Improving Diagnostic Technique of Moisture Content in Power Transformers Using Adaptive Neuro-Fuzzy Inference System

Effective diagnostic techniques of moisture content are fundamentally meant to maximize proper maintenance planning, reliability, quality and safe operation of power transformers. A literature survey reveals that moisture content drastically affects the insulation system of power transformer by; reducing its dielectric strength, accelerating cellulose ageing and causing partial discharge at high temperatures. Traditional technique such as Dew-Point method of moisture evaluation, although used by power transformers' manufacturers, contains few errors related to inaccurate measurements. This work discusses the improvement of Dew-Point (DP) measurement approach using artificial intelligence, particularly the Adaptive Neuro Fuzzy Inference System (ANFIS). The proposed method has proven to have improved the diagnosis efficiency of Dew-Point techniques by 5%.