Optimal Operation Study of Distribution Network Based on Flexible Multi-State Switch and Shared Energy Storage

The distribution network requires additional flexibility to cope with the large-scale integration of distributed energy sources. Energy Storage Systems (ESS) can smooth the fluctuating output of renewable energy. However, due to high investment and maintenance costs, equipping multiple ESS units within a single system is not practical. To address these challenges, this paper proposes a distribution network optimization scheme based on a shared energy storage strategy.In this scheme, a Flexible Multi-state Switch (FMSS) with integrated energy storage systems is introduced at the endpoints of the distribution network. This enables the sharing of charge and discharge energy between nodes. The optimization of distribution network voltage operation is modeled as a Markov Decision Process (MDP) framework. Additionally, a Q-depth neural network algorithm is proposed to effectively address the issue of high-dimensional decision variables, aiming to determine the optimal charging and discharging actions of battery energy storage systems. The proposed approach is validated using an enhanced IEEE 33-node system as a case study, demonstrating the method’s robustness and effectiveness in reducing power losses and mitigating voltage fluctuations.