Transformer Oil Quality Assessment Using Machine Learning Techniques

Abstract

In this suggested work, difficulties in most engineering applications are overcome, and vector machine learning is used to support the state of the power transformer oil based on test findings. For nonlinear issues, such as higher dimensional problems, the support vector machine kernel function is utilized. In machine learning algorithms, kernelized SVM functions such as sigmoid kernel function (SKF), radial basis kernel function (RBF), Gaussian kernel function (GKF), and Bayesian optimization (BO) are performed. To evaluate the quality of high voltage power transformer oil, AI has been used. The oil test on the power transformer is adequate for continued usage., oils that simply require conditioning filtration for future service, oils in very bad condition, the oil should be reclaimed or disposed of, It is theoretically permissible to dispose of oils that are in such terrible condition. In comparison to existing machine learning algorithms, the suggested Bayesian optimization approach achieved a recognition rate of 99.5 percent. The oil test characteristics were transformed and filtered of the dataset, which enhanced the classification performance of the employed methods; The support vector machine kernel function is beneficial for monitoring transformer oil condition.