Annual energy margin
—Delivery potential, not hourly serviceFrom one machine to a national system
What would it take
to power America?
Build an electricity portfolio. Test its annual energy, then ask whether it can meet a difficult evening. Real energy sources supply the power; storage moves it in time.
Peak + reserve margin
—A selected-duration stress testIllustrative annual cost
—Capital recovery + operations + fuelReliability verdict
UnverifiedAnnual totals cannot establish 24/7 service.New: lower-loss recovery hub
Use the six-paper research brief to compare energy paths at the same delivered load. Keep conversion savings separate from grid power displaced by local sources.
01Define the demand
Starting point: 2024 U.S. grid electricity sales. The peak and reserve settings below are assumptions, not measured national peaks.
Boundary: grid-delivered electricity only. A further 135 TWh of 2024 on-site use is outside this portfolio and assumed to continue separately. This does not cover all fuel used for transport, heating, or industry.
02Choose the generation
An illustrative future portfolio—not today's U.S. fleet. Capacity factor is the annual average output fraction. Peak contribution is a separate, assumed fraction available during the stress event.
| Energy source | Capacity GW | Annual factor % | Peak contribution % | Annual energy TWh |
|---|
Default evening test: solar contributes 0%; wind 5%. Nuclear, hydro, and fuel-backed availability are also assumptions—not accredited firm capacity. Fuel-backed generation requires a continuing fuel supply and has emissions.
03Shift energy with storage
Batteries do not generate new energy. Their losses increase the generation needed, and their stored energy limits how long they can help.
Usable capacity is AC energy delivered by the battery. The peak screen assumes the chosen starting charge; it does not establish that charging is available before the event. Charging and discharging share the selected power rating.
04Keep the economics honest
Every cost below is an editable placeholder in constant U.S. dollars (USD)—not a researched supplier quote, a project budget, or a profit forecast. Full modeled generation costs are retained; surplus energy earns no assumed revenue.
Edit construction, operating, and finance assumptions
| Source | Construction $/kW | Fixed operations $/kW-year | Variable + fuel $/MWh | Cost recovery years |
|---|
Battery power and energy costs are complementary components, not two all-in battery prices. Battery life has its own capital recovery period; costs of replacements or degradation beyond that simplified annualization are not simulated. Grid investment and operating costs are placeholders, not a transmission plan.
How to use this
1. Change demand or the generation mix. 2. Compare the two margins. 3. Try the 72-hour stress test. 4. Review costs and the limits of the model before drawing conclusions.
GW measures power right now. GWh measures stored energy. TWh measures energy across a year. A positive energy margin does not mean electricity is available at every hour. Changes last until you reset or reload; download a report to keep your assumptions.
Sources, assumptions, and model boundary
- EIA Table 2.2 (2024): 3,975.381832 TWh retail sales; 135.040521 TWh direct use. Grid sales form the editable demand baseline.
- EIA Table 4.08.B (2024): default annual factors of solar 23.2%, wind 34.3%, nuclear 90.8%, hydro 34.6%. These historical fleet averages are not future predictions.
- EIA transmission and distribution losses: about 5% in 2018–2022. Used as an editable planning assumption, not a measured 2024 loss rate.
- All plant capacities, peak contributions, peak load, reserve, storage specifications, fuel-backed annual factor, and costs are illustrative inputs. A standardized 8,760-hour year is used, not the 8,784-hour leap-year calendar.
- The portfolio is domestic and aggregated; imports, regional deliverability, seasonal dispatch, future load forecasts, construction timelines, land, water, emissions pricing, and permitting are not calculated.
- Annual capacity factors are not peak capacity credits. E3 / NERC: practical application of effective load-carrying capability describes probabilistic capacity evaluation. The peak contributions here are user assumptions, not calculated ELCC.
Show the calculation rules
- Annual generation (TWh) = capacity (GW) × annual factor × 8.76. Fractions use 0–1, not 0–100.
- Battery discharge = usable GWh × annual cycles / 1,000. Charging = discharge / round-trip efficiency. Storage loss = charging − discharge. Battery discharge is never added as new generation.
- Annual delivery potential = (generation − storage loss) × (1 − grid-loss fraction). Annual energy margin = this potential − annual demand. Actual hourly load served is not calculated.
- Battery stress contribution = min(power GW, usable GWh × starting charge fraction / stress hours). Peak contribution after grid losses = (sum of generator GW × peak fractions + battery contribution) × (1 − grid-loss fraction).
- Peak target = customer peak × (1 + reserve fraction). The chosen peak must be at least the annual average load. The peak screen assumes the same available generator power throughout the event.
- Battery charging must not exceed annual generation; charging hours + discharge hours must not exceed 8,760 at the selected power. These are necessary checks only, not proof that energy is available at the right times.
- Annual cost = capital × capital recovery factor for each asset + operating and fuel costs. CRF = r / [1 − (1 + r)−n]; at zero discount rate it is 1/n. Energy-accounted demand = min(annual demand, nonnegative delivery potential). The displayed $/MWh uses this annual-only denominator, which can overstate service and understate a reliability-adjusted cost.
Next engineering gate: regional, hourly demand and weather; validated storage dispatch; outage and fuel scenarios; grid stability and transmission studies; then a site-specific cost and permitting assessment. This planner is a conversation starter, not an investment recommendation.