Data-driven and physically informed power grid dispatch decision-making method

doi:10.1016/j.segan.2025.101644

Data-driven and physically informed power grid dispatch decision-making method

2025-06-01

0

PRE

AI

DOI：10.1016/j.segan.2025.101644

Original

Original request for help

Save

Abstract

En 中文

This paper introduces an innovative approach, namely the Action Generation Network (AG-Net), designed for power system Security Constrained Economic Dispatch (SCED). In contrast to purely data-driven methodologies, our proposal incorporates a Physical Information Judgment Network (PIJ-Net), effectively integrating essential physical information into the model. This strategy simplifies the economic dispatch model's intricacies while facilitating the network's grasp of the model's underlying physical dynamics. The collaborative operation of these two networks is geared towards achieving highly accurate decision-making. Notably, experimental evaluations conducted on the SG-126 bus system demonstrate that our proposed method surpasses both model-based and neural network relaxed solutions. The results highlight the method's capacity to deliver more dependable and efficient dispatch decisions. This underscores the significance of marrying data-driven approaches with physical insights for enhanced performance in power system economic dispatch.

Keywords:

Physical information

Neural networks

SCED

Data-driven and physically informed power grid dispatch decision-making method

Abstract

Journal

Organization