A General Methodology for Evaluation and Classification of Oil-Immersed Power Transformers: Application to Electrical and Physicochemical Parameters

Condition monitoring of power transformers is of vital importance to prevent electricity supply stoppages and reduce power plant maintenance costs. To that end, the use of techniques to evaluate and classify the condition of these devices is highly recommended in order to obtain good quality informa...

Full description

Saved in:
Bibliographic Details
Published in:Journal of control, automation & electrical systems automation & electrical systems, 2019-10, Vol.30 (5), p.832-839
Main Authors: da Cunha Brito, Leonardo, Marques, André Pereira, de Jesus Ribeiro, Cacilda, de Moura, Nicolas Kemerich, Dias, Yuri Andrade, Azevedo, Cláudio Henrique Bezerra, dos Santos, José Augusto Lopes, da Silva Palhares, Pedro Henrique
Format: Article
Language:English
Subjects:
Artificial intelligence
Classification
Condition monitoring
Control
Control and Systems Theory
Electrical Engineering
Electricity pricing
Engineering
Maintenance costs
Mechatronics
Methodology
Optimization
Plant maintenance
Robotics
Robotics and Automation
Statistical methods
Transformers
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Condition monitoring of power transformers is of vital importance to prevent electricity supply stoppages and reduce power plant maintenance costs. To that end, the use of techniques to evaluate and classify the condition of these devices is highly recommended in order to obtain good quality information for their proper maintenance planning. This article presents and details a general methodology for the creation of methods to evaluate and classify these devices, by means of computational modeling and optimization. The results indicate a higher than 93% accuracy rate compared to that of numerical evaluations and symbolic classifications expected by experts, thus demonstrating the applicability of the proposed methodology, which is found to be superior in comparisons against Computational Intelligence and Statistical Learning methods.
ISSN:2195-3880
2195-3899
DOI:10.1007/s40313-019-00487-6