Fully parallel decentralized load restoration in coupled transmission and distribution system with soft open points

• Published in: Applied energy 2023-11, Vol.349, p.121626, Article 121626
• Main Authors: Zhang, Tao, Mu, Yunfei, Dong, Lei, Jia, Hongjie, Pu, Tianjiao, Wang, Xinying
• Format: Article
• Language: English
• Subjects: Analytical target cascading; Decentralized load restoration; Diagonal quadratic approximation; Soft open point; Transmission and distribution coordination
• ISSN: 0306-2619
• DOI: 10.1016/j.apenergy.2023.121626

Collaboration between transmission and distribution system operators (TSOs and DSOs) is crucial to improve power supply reliability during a massive blackout. However, these two systems usually operate separately in reality. To coordinate them, a new fully parallel decentralized TSO + DSO service restoration (D-TDSR) scheme is proposed in this paper. Based on analytical target cascading (ATC) method, the proposed D-TDSR is able to formulate the local restoration model for each TSO and DSO by decomposing shared boundary information in an iterative process. By introducing the diagonal quadratic approximation in ATC, all local restoration models are solved in a fully parallel manner with no need for a central coordinator. This parallel implementation increases the computation efficiency of decentralized restoration procedure. Furthermore, to enhance the ability of TSO and DSOs to support each other, the potential benefits of flexible regulation in DSO, including soft open points and network reconfiguration, are fully exploited, thus improving critical load restoration speed, and reducing voltage deviations. The effectiveness of the proposed D-TDSR is validated using an IEEE standard test system. •A decentralized load restoration for transmission and distribution system is proposed.•SOP flexibility is fully exploited to improve load restoration level and speed.•The fully parallel restoration procedure is proposed to reduce the computation time of each iteration.•The warm start mode is developed, which further accelerates the convergence speed of the restoration model.