Mathematical models of power transformers winding faults diagnosis based on voltage-current characteristics

Power transformers are important components of electrical systems. Their failures are very costly, mainly because of the unavailability of electrical service they cause. Rapid and accurate diagnosis of internal transformer faults is a key factor of efficient and safe operation. Such diagnosis is usually carried out by experts. Several methods of power transformer windings diagnosis exist. The online diagnosis, the interpretation of results and the identification of various internal winding faults are criteria to be taken into account when choosing a method. The voltage-current method is a powerful method that has been successfully used as a diagnostic technique to detect power transformer winding faults. In this work we propose a mathematical models of power transformers winding faults based on voltage-current method. To achieve our goal, we modelled the transformer winding as a distributed network of similar R-L-C circuits. Then we used Matlab/Simulink to simulate the different winding faults on a number of discs ranging from 5 to 120. We defined four failures indicators and we studied their evolution for each fault. At the end we give for each fault the corresponding mathematical model. The transformer used in this work is a 15 MVA 22/0.415 kV distribution power transformer.