7th International Conference on Electric Power Equipment– Switching Technology (ICEPE 2024)

98 / 2024-04-12 09:47:50

Intelligent Breakdown Diagnosis System for Vacuum Interrupter in Conditioning Based on Zynq SoC and Deep Learning

breakdown diagnosis, vacuum interrupter, Zynq, deep learning, conditioning

Breakdown diagnosis in conditioning had a great significance for vacuum interrupter (VI) insulation improvement. This paper proposed an intelligent breakdown diagnosis system in conditioning based on Zynq System-on-Chip (SoC) and deep learning to diagnose breakdown accurately and efficiently. The deep learning model was constructed and trained on PC to diagnose breakdown through pre-breakdown current waveforms, and the weights parameters of deep learning model for breakdown diagnosis were acquired. The intelligent breakdown diagnosis system was implemented on Zynq SoC with acquired weights parameters through software-hardware co-design. It integrated the image capture module and breakdown mechanism classification module, which could diagnose breakdown automatically without manual operation and extern PC devices. The results indicated that the real-time breakdown mechanism could be obtained accurately and efficiently with the intelligent breakdown diagnosis system, whose accuracy was higher than 85%. The intelligent breakdown diagnosis system could have a promising application in engineering for VI insulation improvement.