Multi-Time-Scale Distributed Voltage Optimization for AC/DC Hybrid Distribution Networks with High-Penetration Photovoltaics

After high-penetration distributed photovoltaics (DPVs) are integrated into AC/DC hybrid distribution networks, stochastic source-load fluctuations, AC-DC coupling, and heterogeneous response characteristics of voltage-regulation devices jointly aggravate voltage violations and rapid voltage fluctuations. This paper proposes a spatio-temporal coordinated hierarchical distributed voltage-optimization framework. Spatially, the AC and DC regions are first separated according to voltage-source-converter (VSC) interfaces, and an electrical-coupling-aware modularity index is then constructed for the AC network by combining normalized bidirectional reactive-power-voltage sensitivities, available fast reactive-power support, and intra-cluster compactness. Temporally, an 1 h day-ahead model coordinates slow, discrete, or intertemporally coupled resources, including on-load tap changers, capacitor banks, energy storage systems, and flexible loads, while a 15 min intra-day rolling model coordinates DPV inverters, static var generators, and VSCs. The synchronous alternating direction method of multipliers (SADMM) is tailored to the resulting AC clusters, DC subnetworks, and VSC boundary variables to enable synchronous regional solution and boundary-consensus coordination. Studies on a modified 50-node AC/DC test system show that, under the investigated operating conditions, the framework mitigates voltage violations and intra-day fluctuations while obtaining favorable network-loss, DPV-curtailment, and solution-time performance.