Thermostats for the Smart Grid: Models, Benchmarks, and Insights

We model two existing thermostats and one novel thermostat to see how well they operate under dynamic pricing. The existing thermostats include a traditional thermostat with set temperature goals and a rigid thermostat that minimizes cost while always keeping temperature within a rigid predetermined range. We contrast both with a novel optimizing thermostat that finds the optimal trade-off between comfort and cost. We compare the thermostats' performance both theoretically and via numerical simulations. The simulations show that, under plausible assumptions, the optimizing thermostat's advantage is economically large. Importantly, the electricity demand of the rigid thermostat (but not the optimizing thermostat) ceases to respond to electricity prices on precisely the days when the electricity grid tends to be near capacity. These are the times when demand response is the most socially valuable to avoid massive price spikes. The social benefits of the optimizing thermostat may provide incentives for utilities and regulators to encourage its adoption.