Ivan Buntic ist Mitarbeiter am Lehrstuhl für Hydromechanik und Hydrosystemmodellierung (LH2). Er ist Mitglied der SimTech Graduate School (GS SimTech) und wird am Donnerstag, den 12. März 2026 seine Doktorarbeit mit dem Titel "Data-driven local dynamic model adaptivity for subsurface gas storage, complemented by experimental investigations of microbial processes " verteidigen.
Datum: Donnerstag, den 12. März 2026
Uhrzeit: 9:30 Uhr
Ort: MML, U. 1.003, Pfaffenwaldring 61, 70569 Stuttgart
Abstract
In the course of the transition from fossil to renewable energy sources, new challenges arise for the efficient storage and utilisation of energy. One promising option is to convert surplus energy into chemical form, such as hydrogen or synthetic methane, which can then be stored underground and extracted when needed. This thesis examines underground gas storage in aquifers from multiple perspectives to enable more reliable and efficient planning of such processes. It makes three core contributions that combine numerical simulation, data-driven optimisation, and experimental analysis.
The first contribution concerns the numerical simulation of methane injection into aquifers. To this end, a classical full-dimensional (FD), two-phase model is coupled with a simplified vertical equilibrium (VE) model. While the FD approach provides highly accurate results, it is computationally demanding. The VE model is much more efficient but produces reliable results only under phase-equilibrium conditions. The scheme developed in this thesis adaptively decomposes the simulated domain during runtime so that the FD model is applied in regions dominated by strong flow dynamics, while the VE model is used elsewhere. The resulting system of equations is assembled monolithically and solved implicitly.
The second contribution focuses on data-driven acceleration of the adaptive scheme. Although the coupled model already reduces computational effort, complex coupling interfaces can introduce significant overhead. By selectively embedding data-driven surrogate models to replace the most expensive components, we achieve an overall speed-up of up to 75% while maintaining accuracy and mass conservation.
The third contribution investigates the influence of microbes on the wettability of stored hydrogen. We used a captive-bubble cell to measure contact angles between brine, hydrogen and quartz/sandstone surfaces. After enriching the brine with sulphate-reducing bacteria and allowing a biofilm to form, we observed a shift in wettability towards more brine-wet conditions. However, the magnitude of this effect depends strongly on the surface roughness of the samples, the specific bacterial strain, and the experimental conditions.
Kontakt
