Open Access

ARTICLE

Integration of Flexible Interconnection Device in the Reconstruction of Medium and Low Voltage Distribution Networks Using DRL

Ruosong Hou^1,2,*, Jiakun An¹, Zihao Zhao¹, Wei Guo¹, Hua Shao²

1 Economic and Technological Research Inst., State Grid Hebei Electric Power Co., Ltd., Shijiazhuang, China
2 Hebei Huizhi Electric Power Engineering Design Co., Ltd., Shijiazhuang, China

* Corresponding Author: Ruosong Hou. Email:

Energy Engineering 2026, 123(8), 8 https://doi.org/10.32604/ee.2025.069068

Received 13 June 2025; Accepted 11 October 2025; Issue published 12 July 2026

Abstract

This article evaluates the connectivity with energy sharing in low-voltage distribution areas. Indicators like wind-solar complementing effectiveness, source-load energy sharing possibility, or transformer capacity interconnection measurements are part of the assessment index framework for interconnection capacity that is established after an analysis of the features of linked scenarios. Radial and inflexible, conventional distribution systems can’t handle bidirectional power flow, fluctuating demand, or grid disruptions. Using real-world examples, we can see that the suggested strategy improves power supply efficiency across zones and increases the usage of distributed energy resources, proving the method’s validity. With the help of Flexible Interconnection Devices (FIDs), MV/LV networks may be reconfigured, power quality is improved, DERs are supported, and reliability is increased. With so many distributed PVs connected to distribution substations, managing low- and medium-voltage distribution networks is a real challenge. One novel kind of power gadget that permits adaptable connections between distribution substation segments is the soft open point (SOP). This article presents learning algorithm for low and medium voltage networks for power optimization, which takes into consideration the dynamic connectivity of different areas of substation. The next stage is to construct a multi-agent deep reinforcement learning (DRL) suitable for low and medium voltage distribution networks using Deep Q Network (DQN) model in DRL. The low and medium voltage distribution network employs flexible interconnection device for power loss reduction. Finally, the case studies show that the proposed approach has a good operating strategy for distribution networks with medium and low voltages, and it may lessen voltage fluctuations caused by high PV integration.

---

Keywords

---