Quantifying the Moisture Content of Transformer Oil Using Dielectric Properties and Artificial Intelligence
Hashem Al–Mattarneh, Adnan Rawashdeh, Rabah Ismail
Abstract
Moisture content may significantly affect the performances of the transformer oil. Therefore, this paper presents an investigation of using the measured dielectric constant and loss of various transformer oils using electromagnetic electrode capacitive sensor which has been developed for this task. The proposed sensor was used to detect and quantify the water content in transformer oil. The electromagnetic sensor was used to measure the dielectric constant and dielectric loss of transformer and vegetable-based oils and relate these dielectric properties to oils moisture content. The dielectric sensor was validated and calibrated to enhance its accuracy in measuring the dielectric characteristics of transformer and vegetable-based oils. Dielectric constant and dielectric loss of transformer oils with varying moisture content were computed. The findings illustrate that both the dielectric loss and dielectric constant of vegetable-based and transformer oils increase as the water content rises The microwave dielectric sensor effectively determines the water content in transformer oil. The testing is fast, invasive, low cost and time consuming. advantage of this sensor. These advantages could have made this sensor as a novel quality control method for transformer oil. AI using Cascade ANN model shows the potential of prediction oil moisture content more than empirical regression models with R2 equal 0.998.