Power transformer health condition evaluation: A deep generative model aided intelligent framework

This paper presents a deep generative model-aided intelligent framework for effective health condition evaluation of power grid transformers. The health assessment of a power transformer is required to guarantee the stable and sustainable operation of the grid and to precisely convert the electrical energy. A power transformer must undergo a series of tests to determine its state of health and identify its health index. In this paper, we develop a novel approach to identify and classify the health condition of power transformers using a machine learning approach. The proposed framework is structured by using a multi-layer perception generative model with a logistic regression classifier. The developed model uses the twelve input layers which enables the model to effectively compressed the dataset and eight categories in the output classification layers. The effectiveness of the proposed model is examined on the real-world testing data set of 31 categories of six hundred and eight transformers. The obtained performance using the proposed framework confirms its efficacy in precisely evaluating the transformer's health condition. The obtained results have also been compared with the existing machine-learning models. The comparisons show that the proposed model outperforms the state-of-the-art models by achieving 99% of accuracy.