An ANN-based energy forecasting framework for the district level smart grids

This study presents an Artificial Neural Network (ANN) based district level smart grid forecasting framework for predicting both aggregated and disaggregated electricity demand from consumers, developed for use in a low-voltage smart electricity grid. To generate the proposed framework, several experimental studies have been conducted to determine the best performing ANN. The framework was tested on a microgrid, comprising six buildings with different occupancy patterns. Results suggested an average percentage accuracy of about 96%, illustrating the suitability of the framework for implementatio