Green Hydrogen from PV: Electrolyser Size and Cost
Size a solar-fed electrolyser using hourly PV power and minimum-load limits. Estimate hydrogen output, curtailment and levelized cost with your own costs.
This green hydrogen cost calculator helps engineers, students and project developers compare a photovoltaic array with a PEM or alkaline electrolyser. It estimates hydrogen output, unused solar energy and a levelized production cost from an hourly dispatch model. Use it for early feasibility studies and teaching, before commissioning a site-specific production and financial model.
Start with PV nameplate power and electrolyser system input rating. Their ratio describes solar oversizing: a ratio of two means twice as much nominal PV power as electrolyser power. The representative profile uses a half-sine daylight curve centred at solar noon. Enter net daily PV yield as equivalent full-power hours and set daylight duration. The calculator integrates that curve into 24 hourly average powers. Alternatively, enter 24 hourly average outputs as percentages of PV nameplate. These entries can represent a measured day or a selected scenario.
Select the minimum-load preset or enter a vendor-specific cutoff. The PEM preset uses 5%; the alkaline preset uses 30%. These are screening assumptions, not guaranteed technology limits. Change system electricity consumption and installed cost independently when comparing technologies. The default 55 kWh/kg includes operating balance-of-plant consumption. Below the minimum load the electrolyser shuts down; above its rating, surplus PV power is unused. The daily chart shows available PV, electrolyser draw and the lower and upper running limits.
Annual hydrogen assumes the same day repeated 365 times, with the chosen electrolyser availability. Averages alone hide seasonal and cloudy-day effects, especially near the minimum-load threshold. An actual project needs a full weather year and an operating strategy. The hourly table makes each energy and hydrogen balance inspectable.
For electrolyser sizing with solar, both sweep charts keep PV nameplate and its hourly profile fixed. Moving right increases the PV/electrolyser ratio by reducing electrolyser rating; it does not add PV. A very large electrolyser can miss low-power hours because its absolute minimum load is high. Reducing its rating can initially recover those hours, but further reductions clip more PV and reduce hydrogen output. LCOH balances reduced equipment cost against this lost production. The lowest sampled cost is confined to the displayed range and spacing, not a continuous optimum or vendor selection. If no sampled size is feasible, sampled size and cost results are withheld with an explanatory message; valid selected-design results remain available.
This LCOH calculator annualises initial electrolyser investment, discounts stack replacements and adds fixed O&M, electricity and variable operating costs. Choose whether electricity is charged on all PV generation or only on electrolyser consumption. A PV LCOE already includes solar investment and O&M: do not add those costs again. All monetary entries use one currency and price year; CU means your chosen currency unit. Cost bases remain CU/kW, CU/kWh and CU/kg when physical results switch units. Defaults are illustrative assumptions dated 29 October 2026, not supplier quotations.
Results stop at the electrolyser outlet. They exclude storage, delivery compression, transport, taxes, subsidies, export revenue, battery dispatch, part-load efficiency changes, startup energy and explicit degradation. Stack life and availability approximate maintenance effects. Renewable certification and emissions accounting require separate assessment.
How it is calculated
For the smooth profile, during daylight and zero outside it, with . Each hourly average is its exact integral divided by one hour. The alternative profile directly uses entered hourly percentages of PV rating.
Hourly dispatch is below , otherwise . Sum power × one hour to obtain energy. Hydrogen equals accepted system energy divided by specific system consumption. Annual production is the available-day output × 365 × availability. Capacity factor uses 8,760 calendar hours; online hours count every productive hour, including part load.
Capital recovery is , with at zero discount. NREL 90103 (2024), Eq. 3, provides this annualisation method. A replacement occurs after each input stack-life interval of online hours, at a fractional project year, strictly before project end. Discount each replacement to year zero and annualise its present value with the same CRF.
All-generation billing uses total available annual PV energy, including curtailment and electrolyser downtime. Consumed-only billing uses actual annual electrolyser energy. The 31-point sizing sweep holds PV and all other assumptions fixed while varying PV/electrolyser ratio from 0.5 to 5 in increments of 0.15; infeasible candidates are chart gaps. When every sampled cost is non-finite, no sampled size or cost is reported; the selected design remains independent. No export credit is applied.
The 55 kWh/kg default and LHV efficiency benchmark follow DOE's PEM technical-target table (2022 status). IRENA (2020), pp. 32 and 65, documents load-range variability and system boundaries. NREL 70380 (2017), pp. 28 and 52–58, distinguishes stack specifications, whole-system performance and manufacturing costs; its historical costs are not treated as installed prices today.
Assumptions
- One lumped electrolyser system with constant kWh/kg, no overload, no battery and no grid import.
- A single representative day repeats for a non-leap year. Availability uniformly reduces production and running hours; PV generation continues during downtime.
- Smooth-profile PV yield is net available electrical energy divided by nominal PV power. Its peak must not exceed nameplate. Entered profiles are hourly averages between 0% and 100%.
- PEM 5% and alkaline 30% cutoffs are illustrative. The broad screening ranges in the brief are PEM 3–10% and alkaline 20–40%; IRENA reports some alkaline designs down to 15%. Verify vendor limits, gas purity and operating conditions. The preset changes only cutoff.
- Stack replacements restore performance at constant real cost, scheduled by online hours. Availability includes maintenance downtime; cycling wear is not separately calculated.
- Constant real annual costs and production, initial capital at year zero, end-of-year annual payments, no salvage, tax, subsidy or escalation. Default costs are author-selected examples as of 29 October 2026.
- CU is any single chosen currency. Monetary quantities are numerical accounting values; reference bases of 1 kW, 1 kWh and 1 kg keep units explicit. Monetary input and output bases stay metric under the physical unit switch.
Where it applies
Use for preliminary energy and production-cost comparisons of low-temperature PEM or alkaline electrolysis. Hourly averaging can miss brief cutoff crossings, rapid clouds and startup losses. A representative day cannot reproduce a year's weather or guarantee annual supply.
The lowest sampled LCOH is limited to ratios 0.5–5; it may occur at an endpoint and need not be the continuous optimum. High minimum loads can cause discontinuous changes in online hours, stack replacements and costs. The selected design can lie outside the sweep range.
Use vendor system consumption on the same electrical boundary as PV supply and electrolyser rating. Do not combine a DC stack-only rating with an AC whole-system kWh/kg value. Constant consumption omits part-load and age effects. Validate operating-hour stack life for intermittent operation. No oxygen revenue, water balance, hydrogen storage sizing, delivery-pressure calculation, renewable certification or emissions claim is provided.
Gas separation, ventilation, hydrogen detection and pressure protection are not designed here (hydrogen is flammable and may be produced under pressure); qualified specialists design and install the hydrogen and electrical systems.
References
- NREL/PR-6A20-70380, Ruth, Mayyas and Mann (2017), Manufacturing Competitiveness Analysis for PEM and Alkaline Water Electrolysis Systems. https://docs.nlr.gov/docs/fy19osti/70380.pdf
- IRENA (2020), Green Hydrogen Cost Reduction: Scaling up Electrolysers to Meet the 1.5°C Climate Goal, especially pp. 18, 32, 45 and 65. https://www.irena.org/-/media/Files/IRENA/Agency/Publication/2020/Dec/IRENA_Green_hydrogen_cost_2020.pdf
- U.S. Department of Energy, Technical Targets for Proton Exchange Membrane Electrolysis, 2022 status and 2026/ultimate targets, accessed 29 October 2026. https://www.energy.gov/cmei/fuels/technical-targets-proton-exchange-membrane-electrolysis
- NREL/TP-5400-90103, Penev et al. (2024), Capital Structure for Techno-Economic Analysis of Hydrogen Projects, Table 1 and Eq. 3. https://docs.nlr.gov/docs/fy24osti/90103.pdf
Inputs
| Input | Symbol | Unit | Default |
|---|---|---|---|
| PV nameplate power PV nominal rating. The profile is the net electrical power available at the electrolyser system inlet. | P_pv | kW | 1000 |
| PV profile A representative day repeats 365 times. Hourly entries are percentages of PV nameplate, not irradiance. | profile | Representative half-sine day | |
| Net daily PV yield (equivalent full-power hours) Daily available kWh divided by PV nameplate kW. Includes PV and conversion losses up to the electrolyser system inlet. | Y_day | h | 5 |
| Daylight duration Half-sine production window centred at solar noon; hourly average powers are integrated exactly. | D_day | h | 12 |
| Electrolyser system rating Total electrical input rating including balance of plant, on the same basis as the specific energy input. | P_el | kW | 500 |
| Minimum-load preset Screening presets only. PEM commonly spans 3–10%; conventional alkaline 20–40%. Check the vendor guarantee. This choice changes only the cutoff, not efficiency or cost. | tech | PEM: 5% minimum load | |
| Custom minimum load (%) Percentage of total system rating below which the whole electrolyser is off. | cut_pct | 5 | |
| System electricity consumption AC system electricity per kg H₂, including operating auxiliaries. DOE 2022 PEM status is 55 kWh/kg. Enter a lifetime-average value if available. | epsilon | kWh/kg | 55 |
| Electrolyser availability (%) Uniform availability applied to daily hydrogen and running hours. PV still generates during electrolyser downtime. | avail_pct | 97 | |
| Installed electrolyser cost (CU/kW) Currency units per kW of system input. Include installation, power electronics and balance of plant; exclude PV costs already in the power price. | c_el | 1200 | |
| PV electricity cost (CU/kWh) Use the PV LCOE or the applicable PPA price, in one chosen currency and price year. | c_power | 0.035 | |
| Electricity cost basis All generation allocates curtailment and downtime costs to hydrogen. Consumed-only represents a contract that charges only delivered energy. | billing | All available PV generation | |
| Annual fixed O&M (% of initial electrolyser cost) Real annual operating cost. Stack replacements are added separately; avoid counting them again here. | om_pct | 3 | |
| Other variable operating cost (CU/kg H₂) Water, consumables and other costs per kg. Add downstream costs only if you also include their energy and equipment consistently. | c_variable | 0.15 | |
| Project life (years) Years of constant annual production and real operating costs; zero salvage value. | N_year | 20 | |
| Real discount rate (%) Use a real WACC consistent with costs in constant money; zero is supported. | rate_pct | 8 | |
| Stack service life (running hours) Actual online hours, not equivalent full-load hours. Cycling and degradation require vendor-specific adjustment. | stack_h | h | 60000 |
| Each stack replacement (% of initial cost) Replacement cost including work, as a fraction of initial installed electrolyser cost. Replacements before project end are discounted individually. | stack_pct | 35 | |
| 00:00–01:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_00 | 0 | |
| 01:00–02:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_01 | 0 | |
| 02:00–03:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_02 | 0 | |
| 03:00–04:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_03 | 0 | |
| 04:00–05:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_04 | 0 | |
| 05:00–06:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_05 | 0 | |
| 06:00–07:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_06 | 8.5185 | |
| 07:00–08:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_07 | 24.9751 | |
| 08:00–09:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_08 | 39.7297 | |
| 09:00–10:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_09 | 51.7767 | |
| 10:00–11:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_10 | 60.2952 | |
| 11:00–12:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_11 | 64.7048 | |
| 12:00–13:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_12 | 64.7048 | |
| 13:00–14:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_13 | 60.2952 | |
| 14:00–15:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_14 | 51.7767 | |
| 15:00–16:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_15 | 39.7297 | |
| 16:00–17:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_16 | 24.9751 | |
| 17:00–18:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_17 | 8.5185 | |
| 18:00–19:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_18 | 0 | |
| 19:00–20:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_19 | 0 | |
| 20:00–21:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_20 | 0 | |
| 21:00–22:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_21 | 0 | |
| 22:00–23:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_22 | 0 | |
| 23:00–24:00: PV output (%) Hourly average available power as a percentage of PV nameplate, after upstream electrical losses. | pv_23 | 0 |
Results
With the default inputs above, the calc gives:
| Result | Value |
|---|---|
| Levelized hydrogen cost (CU/kg H₂) | 5.38 |
| Annual hydrogen | 28,739 kg |
| Electrolyser capacity factor (%) | 36.1 |
| Annual PV energy used (%) | 86.6 |
| Hydrogen per available day | 81.2 kg |
| PV / electrolyser power ratio | 2 |
| Electrolyser at lowest sampled LCOH | 500 kW |
| Lowest sampled LCOH (CU/kg H₂) | 5.38 |
Checked against sources
Each case below was calculated by this calc and agrees with the value from its source (within the tolerance where one is shown).
IRENA Figure 1: 51.2 kWh/kg and 65% LHV efficiency
Source: IRENA (2020), Figure 1 note, printed/PDF p. 18: specific electricity use 51.2 kWh/kg corresponds to 65% LHV efficiency. The note calls this LHV; it is the electricity input, not the heating value of hydrogen. Compare efficiency only, not the report's complete cost curves.
Inputs: System electricity consumption = 51.2 kWh/kg; the others as above.
| Result | Source | This calc |
|---|---|---|
| System efficiency on an LHV basis | 0.65 ± 0.005 | 0.651 |
DOE PEM 2022 system benchmark: 55 kWh/kg, 61% LHV
Source: DOE Technical Targets for Proton Exchange Membrane Electrolysis, system energy efficiency, 2022 status: 55 kWh/kg and 61% LHV. Rounded independent published pair; this checks the system energy-to-hydrogen conversion, not PV dispatch. https://www.energy.gov/cmei/fuels/technical-targets-proton-exchange-membrane-electrolysis
Inputs: the default inputs above.
| Result | Source | This calc |
|---|---|---|
| System efficiency on an LHV basis | 0.61 ± 0.01 | 0.606 |
NREL 90103 Table 1: 30-year real capital recovery factor
Source: NREL 90103 (2024), Table 1, printed p. 4/PDF p. 10: real WACC 7.6% and 30-year CRF 8.5%. Published figures are rounded; 1% relative tolerance allows the table precision.
Inputs: Real discount rate (%) = 7.6; Project life (years) = 30; the others as above.
| Result | Source | This calc |
|---|---|---|
| Capital recovery factor | 0.085 ± 0.01 | 0.0855 |
NREL 90103: 30-year factor at a 10% discount rate
Source: NREL 90103 (2024), Table 1 and Eq. 3: nominal 10% WACC gives a rounded 10.6% 30-year CRF. This isolates the algebraic factor; real and nominal rates are not mixed in an actual project.
Inputs: Real discount rate (%) = 10; Project life (years) = 30; the others as above.
| Result | Source | This calc |
|---|---|---|
| Capital recovery factor | 0.106 ± 0.005 | 0.1061 |
Reviewer regression: selected design works outside an infeasible sweep
Source: Independent reviewer case supplied with this calculator's review: Y_day=0.4 h, P_el=100 kW, alkaline 30% cutoff, other defaults unchanged. Selected annual production is 2,229.94 kg; every ratio 0.5–5 is infeasible. Both sampled size and sampled cost must therefore be unavailable while selected production remains valid. The 0.005 kg tolerance is half the review's final reported decimal place.
Inputs: Net daily PV yield (equivalent full-power hours) = 0.4 h; Electrolyser system rating = 100 kW; Minimum-load preset = Alkaline: 30% minimum load; the others as above.
| Result | Source | This calc |
|---|---|---|
| Both sampled size and sampled cost are withheld | 1 | 1 |
| Annual hydrogen | 2,230 kg | 2,230 kg |
Related Resources
Related courses
IEC Electrical Schematics and Industrial Control Panels
Basic Electrical Controls & Diagram Reading

Responses