With the rapid development of digital technologies such as IoT, AI, and big data, electrical energy consumption is rapidly increasing. Electrical facilities that supply electrical energy are operated with high reliability and stability for end-of-life time. In addition, depending on the type of electrical load that consumes electrical energy in various forms, electrical insulation systems deteriorate due to electrical and thermal stress, which reduces electrical and mechanical insulation strength. Due to such continuous stress and electrical transient phenomena, electrical facilities may experience electrical accidents due to electrical insulation breakdown before the expected design lifetime. In addition, periodic inspections according to related regulations must be conducted to prevent unexpected electrical accidents, but this leads to problems in which the electrical facilities cannot be turned off. Therefore, it is believed that an uninterruptible diagnostic judgment technique that determines compliance with related regulations such as electrical facility technology standards, internal wiring regulations, and inspection regulations without turning off the electrical facilities and at the same time detects abnormal conditions of the facilities early, it is possible to prevent electrical accidents and improve the efficiency of electrical facilities. In this paper, we propose an uninterruptible power diagnosis judgment technique that can prevent or reduce electrical accidents in cast-iron transformers by applying judgment criteria of diagnostic sensors for various types of measurement parameters that can diagnose and evaluate the presence or absence of abnormalities in electrical equipment, including partial discharge, and AI algorithms learned from data of diagnostic sensors.

In this research, it have developed a sensor that could diagnose inner deterioration of covered wires. With this sensor it observed results from simulation, and the attribute required for realization. For simulation it have used FLUX, it have considered all of geometric and electromagnetic information from coil and base metal that influences eddy current sensor`s property in order to predict the final result. It assumed there is no mutual inductance in the coil with N number of turns, because equivalent current flows in coil that is continuously connected in eddy current sensor. It assumed circular coil loop draws a circle, always have self inductance, and they are connected in series and overlapped according number of turns (N) in coil, and bobbin configuration. Actual sensor was produced with consideration of inductance and number of turns (N). In conclusion, it were able to test the dependency through results from simulation, actual measurement, and modeling of simulation. It is considered that attributes of respective base metal and structure can be predicted by simulating in advance.