Classification of transformer faults and identification of faults using smart methods

Instant detection of malfunctions in power transformers, one of the equipment used in the production and operation of electrical energy; It ensures the protection of the system and equipment. By reporting malfunctions and preventing power outages, continuity of energy is ensured and financial losses are prevented. Some gases are formed as a result of malfunctions or malfunctions in the transformer oil used for insulation. With the oil dissolved gas analysis used to eliminate faults, the fault can be intervened at the initial stage of the fault. These gases formed as a result of malfunctions are stored in DGA systems. DGA is basically implemented in two ways. Among these, there are known classical methods and, more recently, smart systems, that is, flexible models. Intelligent systems are used to ensure the stability of classical systems in detecting faults. In this way, transformer faults that occurred between 2018 and 2023 were examined and classified according to where the faults occurred, that is, according to the fault condition. However, two different studies have been conducted on the malfunctions that occur in these power transformers. Since the initial billing rates were not very high, the faults were calculated using the Duval triangle and IEC rates. It has been observed that the Duval ratio method gives better results than the IEC ratio method. The last fault was evaluated by combining the limit values of the gases formed as a result of the fault. Five different methods are used here. Due to the limitations of the IEC Ratio and Rogers methods, which are the classical methods of DGA, in fault detection, smart parameters analytical IEC, module Rogers and module Duval methods were used to ensure the use of correct analysis. Fuzzy Duval was detected by more accurate fault analysis than the classical methods of the main IEC and analytical Rogers methods. Again, this general information was observed in detail by performing a more accurate failure analysis than the IEC and comprehensive Rogers methods.