Method for Optimized Analysis of Electrical Tests of Insulation Power Factor in Power Transformers

In this paper is presented a method for the classification of electrical tests of insulation power factor in power transformers, which performs the assignment of individual scores and ratings to each one of the measured variables through these tests and a final rating. To do this, it uses injective functions with which, according to the value of measured quantities, appropriate scores and ratings are determined. The equations of these functions are defined using maintenance engineering criteria in conjunction with computational optimization techniques ‑ the Hill Climbing algorithm and the 1/5-th Success Rule. It's also presented the equations for the correction of the power factor in function of the temperature of the tests, in view of the adequate classification of these. Furthermore, the proposed method, due to its efficiency, corroborated by the high percentage of correct answers provided by it, is already being successfully implemented in the company CELG Distribuição S.A. (CELG D), helping the decision-making and, thus, contributing to an efficient maintenance of the company's equipments.