Enhancing Power Systems Transmission Adequacy via Optimal BESS Siting and Sizing using Benders Decomposition with Feasibility Cuts
New algorithm solves complex grid storage planning 10x faster with rigorous convergence guarantees.
A team from EPFL (École Polytechnique Fédérale de Lausanne) has published a significant paper on arXiv detailing a novel computational framework for strategically placing and sizing battery energy storage systems (BESS) within power transmission grids. The work, led by Ginevra Larroux, Matthieu Jacobs, Keyu Jia, Fabrizio Sossan, and Mario Paolone, tackles the critical challenge of enhancing grid 'resource adequacy'—ensuring the system can meet demand under stress. Their approach is operationally driven, considering multi-year planning horizons and enforcing strict network constraints like line capacities and voltage limits at high temporal resolution, which is essential for integrating variable renewable energy sources.
The core innovation is reformulating the computationally monstrous mixed-integer non-linear programming (MINLP) problem into a more tractable mixed-integer second-order cone programming (MISOCP) model. This is then solved using Generalized Benders Decomposition, enhanced with 'feasibility cuts' that proactively manage grid congestion and voltage issues. A final heuristic recovers a fully AC power-flow-feasible operating plan from the relaxed solution. The framework is designed for parallel computation, offering excellent performance and provable convergence guarantees for massive, real-world meshed networks with thousands of nodes (demonstrated on a 6,428-bus system). This provides grid planners with a powerful, scalable tool to determine where and how large to build grid batteries for maximum reliability and efficiency.
- Solves complex BESS placement (siting & sizing) as a mixed-integer second-order cone program using Generalized Benders Decomposition with feasibility cuts.
- Ensures AC power-flow feasibility for large-scale meshed grids, managing congestion and voltage regulation under binding network limits.
- Parallelizable framework demonstrates excellent computational performance and rigorous convergence guarantees, enabling planning for systems with over 6,000 nodes.
Why It Matters
Provides grid operators a scalable tool to optimize billion-dollar battery investments, boosting grid resilience and renewable integration.