Support vector machine and tree models for oil and Kraft degradation in power transformers. (September 2021)

• Record Type: Journal Article
• Authors: Arias Velásquez, Ricardo Manuel
• Is Part Of: Engineering failure analysis. Volume 127(2021)
• Publication Date: 2021-09
• Subjects: Kraft paper -- Oil quality -- Power transformer -- Reactor -- Support vector machine; System failures (Engineering) -- Periodicals; Fracture mechanics -- Periodicals; Reliability (Engineering) -- Periodicals; Pannes -- Périodiques; Rupture, Mécanique de la -- Périodiques; Fiabilité -- Périodiques; Fracture mechanics; Reliability (Engineering); System failures (Engineering); Periodicals; Electronic journals; 620.112
• DOI: 10.1016/j.engfailanal.2021.105488
• Languages: English
• ISSNs: 1350-6307

Abstract

Highlights: Support vector machine and tree model for degradation analysis in power transformers. A cluster analysis is proposed for partial discharge analysis and to improve the evaluation. Remaining life is created with tree model technique associated to decision making process. Abstract: The power transformer analysis focused internal fault identification is important for the energy efficiency in all the countries, the partial discharge is one of the main failure modes, it affects the remaining life until the failure. In this research article, the new approach by using soft techniques as support vector machine (SVM) and tree models (TM) allow to detect partial discharge with high accuracy (97.55%); dissolved gas analysis (DGA) for the quality of the oil-Kraft, degradation and evaluate the remaining life. An important achieve is the highest accuracy for the thermal influence in oil degradation (100%). Finally, this research article has provided an algorithm for the highest accuracy obtained from the DGA for oil and Kraft degradation with a crosscheck process by using the cluster analysis.