Identify moisture distribution in multilayer oil-immersed paper by Frequency Domain Spectroscopy

Moisture content of oil and paper is one of the most important factors influencing the dielectric behavior of insulation systems. Moisture estimation in insulating paper can provide a way to understand the real condition of insulation and helps to assess its remaining useful life. In recent years, Frequency Domain Spectroscopy (FDS) is widely used by power utilities for moisture estimation and condition assessment of oil-paper insulation equipment, such as transformer, bushing and cables, and the estimated moisture value is an equivalent one of the whole facility. Actually, moisture distribution in the insulation is very complicated due to temperature and load variation, and failures usually happen in the weakest place with high moisture or potential defects. In this paper, some single layer papers with the same average moisture content but different moisture distribution were prepared, and FDS measurements are performed on these samples to identify the different moisture distribution. Although the dissipation factor is used to reveal the whole insulation condition in traditional concept, the results show that different distribution has different performance in FDS curves. The curves are mainly determined by the highest moisture part, and the distribution difference displays in the low and high frequency range.