ANN Based Methodology for Determination of Distribution Transformer Health Status

The objective of this research paper is to utilize artificial neural network for evaluation of the distribution transformer health status by utilizing real measurement from working distribution transformers. The calculated health status is utilized for assessment of the distribution transformer health condition. The author has utilized the real data collected from the field to find out the health status of distribution transformer. The real data for diagnostic tests is taken from power distribution utility in the Nagpur rural area. The diagnostic data are voltage, load current, transformer oil level and top oil temperature. The training of the ANN has done by using real measurements of the distribution transformers. The real data is utilized for testing of trained neural network performance for working distribution transformer. The performance evaluation is tested on specially designed 15 kVA, 400/400 V three phase distribution transformer in the laboratory which indicates that the trained neural network is reliable for detecting the health status of any working distribution transformer.