Multiscale modelling of miscible interfaces: Application on surfactant-enhanced aquifer remediation

In situ chemically enhanced solubilisation through the application of surfactants in injection solutions is a promising method to recover these contaminants. The process, nonetheless, involves dynamic evolution of phase boundaries, and multiple driving forces interacting with each other including capillary and viscous forces. On one hand, quantifying the contribution of the main mechanisms of NAPL recovery (mobilisation of NAPL ganglia through reduced interfacial tension and enhanced solubilisation), requires integration of fundamental physics of interfaces with continuum scale predictive tools. On the hand, the textural heterogeneities that span across multiple length scales of the porous medium complicates predicting the migration paths of the contaminant species, impacts the displacement mechanisms and adds uncertainty to the success of remediation operation in terms of recovery efficiency. Despite accumulated interest in modelling of the processes that involve dynamically evolving interfaces and underpinned science of miscible displacement through recent advances in Computational Fluid Dynamics (CFD), microfluidic studies, micromodel experiments and thermodynamics of interfaces, the impact of multiscale heterogeneity on the mechanisms of NAPL recovery has not been systematically quantified. Our overall aim is therefore to address incorporation of micro- and macro-heterogeneities of the porous system, by devising a novel multiscale computational apparatus that integrates dynamics of miscible displacement in the context of surfactant-enhanced aquifer remediation.

In WP1 we numerically undertake a dimensionless analysis of interacting driving forces in Darcy scale. In particular we will utilise dimensionless numbers including Damköhler number (to represent mass transfer to advection ratio), Péclet number, viscosity ratio, and geostatistical parameters of the absolute permeability distribution such as spatial correlation lengths. We construct a flow-regime diagram and delineate extent of interacting viscous and chemical dissolution fronts instigated by, respectively, viscosity difference between fluids and permeability-feedback mechanism. We demonstrate the interplay of permeability heterogeneities (in various forms such as channelised fluvial systems, long spatially correlated distributions, Gaussian permeability realisations, etc.) on the interaction of viscous and chemical dissolution fingering, and overall NAPL recovery.

In WP2 we seek innovative pore network modelling to underpin the physical processes (ganglia snap-off and mobilisation vs. interphase diffusion and mass transfer) that crucially shape the displacement mechanisms at microscale. We use a CFD theoretical model of interface evolution and rigorous transport model of viscous and chemical displacement. We upscale the results of flow and transport solutions from pore-scale to obtain Representative-Elementary-Volume-averaged multiphase flow and transport macroscopic properties. Through novel pore network generation techniques, we delineate the effect of pore-level statistics, morphology and structure on upscaled properties, and reduce the reliance over from commonly used empirical correlations.

In WP3 we integrate the two-scales of modelling through a novel spatio-temporal adaptive computational apparatus that will provide unique insights into underlying physical phenomena that determine the efficiency of surfactant-enhanced aquifer remediation processes. Beyond the specific application of the novel multiscale tool for aquifer remediation, the computational apparatus will serve the purpose of various disciplines of engineering, such as waste treatment, geological carbon sequestration, enhanced oil recovery, drug delivery, etc. where interphase mass transfer across dynamic interfaces is a ubiquitous feature.

Underpinning science of flow and transport in porous media for groundwater management and decontamination, efficient recovery from hydrocarbon reservoirs, sea water intrusion and safe storage of carbon dioxide to mitigate greenhouse gas emissions are parts of the challenge to secure sustainable access to water, energy and food. This project aims to enhance our understanding, reduce the uncertainty and increase the accuracy of one of the most complex phenomena of subsurface flow in porous media: miscible displacement with mobilisation of ganglia of a previously stagnant phase and interphase mass across interfaces of two phases under presence of micro- and macro-heterogeneities. Miscibility ubiquitously occurs for systems where the mass transfer between phases are hydrodynamically and thermodynamically significant. In miscible displacement, phases of fluid mix and segregate according to the coupled hydrodynamic-thermodynamic conditions, e.g., surfactant-enhanced water mixes with contaminant oil phase (NAPL) or oil left-in-place and mobilises them, or CO2 dissolves into aquifer brine and is safely stored in aquifers. The UK has strategic energy plans for future water, enhanced oil recovery and carbon storage. The project will have positive feedback and impact on the accuracy of predictive tools used for modelling these operations. Moreover UK has a strong expertise on contaminated land revival and remediation as it is one the first countries to suffer from environmental issues due to industrialisation (the UK has over 400,000 hectares of contaminated land according to the UK Government). Therefore the project serves the legacy of clean-up by an innovative, computationally efficient, multiscale methodology. The multiscale algorithm developed in this research will have an impact on other disciplines of engineering where multiscale nature of physical processes require resorting to adaptive switching between grid resolutions to compromise optimally between accuracy and computational expenses.

On a shorter timescale, environmental agencies active in the area of soil remediation can benefit from this work by cost reduction: the experiments to calculate the macroscopic multiphase flow and transport properties are expensive and dynamic interfaces infer variations in these properties with respect to change in species concentrations and thermodynamic states of phases, therefore even further experimentations will be necessary to capture the dynamics over macroscopic properties. Multiscale nature of this work at the same time is of interest to research centres. With the advent of computational power, linking between multiple length scales of physical processes is becoming feasible. Our development aligns with this trend.

On a longer timescale, industrial interest and linkages will be made through knowledge transfer to companies, via for example, the Knowledge Transfer Networks, where the PI is a member of its Energy group/community with activities towards Carbon Abatement and Oil & Gas which this project offer services to. A matter of interest for application of multiscale modelling tool is for groundwater remediation and protection with a full-scale economic analysis. Such analyses inevitably require full consideration of hydrogeological, hydrodynamic, geophysical, operational, and economic uncertainties. The full-scale predictive modelling tool for NAPL removal in multiscale heterogeneous aquifers developed as a result of this project will allow computationally efficient implementation of a stochastic analysis for cost-effective operational design, incorporating spatial and parametric uncertainties. In economic terms, the project has an ultimate goal in serving the economy of aquifer remediation as the societies value groundwater and pay for its protection for future generations. The economy of aquifer remediation processes will be benefited by enhanced modelling accuracy and predictive capacity that this project brings forth.