Intelligent Machine Learning Techniques for Condition Assessment of Power Transformers

Abstract

This paper introduces a condition assessment of power transformer in term of percentage of health index (%HI) by using regression models. The conditions of major components of power transformer are assessed by using input datasets from visual inspection, electrical test as well as paper and oil insulation test. 90 features of these input datasets are tested in regression models for determining the predicted HI. Six regression models such as linear regression, Ridge regression and Lasso regression, random forest regression, support vector regression, and deep neural network regression are tested to predict %HI. Actual input datasets related to actual %HI of 317 power transformers are used to teach such learning regression models. The random forest regression performs the best model providing the best output dataset with the lowest errors.