The power grid is a large interconnected system, and is increasing in complexity every day. Apart from the thousands of conventional generators, millions of smaller energy resources - solar panels, batteries, EVs, smart consumer loads - have to be controlled to ensure stable and reliable operation. The control action for each of these agents must be locally computed. How do you design control algorithms for a large networked systems with provable performance guarantee?
Buildings consume 75% of electricity and account for 33% of CO2 emissions in the USA. We are developing control algorithms to fix this problem: the algorithms will continuously change setpoints of heating, ventilation and air conditioning (HVAC) systems to reduce energy use while improving indoor climate.
Due to the thermal inertia of buildings, the electricity demand of their HVAC systems can be varied within limits without affecting their indoor climate. From the point of view of the power grid, this variation is the same as the charging and discharging of a battery. The same game can be played with almost every electric load. The resulting VES (virtual energy storage) potential of all these loads is huge. It is also a lot cheaper than a real battery of the same size.