Reliability assessment of transformer thermal model parameters estimated from measured data

Abstract

This paper presents a methodology to assess the reliability of substation distribution transformer thermal model parameters estimated from measured data. The methodology uses statistical bootstrapping to assign a measure of reliability to the estimated parameters using confidence levels (CL) and confidence intervals (CI). The bootstrapping technique, which is used to make a small data sample look statistically large, allows a precise estimate of transformer reliability. The proposed methodology is tested on a 28 MVA transformer for which different data sets are available. The CTs are evaluated for both cases: with and without bootstrapping and the reliability indices compared. The results show that the CI values with bootstrapping are more consistently reproducible than the ones derived without bootstrapping.