A nonlinear regression-based approach to assess transformer insulation condition using dielectric response recorded for short duration /

Abstract

Analysis of polarization current data is generally done to assess power transformer insulation condition. Due to considerable measurement time and low magnitude, noise and other environmental factors often affect the recorded dielectric response data. The influence of these factors cannot be avoided during field measurement. Once the recorded data are affected by external factors, it becomes difficult to analyze the data. Available literature has reported some techniques to reduce the testing time. However, these reported methods are ineffective in addressing the practical issues experienced during field measurement. This present article proposes a nonlinear regression-based approach to reduce insulation response measurement time significantly. Data collected from various in situ transformers have been analyzed to test the effectiveness of the proposed method.