HyPTraDe: Global Hydrogen Production Transport and Demand Model Chain

What is HyPTraDe used for?

The HyPTraDe model landscape consists of a hydrogen production model (HyP), a hydrogen transport model (Tra), and a demand regionalization model (De). Using this model chain, costs along the green hydrogen value chain up to the importing country can be determined. For this purpose, pure production costs, land potentials, transport and conversion costs, as well as the worldwide hydrogen demands of different countries are considered to depict  a global hydrogen market. The flexible handling of the model  allows the calculation of  sensitivities and thus determine the influence of input parameters on hydrogen costs. Modifications regarding techno-economic input parameters, capital costs as well as hydrogen demand patterns can be considered.

Which research questions can be answered?

  • How do country-specific capital costs affect the total cost of hydrogen?
  • How are the components of the production system optimally dimensioned in terms of cost?
  • How much hydrogen is available globally and at what cost?
  • What are the current and future costs of producing and transporting green hydrogen?
  • How large are the export and import flows between countries and continents?
  • At what cost can countries or regions supply themselves with hydrogen?
  • What effect does pooling of electricity procurement options at the NUTS-3 level have on the production costs of hydrogen?

Model Structure

The HyPTraDe model chain allows the calculation of global hydrogen costs by using  three initially separate building blocks (production, transport, and demand), which are then integrated into a market model.

Figure 1: HyPTraDe Model chain

The modeling of hydrogen production is based on island systems, represented in cells with a spatial resolution of 50×50 km, incorporating hourly data on renewable energy production capacity factors for each cell. Together with the techno-economic input parameters of the individual components of the hydrogen production system, the cost-optimal size of the system can be determined.  The hydrogen demand for each cell, needed as framework condition in the production model, can be modified as needed  to integrate the requirements of the demand side and assess  its impact on the optimal production system.

Figure 2: Components for the island system in the production model

Furthermore, there are spatial exclusion criteria (e.g., restrictive land use or water scarcity) for each cell, which limit the available area for hydrogen production. By aggregating land availability potentials, hydrogen cost potential curves can be generated for any desired geographical resolution.

Figure 3: Area potential analysis for Argentina

Additionally, the spatial resolution of the production model for Europe can be changed from a cell-based structure to NUTS-3 regions. Based on this, analyses can be carried out for RFNBO-compliant systems.

The transport model simulates the transport of green hydrogen using various transport technologies (ship, pipeline, and truck) and hydrogen carriers or derivatives (gaseous hydrogen, liquid hydrogen, ammonia, and LOHC). It determines the most cost-effective transport route and considers conversion, conditioning, as well as loading and unloading processes. The determination of the cost-optimal transport route can include the combination of multiple transport technologies and hydrogen carriers. It provides insights into transport routes and enables the identification of bottlenecks and requirements for infrastructure expansion.

Figure 4: Example of optimized transport routes

The data basis for the demand regionalization model is formed by internationally published reports and forecasts on future hydrogen demand. To determine country-specific demand, regionalization criteria such as gross domestic product, population, environmental regulations, and hydrogen production and infrastructure projects are weighted and evaluated.

The three main elements of the HyPTraDe model chain are calculated separately and integrated into the market model. This model simulates global hydrogen trade, linking production potentials with demand quantities.  The total costs of global hydrogen demand (production and transport costs) are optimized, in order to achieve a Pareto optimum for the sum of the total costs of all countries.

The potentials per cell serve as the basis for simulating global hydrogen trade and meeting the hydrogen demand of the world countries.

The allocation optimization problem considers additional parameters such as the limitation of transport capacities, national self-supply constraints, and international agreements. The results are presented as a Sankey diagram. Moreover, cost potential curves can be generated for each country or region. Using the market model, questions regarding the development of national and international hydrogen markets, including the integration of derivatives, can be addressed.

Figure 5: HyPTraDe Model chain exemplary result diagrams

Technical details of the model chain

The development of the HyPTraDe model chain is carried out using Python, PostgreSQL, and the FfE-internal geo-database FREM.

The models for production, transport, and the market model are solved through linear programming (LP). The technical implementation of the production model with hourly resolution is done in Python using the PyPSA package (Python for Power System Analysis), which is specifically developed for energy system modeling. This allows the consideration of various hourly renewable energy sources for electricity generation to power the electrolyzer. The optimization problem is defined by Linopy and solved using the Gurobi solver. All components of the island system are dimensioned in this process. A constant, normalized demand of 1kg/h or a more flexible demand to simulate production processes such as ammonia can be considered.

The transport model is solved using the pgRouting algorithm in PostgreSQL to determine the optimal route, the optimal H2 transport carrier, and the most cost-effective transport technology between production and demand clusters. The technical implementation is also based on a cell-level structure. The transport route per cell can run in 8 directions (horizontal, vertical, or diagonal) starting from the center.

The demand regionalization model and market model are technically implemented in Python and directly connected to the FfE geo-database to retrieve input parameters and store results. The regionalization of demand enables its distribution at a national level and utilizes the entropy method to determine the weights of the various regionalization factors.

Ultimately, the market model defines an allocation optimization problem based on input data from the production model, transport model, and demand regionalization modell. The optimization problem is solved using Gurobi. Approximately 400 production clusters and 200 countries are considered for the supply of hydrogen and hydrogen derivatives. Additional constraints to better represent a future hydrogen market, such as transport capacities and self-supply restrictions, have also been implemented in Python.

Current projects