(672d) Grid-Responsive Smart Manufacturing with High-Performance Flywheel Energy Storage

The industrial sector consumes 26% of electricity in the United States but has historically deployed only 7% of the country’s battery energy storage. One of the main barriers to adoption has been the poor return on investment (ROI) for conventional storage technologies in industrial contexts, particularly due to limited revenue streams and the rapid degradation of chemical batteries. This work demonstrated a novel approach to industrial energy storage that significantly improved ROI by combining long-lifetime flywheel storage with a control strategy that stacked multiple value streams.

A flywheel energy storage system was modeled for a rural advanced manufacturing facility. The system was operated by an intelligent automation platform that responded to peak demand strategies and real-time signals from the grid. A machine-learning-based controller dynamically balanced five concurrent use cases: (1) real-time utility dispatch, (2) advanced-notice dispatch, (3) facility peak demand reduction, (4) load shifting under time-of-use pricing, and (5) process stability improvements during voltage sags and short outages.

Flywheel energy storage was selected for its fast response rate, minimal degradation, and high round-trip efficiency. Unlike chemical batteries, the flywheel system retained near-100% capacity over time, functioned reliably in all ambient temperatures, and demonstrated a 25-year operational life. Interval data collected over a one-year period confirmed that the system successfully performed peak shaving, energy arbitrage, and demand response participation. Collectively, these operations made the storage system economically viable, an accomplishment not currently attainable to chemical batteries.

The deployment also improved process reliability by eliminating power-related disruptions, which had previously caused costly production upsets. The flywheel provided seamless backup power during grid disturbances, and its integration with automation and forecasting systems allowed for predictive adjustments to manufacturing operations.

This project provided one of the first large-scale demonstrations of long-duration, mechanically-based energy storage for real industrial load data. It showed that a single storage asset could deliver grid services, cost savings, and improved operational resilience with intelligent control—benefits that have historically required separate investments. The system design and control methods were also highly modular, enabling replication across diverse manufacturing sectors. By overcoming economic and technical barriers, this work demonstrated the viability of flywheel storage for industrial decarbonization and flexibility, providing a roadmap for broader adoption of long-duration storage technologies in the sector.