Large Power Transformer Magnetic Core Vibration Model by Using Dynamic Genetic Algorithm

Large power transformers can experience multiple different possible types of faults. One of them is mechanical faults that occur within the windings and magnetic core of the power transformer. In order to carry out diagnostics for this type of fault and evaluate whether there is a mechanical defect present, it is necessary to obtain information about the generated vibrations inside of the structure of transformer from the measured values on the transformer tank. However, these vibration values are distorted since they propagate the transformer structure before being received. Therefore, a simulation of vibrations generated by the transformer is made using a mass and spring system and a dynamic genetic algorithm in order to acquire the corresponding spring stiffness coefficients that generate the measured vibration values on the tank surface. This paper describes a case study created in MATLAB software for a single magnetic core rod of a large power transformer.