Statistical and numerical analysis of transformer oil AC breakdown

Abstract

Transformer oil AC breakdown research is an important part of transformer insulation design. Research presented in this paper consists of statistical and numerical analysis of breakdown data measured in portable oil tester. Statistical analysis is done by modeling measured data as a random process with Gaussian and Weibull probability function. Numerical analysis uses statistical data for calculation of stressed oil volume, stressed electrode area and safety factors of “cumulative stress” method. Both statistical and numerical analysis showed how breakdown withstand depends on different variables and why they are important in measurement interpretation.