Age Estimation of Cellulose Paper Insulation in Power Transformers Using ANN

• Main Authors: Navamany J.S., Ghosh P.S.
• Other Authors: 6506106930
• Format: Conference Paper
• Published: 2023
• Subjects: Algorithms; Cellulose; Correlation methods; Data acquisition; Degradation; Insulation; Learning systems; Mathematical models; Neural networks; Service life; Age estimation; Cellulose paper estimation; Dissolved gas analysis (DGA); Electric transformers

The useful functional life of a power transformer is determined by the life of the paper insulation. Therefore researches on cellulosic paper degradation in power transformers are primarily directed towards the development of a mathematical model to estimate the age of the cellulose paper insulation based on the concentration of dissolved gases and furanic compounds. In this research work, a utility field study has been carried out on selected transmission power transformers with a wide range of ages. Samples of oil collected from the identified transformers were tested for concentration levels of CO and CO2 and furan derivatives 2-furfural. The results of the concentration of the above mentioned parameters clearly show that there is a dependence on age. Therefore, in this paper an attempt has been made to model the age(T) of the cellulose paper insulation in terms of the concentration of CO, CO2 and 2-furfural that T = f(CO, CO2, 2-furfural). The present modelling has been done using Artificial Neural Network(ANN). The estimated results using the proposed ANN model are further compared with the measured data collected during the field study and has shown a good correlation.