Implementation of Cognitive Artificial Intelligence for Dissolved Gas Analysis Interpretation in 150 kV Transformer

Cognitive artificial intelligence (CAI) is a new method in artificial intelligence (AI) that can emulate the human brain's ability for doing Knowledge growing systems (KGS). Previous research has presented the application of CAI based on Doernenburg ratio method (DRM) for Dissolved gas analysis (DGA) interpretation using IEC TC 10 dataset labeled. This paper implements CAI to the 150 kV transformer in Sepatan substations PT PLN (Persero) UPT Durikosambi UIT Jawa Bagian Barat. This paper also adds a comparison with several methods, there are: Fuzzy inference system (FIS), Duval triangle, Roger's ratio method (RRM), and Doernenburg ratio method (DRM). For the results, if referring to all data based on PLN's analysis, the accuracy of CAI shows 91.67%, whereas if it only refers to faults condition, the accuracy of CAI shows 100%. These results are the same as FIS. Thus, it can be concluded that this new method (CAI) can be implemented in DGA interpretation.