A preliminary study of transformer life estimation based on linear cumulative damage theory

Abstract

SUMMARY Transformer life is primarily influenced by thermal, electrical, and mechanical stresses, and each type of stress can be subdivided into time-varying common stress from the normal usage and variable stochastic stress caused by unusual events. Estimating the cumulative effect of variable stress is essential for life estimation. Current research on the cumulative effect of time-varying common stress or variable stochastic stress varies without a unified theoretical framework, resulting in the separate studies on common stress and stochastic stress. In this paper, the combined effect of time-varying common and variable stochastic stresses on transformer life is studied based on linear cumulative damage theory, where the concept of damage is defined as a universal measurement of the deterioration from common or stochastic stress, and it is linearly incorporated to obtain the linear cumulative damage model. The reasonability of this model is verified by being compared with various widely accepted and experimentally validated life estimation models, and this model is proved to be a unified theoretical framework for all three types of stresses. Subsequently, the probabilistic form of the model is obtained with the assumption that the failure criterion obeys a Weibull distribution. Two application cases on thermal stress are presented to validate the model proposed in this paper. Copyright © 2013 John Wiley & Sons, Ltd.