Flexible urban energy

Schematic for an integrated district energy system

As they become even more interdependent with the electric sector, urban energy systems will need to become more integrated and flexible. We will need to pay close attention to the way they operate. Decarbonization of electricity generation together with electrification of energy-and-carbon intensive services such as heating, cooling and transportation is needed to address ambitious climate goals. The Stanford campus district energy system (Stanford Energy Systems Innovations project; SESI) is roughly equivalent to a city of population 30,000 and provides a unique source of real data as well as an ideal test-bed for new ideas and control algorithms. We explored whether city-scale electrification of heat with large-scale thermal storage also cost-effectively unlocks operational benefits for the power sector. We built an optimization model of fully electrified district heating and cooling networks integrated with other electric loads and leveraged real-world consumption and operational data from the campus. Using our modeling approach, we computed optimal operational strategies for the controllable loads and thermal storage in this system under different economic hypotheses such as least-cost scheduling and carbon-aware scheduling, that takes advantage of variations in power grid carbon intensity. We also explored the interactions between urban energy systems and large-scale electric vehicle charging. Our modeling efforts on capacity-based demand response led to a megawatt-scale experiment in the summer of 2018, when the campus participated in one of Pacific Gas & Electric’s demand response programs. During this experiment, the campus energy operations were significantly modified to provide 5 MW load drops during demand response events. Revenue for the summer was approximately 300,000$. Scientific contributions included approaches to (i) determine how much capacity to nominate each month (the planning problem) and (ii) adjust hourly operations schedules in real time, while preparing for possible events (the control problem). The experiments were possible thanks to the generous involvement of the campus’ team of five control room operators (as well as half a dozen SESI engineering staff). In the context of these experiments, we built a prototype software platform to (i) ingest live energy, weather and price data from the campus data historians; (ii) compute optimal operational strategies and (iii) visually communicate these strategies to inform control room operators.