Distribution network energy storage siting planning based on predictive load generation using Prophet and GAN

Authors

• Jiayu Xu, State Grid Beijing Electric Power Company ， Beijing 100031 ， China ;Economic and Technological Research Institute ， State Grid Beijing Electric Power Company ， Beijing 100055 ， ChinaFollow
• Xiaodong Chen, State Grid Beijing Electric Power Company ， Beijing 100031 ， China ;Economic and Technological Research Institute ， State Grid Beijing Electric Power Company ， Beijing 100055 ， China
• Yafeng Wang, State Grid Beijing Electric Power Company ， Beijing 100031 ， China ;Economic and Technological Research Institute ， State Grid Beijing Electric Power Company ， Beijing 100055 ， China
• Zhaoyan Liu, State Grid Beijing Electric Power Company ， Beijing 100031 ， China ;Economic and Technological Research Institute ， State Grid Beijing Electric Power Company ， Beijing 100055 ， China
• Yiyu Gong, State Grid Beijing Electric Power Company ， Beijing 100031 ， China ;Economic and Technological Research Institute ， State Grid Beijing Electric Power Company ， Beijing 100055 ， China
• Fanglan Liu, State Grid Beijing Electric Power Company ， Beijing 100031 ， China ;Economic and Technological Research Institute ， State Grid Beijing Electric Power Company ， Beijing 100055 ， China
• Jiaoxin Jia, Hebei Key Laboratory of Distributed Energy Storage and Microgrid ， North China Electric Power University ， Baoding 071003 ， China

Keywords

energy storage siting and sizing planning, Prophet algorithm, generative adversarial network (GAN), improved Bayesian optimization algorithm, nonlinear programming

Abstract

An energy storage siting and sizing planning method based on load trend extraction and forecasting is proposed to address the challenges of photovoltaic （PV） curtailment and bidirectional power flow overloads posed by distribution networks with high PV penetration.Firstly，the Prophet algorithm is adopted to decompose the net load of each node into trends and seasonal components，and the residual load is obtained.Secondly，a generative adversarial network （GAN ） is employed to model the relatively random residual component and generate load forecast data for the following year.The generated residual data are then combined with the trend and seasonal components to obtain a new annual net load that conforms to the original statistical patterns.Thirdly，based on the forecasted load data，an improved Bayesian optimization algorithm and nonlinear programming are adopted for energy storage siting and sizing to obtain the optimal energy storage configuration scheme.Finally，simulations are conducted based on real data from the following year to verify the predictability and economic efficiency of the proposed model.This study can provide reference for energy storage siting planning of distribution networks.

DOI

10.19781/j.issn.1673-9140.2026.04.023

First Page

278

Last Page

289

Recommended Citation

Xu, Jiayu; Chen, Xiaodong; Wang, Yafeng; Liu, Zhaoyan; Gong, Yiyu; Liu, Fanglan; and Jia, Jiaoxin (2026) "Distribution network energy storage siting planning based on predictive load generation using Prophet and GAN," Journal of Electric Power Science and Technology: Vol. 41: Iss. 4, Article 23.
DOI: 10.19781/j.issn.1673-9140.2026.04.023
Available at: https://jepst.researchcommons.org/journal/vol41/iss4/23