Development of AI (Deep Learning)-based transformer failure diagnosis device that enables fault detection inside the columnar transformer using AI (Deep Learning) for the first and second current waveforms of the columnar transformer. ①Current waveform detection to transformer primary and secondary, ②Real-time transformer current pattern image conversion, ③ It consists of failure pattern analysis/prediction through AI analysis and will be used for fault diagnosis/prediction of transformers at each HANJEON office. It is a diagnostic/prediction system that converts measured transformer primary and secondary data into real-time failure pattern algorithms and applies image processing technology in AI (Deep Learning) engines to analyze/predict transformer failure patterns to provide real-time transformer status information to users