Fracture Modelling and Geothermal Lithium

Lithium is a critical mineral in the fight against climate change:  it is used in electrical batteries for computing, in electric vehicles, and as local electrical storage for smoothing flow from intermittent sustainable power sources. According to the IEA, in 2023 lithium supply was mainly limited to China, Chile and Australia (85% for mining and 96% for refining), associating lithium supply with high geopolitical risk; a risk to which the UK and EU are exposed.

The UK has a world-class lithium resource in Cornwall, as mineable granite, but lithium is also dissolved in geothermal brines occupying fractures. These fluids have lithium concentrations at approximately 100 ppm (at 2000 m), but they also have order of magnitude lower levels of Na⁺, Mg²⁺ and Ca²⁺ compared with other brine deposits, which makes lithium extraction simpler. Furthermore, the geothermal nature of the brines may allow production plants to be powered by sustainable energy. The question remains, how much lithium-rich brine can be extracted? Here petrophysical fracture modelling can help.

This research reports on some of the modelling technology that can be used to understand lithium-rich brine flow during extraction. It is important to consider aspects of fracture connected volume and connectivity, and to find pragmatic quantitative methods for assessing and reporting such data. Fracture connectivity depends on the number of nodes where fractures interact, and the distance between nodes. Studies of these have been found to be fractal. If that is the case in Cornwall, it implies that aspects of the fracture network at different scales can be fractally extrapolated from measurements made at smaller or larger scales. Connected fracture volume is controlled by fracture length and aperture. These are also fractally distributed. Consequently, a reasonably reliable multiscale 3D model can be constructed in Fracman or FracpaQ.

The aperture, and to some extent the fracture length, changes as the stress regime changes. For example, significant brine drawdown could reduce the flow rate because  external stress acts to close fractures when the fracture fluid pressure is reduced, and hence also reduce connectivity. By contrast, a significant injection of brine from which lithium and heat has been extracted would have the opposite effect. Quantification of this can be carried out using electrical methods as well as non-invasive 3D imaging (CT or micro-CT). Consequently, it is important for the fracture model to be responsive to the changing stresses in the model that might result from different stress tensors and production scenarios.

Finally, geothermal brine flow is also controlled by the roughness of fracture surfaces, especially as fractures close during drawdown. The interacting asperities on the surfaces increase the tortuosity of fluid flow significantly, but they also prop fractures open when they would otherwise close. The fracture surfaces are also fractal, and this work shows both models of fractal fracture surfaces and the fluid flow through them. Examples are given which show that uncompressed fractal fracture surfaces with a fractal dimension of 2.349 can reduce fluid flow, in our scenario by 28%.