image data processing, the system employs an improved convolutional neural network architecture with attention mechanisms and combines morphological feature analysis of discharge images to design an intelligent discharge region detection algorithm, enhancing detection efficiency. During CSMC Paschen testing, 21 cameras were deployed for real-time monitoring of insulation weak points in the superconducting magnet system. The ICMS successfully captured 4 discharge flash regions. After implementing insulation repair measures, the CSMC passed the Paschen test. The results confirm that the ICMS effectively monitors insulation conditions and locates breakdown points during CSMC testing, thus robustly ensuring the safe and stable operation of the magnet system.
