Abstract: To address the coordinated control challenges caused by load randomness,low PV accommodation rate,and voltage instability in 380V auxiliary AC power systems of substations with PV integration,this study proposes a PV-storage coordinated control strategy integrating enhanced DBSCAN dynamic scenario clustering and NSGA-III multi-objective optimization.A multi-dimensional dynamic feature system encompassing load mutation rate,meteorological correlation,and maintenance event markers is constructed.The DBSCAN algorithm is improved via hybrid distance metrics and parameter self-adaptation,increasing scenario recognition accuracy by 31.6%.A reliability optimization model prioritizing voltage qualification rate and critical load supply availability is established,embedding economic objectives including revised PV utilization rate (including storage consumption)and transformer load curtailment.Validation using operational data from a 220kV substation demonstrates:PV accommodation rate increases to 95.2%,voltage qualification rate reaches 99.6%,critical load availability achieves 100%,and transformer load curtailment attains 41.2%.This research provides a theoretical foundation for low-carbon substation retrofitting through dynamic scenario adaptation and multiobjective synergistic