Geologic Hydrogen Extraction

Enhanced Hydrogen Production (EHP) requires injection of reactive fluids in a reservoir rock from one
well, and extraction of reaction products, including gaseous hydrogen, from a neighboring well. In this project, we propose to leverage geomechanics modeling successes for Enhanced Geothermal Systems (EGS) and Carbon Capture Underground Storage (CCUS) to simulate microstructure changes and fracture propagation in mafic rocks during EHP, upscale microscopic predictions to the metric scale, and assess the feasibility of EHP under realistic conditions. A Phase Field (PF) model of serpentinization was developed to simulate the production of dissolved hydrogen by oxidation or iron oxide. Simulations captured the production of solid magnetite and the production and diffusion of liquid hydrogen upon oxidation of grains of various shapes and composition, and upon oxidation of various grain packings. A deep learning model was also trained and tested to infer complex fracture paths in homogeneous and pre-cracked media, which will accelerate simulation of chemo-hydro-mechanical crack propagation in future developments of the PF serpentinization model.

Enhanced Geothermal Systems

An Enhanced Geothermal System (EGS) refers to the technology that is used to circulate a fluid
through a hot rock formation to harvest heat when that rock does not naturally produce a pressurized heat carrier. An EGS is a conduction-based installation made of injection and production wells connected
by a network of fractures, which are most often artificially created by rock stimulation to ensure a sufficient flow rate. The wells are usually drilled vertically up to a depth close to that of the geothermal
reservoir, and then laterally, to increase the specific area of the fracture network in the reservoir layers. Because de-risking the EGS technology for direct heat production requires drilling optimization, we are developing an algorithm to optimize the alignment of the lateral wells that would be drilled if Cornell were to build an EGS on campus. We are also building an EGS forward modeling framework using the Finite Element Method (FEM), from which we aim to calculate stress and fracture distributions by minimizing the difference between in situ measurements and FEM simulation results.

Hydraulic Fracturing

It has been observed that hydraulic fractures that interact with weak interlayers sometimes deflect and propagate laterally before reinitiating into adjacent layers. Despite being commonly noted in layered formations, this phenomenon, known as offset propagation, remains poorly understood. To bridge this gap, a novel fully coupled damage-hydro-mechanical phase-field model has been formulated. The model captures mixed mode fracture propagation, damage-dependent poroelastic degradation, and anisotropic permeability evolution induced by fracturing. The model is implemented in the open-source finite element framework MOOSE, designed for large-scale multiphysics simulations. Numerical simulations reveal that fracture propagation within weak interlayers is governed by shear-dominated failure, while reinitiation into the overlying layer may result from a local competition between fluid pressure evolution and in-situ stress contrast near the fracture tip. Results provide new insights into the interactions between hydraulic fractures and weak or soft interlayers, and highlight the mechanical and hydraulic conditions that govern offset propagation in layered formations.

Aggregate Crushing – Mining

This project aims to unravel the impacts of mineralogy on aggregate crushing and understand sequential fragmentation mechanisms in granular assemblies via multi-scale experimental, numerical, and machine learning (ML) investigations. Particle crushing occurs in railway ballast, granular fault gouge, high tailing dams, and is also relevant to pile installation and offshore foundation design. Crushing and grinding are essential in mining operations, as well as manufacturing processes in the pharmaceutical, agricultural and food sectors. Most natural geomaterials are polymineralic, and yet, there is no known experimental method that can disentangle the effects of morphology and mineralogy on aggregate crushing. Despite the insights gained by statistical fracture mechanics and thermodynamic models, there is still a lack of understanding of the fundamental mechanical processes that govern the sequential fragmentation of natural polymineralic particulate assemblies. Grinding and crushing operations are still highly energy inefficient. We are working on the 3D microstructure reconstruction of polymineralic aggregates and on the numerical modeling of polymineralic crushing with the Discrete Element Method (DEM).