Assessment of moisture content in power transformer based on traditional techniques and Adaptive neuro-fuzzy interference system

Abstract

The use of traditional measurement techniques for condition monitoring of power transformer is still a common practice in the power industry. These techniques have proven to be unreliable as a result of sampling and analysis errors. Given the unequal moisture distribution between cellulose and mineral oil in power transformers, the dryness correlation between the two liquid insulators is not always accurate. The aim of this work is to advice the manufacturer of power transformers to continue use the Dew point measurement or move to the modern methods, like frequency domain spectroscopy (FDS). Dew point measurements have been used to estimate the dryness of power transformers, model the data by adaptive neuro-fuzzy inference system (ANFIS) as is proven to solve complex data and validate the results by Frequency Domain spectroscopy (FDS).