ESC Lab simulates the influence of pressure and dilution on methane synthesis

Methane synthesis reactors convert green hydrogen and CO₂ from industrial processes into synthetic methane - a carbon-neutral energy source with a high calorific value. The reactors operate at 300-350 °C and 1-10 bar using Ni/Al₂O₃ catalysts. One-dimensional simulation models that map complex chemical and physical processes are used to optimise them. Despite high computing times, they provide important insights. The ESC Lab is developing a metamodel to predict the effects of pressure and dilution, which is to be integrated into power-to-X applications and digital twins.

To train the meta-model, a data set is generated from 5,000 simulations of the reactor model. Among other things, the gas inlet composition, temperature, pressure, flow rate, reactor length and reaction progress are varied. A 6th order polynomial, a feedforward neural network and Gaussian processes are used for training.

The aim is to develop a fast, data-driven prediction model for optimising power-to-X processes and digital twins.

Key Insights:

• With over 220,000 simulation points, we have created a robust training dataset.
• The neural network provides the best predictions for the reactor temperature over the reactor length.
• The effect of pressure and dilution on the methane content is well modelled by the neural network.
• The Gaussian processes provide the best results for hydrogen.
• Overall, the metamodels capture the initial parameters well, but improvements are needed, in particular for the temperature distribution in the reactor. A higher data set quality could further increase the model performance.

The work of our scientists Tim Franken, Monang Vadivala, Saurabh Sharma, Tobias Gloesslein, Arnim Brüger and Fabian Mauß supports the development of sustainable, CO₂-neutral energy carriers based on green hydrogen and CO₂ from industrial waste gases.

Last week, they presented their results at the 2nd General Meeting of COST Action CYPHER in Krakow, May 2025 https://cypher.ulb.be/event/cypher-2nd-general-meeting/