We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
An investigation on characteristics of tracking failure in epoxy resin with harmonic and fractal dimension analysis.
- Authors
İSPİRLİ, Mehmet Murat; ERSOY YILMAZ, Aysel
- Abstract
Epoxy resins have excellent mechanical and electrical properties; thus, they are commonly preferred in the electrical{electronic sector as insulation materials. These insulation materials experience degradation such as breakdown, tracking, or treeing. Therefore, safety and reliability tests of insulation materials have great importance in estimating the performance of insulation. These tests must be carried out to decrease the risk of electrical insulation fault. In this study, the relationship between tracking failures and the data obtained during experiments was evaluated. Tracking failure performance of epoxy resin was investigated using the IEC 60112 test standard, which is known as the comparative tracking index test. Leakage current data, collected from the experimental setup recorded from a ground electrode, were decomposed by using the Fourier transform method. The 3rd, 5th, and 7th harmonic distortion components of the leakage current data were calculated and their relationship with the number of drops falling on the sample was interpreted. Then the harmonic distortion components relationship was explained under dirty and clean surface conditions. We explained the tracking pattern formed on the surfaces of the samples after completion of the experiments was decomposed using image processing techniques. The capacity, information, and correlation dimensions of the resulting tracking patterns were calculated using fractal dimension analysis. Shapes in the same terms were examined using fractal dimension analysis.
- Subjects
EPOXY resins; FRACTAL dimensions; ELECTRICAL insulation failure; HARMONIC analysis (Mathematics); INSULATING materials; MATERIALS testing
- Publication
Turkish Journal of Electrical Engineering & Computer Sciences, 2018, Vol 26, Issue 1, p245
- ISSN
1300-0632
- Publication type
Article
- DOI
10.3906/elk-1705-32