Physical heterogeneity in rocks is ubiquitous over a wide range of scales, from the grain-pore structure itself up to the field scale, with features such as deformation bands and faults which may span hundreds of metres.
These heterogeneities can strongly influence the flow of fluids in the subsurface, placing leading order controls on physical and chemical transport. Understanding how these heterogeneities impact multiphase flow is of therefore of critical importance to many subsurface processes, including the storage of CO2 during sequestration operations and during groundwater remediation.
In this talk, I will present work highlighting the impacts of physical heterogeneity across scales using a combination of experimental and modelling work undertaken in the past three years. Firstly, I will discuss the role of heterogeneities from the pore to continuum scale, primarily through micro and medical X-Ray CT experiments, which allow the direct visualisation of interfacial movement at the pore scale. Optimising these imaging techniques allows us to image over large, deca-centimetre core samples, meaning we can directly relate pore-scale, immiscible displacement mechanisms to continuum scale properties such as relative permeability and capillary pressure. 3D continuum-scale numerical models with hysteresis, and pore-network models, are directly validated using imaged fluid configurations and connectivity, across drainage and imbibition with simultaneous differential pressure measurements.
In the second half of the talk, I focus on how heterogeneities impact multiphase flow at larger, field scales. Using the previous characterisation methods, on cm – m scale samples, we characterise the multiphase flow properties through 60m of a target CO2 storage site in the North Sea, UK. Alongside well logs from the field, this allows the generation of realistic reservoir models, with defined heterogeneities constrained by field measurements. These models are used to run high-resolution multiphase flow simulations, elucidating the impacts of heterogeneities on the migration of CO2. Further to this, I show how macroscopic invasion percolation upscaling methods allow the incorporation of these impacts using heavily reduced computational meshes - a necessity for tractable simulations at large scale in 3D. These methods are applied to a km-scale model of the Goldeneye reservoir, North Sea, UK to understand the rapid CO2 plume migration and elongation that has been seen at many field sites such as Sleipner, In Salah and Otway.