Energy in motion for a more resilient grid

Inside the machine

The reference design places permanent magnets on the rotor and copper windings on a stationary stator. A rectifier, DC link and inverter adapt the electrical output to the load. Vacuum can reduce aerodynamic drag, but it complicates heat removal, sealing and service access. The optimum pressure is the one that improves the whole system, rather than simply the lowest achievable pressure. ^{[1], [3]}

A small example makes the scale visible

An ideal thin ring of 10 kg and radius 0.20 m has J = 0.40 kg·m². At 3,000 rpm it contains about 5.48 Wh, equivalent to 19.7 seconds at 1 kW only if all its kinetic energy could be recovered without losses. A real machine retains a minimum operating speed, so less energy is available. This example is unrelated to the grid-scale cost scenario on the next page. [1; recalculated here]

J is rotational inertia; ω is angular speed in rad/s. Delivered electrical energy is lower after conversion, standby and auxiliary losses.

Power (kW or MW) describes the delivery rate. Energy (kWh or MWh) describes the amount available. A high power rating alone does not establish a long discharge duration.

The cost of using it

A storage project must pay for the energy it receives and for the equipment that returns it. Compare alternatives at the same power, useful duration, response requirement, waiting time and annual duty. DOE's storage-cost methodology includes charging and replacement costs. Its assessment provides a framework, not a quotation or a validation of this machine. ^[4]

Illustrative scenario from the paper

1 MW output; 0.25 MWh delivered per full discharge (15 minutes); €800,000 installed capital; €20,000 annual routine maintenance; 20 years; 7% discount rate; €80/MWh charging-input cost; 85% assumed AC-to-AC cycle efficiency, including auxiliaries for the specified duty. These are modeling assumptions, not measured performance or a supplier offer. ^[1]

Calculated with unrounded values: capital recovery factor 0.0943929; annual fixed cost €95,514.34. Q = 0.25 × 365 × c. Indicative unit cost = €95,514.34 / Q + €80 / 0.85.

With these assumptions, falling below €150/MWh requires about 18.73 full cycles per day. At 24 cycles per day the model implies 175,200 full cycles over 20 years, before downtime. That duty, lifetime and charging schedule must be demonstrated. Doubling capital to €1.6 million raises the 24-cycle estimate to about €172/MWh. Grid connection, taxes, exceptional repairs and end-of-life costs remain outside this simplified case. Efficiency must be measured again for each waiting-time profile.

Use market comparisons carefully

IRENA reports global weighted-average generation costs for new 2025 projects of USD 33/MWh for onshore wind and USD 44/MWh for solar PV, in 2025 dollars. These describe generation, not the delivered cost of this storage system. EIA's AEO2026 estimates concern US resources entering service in 2031 under stated policy and financing assumptions. Neither is a local tariff, a guaranteed price or a single ranking of interchangeable services. ^{[5], [6]}

What would count as evidence?

A credible next step is a contained laboratory demonstrator, evaluated by qualified engineers before a field pilot. Define the service and the pass/fail criteria before seeing the results. The supplied paper presents a design assessment and scenarios; it does not supply independent prototype measurements proving net savings or commercial viability. ^[1]

1. Define the boundary

   Record every external electrical and mechanical input, useful output and auxiliary load. Include batteries, the DC bus and initial rotor energy.

2. Measure synchronized data

   Log active power, shaft torque, speed, pressure and temperature together. State calibration, sampling, integration method and uncertainty.

3. Compare air and vacuum

   Use matched load and speed trajectories. Measure pump-down and standby demand, and compare a conventional machine on the same useful task.

4. Close the cycle

   Start and end at matched speed and other stored-energy states, or correct explicitly for their changes. Integrate energy over charging, waiting and discharge.

5. Publish a decision

   Require a repeatable net benefit larger than measurement uncertainty, acceptable temperatures and containment, and a competitive lifecycle cost for the defined service.

For an electrical-only storage cycle at matched internal states: E_in = E_out + E_losses, with auxiliaries counted once. Report both AC-to-AC cycle efficiency and standby loss. Generator shaft-to-AC efficiency is a different boundary and must be identified separately.

From an Italian test bench to a worldwide discussion

A pilot near a variable industrial load or solar plant can make the service measurable. PVGIS can support solar-resource assessment where relevant; measured load data and local network requirements complete the picture. Grid connection and market access need current, jurisdiction-specific approval. ^[7]

Environmental comparison should include manufacturing, magnet sourcing, lifetime, repairs, containment and the electricity used for charging and auxiliaries. A flywheel has mechanical failure risks and ongoing losses. Batteries, supercapacitors and other storage options should be compared on the same service. More renewable generation and effective storage perform complementary roles.

I welcome critical review from researchers, engineers, students and anyone interested in resilient energy. The contribution I want to make is practical: a transparent hypothesis, reproducible measurements and a decision that others can check. Shared knowledge and care for Mother Earth belong together.

Sources, methods and edition notes

Selected primary sources supporting this article. The original paper contains a broader 25-item bibliography. Links were checked on 21 October 2026. External sources describe established principles or other projects; they do not endorse this proposed machine.

1. Giuseppe Carilli · Electromagnetic machine with a vacuum enclosed rotorAuthor-supplied technical paper · 22 pages · IT / EN / ES
2. U.S. Department of Energy · Grid-Scale Flywheel Energy Storage PlantProject fact sheet · October 2012
3. NASA · 7.0 Thermal ControlSmall Spacecraft Technology State of the Art · 2026
4. U.S. Department of Energy · 2022 Grid Energy Storage Technology Cost and Performance AssessmentStorage methodology and lifecycle cost boundaries · 2022
5. IRENA · Renewable Power Generation Costs in 2025: Executive Summary2026 · 2025 USD · pp. 6-7
6. U.S. Energy Information Administration · Levelized Costs of New Generation Resources in AEO2026April 2026 · 2031 online year · pp. 2-3, 6 and 8
7. European Commission, Joint Research Centre · Photovoltaic Geographical Information System (PVGIS)Site-specific solar-resource tools

What this edition clarifies

• Adds the usable-energy relation between maximum and minimum speed, and separates the small-ring example from the hypothetical 1 MW scenario.
• Recomputes the financial example, identifies excluded costs and makes the high-cycle lifetime requirement explicit.
• Separates documented technology, design choices, assumed economics and evidence still required. It does not claim a new physical principle, a validated prototype, peer review or commercial readiness.

Equations and scenario results were recalculated from the paper's stated inputs. No experimental dataset was supplied or generated for this edition. This is an editorial companion article, not a replacement for the original technical paper.