Robust Joint Planning of Distributed Resources in Distribution Networks Based on Cluster Partitioning
The integration of high-penetration distributed photovoltaics (DPV) into distribution networks introduces significant operational uncertainties and challenges in maintaining voltage profiles and reliability. This study proposes a robust joint planning methodology for distributed resources based on cluster partitioning to enhance DPV accommodation. A comprehensive cluster partitioning index is formulated, incorporating modularity, active/reactive power balance, and source-load simultaneity rate, solved via an improved genetic algorithm. Subsequently, a bi-level robust joint planning model is established. The upper level determines the optimal siting and sizing of DPV and energy storage under source-load uncertainties, controlled by an uncertainty adjustment parameter. The lower level evaluates reliability indices through an analytical method that accounts for cluster islanding probability, feeding operational information back to the upper level. Iterative optimization balances robustness and reliability. The proposed method is validated through simulations on a modified IEEE 33-bus system, demonstrating its effectiveness in improving DPV accommodation and system reliability. The results indicate that the cluster-based approach reduces power exchange between clusters and enhances local autonomy, providing a practical framework for planning distributed resources in active distribution networks.