Univariate Time Series Prediction of Reactive Power Using Deep Learning Techniques

Abstract

To surmount the issue related to reactive part of produced power, it is imperative to predict its quantity of generation. Optimizing this back and forth energy flow will enlarge actual part of energy flow. Generating reactive part motives to gain the magnetic field voltage and conversely, no flow of reactive part impedes sending of requested energy. In this paper, two popular deep learning frameworks, Long Short-Term Memory (LSTM) and Artificial Neural Network (ANN) are adopted to forecast reactive part of generated power.