Deploying a Hydrogen Supply Chain at Regional Scale
Introduction
I am Chugoun Mustafa, research engineer at the University of Corsica. This work was funded by ANR France 2030 under the UNITI project — University for the Transformation of Mediterranean Island Territories — in partnership with the University of Toulouse through the LAPLACE laboratory.
The objective is to develop a multi-criteria optimization approach for deploying a hydrogen supply chain at regional scale, applied to Corsica.
Corsica's Energy Context
Corsica is a ZNI — Non-Interconnected Zone in the electrical sense — a territory not directly connected to the mainland grid, though partially linked to Italy.
These territories share distinct energy characteristics. They import the majority of their energy needs by sea — heavy fuel oil, diesel, gasoline — to produce electricity and power transport. This results in electricity production costs approximately three times higher than mainland France (~€72/MWh vs ~€270/MWh), and a carbon intensity five times higher per kWh produced.
Today, over 40% of primary energy consumed in Corsica is dedicated to the transport sector — light, heavy and maritime mobility.
By 2050, local authorities target a 90% reduction in greenhouse gas emissions, a 50% reduction in final energy consumption and 100% renewable electricity coverage.
A key technical constraint frames these objectives: in Corsica, no more than 35% of electricity from intermittent renewable sources can be injected simultaneously into the grid. Hydrogen emerges as a critical solution to lift this constraint, by valorizing renewable surplus production and reducing the carbon intensity of the local electricity mix.
Structure of the Supply Chain
The study covers the entire hydrogen supply chain from production to end use, focusing on mobility — the main fossil energy consumer on the island.
The model addresses the following questions: which energy sources to use (solar, wind, grid)? Which electrolyzer technology and size? Is storage needed and at what capacity? How to transport hydrogen between zones — compressed or liquefied? Can retrofitted hydrogen-powered trucks handle this transport?
The Corsican territory was divided into nine distinct zones. For each zone, the model assesses local demand, available renewable resources, and determines whether production can be met locally or whether supply from a neighboring zone is required.
Methodology
Mathematical Modeling
All data — demand profiles, techno-economic parameters, production constraints — are translated into mathematical language using the GAMS tool. The model optimizes three criteria simultaneously:
— Total chain cost (CAPEX + OPEX, from production to end use)
— Greenhouse gas emissions (kg CO₂ equivalent/kg H₂)
— A risk index specifically developed to assign a risk level to each link in the chain
Optimization can be single-objective (one criterion at a time) or multi-objective (all three simultaneously). In the latter case, the epsilon-constraint method and TOPSIS identify the best compromise between the three objectives.
Geospatial Approach with QGIS
Upstream: identification of territorial constraints — mountainous areas, nature reserves, waterways, residential zones, proximity to the electricity grid and roads — to locate potential installation sites for electrolyzers and storage systems.
Downstream: mapping of existing service stations with sufficient space to accommodate a hydrogen station, minimizing changes to user habits.
Drinking water availability is also mapped, given the increasingly frequent drought episodes in Corsica.
Techno-Economic Parameters
Four electrolyzer sizes were integrated, ranging from 300 kW to 5 MW. Each technology is characterized by its CAPEX, OPEX, efficiency, CO₂ emissions and risk index.
For transport, diesel truck emissions amount to approximately 63 g CO₂ equivalent/kg H₂. The model also integrates the option of hydrogen-retrofitted trucks to eliminate this emission source.
Scenarios Studied
Three renewable energy deployment trajectories were modeled between 2025 and 2050: from 226 MW (low scenario) to over 400 MW (optimistic scenario) in combined PV and wind capacity.
Hydrogen demand was projected between 5 and 43 tonnes per day by 2050, integrating demographic trends and adoption trajectories for hydrogen and electric vehicles across light, heavy and maritime mobility.
Results
Single-Objective Optimization
— Cost minimization: total chain cost ~€9,000/day, LCOH ~€8.5/kg
— Emissions minimization: minimum of 7.16 kg CO₂ equivalent/kg H₂
— Risk minimization: minimum risk index of 16
Critical finding: the European threshold to qualify hydrogen as low-carbon is 3 kg CO₂/kg H₂. Even when minimizing emissions, Corsica does not reach this threshold when using the current electricity mix — too carbon-intensive — as the electrolyzer power source.
Impact of Electricity Cost
Electricity cost represents nearly 80% of the final hydrogen production cost. It is the single most decisive factor across the entire chain. Using the local Corsican production cost or EDF green tariffs yields significant improvement, unlike ADEME projections which have little impact on LCOH.
Liquefied vs Compressed Hydrogen Transport
Liquefied hydrogen transport is not a viable option for Corsica at this demand level: LCOH exceeds €13/kg and adds over 10 tonnes of CO₂ per day to the chain.
Electron Transport vs Hydrogen Transport
Sending surplus electricity through the existing grid to deficit zones rather than transporting hydrogen by truck saves approximately €0.60/kg and reduces grid imports by 7%. There is no physical guarantee of a green electron, but this is consistent with existing green electricity tariffs.
Grid-Free Scenario with Retrofitted Trucks
By completely excluding the electricity grid as an electrolyzer power source and using hydrogen-retrofitted trucks for transport, emissions fall below the 3 kg CO₂/kg H₂ threshold. The hydrogen produced then qualifies as low-carbon.
By doubling the weight of emissions in the multi-objective optimization, the model achieves a minimum of 1.43 kg CO₂/kg H₂ — entirely from local renewable sources.
Conclusions
The optimal configuration for hydrogen supply chain deployment in Corsica is decentralized: multiple small production and storage sites distributed across the island, with minimal inter-zone transport.
The achievable LCOH ranges from €6.5 to €8.5/kg depending on assumptions — a competitive level for an island ZNI starting from zero infrastructure.
Electricity cost is the dominant competitiveness factor, contributing nearly 80% of the final cost.
These models serve as a direct decision-support tool for local authorities defining their deployment strategy. They are also transposable to other non-interconnected zones — Réunion, the Caribbean, Polynesia — by adapting parameters to local specificities such as available energy resources, industrial applications and demand profiles.
Limitations and Perspectives
Identified improvement areas include: integration of industrial hydrogen uses in end-use applications, addition of other production modes (thermal, geothermal depending on the territory), refinement of the risk index to provide an absolute qualitative value rather than a purely comparative one, and formal integration of social acceptability criteria in facility location decisions.
Future extensions of the study will cover the railway sector and the airport sector, both identified as high-potential applications not yet addressed in the current model.
Q&A
On rail transport in Corsica — The railway sector has not yet been integrated into the model but represents a high-potential application for heavy mobility decarbonization. It will be addressed in future versions of the study, alongside the airport sector.
On the 7 kg CO₂/kg H₂ result — This result corresponds to a mixed grid and renewables approach, not a 100% grid scenario. The ratio varies by period and configuration.
On truck retrofitting — The estimated all-in CAPEX for a hydrogen-retrofitted truck is approximately €800,000.
On migrating existing service stations — The model identifies stations with sufficient space to accommodate a hydrogen station, in a logic of continuity with existing user habits. Social acceptability has not yet been formally integrated as a criterion — this is an identified improvement axis.
On model transposability — The code was designed to be modular and easily adaptable to other island territories, by modifying available energy resources and end-use applications to match local specificities.
On the sector slowdown since 2022 — The initial retreat was linked to an overestimation of hydrogen's potential for light vehicles, in the face of increasingly competitive battery technology. The sector is now refocusing on heavy mobility — trucks, buses, trains, ships — and hydrogen combustion engines, which appear to be regaining momentum.
On ferry shore power — A dedicated study is underway to supply ferries at berth via fuel cells and locally produced green hydrogen. In the logistics model, the ports of Ajaccio and Bastia would constitute the two main demand hubs.
