Hydrogen Storage (community-scale, seasonal)

EDT domain: d01_renewable_energy_storage

Definition

Community-scale hydrogen energy storage — an electrolyser (power → H₂) feeding compressed/stored hydrogen and a fuel cell (H₂ → power). The candidate clean seasonal backstop: it banks surplus summer renewables for winter, spanning the multi-week gap a battery cannot. In the NI engine it is a V1+ / research-horizon option, not a near-term default. CR_030

Role — seasonal storage only

  • Round-trip efficiency is low: ~37% typical, ~48% best-case (80% electrolyser × 60% fuel cell), ~35% commonly cited — against ~95% for batteries. CR_030
  • That poor round-trip means it only makes sense for the seasonal (multi-week / inter-seasonal) gap — never diurnal firming, where a battery wins decisively. CR_030
  • The overbuild consequence: because only ~30–45% comes back, you must build ~2–3× more renewable generation to recover a given amount of energy via hydrogen than via a battery — a large hidden cost for a community funding its own generation. CR_040

Cost

  • Indicative ~USD 0.35/kWh cycled, dominated by electrolyser and fuel-cell stacks; compression to 200–700 bar adds parasitic load. CR_030
  • NZ green-H₂ production-cost anchor (NZD, 2019). Concept Consulting’s NZ reference estimate for producing green hydrogen today is 8.91/kg NZD** including tank storage (= **63/GJ = 0.23/kWh**; electricity is >75% of the cost, at an 85% capacity factor), spread by use case across Table 1 — current **6.80–7.57/kg (gas-pipeline injection), 11.30** (service station) and **12.56/kg (off-grid bulk storage, the highest current value and the row most relevant to a remote community), falling to 5.33–9.22/kg in the "future large-scale" case. Hydrogen-from-hydrocarbons (SMR+CCS) is put at **2.7/kg** (19/GJ) at 100/tCO₂ ⚠ **[correction — OT_151, 2026-07-20: the “2.7/kg (19/GJ)” figure is NOT the SMR+CCS cost. Report 2’s modelled SMR+CCS production cost is ~14/GJ (≈2/kg); 2.7/kg = 19/GJ corresponds instead to a green-H₂ value (Table 3 future-opportunistic Gas-Tx-injection 2.67/kg), not a hydrocarbon/SMR estimate]**. All values NZD (benchmarked vs NZ wholesale gas 6/GJ, electricity $0.075/kWh). OT_122
  • Cost-model decomposition + storage-tech costs (Concept 2019, Reports 2+3 — the technical volumes behind the OT_122 anchor). The 8.91/kg NZD reference resolves to wholesale electricity 4.88 + network 1.91 + electrolyser capex 1.03 + opex 0.59 + storage 0.50 per kg (electrolyser 1,400/kW → 700 future, 70% efficiency, 85% utilisation, storage 0.5/kg-H₂). Storage technology detail: four routes — compressed gas (class-IV carbon-fibre tanks, ~**7,000/GJ** of storage capacity), cryogenic liquid (denser, unavoidable boil-off), chemical bonding (ammonia/toluene carriers), and metal hydride (‘not yet ready to be used commercially’). Distribution: pipeline 1–2M/km; compressed-gas tube-trailer 7,500/GJ (→ ~6,200/GJ before 2030); truck delivery 7.51/GJ (compressed H₂) vs $1.41/GJ (ammonia). NZD cross-check on the international storage figures the page carries from CR_030/CR_040. OT_151

Maturity — the feasibility question

  • Technically demonstrated but commercially immature for community backup. NZ hydrogen activity (H2 Taranaki Roadmap; Taranaki Basin underground storage) targets production / transport / industry, not community off-grid. CR_030
  • TIMES-NZ national scenarios include hydrogen but chiefly for industry and transport firming, not household/community storage. OT_055
  • NZ policy framing (current): the MBIE Hydrogen Action Plan (Nov 2024) + EECA position hydrogen for heavy transport / hard-to-electrify industry, not off-grid community electricity storage — “electrification will be the cheapest and most efficient way to reduce emissions for most activities” (zero mentions of off-grid/community/remote; the Plan’s only “Power” use is grid-scale peaking/back-up). Supersedes the older H2 Taranaki Roadmap framing. OT_081 CR_042
  • The only NZ community off-grid demonstration is academic: the Stewart Island/Rakiura multi-carrier microgrid with a stationary hydrogen store (964 kW electrolyser, 619 kg H₂, 261 kW fuel cell) reaching ~0% loss-of-load. LIT_033
  • Named as an enabler across the NZ community-microgrid literature, but not yet deployed at community scale. LIT_002

Real community-scale examples (EU REMOTE)

The clearest real community-scale case set is the EU Horizon 2020 REMOTE project (H2020 779541; €6.7M, 2018–2023, coord. Politecnico di Torino — URL_016), built to test hydrogen storage in isolated/off-grid communities. Froan Island, Norway (PV + 225 kW wind + 3,333 kWh H₂ tank + 550 kWh Li-ion) was justified partly by avoiding a new submarine power cable — i.e. hydrogen competed against a very expensive grid alternative, not against batteries alone. Ginostra (IT), Agkistro (GR) and Ambornetti (IT) test other mixes. The pre-deployment technical specs are now in OT_155 (REMOTE D2.2); the measured operational-outcome data (achieved efficiency, lifetime cost, downtime) is still thin online (RT_275, kept open). The “Froan logic” is the honest niche: hydrogen wins where the alternative is extraordinarily expensive, not on its own merits vs batteries. CR_040

  • Primary REMOTE deliverable (D2.2) now retrieved — Froan config verified + demonstrator specs added. The project’s own technical-specification deliverable confirms the Froan/Rye figures cited above directly from the primary source: 85 kW PV + 225 kW wind turbine, a 3333 kWh (≈100 kg) H₂ store at 30 barg, and a 550 kWh Li-ion buffer (5 racks × 110 kWh, 96%, SOC 20–90%), with a 55 kW PEM electrolyser (63% LHV) + 100 kW PEM fuel cell (50% LHV). Across all four demos the modelled architecture is “battery does the diurnal work, hydrogen is the longer-term/seasonal store” (Froan modelled coverage: 25.8% battery vs 8.4% fuel cell of load), and at Ambornetti biomass, not the fuel cell, is the effective backstop (36.6% vs 8.3%; full autonomy) — both corroborating this page’s “seasonal-only / battery-plus-backstop-wins” verdict. The design round-trip is ~31.5% LHV (63% × 50%) before converter/BoP/auxiliary losses — at/below the low end of the ~35–48% (CR_030) / ~30–45% (CR_040) range, so treat those as optimistic. ⚠ EU demonstrator DESIGN/MODELLED values (subsidised R&D), not measured operational data and not NZ — the measured achieved-efficiency / lifetime-cost / downtime residual of RT_275 remains open. OT_155

Relevance to Neobiome (D20)

A research-horizon seasonal-backup option, deferred behind the diesel-genset backstop and the LPSP reliability lever. Crucially, if it is self-generated (electrolyser fed by on-site surplus), the stored hydrogen is local and does not count as an import — its cost to self-sufficiency is the capital plus the ~50–65% of energy lost round-trip, not a reduction in SSI like bought diesel (diesel_genset). This makes it the one backstop that is clean and import-free, at the price of round-trip losses and immature cost. CR_030

Open questions

  • Community-scale electrolyser + fuel-cell + storage capex in NZD (RT_178 gave only international indicative figures).
  • Whether any NZ community pilot moves community hydrogen from academic demonstration to deployed practice.
  • An NZ-specific community off-grid hydrogen techno-economic feasibility benchmark (none exists; CAPEX is international-modelled — RT_276); the MBIE Hydrogen Action Plan (Nov 2024) and EU REMOTE operational-results deliverables (RT_274; RT_275 advanced — D2.2 specs retrieved as OT_155, measured operational/economic outcomes still outstanding). CR_040

Connections

Links to

Referenced by

Sources (6): CR_040 · CR_042 · LIT_061 · OT_081 · OT_155 · URL_016

EDT domains (1): D01: Renewable Energy & Storage Systems