To ensure storage security, it is vital we understand, and can effectively model, the physical and chemical trapping mechanisms for CO2 storage. A key trapping mechanism underpinning storage security and immobilising a significant proportion of the CO2 plume is capillary trapping. Capillary trapping is considered to be responsible for 90% of the storage capacity in saline aquifers in the US, the largest potential CO2 storage resource. The ability to model and predict capillary trapping over large spatial scales in complex geological systems is essential to minimise risks and evaluate capacity. This paper investigates the effect of natural rock heterogeneities on capillary trapping across spatial scales in the Captain D Sandstone, from the Goldeneye formation, North Sea. That is, how heterogeneities affect CO2 saturation and distribution within rock core samples, and how those effects manifest, as a pseudo residual trapping, when upscaled to the field. A comprehensive dataset of 48 core plugs over a 65m interval have been studied. The location has industrial application as a target injection site for the discontinued Peterhead CCS project, with the aim to store 10 Mt CO2 over 10 years. These results are also applicable to other storage sites of similar geology. We have carried out steady-state core flooding experiments using medical x-ray CT, providing a detailed characterisation of continuum multiphase flow properties, including residual trapping characteristics, over cm scales. The results show that individual core plugs exhibit a large range in apparent residual trapping at the centimetre scale which averages out over scales of 10s of cm’s or greater. The average Land trapping value at the deca-centimetre scale is 1.7, however, at the cm scale, it varies between 0.8 and 2.8, representing a very large variation in locally trapped CO2. Consequently, at the plug scale for large initial saturations, approaching 1, the apparent residual saturation may vary between 0.2 and 0.5, given the variations in Land trapping parameter. At lower initial saturations, the variations are not as great in absolute terms, but the relative variation is just as large. Continuum numerical models are constructed from observations of residual trapping in heterogeneous rock cores using characterisation methods developed by Jackson et al. (2018, 2020). The trapping stage of a CO2 injection and migration project is simulated over a 65m region, while retaining the impact of heterogeneities at the cm-scale. Results from these models indicate that great care should be taken when upscaling the Land trapping parameter for use in field simulations, to estimate the long-term trapping of CO2 by capillary heterogeneities. Through quantifying the proportion of trapping resulting from capillary heterogeneities, the implications of modelling the system with the average and maximum experimental Land trapping parameter are evaluated. Modelling the system with the maximum experimental core plug Land trapping parameter results in the proportion of trapping caused by capillary heterogeneities to approximately double compared to that obtained using the average experimental core plug Land trapping parameter.
Read full abstract