Onsite Hydrogen for Power Plants: PEM vs Alkaline Electrolysis (2025 Guide)

By Green Gas Turbines Team · Published November 15, 2025 · 13 min read


Why Make Hydrogen Onsite?

Onsite electrolysis avoids trucked deliveries, reduces supply risk, and enables fast, flexible hydrogen for peaker/hybrid gas turbines. It can soak up low-price or curtailed electricity, support black-start, and provide oxygen byproduct for industrial users. Two technologies dominate: PEM (proton exchange membrane) and alkaline electrolysis.

PEM vs Alkaline: Quick Comparison

Attribute PEM Electrolysis Alkaline Electrolysis So What for a Power Site?
Dynamic response / turndown Fast ramp (sub-seconds to seconds), wide turndown (≈10–100%) Slower ramp (minutes), narrower turndown (≈30–100% typical) PEM suits peakers, hybrids, and variable renewables; alkaline favors steady baseload production
Whole-system electricity use ≈50–55 kWh/kg H2 (LHV basis), + compression ≈49–54 kWh/kg H2 (LHV basis), + compression Similar in practice; site parasitics and compression often dominate differences
Outlet pressure (without compressor) Typically 20–30 bar Typically 10–30 bar Either way you’ll size compression to storage/crossover pressure
Water quality High-purity deionized water (≈≤1 μS/cm) Demineralized water; less stringent but polishing still needed Plan a robust water plant either way (see below)
Hydrogen purity (after dryers/polishers) Up to 99.999% 99.8–99.999% (depends on system) Both meet gas turbine needs with proper drying and filtration
Stack life (typical) ≈40–80k operating hours ≈60–90k operating hours Budget stack replacements; alkaline often longer life at steady duty
CAPEX (electrolyzer module) Higher $/kW; compact footprint Lower $/kW; larger footprint PEM pays for flexibility; alkaline wins on steady, low-cost kgs
Materials/catalyst Precious metals (iridium/platinum) Nickel-based

How to Size an Onsite Electrolyzer for a Gas Turbine

Work backwards from turbine hydrogen consumption and your storage strategy.

Quick Sizing Steps for Onsite H2

  1. Estimate H2 use at the turbine

    • Simple cycle (100% H2): ~75 kg/MWh
    • Combined cycle (100% H2): ~52 kg/MWh
    • For blends: multiply by H2% (vol ≈ mass for screening)
  2. Define duty

    Required kg per event = (MW × hours) × (kg/MWh) × H2%
    Continuous daily need = sum of events + reserve
  3. Decide make-rate vs store-rate

    Electrolyzer kg/h × run hours + storage draw ≥ event need
    Storage (kg) ≥ peak deficit between consumption and make-rate
  4. Add efficiency & compression

    Electrical load (kW) ≈ kg/h × (50–55 kWh/kg + compression allowance)

Example: 50 MW simple-cycle peaker, 2 hours at 30% H2 → H2 need ≈ 75 × 50 × 2 × 0.30 = 2,250 kg. If you run a 1 MW PEM (≈20 kg/h) for 24 h you make ≈480 kg, so you’d either upsize to ≈5 MW or add storage (or both).

Planning Heuristics (Screening-Level)

Net load = Load − (Wind + Solar)

Required ramp (MW/min) ≈ max{Δ Net load / Δt}

Peaker coverage:

  • Use batteries to cover first 0–2(–4) hours of the evening ramp.
  • Size peakers to cover residual ramp + extended events (e.g., multi-hour/low-renewable periods).
  • Ensure N−1 security: largest unit outage still covered by remaining flexible fleet.

ELCC note:

  • As renewables share increases, peaker ELCC remains high if fuel-secure and fast-start.
  • Battery ELCC declines with duration < worst-hour length; increase duration or pair with peakers/LDES.

Why not just overbuild renewables?

Overbuild helps energy adequacy but doesn’t reliably solve operability in worst hours or stability. A diversified stack (renewables, storage, peakers, transmission, demand flexibility) is cheaper and more reliable.

Action Checklist for Planners & Owners

Conclusion: Rare Hours, Essential Value

In a 60–90% renewables grid, peakers are no longer bulk generators—they are precision instruments for reliability. Paired with batteries and cleaner fuels, they bridge the hardest hours, keep the system stable, and make high-renewables futures achievable and affordable. Plan for fewer hours, higher stakes, and tighter integration—and peakers will continue to earn their keep in a decarbonized grid.