Tim Brünnette from the Department of Stochastic Simulation and Safety Research for Hydrosystems (LS3) of the University of Stuttgart is a doctoral researcher (research project B04) and member of the SFB 1313 Integrated Research Training Group "Interface-Driven Multi-Field Processes in Porous Media". He visited Stanford University from February to April 2025.
Research Report
I met Prof. Tartakovsky for the first time when he was visiting the University of Stuttgart and our working group as part of his Argyris Professorship 2023 award. Exciting discussions during that time eventually led to this collaboration.
The main aim of the research stay at Stanford was to explore surrogate modeling techniques for irregular fracture propagation, with the goal of improving computational efficiency while maintaining physical accuracy. This is particularly relevant for applications such as hydraulic fracturing, carbon sequestration, and reservoir engineering, where computational cost is a critical bottleneck.
Stanford University’s Department of Energy Resources Engineering has invaluable expertise in the areas of uncertainty quantification and geomechanics. During my stay, I made use of GEOS, Stanford’s in-house simulation code designed for high-performance modeling of fracture and deformation in geological materials. The research focused on creating and evaluating surrogate models that could serve as computationally inexpensive replacements for certain components of the full physical model, especially in contexts where traditional numerical methods become prohibitively expensive.
Fracture processes with heterogeneities pose a significant challenge for both modeling and simulation due to their sensitivities to structural features and boundary conditions. We investigated data-driven approaches like Gaussian process regression, conformalized predictions and spezialized neural networks to eventually tackle the inverse problem: Can we identify heterogeneous structural features and boundary conditions from fracture patterns?
Additionally, we also explored whether our methodology can be transferred towards the application of fractures in batteries, a topic that has been present in Prof. Tartakovsky's group for a while.
It was an honor to spend time at Stanford. The place is rightfully renowned for both its academic excellence as well as the beautiful campus. I greatly benefitted from the open discussions and support from the Tartakovsky group and am thankful for the opportunity to realize this stay abroad.
