Authors:
G. Ravikumar, K. Srinarayani, Anusha Bhattacharya, Vanessa M. Willemse, Mykhailo Paslavskyi, Mohammad Ayaz Ahmad
Addresses:
Department of Energy, SBM Offshore, Marly, Fribourg, Switzerland. Department of Computer Science in Artificial Intelligence and Machine Learning, SRM Institute of Science and Technology, Ramapuram, Chennai, Tamil Nadu, India. Department of Mathematical Sciences, Indian Institute of Science Education and Research Mohali, Mohali, Punjab, India. Department of Education, Northern Cape Department of Education, Kimberley, Northern Cape, South Africa. Department of Computer Science, Ukrainian National Forestry University, Lviv, Lviv Oblast, Ukraine. Department of Mathematics, Physics and Statistics, University of Guyana, Georgetown, Guyana, South America.
The rapid growth of renewable energy sources such as solar and wind power has intensified the challenge of maintaining an accurate balance between electricity supply and demand in modern smart grid infrastructures. Conventional forecasting approaches based on statistical models and rule-based heuristics fail to capture the non-linear, seasonally dynamic relationships inherent in energy consumption and generation data. This paper presents a Renewable Energy Demand–Supply Intelligence System that integrates Random Forest, XGBoost, and Long Short-Term Memory (LSTM) networks to forecast electricity demand and renewable energy supply simultaneously. The system is trained and evaluated on the Spain ENTSO-E hourly dataset spanning 2015–2018, comprising over 35,000 hourly records with weather covariates. Engineered temporal features, multi-horizon lag variables, and cyclical encodings are employed to capture diurnal, weekly, and seasonal demand patterns. XGBoost achieved the best demand forecast accuracy (R² = 0.9868, MAPE = 1.18%), while Random Forest yielded R² = 0.9664 for renewable supply prediction. Gap analysis across 7,961 test hours revealed a persistent mean deficit of 22,309.16 MW, with renewable sources covering only 23.24% of total demand. An interactive, Streamlit-based decision-support dashboard provides real-time gap visualization, feature-importance analysis, and smart energy insights. The proposed framework offers a scalable, data-driven solution for proactive energy management and smart grid optimization.
Keywords: Random Forest; Smart Grid Infrastructures; Renewable Energy Sources; Rule-Based Heuristics; Engineered Temporal Features; Cyclical Encodings; Data-Driven Solution; Extreme Gradient Boosting (XGBoost).
Received on: 24/11/2024, Revised on: 31/01/2025, Accepted on: 26/03/2025, Published on: 05/10/2026
DOI: 10.69888/FTSES.2026.000798
FMDB Transactions on Sustainable Energy Sequence, 2026 Vol. 4 No. 2, Pages: 123-132