Deep learning based fault detection in power transformers

It is one of the most important components of the energy efficiency of transformer cables. Malfunctions in power transformers, which are the most important equipment of these centers, have serious financial consequences such as threatening the amount of energy supply. Faults in power transformers are very important in terms of prediction and detection system and protection of power transformers. In this study, general information about power transformers is given and the characteristics of the faults are mentioned. Information about the dissolved gas analysis method (DGA) used in predicting possible (internal) faults that may occur in transformers and determining the types of faults is given, and various DGA methods are mentioned. It has been shown that the detection of a possible malfunction in power transformers can be made more reliably by deep learning-based systems compared to classical methods. In this way, it has been observed that power transformers with high financial value are better protected, and severe damage that will increase the duration of power outages is prevented.