AI-Powered Energy Consumption Optimization System Using Hybrid GRU–LSTM Model

Rapid growth in energy consumption driven by industrialization, urbanization and the proliferation of digital devices mandates better energy use management systems. Legacy energy use systems rely on static rigid rulesets. They lack agility to adapt to dynamic energy use patterns. To address this limitation, this research proposes an AI-Powered Energy Use Optimization System (AIPECS) using an artificial intelligence (AI)-based deep learning and intelligent optimization methodology (i.e. deep learning via hybrid GRU and LSTM neural networks). The hybrid predictive model assesses both long and short-term energy consumption trends through analysis of paired data from a combination of IoT sensor networks, smart energy meters, and environmental variables such as temperature, relative humidity, and occupancy. In order to determine the predictive accuracy of the hybrid predictive model, various performance metrics (MAE, RMSE, MAPE, R²) were evaluated using the results of several validation tests. A total of 98.3% accuracy was achieved with the hybrid model when compared to the accuracy of the GRU-only and LSTM-only models. In addition, a web-based dashboard developed using Streamlit was created to provide real-time monitoring of energy use and recommendations on how to improve energy efficiency. The integrated use of AIPECS will result in reduced energy use, significant cost savings, improved operational efficiencies, and ultimately, the sustainable use of energy in residential, commercial, and industrial settings.