Determination of effective flow properties of caprocks in presence of multi-scale heterogeneous flow elements

Determining the effective fluid retention capacity of caprocks is of importance to petroleum exploration and production, and to CO2 storage in the subsurface. Thick successions of mud-rich sediments can comprise a range of genetic units (GUs), including mud-dominated hemipelagites, channel-levees, and mass-transport deposits. These GUs are characteristic of slope settings within the deep-sea environment. Based on measured properties of small samples, mud-rich sediments within these GUs ought to be high-quality caprocks with very low permeability and high capillary entry pressure (Yang & Aplin, 2007). However, each GU also contains objects of much more permeable materials such as fine sands and silty sediments. Those bodies are in the form of sub-seismic, typically centimeter thick, thin layers, which may or may not be deformed during the process of basin evolution, but they can have considerable lateral continuity. Moreover, GUs may contain deformation-related flow conduits, such as sandy injectites and open fractures, which function as pathways for vertical fluid migration. These GUs may themselves be arranged in complex patterns due to depositional and deformational processes. Such multi-scale heterogeneity in caprocks can affect their fluid retention capacity.



Estimating the effective fluid retention capacity of caprocks involves the quantification of properties at a sample (10-2m) scale in the laboratory. Sample compressibility, permeability and threshold Capillary Entry Pressure (CEPt) could be constrained well by lithology (e.g. sand/silt/clay ratio) of sediments, and at normal conditions they could encapsulate finer-scale heterogeneity satisfactorily too. However, the direct use of those properties on blocks of 1000mx1000mx100m size in basin simulations is highly undesirable and risky if each block corresponds to full or part of one or more heterogeneous GUs. Therefore, the lab-based properties must be upscaled to the block scales through multi-stage upscaling. To do this, we need: “to recognize the geometric arrangements at appropriate length scales; to construct models that capture these arrangements; to populate them with appropriate small-scale properties; and to run flow simulations from which it is possible to determine the effective properties at the target scales” (Aplin et al., 2012). This approach also provides knowledge about the types of sedimentary and mechanical architecture that forms the critical controls, and the conditions under which these features affect fluid flow at chosen length scales. We have developed a set of techniques to recognize structural arrangements from core, well-log, outcrop, and seismic data, to construct constrained models from these observations, and to calculate effective properties by flow simulation.