A Fuzzy Logic Proposal for Diagnosis Multiple Incipient Faults in a Power Transformer

For the safety and continuity of service in industrial electrical systems, availability transformers is essential. this reason, it necessary to develop intelligent fault diagnosis techniques reduce repair maintenance costs. Recently, several methods have been developed that use artificial intelligence such as neural networks, support vector machines, hybrid techniques, etc., for faults power using gas analysis. These methods, although they present very good results, encounter restrictions determine precise moment before occurrence multiple small magnitude are difficult implement practice. This document proposes a method diagnose incipient transformer fuzzy logic. The proposal, based on historical data from composition gases dissolved oil, achieves performance classification 98.3%. With reliable samples gas, guarantees an overall rate accuracy detecting superior obtained by most successful conventional industry. proposal does not generalization difficulties constitutes simple solution allows determining state without affecting electricity supply.