Characterising the role of parametric functions in the van Genuchten empirical model on CO2 storage performance

Abstract

In the application of two-phase flow in porous media within the context of CO2 sequestration, a non-wetting phase is used to displace a wetting phase residing in-situ to the maximum extent through a network of pore conduits. The storage performance of this physical process can be assessed through numerical simulations where transport properties are usually described using the Brooks & Corey (BC) or van Genuchten (vG) model. The empirical constant, namely the pore geometry index, is a primary parameter in both of these models and experimental evidence shows a variation in the value of this empirical constant. It is, therefore, essential to cast this empirical constant into a ternary diagram for all types of clastic porous media to demarcate the efficiency of two-phase flow processes in terms of the pore geometry index (PGI). In doing so, this approach can be used as a tool for designing more efficient processes, as well as for the normative characterisation of two-phase flow, taking into consideration the predominance of capillary pressure or relative permeability effects. This concept is based on the existence of a PGI estimation for clastic sediments, for which the value for 12 sediment mixtures fall between 1.01 and 3.00. Statistical data obtained from soil physics is used for developing and validating numerical models where a good match is observed in numerical simulations. In this context, a new methodology for the effective characterisation of PGI for different clastic rocks is proposed. This paper presents theoretical observations and continuum-scale numerical simulation results of a PGI characterisation for the prediction of the hydraulic properties of clastic reservoir rocks. The effect of key parameters in the vG empirical model, such as the pressure strength coefficient and the PGI, is incorporated into the simulation analysis. In particular, the model is used to investigate the effects of parameter representation on CO2 storage performance in a saline aquifer. Subsequent analysis shows that the PGI is a very important parameter for defining the flow characteristics of simulation models. It can also be flexibly changed for each rock type and this approach may thus be practical when simulating the evolution of CO2 plume in reservoirs with sedimentary heterogeneities, such as intra-aquifer aquitard layers or graded beds. The use of the realistic PGI boundaries promises a more precise description of the hydraulic behaviour in sandstones and shale when using either the BC or vG model.