Nikjoo - Iran University of Science and Technology
Mirzai - Iran University of Science and Technology
Gholami - Iran University of Science and Technology

This paper presents the prediction of chemical characteristics of the transformer oil by neural networks. The aim of this work is to reduce the aging experiments time, with obtaining the same results as in the case of long duration tests, by using neural networks. This model is composed of some characteristics of the transformer oil such as viscosity, acidity, flash point, water content, gravity and colour that have been predicted by service period. This list of evidences indicates the beginning of the oil degradation process. In this paper we show the power of NN to predict those characteristics according to some mathematical models such as polynomial and multiple linear regression. Also we compare two different softwares as NN1 (Neuro Solution) and NN2 (Matlab) for predicting some of theses characteristics, to show the accuracy of the proposed NN.

Neural Network, transformer oil, aging