Artificial Neural Networks for Power Transformers Fault Diagnosis Based on IEC Code Using Dissolved Gas Analysis

Transformer is the main important equipment in electrical power system. Early stage detection of the transformer faults has great economic significance because it considered expensive equipment and it helps to maintain the continuous operation of the electrical power system. Transformer oil is used for two main purposes, one for insulating liquid and the other for cooling. Some physicalchemical tests are carried out to determine the physical and chemical properties of the oil. Dissolved Gas Analysis (DGA) is now considered a common practice method for detection of the transformer incipient fault. This paper focuses on the employment of the Artificial Neural Network techniques (ANN) to diagnose dissolved gas in transformers, in order to determine the fault causes based on the IEC standard method. The ANN on IEC Code results meets the similar results of the other techniques that use to diagnose the transformer fault. Therefore, this method is very reliable to use as a diagnostic tool for transformer fault detection. Keywords—Transformer faults, Dissolved gas analysis,