Voltage transformers play an important role in the transmission and distribution of energy, both by raising the voltage and decreasing the voltage. A good performance of the transformer has a great role in delivering the electrical energy to the consumer on time and without any problems. Since transformers work with magnetic circuits, leakage fluxes and losses due to them must be considered during the design. While designing the transformer, the designers decide on the most suitable winding type according to the location where the transformer will be used, the power of the transformer, its physical size and the area it occupies. Design programs may offer the ability to select the transformer winding type. However, deciding on the winding type by considering the leakage fluxes may takes a long time because it requires the transformer to be designed from the beginning or loaded through its simulation and the leakages must be calculated. In addition to other factors such as size and power, it is an important requirement to consider leakage fluxes in transformer design. In this study, support vector machines (SVM) technique, which is a machine learning method that can detect transformer winding types and perform quadruple classification, was used. It has been seen that the multiple classification and detection model made with the SVM model can detect the transformer winding types with 99.7% accuracy and the SVM technique can accelerate the decision of the winding type in the transformer design.
Anahtar Kelimeler: SVM, Transformers, ML, winding type, classification
|
