Demand Response Aggregation With Operating Envelope Based on Data-Driven State Estimation and Sensitivity Function Signals
Shuying Lai, Jing Qiu, Yuechuan Tao, Xianzhuo Sun
Abstract
With the increasing penetration of renewable energy sources (RES), the value of the demand response (DR) draws wide attention. In order to realize the coordinated dispatch of widely spread resources, the aggregation of the controllable residential loads is managed by a single entity, namely the DR aggregator. Under the price-based DR programs, the DR aggregators actively respond to the market signals to reach maximum welfare. To avoid the quality of electricity services being jeopardized, the operational constraints of the network should be considered by the DR aggregators. However, DR aggregators are not expected to have access to the monitoring equipment and have limited knowledge of the network states. Hence, in this paper, we proposed a DR aggregation with the operating envelope framework based on the representative signals produced by the distributed network operator (DNO) in the context of big data era. The DNO provides representative signals, including real-time state estimation and sensitivity functions, to the DR aggregators based on the proposed Semi-supervised Coupled Generative Adversarial Imputation Network (SC-GAIN) and big data analysis. The DR aggregators can realize the secure and efficient real-time dispatch of the controllable loads based on the received signals. The proposed framework was verified on the IEEE 33-bus and 123-bus systems. The case studies show that the proposed SC-GAIN algorithm can better deal with the missing data, and the learned sensitivity functions can effectively avoid the overestimation of the true DR potential.