An optimization model for identifying backbone networks to enhance the resilience of distribution networks

The distribution network serves as the terminal network distributing electricity from the transmission network and generation systems to consumers. Enhancing the resilience of distribution networks is essential for improving the survivability of critical loads. However, strengthening the entire distribution network is challenging due to resource constraints and the complexities associated with various voltage levels, topologies, and diverse load types. An optimization model was proposed to identify the core backbone network within the distribution network, defined as the minimum subnetwork capable of serving the most critical loads under substation quantity constraints. The proposed approach offered an efficient and effective solution by taking account of load classification, network topology requirements, and substation capacities. It also incorporates the multi-line security principle, which enhances the topological robustness of the distribution network by increasing the power supply paths for critical loads. A linearized hierarchical decomposition strategy decoupled the NP-hard problem into 110kV and 220kV subproblems reducing computational complexity while ensuring global optimality. Additionally, the model integrated distributed generation (DG) into the backbone network to optimize joint scheduling, reducing the scale of backbone network while enhancing critical node redundancy. Case studies verified the minimum capacity threshold of backbone network for reliable class I and II load supply. The prioritization of peripheral loads is critical for backbone network identification, and DG integration reduces the demand for 110kV substations. Moreover, a hypothetical case study was conducted to demonstrate the importance of multi-line security principle for safeguarding critical loads during extreme events. The proposed method provides a scientific basis for emergency resource allocation during extreme events and can be integrated into power system planning models.