Lattice Boltzmann modelling of nonlinear coupled pore scale processes in porous media systems and the applications in geological engineering

Abstract

The coupled subsurface flow involving heat transfer (T), hydraulic flow (H) and chemical reactions (C) in highly heterogeneous geomaterials are essential but also challenging to model in geosciences and geological engineering. To study such complex subsurface processes, multiple scales spanning from the nanoscale to basin scales are required regarding diverse applications. The pore scale is the dominant physical mediator of the above coupled fluid dynamics; hence many scientific and practical engineering decisions for the basin, reservoir and near-wellbore models have been driven by the averages of the pore scale physical processes governing fluid flow.With respect to the pore scale studies, the following key issues still remain as problems:1) How to appropriately describe and model the transport phenomena at the physically important scales, and provide a unifying framework to deal with a range of such multi-physics (T, H, C) coupled issues, rather than typical models tailored to address a single problem?2) How to fundamentally describe and investigate the resultant effect of pore scale coupled (T, H, C) processes on the representative parameter - permeability? Is it size dependent and how to define it if yes?The thesis aims to study the System (T, H, C) as a whole coupled problem. A general numerical framework has been proposed by applying and extending multi-disciplinary technologies for the solutions, which is capable of simulating coupled processes in seemingly different geological problems. The contributions are summarized as:1) Measuring porous media tortuosity and mineral distributions regarding different micro-CT scan resolutions. The advanced modern imaging (as revealed by the micro-CT scanning and QEMSCAN mineralogy mapping) is used to derive the 3D digital sample including materials analysis. The relative results provide the pore scale characterisation of complex real-world materials (e.g. sedimentary rocks), definition of the conductive pore network, properties of complex materials, and geometry heterogeneity. This serves as the first step for the whole thesis study, for building a pore scale heterogeneous geomaterials map with diverse intersecting pores and minerals for further fluid transport simulations;2) Describing the coupled transport phenomena of the System (T, H, C) at the pore scale for heterogeneous sedimentary rocks. A lattice Boltzmann method (LBM) based multi-physics computational model is developed and implemented here as an effective algorithm to address these three fully coupled geological parameters. The model extends the conventional lattice Boltzmann model, and provides simulations including: fluid transport through a combination of different matrix structures with disparate permeability, heat transfer in bulk fluids and solid matrix, and the simultaneous presence of chemical reactions with reactive minerals. Through benchmarks and experimental validation, the model enables coupled System (T, H, C) modelling with the temporal evolution at the pore scale, subjected to thermal and chemical reaction (dissolution/precipitation) effects in ways that honor subsurface petrophysical conditions;3) Effective parallel computing for massive data sets generated from high resolution micro-CT scans.A parallel supercomputing algorithm is developed to address massive data sets and the large scale simulations on supercomputers. Through various benchmarks and practical applications, the effective implementation of the parallel computational XLBM (extended LBM) platform has achieved a nearly-linear speedup and nearly-orthogonal extensibility, and it is effective to handle massive data sets from Gigabytes to Terabytes as an initial input in high-resolution or large scale computational analysis. The computing efficiency still reaches around 93.8% when 512 processors are used, which confirms the high scalability of the custom XLBM and provides capability for massive data sets analysis; 4) Evaluating the 4D permeability and the sensitivity of sample size (scale).A reliable 4D permeability evaluation model is introduced with respect to the modelling of the general coupled problem System (T, H, C) and the resultant effect on the pore scale geostructures (dynamic porosity change). Through the available experiments, the developed computational software provides reliable prediction for specific carbonic acid reactions: the 4D permeability approaches a constant value when the whole system reaches the reactive transport equilibrium status over a long period of time (e.g. three years); and the sensitivity analysis of sample size (scale) on the representative parameter – permeability, shows that the permeability can be determined as the average parameter from the pore scale, when the size is large enough to reach the representative size.The thesis presents a coordinated and extended multidisciplinary approach to coupled System (T, H, C) in fractured porous media. The methods and the algorithm developed in this study are general and can be extended to deal with a wide range of nonlinear coupled flow problems in porous media.