2026-05-10T08:04:24.965Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:11137 2025-02-08T20:29:39Z 3

Optimizing renewable energy utilization through production and demand forecasting using optimized recurrent networks 2025 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/11137 https://www.sciencedirect.com/science/article/abs/pii/B9780443298691000118 N. Bacanin M. Pavlov-Kagadejev S. Kozakijevic L. Jovanovic M. Zivkovic M. Antonijevic Renewable energy utilization is vital for reducing societies’ dependence on fossil fuels. Renewable energy is usually produced from natural sources which are influenced by several external factors. This can result in less reliable production compared to burning fossil fuels. An advantage of renewable energy is the ability to produce it in situ, reducing emissions and losses associated with transport. Furthermore, effective means of storage improve the reliability of power generated from renewable sources such as wind and solar. Utilization can further be improved through the application of artificial intelligence for forecasting both power demand and power production. This work aims to explore the use of time-series forecasting with recurrent neural networks (RNNs) for forecasting power produced from renewable sources, as well as household power consumption presenting a unified system for more sustainable use of generated renewable power. However, as the performance of RNN is heavily dependent on proper architecture and training parameter selection, this work also introduces a modified optimization algorithm specifically for the purposes of improving energy forecasting of renewable sources. The most effective models achieved a mean absolute error (MAE) of 0.145222 for production forecasts and 0.031544 for household load forecasting. bookPart Elsevier 203 225 10.1016/B978-0-443-29869-1.00011-8 Chapter in Renewable Energy Projects and Investments, Interdisciplinary Knowledge, Analysis, Opportunities, and Outlook