Wasserstein Metric‐Based Distributionally Robust Optimisation for Risk‐Aware Coordination of Virtual Power Plants in Power Distribution Systems With Uncertainties

ABSTRACT As distributed energy resources are deployed at scale, the resulting forecast errors and stochastic fluctuations increasingly challenge the economic performance and secure operation of both virtual power plants (VPPs) and distribution system operators (DSOs). This article develops a distributionally robust optimisation (DRO) based risk‐aware VPP‐DSO coordination framework. Initially, each VPP performs a dispatch and then submits price‐quantity offers up and down from its flexible resources to the DSO. Subsequently, to ensure distribution‐network security, the DSO solves a probabilistic optimal power flow formulated on an unbalanced three‐phase feeder (U‐OPF), aiming to maximise distribution‐level social welfare. The DSO adopts an improved Wasserstein metric‐based DRO technique to handle the uncertainty in locational marginal prices (LMPs) provided by the transmission operator. It also employs distributionally robust chance constraints (DR‐CCs) to manage day‐ahead uncertainties of flexible loads and photovoltaics (PV). At the DSO layer, the nonconvex U‐OPF is handled through a combination of an affine reserve decision rule and sensitivity‐based system responses. With this construction, the DR‐CCs admit a second‐order cone programming reformulation, enabling an efficient solution. Case studies on a modified IEEE 33‐bus feeder show that the proposed method reduces ex‐post balancing costs for PV/load deviations and achieves higher social welfare in worst‐case scenarios.