Statistical Techniques for Prediction of Breakdown Voltage of Liquid Dielectrics used in Power Transformers

Abstract

Due to growing energy demand and high transmission voltage, it is desired to have high breakdown strength of the insulating oil to avoid failure of the transformer, which may lead to power supply disruption. Dielectric properties of the conventional insulating oil can be improved by substituting a suitable nanofluid. This paper analyses the AC breakdown strength of transformer oil and TiO2 nanofluid using two statistical techniques, i.e., Weibull distribution and Normal distribution. From the normal distribution, mean value and standard deviation are estimated, while the Weibull distribution is used to estimate the scale and shape parameters. The estimated skewness factor for the normal distribution is observed to be non-zero. The experimental data are distributed asymmetrically according to this non-zero skewness. The asymmetrical form of the Weibull distribution fits the experimental data nicely. The typical breakdown voltage magnitude is provided by the Weibull distribution and is considered more favorable for predicting dielectric breakdown.