Development of nominal rules on the Fuzzy Sugeno method to determine the quality of power transformer insulation oil using Dissolved Gas Analysis data

Abstract

This paper aims to develop the nominal rules on the Fuzzy Logic Method using the Sugeno-Fuzzy Inference System (FIS) for Dissolved Gas Analysis (DGA) and determine the quality of the power Transformer 1 and Transformer 6 insulating oil at the Buduran 150 kV substation. The nominal number of proposed fuzzy rules is 1920 rules. Implementing the Fuzzy-Sugeno method on Transformers 1 and 6 shows that the six input variables from the DGA test can produce a Total Dissolved Combustible Gas (TDCG) output value of 32.67 and 26.19 ppm, respectively. Both values indicate that the insulating oil of Transformers 1 and 6 are in condition one and, at the same time, indicates that the dissolved gas composition is in Normal status. Furthermore, the TDCG value, condition, and quality status of the insulating oil have the same or 100 % accuracy compared to the DGA test by PLN (UPT Surabaya). Thus, the nominal development of fuzzy rules using the Fuzzy-Sugeno method can perform DGA analysis more accurately to determine the quality of power transformer insulation oil compared to previous studies.

Keywords

• dissolved gas analysis
• fuzzy inference system
• fuzzy rule
• fuzzy-sugeno
• membership function
• total dissolved combustible gas