Evaluating BDV in Transformer Oil Combined MFU and GRNN

Power transformers are regarded one of the crucial part of transmission and distribution systems. Insulating transformer oil quality directly affects the operation of the transformer, and breakdown voltage is the main parameters to estimate transformer oil quality. Thus, monitoring the breakdown voltage (BDV) of transformer oil is considered by the power industry as an important means to ensure the safe operation of power systems. This work proposes the measurement of the BDV of transformer oil using multi frequency ultrasonic and generalized regression neural network (GRNN). All 210 samples were measured for breakdown voltage according to traditional testing methods, and were tested by multi-frequency ultrasonic. Then, 210 samples were divided into training sets and test sets. After that, multi-frequency ultrasonic data as the input of GRNN, and the BDV as the output of GRNN. Afterwards, the 20-fold-cross-validation was incorporated to get the best smoothing factor $\sigma$ for GRNN. Then, the GRNN with the best smoothing factor $\sigma=4.54$ was trained with the training sets, and the model was verified with the test sets. The experimental results show that the GRNN model can predict the BDV of transformer oil from multi frequency ultrasonic testing data, and it also provides a new method for evaluating the health of transformer oils.