An Iterative Optimization and Learning-Based IoT System for Energy Management of Connected Buildings

Buildings account for nearly 40% of primary energy and 36% greenhouse emissions, which is one the main factors driving climate change. Reducing consumption in buildings toward zero-energy a vital pillar to ensure that future targets are reached. However, due high uncertainty building loads customer comfort demands, extremely nonlinear thermal characteristics, developing an effective management (BEM) technology facing great challenges. This article proposes novel learning-based iterative Internet Things (IoT) system address these challenges achieve objective BEM connected buildings. First, all IoT-based share their operation data with aggregator. Second, aggregator uses historical train deep reinforcement learning model based on deterministic policy gradient method. The generates precooling or preheating control actions heating ventilation air conditioning (HVAC) systems. Third, solving coupling problem between HVAC systems internal heat gain loads, optimization algorithm developed integrate physics-based models minimize deviation on-site solar photovoltaic generated actual by properly scheduling electric vehicle charging cycles, energy-storage system. Finally, optimal load considering customers’ requirements. All then operate schedule issued proposed IoT validated via simulation real-world from Pecan Street project.