DMPC-Based Two-Level Cross-Regional Optimization for Power Systems with High Renewable Penetration

High renewable integration promotes cross-regional power exchanges to enhance operational flexibility, but also propagates significant uncertainty across interconnected areas. To address this challenge, this paper proposes a two-level distributed model predictive control framework composed of an upper-level coordinator and lower-level regional operators. At the upper level, the coordinator optimizes tie-line flows and reserve allocations through shadow-price signals to mitigate congestion. Inter-regional consensus is enforced through the alternating direction method of multipliers, which ensures rapid convergence and protects data privacy. At the lower level, each regional operator solves a receding-horizon model predictive control problem subject to DC network limits, unit and ramping constraints, and optional storage dynamics. Boundary tie-line references from the coordinator are tracked, while renewable uncertainty is hedged by embedding a conditional value-at-risk measure into the rolling horizon optimization. This formulation yields a tractable second-order cone program that can be updated with new forecasts at each step. Case studies on the IEEE 118-bus system demonstrate that the proposed framework enhances renewable utilization, lowers cross-regional exchange costs, and strengthens overall operational security, offering a practical pathway for secure and economical coordination under high renewable penetration.