Overview of Data-driven Power Flow Linearization

The accuracy limitation of physics-driven power flow linearization approaches and the widespread deployment of advanced metering infrastructure render data-driven power flow linearization (DPFL) methods a valuable alternative. While DPFL is still an emerging research topic, substantial studies have already been carried out in this area. However, a comprehensive overview and comparison of the available DPFL approaches are missing in the existing literature. This paper intends to close this gap and, therefore, provides a narrative overview of the current DPFL research. Both the challenges (including data-related and power-system-related issues) and methodologies (namely regression-based and tailored approaches) in DPFL studies are surveyed in this paper; numerous future research directions of DPFL analysis are discussed and summarized as well.