Voltage Regulation and Loss Minimization of Active Distribution Networks With Uncertainties Using Chance-Constrained Model Predictive Control

The high integration of photovoltaics (PVs) and electric vehicles (EVs) introduces significant uncertainty to distribution networks (DNs), leading to frequent and uncertain voltage fluctuations. The active/reactive power from photovoltaic units and EV charging stations can effectively manage real-time voltage control through multi-time scale coordination. This paper proposes a stochastic real-time control model based on chance-constrained model predictive control (CC-MPC) for coordinated voltage control. The proposed model adapts multi-time scale coordination among control devices, encompassing PVs, EVs, and on-load tap changer (OLTC), using a multi-step optimization model. A scenario-based approach is used to account for the nodal power uncertainties (with different levels of uncertainties for conventional loads, EVs, and renewables), using receding horizon control while ensuring network security constraints. The computational efficiency is increased by employing the backward scenario reduction technique while maintaining the solution accuracy using a pre-defined confidence parameter. Detailed case studies are carried out using 33-bus network and IEEE-123 bus distribution network, to validate the efficacy and scalability of the proposed model. The comparison with a deterministic MPC-based control framework validates the effectiveness of uncertainty handling and control cost reduction.