PESA ETSIG/CSIRO CCS Knowledge Transfer Series: Installment 3

The practice of reservoir engineering and simulation in CCS begins with the standard physics of multiphase flow in porous media, but with CO₂-specific properties to be represented, especially the solubility of CO₂ in brine and the matching changes in brine density, and relative permeability effects. Four major code comparison studies have been carried out over the last twenty years, mostly to cross-validate simulator performance (of both people and software), as well as the most recent comparison with laboratory experiments.

The questions addressed by CCS simulation are quite distinct from most hydrocarbon recovery work: one is modelling large volumes of CO₂ injection (millions of tonnes) into saline formations, where the resulting CO₂ plume migrates laterally over distances of kilometres, and in timeframes of up to a thousand years after injection ceases. Feasibility studies need to address the range of uncertainties in the plume footprint, which stem from the uncertainties in geological characterisation of surface topography, sub-seismic faults, and permeability heterogeneity both laterally and vertically. The pressures increase due to injection may necessitate the design of relief wells. Simulations are needed to design monitoring and verification plans, and to interpret the data which is gathered. Regulatory compliance requires forward predictions of plume evolution, which can be checked against monitoring results to ensure the CO₂ is ‘behaving as predicted’. It may also be necessary to assess the risk of CO₂ injection affecting other resources, such as groundwater, hydrocarbons, geothermal or storage projects (for natural gas or hydrogen).

The least conventional aspects of CCS simulation involve the coupling of fluid flow to additional physics. Thermal effects are important in the wellbore and the near-well environment, with the cooling effects of injection reducing the maximum allowable injection pressure. Geochemical interactions with the reservoir rock or seal can potentially aid storage by mineralising CO₂, and this is the focus of in-situ carbonation research. Injection can also induce seismicity, and this potential can be modelled as well as monitored during operations. Finally, the rise of artificial intelligence and machine learning has prompted research into ways to accelerate the modelling process and allow fast history-matching.

Overall, CCS provides many fascinating challenges for subsurface modelling, and the scope for this work is likely to expand significantly in Australia as more projects progress through feasibility studies towards implementation.