- The paper presents a high-resolution optimization method to design cost-optimal hydrogen production systems powered by renewable energy.
- The paper demonstrates that incorporating country-specific investment risks can significantly increase LCOH, particularly in high-risk regions.
- The paper shows that hybrid systems combining PV and wind are crucial, with 76.4% of designs benefiting from diversified renewable resources to lower costs.
This paper presents a method for modeling the Levelized Cost of Hydrogen (LCOH) production globally with a high spatial resolution of approximately 50x50 km. It uses an optimization approach to design cost-optimal hydrogen production systems (HPSs) powered by renewable energy, specifically considering the impact of country-specific investment risks (Country Risk Premiums, CRPs).
The core of the methodology is the Hydrogen Production System Model (HPSM), implemented using the open-source energy system modeling framework PyPSA. For each of the 51,677 inland grid cells derived from the MERRA-2 weather dataset, the model optimizes the capacity of various components to meet a constant hydrogen demand of 1 kg/h at the minimum total system cost. The available components for the HPS include ground-mounted photovoltaic (PV) systems, onshore wind turbines (weak and strong wind variants based on local wind conditions), Polymer Electrolyte Membrane (PEM) electrolyzers, hydrogen storage units (tubular storage compressed to 100 bar), hydrogen compressors, and stationary battery storage units. The model assumes off-grid operation, meaning no connection to the local electrical grid for buying or selling electricity.
The total annual cost for each technology is calculated using its CAPEX, OPEX (as a percentage of CAPEX), and lifetime, annualized via the Annuity Factor (AF). A key novelty is the inclusion of country-specific investment risk in the discount rate (r). The discount rate is calculated as a baseline Weighted Average Cost of Capital (WACC) of 3.5% plus the country-specific CRP. The CRPs used are derived from bond default spreads and volatility factors, sourced from external data for 2020. This approach allows for a more realistic assessment of financing costs in different countries.
The optimization problem is formulated as a linear program and solved using the Gurobi solver. To manage the computational effort for over 50,000 individual optimization problems, parallelization is employed, significantly reducing the total computation time.
The study analyzes four scenarios:
- cWACC: Uses a constant WACC of 3.5% for all locations, isolating the impact of renewable resource availability.
- BASE: Incorporates country-specific CRPs into the WACC for each location, providing a more realistic cost assessment.
- PVonly: Same as BASE, but only PV is available for electricity generation.
- WINDonly: Same as BASE, but only wind onshore is available for electricity generation.
The results highlight several practical implications:
- Spatial Variation: LCOH varies significantly globally, ranging from 2.7 €/kg to 28.4 €/kg in the BASE scenario, with a mean of 9.1 €/kg. Even within countries, LCOH can differ substantially due to varying meteorological conditions.
- Impact of Country Risk Premiums: CRPs have a strong influence on LCOH, often counteracting the advantage of good renewable resources in high-risk countries. Comparing the cWACC and BASE scenarios shows that LCOH can more than double in countries with high CRPs, such as Argentina (101% to 117% increase). Countries with low CRPs and good resources (e.g., Chile, parts of the USA, Australia) tend to have lower LCOH.
- Hybrid Systems are Cost-Optimal: In the BASE scenario, 76.4% of the optimized HPS designs are hybrid (combining PV and wind). Hybrid systems are crucial for reducing LCOH, particularly in areas with less ideal conditions for a single renewable technology. The LCOH increase significantly in the PVonly and WINDonly scenarios compared to BASE, demonstrating the cost benefit of hybrid designs. This cost reduction is most pronounced in cells with moderate LCOH and less significant in cells with already very low LCOH (often dominated by one technology).
- Component Contributions: Renewable electricity generation is the largest cost component (over 50%) in all HPS designs. The electrolyzer is the second largest. Hydrogen storage is predominantly deployed to balance the intermittent renewable supply for the constant demand, whereas battery storage is rarely used due to its higher cost relative to other component combinations.
The paper also discusses limitations for practical implementation. The 50x50 km resolution is still broad for site-specific planning. The generic CRP/WACC assumptions may not reflect project-specific financing costs or the impact of government support measures. The analysis uses a single weather year, and a more robust assessment would require multiple years of data. Land and water availability are not considered, which would impact potential production volume rather than LCOH for a given demand. The assumption of constant demand and no grid connection also influences the results; incorporating demand flexibility or grid interaction could lower calculated LCOH.
Overall, the study provides a valuable global assessment tool, demonstrating the critical roles of both spatially explicit renewable resources and country-specific investment risks in determining the economic feasibility of green hydrogen production. It emphasizes the advantage of hybrid renewable systems and the importance of hydrogen storage for ensuring a stable supply. The results and underlying data are made open-source to support further analysis and policy decisions in the global energy transition.