Published:
Dynamic fracture.LEFM. Phase-field and continuum damage modeling of different localized failure modes in porous, fluid-saturated geomaterials. Focus on ice calving in glaciers and ice sheets, hydraulic fracking, and compaction band propagation in geological reservoirs.
Published:
Increasing climate volatility demands understanding post-failure behavior and designing resilient, safe-to-fail infrastructure. Large deformations (SPH): Coupled hydromechanical SPH for geomechanical problems such as fault rupture, retrogressive landslides and embankment failure, tailings dam stability/failure, dynamic & static liquefaction. Wave/debris flow–structure interaction: Simulating waves impacting coastal/offshore structures (e.g., wind turbines). Scalability: Large-scale 3D simulations leveraging HPC. Future: Regional-scale simulation of dam and levee failure (e.g., overtopping, storm surge), and 2D/3D dam-break flood modeling to support emergency action planning.Figure showing development of a retrogressive flowslide type landslide: 
Published:
▶ Porous sedimentary rock, compaction bands form perpendicular to 𝜎1: grain crushing, cataclastis, porosity/permeability reduction. Geological reservoirs, CO2 storage, and hazardous waste. ▶ Elastoplastic damage + grain crushing + pore collapse instability to model compaction bands in porous rocks. ▶ Constitutive models often phenomenological. User-bias. ▶ Shale sedimentary rock, clay mineral makeup give ductility and self-healing. Low porosity/permeability → seal or cap rock. ▶ Shale is laminated and transversely isotropic. But hyperelastic models used in shale computational modeling designed for rubbers. ▶ Constitutive artificial neural networks (CANNs) automatically discover hyperelastic model. CANNs (Linka and Kuhl, 2023) ensure physically interpretable, thermodynamically and mechanistically constrained neural networks. ▶ Design CANN architecture to incorporate transverse isotropy. Energy Geomechanics: Using my coupled solid-fluid hydromechanical models and damage model in SPH method to model fluid flow reduction across compaction bands. Upscale to basin-scale modeling of fluid injection or recovery
Published:
Using DEM method and experiments + PIV, failure and localization precursors in granular media, granular force chains. Dynamic loading of granular materials. Circumvent phenomenological constitutive models Particulate nature of discrete element method (DEM) avoids constitutive models. Emergent behavior. Focus on high-resolution models of accretionary wedges, and fault propagation direction. Failure precursors and intermittent or “weak shear bands
Published:
Volcano caldera collapse, Global-scale simulations of fault systems or tectonic plates (microplate models). Coupled SPH-DEM modeling to allow for emergent behavior.