Adaptive fuzzy logic expert system for predictive maintenance of transformers installed in cold climatic conditions

• Published in: Insight (Northampton) Vol. 67; no. 10; pp. 620 - 627
• Main Authors: Wani, N U I, Ranga, C, Gupta, N
• Format: Journal Article
• Language: English
• Published: The British Institute of Non-Destructive Testing 01.10.2025
• Subjects: Cold Climatic Issues; Health Index; Insulations; Machine Learning; Transformer
• ISSN: 1354-2575
• DOI: 10.1784/insi.2025.67.10.620

In this paper, a novel multi-criterion analysis (MCA)-based fuzzy logic (FL) model is proposed to determine the overall health index (OHI) for transformers specifically operating in cold climatic regions. The model incorporates MCA, FL and various diagnostic parameters associated with cold climatic issues such as low temperatures, pour point, kinematic viscosity, water content (WC), cold start issues, oil age, etc. The proposed model is validated using data corresponding to 200 similar in-service transformers installed in different cold climatic regions in India. Furthermore, a comparative analysis is conducted between the outputs of the proposed model and the decisions taken by the transformer diagnostic experts in the field. Evaluation reveals that 91% of the results match. Decision-making using the conventional method is limited to very few diagnostic test outcomes, whereas the proposed diagnostic model assesses the same based on a large number of significant test parameters relevant to cold environments to enable a more comprehensive health assessment. As a result, the proposed model produces an accurate and reliable output. The integration of MCA addresses the rule-overload problem inherent in conventional fuzzy logic systems, enabling a more structured and reliable decision-making process. Moreover, the model is computationally efficient, easily implementable and specifically tailored to the unique challenges faced by transformers in cold climates, making it a valuable tool for utilities and maintenance engineers.