Buildings account for nearly 70% of the electricity usage in the U.S. To improve the energy efficiency of buildings and the overall power grid, it is critical to intelligently manage various energy demands of buildings and coordinate such management across buildings in the smart grid.
At the building level, a key aspect in improving building energy efficiency is to leverage the scheduling flexibility provided by various energy loads and supplies, including HVAC (heating, ventilation and air conditioning), EV (electric vehicle) charging, datacenter computing loads, battery storage and solar power, etc. Among various energy loads, HVAC system accounts for 50% of the total energy consumption. The thermal flywheel effect allows buildings to provide significant flexibility by temporarily unloading the HVAC systems without immediate impact on occupants. However, the building temperature often exhibits randomized behaviors under incomplete modeling due to various building structure and materials, and disturbances from environment and occupants. Therefore, it is essential to develop a data-driven approach for HVAC control and co-schedule heterogeneous energy demands and supplies in a holistic framework.
At the power grid level, it is important to fully leverage the scheduling flexibility from buildings. Most of the previous works, however, focus on developing price-based or incentive-based demand response (DR) strategies to improve power grid efficiency, in which the building energy management system passively follows load reduction signals from the utilities. Little work has been done to consider integrating the intelligent building energy scheduling process with the electricity market economic dispatch strategy in a holistic framework. To further exploit the huge potential of demand response in improving power system efficiency and facilitate customers' engagement level in electricity market, we propose an innovative demand response scheme based on proactive demand participation from smart buildings.
[Click figure to enlarge]
