Abstract
The measurement and subsequent analysis of partial discharges (PD) have become a fundamental tool for diagnosing insulation systems, and thus, assessing the reliability of power electrical equipment. PDs are used to characterize the evolution of progressive defects such as electrical trees, considered one of the main mechanisms of degradation in polymeric insulation, that leads to failure. In this work, a new methodology for the identification of the aging stage of electrical trees is proposed, based on the measurement and analysis of each PD pulse waveform by means of the discrete wavelet transform (DWT) and principal component analysis (PCA). A total of 24 treeing samples of epoxy resin were analyzed and tested under AC voltages of 10, 12, and 14 kV with frequencies of 50 to 550 Hz. The results showed that it is possible to characterize the development of electrical trees into three stages: initial, valley, and post-crossing; showing different behaviors in terms of discharge and tree structure. It was found that PD can be classified into two groups, in which one group (‘Group (b)’) served as an identifier of the post-crossing stage, and thus, an alert of the imminent insulation failure. Regarding the structure of trees, it was observed that growth in the valley stage can be associated with the development of filamentary branches with low PD activity, while growth in the post-crossing stage can be associated with the development of dark branches with higher PD activity. In some cases, reverse trees grew in this last stage. Additionally, it was observed that using the methodology proposed, the development of dark branches was associated with the appearance of Group (b) PD. The ability of the presented measurement and evaluation technique to identify growth-stage of electrical trees at different applied voltages and frequencies shows its potential for the implementation to insulation assessment of real power equipment in service, such power cables.
Funding Information
  • Agencia Nacional de Investigación y Desarrollo (ANID) through the Fondo Nacional de Desarrollo Científico y Tecnológico (11181177)
  • Universidad Técnica Federico Santa María
  • Programa de Incentivos a la Iniciación Científica

This publication has 31 references indexed in Scilit: