Detection of Defaulting Participants of Demand Response Based on Sparse Reconstruction

In demand response (DR) based on contracts with consumers, some participants have the potential to default on providing their scheduled negawatt energy due to demand-side fluctuations. Thus, the detection of defaulting participants is an important function of the aggregator. In particular, it is preferable to detect them with limited information and not by real-time continuous metering because of communication costs and social acceptance. This paper addresses the problem of detecting defaulting participants in contract-based DR, provided that the aggregator can inspect the total negawatt energy and the negawatt energies of a limited number of participants via smart meters. By focusing on the property that only a few participants are defaulting under their contracts, we propose a detection method based on sparse reconstruction, i.e., reconstructing a sparse vector from a small number of scalar equations. The proposed method is iterative, and each iteration improves the sparse reconstruction by including the inspection data from the previous iteration. It is theoretically guaranteed that the proposed method derives the exact solution under practical conditions. Finally, direct load control is incorporated into the detection method to eliminate defaulting participants.