STTR Phase I: An Agent-based Self-learning Technology for Efficient Building Operations and Automated Participation in Electricity Markets

Period of Performance: 07/01/2016 - 06/30/2017


Phase 1 STTR

Recipient Firm

1224 Providence Ter
Mc Lean, VA 22101
Firm POC, Principal Investigator

Research Institution

Virginia Institute of Technology
Institution POC


The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is the capability to enable small- and medium-sized commercial buildings to operate more efficiently and facilitate their participation in electricity markets. It will result in energy consumption reduction in buildings, thus providing savings in electricity bills and lowering carbon footprints, without compromising occupant comfort. The proposed platform also enables revenue generation by allowing buildings to participate in ancillary electricity markets and respond to pricing or demand response signals from electric utilities. By integrating intelligent energy management and automatic demand response features to the traditional building operation, this helps electric utility companies to avoid/defer extensive upgrades of their electrical infrastructures, such as generation, transmission or distribution facilities, with the growing demand for electricity. The proposed software platform demonstrates the benefits of building automation to building owners/operators, and provides them the ability to cross-reference their buildings with best practices in building operations. It also bridges the knowledge transfer process between a small business and an institution of higher education by facilitating both sides to meet, work and exchange experiences with each other. This Small Business Innovation Research (SBIR) Phase I project is built on an agent-based platform for building automation systems where advanced algorithms enable the software platform to be proactive and help optimize building operations. It can also perform fault detection, diagnostics and predictive maintenance, integrate emerging Internet of Things (IoT) devices and enable buildings to interact with the grid in an intelligent manner. Under the proposed project, a reinforced machine-learning algorithm will be developed that uses historical building energy consumption data and ambient conditions - including occupant preferences - to optimize building operations based on usage patterns of major loads and power sources in each zone of the building. Another major research contribution lies in novel algorithms that enable buildings to transact with the grid. These allow buildings to respond to system stress conditions and electricity price signals by automatically adjusting the operation of major loads and internal generating sources. Overall, the proposed software platform allows a building to optimize its operation, perform peak demand management during grid stress conditions, and participate in ancillary service markets.