INTERNATIONAL ENERGY JOURNALVolume 22, No. 03, Month SEPTEMBER, Year 2022, Pages 269 - 280
Determination of copper losses for substation transformers in special region of yogyakarta using the fuzzy system
Agus Maman Abadi, Toto Sukisno, Nurhayadi Nurhayadi, Wredha Pringga Kusuma Hano
Abstract Download PDFTransformer’s efficiency is essential and significantly affects the transformer’s performance in delivering power. The efficiency is affected by copper losses, of which value is not constant. To improve a state-owned electricity company in Indonesia (PLN) in increasing the optimization of transformer loading, a more accurate calculation of transformer copper losses is needed. In this study, a new method was proposed to assess the value of copper losses using a fuzzy system, namely the first-order Takagi-Sugeno-Kang (TSK) method combined with Singular Value Decomposition (SVD). The fuzzy system was built based on training data. Upon this data, clustering was carried out with fuzzy c-mean (FCM). The results of the FCM were cluster centers, which were then used to construct membership functions and fuzzy rules. The parameters on the consequent of each first-order TSK fuzzy rule were determined using SVD. The last step was defuzzification to obtain the value of the transformer copper losses. The defuzzification method used was the weight average. The fuzzy system that had been built was tested on all data to determine and obtain the accuracy of the copper losses values. The results of this study indicated that the determination of transformer copper losses with a fuzzy system for training and testing data have accuracies of 99.3354% and 99.6490%, respectively. Furthermore, the first-order TSK method gives better results than that of the zero-order TSK and Mamdani methods.
Fuzzy system; First-order TSK; Singular value decomposition; Transformer copper losses; Transformer efficiency