AUTHOR=Yu Yixiao , Yang Ming , Zhang Yumin , Ye Pingfeng , Ji Xingquan , Li Jingrui 

TITLE=Fast reconfiguration method of low-carbon distribution network based on convolutional neural network

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 11 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2023.1102949

DOI=10.3389/fenrg.2023.1102949

ISSN=2296-598X

ABSTRACT=The existing meta-heuristic distribution network reconfiguration (DNR) algorithm has excellent optimization ability. However, it is difficult to realize the large-scale fast computing and online real-time response of DNR solutions since the iterative optimization method. In order to effectively mine the distribution network (DN) historical data and improve DN security and low-carbon economic operation, this paper proposes a fast distribution network reconfiguration algorithm based on the convolutional neural network (CNN). Firstly, a multi-branch CNN model based on the loop structure of DN is established, which can reduce the dependence on the concrete structure of DN in the process of modeling, and easy to expand the model. Secondly, the CNN model is trained based on the hybrid training method, the correlation between the representative load mode (LM) of the DN and its optimized topology can be learned, and the mapping relationship between the two is formed, so that the model has the ability to quickly reconfigure the DN with different LMs. Finally, the performance of the model is analyzed on IEEE 33-bus system to verify the effectiveness of the proposed method.