The Estimation of Breakdown Voltage of Vegetable Oil using Support Vector Machine

Abstract

Oil as an insulating medium is widely used in power apparatus and it is important to have knowledge about its breakdown characteristics. Support Vector Machine (SVM) can be a fruitful tool for estimation of breakdown voltage (BDV). In this work, the objective is to explore the application of SVM to estimate breakdown voltage of vegetable oil. Experiments are carried out on vegetable oil to obtain its characteristic breakdown voltage using Weibull distribution. Experiments are carried out using different electrode geometry and electrode gap. At the breakdown condition, the electric field distribution is simulated using FLUX software and various electric field features such as electric field intensity, energy density, etc. are extracted. These electric field features are preprocessed and used to train SVM. The optimum value of SVM parameters are obtained using grid search and K – fold cross validation technique. The trained SVM model is used to estimate the breakdown voltage of the oil medium under different electrode gap and shape. It is seen that the estimated BDV fairly matches with the experimental results. © 2024 Author(s).