Achievable Privacy in Aggregate Residential Energy Management Systems

Demand-side energy management systems consist of communication links between consumer homes and a centralized scheduler, whose main responsibility is to make decisions regarding curtailment, shifting, or modification of individual loads so that the aggregate load profile closely matches a predetermined supply function for a given time horizon. One challenge for aggregators using a centralized energy management system is the consumer's wariness of exposing usage data to third parties. Thus, the transmission of load requests from individual consumers to the centralized scheduler must support the flexibility of anonymizing usage data from potential eavesdroppers. This paper considers how load request times can be hidden from any third party entity by incorporating a packet scheduler that collects individual load requests within a consumer's home and reschedules the load request packets by inserting dummy packets to hide the true load request instances. The packet scheduler impacts performance of the aggregate energy management system and the consumer's quality of service by introducing a delay that helps distort true load request times. Furthermore, it impacts the design of the underlying communication infrastructure in terms of required bandwidth. The paper provides an analytical framework for quantifying the trade-off between the system metrics of privacy, bandwidth, and utility of the centralized scheduler when anonymity is enabled via delay and dummy packet insertions.