Neural Two-Stage Stochastic Volt-VAR Optimization for Three-Phase Unbalanced Distribution Systems with Network Reconfiguration

TL;DR

The paper tackles voltage regulation and reactive-power management in distribution networks with high DER uncertainty by proposing a neural two-stage stochastic Volt-VAR optimization ($2S\text{-}VVO$) framework that accounts for network reconfiguration. It replaces the computationally intensive second-stage MILP with a learned surrogate embedded into a first-stage MILP, enabled by a topology-aware encoder (GCN) and a compact main evaluator (MLP). A voltage-violation duration constraint strengthens reliability under uncertainty, and an accelerated sampling strategy improves data generation efficiency. Empirical results on a 123-bus system show over $50\times$ speedup with an average optimality gap below $0.30\%$, demonstrating scalability to large scenario sets while maintaining solution quality. This approach enables real-time, large-scale stochastic VVO in practical distribution networks with reconfiguration capabilities.

Abstract

The increasing integration of intermittent distributed energy resources (DERs) has introduced significant variability in distribution networks, posing challenges to voltage regulation and reactive power management. This paper presents a novel neural two-stage stochastic Volt-VAR optimization (2S-VVO) method for three-phase unbalanced distribution systems considering network reconfiguration under uncertainty. To address the computational intractability associated with solving large-scale scenario-based 2S-VVO problems, a learning-based acceleration strategy is introduced, wherein the second-stage recourse model is approximated by a neural network. This neural approximation is embedded into the optimization model as a mixed-integer linear program (MILP), enabling effective enforcement of operational constraints related to the first-stage decisions. Numerical simulations on a 123-bus unbalanced distribution system demonstrate that the proposed approach achieves over 50 times speedup compared to conventional solvers and decomposition methods, while maintaining a typical optimality gap below 0.30%. These results underscore the method's efficacy and scalability in addressing large-scale stochastic VVO problems under practical operating conditions.

Figures (3)

• Figure 1: Architecture diagram of the proposed neural network.
• Figure 2: Diagram of the modified IEEE 123-bus system.
• Figure 3: Evolution of the optimality gap over computation time for the 123_Bus_FM case with 100 scenarios.