Techno-Economic Capacity Screening of Battery–Flywheel Storage at a Metro Traction Substation with a Perfect-Information Linear Programming-Based Model Predictive Control (LP-MPC) Benchmark

Wayside energy storage can recover braking energy and moderate traction substation demand, but a capacity chosen under one dispatch policy may not retain its economic advantage under another. We examine battery–flywheel capacity screening using a processed, field-data-based one-second input from a winter week at a metro traction substation. A source-mode-constrained, perfect-information rolling linear program is compared with causal rule dispatch across the complete capacity grid. The evaluation distinguishes converter-side regenerative intake from retained energy and accounts for battery replacements at their assumed service life dates. The preferred configuration changes with controller restrictions, inventory screening and financial assumptions; additional flywheel capacity is not consistently valuable. Thus, a high energy capture rate or a favorable result for one controller does not by itself establish an investment case for hybrid storage. The contribution is an auditable screening workflow that exposes which operating and cost assumptions drive the ranking. Longer records, documented installed costs and a controller validated without future-load access are necessary before equipment selection.