Modelling the Hidden Fractures of the Subsurface

As a senior researcher at DTU Offshore, Michael Welch is working at the intersection of geoscience and energy technology. With a background in structural geology and a deep expertise in fracture modelling and geomechanics, his work plays a crucial role in ensuring the safe and efficient development of carbon storage and geothermal energy systems.

Understanding the Role of Fractures in Subsurface Systems

Fractures - cracks within rock formations - may be invisible to the naked eye in seismic surveys, but they have a significant impact on how fluids like water, CO₂, or even pollutants move underground. Michael’s research focuses on predicting the location, geometry, and connectivity of these natural fractures, especially in low-permeability rocks such as chalk or granite.

"Fractures often dictate how fluids travel underground. Whether we’re storing CO₂, extracting geothermal energy, or managing aquifers, understanding these fractures is essential for designing effective systems,” he explains.

Because direct imaging of fractures is limited, Michael develops tools and models that use available geological and mechanical data to estimate their presence. One such tool is a Petrel plug-in, created in collaboration with his team, that integrates mechanical rock properties, subsurface geometry, and stress conditions to predict likely fracture zones. Originally developed for chalk reservoirs, the tool is now being adapted for CO₂ storage top seals and geothermal formations.

Michael’s work supports a range of energy transition technologies. In carbon storage, understanding the potential for fractures in top seals - the layers meant to trap injected CO₂ - is critical to avoiding leaks and ensuring long-term containment. In geothermal energy, especially in crystalline rocks like granite, predicting and harnessing natural fractures can improve energy extraction efficiency.

In both cases, Michael emphasizes the balance between scientific rigor and practical application. “We always need to strike a balance between the complexity of our models and the availability of data. It’s about making our research usable and impactful,” he says.