Stochastic control of cooling appliances under disturbances for primary frequency reserves

To stabilize power system frequency both in normal operation and after a contingency, a hierarchy of distributed-proportional and centralized-integral controllers is employed in most power systems. Providing proportional control-commonly referred to as primary frequency control or droop-with conventional generators constitutes significant cost. This led to strong interest in utilizing demand response as a cost-effective primary control reserve. This paper investigates an approach that allows a population of domestic refrigerators to reliably and continuously adjust its demand proportional to system frequency. The control relies exclusively on stochastic switching of refrigerators, and thus avoids synchronization and implementation issues associated with control strategies based on temperature-band adjustments. The scheme is tested on a realistic two-area power system model to investigate interaction with system frequency. The modeling of refrigerators is refined by analyzing the effect of door openings on cooling demand, and response of the refrigerator population under these conditions is described both analytically and with simulations.