Learning from data — Calibration and improvement of modeling thermo-osmosis effects in THM simulations based on the Mont Terri Deep Borehole experiment

This study presents results of an effort to validate the numerical implementation of the thermo-hydro-mechanical model with thermo-osmosis in the finite-element simulator OpenGeoSys. The simulator’s capability to capture thermo-osmotic effects on pressure fields under thermal gradients is evaluated by comparison with a reference analytical solution and field data from the Mont Terri Deep Borehole experiment. The results demonstrate that the thermo-osmosis significantly influences subsurface pressure dynamics, and OpenGeoSys effectively reproduces this phenomenon. Additionally, an improved formulation of the THM process is outlined, and a model of the experiment is extended by inhomogeneous effects to reproduce the observed data better. This work illustrates how combining analytical verification, numerical modeling and field observations can improve the interpretation of complex THM processes in clay formations. • Validation of simulation software for thermo-osmosis based on borehole data. • Discussion of sources of uncertainty in thermo-osmosis models. • Tests of different hypotheses on physical phenomena to improve fit to data. • Variable geothermal gradients and permeability heterogeneity act synergistically. • Heterogeneity is important and has to be considered appropriately in the models.