Modelling Energy Systems through Faulted and Fractured Media

Faults and fracture systems are critical components within naturally fractured reservoirs which can influence fluid flow within the subsurface, and therefore it is vital to understand the organisation and interactions of these networks. Developing suitable methodologies to acquire and analyse structural data and to simulate fluid flow through the fractured media is crucial for improving efficiency and reducing uncertainties in modelling subsurface energy systems. Fracture data can be acquired from various sources across the surface domain (outcrops) and the subsurface domain (seismic and wellbore). Outcrop analogues provide excellent exposures for the analysis and characterisation of fracture networks that occur within reservoir units where subsurface data is limited or unavailable. Network properties such as fracture arrangement, distribution, and connectivity can be obtained through imaging fractured outcrop sections using various photogrammetry and sensor techniques. However, key challenges remain concerning the integration of outcrop data within subsurface models due to the complexities inherent in outcrop fracture networks, local and regional variations in outcrop geometry, and scale invariances between the surface and the subsurface. As such, it is important to develop and utilise methods to reduce uncertainties associated with combining the different datasets across scales to improve our understanding of subsurface processes. This thesis utilises outcrop studies from two underexplored regions (northern Bavaria and southern Tunisia) and applies novel techniques to characterise and homogenise outcrop fracture networks, to integrate surface and subsurface data within fracture models that simulate fluid flow and to assess spatial variability in outcrops with varying network parameters. To integrate outcrop fracture properties into subsurface reservoirs, this thesis introduces a new approach to upscale and homogenise traced fracture networks into a 2D permeability tensor using numerical simulation. The method incorporates efficient processes for meshing and simulation to reduce computation time and enable variation of individual and grouped fracture permeability for a more realistic output. Application of this method in northern Bavaria demonstrates the effect of seismic-scale faults on the permeability tensor of outcrop fracture networks. A critical distance is observed at which deformation occurs around large-scale faults, indicating that the connected fault system present in the region may act as an important control on fractured reservoir flow. Secondly, a key challenge lies in integrating useful fractured outcrop data with available subsurface data within fracture models. Outcrop data can provide important parameters for small-scale networks that may not be visible at the resolution of subsurface datasets or may be overlooked due to limited spatial coverage (i.e., seismic and well data). To address this challenge, this thesis introduces a workflow that combines several methodologies and data types to model subsurface fracture networks for numerical fluid flow simulation. Outcrop data obtained from fractured pavements is combined with subsurface FMI data to drive a stochastic DFN fracture model representing small-scale networks within the reservoir. This output is integrated with a deterministic fracture model derived from unconventional 3D seismic attributes that visualises the large-scale fracture and fault framework present in the subsurface. The resulting multiscale fracture network model is then used to derive additional network properties through analytical aperture modelling and the application of the 2D network upscaling approach to derive permeability tensors that define the preferred fluid flow orientation. Using a case study from southern Tunisia, this approach to structural characterisation and modelling found that large-scale features are connected through small-scale fractures orientated at 212° which contribute significantly to fluid flow direction and reservoir permeability. This highlights the importance of small-scale networks within models and demonstrates how the workflow integrates large- and small-scale networks in naturally fractures reservoirs. The integration of surface data within subsurface fracture characterisation and modelling can also present challenges due to the inherent variability of fracture networks at the outcrop scale. Therefore, quantifying variability within complex fracture networks is crucial to reducing the uncertainties associated with incorporating outcrop data into subsurface models. Spatial graph-based fingerprinting and hierarchical clustering of sampled outcrop data from the northern Bavaria case study indicates that intra-network and adjacent faulting play a key role in controlling spatial network heterogeneity. The effects of network variation on fluid flow are further investigated using the 2D upscaled permeability approach combined with analytical aperture modelling. The results of these simulations demonstrate that network variability is influenced by both sampling orientation and the presence of faults, suggesting a potential relationship within the sampled region. This understanding of outcrop network variation is important for the integration of data, as well as for the continued modelling and upscaling of complex fracture networks within reservoir-scale realisations. The methods and workflows presented in this thesis contribute to reducing uncertainties associated with integrating multi-scale surface and subsurface data into subsurface reservoir models. By improving the characterisation of fracture networks and their controls on fluid flow, this work enhances our understanding of regional tectonic influences and provides a robust framework for modelling fluid flow behaviour within complex naturally fractured energy systems.