Real-time demand response potential evaluation: A smart meter driven method

Abstract

This paper proposes a real-time demand response potential evaluation (RTDRPE) method that is purely driven by smart meter data. Probabilistic behaviors of electricity consumers are learned from historical data based on Gaussian Mixture Model (GMM). Then, with GMM of each consumer, RTDRPE is conducted on both the individual level and the aggregated level, which facilitates DR-related and market-related decision making for DR implementers in a real-time manner. Case studies using real-world data demonstrate the efficacy of the proposed method.