Energy-Aware Buildings Reconfigure Internal Systems to Reduce Peak Demand

The rapid increase in urban energy demand highlights the need for intelligent systems that reduce building peak loads without compromising occupant comfort. This study proposes EARS, an LSTM + DyPeS-EAR control framework designed to forecast and dynamically reconfigure building subsystems such as HVAC, lighting, and appliances. The LSTM model predicts short-term energy demand using environmental and occupancy data, while the DyPeS-EAR controller optimally adjusts loads in real time based on priority and comfort constraints. Simulations conducted on a six-zone commercial office building using EnergyPlus and Python demonstrate significant efficiency gains, achieving a 38% reduction in peak demand and a 30% decrease in daily energy consumption. Results confirm that integrating predictive learning with adaptive control enhances system responsiveness and grid stability. The proposed framework provides a scalable pathway for energy-aware smart buildings, supporting sustainable and cost-effective urban energy management.

Перевод пока недоступен