Under the impetus of large-scale distributed generator (DG) development

the deployment scope of DGs in distribution networks continues to expand

gradually establishing them as critical components of multi-source power supply systems. Meanwhile

high DG penetration enhances system flexibility and reliability; however

traditional current protection schemes encounter challenges such as protection miscoordination

maloperation

and failure-to-operate

due to DG output fluctuations and wide-range load variations. To address these issues

this study proposes a K-means++ clustering algorithm that categorizes the operating states of the distribution network into multiple typical scenario sets based on load demands

DG output status

and network topology. For each identified scenario

refined protection settings are calculated offline

taking into account DG topological connections and output ranges. Moreover

the state perception technology is employed to monitor the real-time DG output and network topology

enabling dynamic matching of operating scenarios and switching of setting groups

thereby achieving adaptive optimization of protection configurations. Simulation results by using PSCAD/EMTDC demonstrate that the proposed scheme enhances protection reliability and effectively resolves protection miscoordination issues under high DG penetration. Through the coordinated mechanism of “scenario clustering-offline setting-dynamic switching”

this method provides a systematic solution for protection configuration in active distribution networks.