Aiming at the problems of frequent voltage violations in distribution networks and difficulties in centralized optimal dispatch due to the significant uncertainties of distributed power sources and loads

this paper proposes a two-stage robust voltage/var control (VVC) strategy for distribution networks based on cluster partitioning

taking into account the adjustable reactive power resource allocation in distribution networks. Firstly

the modularity index of integrated electrical distance is constructed by considering the adjustable reactive power regulation capacity of the nodes

and the genetic algorithm adaptability function is designed to ensure the physical connectivity constraints between the nodes. Thus the cluster partition of the distribution network is completed. Then

the distribution network cluster VVC strategy is formulated by using the voltage deviation of each cluster as an index. Furthermore

a stochastic robust weighted set is developed to establish a second-order cone robust optimization model for multi-timescale cluster VVC. The model

aiming at minimizing distribution network losses and power purchasing cost from the main grid

is subsequently transformed and solved based on the strong duality theory and the column-and-constraint generation algorithm. Finally

simulations on the IEEE 33-bus distribution network demonstrate that the proposed cluster VVC strategy effectively mitigates node voltage violations

reduces the number of controlled devices

and enhances solution efficiency.