Privacy-Preserving Voltage Control via Aggregated Residential Flexibility and Feedback Optimization

This paper proposes a framework for real-time management of distributed energy resources in low-voltage networks. It is based on the combination of a decentralized predictive control strategy, enabling local cost minimization, and a centralized online feedback optimization, which adjusts the nodal setpoints to manage low-voltage networks. The proposed approach relies on nodal computation of available flexibility using multiparametric optimization, accounting for forecasts of load, photovoltaic generation, and electricity price. This results in control policies and cost functions that reflect the time-varying value of active and reactive power flexibility. At the central level, an online feedback controller optimizes power flows, requiring only real-time voltage measurements and aggregated local information. This eliminates the need to share complete cost functions, enhancing prosumers’ privacy. The resulting approach allows for cost-effective control of power systems by prioritizing low-cost flexibility. The method is tested on three low-voltage networks of 5, 20, and 43 buses by comparing it with existing relevant techniques.