Dashboard › Deep Learning › Power Systems Analysis › Energy storage › Long-duration energy storage — pumped hydro, CAES, flow, iron-air, thermal

Long-duration energy storage — pumped hydro, CAES, flow, iron-air, thermal

LDES = >10 hr. Pumped hydro 95% of global (180 GW, RT 70-85%, 50-100 yr, geography-limited). Flywheel sub-sec to 30 min (E=½·J·ω², niche). CAES (Huntorf 1978, McIntosh 1991; adiabatic emerging). Flow batteries decouple P and E (VRFB, all-iron ESS). Iron-air Form Energy 100 hr at $20/kWh target. Thermal molten salt at CSP. DOE LDES Storage Shot: 90% cost cut by 2030. WoodMac: 1500 GW LDES by 2050.

Senior ~16 min

Step 1 — Why long-duration energy storage (LDES) — beyond 4-hour batteries

Animation speed 0.55×

tech — duration — round-trip —

Reference notes

As renewables approach 80-100% grid share, the residual demand has multi-day low-renewable periods that lithium-ion cannot economically cover. Long-duration energy storage (LDES, > 10 hours) requires fundamentally different technologies: pumped hydro, CAES, flow batteries, iron-air, thermal, hydrogen. The DOE Long-Duration Storage Shot (2021) targets 90% cost reduction by 2030 ($50/kWh delivered).

Duration regimes

Regime	Duration	Dominant technology
Fast frequency response	sub-second to seconds	Flywheel, supercapacitor, fast battery
Short-duration	0.5-4 hr	Lithium-ion (daily arbitrage, regulation, peak-shaving)
Medium-duration	4-10 hr	Lithium-ion + emerging flow batteries
Long-duration (LDES)	> 10 hr	Iron-air, flow, CAES, thermal — the LDES challenge
Very-long / seasonal	> 100 hr	Hydrogen, large pumped hydro, salt caverns

Technology comparison

Technology	Duration	Round-trip	Life	Notes
Pumped Hydro (PHS)	4-24 hr	70-85%	50-100 yr	95% of global storage; geography-limited
Flywheel	sec-30 min	80-90%	100k+ cycles	High power; freq regulation niche; not bulk energy
CAES (diabatic)	4-24 hr	42-55%	40+ yr	Huntorf 1978, McIntosh 1991; salt cavern + gas reheat
CAES (adiabatic)	4-24 hr	60-72%	~30 yr	Captures compression heat; no gas needed; emerging
Vanadium flow (VRFB)	4-12 hr	70-80%	20k+ cycles	Decoupled P/E scaling; Sumitomo, Invinity, Largo
All-iron flow	8-12 hr	60-70%	20k+ cycles	ESS Inc.; cheaper than vanadium
Iron-air (Form Energy)	100 hr	~50%	15+ yr	Rust chemistry; $20/kWh target; multi-day
Thermal (molten salt)	10-15 hr	40-50%	30+ yr	CSP plants; Crescent Dunes 110 MW
LAES (liquid air)	6-24 hr	50-60%	~30 yr	Cryogenic; Highview 50 MW Manchester UK
Hydrogen	weeks	30-45%	20+ yr	Seasonal; lowest cost/kWh long-term
Li-ion (reference)	0.5-4 hr	85-95%	3-6k cycles	Short-duration baseline; not LDES

Pumped Hydro Storage (PHS) — 95% of global storage

Two reservoirs at different elevations; reversible Francis turbine + motor-generator pumps UP (low-price hours) and releases DOWN (high-price hours).
Power output P = ρ·g·H·Q·η (water density, gravity, head, flow rate, efficiency).
Typical: 4-24 hr duration, 100 MW to 3 GW power. Examples: Bath County VA (3 GW, 31 GWh), Ludington MI, Goldisthal Germany, Yangjiang China.
180 GW installed globally. China target 120 GW by 2025.
Variable-speed PHS (Goldisthal, Linthal, Tehri): DFIG / full-converter motor-generators provide reactive support + frequency response while pumping.
Constraints: 100-1000 m elevation difference; 10-15 yr permitting; large environmental impact. Closed-loop PHS (off-river) reduces impact.

Flywheel — fast power, not bulk energy

Energy E = ½·J·ω²; modern flywheels spin 10,000-50,000 RPM.
Architecture: carbon-fiber composite rotor + magnetic-levitation bearings + vacuum chamber + integrated PM motor-generator.
100 kW-2 MW power, 15-30 min duration, sub-second response, 100,000+ cycle life.
Niche: frequency regulation (Beacon Power 20 MW Stephentown NY), data-center UPS, metro regenerative braking.
Suppliers: Beacon Power, Amber Kinetics, Active Power.
NOT bulk energy storage — high cost per kWh.

Compressed Air Energy Storage (CAES)

Compresses air into underground cavern (salt dome) at 1-8 MPa during charge.
Expands compressed air through turbine to generate during discharge.
Two legacy plants:
- Huntorf Germany (1978) — 290 MW / 2 hr discharge, still operating.
- McIntosh Alabama (1991) — 110 MW / 26 hr discharge.
Diabatic CAES: dumps compression heat, burns natural gas to reheat air at discharge. RT 42-55%.
Adiabatic CAES: captures and reuses compression heat. RT 60-72%. Emerging (China Zhangjiakou 100 MW; ADELE paused).
Isothermal CAES: research stage (Lightsail defunct).
Geography-limited (salt domes, abandoned mines).
New projects: Australia (Hydrostor Strathalbyn), Northwestern US, China.

Flow batteries

Key advantage: DECOUPLED scaling of power (cell stack) and energy (electrolyte tank). Adding duration just means larger tanks — cheap.
Vanadium Redox (VRFB): most mature. V²⁺/V³⁺/V⁴⁺/V⁵⁺ chemistry avoids cross-contamination. 4-12 hr duration, RT 70-80%, 20,000+ cycles. Suppliers: Sumitomo Electric, Invinity, Largo Clean Energy.
Zinc-bromine: lower cost; cycle life and dendrite formation challenges. ESS Inc., Redflow.
All-iron flow: ESS Inc. commercial product. Iron-based electrolyte cheaper than vanadium. 8-12 hr.
Other chemistries in development: polysulfide-bromine, organic-aqueous, semi-solid lithium.

Iron-air battery (Form Energy)

NOT a flow battery — different architecture.
Iron electrode OXIDIZES (rusts) during discharge: Fe + O₂ → Fe₂O₃ + electrons. Air provides oxygen.
During charge: reduces back to iron — Fe₂O₃ + electrons → Fe + O₂ released.
Form Energy commercial product targets 100-hour duration at ~$20/kWh stored energy.
Round-trip efficiency ~50% — economical only with very-low-cost charging energy.
Pilots: Great River Energy MN (1 MW / 150 MWh, 2023), Xcel Energy Colorado (10 MW / 1000 MWh planned 2025), Georgia Power, Dominion Energy Virginia.
Critical for high-renewable grids: charges during multi-day windy/sunny stretches; discharges during multi-day calm/cloudy stretches.

Thermal energy storage

Molten salt (60% NaNO₃ + 40% KNO₃ at 565 °C) at CSP plants: Crescent Dunes (NV, 110 MW, 10 hr), Cerro Dominador (Chile), Noor (Morocco). 30+ year asset life.
Rock beds / sand / ceramic at 300-600 °C — cheaper. Companies: Antora Energy, Brenmiller Energy, MGA Thermal (Australia, repurposing decommissioned coal plants).
Liquid Air Energy Storage (LAES): cryogenic at −196 °C. Highview Power 50 MW Manchester UK (2023).
Round-trip 40-50% (heat-to-electric limited by Carnot efficiency).

Gravity storage (mostly struggling commercially)

Energy Vault — original cranes lifting concrete blocks; pivoted to building-style designs.
Gravitricity — heavy weights in abandoned mine shafts.
ARES (Advanced Rail Energy Storage) — rail cars on gradient.
Most gravity-storage concepts have struggled with cost per kWh.

Hydrogen storage (see L130)

Production via electrolysis during surplus renewable hours.
Stored as compressed gas, liquid (cryogenic), or in metal hydrides / salt caverns.
Consumed via fuel cell or combustion turbine.
Very-long-duration (weeks to months). Low round-trip (30-45%). Highly scalable but expensive.

Policy and market

DOE LDES Storage Shot (2021): 90% cost reduction by 2030, $50/kWh delivered target.
IRA 2022: storage ITC 30% + 10% domestic content + 10% energy-community + 10% low-income — up to 60% effective ITC. Most aggressive ever.
WoodMac forecast: 250 GW LDES globally by 2030; 1500 GW by 2050.
Major utility pilots: Xcel, Duke, Dominion, PG&E, NextEra, Berkshire Hathaway Energy, FPL, Georgia Power, Entergy.
FERC Order 841 (2018) opened markets to storage; FERC 2222 (2020) DER aggregation.

Take-away. LDES (Long-Duration Energy Storage, > 10 hr) is critical for 80-100% renewable grids — lithium-ion is economical only to ~4 hr. Duration regimes: fast (flywheel), short (Li-ion), medium (Li-ion + flow), LDES (iron-air, flow, CAES, thermal), seasonal (hydrogen). Pumped hydro is 95% of global storage today (180 GW, 70-85% RT, 50-100 yr life, geography-limited). CAES legacy: Huntorf 1978, McIntosh 1991; adiabatic emerging. Flow batteries decouple P and E for cheap duration scaling (VRFB, ESS iron). Iron-air (Form Energy): 100 hr at $20/kWh target, ~50% RT, multi-day chemistry. Thermal: molten salt at CSP plants, rock beds, LAES. DOE LDES Storage Shot: 90% cost reduction by 2030. IRA 2022 storage ITC 30% + adders. WoodMac forecast 1500 GW LDES globally by 2050.