# InterPore2018 New Orleans

US/Central
New Orleans, USA

#### New Orleans, USA

Description

REGISTRATION FOR INTERPORE 2018 IS AVAILABLE NOW. YOU CAN REGISTER HERE.

Louisiana State University and Tulane University are organizing the Interpore 10th Annual Meeting and Jubilee Conference from May 14th to 17th, 2018, in the historic city of New Orleans in the USA.

Frenchman, Sieur de Bienville founded the city of La nouvelle Orleans in the year 1718. Following a series of swaps between Spain and France, the New Orleans became part of the US in the Louisiana Purchase of 1803 New Orleans has since grown into a city of global culture with strong French, Spanish, African, and American influences. It is a city of great food, music and culture.

The 10th Annual Meeting and Jubilee celebration will be held at the New Orleans Ernest N. Morial Convention Center, in walking distance from the historic French Quarter and the Riverwalk. Conference hotels will be adjacent to Convention Center in the Warehouse and Arts Districts. Plan to attend, engage in cutting edge research on porous media, and enjoy one of the most historic cities in America.

Conference technical program, important dates, accommodation details will be available shortly on the InterPore site.

The local organizing committee looks forward to welcoming you to New Orleans in 2018.

Topics and applications

 Transport phenomena Mathematical and computational modeling Interfacial behavior and multiphase flow Multiscale and multiphysics processes Experimental advances CO2 sequestration Imaging, microscopy, and micromodels Swelling porous media Reservoir simulation Groundwater hydrology Geomechanics and fractured materials Soil mechanics and engineering Pore-scale modeling and upscaling Reactive flows Wave propagation Filters and membranes Fibers, wood, paper, and textiles Cements and construction materials Food and consumer products Biofilms, bone, & tissue Geothermal energy Biotechnology Composites and foams Fuel cells and batteries Novel porous media applications Nanoporous materials

Registration
InterPore 2018 Participant Registration (Payment in Euro)
InterPore 2018 Participant Registration (Payment in USD)
Short Course Registration
Contact
• Monday, 14 May
• 09:45 12:30
Parallel 1-A
• 09:47
Modeling CO2 Storage in Fractured Reservoirs: Fracture-Matrix Interactions of Supercritical and Dissolved CO2 15m

The injection and storage of supercritical CO2 (scCO2) have been conducted in fractured sandstone reservoirs at In Salah, Algeria and Snøhvit, Norway, and planned in fractured sandstone, carbonate, and dolomite reservoirs at Longyearbyen, Norway, Hontomin, Spain, and Kevin Dome, USA, respectively, with matrix permeability varying from 0.01 to 60 md. For densely fractured reservoirs with low matrix permeability (e.g., at Longyearbyen, Norway), injected scCO2 can dissolve into the resident brine at fracture-matrix interfaces and the dissolved CO2 (dsCO2) can diffuse into the rock matrix making solubility trapping the dominant trapping mechanism. For fractured reservoirs with intermediate matrix permeability (e.g., at In Salah, Algeria), the storage of scCO2 in the rock matrix dominates with strong fracture-matrix interactions observed through field monitoring at In Salah. We developed a comprehensive conceptual model for enhanced CO2 storage to account for global migration of scCO2 in the fracture continuum, local storage of scCO2 and dsCO2 in the matrix continuum, driving forces for scCO2 invasion and dsCO2 diffusion from fractures, and brine outflow through connected matrix blocks.

For the dominant matrix scCO2 storage, we developed high-resolution fracture-matrix models for individual matrix blocks, homogeneous columns of fractures and matrix blocks, and heterogeneous REVs consisting of multiple columns of matrix blocks with varying flow properties and sizes. The multiscale modeling results show that the equilibrium efficiency of local scCO2 storage strongly depends on matrix entry capillary pressure, matrix-matrix connectivity, and reservoir thickness, while dynamic efficiency and transfer function are also sensitive to fracture spacing and matrix flow properties. The transfer functions calculated for various REVs were used along with reservoir-scale dynamics of scCO2 plume flow in fractures, showing that the preferential migration of scCO2 through fractures is coupled with bulk dsCO2 storage in the rock matrix that in turn retards the scCO2 fracture plume. The bulk matrix storage is mainly driven by buoyancy between fracture scCO2 and matrix brine and facilitated by matrix-matrix connectivity that allows displaced brine to outflow, enabling the rock matrix to act like an open system. Conventional dual-continuum models cannot capture these processes because they model isolated matrix blocks with no capillary continuity, thereby underestimating storage efficiency.

For the dominant matrix dsCO2 storage, we developed the unified-form equations of diffusive flux of dsCO2 into brine-bearing matrix blocks of varying shapes (i.e., spheres, cylinders, slabs, squares, cubes, rectangles, and rectangular parallelepipeds) and sizes (Zhou et al., 2017a, b). We then applied the flux equations to a fractured reservoir with various scenarios of matrix blocks by assuming 1-D and 2-D radial scCO2 flow in fractures and by using diffusion of dsCO2 from fracture-matrix interfaces into matrix blocks as the sink for scCO2 in fractures. For each scenario, the dynamic dsCO2 plume with different mass fraction was produced analytically, showing that solubility trapping is significant in fractured reservoirs with low matrix permeability and small fracture spacing.

Speaker: Quanlin ZHOU (Lawrence Berkeley National Laboratory)
• 10:05
The impact of drainage displacement patterns and Haines jumps on CO2 storage efficiency 15m

Injection of CO2 deep underground into porous rocks, such as saline aquifers, appears to be a promising tool for reducing CO2 emissions and the consequent climate change. During this process CO2 displaces brine from individual pores and the sequence in which this happens determines the efficiency with which the rock is filled with CO2 at the large scale. The aim of this work is to better understand the impact of different flow regimes, during immiscible two-phase flow, on the displacement and storage efficiency of CO2 deep in saline aquifers. Using multi-GPU free energy Lattice Boltzmann simulations we directly solve the hydrodynamic equations of motion on a three dimensional geometry reconstructed from micro-CT images of Ketton limestone and consider fluid flows in a range of capillary numbers Ca and viscosity ratios. We first verify the existence of the three typical fluid displacement patterns, namely viscous fingering, capillary fingering and stable displacement [1]. We examine how these distinctively different flow regimes can affect the displacement efficiency, defined here as the fraction of the defending wetting fluid that has been displaced from the pore matrix when the injected non-wetting phase reached the outlet of the domain. Continuing the injection beyond this point we establish the maximum displacement efficiency or storage capacity. Our results indicate that the maximum displacement efficiency decreases with decreasing Ca. As capillary fingering becomes the dominant displacement process at low Ca, storage efficiency converges to a limiting value irrespective of the viscosity ratio.

Particular focus is given to the low Ca flow regime, where displacements at the pore scale typically happen by sudden jumps in the position of the interface between brine and CO2, Haines jumps. We demonstrate that the method reproduces the expected features of the jumps, i.e. sharp increase in the non-wetting phase velocity, abrupt drop in the pressure signal and significant fluid rearrangement. We quantify the degree of fluid redistribution associated with these sharp events by identifying each event from the pressure signal. Preliminary results from this analysis suggest that pressure fluctuations and waiting times between the jumps follow an exponential distribution, in agreement with theoretical predictions, while the same also applies for the event filling volumes probably due to the extensive fluid redistribution. More importantly a significant decrease in storage efficiency is observed, irrespective of the direction of the jump relative to the overall flow direction, contrary to the arguments by Yamabe et al. [2]. This is due to irreversible fluid rearrangement during Haines jumps that alters the displacement pathways and renders regions of the porous rock inaccessible to the injected non-wetting fluid. This has important implications in the context of geological sequestration of CO2, as Haines jumps become a limiting factor in the storage process.

Speaker: Dr. Ioannis Zacharoudiou (Imperial College London)
• 10:23
Modeling the dissolution-driven convection as a Rayleigh-Benard problem 15m

We examine the linear and weakly nonlinear stability analyses of the dissolution-driven convection induced by the sequestration of carbon dioxide in a geological formation. The mathematical model consists of Darcy's equation, the conservation of mass and the conservation of solute equations. The model accounts for anisotropy in both carbon diffusion and permeability which is modeled by a decaying exponential function of depth. The presence of a first order reaction between the carbon-rich brine and host mineralogy is also included. We prescribe either Neumann or Dirichlet boundary condition for the concentration of carbon dioxide at the rigid upper and lower walls that bound a layer of infinite horizontal extent. We consider a Rayleigh-Taylor-like base state consisting of a carbon-rich heavy layer overlying a carbon-free lighter layer and seek the critical thickness at which this configuration becomes unstable. With this approach, standard mathematical methods that were successfully used in the study of Rayleigh-Benard convection can be applied to this problem. We quantify the influence of carbon diffusion anisotropy, permeability dependence on depth and the presence of the chemical reaction on the threshold instability conditions and associated flow patterns using the classical normal modes approach. The critical Rayleigh number and corresponding wavenumber are found to be independent of the depth of the formation. The weakly nonlinear analysis is performed using long wavelength asymptotics, the validity of which is limited to small Damk\"{o}hler numbers. We derive analytical expressions for the solute flux at the interface, the location of which corresponds to the minimum depth of the boundary layer at which instability sets in. We show that the interface acts as a sink leading to the formation of a self-organized exchange between descending carbon-rich brine and ascending carbon free brine. Plots of the high order perturbation terms for the concentration successfully reproduce the fingering pattern that is typically observed in experiments and full numerical simulations. Using the derived interface flux conditions, we put forth differential equations for the time evolution of the upward migration of the interface as the dissolution process progresses. We solve for the terminal time when the interface reaches the top boundary thereby quantifying the time it takes for an initial amount of injected super-critical Carbon dioxide to be completely dissolved. We also consider the case where the interface migration is accompanied by interface deformations that conform to the convection pattern.

Speaker: Prof. Layachi Hadji (The University of Alabama)
• 11:00
Vertically-Integrated Dual-Continuum Models for CO2 Injection in Fractured Saline Aquifers 15m

Injection of CO2 into a saline aquifer leads to a two-phase flow system, including a supercritical CO2 phase and a brine phase. Various modeling approaches, including fully three-dimensional (3D) models and vertical-equilibrium (VE) models, have been used to study the system in unfractured formations. Three-dimensional models solve the governing flow equations in three spatial dimensions and are applicable to generic geological formations. VE models assume rapid and complete buoyant segregation of the two fluid phases, resulting in vertical pressure equilibrium and allowing integration of the governing equations in the vertical dimension. This reduction in dimensionality makes VE models computationally much more efficient, but the associated assumptions restrict the applicability of VE model to formations with moderate to high permeability.

In this presentation, we extend the VE and 3D models to simulate CO2 injection in fractured aquifers. This is done in the context of dual-continuum modeling, where the fractured formation is modeled as an overlap of two continuous domains, one representing the fractures and the other representing the rock matrix. Both domains are treated as porous media continua and, as such, can be modeled by either a VE or a 3D formulation. The transfer of fluid mass between fractures and rock matrix is represented by a mass transfer function connecting the two domains. Because the fracture domain is usually much more permeable than the matrix domain, we apply VE modeling to the fracture domain but not the matrix domain. We refer to the resulting model as a hybrid VE-3D model, with the VE model applied to the highly permeable fractures and the 3D model in the less permeable rock matrix.

Our hybrid VE-3D model includes both dual-porosity and dual-permeability types. The dual-porosity model conceptualizes the rock matrix as sugar-cubes that are isolated uniformly by vertical and horizontal fractures, or as match-sticks that are isolated by vertical fractures through the entire thickness of the aquifer. In contrast, the dual-permeability model explicitly represents the 3D flow dynamics in the rock matrix. We derive mass transfer functions that couple the VE model in the fracture to the different models in the rock matrix. We then apply the hybrid VE-3D model to simulate CO2 migration in fractured saline aquifers and compare with 3D-3D models where both the fracture and rock matrix are modeled in 3D. The hybrid VE-3D models are much more computationally efficient while providing results that are close to those from the 3D-3D models. These vertically-integrated dual-porosity and dual-permeability models provide a range of computationally efficient tools to model CO2 storage in fractured saline aquifers.

Speaker: Yiheng Tao (Princeton University)
• 11:18
Modification of wettability and interfacial tension by biosurfactant-producing bacteria for geologic carbon storage 15m

Injection of carbon dioxide (CO2) into deep geologic formations has been widely proposed as an effective way for the permanent storage of CO2. Modification of the interfacial properties of CO2 in minerals by using surfactant has been proposed aiming on increasing the mobility of CO2 through porous media. Surfactants are proven to effectively alter the interfacial tension and wettability in both CO2/water/mineral system, improving the displacement and sweep efficiencies of CO2 in porous media. In the meantime, biosurfactants have been drawing much attention as an alternative to the chemical surfactants for their biodegradability, ecological suitability and low toxicity. However, the question as to the extent of microbial alterations in fluid wettability and interfacial tension under reservoir pressure and temperature conditions still warrants further investigation. Therefore, this study investigated the role of lipopeptide biosurfactant on wettability and interfacial tension alterations in a CO2/brine/mineral system for different CO2 phases during the growth of thermotolerant and barotolerant bacteria, Bacillus subtilis, and the production of lipopeptide biosurfactant, surfactin. Quartz, mica and carbonate substrates were selected and used as representative minerals. While monitoring the changes in the interfacial tension and wettability with pH, fluid samples were acquired from the brine phase, and the concentrations of glucose, nitrate, ammonium and surfactin in the acquired samples were quantitatively assessed using various assays and spectroscopic methods. As a result of surfactin production by B. subtilis, we observed the reductions in interfacial tension and increases in contact angle at all tested cases. The concentration of surfactin and the rate of wettability alteration differed with the experimental conditions. The modification of CO2 wettability was the greatest for liquid CO2 while the least of modification was observed for gaseous CO2. The obtained results allow in-depth assessment of the feasibility of using biosurfactant-producing bacteria for effective geologic carbon storage practices.

Speaker: Mr. Taehyung Park (Korea Advanced Institute of Science and Technology (KAIST))
• 11:36
The impact of heterogeneity on the flow and trapping of CO2 in target UK aquifers 15m

The Bunter sandstone formation in the Southern North Sea and the Captain sandstone formation in the Northern North Sea represent two of the largest potential CO$_2$ stores in the UK, with estimated capacities of up to 14 Gt and 1.7 Gt respectively [1, 2]. With current UK CO$_2$ emission totalling ~400 Mt/yr [3], the Bunter and Captain formations alone have the potential to store UK emissions for many years.

In order to determine the long-term fate of the injected CO$_2$ in these systems, accurate characterisation of the multiphase flow behaviour and trapping is needed [4]. Conventionally, multiphase flow functions, namely relative permeability, capillary pressure and trapping, are derived from viscous limit core flood experiments, measured at high flow rates on subsurface rock cores preferentially selected for homogeneity [5], and either used directly in field scale modelling or for further upscaling.

However, for modelling low potential flows characteristic of buoyantly driven CO$_2$ plume migration, it is important to derive properties that capture the impacts of rock heterogeneity. Sub-metre scale capillary pressure heterogeneities will control local fluid distribution, resulting in equivalent relative permeabilties which are dependent on the flow direction, rock heterogeneity, and the capillary number [6]. Modelling studies have estimated that this can have a significant impact on plume migration and trapping from the mm-km scale [7,8]. However, no experimental protocols have been developed to inform the models with appropriate properties measured on heterogeneous rock samples in the laboratory.

To address the impacts of small scale heterogeneity on large scale flow and trapping of CO$_2$, we present a combined experimental and numerical study on rock cores from the Bunter and Captain sandstone formations. We analyse 38 small rock cores covering the entire 100m interval of the Captain D reservoir unit in the Northern North Sea, and a smaller selection of cores taken from the Bunter Sandstone in the Southern North Sea. We use a recently developed characterisation approach [9] to create a 3D numerical model of heterogeneous rock cores, based on laboratory observations. We incorporate hysteresis into the characterisation by building on the recent approach developed by [10]. Once characterised, the numerical cores can accurately predict equivalent relative permeabilties and trapping, dependent on the capillary number and direction of fluid flow.

The numerical models are then used to investigate multiphase flow hysteresis and trapping across the range of conditions estimated to prevail in the CO$_2$ storage reservoirs. Under these conditions, we systematically explore the impact of hysteresis and heterogeneity on flow and trapping at multiple scales. The migration of CO$_2$ may be significantly enhanced by heterogeneity when flow can align with the direction of layers. This situation may arise in gravity currents of plumes underneath a confining caprock layer. In contrast, flow is impeded by heterogeneity when the dominant direction crosses bedding layers, as may occur in predominantly upward buoyantly driven migration. In this case, the lowered mobility results in significant spreading of the plume and residual trapping is also enhanced.

Speaker: Samuel Jackson (Imperial College London)
• 11:54
Adaptive hybrid multilayer model coupling vertically-integrated and full multi-dimensional models for geological CO2 storage 15m

CO$_2$ injection into a saline aquifer leads to a two-phase flow system (supercritical CO$_2$ and brine), which often involves large spatial and temporal scales that require high computational cost. To address the computational challenge, in the past decade, a series of simplified models based on vertical integration of the full multi-dimensional governing equations have been developed. These vertically integrated models either assume a rapid segregation between CO$_2$ and brine due to strong buoyancy (i.e., vertical equilibrium assumption) or solve the one-dimensional vertical two-phase flow dynamics as fine-scale problems on top of the (coarse-scale) vertically integrated equations. The former is ofen referred to as vertical equilibrium (VE) model, while the latter relaxes the VE assumption and is called dynamic reconstruction (DR) model [1,2]. The major computational cost of the VE and DR models comes from solving the coarse-scale vertically integrated equations while the computation associated with the vertical reconstructions (either VE or DR) is minor. As such, they are much more computationally efficient than full multi-dimensional models and have been used to answer many important engineering questions. However, the vertically integrated VE or DR models are often limited to aquifers with homogeneous or layered heterogeneous properties. Thus, for aquifers with strong 3D heterogeneity, the computationally expensive 3D models are to date the only robust option.
In this talk, we present a hybrid multilayer framework to couple full multi-dimensional models with the various vertically integrated models. Such a framework allows us to use full multi-dimensional models in highly heterogeneous layers of an aquifer where full multi-dimensional model is the only robust option, while applying simplified vertically integrated models in layers with homogeneous or layered heterogeneous properties. We develop algorithms to couple the full multi-dimensional model with vertically integrated models (VE or DR), as well as algorithms for the coupling between the VE and DR models. In addition, we develop a local criterion to adaptively switch between VE and DR reconstructors [3], i.e., use VE reconstructor when the two fluid phases are in equilibrium while use DR reconstructor to capture vertical dynamics when the fluids deviate from vertical equilibrium. Comparisons with full multi-dimensional models (MRST [4] is used in our work) show that our adaptive hybrid multilayer model is much more computationally efficient than full multi-dimensional models while providing results with similar accurary, making this hybrid model an attractive tool for modeling of CO$_2$ injection and migration in highly heterogeneous saline aquifers.

Speaker: Tianyuan Zheng (Helmholtz Centre for Environmental Research)
• 12:12
Microfluidic Measurement of CO2-Oil Phase Behavior for Enhanced oil Recovery and CO2 Storage 15m

Carbon capture, storage, and utilization technologies target a reduction in net CO2 emissions to mitigate greenhouse gas effects. The largest such projects worldwide involve storing CO2 through enhanced oil recovery - a technologically and economically feasible approach which combines both storage and oil recovery. Successful implementation relies on detailed measurements of CO2–oil properties at relevant reservoir conditions (P = 2.0–13.0 MPa, and T = 23 and 50 ºC). In this paper, we demonstrate a microfluidic method to quantify the comprehensive suite of mutual properties of a CO2 and crude oil mixture including solubility, diffusivity, extraction pressure, minimum miscibility pressure (MMP), and contact angle. The time-lapse oil swelling/extraction in response to CO2 exposure under step-wise increasing pressure was quantified via fluorescence microscopy, using the inherent fluorescence property of the oil. The CO2 solubilities and diffusion coefficients were determined from the swelling process with measurements in strong agreement with previous results. The CO2–oil MMP was determined from the subsequent oil extraction process with measurements within 5% of previous values. In addition, the oil–CO2–silicon contact angle was measured throughout the process, with contact angle increasing with pressure. In contrast with conventional methods which require days and ~500 mL fluid sample, the approach here provides a comprehensive suite of measurements, 100-fold faster with less than 1 μL of sample, and an opportunity to better inform large scale CO2 projects.

Speaker: Mrs. Atena Sharbatian (University of Toronto)
• 09:45 12:30
Parallel 1-B
• 09:47
Experimental Studies on the Hydraulic Effects of Fungal-Mycelia in Sandy Soil 15m

One gram of soil can contain up to 100 million to 1 billion microrganisms and up to 1 million different species of microorganisms. Despite this fact, geotechnical engineers have, until fairly recently, ignored biological activity in the soil or possible biological amendments that could be introduced. Over the last ten years research has focused on bioaugmentation strategies (i.e. the injection of a single strain of bacteria) to alter hydraulic and mechanical behaviour of porous and fractured media (e.g. microbially induced calcite precipitation). One challenge of bioaugmentation technologies is the transportation of bacteria within the ground. This study investigates for the first time the potential use of fungal networks for ground engineering applications. Fungi produce hyphae, long filamentous structures which collectively are called a mycelium. Mycelium can grow to vast sizes, with individual mycelia (in forest floors) covering areas up to 9km2 in North America. As such there is great potential ‘to grow’ fungal mycelia for earth infrastructure over large areas.

We investigated the hydraulic behaviour of sandy soils treated with fungal mycelia using P. ostreatus (oyster mushroom) in order to: (i) Assess the level of hydrophobicity induced and (ii) understand the influence of P. ostreatus mycelia on water flow through the soil profile. To investigate these, we grew mycelia in petri dishes and conducted water drop penetration tests to ascertain induced hydrophobicity. We also determined the surface water evaporation rates for soils with mycelia and those without, at different starting moisture conditions. Next, we set up a 1-dimensional infiltration column test with mycelia inoculated and incubated to grow overtime throughout the soil profile. The infiltration column was instrumented with tensiometers and Time Domain Reflectometer (TDR) probes for the real-time measurement of suction and water content. Soil infiltration water fronts were obtained for both treated and untreated soils in respective columns. The presence of fungal mycelia resulted in significantly altered hydraulic characteristics of the soils. Mycelia induced extreme hydrophobicity on fine sands and reduced surface water evaporation rates. Infiltration time was slower for fungal-treated soils than untreated soils. These results highlight the potential for fungal mycelia to be used in the creation of semi-permeable or impermeable barriers in a range of ground engineering applications.

Key words: fungal mycelia; soil hydraulics; ground improvement; 1-D infiltration;

Speaker: Mr. Emmanuel Salifu (University of Strathclyde Glasgow and Università di Napoli Federico II, Napoli)
• 10:05
Seismic monitoring of biopolymer accumulation and permeability reduction in sands 15m

Bacterial colonization and the spread of biopolymer, gel-like material, on porous media are known to decrease permeability by several order of magnitude and to cause bioclogging thereby altering the hydraulic flow systems of porous media. Attention to microbial bioclogging has been increasing owing to the increasing demand of microbial soil treatment and soil improvement. Successful microbial bioclogging treatments require geophysical monitoring techniques to provide appropriate spatial and temporal information on bacterial growth and activities in the subsurface; such monitoring datasets can be used to evaluate the status of plugged sections and optimize re-treatment if the plug degrades. Therefore, this study investigated the feasibility of using P- and S-wave velocity and attenuation for monitoring the accumulation of bacterial biopolymers and the permeability variations during bioclogging. In sand-packs, Leconostoc mesenteroides was cultured and stimulated to produce insoluble biopolymer and generate bioclogging. During such bacterial bioclogging, permeability and high-frequency P- and S-wave responses were monitored. P-wave velocity was consistent and S-wave velocity was increased with biopolymer accumulation. Both P-and S-wave attenuation, evaluated by using spectral ratio method, were increased with increasing biopolymer saturation. Increases in seismic attenuation are closely linked to the biopolymer saturation and permeability reduction. Herein, we also presented a theoretical model to correlate biopolymer saturation, permeability, and seismic attenuation by modifying three-phase Biot model and Pride-Berryman double-porosity model.

Speaker: Dong-Hwa Noh
• 10:23
The effect of kinetics on the efficiency of biologically induced carbonate precipitation via urea hydrolysis for soil improvement applications 15m

Microbially and enzyme induced carbonate precipitation (MICP and EICP) via urea hydrolysis are emerging biological soil improvement techniques. In these techniques, a treatment solution including calcium chloride (CaCl2), urea, and either urease-producing bacteria or free urease enzyme are introduced into soil to precipitate calcium carbonate (CaCO3) and modify the mechanical properties of the soil. Both chemical efficiency and mechanical efficiency affect the feasibility, cost, and environmental impact of these techniques. Chemical efficiency is defined as the percentage of substrates converted to the desired product. Mechanical efficiency describes the amount of improvement in the targeted mechanical property as a function of the amount (percent by dry weight) of precipitated CaCO3. . To assess the efficiency of these techniques, the kinetics of urea hydrolysis and CaCO3 precipitation need to be considered. This study evaluates how the kinetics of these techniques affect both chemical and mechanical efficiency.
Chemical efficiency was evaluated through a systematic experimental design using different concentrations of substrate (CaCl2 and urea) and urease (bacterial cell or free enzyme). It was observed that chemical efficiency drastically drops when substrate concentration exceeds a specific value that is a function of the initial urease concentration. It was also demonstrated by these experiments and by experiments conducted by others that chemical efficiency of a microbial urease treatment solution is affected by the presence of seawater and by insufficient nutrient and air for microbial growth. Based upon these observations, a numerical model was developed to predict the effects of initial urease content, degradation and encapsulation of the cells or enzymes, pH, temperature, urea, and calcium concentration on the rate of hydrolysis and precipitation. Model predictions are in good agreement with the experimental results.
Mechanical efficiency is typically characterized by empirical correlations relating strength and stiffness to CaCO3 content. It is difficult to distinguish the effect of the reaction rate on these empirical correlations as the mechanical properties of the improved soil also depend on many other factors, including density, grain size distribution, particle shape and saturation during treatment. However, experiments have been performed to evaluate how reaction rate influences treatment uniformity and crystal morphology, both of which have an influence upon mechanical properties. Uniform treatment requires uniform distribution of urease, substrate, and nuclei in the pores, which depends on the employed treatment method (e.g. injection, soaking, percolation, mix-and-compact) and the recipe for (i.e., the constituents concentrations of) the treatment solution. MICP at a low hydrolysis rate, which occurs at low initial cell concentration and low temperature, prolongs the induction time, delays the onset of precipitation, and reduces the number of crystals. For microbial solutions with high initial cell concentration and high CaCl2 concentration, the solubility product of CaCO3 is exceeded within a short period and the supersaturation for calcium carbonate precipitation remains high for extended periods, resulting in prolonged nucleation time and consequently extended growth of non-stable CaCO3, which may adversely affect strength and durability. The effect of calcite seeds and the presence of seawater on morphology were also investigated. It was observed that presence of calcite seeds lowers the carbonate supersaturation level, facilitating calcite crystal formation. The presence of seawater was found to inhibit calcite crystal growth.

• 11:00
Optimizing field-scale MICP with multi-scale micro-continuum OpenFOAM modeling 15m

Microbially Induced Carbonate Precipitation (MICP) through the urea hydrolysis reaction has been extensively studied in the lab and implemented at field-scale several times, most notably for fracture sealing (Cuthbert et al., 2013; Phillips et al., 2016), for erosion control (Gomez et al., 2015), and for ground improvement (van Paassen et al., 2010). Grouting strategies used in industry are commonly based on experience derived from the injection of Ordinary Portland Cement, (e.g. use of the split-spacing method). Field-scale injection strategies for MICP are likely to differ considerably from traditional cement grout injections as:

i. the low viscosity of the injection fluids allows near-surface grouting with minimal risk of ground heave and the potential for larger soil/rock volumes to be treated around each injection point,
ii. strength improvement occurs without complete permeability reduction and multiple injections are required to incrementally reach the desired strength and permeability,
iii. flow velocity (to control bacteria attachment), pH adjustment (to control CaCO3 saturation state), and temperature (to control the rate of ureolysis) may all be used to limit blocking of the injection points, and
iv. abstraction boreholes may be required for the removal of waste ammonium.

We present here a multi-scale micro-continuum MICP model implemented in OpenFOAM and solving fluid flow with the Navier-Stokes equations. The model is intended to inform the choice of injection strategy used in field-scale pilot projects and solves for 1) bacteria injection, velocity dependent attachment, and encapsulation within precipitating CaCO3; 2) re-agent transport, urea hydrolysis and CO3 production with Michaelis-Menten kinetics and 3) CaCO3 precipitation, porosity reduction, and subsequent flow path alteration.

In this model injections can be driven by a constant flow rate, constant pressure, or stepped flow rate and can be planar flow (e.g. for groundwater movement) as well as radial flow from/ to multiple injection/ abstraction wells. The model includes the ability to import 2D and 3D data from image analysis software ImageJ allowing X-ray micro-CT results from lab scale experiments to be modeled (at pore or micro-continuum scale) for validation purposes, and for heterogeneous site conditions to be modelled (at continuum-scale). A parametric sweep function is included to assess sensitivity to the choice of parameter values.

Results show that grouting with MICP is fundamentally different to grouting with cement. Operators may wish to replace sequential injection through a series of boreholes with simultaneous injection through multiple boreholes, or use abstraction boreholes to both collect waste ammonium and direct the transport of MICP re-agents.

Speaker: James Minto (University of Strathclyde)
• 11:18
Microbially Induced Desaturation and Precipitation (MIDP) via Denitrification during Centrifugal Loading 15m

Microbially induced desaturation and precipitation (MIDP) via denitrification has the potential to mitigate earthquake-induced liquefaction by two mechanisms: biogenic gas production to desaturate and dampen pore pressure changes in soil and calcium carbonate precipitation to mechanically strengthen soil. Lab-scale tests have demonstrated that both desaturation and precipitation are effective mitigation mechanisms. However, small-scale laboratory column tests at ambient pressure lead to gas pockets and lenses, causing upheaval due to low overburden pressures. Therefore, biogenic gas formation, distribution, and retention need to be evaluated with more realistic over-burden pressures to understand the effectiveness of this treatment mechanism. Centrifuge tests of soil desaturated by MICP treatment are currently being performed to simulate field pressures and stresses. In addition, a numerical model was developed to evaluate the scaling effects on biogenic gas generation between the centrifuge model and prototype scale. The centrifuge tests are conducted within a laminar box on the 1-m radius centrifuge at the University of California, Davis NHERI/CGM centrifuge facility. Desaturation is induced in the laminar box prior to acceleration in the centrifuge by augmenting saturated soil with an enriched culture of denitrifying microorganisms. The models are accelerated to 80 g in stages and measurements of soil moisture content are made over time to see the combined influence of steady-state pore pressure and overburden pressure on the degree of saturation. Upon reaching the final centrifuge acceleration, the models are subjected to strong shaking until either liquefaction is triggered or the capacity of the centrifuge is reached. Test results provide evidence of the capacity for MIDP to mitigate the potential for earthquake-induced soil liquefaction by desaturation. Comparison of modeling results to test data suggest that the numerical model does not consider certain pore-scale influences and the effects of mixing from liquid-gas transfer and transport observed in the centrifuge tests. Thus, future work will add these features to the model.

Speaker: Caitlyn Hall
• 11:36
Pore Scale Simulation of Biogenic Gas Formation and Migration in Porous Media 15m

The biogenic gas behavior in porous media, which includes bubble nucleation and growth, migration, coalescence and trapping is affected by the gas generation rate, distribution of reactive sites and the pore scale characteristics of the sediment. In this study, experiments are performed using a micro-fluidic chip in which different gas bubble behavior mechanisms in the porous media are observed. Secondly, the behavior of biogenic gas is simulated using a pore-network model extracted from the 3D X-ray image of an in-situ sediment. The formation of biogenic gas bubbles is modeled using the classical gas nucleation theory. Several numerical algorithms and criteria developed for the expansion of gas bubbles during the biogenic gas formation, size-dependent rising velocity of gas bubbles, bubble coalescence, slug formation and movement, escaping, and trapping in the pore space. The amount of produced gas bubbles, residual gas saturation and hydraulic conductivity are calculated during the simulations. Results of the simulation are qualitatively compared with the microfluidic chip experiments.

Speaker: Dr. Nariman Mahabadi (Arizona State University)
• 11:54
'Microbial Mortar’- restoration of degraded marble structures with microbially induced carbonate precipitation 15m

Ancient stone relics and historic buildings are often subject to significant degradation. The protection and restoration of these monuments is extremely urgent. Here, a method of building repair based on microbial induced carbonate precipitation (MICP) has been tested on marble stone. In previous research, microbial mortar (stone powder treated by MICP) was tested as a filling material to repair cracks within stone. In this paper, the effect of microbial treatment on degraded marble consisting of larger particle sizes is studied. In the experiment, we focus on altering the permeability and porosity of crushed marble grains and show that the porosity and the permeability of the sample are notably decreased by carbonate precipitation.

MICP treatment is carried out in a column filled with marble grains with the injection of six batches over six days. A white CaCO3 precipitate is produced which matches the original marble colour and is sufficiently strong to cement the marble sand together from the inlet up to a depth of 150 mm into the column. To understand the micro-scale distribution of the CaCO3 precipitation within the column, and its effect on flow and transport properties, we analyse the MICP-treated column using X-ray CT with a resolution of around 3 microns. The X-ray CT scan data, support the macro-scale observations of a gradient in the degree of cementation along the direction of liquid flow, indicating that producing an evenly solidified sample is a problem that needs to be resolved.

We use the core-scale experimental data to derive cm-scale fluid transport properties using tracer breakthrough curves taken, prior to, and after MICP treatment. The fitted transport properties show that the fraction of pores containing mobile water decreases with increasing cycles of MICP. Pore-scale modelling using the X-ray CT data supports these findings, showing that cementation leads to a change in the pore network structure, with flow increasingly focussed into a smaller number of faster moving open channels.

Our experiments show that CaCO3 precipitation is greatest at the inlet. It is reasoned that this could be avoided by modification to the injection strategy. Prevention of re-agent mixing outside the marble grains, careful choice of marble grain size distribution, and tailored injection flow rates could deliver re-agents deeper into the media and take advantage of the formation of stable flow pathways to maximise seal uniformity. MICP is a promising technique for the restoration of marble structures and monuments.

Speaker: Prof. Rebecca Lunn (University of Strathclyde)
• 12:12
High Phylogenetic and Physiological Diversity of Ureolytic Bacteria in Native Soils Bio-stimulated for MICP 15m

Worldwide demand for new and sustainable approaches to geotechnical engineering problems has generated novel research opportunities in the emerging field of bio-mediated soil improvement. The most widely researched of these processes is microbially induced calcite precipitation (MICP), which has shown promise for a wide variety of engineering applications. Initially MICP was accomplished by bio-augmentation with a high density of the constitutively ureolytic bacterium, Sporosarcina pasteurii (Stocks Fischer et al., 1999). The amended soil was then supplemented with liquid medium containing calcium salts, urea, and sometimes growth-promoting organic compounds. Bacterial hydrolysis of urea generates a molecule of carbonic acid and two of ammonia. The resulting ammonia, a weak base, equilibrates with water and tends to form ammonium and hydroxide ions. This shifts the carbonic acid-bicarbonate-carbonate equilibrium toward carbonate, which will precipitate as calcium carbonate in the presence of sufficient calcium, ideally in the immediate vicinity of ureolytic bacteria, thereby cementing adjacent soil particles and increasing soil strength and stiffness. More recently bio-stimulated MICP has been fully demonstrated in native sands with prospects for eliminating costs and environmental impacts of propagating and transporting large quantities of bacteria. In our recent column experiments, completed on 14 different sandy soils from different depositional environments -- including several samples obtained from natural deposits as deep as 12 meters -- bio-stimulated MICP was always successful (Gomez et al. 2014; 2017; 2018).

Over 300 bacterial pure cultures were obtained from the most recently bio-stimulated soils and were stored (-80°C) to enable future physiological and genetic studies. A study of the urease kinetics of 8 randomly selected bacteria enriched in a meter-scale stimulated MICP demonstration (Gomez et al., 2017) showed that whole cell rates of urea hydrolysis follow Michaelis-Menten kinetics, with half-maximum values achieved at urea concentrations ranging from 56 to 837 mM and maximum rates varying from strain to strain by 100-fold. In progress 16S rRNA sequencing of the culture collection shows it includes a wide variety of ureolytic strains closely related to, but not identical to, the S. pasteurii (ATCC strain 11859), which has been employed almost universally in bio-augmentation experiments. Certain strains were found repeatedly in all or most of our bio-stimulated MICP experiments. Physiological differences between these strains will be discussed along with their surprisingly high diversity at the end of bio-stimulated MICP treatments. We have also begun to link pure culture physiology with relative abundance of these same strains in bio-stimulation treatments by extracting and amplifying bulk DNA from dilute aqueous bacterial suspension. In progress sequencing of 400 full-length clones is expected to provide a higher resolution but lower density sampling of diversity versus time for a single bio-stimulated MICP column. In parallel, high throughput 16S amplicon sequencing will provide much higher depth but lower resolution snapshots of changes in bacterial diversity throughout this same progression.

Speaker: Mr. Charles M.R. Graddy (University of California, Davis)
• 09:45 12:30
Parallel 1-E
• 09:47
Use of molecular simulations to fit EOS in confined space in order to perform large scale tight oil and shale gas reservoir simulations 15m

Unlike conventional reservoirs where pore size distribution has a micrometer scale (Nelson 2009), tight oil and shale gas reservoirs have predominantly mesopores (between 2 and 50 nm) and micropores (below 2 nm). Volume fraction of micropores is not negligible and can be as high as 20% (Kuila et Prasad 2011). As hydrocarbon molecules range between 0.5 and 10 nm (Nelson 2009), interaction forces between confined fluid and pore wall molecules become as significant as inter molecular interactions within the confined fluid. That is why nanofluidic experiments (Wang et al. 2014) and bubble point measurement on hydrocarbon mixture in mesoporous materials (Cho, Bartl et Deo 2017) have demonstrated that confinement considerably changes fluid phase behavior. Consequently the commonly-used equation of state (EOS) such as Peng-Robinson EOS is not able to describe the confined fluid phase behavior. A pore radius dependent EOS is therefore needed in reservoir simulators for accurate large scale tight oil and shale gas production forecast simulations.

The idea of this work is to integrate first the capillary pressure effect into the classical Peng-Robinson EOS and then to calibrate EOS parameters function of pore radius to fit molecular simulations. The capillary pressure which depends on pore radius adds pressure difference between vapor and liquid phase in the equilibrium computation. It is calculated using the Young-Laplace equation. Molecular simulation is performed using Monte Carlo method in the grand canonical and in the NVT Gibbs ensemble with anisotropic volume change in order to calculate equilibrium properties of several pure hydrocarbon components and mixtures in confinement. Kerogen pores are modelled by graphite slit pores and fluid/wall interaction potential is added. For a given pore radius, the critical temperature and pressure are determined for pure components, and liquid and vapor pressures, densities and molar fractions of components are calculated for both pure components and mixtures at different temperatures for calibration. These values are used as reference fitting data for the Peng-Robinson EOS with capillary pressure. The optimization parameters are the Peneloux volume correction constant, the acentric factor and the binary interaction coefficients. The calibration of these parameters allows getting correlations versus pore radius that will be used to model the confined fluid thermodynamic behavior.

The pore radius dependent EOS calibrated with molecular simulation data can therefore be used in reservoir simulators to accurately forecast tight oil and shale gas production. However the grid cells in a dynamic flow simulation is usually in order of several meters to 100m. Such a cell includes a large pore size distribution, the pore radius value used in the EOS is therefore an issue. In order to consider the pore size variability within a simulation cell, an effective radius function of oil saturation is taken. It is determined from the distribution function of pore size volume. Assuming oil is the wetting phase, during a flow simulation, oil is present in small pores and gas appears in larger pores, then the effective pore radius decreases. Oil and gas production simulations with a dual porosity model for a fractured tight-oil reservoir show that this methodology gives more reasonable results than using an average pore radius and of course than a bulk approach.

Speaker: Nicolas Sobecki (IFPEN)
• 10:05
Storage and recovery of multi-component mixtures in single shale pores 15m

Natural gas production from shale formations has received extensive attention in recent years. While great progress has been made in understanding the adsorption and transport of single-component gas (usually CH$_4$) inside shales’ nanopores, the adsorption and transport of multicomponent shale gas under more realistic reservoir conditions (e.g., considering CH$_4$/C$_2$H$_6$ mixture) only begun to be studied. In this work, we use molecular simulations to compute the storage of CH$_4$/C$_2$H$_6$ mixtures in single nanopores and their subsequent recovery. We show that, surface adsorption contributes greatly to the storage of CH$_4$ and C$_2$H$_6$ inside the pores, and C$_2$H$_6$ is enriched over CH$_4$. The enrichment of C$_2$H$_6$ is enhanced as the pore is narrowed, but is weakened as the pressure increases. These effects are captured by the ideal adsorbed solution (IAS) theory, but the theory overestimates the adsorption of both gases. We show that the recovery of gas mixtures inside the nanopores toward a gas bath approximately follows the diffusive scaling law. The ratio of the production rate of C$_2$H$_6$ and CH$_4$ is close to their initial mole ratio inside the pore despite that the mobility of pure C$_2$H$_6$ is much smaller than that of pure CH$_4$ inside the pores. By using scale analysis and by computing the Onsager coefficients for the transport of binary CH$_4$/C$_2$H$_6$ mixtures inside the nanopores, we show that the strong coupling between the transport of C$_2$H$_6$ and CH$_4$ is responsible for the effective recovery of C$_2$H$_6$ from the nanopores.

Speaker: Prof. Rui Qiao (Virginia Tech)
• 10:23
Molecular dynamics study of the occurrence states of gas-water mixtures near the organic solid in shale reservoirs 15m

Abstract
Understanding the gas occurrence states under real reservoir conditions is the prerequisite to study the mechanisms of gas flow in shale reservoirs, in which large amounts of nanoscale organic pores exist. Besides, water is inevitable when considering the gas flow in shales. Thus molecular dynamics simulations were performed to study the occurrence states of gas-water mixtures near the organic solid. Results indicate that methane will approach the organic surface spontaneously, accumulate at the solid-liquid interface and form a dense gas region finally. This process has little relationship with gas saturation. Potential of mean force (PMF) was calculated to explain the enrichment of methane. We found that both the wall-gas interaction and the water-gas interaction are beneficial to the adsorption of methane and the former takes the leading role, while the gas-gas interaction impedes the adsorption. In addition, the effects of reservoir temperature, pressure, rock wettability, and carbon dioxide (CO2) on the occurrence states of methane-water mixtures near the solid surface were studied. The temperature, rock wettability and CO2 influence the occurrence states near the surface obviously, while the pressure in the simulating range does not. This study also suggests the potential of thermal exploitation, altering the rock wettability and CO2 injection to enhance shale gas recovery.
Key words shale; gas-water mixtures; organic matter; occurrence state; adsorption; potential of mean force

Speaker: Mr. Zheng Li (China University of Petroleum (East China))
• 10:41
Break 18m
• 11:00
An Image-based Micro-continuum Pore-scale Model for Gas Transport in Organic-rich Shale 15m

Gas production from unconventional source rocks, such as ultra-tight shales, has increased significantly over the past decade. However, due to the extremely small pores (~ 1-100 nm) and the strong material heterogeneity, gas transport in shale is still not well understood which poses challenges for predictive field-scale simulations. In recent years, digital rock analysis has been applied to understand shale gas transport at the pore-scale. A widely recognized issue with rock images (e.g., FIB-SEM, nano-/micro-CT images) is the so-called “cutoff length”, i.e., pores and heterogeneities below the resolution cannot be resolved, which leads to two length scales (resolved features and unresolved sub-resolution features) that are challenging for flow simulations. Here we develop a micro-continuum model, modified from the classic Darcy-Brinkman-Stokes framework, that can naturally couple the resolved pores and the unresolved nano-porous regions. Gas flow in the resolved macropores is modeled with Stokes equation. For the unresolved regions where the pore sizes are below the image resolution, we treat them as a continuum and develop an apparent permeability model considering non-Darcy effects at the nanoscale including slip flow, Knudsen diffusion, adsorption/desorption, and surface diffusion. This leads to a micro-continuum pore-scale model that can simulate gas transport in 3D shale images. We present case studies to demonstrate the applicability of the model, where we apply the new micro-continuum model to 3D segmented FIB-SEM shale images that include four material constituents: organic matter, clay, granular minerals, and macropore. We populate the model with experimental measurements (e.g., pore size distribution of the sub-resolution pores) and parameters from the literature, and identify the relative importance of different physics on gas production. Overall, the micro-continuum model provides a novel tool for digital rock analysis of organic-rich shale.

Speaker: Bo Guo (Stanford University)
• 11:18
Efficient molecular simulations of binary gas mixture transport in slit nanopores 15m

A novel approach is suggested to simulate the gas mixture transport in slit nanopores. The proposed method is based on the modification of the dual control volume grand canonical molecular dynamics (DCV-GCMD) method. The conventional method, DCV-GCMD, describes the gas mixture transport with pre-set constant composition. Due to the selective adsorption in the nanopores, the composition of the produced gas mixture will be affected. The modified method provides the composition in the permeate side. Gas mixtures of CH4/He and CO2/CH4 are investigated in graphene, graphite, and clay slit nanopores. The results show that pore size is the dominant factor in the species separation; the solid surface roughness has pronounced effect on gas separation; the influence of average pressure is not pronounced. The effects of pore length, temperature, pressure gradient, and feed composition are also investigated.

Speaker: Dr. Tianhao Wu (Reservoir Engineering Research Institute (RERI))
• 11:36
On the surface diffusion phenomena in the organic nanopores of shale 15m

Adsorbed phase transport (also known as surface diffusion) has been reported to be one of the main transport mechanisms in organic nanopores of shale. Surface diffusion is a mechanism in which adsorbed molecules move over the surface by hopping between adjacent adsorption sites. From a molecular point of view, one of the main evidences for existence of this mechanism is the molecular simulation results of flow in graphite conduits with smooth structureless walls. The objective for this work is to fulfill the need for the investigation of the contribution of the adsorbed phase transport to the total mass flux of organic conduits with pore wall irregularities.
In this study, adsorption and transport of methane are studied using molecular simulations in three-layered graphite nano-scale channels with different degrees of surface roughness. To simulate the surface roughness, a fraction of the carbon atoms are randomly removed from the surface. At each channel roughness level, velocity profiles and mass fluxes of methane are computed. The simulations are performed at different pressures, pressure gradients, channel widths (2 and 4 nanometers), and levels of roughness (level roughness is characterized by the percentage of the removed atoms from the graphite surface). For each roughness level, three realizations are considered (each realization with same percentage of removed atoms) and averaged for the results not to be realization-specific. A comparison between the results in this work with Regularized 13-moments (R13) and Knudsen diffusion model is also performed.
Based on the molecular simulation results, as the pressure of methane increases, the number of gas molecules adsorbing to the graphite walls increases. Density profiles of methane at different pressures show two sets of density peaks; the first peaks correspond to the gas molecules adsorbed to the sections of the channel where the carbon atoms are removed. The second peaks correspond to the density of gas molecules adsorbed to the three-layered sections of the surface, where no carbon atoms are removed. The velocity profiles in the smooth channels (no atoms removed) are significantly larger than those in rough channels. In a case with only one row of the carbon atoms removed from the graphite surface, the velocity values are almost five time smaller than those in the smooth channel. The contributions of the adsorbed phase transport to the total mass flux in the rough channels are shown to be significantly lower than those in the smooth channels. The reason behind the high reported values of the adsorbed phase mass flux seems to be an artifact of using smooth structureless conduits in the previously published studies. Predictions made using the R13 and Knudsen diffusion models are significantly different from the results in the smooth channel but in agreement with the results in the channel with highest level of roughness (50-percent removed atoms).
This work is one of the few in-depth investigations of the contribution of surface diffusion to the total mass flux and also its significance in gas transport in organic nanopores of shale. Based on the results in this work, when graphite conduits are used to represent the organic nanopores of shale, particular attention should be paid to the surface roughness, as they significantly impact the fluid transport by lowering the mass flow flux of the adsorbed phase. The results of this study can potentially modify the multiscale formalism of the gas flow in shale reservoirs.

Speakers: Dr. Mohammad Kazemi (Kansas University) , Ali Takbiri-Borujeni (West Virginia University)
• 11:54
Sorption of Methane and Carbon Dioxide in Type II-A Kerogen Rough Slit Nanopores by Molecular Simulations 15m

Shale gas has redefined energy landscape[1]. The United States (U.S.) natural gas production is expected to increase every year, and in 2035 the U.S. shale gas production may raise to 50% of the total gas production.

Shale rock consists of micropores and mesopores[2]. It is also composed of inorganic minerals (quartz, clays, calcites, and feldspars, etc.) and organic matter (kerogens and bitumens). The organic matter is mainly composed of kerogens and it is considered as the main gas trapping of methane[3] and shows high capacity for carbon dioxide adsorption trapping[4]. An understanding of shale kerogen adsorption characteristics, under the reservoir condition, is required to successfully exploit the shale formations.

Three different types of kerogen depending on its origin can be distinguished: i) type I from a lacustrine anoxic environment, ii) type II from marine shale and continental planktons, and iii) type III from plants in tertiary and quaternary coals. All these types can be classified according to the elemental ratio of Hydrogen/Carbon (H/C), Oxygen/Carbon (O/C), and Sulfur/Carbon (S/C). The physicochemical properties (structure, adsorption, retention, etc.) of the kerogen strongly depend on its origin and on the burial history of the reservoir where it came from[5].

Molecular simulations allow a detailed picture of the structure, thermodynamics and dynamics of the fluid at the interface. In this work the dry structures of type II immature kerogen with different types of dummy particles are simulated. Then, the kerogen media are used to study the sorption of methane and carbon dioxide in the kerogen matrix and at the kerogen rough slit nanopore surface. Our results are compared with the available experimental data.

Speaker: Dr. Stephane TESSON (Reservoir Engineering Research Institute (RERI) & University of California Riverside (UCR))
• 12:12
Determination of Absolute Adsorption Isotherms of C1/C2 Mixtures with the Simplified Local Density (SLD) Theory 15m

Objectives:
Molecular simulations have been recently used to obtain the absolute adsorption isotherms of pure hydrocarbons on shale by considering the surface adsorption in nanopores. But, few attempts are made to obtain the absolute adsorption isotherms of multi-component mixtures, which, however, is important in determining the initial gas resources in place in shale formations. This study measures the excess adsorption isotherms of pure C1, C2, and C1/C2 mixtures on shale, and consequently determines the absolute adsorption isotherms of these mixtures with the simplified local density (SLD) theory.

Methods Procedures/Process:
By applying the thermogravimetric method, we measure the excess adsorption isotherms of pure C1, C2, and C1/C2 mixtures on two shale samples over the temperature range of 303.15-363.15 K and the pressure range of 0-30 MPa. SLD theory is then applied to calculate the density profiles of C1/C2 mixtures in nanopores by considering the local fluid-surface interactions. We calculate the adsorption phase density of the mixtures and such density is employed to calibrate the excess adsorption isotherms, which enable us to eventually obtain the absolute adsorption isotherms of these hydrocarbon mixtures.

Results/Observes/Conclusions:

Novelty:
This work is a study on the adsorption behavior of multicomponent mixtures in shale, which is currently of great interest to the petroleum industry. A pragmatic approach is proposed to obtain the absolute adsorption of hydrocarbon mixtures, helping to achieve more accurate shale-gas-in-place estimations in shale reservoirs.

Speaker: Mr. Yueliang Liu (University of Alberta)
• 14:00 14:30
Coffee Break 30m
• 14:05 16:00
Parallel 2-A
• 14:07
Visualisation of solute transport and determination of its transport properties in porous sintered glass 15m

Solute transport in porous media is important for several industrial applications, i.e.: hydrology, building stone performance and waste management. Spreading and mixing during solute transport is significantly impacted by the pore scale heterogeneity found in natural porous media, which complicates upscaling (Dentz et al., 2011). Therefore, simulations and experiments which investigate the evolution of pore scale solute concentration fields in such materials are very valuable. However, direct visualisation of these concentration fields at the micron-scale in rocks is complicated by the high spatial and time resolutions that are required. Bultreys et al. (2016) and Boone et al. (2016) present first tests on imaging solute transport in a carbonate rock using fast laboratory-based micro-CT. In this study, we extend this work by attempting to quantify micro-CT concentration fields, in order to investigate spreading and mixing under different flow conditions in porous materials with different degrees of heterogeneity.

A significant part of this work is aimed at the methodological challenge of performing in-situ micro-CT scans of solute transport with imaging times on the order of seconds. We use the EMCT scanner of UGCT (www.ugct.ugent.be), a micro-CT system specially designed for in-situ imaging, with a rotating X-ray tube and detector in a horizontal plane (Dierick et al., 2014) and investigate the quantitative correctness when imaging the concentration of a dissolved tracer salt (0 wt%, 2 wt%, 4 wt%, 6 wt%, 8 wt% and 10 wt% CsCl) in porous sintered glass at 12 seconds per scan, with a voxel size of 13 micron. The CsCl-concentration increases the X-ray attenuation coefficient of the fluid, which causes an increase in grey values observed in the reconstructed micro-CT datasets. The high temporal resolution at which the micro-CT images are taken, is inherently linked with a limited signal-to-noise ratio. Despite this drawback, the first experimental results suggest a linear relationship between the grey values of the tracer-solution in the fast scans and the tracer concentration.

Results from the presented experiments can be used to investigate flow structures at the pore scale and to validate pore scale solute transport simulations. Further development of the methodology could also lead to valuable insights in multi-phase solute transport and reactive transport.

Speaker: Ms. Stefanie Van Offenwert (Ghent University, Pore-scale Processes in Geomaterials Research Team (PProGRess))
• 14:25
Transport with Bimolecular Reactions in Fracture-Matrix Systems: Analytical Solutions with Applications to Weathering Reactions and In-Situ Chemical Oxidation 15m

We consider the problem of advection, matrix-diffusion and bimolecular reactions in fracture-matrix systems, with two example applications: (i) Weathering reactions in fractured bedrock and (ii) in-situ chemical oxidation (ISCO) for remediation of fractured rock. In both cases, a reagent (a weathering agent such as H+ or dissolved oxygen, or permanganate in the case of ISCO) are supplied through a fracture, and react with a second species (immobile mineral species in the case of weathering reactions, or TCE/PCE in the case of ISCO) initially contained in the fracture-matrix system. In both cases, moving reaction fronts form and propagate along the fracture and into the rock matrix. The propagation of these reaction fronts is strongly influenced by the heterogeneity/discontinuity across the fracture-matrix interface (advective transport dominates in the fractures, while diffusive transport dominates in the rock matrix). We present analytical solutions for the concentrations of the oxidant/weathering agent, weathering mineral or TCE and natural organic matter; and the propagation of the reaction fronts in a fracture-matrix system. Our approximate analytical solutions assume advection and reaction dominate over diffusion/dispersion in the fracture and neglect the latter. In the ISCO problem, the behavior of the reaction-diffusion equations in the rock matrix is posed as a Stefan problem where the supplied oxidant reacts with both diffusing (TCE) and immobile (natural organic matter) reductants. Our analytical solutions establish that the reaction fronts propagate diffusively (i.e. as the square root of time) in both the matrix and the fracture. Our analytical solutions agree very well with numerical simulations for the case of uniform advection in the fracture. In the context of the ISCO problem, we also present extensions of our analytical solutions to non-uniform flows in the fracture by invoking a travel-time transformation. These non-uniform flow solutions are relevant to field applications of ISCO, which employ forced-gradient flow systems. Our approximate analytical solutions are relevant to a broad class of reactive transport problems in fracture-matrix systems where moving reaction fronts occur, and may be generalized further to consider multiple interacting species.

Speaker: Harihar Rajaram (University of Colorado, Boulder)
• 14:43
Chaotic Fluid Advection in Crystalline Granular Media 15m

We study the Lagrangian dynamics of steady three-dimensional (3D) Stokes flow over granular media consisting of simple cubic (SC) and body-centered cubic (BCC) lattices of closed-packed spheres, and uncover the mechanisms governing chaotic fluid advection. Due to the cusp-shaped sphere contacts, the topology of the skin friction field is fundamentally different from that of continuous (non-granular) media (e.g. open pore networks), with significant implications for fluid advection. Weak symmetry breaking of the flow orientation with respect to the lattice symmetries imparts a transition from regular advection to strong chaotic advection in the BCC lattice, whereas the SC lattice only exhibits weak advective mixing. Using a numerical simulation of the flow at various flow orientations, we quantify the strength of chaotic mixing from the Lyapunov exponent, and examine how it is distributed over the parameter space of mean flow orientation [1]. We furthermore analyze the flow topology and show that the occurrence of chaotic advective mixing is controlled by the existence within the flow of transverse intersections between stable and unstable manifolds originating from the spheres [1]. These insights are used to develop accurate predictions of the Lyapunov exponent distribution over possible flow orientations [2]. The difference of behavior observed for the SC and BCC lattices, which share the same symmetry point group, results from their different space group symmetries: a glide symmetry of the BCC lattice allows the occurrence of chaotic advection [1]. These results point to a general theory of advective mixing and dispersion based upon the inherent symmetries of arbitrary crystalline structures.

Speaker: Prof. Yves Méheust (Géosciences Rennes)
• 15:01
Mixing effects in agrochemical biodegradation networks in variably saturated soils 15m

Few current bioreactive transport solvers currently provide a comprehensive mechanistic description of biogeochemical cycles and allow easy integration of all the involved processes, including flow in variably saturated media, solute transport and kinetic/equilibrium biochemical reactions. The parameterization of these processes is particularly challenging since our knowledge of model parameters and of their spatial heterogeneity is typically incomplete. Therefore, it is crucial to study the impact of each process and related parameters uncertainty in the outputs of interest.
In this communication, we consider reaction networks describing biochemical degradation of herbicides, atrazine and glyphosate (la Cecilia and Maggi, 2017a, 2017b). In particular, we quantify the degradation potential and accumulation of toxic substances and we study the biomass and ecological structure dynamics in soil and groundwater. Our discussion encompasses the assessment of model outputs sensitivity on the identified model structure and related parameters. We focus on the effects of local mixing, which entails the full characterization of spatial and temporal fluctuations of solutes concentration within the time-space domain. We frame these analyses in terms of dimensionless parameters describing the integrated processes of interest, i.e., fluid flow, solute transport, and biogeochemical reactions and we identify global trends and statistical indicators to characterize the system at steady state. This work will address system nonlinearities linked to the interplay of diverse processes with the aim of increasing our understanding of dominant mechanisms involved in agrochemicals bioreactive transport.

Speaker: Dr. Giovanni Porta (Dept. of Civil and Environmental Engineering, Politecnico di Milano)
• 15:19
Semi-Analytical Particle Tracking Scheme For Advective/Diffusive Transport in Porous Media 15m

Semi-Analytical Particle Tracking Scheme For Advective/Diffusive Transport in Porous Media

The particle tracking scheme of David W. Pollock [Ground Water 26(6), 1988] provides a computationally efficient and mass-conservative method for Lagrangian transport in the absence of diffusion. In this work, a generalization of Pollock's scheme that allows for the inclusion of diffusion is presented. The new scheme is based on a semi-analytical representation of the advective/diffusive motion. The scheme does not require stepping at sub-grid-cell time or length-scales and thus is computationally efficient. It is formulated in such a way that it becomes exact for Pe going to zero and infinity, and provides an accurate numerical approximation in the intermediate Pe number range. Application examples dealing with Darcy flow in heterogeneous porous media and Stokes flow in resolved pore-space geometries document the capabilities of our new scheme.

Speaker: Daniel Meyer (Institute of Fluid Dynamics, ETH Zurich)
• 14:05 16:00
Parallel 2-B
• 14:07
An experimental and numerical pore-scale study of bio-enhanced NAPL dissolution in porous media 15m

Nonaqueous phase liquids (NAPLs) are still a major challenge for all traditional groundwater treatment technologies. NAPLs often contaminate the subsurface following an accidental spill or due to a defect in the oil storage tank. These pollutants remain trapped in the form of droplets and / or immiscible clusters within the aquifer, thus constituting a persistent source of pollution that is difficult to decontaminate. Predicting the fate of this pollutant requires characterizing all the mechanisms involved and in particular the biodegradation, which can occur in the vicinity of the pollutant source or further, to the dissolved plume. If a significant research effort has been put into investigating the transport and biodegradation of dissolved contaminants, comparatively very few works (e.g., Bahar et al., 2016) are focused on the study of such processes in multiphase conditions (oil/water/biofilm systems).
In this study, we give an attempt to address this open issue from an experimental and numerical perspective. First, we illustrate impact of bacteria on dissolution of pure organic phase from micromodel experiments. The experimental set-up is made of a micromodel (i.e. 2D transparent flowcell) used to study the dissolution of oil phase. Changes in toluene saturation are directly monitored from recorded two-dimensional images and dissolved concentrations at the outlet are measured by gas chromatograph. Results of toluene dissolution and biodegradation by a toluene-degrading strain (Pseudomonas putida F1) are compared with experiments in abiotic conditions.
In parallel, we present a two-dimensional pore-scale numerical model (Benioug et al., 2017) to investigate the main mechanisms governing biofilm growth and NAPL dissolution in porous media. Fluid flow is simulated with an immersed boundary–lattice Boltzmann model while solute transport is described with an interface reconstruction finite volume approach (Benioug et al., 2015). A cellular automaton algorithm combined with the immersed boundary method was developed to describe the spreading and distribution of biomass. Different conditions are considered (spatial distribution of biofilm, reaction kinetics, biosurfactant production, NAPL toxicity) and their impacts on the dissolution process are analyzed.

Speaker: Fabrice GOLFIER (Université de Lorraine - GeoRessources Laboratory)
• 14:25
Monitoring of bacterial biofilm formation and permeability reduction in sands using complex electrical conductivity 15m

Geophysical monitoring of bacterial activities in subsurface has drawn significant interest in various civil engineering, hydrocarbon recovery, soil remediation practices. This study explored the feasibility of use of complex electrical responses to monitor bacterial biofilm formation in soils. A series of column experiments was conducted, in which the model bacteria Shewanella oneidensis MR-1 were cultured in a sand-pack and stimulated to form biofilms. During the bacterial growth and biofilm formation, the variations in complex electrical conductivity were monitored at a frequency range of 0.1–1000 Hz while measuring the permeability reduction. As a result of the bacterial growth and biofilm formation, it was observed that the imaginary conductivity significantly increased by more than 500% and the real conductivity was reduced by more than 25% over 14 days of the experimental run. At the same time, the permeability was reduced by ~50–60%. However, these variations diminished along the sand column as the distance became farther from the nutrient injection port. It appeared that the imaginary conductivity effectively captured biofilm formation in porous media, while the real conductivity was heavily affected by porosity reduction as well as pore fluid conductivity. The obtained results suggest that complex conductivity can be effectively used to capture the bacterial biofilm formation and the resulting permeability reduction in soils.

Speaker: Ms. Hyun-Woo Joo (KAIST)
• 14:43
A numerical model for reactive transport coupled with microbial growth on Darcy scale 15m

The injection of substrates, e.g. hydrogen with the purpose of energy storage, into subsurface structures could stimulate the growth of all present microbial species which are able to use this substrate for their metabolism. The linkage between transport, the growth of microorganisms, substrate availability and biodegradation results in a strongly coupled dynamic system. The difficulty in the development of a general model is the inclusion of processes which appear on different length and time scales. In this work, a flexible numerical model was developed which uses effective representations of the processes on Darcy scale. The mass exchange between two phases (gas and water) is treated instantaneously by using an equilibrium law. Multiple metabolic reactions can be included kinetically by defining the stoichiometry and kinetic coefficients. Different mathematical models can be selected to describe the substrate-limited microbial growth. Microorganisms can be considered as an immobile biofilm or as partially mobile within the water phase. The jump-like appearance of usual growth models and the strong coupling to the reactive transport equations results in a very stiff equation system. The numerical instability was overcome by a proper adaptive time step selection and a check for the physical possibility of each solution before it is continued to the next one. Example simulations are shown for a near wellbore and a field scale study of an underground hydrogen storage.

Speaker: Birger Hagemann (Clausthal University of Technology)
• 15:01
Soil bacteria: searching for more efficient bio-fertilizers 15m

Life in porous media, as soil bacteria, are used since more than 40 years ago as bio-fertilizer contributing to the development of a sustainable agronomy. Even though they are extensively used due to their low cost, such biotechnology is still far from being efficient and many challenges are opened for basic research in porous media science.

Our microbiological system of study are the Bradyrhizobium diazoefficiens, bi-flagellar bacteria. One of the opened question regarding its efficiency is which is the aim of its bi-flagellum system developed by this specie and not developed in general in Bradyrhizobium. It may be an adaptive trade-off between energetic costs and ecological benefits among different species. We work interdisciplinary on bacteria diffusion in porous media, numerically and experimentally, imitating the complex and structured soil with artificial micro-fluidics devices. With a better visualization in transparent devices, easy to manipulate in a laboratory we aim to understand and control the transport properties of the system for further realistic applications. In this work we were able to report numerically their recent reported strategies to swim [Quelas et at, Sci.Rep. 2016]. Further we simulate their dynamics with those realistic parameters, under a broad spectrum of micro-confinement environments, contributing to micro-fabricate different geometries of porous media. All these studies contribute to understand their diffusion properties versus their flagellar systems (different motility) and versus the porous arrangement. These in vitro contributions hopefully will be useful for further development in a sustainable agronomy.

Speaker: Verónica I. Marconi (FaMAF-UNC and IFEG-(UNC-CONICET))
• 14:05 15:59
Parallel 2-E
• 14:05
Supercritical Methane Sorption on Organic-Rich Shales from Sichuan Basin, China 15m

Fine-grained sedimentary rocks, such as mudstones and shales, contain abundant nanometer- to micrometer-sized pores. These narrow pores create intense fluid-rock interaction that may lead to complicated fluid storage and transport process. Concerns about the accurate evaluation of gas content and diffusion kinetics have led to many experimental studies about gas sorption on shales. However, data on high-temperature high-pressure sorption isotherms of shales are still scare. In particular, the burial depth of the Paleozoic shales in the Upper Yangtze region of China is mostly in a range of 2000–4000 m, which indicates that the temperature and pressure of shale reservoirs are in the range of 60–120 °C and 20–40 MPa. Experimental techniques employed in obtaining sorption data have to be optimized and at the same time the measuring conditions have to be extended to in-situ conditions of deep shales while many published sorption data are limited to moderate pressures and temperatures.
In this work, high-temperature high-pressure sorption data for methane on shales from Sichuan Basin have been obtained at 30–120°C and pressures up to 25 MPa using a specially designed two-temperature-zone manometric setup. Dubinin-Polanyi potential theory was modified to extend to supercritical gas sorption over wide temperature and pressure ranges. A modified adsorption potential method was proposed to calculate the characteristic curves for supercritical gas sorption, and then a rigorous function from the supercritical Dubinin-Astakhov equation was also developed to describe the modified characteristic curve. Furthermore, the physical meaning of characteristic curve has been elucidated by comparing characteristic curves of different kinds of shales and clay minerals.
The measured excess sorption isotherms of shales follow the physisorption trend of decreasing amounts of methane adsorbed with increasing temperature. Characteristic curves of methane on shales at 30–120°C were calculated using a new expression of adsorption potential. It is found that if the thermal expansion of adsorbed phase is considered, these modified characteristic curves are temperature-invariant. The characteristic curve equation is capable to predict methane sorption at other temperatures based on the easily tested isotherm at room temperature. Using the sorption isotherm and characteristic curve equation at 30°C, the predicted isotherms at 120°C agree well with experimental data. The modified characteristic curves comprehensively characterize the available pore space for sorption and the affinity of methane molecules. The later stage of the modified characteristic curves (limited adsorption volume) is mainly controlled by the available pore space provided by organic matter and clay minerals. The limiting adsorption volume of shales in the gas window is larger than shales in the oil window with the same TOC content. The initial stage of the characteristic curves reflects the affinity of methane molecules for sorption on organic matter. According to the characteristic curves, shales in the gas window show higher affinity than shale in the oil window and clay minerals, though the clay minerals may provide comparable adsorption volume. The sorption characteristic energy shows a parabolic-like shape with a minimum approximately around Req =1.1%, which are related with the evolution of porosity of shales.

Speaker: Feng Yang (China University of Geosciences, Wuhan)
• 14:23
Direct observation of phase change in sub-10 nm porous media 15m

Phase change at the nanoscale is critical to many industrial applications including rapidly emerging unconventional oil and gas production from nanoporous shale reservoirs. The thermodynamic behaviour of hydrocarbons confined to these nanopores is expected to deviate significantly from bulk properties and there is little experimental data to validate theories. This research aims to visually observe the evaporation of hydrocarbons in a nanofluidic chip that accurately represents the geometric dimensions and the pressure/temperature conditions observed in shale. The chip consists of a nanoporous network of two-dimensional (2D) nano pores with dimensions down to 8 nm. Using an experimental procedure that mimics pressure drawdown during shale oil/gas production, our results show that evaporation of pure propane takes place at pressures lower than predictions from the Kelvin equation (maximum deviation of 11%). We probe evaporation dynamics as a function of superheat and find that vapor transport resistance dominates evaporation rate. For the transport resistance in the sub-10 nm nanoporous media, the contributions of the Knudsen flow and the viscous flow are found to be approximately equivalent. We also observe a phenomenon in sub-10 nm confinement wherein lower initial liquid saturation pressures trigger discontinuous evaporation resulting in faster evaporation rates. Additionally, we also extend this work to study evaporation and cavitation dynamics in nanofluidic devices with (a) mixture of pore sizes coupled with (b) mixture of hydrocarbons. Collectively, the results presented will aid in increasing the efficiency of shale production and will inform modelling and simulation of shale gas production.

Speaker: Arnav Jatukaran (University of Toronto)
• 14:41
Molecular Simulation of Competitive Adsorption behaviors of CO2/CH4 Mixtures on Shale Clay Minerals 15m

Objectives:
CO2 injection, as one of the effective techniques for enhancing recovery of shale gas, has been widely used and proved economically available. In shale, clay minerals play an important role on methane adsorption due to its large volume of micropores. So far, however, a few attentions have been paid on competitive adsorption of CO2/CH4 Mixtures on clay minerals. In this study, we conduct molecular simulations of CO2/CH4 mixtures to provide a better understanding of competitive adsorption behaviors on clay minerals with the grand canonical Monte Carlo (GCMC) simulation.
Methods Procedures/Process:
We conduct GCMC simulations of CO2/CH4 mixtures adsorption in various clay minerals. Based on the actual conditions of shale gas reservoir, the competitive adsorptions of CO2/CH4 mixtures are investigated at various temperatures of 303.15K, 333.15K, and 363.15K with the pressure range of 0-35Mpa. For comprehensive comparison, the effects caused by pore size, mole fraction of CO2/CH4, and different clay minerals on competitive adsorption are processed. The competitive adsorption behaviors are characterized by selectivity and such key parameter are employed to evaluate the density profiles of CO2 and CH4, the characteristics of CO2 adsorption over CH4, and the timing of CO2 Injection.
Results/Observes/Conclusions:
Due to strong quadrupole moment and higher van der Waals interactions, CO2 possess a stronger affinity for clay minerals than that of methane, which is nonpolar. CH4 has the characteristic of single-layered adsorption, while the CO2 is able to form multi-layers adsorption with higher adsorption amount. Molecular simulations results show that the selectivity of CO2 in competitive adsorption decrease with decreasing of pressure. In addition, the selectivity of CO2 is independent on temperature. Because of negatively charged silicate layers, CO2 adsorption in nanopores of illite and montmorillonite are stronger than that of in kaolinite. As a result, the selectivity of CO2 in kaolinite models is less than that of in illite and montmorillonite models. When pressure is higher than 10Mpa, however, the selectivity of CO2 in three different clay minerals are similar. Due to the more adsorption sites occupied by CO2, the selectivity increases with mole fraction of CO2. By comparison with different pore size adsorption, the selectivity is insensitive to the mole fraction in micropores, while it increases with the increasing of mole fraction of CO2 in mesopores.
Novelty:
This work is a study on CO2/CH4 Mixtures adsorption behaviors on shale clay minerals. The molecular simulations with GCMC are proposed to give an insight in competitive adsorption mechanism, which is expected to provide a more accurate understanding of CO2 injection for enhancing recovery of shale gas.

Speakers: Mr. Xiaofei Hu (1. College of Energy, Chengdu University of Technology, Chengdu, PR China. 2. School of Mining and Petroleum Engineering, Faculty of Engineering, University of Alberta, Edmonton, Canada.) , Ms. Yuanyuan Tian (1. College of Energy, Chengdu University of Technology, Chengdu, PR China. 2. School of Mining and Petroleum Engineering, Faculty of Engineering, University of Alberta, Edmonton, Canada.)
• 14:59
Production Data Analysis from Unconventional Reservoirs with a Novel Data-Driven Drainage Volume Approach 15m

Understanding how pressure fronts propagate (diffuse) in unconventional reservoirs is fundamental to transient flow analysis as well as reservoir drainage volume estimation. We have developed an alternative approach to the solution of the 3-D diffusivity equation by directly solving the propagation equation for the “pressure front” of the transient solution. The pressure front equation is an Eikonal equation, which is obtained from the high frequency asymptotic limit of the diffusivity equation in heterogeneous reservoirs and whose properties are well developed in the literature. Most importantly, the Eikonal equation can be solved very efficiently by a class of solutions called the Fast Marching Methods for a “diffusive time of flight” that governs the propagation of the pressure front in the reservoir. The diffusive time of flight can be used as a spatial coordinate to reduce the 3-D diffusivity equation into an equivalent 1-D formulation, leading to a simplified method for rapid reservoir modeling.

Based on this theory, we may further introduce a novel data-driven approach for production analysis of unconventional reservoirs without the traditional rate transient and pressure transient (RTA/PTA) assumptions of specific flow regimes. Our approach uses a transient generalization of the Matthews-Brons-Hazebroek method for the PSS drainage volume which relies on a w(τ) function to characterize the flow geometry from the transient drainage volume. Together with a calculated instantaneous recovery ratio, it has been successfully used to rank refracturing candidates and to obtain optimal fracture spacing. Given well pressure and flow rate data, we can calculate the transient well drainage volume with time. The time evolution of the drainage volume can be inverted to derive the w(τ) function which then provides a high resolution diagnostic plot that can be used for quantitative analysis to obtain fracture surface area, matrix properties, stimulated reservoir volume (SRV), and additional reservoir and fracture characteristics that are not apparent in the usual rate and pressure transient analysis techniques.

We have applied our methodology to field examples from the Montney and Eagleford shales. The comparison to standard RTA/PTA shows linear flow and fracture interference features more clearly than conventional RTA/PTA. It also provides detailed characterization of complex non-planar hydraulic fracture geometry, partial completion effects, the development and growth of the SRV, leading to the estimation of future decline rate and ultimate recovery.

The major advantage of the proposed approach is the data-driven model-free analysis of production data without the presumption of specific flow regimes. It provides a simple and intuitive understanding of the transient drainage volume and instantaneous recovery efficiency, irrespective of the complexity of the geometry of the reservoir depletion.

Speaker: Mr. Zhenzhen Wang (Texas A&M University)
• 15:17
Microscale Modeling of the Effect of Silica nanoparticles and Surfactants on Heavy Oil Displacements 15m

Chemicals in the form of nanoparticles or surfactants provide opportunities to improve oil displacement from rocks. They increase the rate of hydrocarbon recovery by breaking down the oil trapped in by-passed zones and separating the residual oil from rock surfaces in the form of tiny droplets suspended in the water phase. In this study, a series of heavy oil displacement experiments are conducted by flowing a series of aqueous solutions through an oil-wet and transparent network of microfluidic devices. Micromodels are fabricated by soft lithography techniques on a silicon wafer and replicated with Polydimethylsiloxane (PDMS) polymer. The effect of silica nanoparticles and three different types of surfactants (SDS, Tween 20, and Silwet) on the displacement of heavy oil, removal of oil films, and mobilization of trapped oil droplets are investigated. Furthermore, the patterns of residual oil and final oil recovery factors are explored. Also, the synergism effect between nanoparticles and different types of surfactants are reported.
3D Confocal microscopy coupled with fast speed fluorescent imaging of the displacement process reveals the effect of each chemical additive on oil mobilization. Silica particles show the tendency to reduce or remove the remaining oil film thickness while the surfactants break up the oil phase in the by-passed channels into tiny clusters that can be transported by the displacing fluid. The results demonstrate that the addition of the silica nanoparticles increases the rate of oil recovery up to 60% resulted from the wettability alteration and oil film removal. Moreover, the recovery factors increase upon adding the silica nanoparticles to a constant concentration of Tween 20 and SDS. The silicon-based surfactants improve the oil recovery up to 80 % where the recovery improvement by the addition of nanoparticles is negligible.
The developed microfluidic-based model is a powerful mimetic prototype of real porous media which can clarify the mechanisms underlying the process of chemical-based flooding for oil recovery. Considering the time-consuming and expensive nature of core-flood experiments, the proposed microfluidic approach provides an attractive alternative for rapid and low-cost enhanced oil recovery (EOR) screening studies.

Speaker: Mrs. Parisa Bazazi (Subsurface Fluidics and Porous Media Laboratory; Department of Chemical and Petroleum Engineering; University of Calgary)
• 15:35
Experimental studies on CO2 Enhanced Shale Gas Recovery on the Basis of Competitive Adsorption 2m

Shale gas is one type of the conventional natural gases and widely distributed in the world. After the shale gas revolution in the U. S. A., shale gas has also been developed in Canada and China. Natural depletion method is often used for the development of shale gas. But how to enhanced the shale gas recovery for the depleted shale gas reservoir? The comparison of the adsorption and desorption properties of CO2 and methane were conducted. A new equipment was designed to study the CO2-flooding of gas in shale cores at different injection pressure and rate. The gas recovery factors were compared with that of depleted method. Experiments show that the adsorption of CO2 is higher than that of methane for the same shale. The gas recovery factor increases after CO2 injection compared to that of natural depletion development. The increment of gas recovery factor was related to the injection pressure and rate. The competitive adsorption maybe the main mechanism for CO2-flooding of shale gas.

Speaker: Prof. Renyuan SUN (China University of Petroleum)
• 15:38
Effects of Composition on Canister Desorption Behavior of Upper Paleozoic Shales in the Ordos Basin, NW China 2m

A series of canister desorption tests were carried out on 31 deep (over 3000 m) over-mature Lower Permian-Upper Carboniferous shale cores under atmospheric pressure and at reservoir temperatures of 75 and 80 °C, as well as a higher temperature of 95°C. Organic chemistry and X-ray diffraction were combined to investigate the impact of composition on canister desorption behavior. In order to better understand sorption and emission processes of shale gas, high-pressure methane sorption experiments were conducted at the reservoir temperature and pressures up to 18 MPa . Geochemical measurements show that the total organic carbon (TOC) content ranges from 0.488 wt % to 4.310 wt %. The depositional setting is lagoon and delta. The type of organic matter is mainly Type III. The dominant minerals of the shale samples are clay (25.4-97.0 wt %, average 58.8 wt %) and quartz (1-62.1 wt %, average 33.3 wt %). The content of clay minerals shows a significant negatively correlation with that of quartz with a coefficient of determination, R2, of 0.7260. The results show that an increase from the reservoir temperature to a higher value results in an average of 31% enhancement in desorbed gas volume. The desorbed gas volumes at both temperatures are linearly correlated with total organic carbon (TOC) content, which support the positive relationship with TOC found by the high-pressure methane sorption isotherms. The coefficients of determination, R2, at reservoir temperature and 95°C, are 0.6584 and 0.6444, respectively. The desorbed gas volumes at both temperatures show a slight correlation with clay minerals, indicating that adsorption sites on clay may have an impact on canister desorption testing. In addition, a slight negative correlation between quartz and desorbed gas volume was observed. The shale samples with a lower content of quartz and a higher desorbed gas volume generally have a higher content of clay minerals and TOC, which may indicate that clay minerals and TOC are the dominant contributors to the sorbed gas capacity, even after spontaneous imbibition.

Speaker: Mr. Fengyang Xiong (School of Earth Sciences, The Ohio State University)
• 15:41
N2, CO2, and Ar adsorption to characterize micro- and mesopores of shales 2m

The storage and flow mechanisms in shales depend largely on their microstructure. We use two parameters to characterize microstructures, namely specific surface area (SSA) and pore-size distribution (PSD). We use N$_2$ adsorption at 77K to quantify SSA and PSD of nanopores. There are two limitations of the N$_2$ adsorption method due to (1) uncertainties in molecular area due to the quadrupole moment of N$_2$ molecules result in 20% uncertainty in calculated BET SSA, and (2) kinetic restriction of N$_2$ molecules prevent it to access narrow pores (< 0.7 nm). To circumvent these limitations, we also used other adsorptives, such as CO$_2$ and Ar, for the measurements.

We present results from adsorption measurements of CO$_2$ at 273 K and Ar at 77 K on shales and compare them to N$_2$ adsorption at 77 K. Adsorption measurements with CO$_2$ at 273 K allows for detailed characterization of ultramicropores (< 0.7 nm), which are inaccessible to N$_2$ molecules. Our results from CO$_2$ adsorption reveal significantly larger micropore SSA in comparison to N$_2$ probed SSA. Ar molecules do not have quadrupole moment and resolve the uncertainties of molecular area for BET calculation.

Speaker: Nerine Joewondo (Colorado School of Mines)
• 15:44
Theoretical foundation of fracture permeability to analyse the flow in channel fracturing based on Navier-Stokes equation 2m

Channel fracturing, as a novel technology, has reveived increasing attention in recent years due to its great advantage in promoting the fracture conducitivity as well as reducing consumtion of water and proppant. The open channels created by heterogeneous distribution of proppant are the priority path for oil or gas to pass instead of the pores exist in proppant pack. Since the flow pattern has been changed from Darcy flow to the flow mode dominated by Navier-stokes percolation by this technique, traditional method to calculate and predict its permeabity may not that accurate and appropriate. So this full paper presents a new model to measure the flow capacity of channel fracturing combined with a power law embedment model. The new approach is based on Navier-Stokes equation instead of Darcy-law to compute fracture permeability in channel fracturing. Besides, four influencing, namely, fracture height, propant pillar size, proppant area friction and embedment, are investigated to analyse their effect on permeability. The proposed analytical model is found to in good agreement with the experimental data, which verifies the precision and the feasibility of this model. Based on this model, this paper can provide a theoretical basis for channel fracturing design to evaluate the key factor governing the conductivity(permeability), which is helpful to providing a reference for proper pulse time and pumping rate optimization.

Speaker: Dr. Guoqing XU (Research Institute of Petroleum Exploration and Development(RIPED), PetroChina,)
• 16:00 17:30
Poster 1: Poster 1-A
• 16:00
CO2 Breakthrough Pressure in Resedimented Caprock Seals 15m

Structural trapping is the ultimate barrier for reducing the risk of leaks at CO2 storage sites. Small pores in high specific surface clay-rich caprocks give rise to high capillary entry pressures and high viscous drag that hinder the migration of buoyant carbon dioxide CO2.
In this work we show measurements of the CO2 breakthrough pressure and ensuing CO2 permeability through sediment plugs prepared with sand, silt, kaolinite and smectite. Our experiments and data from the literature demonstrate that the breakthrough pressure can reach ∼6.2 MPa in argillaceous formations, and 11.2 MPa in evaporites. The CO2 relative permeability after breakthrough increases up to a maximum of ~0.2. Our parametric study highlights the inverse relationship between breakthrough pressure and pore size, as anticipated by Laplace’s equation. In terms of macro-scale parameters, the breakthrough pressure increases as the sediment specific surface increases and the porosity decreases.
In addition, we introduce two dimensionless numbers that help pre-asses the safety of storativity condtions in-situ. The “sealing number” and the “stability number” combine the initial fluid pressure, the buoyant pressure caused by the CO2 plume, the capillary breakthrough pressure of the caprock, and the stress conditions at the reservoir depth; these two numbers provide a rapid assessment of potential storage sites.

Speaker: Prof. D. Nicolas Espinoza (The University of Texas at Austin)
• 16:15
The Effect of Original and Initial Saturation on Residual Nonwetting Phase Capillary Trapping Efficiency 15m

Injection of supercritical carbon dioxide (CO2) into geological formations is used for both atmospheric greenhouse gas reduction (climate change mitigation) and enhanced oil recovery. In an effort to fully understand CO2 trapping efficiency, the capillary trapping behaviors that immobilize subsurface fluids were analyzed at the pore-scale using pairs of proxy fluids representing the range of in situ (supercritical) nonwetting and wetting fluids. The pairs of fluids were cycled through imbibition and drainage processes using a flow cell apparatus containing a sintered glass bead column. Computed x-ray microtomography (microCT) was used to identify immobilized nonwetting fluid volumes after imbibition and drainage events.
From the images, the trapped residual (post-secondary imbibition) nonwetting phase was spatially correlated to both the original (post-primary imbibition) and the initial (post-primary drainage) nonwetting phase; relationships referred to as the original saturation dependence (So-dependence) and initial saturation dependence (Si-dependence), respectively. Statistically significant trends of decreasing So- and Si-dependence with increasing wetting and nonwetting fluid phase viscosities were observed. This finding implies that the amount of CO2 injected and ultimately trapped is dependent on the nonwetting phase (e.g. oil or gas) already present in the formation, as well as on the manner in which supercritical CO2 is initially injected.

Speaker: Dorthe Wildenschild (Oregon State University)
• 16:30
A coupled wellbore-reservoir model of CO2 flow and heat transfer during a push-pull experiment at Heletz, Israel 15m

To quantify in-situ CO2 residual trapping for CO2 geological storage, dedicated push-pull experiments have been carried out at the Heletz, Israel pilot CO2 injection site. The site is well characterized and instrumented for CO2 injection and sophisticated sampling and monitoring (Niemi et al., 2016) and residual trapping experiments have been carried out during 2016-2017. The objective of the present work is to develop a simulation model capturing the CO2 transport and trapping behavior consistent with the recorded pressure and temperature data at the injection well, with special focus on the coupled wellbore-reservoir flow. For this purpose, the simulation of the CO2 push-pull (injection-withdrawal) experiment is carried out with the numerical simulator T2Well/ECO2N (Pan et al., 2011) to account for the role of wellbore-reservoir coupling. Of particular interest in this work is to accurately model the period when the well is self-producing fluids and to analyze what conditions are causing the observed gas-release behavior. Comparison of numerical model simulations and the measured data suggests that the gas saturation in the reservoir at the onset of the self-release period is only slightly above the residual gas saturation of the formation. In addition, the results indicate that the effective permeability in the reservoir is small enough to be the controlling factor for the gas inflow rate into the wellbore. This detailed modeling of the well self-release behavior allows a more reliable overall estimation of the in-situ residual trapping at the site.

Speaker: Farzad Basirat (Department of Earth Sciences, Uppsala University)
• 16:45
Stress-sensitivity Modeling during CO2 Flooding and Storage in Tight Sandstone Core 15m

Understanding the influence of CO2 injection on rock stress is one of the key elements to analyze CO2 Enhanced gas recovery and long term CO2-storage in tight sand gas reservoirs. Producing natural gas from reservoir and injecting CO2 to the tight reservoir causes a change in pore pressure, which in turn, changes the three dimensional effective stress state. The stress path followed by the reservoir controls the evolution of the effective stress state, and with it the changes in deviatoric stresses which cause reservoir rock deformation and permeability changes. In order to depict these effects, a new experiment and modeling method-CO2 flooding considering the Interaction between CO2 and Rock - are required. This paper presented a series of stress-sensitive experiments during CO2 flooding. A total of twenty-eight sandstone rock samples, which include sand stones with natural fissures, shear fissures and artificial sanding crack fissures, were selected to study stress-sensitivity during gas production and CO2 flooding. Core flooding experiments were conducted in these rocks. The flow regime used was first depleted follow by injecting CO2. After the experiments, the numerical simulations coupled with nonlinear geo-mechanic model with fluid flow were performed. The simulation results give a detailed understanding of the experimental geo-mechanic system. It is concluded that the experimental and simulation methods can be used in combination to evaluate the potential for stress changes during CO2 flooding in tight gas reservoir. The results show that the evolution of the stress state, captured as a stress path, significantly affects the gas production profile and CO2 storage capability. This research of spatial and temporal changes in stress state laid the basis for studying CO2 storage and enhanced gas recovery.

Speaker: Ying Jia (E&P Research Institute, SINOPEC)
• 17:00
Multi-scale CO2–Brine Core Flooding Under X-Ray CT In Sandstone From Ordos Basin 15m

To understand how does the injected CO2 migration could help increase the available storage capacity in geologic formations, this paper reports a series of experiments of core flooding. To examine the effects of CO2migration pathways in geologic formations, our team have developed a core flooding test of displacing water in porous media with CO2. The samples were obtained from the Ordos Basin, all formations being used for carbon capture and storage(CCS) pilot—Shenhua CCS demonstration project, with a capacity of 100,000 tones CO2 per year, which is the first fully process of CCS in saline reservoir in China. All flooding tests in this paper were performed in the industry CT scanner, medical CT scanner and micro CT scanner.
Experiments were performed over several weeks by injecting CO2-saturated brine through samples. At the same time the samples were scanned with a computed tomography (CT) scanner at regular intervals (0.5mm) during the course of the experiments. Injection flow rates and temperature of the system were varied for each experiment. For the first test of every sample, the helium gas as the flow constant pressure 100 psi at different flow rates (0.72ml/min, 1.48 ml/min, 2.96 ml/min ) was test. Then the brine and CO2 as the flow constant pressure 2000 psi at different flow rates (1.72ml/min, 3.43ml/min, 5.13 ml/min) was test. The constant injection pressure resulted in unstable flow patterns. For the subsequent tests a constant injection rate was set with the Isco pumps and with additional software controls to ensure the pore pressure did not exceed the confining pressure (2500 psi). As long as the injection pressure was less than the confining pressure the flow rate was constant for a constant delta pressure. When the injection pressure increased to a value close to the confining pressure the flow rate was decreased to ensure safe operations.
A review of the findings in common among the studied is presented in the final sections of this paper.

Speaker: Ms. Yan WANG (State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics, Chinese Academy of Sciences)
• 17:15
Impacts of Methane on Carbon Dioxide Storage in Brine Formations 15m

In the context of geological carbon sequestration (GCS), carbon dioxide (CO2) is often injected into deep formations saturated with a brine that may contain dissolved light hydrocarbons such as methane (CH4). In this multicomponent multiphase displacement process, CO2 competes with CH4 in terms of dissolution, and CH4 tends to exsolve from the aqueous into a gaseous phase. Because CH4 has a lower viscosity than injected CO2, CH4 is swept up into a ‘bank’ of CH4-rich gas ahead of the CO2 displacement front. On the one hand, this may provide a useful tracer signal of an approaching CO2 front. On the other hand, the emergence of gaseous CH4 is undesirable because it poses a leakage risk of a far more potent greenhouse gas than CO2 if the cap rock is compromised. Open fractures or faults and wells could result in CH4-contamination of overlying groundwater aquifers as well as surface emissions. We investigate this process through detailed numerical simulations for a large-scale GCS pilot project (near Cranfield, Mississippi) for which a rich set of field data is available. An accurate cubic-plus-association (CPA) equation-of-state (EOS) is used to describe the non-linear phase behavior of multiphase brine-CH4-CO2 mixtures, and breakthrough curves in two observation wells are used to constrain transport processes. Both field data and simulations indeed show the development of an extensive plume of CH4-rich (up to 90 mol%) gas as a consequence of CO2 injection, with important implications for the risk assessment of future GCS projects.

Speaker: Dr. Mohamad Reza Soltanian (University of Cincinnati)
• 17:15
Mixing and reactions: The case of Taylor dispersion in a tube 15m

The progression of reactions in systems where mixing occurs has been the subject of investigation for decades; however, there is still much that is unknown about such systems. One area of particular interest to us is the influence of the initial configuration of a system as it evolves in time. Many (if not most) investigations of mixing are formulated at the long time limit, which requires that a certain amount of relaxation of the system has occurred. Although this is an interesting regime that has relevant to, for example, reactions occurring in the subsurface, it does not describe the initial phases of mixing well.

In this work, we examine how the initial configuration of a system can influence the mixing and reaction process evolution. Such early time conditions have relevance to many systems. For example, in industrial processes the basic design constraints of mixing facilities essentially requires that the mixing regime is dominated by early time behavior (e.g., the effective reaction rate for multi-component injections into tubular, packed bed, or fluidized bed reactors would all generally depend strongly on the initial configuration of the chemicals introduced).

Because mixing is exceptionally complex, we have chosen to examine a two-component mixing and reaction process within a tubular reactor. This choice is motivated in part by the simplifications that this geometry allows, and in part because of recent successes we have had in better understanding the early time dispersion process (pre-asymptotic Taylor dispersion) in such systems. The results of this work will focus primarily upon (1) development of the effective mass transport equations (resulting in an explicit representation of convection, effective dispersion, and the effective reaction rate), and (2) presentation of a closure scheme for this problem, and an prediction of the effective rate of reaction via numerical computations. In particular, we will discuss the need to search for effective empirical dynamical scaling laws even in the presence of a fully-predictive theory. The need for empirical models arises because, in this particular case, the process of upscaling does not reduce the complexity of the problem to the extent that would be the most useful for applications.

Speaker: Brian Wood (Oregon State University)
• 17:15
Optimization of key parameters of carbon dioxide huff-n-puff Process for geological storage 15m

Carbon dioxide huff-n-puff process can be used as an effective geological storage approach for oil reservoirs. In this study, the experiment method of carbon dioxide huff and puff was established by high-pressure physical simulation system for large scale outcrops. The corresponding development effectiveness and influence factors of the huff-n-puff process were analyzed. Such as time of huff-n-puff process, injection pore volume multiple, time of shutting in, fracture density, fracture length, fracture spacing, fracture shape and fracture flow capacity. The results show that in consequence of the parameter of huff-n-puff process and fracture, the carbon dioxide huff-n-puff process in staged fracturing horizontal wells is one of the effective enhanced recovery techniques and the cumulative recovery percent of reservoir rises from 12.37% to 17.35%. Optimized parameters of the huff-n-puff process and fracturing parameters can provide powerful technical support for geological storage of carbon dioxide.

Speaker: Xiangyang Wang
• 17:15
Refractive-light-transmission measurements of density-driven convection with application to solubility trapping of geologically sequestered CO2 15m

Density-driven convection can accelerate the rate of CO2 solubility trapping during geological CO2 storage in deep saline aquifers. We present a bench-scale experimental method based on refractive light transmission (RLT) in an analogue system that enables comprehensive study of solutally induced density-driven convection in saturated porous media. In an analogue system, we investigate density-driven convective mixing under conditions relevant to geological CO2 storage. A range of Ra values relevant to potential storage sites are investigated by varying the grain size and density contrast in the laboratory setup. We show that the method accurately determines the solute concentration in the system with high spatial and temporal resolution. We can thereby quantify the onset time of convection (t_c), mass flux (F) and flow dynamics for the different Ra values tested. Based on our findings, we present a scaling law for t_c. The resulting dependence of t_c on Ra, indicates that t_c is more sensitive to large Ra than previously thought. Our findings can also show why F is described equally well by a Ra-dependent or a Ra-independent scaling law. The new method and findings can serve to improve the understanding of convective mixing processes in saturated porous media, and aid the assessment of CO2 solubility trapping, including potential for trapping under given field conditions.

Speaker: Auli Niemi (Uppsala University)
• 16:00 17:30
Poster 1: Poster 1-B
• 16:00
Pore structure alteration of sands by microbially induced carbonate precipitation via denitrification 15m

Denitrification is one of the key microbial reactions for sandy soils to induce desaturation and calcium carbonate precipitation. As the replacement of urea hydrolysis for microbially induced carbonate precipitation (MICP), the effect by denitrification has been evaluated. Calcium carbonate precipitation and biomass production occur in soil through the reaction process and some of these accumulate in the pore space or on the surface of the soil particles. Due to the accumulation, reducing the porosity, permeability of the soil possibly reduces. Well understanding of pore structure alteration through the reaction makes possible to control permeability efficiently.
This study evaluated pore diameter distributions of three sand samples before and after MICP treatment via denitrification by air intrusion method. By measuring air flow rate applying controlled air pressure into water saturated specimens of the sands, air permeability and pore diameter were calculated. Comparing non-treatment sands, the higher air pressure was required to push pore water against capillary pressure out of pore throats of sands after the treatment. The pore diameter distribution curves were slightly shifted to the smaller pore size range after the treatment. These results indicate that pore water retention ability of the samples was altered by the treatment.

Speaker: Akiko Nakano
• 16:15
Overview of Experimental Systems and Approaches Supporting In Situ Mineral Precipitation Research 15m

The Center for Biofilm Engineering (CBE) at Montana State University has a long, successful history of investigating biofilm and mineral precipitation processes in subsurface environments. This poster summarizes many of the experimental approaches the CBE has taken to develop field-suitable technologies. There are numerous applications for engineered biomineralization. The CBE has largely focused on the sealing of leakage pathways, water remediation, soil stabilization, dust suppression and enhanced resource recovery. The experimental systems to examine these engineering applications have been designed to interrogate biofilm and mineral precipitation processes in a variety of environments, conditions, geometries and chemistries. This poster will highlight experimental systems from low pressure to high pressure, bench scale to field scale and low temperature to high temperature.

Speaker: Adrienne Phillips (Montana State University)
• 16:30
Biocement soil improvement using acidified all-in-one solution by acid buffer 15m

To date, soil bio-cementation via Microbially Induced Carbonate Precipitation (MICP) has been extensively studied as a promising alternative technique for ground improvement to address the growing environmental concerns of traditional chemical cementing agents. This paper presents a new one-phase injection method of biocementation using an acidified all-in-one biocementation solution (i.e., a mixture of bacterial culture, urea, and CaCl2). The key feature of this method is to generate a lag period of the MICP process, which can be controlled by adding acidic pH buffer to the biocementation solution, so that the formation of bio-flocs and CaCO3 crystals is significantly delayed. This feature allows the low-pH all-in-one biocementation solution to be fully injected into the sand column before the biocementation occurs, hence serious surface clogging can be avoided. The duration of the lag phase was evaluated using different amount of HCl or acidic buffer. The performance of biocementation using the acidified all-in-one solution was tested for short and long sand columns bio-stabilization, showing a significant improvement of uniformity. The results of this study show strong potential to scale up the proposed approach to filed applications.

Speakers: Dr. LIANG CHENG (Nanyang Technological University) , Mr. Yang Yang (Nanyang Technological University )
• 16:45
Microbial life in unsaturated porous media: a microfluidic approach 15m

Soil is a complex environment in which the presence of several phases creates numerous interfaces (solid-liquid, liquid-gas and solid-gas). Understanding the local hydrodynamics in soil pores and the biogeochemical processes such as nutrient cycling has been of growing importance in the field of bioremediation and ecology. Besides the coexistence of two immiscible phases (air and water) in the pore space, microorganisms, especially bacteria, are often found in large numbers in natural soil environments. The complex spatial distribution of air and water results in the development of a mosaic of regions of very low water velocity, including areas where water or air is trapped and of preferential channels of high velocity. This landscape of conditions enables microorganisms to live in the free-swimming phase and to form surface attached communities known as biofilms.

At the same time, the biofilms’ structure influences pore geometries resulting in altered hydrodynamics, affecting biofilm development and therefore mass transport. To study influences of soil conditions on biofilms and vice versa, we have studied two soil-born microorganisms, Pseudomonas and Bacillus, at the pore scale using microfluidic devices. We have explored the biofilm forming behavior under different physical conditions such as varied water saturation and flowrate. Carefully designed channel geometries coupled with automated video microscopy allowed us a zoomed-in view on specific interactions while controlling the water saturation by varying the gas flow into the channel. The simplified geometries of the devices resulted in a varied biofilm growth caused by the presence of an immiscible phase.

Speaker: Dorothee Luise Kurz (ETH Zürich)
• 17:00
How biofilm growth affects hydraulic parameters: a reevaluation of the impact of partial in hydraulic conductivity and hydrodynamic dispersion 15m

Microbial dynamics in porous media are drivers for a number of applications in subsurface pollutant remediation. Bioﬁlms are communities of microorganisms that are attached to interfaces (pores-grains), and embedded within a matrix of extracellular polymeric substances (EPS) that they have produced. Growing biofilms have a very small effect on porosity, but a very significant effect on the hydraulic conductivity, that reduces well beyond the value that would be obtained from the Kozeny-Carman updating formula, and further results in order(s) of magnitude increase in the estimated dispersion coefficient. We present a simplified conceptual model that is capable of providing practical expressions for the variations in conductivity and porosity. The advantage of the expressions is that they are written in terms of observables that are relatively easy to measure in the lab or the field, contrarily to most existing expressions. We then tested our simplified expressions in a number of reported experiments. Finally, we see how the simplified model captures the most significant processes of a global multi-compartment mechanistic model recently presented.

Speaker: Prof. Xavier Sanchez-Vila (University Politecnica de Catalunya)
• 16:00 17:30
Poster 1: Poster 1-C
• 16:00
Numerical analysis of viscous oil recovery using micromodel experiments on thermal solvent-based displacement 1h 30m

Hot solvent injection is an in-situ technology which uses heated solvent for efficient and sustainable viscous oil (VO) recovery (cf. 93 kg per barrel less GHG emission than SAGD technology). The process reduces the oil viscosity via mass and heat transfer so that the combined effect of heating and solvent dilution yields a better result than in steam (SAGD) or cold solvent injection (VAPEX) cases.

In this case, the injection of the solvent in gaseous state reduces the amount of solvent required per unit volume of produced oil, and takes advantage of higher rate of solvent diffusion. The solvent will not remain in gaseous state but will condense downstream and release a latent heat making easier its mixing with the VO in-place due to the local temperature increase. Recently the hot solvent injection technology (known also under the name NSolv) had its first successful pilot project designed to recover bitumen from oil sands.

Physically speaking the VO displacement in the solvent-based process is a complex combination of dynamic energy and mass transport and phase transformation phenomena. The rapidly emerging experimental technique of fluid dynamical measurements and observations on micromodels (MM) is proved to be a powerful mean providing quantitative information on multiphysical processes in porous media.

The numerical simulation of solvent injection for VO displacement in a MM setup has demonstrated its feasibility and usefulness both for model design and experimental results analysis. This includes first a pore-scale imaged-based study of micromodel transport properties. Then the Darcy-scale model has been developed and applied for dynamic displacement study. It
has been shown that although being not capable to reproduce in detail the fluid- and thermodynamic diversity of the displacement (especially at pore scale), the developed numerical model has indicated the process key parameters and offered the framework for their quantitative determination.

Finally the dedicated study of process dynamics and corresponding adaptation of the numerical model parameters has been presented and discussed.

Speaker: Marelys Mujica (CHLOE, University of Pau)
• 16:00
Why fractures in Marcellus Shale might be plugged too soon: Case study comparing geochemical and geomechanical data obtained from outcrop vs reservoir cores 15m

Shale rocks play an essential role in petroleum exploration and production because they can occur either as caprocks for subsurface storage in conventional reservoirs or as unconventional reservoir rocks for hydrocarbon extraction via hydraulic fracturing. The ability to produce gas from rocks previously only considered caprocks is an unprecedented and innovative feat, but does not come without an environmental impact and costly issues with permeability reduction of engineered fracture systems. The quantities of water required for hydraulic fracturing and developing these formations for production have been large, and the amounts of flowback and produced water after the hydraulic fracturing processes have been astronomical. These volumes make it imperative that a water recycling solution be found and applied to the development of these fields.

In this study, a batch reaction was conducted with Marcellus shale. Both outcrop and reservoir cores (from different points along wellbore trajectory) were exposed to de-ionized water and a synthetic hydraulic fracturing mixture at reservoir temperature for up to four weeks at a high fluid to rock volume ratio. The chemistry of the created simulated flowback and produced water were analyzed using an ICP. In addition, microstructural analyses were performed in order to establish mineralogical and structural properties, as well as presence of microfractures. Furthermore, indentation tests were conducted at both micro and nanometer level to link the geochemistry and geomechanics of shale rocks, through mechanical properties mapping, the volumetric proportions of each phase can be estimated based on the differential mechanical properties.

The key findings include an analysis of the variation of the simulated flowback water from surface down to cores at a depth of 6420ft, with a focus on heavier mineralogical elements and metals. Less than 1% of the fluid used in these tests consisted of hydraulic fracturing fluid additives, however, even with this small volume of additives used, a significant difference in mineral dissolution compared to the samples treated with water only was observed. The carbonates in the rock samples showed a high level of dissolution, which can cause an increase in permeability, but can also precipitate causing fracture bridging as well as scale buildup in wellbore structure/pipes. The concentration of Pb was found to be significant in the water comparison, posing a potential environmental issue. The indentation results showed a significant difference in mechanical properties as the result of the alteration in microstructures and mineralogical composition during the batch reaction, and the change of microstructure causing by dissolution/precipitation of individual phase were correlated with the alteration in bulk response of the rock.These findings are preliminary and would require and extensive study that would include numerous samples for different location within Marcellus Shale as well comparison to other producing shale formations.

• 16:15
An Assessment of Research Gaps Related to Deep Water Wellbore Integrity 15m

In order for a deep-water wellbore to uphold its integrity under high pressure - high temperature conditions, the wellbore must possess complete zonal isolation while surrounded in an extreme environment. Highly variable temperature and pressure ranges, shallow flow zones, as well as potentially corrosive fluids and gasses all present unique challenges to the job of the cement which maintains that zonal isolation. As such, alternative options to mainstream choices often present themselves as attractive avenues of discovery.

As it is of utmost importance to maintain structural integrity under HPHT conditions, cement slurries are pumped downhole to provide zonal isolation and structural support to offshore wells. The wellbore system potentially faces a variety of temperature and pressure fluctuations from the immediate onset. These fluctuations may affect the hydration properties of the cement. It is also important to consider the chemical interactions that the cement may have at the rock-cement interface where potential degradation or annulus gaps may occur further risking a decrease in zonal isolation. This presentation intends to review some of the important issues regarding zonal isolation in HPHT conditions and to highlight critical knowledge gaps in order to generate important research questions.

Speakers: Mary Tkach , Mileva Radonjic (Louisiana State University)
• 16:30
How 3D Printing maybe used to facilitate the design and testing of hydraulic barriers and establish bonds to formations and casing, as well as data gathering implementation and formation visualization 15m

3d printing in the oil and gas industry is in its infancy. The ability to use 3d printing to not only produce tools and equipment, that could not be manufactured in traditional manners, is only the beginning. There are several possible applications for 3d printing to facilitate this project:
• Controlled deposition of unique and varied barrier materials
• Equipment development for purpose
• Integration of data sensors within casing materials
• Production of planning models for analysis and review

With 3d printing it is possible to “print” in a controlled manner the desired material or material blends to get unique characteristics that would not be possible any other way. The ability to create and blend materials in a planned structure to accommodate the downhole environment would be a step forward in the evolution of hydraulic barriers. 3d printing will allow us to rethink the way oil and gas downhole tools and materials are created and the functions that they perform. The power of 3d printing may allow us to implant long term data gathering sensors right in the barrier materials themselves using technology derived from muti-material deposition from standard FDM 3d printer technology. 3d printing also has many uses in the physical 3d modeling of formation structures and flow analysis for peer review. 3d printings unique layering technology can recreate the details of complex natural geological formations, bringing another dimension to project planning. In 3d printing the user of the tool limits the possibilities of this technology more than the technology of 3d printing itself.

Speaker: Mr. Barry Calnan (President, Calnan Design Group LLC )
• 17:15
Experimental and Computational Tools to Assess Wellbore Integrity: Predicting Failure and Designing Next Generation Seal Repair Materials 15m

We present here our efforts to characterize wellbore interfaces via chemical and mechanical characterization methods. These methods including mechanical “push-out” tests to measure interfacial bond strengths between cement, host media, and polymeric seal repair materials, 3D geochemical modelling of the wellbore environment during subsurface operations, and “mock-wellbore” experimental test setup for simulating downhole stresses and strains and measuring the ensuing in situ permeability. We will show applications of these techniques to case studies, as well as illustrate how a combined experimental and computational approach can be implemented to better understand the multi-scale, coupled processes that are critical to the design and prediction of subsurface seal performance.

Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525. SAND2017-13699 A

Speaker: Edward Matteo (Sandia National Laboratories)
• 17:15
Mathematical model of microbiological oil recovery with wetting inversion by bio-surfactants 15m

• Akerke Mukhamediarova (Institut Elie Cartan, Université de Lorraine)
• Mikhail Panfilov (Institut Elie Cartan – Université de Lorraine ; and
Institut Jean le Rond d’Alembert, Sorbonne Universités)

Oil displacement by water which contains microorganisms able to produce bio-surfactants is one of the most promising methods of oil recovery. The bio-surfactant significantly reduces the surface tension and weakens the negative role of capillary oil trapping. The second effect caused by surfactants is inversion of wetting, which is even more important for oil recovery, since it allows separating oil from pore walls, making it non-wetting (in carbonate reservoirs). We develop the mathematical model of this process, which takes into account both mentioned effects. The model of wetting alternation is its key point. On the macroscale this effect leads to the modification of the relative permeability curves, which may be modeled by special kinetic relationships. The closure relationships for the characteristic time of wetting inversion has been obtained by modelling this process at the pore-scale. The numerical method of diffuse interface was applied to system water-surfactant-oil separated by a meniscus on a solid surface.
For the kinetics of bacterial population grow and decay, we suggest new nonlinear relationships, which enables to model various physiological stages, including the lag stage.
The results of modeling showed the appearance of regimes of self-organization manifested in the form of auto-oscillatory waves in time and in space.

Speaker: Ms. Akerke Mukhamediarova (Institut Elie Cartan – Université de Lorraine)
• 17:15
Pore Scale Dynamics of Gravity-Stable Surfactant Flooding 15m

Surfactants can drastically reduce the water/oil interfacial tension (IFT) to mobilize residual oil. However, surfactant flooding is viscously unstable inherently because of its large mobility ratio. Alkaline/surfactant/polymer (ASP) flooding has been serving as a conventional solution, where polymer plays a vital role in increasing viscosity for a more stable oil bank. Recently, a gravity-stable surfactant (GSS) flooding was proposed as a promising alternative to stabilize the displacement front without mobility control.
In comparison with the sandpack/coreflood experiments, glass-etching model provides a powerful visualization means to examine both static occurrence and flow dynamics of the fluids from a micro perspective. Based on pore network patterns, conceptual models were etched to study the basic mechanism; while actual water-wet models were etched to investigate the effectiveness in heterogeneous scenarios. After the model was vacuumed and NaCl brine-saturated, we took advantage of the gravity, therefore, oil injection (from the top), initial brine injection and surfactant flooding (from the bottom) were vertically performed in turn. Using computer image preprocessing technique, microscopic residual oil in the model is automatically recognized and extracted.
After injecting surfactant solution, microemulsion phase was generated between the oil bank and surfactant slug. Their movement and interaction was visualized and recorded. Firstly, we compared the results among brine at under-optimum, optimum, and over-optimum salinities. Besides, under the condition of optimum salinity, formulation was tuned to reduce microemulsion viscosity to obtain higher critical velocity. Results show that a significant enhanced oil recovery was achieved; and the displacement front was sufficiently maintained with collected oil bank moving ahead under high critical velocity. This validates the modified stability theory and corresponding reservoir simulation. Specifically, according to topological structures, we classify the residual oil into five types for a detailed discussion: membranous flow, droplet flow, columnar flow, multi-porous flow, and clustered flow. And their ratio were calculated through shape recognition. During water flooding, the clustered flow (continuous phase) was mainly transformed into multi-porous flow and columnar flow (non-continuous phase). After the surfactant flooding, yet some reversion occurred due to the in situ emulsification. Moreover, in the follow-up experiments, a less surfactant slug size with a brine post-flush could also realize the desired result, saving more chemical cost.
The novelty of the paper lies in 1) applying microfluidics to the newly designed GSS flooding for microscopic investigation; and 2) quantitatively charactering the residual oil during the flooding. This approach illustrates the phenomena and analyzes the mechanism of the GSS flooding at pore scale, which offers guidance for further effective optimization and economical implementation of the flooding.

Speaker: Mr. Hanxu Yang (China University of Petroleum, Beijing)
• 16:00 17:30
Poster 1: Poster 1-D
• 16:00
Intercalation of trichloroethene by sediment-associated clay minerals 15m

The objective of this research was to examine the potential for intercalation of trichloroethene (TCE) by
clay minerals associated with aquifer sediments. Sediment samples were collected from a field site in
Tucson, AZ. Two widely used Montmorillonite specimen clays were employed as controls. X-ray diffraction,
conducted with a controlled-environment chamber, was used to characterize smectite interlayer dspacing
for three treatments (bulk air-dry sample, sample mixed with synthetic groundwater, sample
mixed with TCE-saturated synthetic groundwater). The results show that the d-spacing measured for
the samples treated with TCE-saturated synthetic groundwater are larger (26%) than those of the
untreated samples for all field samples as well as the specimen clays. These results indicate that TCE
was intercalated by the clay minerals, which may have contributed to the extensive elution tailing
observed in prior miscible-displacement experiments conducted with this sediment.

Speaker: Donald Matthieu (University of Arizona)
• 16:15
Experimental investigations of effective visco-elastic properties of sandstones 15m

The effective visco-elastic properties of reservoir rocks are strongly dependent on characteristics of the pore geometry and of the inherent viscous pore fluids, saturation degree and excitation frequency. Constraining these dependencies is important for the interpretation of seismic data from geothermal or oil and gas reservoirs. Thus, experimental studies are needed that focus on effective visco-elastic properties in the frequency range of seismic field data. For this purpose, some laboratory apparatuses were developed in the past, using the forced oscillation method, in which stress and strain of a harmonically loaded sample are recorded and analyzed. In most of these studies, Young's modulus and seismic wave attenuation were measured at frequencies below 100 Hz (e.g., [1-4]).However, to interpret complete seismic surveys, laboratory data are required up to even higher frequencies. Therefore, we developed a forced oscillation setup aiming at measuring the effective visco-elastic properties of partially and fully saturated rock samples in the seismic frequency range up to 1 kHz. Cylindrical samples are excited to axial oscillations during which the axial force and the axial as well as the lateral strain of the sample are measured to derive Young's modulus and Poisson's ratio and the two corresponding attenuation coefficients by a sliding-window fast Fourier transformation.

We present forced-oscillation experiments with varying amplitude (amplitude sweep) or frequency (frequency sweep) on different materials at ambient conditions. Calibration experiments were performed on visco-elastic polymethyl methacrylate (PMMA), commonly called Plexiglas, and elastic AlCuMgPb-alloy. These well characterized standard materials with contrasting behavior were already used to test similar apparatuses (e.g., [1-4]). The determined visco-elastic properties do not vary with amplitude in the investigated axial strain range of 4e-06 to 5e-05 (PMMA) or 3e-07 to 5e-06 (AlCuMgPb-alloy), but show a frequency dependence for PMMA. For example, the Poisson's ratio of PMMA decreases continuously with increasing frequency, in agreement with previously reported trends [2]. Currently, we perform forced-oscillation experiments on samples of Berea sandstone with a porosity of approximately 18 %. The axial strain range, in which the four mentioned visco-elastic properties show no amplitude dependence, is identified in amplitude sweeps and used in the subsequent frequency sweeps.

Speaker: Prof. Holger Steeb (University of Stuttgart)
• 16:30
Influence of rock micro-pore structure parameters on Remaining oil Distribution 15m

Most of the water-flooding fields in the eastern part of our country have now entered the stage of high-water-cut mining, their actual recovery rates are generally low. Because of this, the research on the remaining oil distribution in the reservoir is urgent. In order to reflect the influence of the pore structure parameters of the core on remaining oil from the microscopic scale, this article starts with sandstone core. The displacement experiment of the selected rock samples and the micro-grayscale images of the cores were obtained after CT scanning. Then images are processed by non-local uniform filtering and watershed segmentation algorithm. Finally, some of the pore structure units were extracted from the remaining oil distribution obtained and a pore network model was constructed. Experimental results show: Rock pore radius is proportional to the degree of enrichment of remaining oil and inversely proportional to the water-flooding effect. The coordination number and shape factor of rock are inversely proportional to the degree of enrichment of remaining oil and proportional to the water-flooding effect. What’s more, the pore-throat ratio value is more intermediate, the lower the enrichment of remaining oil, the better the water-flooding effect. This study is of guiding significance for the design of the remaining oil in the water-flooding oilfield.

Speaker: Ms. Shuyao Sheng (Southwest Petroleum University )
• 16:45
Contact Angle Measurements of scCO2 and Brine in 3D Printed Models with Varying Surface Roughness 15m

Geomaterial pore networks are highly tortuous with intricate geometries and varying surface roughness. It is reported in literature that both pore geometry and surface roughness influence flow through porous media (Ketcham and Carlson, 2001; Noiriel et al., 2016; Lv et al., 2017). Surface roughness is quantified by the deviations in the direction of flow perpendicular to the real surface. Simplified pore networks with known geometric shapes and the quantified surface roughness affords the opportunity to back-calculate internal forces and begin to quantify the effect on contact angles. 3D printed models printed using acrylonitrile butadiene styrene were designed with an internal structure of void geometries to represent a flow path with different geometric interfaces. To look at surface roughness, different techniques were used to add surface roughness to the models. The models were exposed to chemicals that reacted with the material surface to add microscopic surface roughness and macroscopic roughness was added via design and printing techniques. Each model was placed in a core flooding setup and exposed to a series of CO2-saturated brine and scCO2 injections to mimic underground conditions. Once at residual conditions, the core-flooding setup was set to shut-in conditions and scanned used X-Ray micro-computed tomography. 3D reconstructions contain information to measure contact angles, analyze forces, and correlate each to the geometries and surface roughness of each model. Analysis of the local impact to scCO2-brine contact angles within pores with varying surface roughness will be presented.

Speaker: Ms. Laura Dalton (U.S. Department of Energy National Energy Technology Laboratory)
• 16:00 17:30
Poster 1: Poster 1-E
• 16:00
A Dual-site Simplified Local Density Model for Shale Gas Adsorption under Reservoir Conditions 1h 30m

Direct measurement of shale gas adsorption isotherms at high pressures and high temperatures (HPHT) is intricate and requires expensive apparatuses. Most of the documented studies only report shale gas adsorption data at pressures below 12 MPa, which is much smaller than the reservoir pressure, e.g., up to 36 MPa in Eagle Ford shale. Recent studies also suggest that the excess adsorption isotherm of shale gas exhibits distinct features from that observed at low pressures. Therefore, predicting gas adsorption isotherms at reservoir conditions may be useful.
On the basis of the simplified local density (SLD) theory, we developed a novel dual-site adsorption model for shale gas. Shale matrix composed of both organic matter and inorganic minerals and the pores located within kerogen can be smaller than inorganic pores. Our grand canonical Monte Caro (GCMC) simulations also confirm that the adsorption capacity of organic matter is much greater than those of inorganic minerals. Therefore, the model for shale gas adsorption isotherms should distinguishes methane adsorption in kerogen surface from that of the inorganic substrates. Our proposed dual-site SLD model takes into account the different pore sizes and fluid-solid interaction energy parameters of organic matter and inorganic minerals.
We first used conventional SLD model to match the excess adsorption isotherms of CH4 in graphene and montmorillonite slit (pressure: 0-40 MPa). Excellent agreements are observed, which manifest that the SLD model is able to describe gas adsorption in a slit at both subcritical and supercritical states. Then we examined the validity of our proposed dual-site model using high-pressure CH4 adsorption isotherms on shale reported in the literatures. These experimental data were measured at pressures up to 25 MPa and temperatures up to 150 ℃. Our proposed model fit these adsorption isotherms very well. If we use the experimental results measured at low pressures (<12 MPa) to make the fit, the high-pressure isotherms predicted using the fitted parameters are very close to the measured data, which demonstrated the validation of our model.
We also probed the differences of original gas-in-place (OGIP) and production performance estimated using low-pressure adsorption isotherms (always characterized using the Langmuir adsorption isotherms) and high-pressure adsorption isotherms. The great derivations suggest that reliable adsorption isotherms of shale gas under reservoir conditions are very essential. Our proposed dual-site SLD model provides an alternative method to predict shale gas adsorption isotherms under reservoir conditions using low-pressure experimental measurement.

Speaker: Dr. Sen Wang (China University of Petroleum (East China))
• 16:00
Analytical Modeling of Stimulation Fluid Temperature for Hydraulic Fracturing Design 1h 30m

In hydraulic fracturing of unconventional reservoirs, the stimulation fluid is injected at a different temperature than initial reservoir temperature. The dynamic temperature profile of stimulation fluid during the treatment can provide critical information for fracturing design. In this work, an analytical solution to model the stimulation fluid temperature profile during hydraulic fracturing is presented. This analytical model is derived from the energy balance equation for fracture system coupled with a fracture propagation and fluid leak-off model. The procedure to obtain this analytical solution from the governing equation involve Method of Characteristics with valid assumptions. Several important features of the treatment have been preserved, including dynamic fluid leak-off and stimulation fluid velocity inside the fracture. Simple procedures to apply this solution are presented, which provide a convenient way to relate the warm-back temperature profile to the fracture, reservoir, and fluid properties.
The results of the analytical model are presented in terms of the temporal temperature variation inside the fracture. These results are compared and validated in multiple cases with numerical results from commercially available simulation software, as well as simulation results reported in the literature. We identify and analyze the major mechanisms contributing to the temperature signal, which involves the conduction to surrounding stimulated reservoir volume and convection associated with fluid leak-off and varying fluid velocity inside the fracture. The dynamic temperature profile for individual fracture and associated stimulated zone are strong functions of fracture and fluid properties, which include the leak-off coefficient, fracture pore volume, heat transfer coefficient between fracture and stimulated reservoir volume, and density and specific heat of the stimulation fluid. The effect of these factors on temperature distribution is investigated in the sensitivity analysis, which produces several dimensional parameters from this analytical solution. Various types of fracturing treatment design are applied to the developed solution to show its feasibility.
Despite many previous numerical studies on the same issue, this analytical solution brings direct insight into physics behind the stimulation fluid temperature profiles including fracture propagation and fluid leak-off. Besides multiple applications mentioned above, this work can be used as a theoretical basis for a potential analytical approach to address the proceeding warm-back temperature analysis after hydraulic fracturing.

Speaker: Yilin Mao (Louisiana State University)
• 16:00
CO2 interaction with shale: Insights from experiments and literature 1h 30m

Unconventional shale reservoirs with high organic content and swelling clays may have a high affinity for uptake of carbon dioxide (CO2). The pore space and mineral surfaces that sorb/contain petroleum are also potential sorption sites for CO2 and could become available for CO2 uptake once the reservoir is produced and depressurized. Understanding how shales interact with CO2 is important for enhanced resource recovery in the near term, and potential geologic carbon sequestration in the long term.
A series of experiments are discussed where fractured Bakken and Marcellus shale samples were exposed to CO2 at in-situ conditions for extended periods, a computed tomography scanner was used to visualize changes in structure, and simultaneously the fracture permeability was recorded. These measurements allowed for the correlation of hydro-mechanical changes in the fracture which are inferred to be a result of matrix swell and aperture closure.

A detailed examination studies in the literature to extend these results beyond two small samples is then presented. From the experimental CO2/shale interaction results and the growing body of literature on this topic several salient recommendations are presented to unify shale interaction study results into results that can be expanded beyond individual studies. This includes more rigorous characterization of sample constituents, maintaining micro-fabric of the samples, and enhanced control of initial shale water content.

Speaker: Dr. Dustin Crandall (U.S. Department of Energy National Energy Technology Laboratory)
• 16:00
Effect of Salt Dry-out on Shale Gas Reservoir Production Performance 1h 30m

Shale gas reservoirs are typically characterized by nanometer pore throats and very low permeability matrix requiring hydraulic fracturing stimulation of horizontal wells. Water is the main fluid used in hydraulic fracturing and a variety of chemicals are mixed with the water each for a different pur¬pose. Given the very low permeability of shales, very high pressure gradients are experienced in order to achieve economic production rates. The pressure drop results in vaporization of the water by the producing gas. The water being vaporized may be the formation water or a mixture of formation water with the stimulation fluids leaked off into the reservoir. The more pressure drop, the more water is vaporized into the gas and, as a result, the more salt dries out. With the nano-meter size pores in shale gas reservoirs and significant pressure gradients required for economic production, the shale gas reservoir permeability can be significantly damaged due to salt deposition. Salt dry-out can significantly reduce the productivity. Significant adverse impact of salt precipitation on well performance were observed in Marcellus shale play where brine salinities are relatively high.
In this paper, we model a shale gas reservoir having gas-water thermodynamic equilibrium, and investigate the effects of pressure decline induced by production on salt deposition and its consequences on permeability reduction. Reservoir properties representative of Marcellus shale gas play are considered for this study. Peng-Robinson EOS is used for reservoir fluids (methane and water) modeling. A horizontal well is placed in the middle of the reservoir and hydraulic fracturing is considered to create fractures around the well. Conducting hydraulic fracturing, the fractures are invaded by stimulation fluid. Gas is then produced through the fractures under constant rate constraint and water is allowed to be vaporized. Gas-water flash calculations are performed to evaluate the amount of vaporized water. Water vaporization causes the dissolved salt to be dried-out. Change in porosity and then permeability due to salt dry-out is calculated using Kozeny-Carman formula. Saturation of dried-out salt is evaluated and correlated as a time-dependent skin around the wellbore. The method is verified by modeling salt precipitation during CO2 injection into an aquifer and comparing the results with 1D analytical solutions found in the literature. Verifying the method, controlling the gas production from shale gas reservoirs by salt dry-out is elucidated.

• 16:00
Energy exploitation analysis of natural gas hydrate depressurization dissociation in porous media 1h 30m

As a kind of clean and potential energy resource, large quantities of gas hydrates have been proved to exist widely in the permafrost and in deep marine environments with high pressure and low temperature conditions favorable for their formation. Recently, how to develop and exploit the natural gas hydrate reservoir efficiently is one of the critical issues in the energy resource R&D in 21st century. However, gas hydrate dissociation is an endothermic process and the researches on the heat analysis of dissociation process is not enough and the results are often various.
Base on the research findings about the mechanism of gas hydrate dissociation, this study develops the mathematical model including mass conservation, energy conservation, chemical reaction kinetics and geo-mechanic equation and conducts the numerical simulation of gas hydrate dissociation in porous media. Three cases have been simulated in a cylindrical reactor adopting a vertical well in the center with the production well bottom pressure Pin=3.1, 2.5, 2.1 Mpa, respectively. All these cases are well matched with the experiment on the gas recovery factor of methane hydrate in porous of the Konno et al(2014) using the unique apparatus referred to as High-pressure Giant Unit for Methane-hydrate Analyses (HiGUMMA). Both of those Cases (1 to 3) are not dissociated completely in 200 mins because of the higher initial gas hydrate in the porous media. The experiment and the numerical simulation indicate obviously the existence of a “freezing stage” in case of Pin=2.1(Case 3). Four phases of gas, liquid, ice, and hydrate are observed to coexist in the reactor in this freezing stage. Compared with other two cases, the hydrate dissociation in Case 3 is promoted by three kinds of heat: the reservoir sensible heat Qr-sen, the conducted heat from the boundaries Qsur , and the latent heat from ice transition Qi-lat. Meanwhile, the discussions of various heat consumption in Case 3 are represented in detail. It could be conclude that Qr-sen and Qi-lat act as an essential role in inducing the rapid gas production rate in two peaks, which also means the faster dissociation rate of natural gas hydrate in porous media. Furthermore, the entropy of dissociation process is calculated specially according to the second law of thermodynamics. Based on above analysis, the rate of pressure reduction is optimized and the energy efficiency is enhanced. Although the formed ice may cause flow blockage for the gas and water, the released latent heat is still remarkably attractive for the fast hydrate dissociation. In this work, the basic reservoir unit of conceptual model is established and developed, which provides theoretical guidance for the exploitation of the actual hydrate reservoir in the future. Hereafter, the various reservoir conditions would be conducted to explore the transformation patterns of the energy exploitation and entropy.

Speaker: Dr. Didi Wu (School of Petroleum Engineering, China University of Petroleum)
• 16:00
Experimental measurement of permeability in porous medium containing methane hydrate 1h 30m

Natural Gas Hydrate (NGH) widely distributed in marine sediments and permafrost areas has attracted global attentions as potential energy resources. Permeability is a critical parameter that influences the gas production potential from hydrate reservoirs. The hydrate saturation affects the characteristics of the porous media, which is also the key factor determining the permeability. In this study, the absolute permeability and the relative permeability of water were experimentally measured at varying hydrate saturations (0-0.347) in porous medium made of quartz sands and the porosity was 0.3. During permeability measurement, the steady flow and stable differential pressure were obtained under certain water injection rate. Hydrate saturations were controlled and calculated precisely based on the amount of injected and produced gas/water, the system pressure and temperature.
The result indicated that the water relative permeability reduced exponentially with the increase of methane hydrate saturation, and the reduction exponent value of 7.9 was obtained in the Masuda’s permeability model. In addition, hydrate with different saturation in porous media is stored with different forms, which exerts considerable influence on permeability as well. Therefore, a new permeability model based on the weighted combination of pro-filling and grain-coating model was proposed. The weight of pro-filling model is defined as the N-th power of hydrate saturation, and the weight of grain-coating model is defined as the N-th power of non-hydrate saturation. In this work, the calculated index N was 5.6. Compared with the Masuda’s model, the new model not only shows the relationship between permeability and hydrate saturation, but also reflects the aggregate performance of hydrate in porous media.

Speaker: Mr. Fubo Wu (School of Petroleum Engineering,China University of Petroleum(East China))
• 16:00
Laboratory investigations of geochemical evolution in unconventional reservoirs during hydraulic stimulation 1h 30m

Hydraulic fracturing fluids (HFF’s) have been used for several decades to control mechanical, hydraulic, and geochemical behavior in unconventional reservoirs during stimulation. The interactions that occur in these environments during stimulation (hydrofracturing) are designed to prevent scaling, improve production, and prevent damage to formations. However, there is still uncertainty with regards to near fracture geochemical reactions and the evolution of the fluids temporally as they interact with the reservoir rock1.
Rock cores taken from the Marcellus shale, both in outcrop and from a production well, were exposed to simulated HFF’s and simulated formation brines (SFB). The fluids were designed based on regional averages obtained from operators. The tests were conducted at pressures and temperatures representative of regional reservoir conditions, 19.3 MPa and 71°C respectively, and with all fluids under a nitrogen atmosphere to limit free oxygen. Core samples were fractured and loaded with 40/70 US Silica White™ quartz proppant. Tests lasted approximately 96 hours with low flow rates to represent a shut-in period. Geochemical samples were taken daily and analyzed using Inductively Coupled Plasma Mass Spectrometry and Ion-Chromatography. Imaging of the cores was done before exposure using Computed Tomography (CT) scanning and after exposure using CT, Scanning Electron Microscopy (SEM), and Raman Spectrometry.
Control samples, using deionized water as the flow medium, exhibited little to no change in the rock core or in the effluent of the system, dominated by Ca and SO4. Experiments using only SFB showed minor increases in major elemental chemistry in the effluent, consistent with minor dissolution or entrainment of free particles, and minor precipitation of barite/calcite on the fracture surface observed with SEM. HFF chemicals, without SFB, resulted in increases in most elemental constituents in the effluent, with few exceptions including barium which showed an increase followed by a major decrease in concentration. The rock core exhibited significant reaction in CT images and only minor traces of barite/calcite precipitation. The combination of HFF and SFB resulted in large increases in most metal constituents in the effluent during the reaction with the core and significant alteration of the rock matrix adjacent to the fracture. However, this mixture resulted a continual decrease in Ba and SO4 during the experiment, signaling potential deposition of these constituents.
In general, experiments indicated minor pyrite oxidation/dissolution, dissolution of carbonates, and minor precipitation of barite. The degree of precipitation was not of the magnitude observed in Paukert et al. (2017), but there is evidence that carrier fluid composition (>90% of the total volume) is an important consideration in precipitation within unconventional systems and may provide the nucleation surfaces for precipitation. Work continues analyzing precipitates, fracture surfaces, and base-fluid importance in precipitation.

Speaker: Johnathan Moore (AECOM, National Energy Technology Laboratory)
• 16:00
Lattice Boltzmann Simulation of Liquid Flow in Nanoporous Media 1h 30m

A multi-relaxation-time lattice Boltzmann (LB) model for nanoscale liquid flow is developed to investigate the liquid flow characteristics in nanoporous media. The slip length and effective viscosity obtained from molecular dynamics (MD) simulations are adopted to account for the nanoscale effect. First, the LB model for water flow in nanopores is built and water flow characteristics in nanoporous media are investigated. The results show that: (1) the nanoscale effect can either increase or decrease the water flux in nanoporous media, depending on the fluid-solid interaction force; (2) the nanoscale effect impacts the velocity distribution in porous media, making it more uniform in hydrophobic porous media while more heterogeneous in hydrophilic porous media; (3) the end effect caused by the bending of streamlines plays a significant role in water flow in nanoporous media, and neglecting the end effect can greatly overestimate liquid flow ability; and (4) the pore structure also has significant influence on water flow in nanoporous media. With the increase of specific interfacial length, the nanoscale effect increases. In addition, the LB model for oil (octane) flow in quartz nanopores is also established by incorporating the MD simulation results [1]. Oil flow simulation in quartz nanoporous media shows that the conclusions obtained for water flow are also applicable for oil flow.

Speaker: Qinjun Kang (Los Alamos National Laboratory)
• 16:00
Numerical study of effective thermal properties of granular porous medium using Lattice Boltzmann methods 1h 30m

Heat conduction in granular porous media is a phenomenon that is relevant to a broad spectrum of problems in science and engineering disciplines including physical, earth, and biological sciences, to name a few. Effective thermal conductivity in granular porous media is a function of morphological features of the medium such as grain shape, grain size, and geometrical structure. Thermal contact resistance can also affect heat conduction due to topological features such as the surface profile of the grain. Furthermore, the compressive pressure and presence of different fluids in the pore space along with partial saturation can also dictate the nature of the effective thermal properties.
To study the effect of all these factors on the effective thermal conductivity of granular porous medium, we simulate heat conduction by developing a two-dimensional, parallel and thermal Lattice Boltzmann Method (T-LBM) simulator using existing open source libraries. We use this simulator on a digitally reconstructed, two-dimensional granular porous medium that is generated with an existing packing algorithm. We then conduct a progressive investigation by first, introducing thermal contrast resistance as surface roughness on the grains and study its effect on thermal conductivity. Second, we introduce thermal anisotropy in the system by inclusion of elliptical grain packing in the medium. Third, we investigate the effect of partial saturation of water and air in pore space. We use an LBM single component multiphase model to simulate phase segregation in the pore space. We also incorporate elastic deformation of grains based on an existing model, which depicts the surface topology of grains as a self-affine fractal function. This elastic deformation is a function of Young's modulus of the grains and the external compressing pressure.
Based on our investigation, we observe that thermal contact resistance due to the surface roughness of grains reduces effective thermal conductivity. Elliptical packing of grains, manifest thermal anisotropy in the system and causes local heat flux deviations especially when the grain orientation angle changes. External compressive pressures cause elastic deformation of the grain surface and enhance the thermal conductivity of grains with lower Young’s modulus. Introducing partial saturations of water and air in the pore space offsets the effective contribution in heat conduction from the grains as well as the effect of compressing pressure. All of these observations are further accentuated if the thermal contrast ratio of the granular porous medium is changed. We also compare results for selected observations for consistency. A qualitative agreement is obtained with the existing experimental data.

Speaker: Mr. Muhammad Fowaz Ikram (Subsurface Fluidics and Porous Media Laboratory, Department of Chemical and Petroleum Engineering, University of Calgary, Calgary, Alberta, T2N 1N4)
• 16:00
Proppant embedment in shale during exposure to hydraulic fracturing fluids 1h 30m

Proppants are small, granular additives used in hydraulic fracturing to keep induced fractures open and permeable after the reservoir pressure is lowered; typically sand is used. These materials are designed to resist the closure force across a fracture face and allow fluid to migrate out of the system. While the simple mechanical support of the proppant keeping a fracture open is well understood, the interplay between hydraulic fracturing additives, rock strength, and proppants is still lacking.
Experiments were conducted on Marcellus Shale samples that were cored and saw cut to create an artificial, uniform fracture. 40/70 US Silica White™ proppant was loaded into the fracture. Cores were then subjected to a confining pressure of 20.7 MPa at a temperature of 71°C. Each core was then exposed to either deionized water, simulated hydraulic fracturing fluids, or air and allow to react for a period of 5 days. Each core was scanned using computed tomography (CT) before and after the test to evaluate for proppant embedment. Scanning Electron Microscopy (SEM) was used to take high resolution images of the fracture surfaces after the experiment to evaluate embedment features that occur below the CT scan resolution.
Fractured cores that were filled with air or water did not exhibit mechanical signs of change in the CT or SEM. Rock integrity was not compromised and the proppants performed as designed with very little in the way of crushed proppant or rock material. When the hydraulic fracturing components were added, there were clear signs of indentation on the fracture surface.

Speaker: Dr. Sarah Brown (AECOM at NETL)
• 16:00
Study on pressure propagation of methane hydrate decomposition by depressurization in porous medium 1h 30m

Natural gas hydrate is an ice like crystalline compound with a cage structure under high pressure and low temperature. The hydrate will decompose when the pressure is lower than the hydrate equilibrium pressure, so the pressure propagation rule is different from the porous medium without hydrate. Based on the theoretical analysis method of fluid mechanics in porous medium and considering the influence of hydrate existence on the law of influent in porous media, a seepage flow model of hydrate in porous media is established. Analytical solution of pressure distribution in porous media when the hydrates are decomposed can be obtained. Using the numerical simulation software CMG to simulate Gas Production from methane hydrate reservoir by depressurization in Shenhu Area . The law of pressure propagation in the process of decomposition by depressurization mining is obtained, and the results obtained by theoretical analysis are verified.

Speaker: Mrs. Lingxi Qiao (Institute of Petroleum Engineering，China University of Petroleum （East China）)
• 16:00 17:30
Poster 1: Poster 1-F
• 16:00
A domain decomposition method to couple nonisothermal compositional gas liquid Darcy and free gas flows 15m

A domain decomposition algorithm is introduced to couple non isothermal compositional gas liquid Darcy and free gas flow and transport. At each time step, our algorithm solves iteratively the nonlinear system coupling the nonisothermal compositional Darcy flow in the porous medium, the RANS gas flow in the free-flow domain, and the transport of the species and of energy in the free-flow domain.In order to speed up the convergence of the algorithm, the transmission conditions at the interface are replaced by Robin type boundary conditions.The Robin coefficients are obtained from a diagonal approximation of the Dirichlet to Neumann operator related to a simplified model in the neighbouring subdomain.The efficiency of our domain decomposition algorithm is assessed on several test cases focusing on the modeling of the mass and energy exchanges at the interface between the geological formation and the ventilation galleries of geological radioactive waste disposal.

Speaker: Roland Masson (University Côte d'Azur, LJAD-CNRS-Inria)
• 16:15
Effect of Non-Newtonian Foam on SAG Foam EOR 15m

Objectives/Scope:
Foam can improve sweep efficiency in gas-injection enhanced oil recovery. Surfactant-alternating-gas (SAG) is a favored method of foam injection due to injectivity and operational considerations. Laboratory data indicate that foam can be non-Newtonian in the high-quality regime, and therefore during gas injection in a SAG process. We investigate the implications of this finding for mobility control and injectivity, by extending fractional-flow theory to gas injection in a non-Newtonian SAG process in radial flow.
Methods, Procedures, Process:
Non-Newtonian behavior in the high-quality regime means the limiting water saturation for foam stability varies as superficial velocity decreases with radial distance from the well. We discretize the domain radially and perform Buckley-Leverett analysis on each ring; solution characteristics are of constant foam quality. For the first time, we show the implications of this behavior for mobility control at the displacement front as well as injectivity. We base the foam-model parameters and the extent of non-Newtonian behavior on laboratory data in the absence of oil. We compare results to mobilities determined by conventional simulation, where grid resolution is limited.
Results, Observations, Conclusions:
For shear-thinning foam, mobility control improves as the foam front propagates from the well, but injectivity declines somewhat with time. The change of mobility ratio at the front can be considerable, given the huge velocity difference between the wellbore and further out. This change is not simply that measured at steady state at fixed foam quality in the laboratory, however, because the foam front in a non-Newtonian SAG process does not propagate at fixed foam quality. Injectivity benefits from the increased mobility of shear-thinning foam near the well. The foam front, which maintains a constant dimensionless velocity for Newtonian foam, decelerates somewhat with time for shear-thinning foam. For shear-thickening foam, mobility control deteriorates as the foam front advances, though injectivity improves somewhat with time. Overall, however, injectivity suffers from reduced foam mobility at high superficial velocity near the well. The foam front accelerates somewhat with time. Overall, injectivity is a complex result of changing saturations and varying superficial velocities very near the well. Conventional simulators cannot adequately represent these effects, or estimate injectivity accurately, in the absence of exceptional grid resolution near the injection well.
For the first time we extrapolate laboratory steady-state foam data for non-Newtonian foam to investigate the implications for injectivity and mobility control in gas injection in SAG in the field.

Speaker: Rodrigo Salazar Castillo (TU Delft)
• 16:30
Study of foam generation mechanism at the pore scale 15m

Foam injection into the subsurface is generally performed to improve gas mobility control during enhanced-oil recovery (EOR) and contaminated site remediation (Lake et al., 1989; Hirasaki et al., 2000; Mulligan et al., 2006). Several experiments have been conducted to study the foam generation mechanism at both the pore and continuum scales (Kovscek et al., 1994; Kam et al., 2003; Gauteplass et al., 2015; Prigiobbe et al., 2016). Pore scale experiments allow to understand the mechanism of bubble formation with a potential to help formulating constitutive equations for foam flow models improving their accuracy of prediction. However, pore scale studies have not been used to formulate foam generation rate, yet. Here, we present an experimental and modeling work on foam generation mechanism with a porous medium chip. Systematic tests at different flow conditions were performed using various chemicals to stabilize the foam, such as the surfactant, nanoparticles, and a combination of them. The pressure drop and the foam texture were monitored continuously using a pressure transducer and a high-speed high-resolution camera. We observed that to generate a foam in the presence of nanoparticles requires larger energy than when the surfactant is used to stabilize the lamellae. Possibility due to the larger critical capillary pressure for bubble rupture (Pc*) that can be reached in the presence of nanoparticles. Upon image processing, the results show that the generation rate and, therefore, the total number of bubbles increase with the injection rate, creating a more uniform bubble size distribution. We observed that nearby the gas injection the controlling mechanism of the bubble formation is snap-off, while afar from that lamella division dominates.

Speaker: Qingjian Li (Stevens Institute of Technology)
• 16:45
How to Predict CO2 Foam Propagation Distance by Using Bubble Population Balance Model 15m

Although foams are known for effectively reducing gas mobility and enhancing oil recovery in many field applications, it is still not clear how far the injected fine-textured foams will propagate into the reservoirs. Lacking such a knowledge makes the design of foam field treatments difficult and often unreliable. The purpose of this study is to investigate CO2 foam propagation distance as a function of injection foam quality and injection total rate by using bubble population balance model. This study is believed to cover the steps needed from the pore-scale to field-scale events.

In order to meet the purpose, this study performs the following tasks: (i) fitting bubble population balance model to lab coreflood experiments and determining model parameters; (ii) establishing the mathematical framework to determine foam propagation distance during EOR processes; and (iii) characterizing foam propagation distance at different injection strategies. The laboratory data consists of three foam states (weak-foam, strong-foam, and intermediate states) as well as two different flow regimes (high-quality and low-quality regimes) of the strong-foam state.

The mobilization pressure gradient is one of the key model parameters to distinguish gaseous CO2 foams from supercritical CO2 foams. It is because, the mobilization pressure gradient being proportional to the interfacial tension, supercritical (or dense) CO2 foams exhibit much lower mobilization pressure gradient compared to gaseous CO2 foams, often with a couple of orders of magnitude difference.

The results show that the presence of three different foam states as well as two different strong-foam flow regimes (high-quality and low-quality regimes) plays a key role in model fit and field-scale propagation prediction. More specifically, this study finds that supercritical CO2 foams can propagate a few hundreds of feet easily, which is a few orders of magnitude higher than gaseous CO2 foams. For dry foams (or, strong foams in the high-quality regime), higher injection gas fractions result in shorter foam propagation distance, while for wet foams (or, strong foams in the low-quality regime) the propagation distance is not really sensitive to injection gas fractions. In addition, the higher injection rates (or pressures), the farther foams propagate – such an effect is shown to be much more pronounced for dry foams.

• 16:00 17:30
Poster 1: Poster 1-G
• 16:00
A fractal model of permeability for shale gas in fractal fracture networks 15m

Abstract
Random fractures widely exist in water/oil reservoirs, soils etc. Study of the permeability of the fractured networks has been one of focuses in the area of mass transfer in the past decades. Generally, the fractures in scale reservoirs distribute randomly and have statistical self-similarity and fractal characteristic. In this paper, the permeability model for gas flow in the fractured networks in shale reservoirs is derived based on the fractal geometry theory with gaseous slip flow included. The validity of the proposed model is verified by comparisons between the model predictions and experimental data, and the parametric study is also performed in detail. The present results show that the proposed permeability model can reveal more mechanisms of seepage characteristics in the media than the traditional models.

Speaker: Dr. Tongjun Miao (Xinxiang University)
• 16:15
Quantitative evaluation of carbonate reservoir pore structure based on fractal characteristics 15m

Pore structure of large scale porous limestone reservoir with strong heterogeneity is very complex，so it is difficult to evaluate its pore structure of Mishrif Formation of W oilfield in Iraq. Based on thin section observation，porosity and permeability test and mercury injection capillary pressure test，fractal theory was applied to quantitative pore structure evaluation，and the pore fractal dimension criterion for reservoir type classification was established. There are two types of reservoir pore structure fractal characteristics. Some samples called“single segment”perform obvious fractal character overall. Others called“multiple segments”have distinct large pore throat system and small pore throat system which perform unique fractal characters respectively while have no uniform fractal character overall. The complexity and heterogeneity of pore structure of porous limestone can be reflected by fractal dimension，the greater the fractal dimension，the more complex pore structure，and the more conspicuous segmental character in the relationship between capillary pressure and water saturation，the stronger the heterogeneity. The samples were classified based on the fractal dimension combined with porosity and permeability distribution of the samples. The majority of type Ⅰ-Ⅱ and type Ⅲ-Ⅳrespectively corresponded to“multiple segments”and“single segment”fractal characteristics. It has an important guiding significance for the quantitative evaluation of pore structure to similar carbonate reservoir.

Speaker: Mr. Hangyu Liu
• 16:30
Fractal model of gas diffusion coefficient through porous nanofibers with rough surfaces 15m

Fractal model of gas diffusion coefficient is derived for porous nanofibers, which are assumed to be composed of a bundle of tortuous capillaries whose pore size distribution and roughness of wall surfaces of capillaries follow the fractal scaling laws. The analytical expression for gas relative diffusion coefficient is a function of the relative roughness, fiber radius and microstructural parameters (porosity, the fractal dimension for pore size distribution and tortuosity, the maximum and minimum pore diameter and the characteristic length). The proposed fractal model is validated by comparison with available experimental data and correlations. At the same time, the effect of microstructural parameters of porous nanofibers on gas diffusion has been studied in detailed. The results show that roughness of wall surfaces of capillaries in porous nanofibers should not be neglected. It is believed that the current work can reveal gas diffusion mechanism in porous nanofibers and may be applied in other porous materials.

Speaker: Dr. qian zheng
• 16:45
Porosity Characteristics of Coal Reservoir in Daqing Exploration Area 15m

Zhang Yongfeng, Jiang Yongxu, Lu Guoqiang
Exploration and Development Research Institute of Daqing Oilfield Company Ltd.,Daqing,China

ABSTRACT: This papers identity the coal structure in Daqing exploration area, and to discuss types of pore in different basins. Coal is a complicated porous medium. Adsorbability and permeability of its pore structure for coalbed methane (CBM) has drawn extensive attention increasingly. Porosity of coal enables the coal reservoir to store gas and allows CBM to desorb, diffuse and percolate. For this reason, it is of significance to study the pore structure characteristics for exploration and development of CBM as well as evaluation of its mineability. In the three basins of Daqing exploration area, the coal basins include Hailar basin in the west and Jixi and Higang basins in the Sanjiang region in the east as representatives. The coal from the Huhehu depression belongs to lignite. A great deal of intact plant tissue pores can be observed in the SEM. The cell cavities deform to different extents because of compaction effect, but they arranges in a uniform direction with similar shapes, which indicates feature of plant tissues. The fine stratification and fissures can be seen in locality. Its fractures occur in interlaminations, but the fractures connecting the pores are rare, which contribute less to the permeability of coal. So, this is to the disadvantage of migration and deposit of CBM. The Huhehu depression is the one where original texture coal mainly develops. The coal in Jixi and Hegang Basins develops better between gas coal and coking coal. Though parts of plant tissue pores remain in the coal, the pores are filled nearly with minerals. A great many blowholes and emposieu exist in the coal with lithification. Some microstructures such as friction surfaces occur with anabatic deformation. The original texture is destroyed in the tectonic coal so that the pore structure becomes complicated. A large number of micro-fractures and shrinkage joints form connected bridges between pores, which improves seepage capability among coal pores to some extent. There are three types of low temperature nitrogen adsorption cures in Daqing exploration area, which represents different types of pore structure. There is type I in Huhehu drepression. The type I indicates that the coal reservoir has a great many open breathing holes and a few of non-air holes whose one end is closed. The pore size is distributed in twin peaks. Type II and type III can be find in Jixi and Hegang basins. Type II indicates their pores are in composite with multi-pore form and contain ink bottle pores and non- breathing pores whose one end is closed. Type III adsorption and desorption curves have distinct hysteresis loop. The pore volume exists two peaks, but pores at a size of 3—4nm have greatest specific surface area. The micro-pore becomes a greatest contributor to specific surface area. Occurrence of a large number of ink bottle micro-pores is a major reason for difference of adsorption capacity.
KEYWORD: Daqing exploration area; Pores characteristics; Low temperature nitrogen adsorption method; Pore structure

Speaker: Mr. yongfeng zhang
• 17:00
On modeling scale-invariant dual-porosity media based on general fractal topography 15m

Dual-porosity media widely exist in natural reservoirs and have been received much attention in heat and mass transfer. Due to multiplicative cascade effects, the microstructure might be disordered and complicated, with fractures/pores scale-invariantly distributed. In this study, we briefly introduce the concept of General Fractal Topography proposed recently which not only reduces modeling complexity significantly but also admits scaling objects and fractal behaviors as complex as possible. And then, we developed an algorithm to model fractal fracture-pore porous media according to the scaling-invariant topography based on Voronoi tessellations. The original complexity of the fractures and pores distribution is wrapped in the determined phase of scaling object, while the behavior complexity is defined by the fractal topography. Our investigation provides an open framework to unify the definition and modeling of pore, fracture network, and dual-porosity media.

Speaker: Prof. Yi Jin (Henan polytechnic university)
• 17:15
A universal visco-inertial flow model in geologic porous media 15m

Fluid flow though geologic fractured/porous media tends to become non-Dacian as a result of the competition between viscous and inertial forces and the effect of pore geometry variation. The Forchheimer equation has been widely shown to apply in these situations, in which the coefficient of viscous permeability (kv) is largely predictable, but this is not so for the coefficient of inertial permeability (ki). Synthesizing thousands of pore-scale flow models and field and laboratory observations, we show that ki can be predicted from kv via the equation ki~(kv)^(3/2) across twelve and sixteen orders of magnitude in ki and kv, respectively. kv is thus sufficient for predicting flow across viscous-to-inertial regimes for most geologic media.

Speaker: Yi-Feng Chen (Wuhan University)
• 16:00 17:30
Poster 1: Poster 1-H
• 16:00
Hybrid machine learning/adjoint sensitivity model for source zone sampling optimization 15m

Modeling DNAPL source zone plume evolution using traditional flow and transport models is a computationally intensive process that requires specification of a large number of material properties and hydrologic/chemical parameters. Given its computational burden, Monte Carlo simulation using such models is particularly ill-suited for uncertainty assessment and/or subsurface sampling optimization in real field applications. In this work, we present an innovative approach that couples machine learning, adjoint sensitivity theory, and statistical analysis to optimize borehole sampling for quantification of the evolution of down gradient flux-averaged concentration.

Probabilistic models, based on discriminative random fields (DRF), are first employed to synthesize stochastic realizations of a subsurface source zone consistent with known, limited, site characterization data. Using a suite of full scale simulations as training data, a statistical model is developed to predict the spatial distribution and uncertainty associated with key features (i.e. permeability and sequestered contamination [aqueous, sorbed, and NAPL]) that control plume evolution and persistence. Given an initial spatial distribution of contaminant mass, conditioned on measured field data, the adjoint state sensitivity method is then employed to quantify the importance of local system properties on down gradient flux-averaged concentration. The optimal sampling design problem is then addressed using first-order second-moment uncertainty analysis. In the decision process, the costs of additional measurements are justified by a sufficient decrease in the uncertainty, selecting measurements associated with the highest expected worth.

The utility of this probabilistic statistical modeling approach is demonstrated using numerically generated, two-dimensional, heterogeneous DNAPL source zones. Results reveal that down gradient flux-averaged concentration sensitivities to initial contaminant mass compartments are strongly affected by local permeability values. In addition, initial aqueous and sorbed concentrations and their corresponding variances have a major impact on down gradient flux-averaged concentration at early times, while the influence of initial NAPL saturation persists for a longer period. This innovative sampling strategy, coupling sensitivity analysis and uncertainty quantification, shows promise for enhancement of our ability to guide characterization of source zones under realistic field conditions.

Speaker: Tian Tang (Tufts University)
• 16:15
Solute dispersion for stable density-driven flow in randomly heterogeneous porous media 15m

We present a theoretical investigation on the processes underpinning the reduced longitudinal spreading documented in stable variable density flows, as opposed to constant density settings, within heterogeneous porous media. We do so by decomposing velocity and pressure in terms of stationary and dynamic components. The former corresponds to the solution of the constant density flow problem, while the latter accounts for the effects induced by density variability. We focus on a stable flow configuration and analyze the longitudinal spread of saltwater injected from the bottom of a column formed by a heterogeneous porous medium initially fully saturated by freshwater. We adopt a perturbation expansion approach and derive the equations satisfied by section-averaged concentrations and their ensemble mean values. These formulations are respectively characterized by a single realization and an ensemble dispersive flux, which we determine through appropriate closure equations. The latter are solved via semi-analytical and numerical approaches. Our formulations and associated results enable us to discriminate the relative impact on the density-driven solute displacement of (a) covariance of the permeability of the porous medium, (b) cross-covariance between permeability and concentration, which is in turn linked to the coupling of flow and transport problems, and (c) cross-covariance between the dynamic and stationary velocities.

Speakers: Ms. Monica Riva (Dipartimento di Ingegneria Civile e Ambientale, Politecnico di Milano) , Mr. Alberto Gudagnini (Dipartimento di Ingegneria Civile e Ambientale, Politecnico di Milano)
• 16:30
Effects of wettability on two-phase relative permeability estimates from direct pore-scale simulations 15m

We present a numerical analysis of fluid phase distributions and relative permeabilities obtained from direct pore-scale simulations of two-phase flow through a real pore space with diverse conditions of wettability. Exploring the effects of wettability on the fluid behaviors within porous media is of fundamental relevance for a variety of engineering as well as environmental applications, including, e.g., conventional and unconventional hydrocarbon extraction, along with CO2 storage in subsurface systems. The pore space investigated is a three-dimensional limestone rock reconstructed via X-ray micro-tomography at 2 microns voxel size resolution. The one millimeter-scale rock sample is associated to a connected porosity Phi = 0.128 and an absolute permeability k = 357 mD. We employ a simulation procedure mimicking steady-state protocols implemented to estimate multiphase relative permeability by laboratory-scale experiments. We first simulate a primary-drainage (oil injection) process until steady-state conditions are reached. Then, we simulate a series of secondary-imbibition (water injection) processes by injecting diverse amount of water/oil volumes until steady-state conditions are achieved. The ratios between oil and water densities and viscosities are set equal to 0.78 and 2.87, respectively. In all our simulations the capillary number is constant and equal to 1e-5. Flow field and fluid phase distributions are calculated within the explicit three-dimensional pore space using a finite volume-based solver and implementing the so-called Volume-Of-Fluid technique. Capillary end effects that may arise at the inlet and outlet boundaries are removed by implementing ad-hoc periodic conditions (see ref [1]). In this context, adjustment of fluid-phase saturations is achieved through local injection of one of the two fluids. The total fluid flow rate is indirectly set through a pressure jump across the periodic inlet and outlet boundaries. Simulations were achieved using HPC resources at CINECA computing center [2], within the LISA-Prod Project HPL13P71U0.
In order to assess the impact of wettability conditions on the resulting two-phase relative permeability curves, multiphase flow simulations are performed by using diverse contact angle values at the solid-fluid boundaries resulting in water-, neutral- and oil-wet conditions. The numerical errors associated with both geometry and pore space discretization are analyzed by performing a sensitivity analysis reconstructing the pore-space with diverse spatial resolution scales. We observe that errors in the geometrical representation have an impact on the averaged capillary pressure as well as on residual fluid saturations, while errors originating from the pore-space discretization impact water and oil relative permeabilities. Our results document that varying the contact angle has a strong influence on residual oil and water saturations, relative permeability curves and spatial distribution of fluid phases observed at the end of the primary-drainage and secondary-imbibition processes.

Speaker: Dr. Gaël Raymond Guédon (Politecnico di Milano, Dipartimento di Energia)
• 16:45
Image-based modeling of flow and transport in porous media 15m

Flow and transport in porous media is encountered in many industrial and hydrogeological applications, such as hydrogen fuel cells, inkjet printing, hydrocarbon exploration, and subsurface remediation. The relevant study domain can cross multiple scales from a few nanometers to hundreds of kilometers. Therefore, in porous media research, the upscaling and multiscale techniques have been widely used to bridge the hierarchical scales. Recently, with the increase of powerful computational resources and the improvement of characterization methods of pore structures, image-based modeling is attracting much attention. It sheds light on fundaments of pore-scale flow dynamics, and assists in the upscaling of porous media models. In this work, we will discuss the current challenges of image-based modeling of single-phase reactive transport and two-phase flow in various porous materials (e.g., paper and shale). Then, we will present our PDMS-based micromodel benchmark experiments of two-phase flow in porous media. The obtained data will be used to calibrate and validate pore-scale models.

Speaker: Dr. Chao-Zhong Qin (Eindhoven University of Technology )
• 17:00
The Study of Solid Phase Particles Blocking Process based on CT scanning technology 15m

The pore blocking caused by solid particles migration is the major reason to formation damage. In order to further describe the solid particles blocking process, the realistic pore network model is established based on the results of micro-CT scanning. At the same time, the granularity distribution model is generated according to the solid phase particles size distribution. Then the “blocking volume” is introduced to judge whether or not pore blocking happen. Based on the conception of “blocking volume”, pore element is generated and all these characteristic parameters of pore element are calculated such as the flow distribution, pore size, particles granularity, blocking volume. According to the judgement standard, the intrusion ratio of solid particles, the sediment ratio of solid particles, the blocking ratio of solid particles, the sediment depth and the blocking depth are all obtained. In addition to this, the parameter sensitivity analyses of influence factors are taken. Through the study of pore blocking simulation, the aim to predict the probable formation damage caused by solid particles based on the structure of core and particles parameters have been achieved. It provides

Speaker: Zhenglan Li
• 17:15
Observation and Evaluation of Multiphase Flow in Heterogeneous Porous Media of Tight Gas Reservoir with Super-normally Saturated Water 15m

The phenomena of tight gas reservoirs with super-normally saturated water, which exhibit the unique combination of very high initial water saturation combined with low to very low permeability (permeability of less than 0.1 mD) exist extensively in a number of regional sedimentary basins. Traditionally, multiphase flow of natural gas and connate water is evaluated in laboratory by conducting flow experiments. However, due to the lack of sufficient insight vision, this traditional method only reflects the macroscopic phenomena but fails to reveal some important microscopic behaviors inside the cores of different type of reservoir during multiphase flow process. In this paper, a new method, which combined the multiphase flow experiment and nuclear magnetic resonance experiment, was introduced to effectively observe and characterize the water flow to various multiphase flow scenarios in different types of rocks. First of all, a series of core displacement experiments were carried out. Then, based on the displacement experiments, the pore size distribution of rock samples is converted by NMR T2 spectrum by T2 spectrum experiment. The distribution of the remaining water under different water saturation of different cores was analyzed and the water production mechanisms of tight high water saturation reservoir were revealed. Finally, we prove the relationship between gas & water production and pore structure using effective flow formula. Through theoretical analysis and calculation, the physical significance and the value of two main parameters in gas production formula are made clear. This experiment and evaluation methods provided a valuable tool in the evaluation of whether given pay zones in an ultra-tight gas reservoir situation are worthy of completion and what kind of pressure should be provided for production. It also provides information on how this new method can be used in the exploitation of this ever increasing area of tight gas reservoir production.

Speaker: Dr. Ying Jia (E&P Research Institute, SINOPEC)
• 17:15
Pore Network Stitching for Pore-to-Core Upscaling of Capillary-Dominated Two-Phase Flow in Heterogeneous Natural Reservoir Rocks 15m

Physics of two-phase flows in heterogeneous natural rocks plays an important role in many applications, such as carbon sequestration in deep saline reservoirs and recovery of oil from hydrocarbon reservoirs. Although pore-scale models are used to compute macroscopic average properties required in field-scale simulators, most work is limited to small sample size. There is a need for pore-scale models that can accurately represent the 3D complex pore structure and heterogeneity of real media. Pore network modeling (PNM) simplifies the geometry and flow equations at pore-scale, but can provide characteristic curves in capillary-dominated systems on fairly large samples with huge saving on computational costs compare to direct numerical simulation methods. However, there are limitations for attaining a large representative pore network for heterogeneous cores such as technical limits on scanning size to discern void space and computational limits on pore network extraction methods. To address this issue, we propose a novel pore network stitching method to provide large-enough representative pore network for a core.

In this study, we use industrial (as core-scale) and micro-CT (as pore-scale) scans of actual reservoir rock samples to characterize the pore structure of a core. Few signature parts of the core are selected from industrial scans, and their micro-CT scans are taken. Equivalent 3D pore network of each signature part is extracted by applying maximal ball pore network extraction algorithm. The space between signature networks is filled by using stochastic random network generator that uses statistics (radius, shape factor, connection number, and length of pore elements) of all signature networks and a layered stitching method that glues network pieces. The outcome is a large pore network that can be used in a fast quasi-static PNM solver to obtain absolute permeability, relative permeability and capillary pressure curves. We have tested the developed method on various generated and extracted networks by 1D stitching direction, and it will be improved and extended to 2D and 3D stitching directions.

This work is primarily supported as part of the Center for Geologic Storage of CO2, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science and partially supported by the International Institute for Carbon-Neutral Energy Research (WPI-I2CNER) based at Kyushu University, Japan.

Speaker: Amir Kohanpur (University of Illinois at Urbana-Champaign)
• Tuesday, 15 May
• 09:30 10:44
Parallel 3-A
• 09:32
Methane Migration in Water Saturated Formations --- Applications to CO2 Sequestration and Groundwater Contamination from Leaky Natural Gas Wells 15m

Target formations for large-scale CO2 sequestration are often saturated with brines that contain dissolved methane and other light hydrocarbons. When CO2 is injected in such deep formation, non-trivial phase behavior may result in methane exsolving from the brine and forming a free gas phase. Because methane has a lower viscosity than the (generally supercritical) CO2, this methane-rich gas is swept up ahead of the CO2 front. On the one hand, such a bank’ of methane may provide an early tracer warning of the approaching CO2 plume, e.g. in observation wells. On the other hand, the emergence of gaseous methane poses risks of releasing a potent (greenhouse) gas contaminant in overlying groundwater and potentially the atmosphere if the formation integrity is compromised by (open) fractures, faults, or leaky wells.

The transport of methane in water-saturated formations is also important in the context of natural-gas production from deep reservoirs. If a producing well is compromised, e.g., at shallower depths, leaking stray natural-gas may contaminate groundwater resources. Whether this contamination occurs in a small radius around the well (due to buoyancy) or travels significant distances laterally before contaminating groundwater wells depends strongly on the formation heterogeneity, notably fractures.

The modeling of these important processes is complicated by 1) strong heterogeneity in fluvial target formations for CO2 storage and fractured groundwater aquifers, and 2) the complicated phase behavior of mixtures of water and hydrocarbons. The latter can be modeled accurately by the cubic-plus-association (CPA) equation of state (EOS), which takes into account the polar nature of water molecules, its self-association, and the polar-induced cross-association between water, CO2, and methane molecules (as well as compressibility of the aqueous phase). While accurate, the CPA EOS is highly non-linear and computationally expensive. In this work, we develop new efficient algorithms to adopt the CPA EOS for large-scale simulations. Flow and transport are modeled by the mixed hybrid and discontinuous Galerkin methods, respectively, and discrete fractures are incorporated through a cross-flow equilibrium approach.

Simulation results are presented for 1) the Cranfield large-volume CO2 storage pilot project, and 2) for lateral migration of stray methane leaking from a compromised natural-gas well into shallow fractured groundwater aquifers for conditions representative of those overlying the Barnett formation in Texas.

Speaker: Prof. Joachim Moortgat (The Ohio State University)
• 09:50
Robust algorithms for stability analysis and flash calculation of reservoir fluids at constant moles, volume and temperature 15m

Accurate modeling and robust computation of the phase behavior is essential for optimal design and cost-effective operations in petroleum reservoirs as well as in petroleum processing plants, where we need to understand the fluid flow of partially miscible multi-component multi-phase mixture in free spaces or in porous media. Phase behavior calculation of fluid mixture consists of stability analysis and flash equilibrium calculation. The goal of the NVT stability analysis is to determine whether a phase is stable at specified volume, temperature, and mole numbers. If it is not stable, the NVT flash equilibrium calculation is to establish the composition and amount of each stable phase after phase splitting. Conventional algorithms for stability analysis and flash calculation are based on fixed-point iteration, Newton-type iteration, or their combination, and the convergence has never been guaranteed. In this work, we propose an energy-stable iterative method for NVT stability analysis and NVT flash equilibrium calculation. We consider fluid mixture modeled by the Peng-Robinson equation of state, and our proposed algorithm is an iterative algorithm motivated from the dynamics of two-phase fluid system with Fick's law of diffusion for multi-component fluids. The proposed iterative procedure is proven to be energy stable under certain conditions. Numerical examples are tested to demonstrate efficiency and robustness of the proposed method. We also discuss the extension of the algorithm to NVT flash in multiple spatial dimensions, which can be used to model the interface of nonzero thickness between two phases.

Speaker: Shuyu Sun (King Abdullah University of Science and Technology (KAUST))
• 10:08
Analysis of enhanced gas transport in fractured rock due to barometric pressure variations 15m

Barometric pressure variations are often one of the main drivers of gas transport in fractured rock, a process that is referred to as barometric pumping. Barometric pressure variations are complex, multi-frequency signals influenced by latitude, weather, elevation, lunar phase, time of year, and diurnal and semi-diurnal earth tides. However, our results indicate that it is often a subset of the pressure frequencies that lead to the vast majority of transport while the majority of frequencies result in minor or even insignificant transport. Identifying the dominant pressure frequencies for transport allow us to more simply and effectively characterize the potential for gas transport to the surface at different geographic locations. We will present barometric pressure decomposition analyses on gas transport in fractured rock.

Speaker: Dr. Dylan Harp (Los Alamos National Laboratory, Earth and Environmental Sciences Division)
• 10:26
Multiphase Flow and Underpressured Shale at the Bruce nuclear site, Ontario, Canada 15m

A deep geologic repository (DGR) for low- and intermediate-level radioactive waste has been proposed at the Bruce nuclear complex on the eastern flank of the Michigan Basin in southeastern Ontario, Canada. The proposed location for the repository is at a depth of ~680 m, in the middle of a ~450 m-thick sequence of Ordovician-aged shale and limestone with extremely low porosity and permeability, which makes fluid flow and mass transport processes very slow. Significant underpressure exists in these rocks, and questions have been raised about whether gas phase methane is present and how it relates to the generation and persistence of the underpressure here, as well as those in numerous other shale- and gas-rich sedimentary basins around the world. Multiphase flow simulations have suggested that water can become underpressured in the presence of gas phase due to transient glacial loading cycles, and a previous modeling study of the Bruce site, in which the presence of gas phase was approximated using ad hoc adjustments to single-phase flow parameters, showed that underpressures can persist for geologically significant periods of time. However, while multiphase interaction and migration processes have been studied extensively for conventional petroleum and environmental engineering applications, they are relatively poorly understood in low-permeability argillaceous rocks such as those at the DGR. The goals of this study are to: (1) determine which rock and fluid parameters are most critical for understanding the multiphase flow processes that may have occurred in the low-permeability formations at the Bruce site through geologic time, (2) assess uncertainty in our understanding of those parameters, and (3) investigate, using the multiphase flow simulator iTOUGH2-EOS7C, whether the presence of gas phase methane could have generated or contributed to the underpressure. Results suggest that the presence of gas phase methane does not by itself fully explain the underpressure.

Speaker: Michael Plampin (U.S. Geological Survey)
• 09:30 10:41
Parallel 3-B
• 09:32
Filter media design for Dust Holding Capacity by computer simulations 15m

The goal of this study is to optimize a dual-fiber filter media by increasing the dust holding capacity (DHC), while maintaining the initial pressure drop and initial filter efficiency. Three main parameters define the performance of a filter, namely the DHC, filter efficiency and pressure drop. The DHC defines the quantity of solid particles which a filter media can trap and hold before the maximum allowable pressure drop is reached. The key idea is the use of micro structure models to optimize the filter media. To generate various filter media, the FiberGeo module of GeoDict® software is used, while the FlowDict and FilterDict modules simulate the flow behavior and the particle filtration behavior of the media, respectively.
Three different dual-fiber filter media are modelled. They all consist of the same two types of glass fibers (bi-modal diameter distribution). The distribution of the coarser fibers is uniform for all three models. They provide stiffness to the media and support the finer fibers. However, the finer fibers are distributed differently over the through-direction of the filter media [1]. The three models are called homogeneous model, linear increasing model, and exponentially increasing model. The naming scheme is based on the distribution of the finer fibers in through-direction of the filter media. In order to make the three models comparable, they were designed to have the same pressure drop in the clean state and to have the same initial filter efficiency. To ensure this, FlowDict and FilterDict simulations were used to guide the choice of geometric model parameters.
The first simulation step is to model the filter media with different spatial distributions of the two types of fibers. The next step is to simulate the clean fluid flow through the filter media. The final step is to simulate the particulate flow and particle deposition.
Filtration simulations on the three different models were done using the FilterDict module of GeoDict®. Filter life time simulations were carried out using the multi pass mode. In a multi pass simulation, fluids move in a circuit through the system, and the particle size distribution and concentration in front of the filter change over time. Outputs of these simulations are pressure drop, filtration efficiency and count/mass of the deposited dust as a function of time. The results show that by changing the distribution of the thin fiber type over the through-direction, the DHC of the filter media changes as well. In this way, the structure of the dual-fiber filter media can be optimized to achieve higher DHC, while keeping the initial pressure drop and initial filter efficiency the same.

Speaker: Sven Linden (Math2Market GmbH)
• 09:50
Membrane morphology and topology: Fouling control in filtration systems 15m

Reverse Osmosis Membrane (ROM) filtration systems are widely utilized in waste-water recovery, seawater desalination, landfill water treatment, etc. During filtration, the system performance is dramatically affected by membrane fouling which causes a significant decrease in permeate flux as well as an increase in the energy input required to operate the system. Design and optimization of ROM filtration systems aim at reducing membrane fouling by studying the coupling between membrane structure, local flow field and foulant adsorption patterns. Yet, current studies focus exclusively on oversimplified steady-state models that ignore any dynamic coupling between fluid flow and transport through the membrane. In this work, we develop a customized solver (SUMs) under OpenFOAM to solve the transient equations. The simulation results not only predict macroscopic quantities (e.g. permeate flux, pressure drop, etc.) but also show an excellent agreement with the fouling patterns observed in experiments. It is observed that foulant deposition is strongly controlled by the local shear stress on the membrane, and channel morphology or membrane topology can be modified to control the shear stress distribution and reduce fouling. We demonstrate how channel morphology and membrane topology can be jointly optimized in order to increase the efficiency of the system. Finally, we identify optimal regimes for morphological and topological modifications in different operation conditions.

Speaker: Mr. Bowen Ling (Stanford University)
• 10:08
Comparative simulation of reactive flow in catalytic filter using 3D pore-scale model on CT image and 1D effective model 15m

In order to improve the performance of the exhaust after treatment system and keep reasonable complexity, the number of the used devices is reduced by enhancing wall flow particulate filters with a catalytic functionality, like selective catalytic reduction in diesel or three way catalysis in gasoline vehicles. In this case the solid matrix of the filtering media consists from inert grains and active grains. For the simulation of such devices, effective 1D models are regularly used, because they are relatively simple and fast in comparison to higher dimensional models, while reproducing most of the experimentally observed phenomena with sufficient accuracy. The reduction of the complexity in the modeling, however, is achieved by introducing many simplifying assumptions.

As it will be seen in this presentation, 1D models do not always reproduce detailed 3D simulations. Here we compare a standard 1D homogenous wall model with a 3D pore-scale model. The latter describes convection and diffusion in the pores and in washcoat grains, and absorption in the washcoat grains. In fact, washcoat particles are nanoporous and surface reaction (adsorption) occurs at this scale. Softare tool called PoreChem [1] has been used for simulating the reactive flow at pore scale. A wall segment of a real particulate filter was used in the simulation. The three dimensional structure of the wall segment was obtained by X-ray microtomography, in which we resolved the different materials: Pores, (inert) Substrate and (active) Washcoat. A first order reaction was studied to examine if 1D simulations based on effective (Darcy scale) model can describe the behavior of a catalytic filter wall sufficiently well. In simulations the adsorption rate constant varied while the other properties, like temperature and flow speed remained constant. The conversion (adsorbed amount) was computed for different reaction rate. The discrepancy between the results obtained with the 3D and 1D models increase with higher reaction rates. They are caused by the inhomogeneous flow and washcoat distribution in the 3D system that cannot be described well with a homogenized 1D model. 3D models are therefore useful to optimize the microstructure of catalytic particulate filters.

Speaker: Prof. Oleg Illiev (Fraunhofer ITWM, Kaiserslautern, Germany )
• 10:26
Influence of the microstructure of non-woven media on filtration performance at different operational regimes. 15m

In this work, we study the efficiency of the filter media at different operational regimes. In particular, we focus on how the media microstructure influences the efficiency performance, which is quantified by macroscopic parameters. We investigate the microscale characteristics since the filtration is an intrinsically multiscale process. On one hand, contaminant adsorption onto the fibre surface occurs at the pore-scale of the non-woven filter media, i.e., it is a microscale process. On the other hand, we are interested in the overall performance of the filter media, which is a macroscale characteristic, and how the performance changes for different macroscopic characteristics, such as porosity and thickness of the filter media and operational conditions. Filtration processes at both scales are fully coupled. Therefore, to model this problem we use the method of multiple scales, which is an upscaling technique that models variations at macroscale while accounting for filtration processes at microscale. The advantage of this multiscale method is that the coupled micro- and macroscale behaviour is captured without the computational expense of globally resolving the microscale filtration problem. We account for: a single-phase fluid flow through the filter medium, the contaminant transport with convection, diffusion and adsorption and the evolution of the filter medium microstructure due to the contaminant adsorption. Using developed simulation framework, we perform extended numerical study to investigate the influence of microstructure on the filtration performance at different filtration regimes.

Speaker: Dr. G. Printsypar (University of Oxford)
• 11:17 12:30
Parallel 4-A
• 11:17
Pore scale modeling of water flow in limestone by cascaded lattice Boltzmann method 15m

Lattice Boltzmann methods (LBM) has established itself as strong candidate to simulate pore scale fluid flow in porous media. We present pore scale modeling of water flow in limestone porous rocks using central moment based cascaded collision schemes on a three dimensional D3Q27 lattice model. Cascaded collision schemes (CLBM) enhance the stability of the lattice Boltzmann scheme, and central moment approach reduces the insufficient degree of Galilean invariancy. For pore scale characterization, we use X-ray microcomputed tomography to generate 3D images of limestone and perform water flow simulations. CLBM performs very well producing results with high stability and accuracy.

Speaker: Dr. Robert Straka (AGH university of science and technology)
• 11:35
Sensitivity of bare non-vegetated soil moisture dynamic simulations to prescribed soil-atmosphere interface boundary condition forcings 15m

Soil moisture is closely linked to the near-surface heat and mass transfer that couples the land and atmospheric states. The accurate simulation of the spatiotemporal distribution of soil moisture is constrained by existing knowledge gaps with respect to the mechanisms and processes linking the atmospheric and soil states, their magnitude, and sensitivity to applied soil conditions. In this study, we explore the effects of variations in microclimate conditions on state variable distributions and water balances for different bare soil conditions. A series of evaporative experiments were conducted at the Center for Experimental Study of Subsurface Environmental Processes (CESEP) wind tunnel-porous media test-facility to generate atmospheric and subsurface datasets that were in turn applied in the prescribed soil-atmosphere boundary conditions of a heat and mass transfer numerical model. Experimental results showed that for the length scale and edaphic conditions tested, variations in local soil-atmosphere coupling had a slight impact on the lateral distribution of soil moisture. This localized soil moisture variability could not be reproduced with numerical model. From a water balance perspective however, cumulative water loss could be adequately captured with little loss of fidelity. This demonstrates that heat and mass transfer models may be insensitive to the local microclimate driving bare-soil evaporation but are strongly influenced by local soil properties (i.e., heterogeneity). Together, these findings suggest that greater focus should be given to characterizing subsurface conditions and subsurface constitutive models and heat and mass transfer theory than localized near-surface atmosphere conditions.

Speaker: Dr. Tissa Illangasekare (Colorado School of Mines)
• 11:17 12:47
Parallel 4-B
• 11:17
DLVO Interaction Energies between Hollow Spherical Particles and Collector Surfaces 15m

The surface element integration technique was used to systematically study Derjaguin-Landau-Verwey-Overbeek (DLVO) interaction energies/forces between hollow spherical particles (HPs) and a planar surface or two intercepting half planes under different ionic strength conditions. The inner and outer spheres of HPs were concentric (CHP) or in point contact (PHP). In comparison to a solid particle, the attractive van der Waals interaction was reduced with increasing inner radius of the CHP, but the reduction effect was less significant for the CHP at smaller separation distance. Increasing the inner radius for CHP therefore reduced the depths of the secondary minima, but had minor influence on the energy barrier heights and depths of the primary minima. Consequently, increasing inner radius reduced the potential for CHP retention in secondary minima, whereas did not influence the retention in primary minima. For PHP these interaction energy parameters and colloid retention depended on the orientation of the inner sphere relative to interacting surface. In particular, the van der Waals attraction was significantly reduced at all separation distances when the inner sphere was closest to the interacting surface, and this diminished retention in both secondary and primary minima. The PHP retention was similar to that of CHP when the inner sphere was farthest from the interaction surface. These orientation dependent interaction energies/forces resulted in directional bonds between PHPs and the formation of aggregates with contact points of the primary PHPs facing outward. The findings in this study have important implications for the design and utilization of HPs in soil remediation and colloid assembly.

Speaker: Chongyang Shen (China Agricultural University)
• 11:35
Interfacial Processes Control Microbial Contamination and Cleaning of Fresh Produce 15m

Foodborne illnesses involving fresh produce have been increasingly causing concerns around the world. Pathogenic bacteria can attach to and colonize the surfaces of fresh produce, leading to contamination and illness outbreaks. however, mechanistic interactions between produce surface properties (e.g. roughness, topography and hydrophobicity) and bacterial retention remain poorly understood. As a result, effective strategies eliminating pathogenic contaminants for fresh produce are not yet available. Using produce surfaces and their replicas, we systematically evaluated bacterial/colloid retention and removal as a function of physicochemical properties (roughness and hydrophobicity) of these surfaces as well as characteristics of water retention and distribution on the surfaces. We found that water retention and associated interfacial behavior associated with water on produce/replica surfaces are the key parameters that dominantly affect bacterial/colloid retention and removal, and those parameters in turn are collectively governed by surface roughness, topography and hydrophobicity. Based on these insights, we are developing new methods/strategies for more effective cleaning of fresh produce and other contaminated surfaces.

Speaker: Yan Jin (University of Delaware)
• 12:11
On the macroscopic momentum balance equation for the fluid-fluid interfaces in two-phase porous media flows 15m

Although two-phase flow in porous media is an established research field since decades, its theoretical background is still incomplete. In particular, while a universal definition of capillary pressure exists at the micro-scale, its upscaling to the macro-scale is still rather vague and a rigorous theory of capillarity at the macro-scale is missing. In this work, a new macroscopic theory of capillarity based on the volume averaging method is presented. The novel feature of the proposed averaging approach is the use of the superficial surface average for upscaling the relevant conservation equations for a surface. This allows for rigorous derivation of the macroscopic momentum balance equation for all the fluid-fluid interfaces contained within the Representative Elementary Volume (REV), thus resolving most of the shortcomings of previous studies, such as the averaging-scale inconsistency and the accounting for the different orientation of interfaces within the averaging volume. This latter aspect is described by an additional parameter arising in the proposed derivation, namely the intrinsic surface average of interface normal vectors $\langle \mathbf{n}_{nw}\rangle^{nw}$. Furthermore, defining the macroscopic capillary pressure as the difference between the intrinsic surface averages of the bulk pressures, it is shown how the capillary pressure-fluid phase saturation curve can be determined in a more consistent manner by upscaling results of pore-scale simulations as oppose to traditional coreflooding experiments. This sets new challenges and opportunities for modelling unsaturated porous materials.

Speaker: Michele Starnoni (University of Bergen)
• 12:29
Hydromechanical couplings in multiphase granular systems: recent advances and perspectives 15m

This presentation report recent advances in the framework of the discrete element method (DEM) for multiphase granular media. Computationally efficient methods based on the DEM have been developed for a while for partially saturated materials but they have been generally limited to the pendular regime. In contrast, one hardly avoid expensive direct resolutions of 2-phase fluid dynamics problem for mixed pendular-funicular situations or even saturated regimes. Following previous developments for single-phase flow, a pore-network approach of the coupling problems is described. The geometry and movements of phases and interfaces are described on the basis of a tetrahedrization of the pore space, introducing elementary objects such as bridge, meniscus, pore body and pore throat, together with local rules of evolution [1]. As firmly established local rules are still missing on some aspects (entry capillary pressure and pore-scale pressure-saturation relations, forces on the grains, or kinetics of transfers in mixed situations) a multi-scale numerical framework is introduced, enhancing the pore-network approach with the help of direct simulations [2]. Small subsets of a granular system are extracted, in which multiphase scenario are solved using the Lattice-Boltzman method (LBM). In turns, a global problem is assembled and solved at the network scale, as illustrated by a simulated primary drainage.

Speaker: Dr. Bruno Chareyre (Univ. Grenoble Alpes, 3SR lab. - France)
• 14:37 17:15
Parallel 5-A
• 14:37
Effects of Wettability and Permeability on Viscous Fingering during Unstable Immiscible Displacements 15m

During water flooding of viscous oil reservoirs, adverse mobility ratio leads to an unstable displacement and thus viscous fingering. Previous research in viscous fingering has focused on low flow rates and high permeability systems (above 1 Darcy). This paper provides a more thorough and systematic study of the factors affecting viscous fingering including lower permeability, different wettability and high flow rate. Homogeneous cores with permeability ranging from 20 md to 6 Darcy are selected for unstable coreflood experiments. For water-wet systems, water is used to displace viscous mineral oil of different viscosity. For weakly oil-wet (mixed-wet) systems, the core is aged in a crude oil and water displaces a viscous oil. To study strongly oil-wet system, the fluid phases are switched. A light hydrocarbon is used to displace a viscous water in a water-wet core, which is invasion of a strongly non-wetting fluid (similar to water invasion in an oil-wet rock). The effects of permeability, wettability and flow rate on fingering and oil recovery are studied. Unstable displacements are also conducted in micromodels to allow easy visualization of pore-scale mechanisms.
For water-wet systems, decreasing permeability leads to more intense fingering and lower recovery. Increasing flow rate leads to consistently lower recovery due to less time for the imbibition of water into oil-filled pores. Nvf = Nc (μr^2) (D^2)/K can be used to correlate recovery in the presence of viscous finger for water-wet systems. Weakly oil-wet (mixed-wet) systems show distinctly different flow patterns than strongly oil-wet and water-wet systems. For strongly oil-wet systems, permeability does not have a significant effect on fingering and recovery. At low flow rates, increasing flow rate leads to a higher recovery because the higher pressure in the finger can overcome the capillary pressure and invade oil-filled pores. However, further increase in flow rate enhances fingering and results in lower recovery. There exists an optimum flow rate to yield the highest recovery in strongly oil-wet systems. Nvs=[Nc^(-1)] μ_r^2 D^2 can be used to correlate recovery in the presence of viscous finger for strongly oil-wet system. Viscous fingering is more intense in oil-wet systems than in water-wet systems. Water film flow in water-wet systems damps viscous fingering, but there is little oil film flow in strongly oil-wet systems due to the high oil viscosity.

Speaker: Mr. Bochao Zhao (The University of Texas at Austin)
• 14:55
Pore-scale Analysis of Nanoparticle Flow Using Lattice Boltzmann Method: Effect of Nanoparticle Properties on Retention 15m

With the advent of nanotechnology many researchers attempted to adopt nanoparticles to improve fluid flow behavior in porous media. Despite potential abilities of nanoparticles to diffuse in small pores, in most of the cases, the injection of nanofluid in porous media is either unsuccessful or damaging to the system. Hence, the behavior of the nanoparticles at pore scale should be carefully securitized and a specific nanoparticle should be designed for the porous medium of interest to guarantee the best performance. This study examines nanoparticles flow in a designed pore-scale porous medium and considers the effect of particle and rock surface properties on nanoparticle retention. The prediction of nanoparticles behavior in such a scale entails considering the interactions and surface properties of the particles and porous media while avoiding expensive computational cost of molecular dynamics approach and approximations of macroscopic approach. To this end, a mesoscopic approach, Lattice Boltzmann Method, has been adopted to simulate nanoparticles flow in a pore-scale porous medium. The behavior of nanoparticles has been studied considering deposition and adsorption of the nanoparticles in the system. It is shown that the size and surface energy of nanoparticle can affect the hydrodynamic behavior of the injecting fluid. The impact of temperature rise on the adsorption and deposition also shows that the kinetic vibration of the molecules overcomes attraction energies between nanoparticles and rock surface and consequently enhances stability of nanoparticles.

Speaker: Mr. Mohammad Zargartalebi (Department of Chemical and Petroleum Engineering, Schulich School of Engineering, University of Calgary)
• 15:13
Viscous fingering and nonlinear waves in a Langmuir adsorbed solute 15m

Waves in chromatography are well known to the practitioners. Characteristics of these waves are strongly correlated to the nature of the adsorption isotherm. We model the displacement of a finite slice containing an adsorbed solute by a carrier liquid flow. We examine the nonlinear dynamics that emerge from the interactions of rarefaction, shock layer and/or viscous fingers (VF) in the finite solute slice that adsorbs on a porous matrix according to a Langmuir isotherm. The differences between a linear and a Langmuir isotherm are discussed. In the absence of VF, a shock layer (rarefaction) wave appears at the frontal (rear) interface of the solute. VF at a viscously unstable rarefaction interface propagates through the finite sample to preempt the shock layer. However, no such incident is observed when the shock layer front features VF. Various quantities (e.g., the onset of VF, shock layer thickness, etc.) are calculated as a function of the saturation rate and the results a
re supported by mathematical analysis.

Speaker: Chinar Rana
• 15:40
Non-modal growth of perturbations in miscible displacements with non-monotonic viscosity pro files 15m

We study the effect of a non-monotonic viscosity profile on miscible viscous fingering in porous media. This hydrodynamic instability is studied by coupling the continuity and Darcy equations with a convection-diffusion equation for solute concentration that determines the viscosity. A toy viscosity model composed of a sequence of transformation in a sine function is considered. Parametric studies are performed in terms of the end-point viscosity contrast, maximum viscosity, and the corresponding value of the concentration. We employ a non-modal analysis (NMA) based on the singular value decomposition of the propagator matrix approach to perform the stability analysis of the non-autonomous linear system. NMA facilitates to identify the optimal amplification of the perturbations and its spatial structure. We demonstrate that there is a disagreement of previous linear stability analyses and NMA. This disagreement is inherited from the perturbation structure and the parameters involved in defining the non-monotonic viscosity profiles. Our study shows that miscible displacements with non-monotonic viscosity profiles can successfully be analyzed using NMA and paves the way for future work to understand the displacements with a non-monotonic viscosity in miscible reactive flows.

Speaker: Dr. Tapan Kumar Hota (Indian Institute of Technology Ropar, India)
• 15:58
The effect of pore scale disorder on unstable multiphase flow at the pore scale. 15m

The flow of multiple immiscible fluids at the pore scale is sensitive to local porosity fluctuations that can be measured as pore scale disorder. Our high fidelity direct numerical simulations of pore scale multiphase flow indicate that the degree of disorder governs the expression of viscous instability at the pore scale. Instability is suppressed when porosity is highly ordered even for large viscosity contrasts. While instability appears for even minute deviations away from the ordered state, it saturates beyond a certain threshold that depends on the viscosity ratio. The importance of the length scale of flow paths associated with such behaviors will be reported. Implications for permeability measurement and the construction of REV scale under time varying flow will also be examined.

Speaker: Prof. Amir Riaz
• 16:16
Thermo-Viscous Fingering of Nanoflow Displacements in Homogeneous Porous Media 15m

The interface of two approaching fluids in porous media is unstable if a less viscous fluid displaces a high viscous one. These instabilities known as viscous fingering (VF) or Saffman-Taylor (ST) instabilities are favourable when a high mixing of the fluids is desired e.g. micro-mixers while they are unfavourable when the aim is to sweep the highest amount of the displaced fluid e.g. oil industry.
In a homogeneous porous medium with constant permeability, viscosity controls the instabilities. Accordingly, the presence of nanoparticles and heat transfer may change the viscosity distribution in the system and as a result the growing of VF instabilities. The aim of this study is to investigate the role of any parameters which can affect the creation and control of VF instabilities after addition of nanoparticles to the system under non-isothermal conditions. The analysis is important due to the considerable application of nanofluids in porous media, e.g. oil recovery and ground water remediation, the critical role of nanoparticles in heat transfer augmentation and finally the lack of studies about their coupled effect on controlling VF instabilities.
The governing equations include the conservation of mass, conservation of momentum in the form of Darcy’s law, two convection-diffusion equations representing the transport of the fluids and nanoparticles and energy equation. The equations are solved numerically using the Hartley-based pseudo-spectral method. Using this method, the equations are transferred to the Hartley space where the special derivatives are eliminated from the equations and accordingly the errors encountered with the numerical derivatives will be reduced. Furthermore, the governing equations are changed to a system of ODEs and algebraic equations in the Hartley space, and as a result are easier to solve with a high degree of accuracy.
The results are used to analyse the effects of Brownian diffusion, thermophoresis, nanoparticle deposition and the thermal lag coefficient on the instabilities. The analysis is conducted both qualitatively, by the shape and number of fingers and quantitatively, by the mixing length, the ratio of the length within a specific concentration range to the length of the whole domain, breakthrough time, the time for fingers to reach the end of the domain, and contact line length, the length of the iso-surface of a specific concentration.

Speaker: Mr. Behnam Dastvareh (Department of Chemical and Petroleum Engineering, Schulich School of Engineering, University of Calgary)
• 16:34
Salinity Effects During Two-Phase Flow in Porous Media: Electrokinetic Control of Viscous Fingering 15m

Pattern formation is ubiquitous in many physical or chemical processes and has been at the center of attention for the past couple of decades. In many instances, interfacial instabilities play a central role in creating these patterns and controlling their spatiotemporal evolution. Perhaps one of the most well-known examples is the striking figures generated when a high-viscosity fluid is displaced by a low-viscosity fluid. This “viscous fingering” phenomenon was originally described by Saffman and Taylor in the context of Hele-Shaw flows but is also observed in porous media flows, where it leads to residual trapping during secondary or tertiary oil recovery and reduces extraction efficiency.

In the classical theory, the onset of instability is only controlled by a single parameter, i.e. the viscosity ratio. However, coupling with other physiochemical processes could enhance or suppress viscous fingering. For instance, many rock formations contain chemically active surface groups that dissociate in the presence of water and lead to the formation of surface charge. These surface charges interact electro-statically with mobile ions in the solution and can affect the flow behavior through the so-called electrokinetic coupling.

Here, we use linear stability theory as well as nonlinear numerical simulations to study the role of electrokinetic coupling and salinity effects on the interface stability. Our results indicate that viscous fingering may be controlled, and even suppressed, by applying external electric fields. Furthermore, even in the absence of electric fields, strong electrokinetic coupling (present in nanopores where the electric double layers overlap) can reduce viscous fingering by enhancing the `effective viscosity'' of the injected fluid through the electro-viscous effect.

Our findings can have implications for electrically enhanced oil recovery application, as well as low salinity water flooding, and might help with understanding other similar multi-field driven interfacial instabilities.

• 14:37 17:37
Parallel 5-B
• 14:37
Capillary imbibition in wood governed by water adsorption in walls 15m

Water transfers through wood structure play a major role in wood behavior under various conditions such as drying, imbibition, sapflow, for which various problems may occur, such as shrinkage, cavitation, fracture, swelling. The physical understanding of these transfers is in some cases rather poor, as illustrated by the fact that it is still sometimes considered that the permeability of wood varies with the sample length, which would mean that it is not an intrinsic material property.
Here we focus on imbibition properties of wood (here hardwood) along the longitudinal direction. We first show that the dynamics of water penetration in hardwood exhibits a contradiction when it is considered within the standard frame of capillary imbibition: on one side we have a very slow dynamics a priori associated with an extremely poor wetting, on the other side the water is finally able to climb at a high level against gravity as for good wetting. This contradiction is confirmed by 3D Synchrotron images of the internal structure obtained during imbibition, which show that the liquid-air interfaces in the capillary vessels remain planar, which implies negligible Laplace pressure, but significantly advance along the vessels, again unexpectedly.
From further examination of the dynamics of water penetration and wood microstructure evolution from Synchrotron images and MRI measurements allowing to distinguish bound and free water, we show that this contradiction is explained by the adsorption of a slight amount of bound water in the cell walls, and at the origin of wood swelling. This adsorption governs the process: it momentarily damps wetting then allows further advance when the walls are saturated with bound water. Finally we definitely prove this mechanism with the help of experiments with model materials, i.e. hydrogels, from which both the ascent of free water and the adsorption and propagation of bound water may be directly observed.

Speaker: Mr. Philippe COUSSOT (univ. Paris-Est)
• 14:55
TUNING PORE MORPHOLOGY AND FUNCTIONALITY IN SOFT POROUS MATERIALS 15m

Over the last decade, the generation of organic porous (nano)materials with tunable pore sizes and desired functionalities has been the subject of increasing attention in materials science. Interest in such porous frameworks originates from the large variety of applications in which they are involved, e.g. size/shape-selective nanoreactors, monoliths for advanced chromatographic techniques, nanofiltration membranes, high specific area catalytic supports, as well as 3-D scaffolds for tissue engineering.
This lecture examines the scope and limitations of three different approaches to porous polymers with controlled porosity and functionality at different length scales. The first approach relies on the synthesis of polystyrene-block-poly(D,L-lactide) diblock copolymers with functional groups at the junction between both blocks, followed by their macroscopic orientation, and the subsequent selective removal of the polyester block to afford ordered nanoporous materials with channels lined with chemically accessible functionalities (e.g., COOH, SO3H, SH, COH) [1-4].
The second strategy entails the preparation of biocompatible doubly porous crosslinked polymer materials through the use two distinct types of porogen templates, namely a macroporogen in combination with a nanoporogen. To generate the macroporosity, either NaCl particles or fused PMMA beads are used, while the second porosity is obtained by using a porogenic solvent [5]. Alternatively, a straightforward and versatile methodology for engineering doubly porous polymers is implemented through a thermally induced phase separation process [6].
Finally, 3-D macroporous scaffolds based on biodegradable polyesters have been engineered by electrospinning to generate nanofibrous biomaterials that mimic the extracellular matrix [7,8]. The potentialities afforded by these approaches will be addressed, and some typical applications of the resulting porous materials will be illustrated.

Speaker: Daniel Grande (Institut de Chimie et des Matériaux Paris-Est)
• 15:13
Origin of sorption hysteresis of micro-porous polymers: an explanation based on hydrogen bonds 15m

Soft nanoporous matter encompasses man-made materials such as compliant porous solids, intrinsically porous polymers and organic membranes as well as natural materials such as wood, bamboo, cotton and other plant-derived materials. These materials can undergo significant deformation during water adsorption because of strong coupling between the adsorption and mechanical properties. In addition, hysteresis is observed in the sorption isotherms of these materials. We find that sorption-induced deformation is the cause of hysteresis and analyze the special role played by the different types of hydrogen bonds.
We prepare three samples of amorphous cellulose as the host material considering its wide presence in nature. Water sorption and deformation are simulated with a hybrid Grand Canonical ensemble (GCMC)/ Molecular Dynamics (MD) method. The sorption isotherms simulated for each of the three samples exhibit significant hysteresis and the simulation results agree well with experimental data on low-crystalline cellulose. Significant hysteresis is observed as also seen in a variety of micro-porous polymers. The volumetric strain is monitored and shows significant swelling strains reaching as high as 36% at the saturation point. We note that, if we eliminate the deforming effect by freezing the cellulose at the dry state or the saturated state and then conduct an additional adsorption or desorption loop, the simulated isotherms significantly deviate from the experiments and the hysteresis disappears. This shows that the hysteretic sorption strongly depends on the deformation of the system.
We explore the mechanism of sorption hysteresis by further interrogating the systems. As amorphous cellulose is a non-cross-linked polymer, hydrogen bonds play a crucial role in both sorption and deformation processes. Hydrogen bonds are analyzed to monitor the interactions inside the system, revealing different configurations during adsorption and desorption. We find that the origin of sorption hysteresis in amorphous cellulose can be attributed to the process of creation of new adsorption sites due to hydrogen bond breaking during swelling. During adsorption process, some cellulose-to-cellulose hydrogen bonds (HBCC) change into cellulose-to-water hydrogen bonds (HBCW). However, the recovery of these new adsorption sites back to HBCC during desorption is less likely since they more likely remain occupied by water molecules. Consequently, the system accommodates the same amount of water molecules distributed over more sorption sites at lower energy state during desorption process, leading to hysteresis in the sorption isotherms observed in both experiments and simulations. Moreover, sorption hysteresis is also shown to have a significant influence on other important properties such as pore size distribution and bulk modulus.

Speaker: Mingyang Chen (ETH Zurich)
• 15:40
First and second order transition during water adsorption in hemicellulose and its consequence on hygro-mechanical behavior 15m

Arabinoglucuronoxylan (AGX) is one of the most abundant hemicellulose of softwood. It is formed by β-1,4-linked β-D-Xylp units, partially substituted at O-2 by 4-O-Me-α-D-GlcpA and at O-3 by α-L-Araf. It has diverse potential use in many industries, such as packaging of cosmetic or pharmacy, plastic additive and bio-refinery.
Despite its versatile potential, most of the key properties of xylan remain mysterious as it is difficult to isolate this polymer from wood sources without chemically altering it. Hemicellulose interacts strongly with moisture and induces many physical or mechanical changes. In this study, we apply molecular dynamics to investigate the mechanism of moisture influence on AGX.
Moisture induces various changes to AGX, such as isotropic swelling, enhancing diffusion, shifting pore size distribution, decreasing stiffness, etc. Besides the first order changes of properties upon moisture adsorption, there also exists a second order transition which means that properties considered as function of moisture develop two regimes separated by a transition zone. Water population distribution analysis shows that multi-layer adsorption is the mechanism of the second order transition. Moreover, we show that the main polymer of wood, e.g. amorphous cellulose (AC), does not show such second order transition as AC is more water miscible compared to AGX.

Speaker: Chi Zhang (ETH Zurich)
• 15:58
Kerogen flexibility, a key to understand hydrocarbon expulsion from shale reservoirs? 15m

A common assumption in recent molecular simulation based investigations of hydrocarbon transport in kerogen is that the latter behaves as a rigid matrix [1-4]. In other words, its porosity remains relatively constant and irrespective of temperature, lithostatic (or external) pressure or fluid (or adsorption) pressure. This implies that the matrix isolates the fluid from external pressure effects and that, diffusion and transport are only affected by the free pore volume, a lower amount of adsorbed fluid implies a higher amount of free volume and thus a higher diffusion coefficient [3]. However, it was shown that, immature kerogen, in particular, can retain a high (soft) aliphatic content and present low stiffness [5], typical of compressible media. Here we use an amorphous hydrogenated carbon model with high aliphatic content as a proxy for immature kerogen and characterize its poroelastic behavior in large ranges of temperature, lithostatic pressure and fluid pressure and determine the sorption isotherms of methane accounting for swelling. Results show that the pore space depends considerably on the three parameters and that the matrix behaves as an ideal adsorbent (linear increase of sorbate amount with the fluid pressure) up to very high pressures due to considerable swelling. An important consequence of swelling is that instead of decreasing with loading as is the case in rigid frameworks, the self diffusion coefficient of methane actually increases with loading. Furthermore, simulation of methane desorption at kerogen/macropore interfaces reproduces the well-known productivity decline of shale-oil plants when matrix flexibility is accounted for while the conventional fickian regime is obtained in the rigid approximation. We associate this anomalous transport phenomenon to a macroscopic deformation of the matrix taking place during fluid desorption.

[1] Collell J., Galliero G., Vermorel R., Ungerer P., Yiannourakou M., Montel F., Pujol M., J. Phys. Chem. C, 119 (2015) 22587.
[2] Falk K., Coasne B., Pellenq R., Ulm F.-J., Bocquet L., Nat. Commun. 6 (2015) 6949.
[3] Obliger A., Pellenq R. J.-M., Ulm F.-J., Coasne B., J. Phys. Chem. Lett. 7 (2016) 3712.
[4] Ho T. A., Criscenti L. J., Wang Y., Sci. Rep. 6 (2016) 28053.
[5] Bousige C., Ghimbeu C., Vix-Guterl C., Pomerantz A. E., Suleimenova A.,Vaughan G., Garbarino G., Feygenson M., Wildgruber C., Ulm F.-J., Pellenq R. J.-M., Coasne B., Nat. Mater. 15 (2016) 576.

Speaker: Dr. Pierre-Louis Valdenaire (Massachusetts Institute of Technology, MSE2 MIT-CNRS International Joint Unit)
• 16:16
Hydrogen storage, adsorption induced deformation and the role of confinement dimensionality in CAU metal organic frameworks 15m

We have studied mechanism of hydrogen storage in CAU-1 and CAU-8 metal organic frameworks synthesized at Christian-Albrechts-University of Kiel, Germany 1, [2]. The structures of these compounds are built from aluminum polyhedra with fully coordinated metal ions. Using various types of organic linker molecules tunable structures with channels of different geometries, but very similar chemical composition can be formed. Thus, CAU-1 consists of a three-dimensional confinement network with cavities of 10 Å and 5 Å of average cross section, while the structure of CAU-8 contains one-dimensional channels of about 8 Å diameter.

Speaker: Dr. Margarita Russina (Helmholtz Zentrum Berlin )
• 16:34
In-situ control of soft adsorbents pore size for optimal separation properties 15m

Gas separation processes involving adsorption may have advantages over other separation methods although one drawback still remains the need for increased selectivity. A simple way to improve the selectivity of a separation process is to increase the pore confinement in the adsorbent. For instance, this can be achieved by the inclusion of large cations in the case of zeolites or by controlling the activation conditions of carbons. With flexible Metal-Organic Frameworks (MOFs), one can modify the chemistry to control the pore size or shape. These methods allow a primary control of the pore structure from the synthesis step. However, the question arises as whether it is possible to control the pore size or shape of existing porous adsorbents during the adsorption process via utilization of an external stimulus.
In this fundamental study, we have applied an external mechanical pressure to compress a flexible Metal-Organic Framework, MIL‐53. Indeed, some MOFs are considered as soft porous crystals which can change reversibly their structure when they are exposed to stimuli such as adsorbed molecules1, pressure2 or temperature. This structural flexibility, leading to a change in the pore size, strongly influences the selectivity of the adsorbent for some gas mixtures. For example, the MIL‐53 solid in its narrow pore form exhibits a very good CO2/CH4 selectivity, which is lost when the structure switches to the large pore form with the increase of gas pressure3. By maintaining the narrow pore form via mechanical compression, it may be possible to control its selectivity over a wider range of pressure or even induce molecular sieving. Releasing the mechanical pressure will lead to the large pore form which could be easier to regenerate.
For this purpose, we have developed a novel methodology to tune the adsorption behavior of mechanically responsive materials by coupling the effects of ‘internal’ gas adsorption pressure and ‘external’ mechanical pressure. In order to pilot the structural flexibility of the adsorbent during the gas adsorption, we built an experimental device to apply a mechanical pressure (up to 25 tons) on the porous material via a uniaxial press system which equally allows gas adsorption up to 15 bars.
Results showing proof of concept with the MIL‐53 will be given along with some openings towards more applied subjects.

Speaker: Nicolas Chanut (Massachusetts Institute of Technology)
• 17:15 18:45
Poster 2: Poster 2-A
• 17:15
Study of Gas Production from Shale Reservoirs with Multi-stage Hydraulic Fracturing Horizontal Well considering Multiple Transport Mechanisms 15m

Development of unconventional shale gas reservoirs (SGRs) has been boosted by the advancements in two key technologies: horizontal drilling and multi-stage hydraulic fracturing. A large number of multi-stage fractured horizontal wells (MsFHW) have been drilled to enhance reservoir production performance. Gas flow in SGRs is a multi-mechanism process, including: desorption, diffusion, and non-Darcy flow. The productivity of the SGRs with MsFHW is influenced by both reservoir condition and hydraulic fracture properties. However, rare simulation work have been conducted for multi-stage hydraulic fractured SGRs. Most of them use well testing method, which have too many unrealistic simplifications and assumptions. Also, no systematical work has been conducted considering all reasonable transport mechanisms. And there are very few work on sensitivity studies of uncertain parameters using real parameter range. Hence, a detailed and systematic study of reservoir simulation with MsFHW is still necessary.
In this paper, a dual porosity model was constructed to estimate the effect of parameters on shale gas production with MsFHW. The simulation model was verified with the available field data from the Barnett Shale. Following mechanisms have been considered in this model: viscous flow, slip flow, Knudsen diffusion, and gas desorption. Langmuir isotherm was used to simulate the gas desorption process. Sensitivity analysis on production performance of tight shale reservoirs with MsFHW have been conducted. Parameters influencing shale gas production were classified into two categories: reservoir parameters including matrix permeability, matrix porosity; and hydraulic fracture parameters including hydraulic fracture spacing, fracture half-length. Typical ranges of matrix parameters have been reviewed. Sensitivity analysis have been conducted to analyze the effect of above factors on the production performance of SGRs. Through comparison, it can be found that hydraulic fracture parameters are more sensitive compared with reservoir parameters. And reservoirs parameters mainly affect the later production period. However, the hydraulic fracture parameters have significant effect on gas production from the early period. Result of this study can be used to improve the efficiency of history matching process. Also, it can contribute to the design and optimization of hydraulic fracture treatment design in unconventional tight reservoirs.

Speaker: Dr. Chaohua Guo (Dapartment of Petroleum Engineering)
• 17:30
Comparison of Model Approaches for Gas Transport in Compacted Bentonite: A Current Task in the International DECOVALEX Project 15m

The DECOVALEX project is an international research and model comparison collaboration for advancing the understanding and modeling of coupled thermo-hydro-mechanical (THM) processes in geological systems. Prediction of these coupled effects is an essential part of the safety assessment of geologic disposal systems for radioactive waste and spent nuclear fuel, but also for a range of other sub-surface engineering activities such as carbon dioxide subsurface storage, enhanced geothermal systems, and unconventional oil and gas production through hydraulic fracturing. DECOVALEX involves analysis and comparative modeling of state-of-the-art laboratory and field experiments, with various international research teams providing a wide range of perspectives and solutions to these complex problems.

One of the analysis and modeling tasks of the current DECOVALEX phase centers on the transport of gas through low-permeability clay-based materials such as bentonite or argillite rock. In a geological repository for radioactive waste, the corrosion of the ferrous materials, radioactive decay of the waste, radiolysis of organic materials and water, and the microbial breakdown of organic materials will produce gas, the most important of which (by volume) is hydrogen. As gas is produced, it will initially accumulate until gas pressure is eventually high enough to drive gas away from its source. Understanding the long-term fate of such gas transport and its impact on the surrounding materials is therefore important in the development of a geological disposal facility for radioactive waste.

Research teams involved in the modeling task use a range of different approaches to simulate gas transport data from a series of well-controlled laboratory tests, in a staged manner building in complexity (both in terms of the experimental and modelling approaches). Special attention is given to the THM mechanisms controlling factors for initiation and evolution of gas flow, such as gas entry, mechanical damage, pathway dilation and flow, as well as pathway stability and sealing, all of which will impact barrier performance. This presentation will provide an overview of process understanding and model approaches, and then will discuss in more details two specific numerical models to analyze gas-migration laboratory experiments on mechanically confined bentonite samples. The first model is a continuum model based on multiphase fluid flow, linear poro-elasticity and moisture swelling, with gas permeability related to the minimum effective compressive stress. The second model is a discrete fracture model based on continuum multiphase fluid flow linked with a lattice model to represent discrete fracture developments. In the continuum model, the key to capture observed responses is the stress-dependent gas permeability and moisture shrinkage. In the discontinuum model, the key processes involved are shear failure creating a dense discrete fracture network and abrupt permeability enhancement. Both models capture reasonable well the measured evolution of gas flow, pressure and stress, including the abrupt responses observed once an apparent threshold gas pressure is exceeded. Further modeling of other experiments conducted on the same type of bentonite (e.g. spherical flow versus linear flow) will be necessary to better distinguish between different underlying and controlling processes occurring within the sample.

Speaker: Dr. Jens Birkholzer (Lawrence Berkeley National Laboratory)
• 17:45
Interfacial Impacts on Slickwater Imbibition and Gas Production in the Marcellus Shale 15m

Development in the dry unconventional gas-bearing Marcellus Shale in the Eastern United States has grown rapidly over the past decade. When a well is fractured in the Marcellus, only a small proportion of the slickwater fracturing fluid, typically <10%, is produced back following well completion. Most analyses also suggest that existing fracking and production practices only produce a fraction, typically <25%, of the Original Gas in Place (OGIP) over the life of the well. The connection between slickwater fate and gas production is poorly understood but it is generally assumed to involve trapping of imbibed water due to high capillary pressures, which impacts mobility of natural gas.

This work seeks to understand the connection between slickwater fluid properties, shale minerology and pore structure, and gas migration through fractured shales. Experiments were carried out to measure the fluid contact angles of Marcellus Shale mineral surfaces, representative fracturing fluids, and high pressure methane and the interfacial tension between fluids. These interfacial data were then integrated into a modeling framework developed using the EOS7C-ECBM equation of state modules within the TOUGH2 code, which includes the effect of non-Darcy flow regimes and sorption.

Our experimental results suggest that pre-wetted shale surfaces have a significantly lower static contact angle in the methane-water-shale system. Advancing contact angle on dry shale and drainage with receding angle on wet shale resulted in a difference of 70 degrees, illustrating the potential contribution of interfacial properties to relative permeability (Kr) and capillary pressure (Pc) hysteresis during imbibition and drainage in shale systems. These effects were correlated to specific Total Organic Carbon (TOC) content in the mineral. The experimental data were used to create synthetic inputs for using in the TOUGH2 code and the model was run to simulate flow behavior through the shale matrix, which was then used to represent different well configurations. Historical data of slickwater use and gas production from representative wells in the Marcellus region were used to benchmark the modeling output. The results suggest that there are key ways in which slickwater chemistry might be manipulated to increase the ultimate recovery of natural gas.

Speaker: Andres Clarens (University of Virginia)
• 18:00
Multi-component diffusion in a coupled free-flow porous-medium system 15m

A standard approach to model diffusion in porous media is the assumption of the validity of Fick‘s Law. Although widely used, that description can only be employed for binary mixtures or low concentrations of the components as it neglects molecular interactions of the different species.
When looking e.g. at gas migration of an organic component in soil where higher concentrations of components can occur, more complex laws need to be employed. In this work we present a multi-phase, multi-component model incorporating the Maxwell-Stefan‘s approach to diffusion which takes into account all interactions between the molecules of different species. Therefore it is possible that the diffusion behaviour is very different to the one seen in the standard continuum advection-diffusion description.
We present a coupled free-flow porous-medium model, where the Maxwell-Stefan diffusion approach is employed in both domains. Consistent coupling conditions for coupling the free-flow domain to the porous medium are presented as well.
The model is implemented in the numerical software framework DuMuX and can be used for various applications. The example application presented is a study on evaporation together with gas migration in the porous medium and across the porous-medium free-flow interface. Different concentrations of the gas components and their influence on evaporation rates and gas migration rates across the porous-medium free-flow interface are compared.

Speaker: Katharina Heck (University Stuttgart)
• 17:15 18:45
Poster 2: Poster 2-B
• 17:15
Mathematical modeling of BTEX concentrations on the unsaturated zone using a simple finite differences model: evaluation of the mass distribution between phases 15m

The impact of fuel spills on the unsaturated zone are one of the main environmental issues when licensing new fuel stations or industrial facilities where Underground Storage Tanks (UST) are used. The development and use of fate and transport models of organic pollutants (BTEX) on the vadose zone can therefore be used to understand the behavior of these pollutants under different scenarios.

This paper describes the results obtained when using a simple one-dimensional finite different vadose zone leaching model that describes the movement of organic contaminants within and between three different phases: (1) as a solute dissolved in water, (2) as a gas in the vapor phase, and (3) as an absorbed compound in the soil phase. The model uses a numerical approximation of the Millington Equation, a theoretical based model for gaseous diffusion in porous media. This equation has been widely used in the ﬁeld of soil physics and hydrology to calculate the gaseous or vapor diffusion in porous media.

Initially, the equilibrium distribution of contaminant mass between liquid, gas and sorbed phases is calculated. Transport processes are then simulated. Liquid advective transport is calculated based on values deﬁned by the user for inﬁltration and soil water content. The contaminant in the vapor phase migrates into or out of adjacent cells based on the calculated concentration gradients that exist between adjacent cells. After the mass is exchanged between the cells, the total mass in each cell is recalculated and re-equilibrated between the different phases. At the end of the simulation, (1) an overall area-weighted groundwater impact for the entire modeled area and (2) the concentration proﬁle of BTEX on the vadose zone are calculated.

The distribution of total mass of pollutants between the three phases is shown. A sensitivity analysis of the model parameters to a set of soil contamination scenarios caused by a set of BTEX spills from synthetic underground storage tanks is presented. Results demonstrate the applicability of simple numerical models for the environmental analysis of new industrial sites where soil contamination may be caused by organic pollutants.

Speaker: Dr. JAVIER RODRIGO-ILARRI (UNIVERSITAT POLITECNICA VALENCIA)
• 17:30
DEM-CFD coupling for the simulation of filter cake formed due to poly-dispersed particles. 15m

In this study, we analyze the filter cake formed due to mono and bi-dispersed spherical particles. The particle-particle, particle-filter interactions are simulated using Discrete Element Method (DEM) and the fluid flow is simulated using Finite Volume Method (FVM). The computation of the two-way particle-fluid interaction is the challenging part in the numerical studies mainly due to the calculation of the drag force. The fluid drag force in this study is calculated by using the poly-dispersed drag model suggested by Beetstra (2005). The numerically predicted void fraction of the filter cake formed due to the mono and the bi-dispersed particles is compared with the sedimentation experiments in the literature. We then analyze the effect of factors affecting the filter cake formation i.e. the particle-particle interaction parameters (the coefficients of the sliding and the rolling friction, the surface energy) and the poly-dispersity (the particle mass fraction ratio).
We observed that, in general with the increase of the coefficient of sliding and rolling the predicted void fraction increases. Further neglecting the attractive forces between the particles under predicts the void fraction. At sufficiently higher particle Reynolds number the filter cake formed due to the mono and the bi-dispersed particles undergoes consolidation. At a constant Reynolds number pressure drop across a filter cake formed due to bi-dispersed particles increase with the increase in the mass fraction of the bigger particles.

Speaker: Ruturaj Deshpande (Fraunhofer ITWM)
• 17:45
Field-scale modelling of nanoparticle injection and transport for nanoremediation design and particle fate assessment 15m

The design of a field-scale injection of engineered nanoparticle (NP) suspensions for the remediation of a polluted site requires the development of quantitative predictive models for the system design and implementation.
In general, micro- and nanoparticle transport in porous media is controlled by particle-particle and particle-porous media interactions, which are in turn affected by flow velocity and pore water chemistry. During the injection, a strong perturbation of the flow field is induced around the well, and the particle transport is mainly controlled by the consequent sharp variation of pore-water velocity, and by the hydro-chemical properties of the injected fluid. Conversely, when the injection is stopped, the particles are transported solely due to the natural flow, and the influence of groundwater geochemistry (ionic strength, IS, in particular) on the particle behaviour becomes predominant. Pore-water velocity and IS are therefore important parameters influencing particle transport in groundwater, and have to be taken into account by the numerical codes used to support nanoremediation design.
Several analytical and numerical tools have been developed in recent years to model the transport of colloidal particles in simplified geometry and boundary conditions. For instance, the numerical tool MNMs was developed by the authors of this work to simulate colloidal transport in 1D Cartesian and radial coordinates. Only few simulation tools are instead available for 3D colloid transport, and none of them implements direct correlations accounting for variations of groundwater IS and flow velocity.
In this work a new modelling tool, MNM3D (Micro and Nanoparticles transport Model in 3D geometries), is proposed for the simulation of injection and transport of nanoparticle suspensions in generic complex scenarios. MNM3D implements a new formulation to account for the simultaneous dependency of the attachment and detachment kinetic coefficients on groundwater IS and velocity. The software was developed in the framework of the FP7 European research project NanoRem and can be used to predict the NP mobility at different stages of a nanoremediation application, both in the planning and design stages (i.e. support the design of the injection plan), and later to predict the long-term particle mobility after injection (i.e. support the prediction of final fate of the injected particles). The application of the model in the framework of a novel approach for risk assessment at particle-contaminated sites is also reported.

Speaker: Tiziana Tosco (DIATI, Politecnico di Torino)
• 18:00
Krypton Adsorption Characteristics of Activated Charcoal at Ambient and Cryogenic Temperatures 15m

The adsorption characteristics of krypton on activated charcoal with nitrogen and helium carrier gases were studied at various temperatures using a bench-top gas test system. The adsorption characteristics were quantified using a height of mass transfer unit adsorption model. The adsorption model is a function of the filter’s length, activated charcoal mass, and dynamic adsorption coefficient. The dynamic adsorption coefficient characterizes the retention of krypton as a function of the activated charcoal’s temperature. Previous literature values for the dynamic adsorption coefficient of krypton were recorded at temperatures as low as -140 C for helium carrier gas, -120 C for nitrogen carrier gas, and -40 C for hydrogen carrier gas. In this paper, measurements at temperatures as low as -150 C are described for nitrogen and helium carrier gases, as well as at temperatures in the literature range for validation. Two different concentrations of krypton were used as well as various flow rates and operating pressures. Pulse measurements of krypton in ultra-high purity nitrogen and helium were used to measure the dynamic adsorption coefficient, and continuous flow measurements of krypton were used to measure complete breakthrough to fit a value of the height of mass transfer unit. The change in krypton concentrations due to filter adsorption was measured using a quadrupole mass spectrometer.

Speaker: Glen Guzik
• 17:15 18:45
Poster 2: Poster 2-C
• 17:15
Benchmark Analytical Solutions to Advection-Dispersion in Discrete Fractures Coupled with Multirate Diffusion in Matrix Blocks of Varying Shapes and Sizes 15m

The current state-of-the-art modeling approaches for contaminant/heat transport in fractured rock include (1) discrete fracture-network (DFN) and discrete fracture-matrix (DFM) models with the fracture network and matrix blocks randomly generated, (2) numerical models based on conventional dual-continuum models, and (3) analytical models with simplified parallel fractures and slab-like matrix blocks. These models differ in the complexity of the fracture network and matrix blocks, modeling accuracy, and computational efficiency, making it difficult to compare their results through benchmark problems.

We developed several benchmark problems of contaminant/heat transport under different flow (e.g., linear, radial, and dipole) fields and provided corresponding analytical solutions to global advection-dispersion coupled with multirate diffusion in the rock matrix. The benchmark problems consist of (1) a fracture network of one, two, and three sets of orthogonal fractures with varying fracture spacing and (2) matrix blocks of various shapes (slabs, squares, rectangles, cubes, and rectangular parallelepipeds) and sizes bounded by the fracture network. The matrix blocks can be isotropic with the same fracture spacing in each direction and anisotropic with varying aspect ratios.

The multirate diffusion caused by different shapes and sizes of matrix blocks was accounted for by using a unified-form diffusive flux equation for 1D isotropic (spheres, cylinders, slabs) and 2D/3D rectangular matrix blocks (squares, cubes, rectangles, and rectangular parallelepipeds) in the entire dimensionless time domain (Zhou et al., 2017a, b). For each matrix block, this flux equation consists of the early-time solution up until a switch-over time after which the late-time solution is applied to create continuity from early to late time. The early-time solutions are based on three-term polynomial functions in terms of square root of dimensionless time, with the coefficients dependent on dimensionless area-to-volume ratio and aspect ratios for rectangular blocks. For the late-time solutions, one exponential term is needed for isotropic blocks, while a few additional exponential terms are needed for highly anisotropic blocks. These solutions form the kernel of multirate and multidimensional hydraulic, solute, and thermal diffusion in fractured reservoirs.

The transient flux equation for multirate diffusion was transformed to develop the analytical solutions to the benchmark problems in the Laplace domain with typical functions (e.g., Airy functions) for the global advection-dispersion equation. The benchmark solutions can bridge the gaps between the three modeling approaches with reasonable complexity of fracture network and matrix blocks, as well as high modeling accuracy and efficiency. They are very useful for benchmarking the DFN/DFM modeling, whose accuracy depends on how to capture the fracture-matrix diffusive transfer and diffusion within each matrix block.

Speaker: Quanlin ZHOU (Lawrence Berkeley National Laboratory)
• 17:30
Upscaling of mass transfer in field-scale discrete fracture networks using fractional-derivative models 15m

Mass transfer in field-scale discrete fracture networks (DFNs) is affected by the erratic internal structure and hydrogeological properties of the fractured media, which can result in non-Darcian flow due to channeling flow and non-Fickian transport due to matrix diffusion competing with fast displacement along fractures. This study explores flow and transport dynamics in various DFNs with a wide range of physical properties using the Monte Carlo simulation approach. The resultant mass transfer dynamics are then quantified by fractional-order derivative models built upon the promising fractional calculus. We will report results implying information transfer from non-Darcian flow to non-Fickian transport, and we will also try to explore the quantitative linkage between these two related processes. The goal is to develop efficient upscaling approaches using the spatiotemporally non-local fractional-derivative equations to characterize mass transfer in field-scale fractured media, without the need to map individual rock fractures.

Speakers: Ms. Bingqing Lu (University of Alabama) , Dr. Yong Zhang (University of Alabama) , Dr. Donald Reeves (Western Michigan University)
• 17:45
Direct inversion for joint parameter and boundary conditions estimation for fractured aquifer 15m

A new direct inverse method is developed that is capable of simultaneous estimation of hydraulic conductivity (K), boundary conditions, and flow field for both confined and unconfined aquifers. In this research, the direct inverse method is applied to the inversion of discrete fractured aquifers. By sampling synthetic aquifer problems with different fracture patterns, the inverse method is tested under varying measurement data quality, data density, and the ratio of fracture (Kf) to matrix hydraulic conductivity (Km). The method achieved stable K estimations under measurement errors up to +/-10% of the total hydraulic head variation of a given problem. The accuracy of K estimation, however, is sensitive to measurement density. Given sufficient and high quality measurements, inversion is successful for Kf/Km up to 10^6. Moreover, hydraulic heads, Darcy fluxes, streamlines, and boundary conditions are also satisfactorily recovered. For a set of test problems, direct inverse solutions were compared to those obtained with PEST and the importance of boundary conditions estimation is highlighted. Assuming that aquifer is homogeneous, directional equivalent K that can represent bulk flow in fractured aquifers is also successfully estimated.

Speaker: Prof. Ye Zhang (University of Wyoming)
• 18:00
Fast large-scale joint inversion for deep aquifer characterization using pressure and heat tracer measurements 15m

Characterization of geologic heterogeneity is crucial for reliable and cost-effective subsurface management operations, especially in problems that involve complex physics such as deep aquifer storage of carbon dioxide. With recent advances in computational power and sensor technology, large-scale aquifer characterization using various types of measurements has been a promising approach to achieve high-resolution subsurface images. However, large-scale inversion requires high, often prohibitive, computational costs associated with a number of large-scale coupled numerical simulation runs and large dense matrix multiplications. As a result, traditional inversion techniques have limited utility for problems that require fine discretization of large domains and a large number of measurements to capture small-scale heterogeneity, like CO2 monitoring in the subsurface.

In this work, we apply the Principal Component Geostatistical Approach (PCGA), an efficient inversion method, for large-scale aquifer characterization. The domain considered is a synthetic three dimensional deep saline aquifer intended for CO2 storage with 24,000 unknown permeability grid-blocks. Transient pressure and heat tracer measurements from multiple dipole pumping tests are simulated with the TOUGH2 simulator and are used to estimate the heterogeneous permeability field and the corresponding uncertainty. For this scenario, we investigate the worth of combining heat and pumping tracer data for characterization. We demonstrate that with the PCGA, the inversion can be performed at a reasonable computational cost, while also resolving the main features of the permeability field. This presents opportunities for using inverse modeling to improve monitoring design and data collection strategies in field applications.

Speaker: Jonghyun Lee (University of Hawaii at Manoa)
• 18:15
A comprehensive simulation model for solvent-aided thermal recovery of heavy oil and bitumen—Analyzing the impact of diverse factors on productivity and product selectivity 15m

A new simulation model for solvent-aided thermal recovery of heavy oil and bitumen has been developed. The simulation model describes non-isothermal, multiphase, and multicomponent reservoir systems involving multiple kinetic reactions of heavy oil cracking. In the development of numerical simulator, we include 12 fluid-and-solid components in four phases of 1) aqueous, 2) liquid organic, 3) gaseous, and 4) solid phases. The 12 fluid-and solid components are 1) water, 2) heavy oil, 3) light oil, 4) asphaltene, 5) methane, 6) ethane, 7) propane, 8) hydrogen, 9) carbon monoxide, 10) carbon dioxide, 11) hydrogen sulfide, and 12) coke. It describes relevant physical and chemical phenomena during in-situ heating and production, such as dynamic changes of rock-and-fluid properties as a function of system conditions, phase transition thermodynamics, heat transfer by conduction and convection, pore clogging by coke generation from reactions, and porosity and permeability alteration.
The application cases of numerical simulation are categorized into four as follows.
Firstly, we conduct local sensitivity analysis of productivity and product selectivity to diverse uncertain parameters. They include reaction parameters, formation water saturation, and rock permeability. The effect of each parameter has been quantified, and the most influential parameters to the hydrocarbon productivity have been figured out. Product selectivity, especially for unwanted gases of carbon dioxide and hydrogen sulfide is analyzed, as being affected by uncertain parameters. The delivered most influential parameters are the most important data to be measured, to reduce the prediction uncertainty of unwanted acid gases of carbon dioxide and hydrogen sulfide.
Secondly, we analyze the diverse heating methods. The heating methods include 1) electrical heating, 2) heating by hot water drive, 3) heating by hot water drive containing condensate, 4) electrical heating with hot water drive, and 5) electrical heating with hot water drive containing condensate. In each case, both physical and chemical changes of system are considered and compared, such as viscosity, density, and composition of the fluid phases; and optimized heating method for maximizing productivity has been figured out.
Thirdly, we also quantify the effect of heating temperature to the productivity and product selectivity. High heating temperature activates the reactions of heavy oil cracking, but also accelerates the generation of solid product, coke. Here, the effect of pore clogging by coke generation to the permeability alteration and subsequently altering fluid flow and heat convection is analyzed. Through a case study, we find an optimal heating temperature for maximizing productivity.
Fourthly, we conduct the simulation runs using diverse ratios between condensate and water in the cases of hot water drive containing condensate. It is found that the optimal ratio between condensate and water is affected by the wettability of the porous media and initial water saturation.
From the plentiful simulation cases, the optimization of in-situ heating and production in heavy oil and bitumen reservoirs has been realized. The developed simulator provides the powerful tool to investigate the impacts of various unknown parameters and controlling factors, and hence enables us to increase the success-likelihood of hydrocarbon production from thermally-cracked heavy oil/bitumen reservoirs.

Speaker: Dr. Kyung Jae Lee (University of Houston)
• 18:30
A multiscale method for poroelasticity problems in heterogeneous porous media. 15m

Keywords:

Generalized Multiscale Finite Element Methods, poroelasticity, heterogeneous porous media.

Abstract:(<500)

In this work, we develop a coarse-grid solution technique for poroelasticity problems in heterogeneous media. First, we will show the challenges associated with mechanics and flow problems in heterogeneous media. Because of coupling and disparate scales, the direct numerical solution is expensive and some type of coarse-grid models are needed. For the coarse-grid numerical solution, we develop and implement a Generalized Multiscale Finite Element Method (GMsFEM) that solves the underlying multi-physics problem on a coarse grid by constructing local multiscale basis functions. The procedure begins with the construction of multiscale basis functions for both displacement and pressure in each coarse block. We present numerical results, for a test problem with highly heterogeneous flow and mechanics properties and compute error between the multiscale solution with the fine-scale solutions. The ingredients of the multiscale method are presented. Some applications to stochastic problems will also be discussed.

Speaker: Aleksei Tyrylgin (Multiscale model reduction Laboratory, North-Eastern Federal University:)
• 18:30
Development of Embedded Discrete Multi-Fractures Model for Simulation of Fractured Reservoirs 15m

Accurate and efficient numerical simulation of fractured reservoirs is important and challenging. Conventional dual porosity and dual permeability(DP/DK) models are efficient but not accurate, especially when fracture-diagnostic tools make it easier to get the detail of the complex fracture networks. Discrete-fracture models(DFM) have been developed to use information of fracture networks, which is still limited for its computational inefficiency. Recently, Embedded Discrete Fracture Model (EDFM) became a promising study orientation to overcome such problems.
In this study, we improve the EDFM approach by embedding discrete multi-fractures instead of simple fracture pieces in the matrix domain. Multi-fractures here stand for parts divided from fracture network, each consisting of multiple fractures and their intersection. This model has a comparative advantage: the embedded fracture network can be divided into larger and more complex parts with arbitrary shapes and sizes when meshing grid. And each complex part can be considered as a whole.
The calculation of conductivity between NNCs (Non-neighboring Connections) is the core of EDFM. We expound that the key to calculate conductivity of NNC is to approximate a local pressure field. Thus a new calculation method based on a more reasonable local pressure distribution has been developed. This method is suitable for the multi-fractures model and more accurate than the one used in original EDFM especially when matrix grid is coarser.
We demonstrate the accuracy and efficiency of the new conductivity calculation method and the embedded discrete multi-fractures model by performing a series of case studies with CarstSim simulator and comparing the results with the original EDFM and fine-grid models. We also present two numerical case studies to demonstrate the applicability of our method in naturally fractured reservoirs.

Speaker: Mr. Renjie SHAO (Peking University)
• 18:30
Direction Dependency of Relative Permeability for Oil-Water Two Phase Flow in Vugular Porous Medium 15m

Multi-phase flow process in porous medium are generally simulated with in introduction of relative permeability, which is assumed to be a scalar function of phase saturation. Previous research have demonstrated this assumption might not be suitable for capillary force dominated heterogeneous porous medium. Similarly, in vugular porous medium, the free flow region in vugs would introduce vertical velocity component, causing the flow field to differ. Two-dimensional experiments are conducted to examine such effect. The physical model are artificial homogenous and isotropic porous medium square plates, and vugs are designated as cylindrical holes. A group of models with axisymmetric vug distribution are designed so that the fluid flow in two principal directions are the same if the plate is placed horizontally but different if the plate is placed vertically. Also for general purpose another group of models with random placed vugs are designed to simulate actual reservoirs. The upscaled relative permeability curves in two principal directions are measured and compared. Supplemental numerical simulation are also conducted on two simulators. The first is a finite volume simulator KarstSim with the assumption of gravity segregation and instantaneous establishment of steady-state in vugs, the second is a commercial finite element software. The experiment result indicates that the direction of upstream fluid flow have a significant influence on the flow field within the physical model, and is presented in the relative permeability curves. This effect is pronounced in the cases where the vugs are distributed in specific orders. The numerical result are in good agreement with the result of experiments. Based on the results, we prompt the need to introduce a secondary variable in the relative permeability function, which is the angle between the upstream flow direction and the vertical direction to better simulate the multi-phase flow process in vugular porous medium.

Speaker: Mr. Shihan SONG (Peking University)
• 18:30
Influence on Oil-water Flow Mechanism with Hydraulic Fracture Existed in Low-permeability Reservoir 15m

Hydraulic fracturing is one of the most effective treatment methods in development of low permeability reservoir which improve the conductivity of the formation such that the reservoir liquids seepage capacity is enhanced with flow friction reduced, which highly increase the withdrawal of underground liquid. However, with the existing of hydraulic fractures and due to the complication of their morphology, seepage of water and oil in the porous is getting complicate. In order to figure out the oil and water flow mechanism, physical and numerical simulation are designed to research the oil-water seepage law of low-permeability oil reservoir and the influences of fracture on reservoir development effect. Experiments of oil-water displacement are conducted and sample cores of three kinds of fracture morphology (no fracture, horizontal fracture and vertical fracture) and four permeability level (5,10,30,50×10-3μm2) are used from low permeability turbidite reservoir. Experimental results are discussed and compared with a coupled fracture and flow model. Variation among different kinds of fractures and levels of permeability are presented. Results show that with the presence of the artificial fracture, the threshold pressure gradient decreases, the oil relative permeability curve drops, the water relative permeability curve rises, the saturation of remaining oils increases, and the two-phase flow area of tested cores becomes narrower. The water flooding recovery of the core with vertical cracks is larger than that without fracture which improves the effect of water-flooding. While the flooding effect turns out to be poorer with the presence of horizontal fracture through the samples. Results can be used in numerical simulation of developing low permeability reservoirs.

Speaker: Mr. Mingjing Lu (China University of Petroleum(East China))
• 18:30
NMR study on multi-layer waterflooding of middle-east low permeability carbonate reservoirs 15m

Aiming the heterogeneity of low permeability carbonate reservoir in M group of H oilfield, carried out the physical simulation experiment of multi-layer waterflooding, and based on the T2 relaxation spectrum technique, studied the micro distribution of residual oil and production degree of different permeability combination. Results show that to the different pore structure types, NMR tests have different responses. The permeability is larger, the T2 relaxation spectrum is more, the peak value is higher, and the movable fluid is more. The difference of permeable grade has certain influence on the residual oil and recovery degree of multi-layer waterflooding, and the residual oil and recovery degree of rock samples with different permeability are different in absolute value and relative value. The absolute value of the absolute value and recovery degree of the residual oil are mainly distributed in the large pores, and the low porosity is very small, the smaller the relative value of the residual oil in the small pores, the larger the relative value of the recovery degree. The residual permeability of rock samples with the same permeability difference is higher than high permeability rock samples, and the exploitation potential of macro-pores is still large.

Speaker: Mr. Xingwang Shi (Institute of Porous Flow and Fluid Mechanics, Research Institute of Petroleum Exploration & Development)
• 18:30
Numerical Simulation of Shale gas reservoirs with embedded DFN model 15m

In this paper, we studied the numerical simulation of shale gas reservoir with both hydraulic fractures and natural fractures using the embedded discrete fracture system. First, the 3D DFN (discrete fracture network) model was built according to the real geological state. Then, transmissibilities between the embedded fracture grid and the matrix grid are calculated using two different methods. These two methods are adaptive to both Cartesian grids and corner point grids. Last, several real oilfield cases were studied in shale gas reservoirs with hydraulic fractures and natural fractures considering the Klinkenberg effect and gas adsorption effect. The result shows that using the method provided by this paper to solve the numerical simulation problems of shale gas reservoirs with fractures can simplify the calculation as well as ensure the accuracy.

Speaker: Mengyin Liang (China University of Petroleum (East China))
• 18:30
Study on water flooding seepage regularity of low permeability carbonate reservoir —Taking Middle East H oilfield as an example 15m

Abstract: Low permeability carbonate reservoir generally contain multiple types pore structure as matrix pore, hole and crack, which lead to significant difference in water flooding seepage regularity and great perplex in oilfield production. In order to relate pore structure to seepage regularity of core samples preferably, micro-CT scanning and scanning electron microscope advanced experimental technology were used to quantitatively describe pore structure parameters of carbonate reservoir, and divided reservoir types according to space proportion of different pore structure types. On this basis, same rock samples water flooding experiment was carried out to study the seepage characteristics regularity. Results show that this three methods are complementary in reservoir type division, and are able to meet the need of low permeability carbonate reservoir layer type in the Middle East by comprehensive analysis of the experimental data. The reservoir was divided into three types of pore structure including pore type, fracture pore type and fracture type, pore type and fracture pore type reservoir take up the large part, a few of crack type. Oil-water relative permeability curve is nearly X. Irreducible water saturation is 25% on average, and the average residual oil saturation is 32%. According to water flooding seepage characteristics, the anhydrous displacement efficiency is 15% on average, and the total displacement efficiency is 48% on average. The rise of water cut is relatively slow. The research have important guiding significance for understand and guide the production regularity of such reservoirs.
Key words: carbonate reservoir, X-CT, reservoir type, seepage regularity

Speaker: Mrs. yapu zhang
• 17:15 18:45
Poster 2: Poster 2-D
• 17:15
Single-scale heterogeneous pore network modelling with microporosity upscaling. 15m

One of the longstanding challenges of the oil and gas industry is the problem of scale and hence, the term “upscaling” is used frequently in literature. In this work, we investigate the ways to represent connected regions with substantially different pore sizes. For this purpose, pore-scale simulations are combined with conventional continuum scale models. Our primary objective is to run sensitivity analysis considering topology changes for the pore space above and below a given resolution. Inspired by conventional upscaling practices, we assign upscaled flow properties to a group of pores and throats that are smaller than a given resolution. We show if the back calculated upscaled properties are solely a function of the pore space topologies at smaller scale. To make our study more general, we also investigate different pore-to-pore connections at larger scales. First, we start by considering the simplest case where two macro-pores connected through cuboidal regular lattice network of micro-pores that represents microporous region. Calculating upscaled properties of the cuboidal lattice, we compare results from both conventional pore-scale and hybrid upscaled models. Consequently, we vary pore space connectivities within the cuboidal lattice as well as the connectivity with macro-pores to study a wide range of possible scenarios for these two different scales of pores. The results could provide insight to our understanding of multiphase flow in rocks with different scales of importance and upscaling large pore networks to speed up pore-scale simulations.

Speaker: Mr. Nijat Hakimov (University of Kansas)
• 17:30
Multi-scale analysis on coal permeability using the Lattice Boltzmann Method 15m

As a mesoscopic kinetic approach, the lattice Boltzmann method (LBM) has been widely applied to characterize the flows in porous medium. However, for larger scale flow, upscaling microscale technique is the key and difficult point. Based on microtomographic images of an actual coal sample, numerical simulations were carried out using the LBM at the pore scale. The velocity/pressure distributions and coal permeability were obtained. The local rate of mechanical dissipation is applied to determine an appropriate REV. Then based on the WBS-LBM, numerical simulations were carried at the REV scale and the pressure distribution is almost identical with the solution of the LBM simulations at the pore scale. The simulation results of REV with different sizes indicate that: The size of REV determined in this paper is reasonable. The relative error between pore scale and REV scale simulation for permeability prediction is less than 5%. In addition, the REV scale simulation can greatly improve the computational efficiency, and provides an effective approach for large-scale flow simulation.

Speaker: Dr. Yanlong Zhao (China University of Petroleum, Beijing,)
• 17:45
ROCK ABSOLUTE PERMEABILITY ANALYSIS USING IMAGE-BASED DIRECT PORE-SCALE SIMULATIONS 15m

The direct numerical simulations (DNS) experience of pore-scale flow is still relatively scarce and laborious due to the numerous practical challenges. They include typically huge model size and high computational expenses, some uncertainties in geometrical description related to resolution size and other factors remaining much in common for single and multiphase flow cases. Sometimes this makes challenging an unambiguous definition of rock permeability based on numerical simulations.

The advantage of DNS comprises the ability to model in detail dynamic physical fields interaction in “real” pore volume (and/or solid matrix) geometry. In our current work we address the incompressible steady single-phase flow in voxel-based geometry of 3D real rock image (more precisely, stack of images) with the main objectives to analyze and quantify the impact of image processing workflow on permeability computation.

Speakers: Clément Varloteaux (CHLOE) , Igor Bondino (Total) , Igor Bogdanov (laboratoire CHLOE, Université de Pau)
• 18:00
Microstructural characterization via Minkowski-functional-based global descriptors 15m

The densities of Minkowski functionals (volume, surface, mean curvature and total curvature) represent a complete set of independent global microstructural descriptors [1-3], as a consequence of Hadwiger’s characterization theorem [4]. Similar to correlation functions [5, 6] they offer a systematic and principally automatizable approach for the quantitative description of microstructures, but unlike the latter they form a complete set of simple parameters that are readily determined on real microstructures and, if appropriately implemented into microstructure-property relations, could provide more accurate predictions of effective properties than micromechanical bounds or model predictions based on volume fractions alone. In this contribution we present examples of the application of Minkowski-functional-based global descriptors for the quantitative description of porous ceramics. We show that, apart from the porosity (pore volume fraction) and mean chord length (based on the phase-specific surface density) also a generalized Jeffries size (based on the mean curvature integral density) can be determined from planar sections [1,7,8]. The correlation between these two independent size measures is analyzed and the average pore size thus determined is compared to the characteristic values (quantiles and mean values) extracted from pore size distributions (number- and volume-weighted) determined via microscopic image analysis, after correcting for the random section problem (Wicksell’s problem [9]) via appropriate transformation matrices [10]. Moreover, it is shown how the 3D Euler characteristic can be determined on (appropriately binarized) serial sections of spatial images (obtained by X-ray computed tomography).

Speaker: Willi Pabst (University of Chemistry and Technology, Prague)
• 17:15 18:45
Poster 2: Poster 2-E
• 17:15
A Model for Gas Transport in Inorganic Nanopores of Shale Gas Reservoirs 15m

The shale gas reservoirs are rich in organic and inorganic nano-sized pores. Generally, both adsorbed gas and free gas are considered to be exist in organic matrix pores, while there is no adsorbed gas in inorganic matrix pores. Therefore, gas transport mechanism is quite different in these two types of nanopores. Many researchers have provided the presence of water film on the inorganic pore surface due to its strong hydrophilic ability. However, most of the models for gas transport in shale inorganic nanopores ignore the effect of water distribution on gas transport, which lead to overestimating of gas transport capability.
In this paper, a new gas transport model for inorganic nanopores in shale gas reservoirs is proposed. First, considering the gas transport capability varies with pore size, the logarithmic normal distribution function is utilized to describe the nanopores distribution in shale inorganic porous media. Then, the influence of real gas effect, stress dependence and water distribution are all considered to derive the model. The validation results show that the proposed model and published experimental data can be well fitted. Finally, the effect of each factor on gas transport capability in shale inorganic nanopores is analyzed and discussed.
The results indicate that the gas transport capability will decrease with the increase of relative humidity. When the relative humidity increases to a critical value, the nanopores will be blocked with capillary water. During depressurization development process, the effective pore size will apparently reduce due to the influence of stress dependence, which cannot be ignored. Furthermore, when the shale inorganic matrix pores have the same mean pore size, the gas transport capability under various pore distribution probability is quite different, and the lower the peak frequency, the higher the transport capability. Meanwhile, under high temperature and low pressure conditions, methane transport capacity is significantly higher than ethane and carbon dioxide.
The research results of this paper can provide a reference for the analysis of nanoscale gas flow mechanism in shale matrix, and also provide a theoretical basis for more accurate production prediction of shale gas wells.

Speaker: Ms. Shan Wang
• 17:30
Soft fillings in nanoporous solids: Electro-polymerization and mechanical characterization of polypyrrole in nanoporous silicon 15m

We investigate the properties of the electrically conductive polymer Polypyrrole (PPy) in tubular pores of monolithic micro (pore diameter D < 2 nm)-, meso- (2 nm < D <20 nm) and macroporous (D > 50 nm) silicon. We successfully demonstrate a homogeneous filling via electro-polymerisation for the extremely anisotropic confinement of 12 nm pore diameter and 180 µm pore length. The kinetics of this process are explored experimentally and related to phenomenological models for polymerization in confined geometries. First experiments on the enhancement of mechanical properties of the resulting soft-hard nanohybrids are also presented.

Speaker: Patrick Huber (Hamburg University of Technology)
• 17:45
A ferroelectric liquid crystal confined in cylindrical nanopores: Reversible smectic layer buckling, enhanced light rotation and extremely fast electro-optically active Goldstone excitations 15m

The orientational and translational order of a thermotropic ferroelectric liquid crystal (2MBOCBC) imbibed in self-organized, parallel, cylindrical pores with radii of 10, 15, or 20 nm in anodic aluminium oxide monoliths (AAO) are explored by high-resolution linear and circular optical birefringence as well as neutron diffraction texture analysis. The results are compared to experiments on the bulk system. The native oxidic pore walls do not provide a stable smectogen wall anchoring. By contrast, a polymeric wall grafting enforcing planar molecular anchoring results in a thermal-history independent formation of smectic C helices and a reversible chevron-like layer buckling. An enhancement of the optical rotatory power by up to one order of magnitude of the confined compared to the bulk liquid crystal is traced to the pretransitional formation of helical structures at the smectic-A-to-smectic-C* transformation. A linear electro-optical birefringence effect evidences collective fluctuations in the molecular tilt vector direction along the confined helical superstructures, i.e. the Goldstone phason excitations typical of the para-to-ferroelectric transition. Their relaxation frequencies increase with the square of the inverse pore radii as characteristic of plane-wave excitations and are two orders of magnitude larger than in the bulk, evidencing an exceptionally fast electro-optical functionality of the liquid-crystalline-AAO nanohybrids.

Speaker: Patrick Huber (Hamburg University of Technology)
• 18:00
Synthesis and characterisation of B-substituted nanoporous carbons with high energy of hydrogen adsorption. 15m

The world is running out of fossil fuels and the products of their burning in air (mostly CO2) have already impacted global climate. Today it is clear that in near future we need to convert the global energy economy towards cleaner and renewable fuels (like hydrogen). However, to efficiently store hydrogen at ambient temperature and not too high pressures, we need to develop the hydrogen sorbent with simultaneously optimized specific surface and adsorption energy.

Here we report the first studies of the potential effectiveness of arc-discharge procedure to synthetize nanoporous, carbon based sorbents with characteristics required for hydrogen storage in vehicular applications. The arc-discharge, successfully used in the past to synthetize fullerenes and nanotubes, provides a relatively easy way to incorporate heteroatoms into pure carbon structures. Therefore we have assumed that we can adjust the synthesis parameters to prepare other graphene-based structures, with a variety of shapes, sizes, and interconnections between graphene fragments.

The properties of first boron-substituted carbons obtained by this method are promising: the prepared carbon soot contains a variety of organized, graphene based structures, and the HRTEM and NMR study confirm the presence of boron nanoclusters, partially incorporated into graphene layers. The energy of hydrogen adsorption is the highest ever observed experimentally in carbon-based sorbents: at least 10 % of adsorption sites adsorb hydrogen with the energy higher than 6.5 kJ/mol, and the strongest adsorption occurs with energy higher than 10 kJ/mol. These values are significantly larger than hydrogen adsorption energy in activated carbons (~4.5 kJ/mol). However, the specific surface of as-prepared samples is low (~ 200 m2/g), even after thermal activation. Therefore the samples are currently activated chemically (with KOH); this procedure should increase the surface accessible for adsorption by one order of magnitude.

Speaker: Ms. Katarzyna Walczak (Laboratoire Charles Coulomb, University of Montpellier, France)
• 17:15 18:45
Poster 2: Poster 2-F
• 17:15
A Transient Productivity Model of Multi-stage Fractured Horizontal Wells in Shale Gas Based on the Continuous Succession Pseudo-steady State Method 1h 30m

The multi-stage fractured horizontal wells(MFHW) is the key technology for developing shale gas reservoirs. After the stimulated reservoir volume is fractured, the gas flowing in matrix is non-linear seepage controlled by the nano-scale pores, while the seepage in stimulated region is transformed into Darcy flow controlled by the micro-scale fracture network. In this paper, the steady-state productivity model of MFHW is firstly established by comprehensively considering the multi-scale flowing states, shale gas desorption and diffusion after shale fracturing, which coupled flows in matrix and stimulated region. On this basis, for the first time, a transient productivity calculation model of MHFW combined the material balance equation is obtained with the continuous succession pseudo-steady state method (SPSS), which considered the unstable propagation of pressure wave. And the horizontal well productivity prediction and factors analysis are carried out by using the SPSS. The results show that the SPSS has the advantages of simple process of calculation, fast calculation speed and high agreement with numerical simulation results. During the production process, the desorption effect of shale gas is the key factor affecting the transient productivity of gas wells. During the production process, the desorption effect of shale gas is the key factor affecting the middle and late stage production of gas wells. With the increase of the radius and permeability of the mass fractures, the diffusion coefficient and Langmuir volume, the productivity of shale gas wells would increase, while the increasing rate would decrease. And the effect of Langmuir pressure on productivity is less. It is concluded that this method provides a theoretical basis for the calculation of transient productivity of shale gas fractured horizontal wells.

Speaker: Mr. Fanhui Zeng (Southwest Petroleum University)
• 17:15
Tuned Nanoparticle Deposition In Porous Media To Improve Efficiency of Nanoremediation 15m

Nanoremediation is an innovative environmental nanotechnology aimed at reclaiming contaminated aquifers. It consists in the subsurface injection of a reactive colloidal suspension for the in-situ treatment of pollutants. The greatest challenges faced by engineers to advance nanoremediation are the effective delivery and the appropriate dosing of the nanoparticles into the subsoil. These are necessary for the correct emplacement of the in situ reactive zone and to minimize the overall cost of the reclamation and the potential secondary risks associated to the uncontrolled migration of the injected particles.
In this study, a model assisted strategy, called NanoTune, is developed to control the distribution of colloids in porous media. The proposed approach consists in the sequential injection of a stable suspension of reactive nanoparticles and of a destabilizing agent with the aim of creating a reactive zone within a targeted portion of the contaminated aquifer. The controlled and irreversible deposition of the particles is achieved by inducing the mixing of the two fluids in the desired portion of the aquifer.
This approach is here exemplified by the delivery of humic acid-stabilized iron oxide nanoparticles (FeOx), a typical reagent for in situ immobilization of heavy metals. Divalent cations, which are known to cause rapid aggregation of the suspension because of their strong interaction with the humic acid coating, are used as destabilizing agents. The injection strategy is here applied in 1D columns to create a reactive zone for heavy metal removal in the central region of the sandy bed. The software MNMs was used to assess the correct sequence and duration of the injection of the different solutions in the 1D medium. Moreover, the numerical code MNM3D (MNM3D - Micro and Nanoparticle transport Model in 3D geometries) was developed by the authors of this work to support the case-specific design of the injection strategy during field scale applications.
The NanoTune approach represents an advancement in the control of the fate of nanomaterials in the environment, and could enhance nanoremediation making it an effective alternative to more conventional techniques.
Co-funded by: EU H2020 Reground Grant Agreement No. 641768

Speaker: Tiziana Tosco (DIATI, Politecnico di Torino)
• 18:30
3D Reconstruction and permeability calculation from 2D thin sections 15m

Permeability prediction of porous media is be of great significance for both petroleum and environment fields. The permeability of a porous medium can be directly calculated based on one 2D thin section. However, it is often doubtful. Therefore, a new technique of the permeability prediction from a 2D thin section is proposed. First 3D porous media is reconstructed from a 2D image using multiple-point statistics. Then the single- and two-phase flow simulations are carried out based on the reconstructed 3D porous media. The absolute permeability is calculated by computing Navier-Stokes equation and Darcy's law. The relative permeability is predicted using pore network method. In order to validate the method, the permeability calculation results are compared with them which are computed from the 3D real porous medium obtained using micro-CT scanner. The comparison shows that the technique is reliable, which offers petroleum and environment researchers a novel method for predicting the permeability when a 2D thin section is available.

Speaker: Yuqi Wu (China University of Petroleum (East China))
• 18:30
A New Method to Establish A Full Scale Diagram For Unconventional Oil Reservoir 15m

The pore and throat size of the global unconventional oil and gas reservoir is mainly nano-scale, and the minimum radius is evenly about 1 nm. The characteristics of the unconventional oil reservoir depend greatly on the micro pore structure of the reservoir. Its main contents include the shape, size, quantity, connectivity and distribution characteristics of reservoir pores and throats. The research methods of microscopic pore structure of reservoirs mainly include: digital core technology and experimental testing which is widely used in the practical work. has great limitations. For example, Low-temperature nitrogen adsorption method is used in the study of pore structure and its effective range is 2 nm - 50 nm, which has limited significance for the study of tight oil exploration; The test range of pore and throat radius of high pressure mercury technology is extended to 1.8nm-500um, but it may cause fracture easily when the mercury pressure reaches high level and have a large error in the test of small pores. In order to enhance oil recovery, it is necessary to establish a full-scale distribution diagram to complete the different oil zones in the tight oil zone reservoir by combining high-pressure mercury intrusion, low-temperature nitrogen adsorption and other measuring means. By using a full-scale distribution diagram,we could put forward the comprehensive classification evaluation method, establish micro-nano flow mathematical model, reveal the permeating law and provide theoretical support and technical guidance for the unconventional reservoir.

Speaker: Taiyi Zheng ( Institute of Porous Flow and Fluid Mechanics)
• 18:30
A Novel Mehtod to Correct Steady-State Relative Permeability for Capillary End-Effects Based on Simulation Approach 15m

In laboratory steady-state measurements of relative permeability, capillary discontinuities at sample ends give rise to capillary end-effects (CEE) and keep a higher water saturation toward the core end. The water saturation measured is higher than ideal saturation without CEE and result of erroneous relative permeability curves[1] finally. Especially in tight sand cores, high capillary force and low flow rates can cause end-effects to become more important in the interpretation of the steady-state experiment tests.
On the basis of capillary force measurement data，a novel method was initiated to make correction to relative permeability data combined experiment data and numerical simulation[2] for tight sand cores. First, one dimensional numerical simulator was created for oil and water two phase flow considering the mechanism of CEE. Water saturation distribution and CEE region could be clarified corresponding to the relative permeability curve input. And then correct relative permeability input constantly until the average water saturation simulated consistent with the saturation measured by experiment. The new method could make correction to relative permeability based on the traditional experiment data of multiple fractional flow and same total flow rate, which is more practical than “intercept method”[3]. The impact of end-effects and practical means of reducing the end effects for tight sand cores are also discussed in this paper.

Speaker: Dr. Shiyuan Zhan (School of Petroleum Engineering, China University of Petroleum (East China))
• 18:30
A numerical simulation study on the hydraulic fracture propagation in heavy oil reservoir with the THM coupling 15m

For further research on the effect of heavy oil viscosity on the fracture geometry, this paper establishes heavy oil fracturing model and conventional fracturing model based on thermal-hydraulic-mechanical (THM) coupled theory, Walther viscosity model and K-D-R temperature model. We take viscosity and density within heavy oil fracturing model as functions of pressure and temperature, while that as constants within conventional fracturing model. A heavy oil production well is set as an example to analyze the differences between the two models in account of thermo-poro-elastic effect. The results show that temperature has the greatest influence on heavy oil viscosity, while pressure presents the least influence; A cooling area which has 0～1meters width and varied length generates near the fracture. Heavy oil viscosity increases sharply in this area, presenting an area of viscosity increment. Heavy oil viscosity increases faster closer to wellbore, and a large viscosity increment will reduce the mobility of heavy oil and prevent fracturing fluid from bumping into reservoir; The special viscosity distribution results in significant differences of pore pressure, oil saturation and changing trend between these two models; In heavy oil reservoir fracturing model, the thermal effect completely exceeds the influence of pore elasticity, and the values of the fracture length, width and static pressure are larger than those calculated in conventional fracturing model. Thus the change of heavy oil viscosity plays a dominant role in influencing the expansion of hydraulic fracture.

Speaker: Qiang Wang
• 18:30
A Semi-analytical Model for Production of Volume Fractured Vertical Wells in Tight Oil Reservoirs under the Influence of Wettability Alteration 15m

It is one of important technologies for improving the recovery of volume fractured vertical wells (VFVWs) in tight oil reservoirs that adding surfactant in fracturing fluid to alter the wettability of reservoirs， which can reduce the viscosity of crude oil, increase the strength of imbibition and improve permeability of oil phase in stimulated-reservoir-volume (SRV) region. In this paper, a multi-linear fractal model of VFVWs in tight oil reservoirs including the influence of wettability alteration was established. In this model, mathematical characterizations of wettability alteration, imbibition and distribution of fractures networks were considered in SRV region though the fractal theory and mechanism of imbibition. And the semi-analytical solution was given by using the Laplace transformation and the pressure responses in the domain of real time were obtained with Stehfest numerical inversion algorithms. Though this model, the influence of the wettability alteration was analyzed. The results show that technology of wettability alteration can improve the oil recovery, which is consistent with the experimental results. In addition, the fractal distribution and scale of fractures network have great effect on the almost whole period of the porous flow process, and the imbibition causing by wettability alteration mainly influence the period of cross-flow.

Speaker: Zhiyuan Wang (Institute of Porous Flow and Fluid Mechanics，University of Chinese Academy of Sciences)
• 18:30
A Two-Phase Flow Model for Pressure Transient Analysis of a Water Injection Well Considering Water Imbibition in Natural Fractured Reservoirs 15m

Water imbibition is an important stimulation mechanism especially in natural fractured reservoirs. Tremendous attention has been focus on oil-water two phase flow mathematical model according to the laboratory experiments of water imbibition. However, the pressure transient analysis model so far has rarely considered water imbibition. In this paper, we present a novel two-phase flow imbibition model for pressure transient analysis of a water injection well in natural fractured reservoirs. The model was solved by using Laplace transforms and Stehfest numerical inversion. Based on the solution, the model was validated by selecting a case in literature. The features of type curves were studied comparing to the two phase flow model without considering water imbibition. In addition, the influences of wellbore storage coefficient, skin factor, matrix permeability, shape factor, initial water saturation, fracture water saturation and storativity ratio on the typical curves were analyzed by studying the new model with different seepage regimes. The novel well test model provides important references for some reservoir engineers in the design and evaluation of stimulation treatments in natural fractured reservoirs.

Speaker: Dr. Mengmeng Li
• 18:30
An analytical model of apparent permeability for shale gas reservoir considering characteristics of nanopore distribution 15m

The structure of nanopore in shale is complex, multiple gas migration mechanisms coexist. In this paper, we have used a bundle of tortuous capillary tubes with different diameters to represent porous structure of shale considering slippage effect and Knudsen diffusion and surface diffusion. Fractal theory is applied to mathematically express the capillary diameter distribution and their tortuosity. A multi-scale gas flow model for shale is established, and an analytical model of apparent gas permeability for shale is deduced. Simulation results show that the ratio of apparent permeability to Darcy permeability increases with the decrease of pressure and the increase of temperature. Apparent permeability increases with the increase of fractal dimension for pore size and the decrease of fractal dimension for tortuosity. Apparent permeability decreases rapidly as pressure increases, and then stays steady. With the increase of pressure, the contribution of slip flow to apparent permeability increases quickly first and stays stable after, but the contribution of surface diffusion to apparent permeability is inverse. The contribution of slip flow to apparent permeability dominates, followed by surface diffusion. The contribution of Knudsen diffusion is minimum. The contribution of surface diffusion to the apparent permeability is underestimated with constant surface diffusion coefficient. The model provides a kind of method and theoretical basis for analysis of shale gas flow behavior.

Speaker: Mr. Xieyang Pu (Sinopec & Southwest Petroleum University)
• 18:30
Characteristic of Coal Pore Structure and Its Relationship with Sedimentary Environment in Hegang Basin 15m

Characteristic of Coal Pore Structure and Its Relationship with Sedimentary Environment in Hegang Basin
Wang You-zhi1 , MaoCui2
（1.Exploration and Development Research Institute of Daqing Oilfield Company Ltd.，Heilongjiang, Daqing 163712, China
2.School of Geosciences, Northeast Petroleum University, Heilongjiang, Daqing 163318, China）
Abstract:Base on low temperature nitrogen adsorption method, Argon Ion milling Scanning Electron Microscopy and Nuclear Magnetic Resonance, the characteristics of pore structure and sedimentary environment are discussed to find the relation between them.The result show that four pore models are established through analysis on pore structure features of Hegang Basin, and they are corresponding to different mine areas respectively. Micropores are dominated for the most part and the small pores take the second place in Hegang Basin, with stronger adsorption capacity. However, the significant difference exists in the influence of different pore models on CBM desorption and filtration. Pore Model I has relatively independent pores with easy desorption but poor filtration capacity. Pore Model II, III and IV have relatively strong adsorption capacity, but the latter two enjoy more developed net fractures. This is of great significance for improving permeability, and beneficial to CBM output. The coal facies are forest and marsh covered area in the south of Hegang Basin, which is in transition to shallower and deeper water-covered marshes. Vitrinite/inertia (V/I) ratio increases from south to north, while ash content and total sulfur content are on the contrary, indicating the burial speed accelerates coal forming environment with water-covered deepening. The fan delta plain is the main coal accumulating environment in the middle of the Basin, and the provenance pours in from the south of the Basin, extending to the northeast through a braided riverway, so that stable thick coalbed can be formed easily in the north mine areas in the Basin.The sedimentary environment has a control effect on coalbed pore structure and fracture development to some extent. From south to north in Hegang Basin, the water body becomes deeper and deeper, and the coal model IV has a gradual transition to Model I; the coarse lithological association become the fine one, and ash content decreases. The high ash content causes part of pores to be filled, so that the pore system becomes diversified and complicated. The water power is ever-increasing in the south, and near provenance, turbulent water body allows peat cracks to gradually increase in quantity and scale in the burial process. The coal seam is thick in the north and thin in the south in Hegang Basin, and filtration capacity and operability of late transformation are strong in the south and weak in the north. Thus, when seeking advantageous targets in Hegang Basin, we should take into account macroscopic tectonic setting and microscopic features in an all-around way.
Key word：pore structure; pore model; coal facies; sedimentary environment; Hegang Basin

Speaker: Mrs. cui Mao
• 18:30
Characteristics of Remaining Oil Micro⁃Distribution in Laojunmiao Oilfield after Waterflooding 15m

Using the fresh oil⁃bearing core samples, this paper studied the remaining oil micro⁃distribution or occurrence and the influence of long⁃term waterflooding on the reservoir pore structures in major reservoir of Laojunmiao oilfleld in Yumen. The study indicates that the remaining oil occurrence in the reservoir is dominated by film or interstitial form, and the pore stucture is the key factor influencing the remaining oil distribution and oil displacement efficiency. After long⁃term water flooding, the interstitial material and flowing resistance are decreased in the pores with better physical property. For L layer, its remaining oil is still distributed in big pores，while for M layer, its one third of remaining oil is in the small pores or disseminated on rock surface, and hard to flow. The content of movable water in the big pores is directly related to the total water cut in the reservoir.

Speaker: Mr. Guoqiang Sang
• 18:30
Micro-scale effect of CO2 diffusion on two-phase flow in dual-porosity of tight oil reservoirs 15m

Carbon dioxide (CO2) diffusion in dual-porosity plays a great important role for effective flow in tight oil reservoir. The CO2 diffusion coefficient in matrix is different with the coefficient in fracture because of micro-scale effect. Matrix diffusion coefficient and fracture diffusion coefficient was introduced and respectively used into matrix flow model and fracture flow model. Using pressure drop method, matrix diffusion coefficient in tight porous media was determined by soaking the oil-saturated core in CO2 filled container with constant temperature. This paper developed a two-phase flow model in dual-porosity media coupling with two CO2 diffusion equations, and solved by Finite Difference Method (FDM). To reveal the scale effect, the results of two different numerical models were compared: (1) matrix diffusion coefficient equals to fracture diffusion coefficient; (2) matrix diffusion coefficient differ to fracture diffusion coefficient. Finally, this study verified the micro-scale effect on fluid flow in tight formations.

Speaker: Mr. Shouya Wu (School of Petroleum Engineering, China University of Petroleum (East China))
• 18:30
Microcosmic Visual Experimental Study of CO2 Huff-n-Puff injection to Enhance Oil Recovery in Liquid-Rich Shale Reservoirs 15m

CO2 huff-n-puff is an effective way to recover oil in liquid-rich shale reservoirs with multistage fractured horizontal wells. Numerous core-scale lab experiments have been conducted to prove its potential. There are many mechanisms during the CO2 huff-n-puff process such as pressure repressurization, miscibility, molecular diffusion, relative permeability hysteresis, gas dissolve and oil swelling, etc. However, the roles these mechanisms played are still in blur. The purpose of this study is to investigate the functions of related mechanisms through experimental and numerical methods.
In this experimental study, a new artificial visual microscopic model was built to observe the CO2 diffusion and shale oil percolation in the nanoscale porous media during the CO2 huff-n-puff process. Different parameters such as oil composition, reservoir properties (permeability, porosity) and operation conditions (injection pressure, soaking time, depletion rate, gas injection cycles) were analyzed. After that, micro-nanofluidic chips were designed to mimic different cases during CO2 huff-n-puff: homogenous / heterogenous with / without natural fracture. Numerical study was also applied to simulate the CO2 huff-n-puff process to investigate the functions of different mechanisms and to be applied to fit some of CO2 field plots which were performed in Bakken formation of North Dakota. Different simulation cases were designed and established to investigate different mechanisms such as pressure repressurization, miscibility, relative permeability hysteresis, capillary pressure and oil swelling.
This work clearly shows the CO2 diffusion and shale oil percolation process in the shale oil reservoir media through the microscopic experimental model. The micro-nanofluidic chips can simulate the nanopore to investigate the shale oil phase behavior in nanopores combined with hydraulic fracture and natural fractures. Visual observation of the CO2/shale oil displacements indicate that there is an interaction between phase behavior and microscopic heterogeneity. The experimental and numerical results show that the most important mechanism is pressure repressurization, followed by miscibility, diffusion, relative permeability hysteresis, and oil swelling. This study explains how CO2 diffusion in liquid-rich shale oil reservoirs is different in lab core scale level from microscale level, and how different mechanisms affects CO2 performance in improving oil recovery from unconventional resources.

Speaker: Dr. lei li
• 18:30
Numerical Modelling of Microbial Enhanced Oil Recovery under the Effect of Environment 15m

Considering the sensitivity of microorganisms to the different environment, the microbial growth kinetics equation was improved, and a 3D two-phase five-component mathematical model which could fully reflect the microbial flooding process in the reservoir medium was established. The components of the model involve oil, water, microbe, nutrient and metabolite. The model integrates the effects of microbial growth / death, nutrient consumption, metabolite production, chemotaxis, convection, diffusion, oil viscosity reduction, adsorption, desorption, oil-water interfacial tension change and other properties. Taking into account the effect of environmental factors on the microbial growth model and inconsistent growth rate of microbes on the ground and underground, the microbial growth kinetics equation was improved which was based on the Monod model. In this paper, the microbial growth models, mixed solution and maximum specific growth rate were analyzed. The results showed that the metabolite concentration calculated by the Khan model is lower than that of the Monod model, but the difference in the metabolite concentration of the two models is not sufficient to have a significant effect on the recovery factor. With the increase in the amount of nutrient mixed solution and the injection time, this difference will gradually become apparent.

Speaker: Mr. Ma Yuandong
• 18:30
One-Dimensional Transient Inter-Porosity Flow Model in Tight Porous Media with Consideration of Fracture Pressure Depletion 15m

Tight porous media have the characteristics of extremely low permeability and the permeability of it is sensitive to the effective stress. Stimulated reservoir volume (SRV) is usually performed to improve the production of well in tight porous media. The SRV zone is usually considered as a dual-porosity medium in well test or numerical simulation due to the computationally efficient. The shape factor is the key of the dual-porosity model, which determines the ability of mass transfer between the matrix block and fracture. However, the conventional shape factor model which is usually obtained based on the assumptions of pseudo-steady state and constant fracture pressure, which lead to a poor application in the characterization of mass transfer between matrix and fracture in tight porous media.
In this paper, a new model was established by considering the effect of stress sensitivity and time-dependent fracture pressure boundary condition. Pedrosa substitution and perturbation method were applied to eliminate the nonlinearity of the model. By using the Laplace transformation method, the analytical solution in the Laplace domain was obtained. According to the Duhamel principle, the solution for time-dependent fracture pressure boundary condition was obtained. Then validation was performed to show that the model is valid. Finally, influence of stress-sensitivity and fracture pressure depletion on shape factor and inter-porosity rate were discussed. Study shows that: Stress sensitivity of the matrix has an obvious influence on the inter-porosity flow. As the value of stress sensitivity coefficient increases, the value of the shape factor and inter-porosity rate decrease. After considering the impact of fracture pressure depletion, the effect of stress sensitivity on the shape factor is more significant. The faster the decreasing rate of fracture pressure, the larger the value of the shape factor at the initial time. However, the value of the shape factor is smaller at the later time due to the earlier and faster decreasing of fracture pressure. The inter-porosity rate will rise first to reach equilibrium and then decrease when the decreasing rate of fracture pressure is small. The new model can be used in the study of well test interpretation and numerical simulation, which provides a theoretical guidance for the reasonable development of tight porous media.

Speaker: Shan Huang (China University of Petroleum-Beijing)
• 18:30
Petrographic characterization of low-permeable to tight turbidite sandstone from Eocene Shahejie Formation using micro-CT. 15m

Pore scale flow simulations in reservoir rocks heavily depend on characterizing and modeling of the pore space. Single scale and multiscale pore network extraction from micro-CT images are going through extensive development. However, the choice of pore network extractions method is sensitive to the rock nature (homogenous, complex or microporous). Additionally, the success of the pore network to predict flow properties relies on image quality and image segmentation. In this study, we characterized four samples from a low permeable to tight sandstone reservoir using micro-CT at the expertise Centre for X-ray Tomography at Ghent University (www.ugct.ugent.be) to assess the impact of microporosity on the rock model. All samples were scanned at a resolution of 1.5 to 1.7 micron. As the samples can be categorized as illite rich and kaolinite rich, attention towards the clay minerals were given as they play a vital role to influence microporosity. (pore size < 1.7 micron) Image segmentation analysis from micro-CT images indicated that 5-6 % of microporous regions were present in kaolinite rich sandstone, while illite rich sandstone displayed 1.7-1.8 % microporous regions. Correlation of mineral phase data from micro-CT and XRD revealed that the microporosity consisted mainly out of dissoluted feldspar grains and clay mixed cement. In kaolinite rich, macropore system does not percolate without micropores, while in illite rich sandstone the pore system percolates without micropores. In illite rich sandstone, the total MICP porosity is equal to the macroporosity (pore diameter >1.7 micron) determined on the 3D micro-CT, this means that the macropores are well connected and microspores do not play any role in the flow process. However, in kaolinite rich sandstone, the macroporosity determined from the 3D X-ray micro-CT images is far less (almost 50 %) than the total MICP porosity which means that almost 50 % of the porosity consisted out of pores that were not detected by the micro-CT scan. In these kinds of rocks, the pore system does not percolate without the micropores and a multiscale approach is needed to characterize such complex rock.

Speakers: Mr. Muhammad Jawad Munawar (China University of Petroleum (East China)) , Prof. Veerle Cnudde (Department of Geology, PProGRess/UGCT, Ghent University, ) , Prof. Chunmei Dong (China University of Petroleum (East China)) , Mr. Yuqi Wu (China University of Petroleum (East China))
• 18:30
Physical simulation experiment of different injected media huff and puff for tight porous media 15m

Different injected media huff and puff is a promising enhanced oil recovery approach in tight oil reservoirs, which can effectively supply formation energy and enhance the productivity of individual well. And there still exists many uncertainties of oil recovery mechanism in the process. In this study the experiment method of cyclic water injection and carbon dioxide huff and puff in tight porous media were established by high-pressure physical simulation system for large scale outcrops. Water, surfactant and carbon dioxide were employed on the physical simulation experiments, the corresponding development effectiveness and influence factors of different injected media huff and puff were analyzed, and the feasibility of carbon dioxide huff and puff after cyclic water injection was studied. The results show that the development effectiveness of carbon dioxide huff and puff is better than surfactant huff and puff, and cyclic water injection is the worst. Injection pore volume multiple is the important factors to effect development effectiveness, compared with water and surfactant, the injection ability of carbon dioxide is strongest, and carbon dioxide can effectively replenish formation energy. For different injection media, the higher the injection rate, the more pronounced the fingering phenomenon, the fluid can enter the model deep, and the better the effect. The dispersed phase of oil and water influence subsequent carbon dioxide injection. Oil component differentiation could occur after oil contacted with carbon dioxide, Aromatics in oil were extracted rapidly by carbon dioxide, and made component differentiation in the process of carbon dioxide huff and puff.

Speaker: Dr. Xiangyang Wang (Institute of Porous Flow and Fluid Mechanics，University of Chinese Academy of Sciences)
• 18:30
Productivity forecast model of vertical hydraulic fracturing well with varying conductivity in tight oil reservoir 15m

Hydraulic fracture has become an essential well stimulation technique in tight oil reservoirs. Large-scale vertical well fracturing can generate longer vertically oriented fractures to increase drainage area and, therefore enhance production and recovery efficiency of single well while also save costs[1]. The aim of this study is to develop a more practical productivity forecast model that can take into account of the heterogeneity of fracture distribution. Firstly, the fracture geometry model is established based on the features of pseudo three-dimensional model (P3D)[2] and the aperture changes along the length of hydraulic fracture. The hydraulic fracture is then divided into N segments. For each segment, fracture permeability is calculated by analyzing porosity change caused by proppant embedment, deformation, crush and diagenesis under the influence of effective closure pressure[3]. Finally, the productivity forecast model of vertically fractured well is established to deal with practical situations where the aperture and permeability of fractures cannot be neglected. The heterogeneous fracture model shows a good precision in the proof-test, and proved to be more credible and practicable than homogeneous model. We then report the results of several numerical simulations conducted for different values of fracture length, effective closure pressure and permeability of each segment, as well as a comparison with the simulated results of fracture model with homogeneity properties.

Speaker: Dr. Mingyu Cai (School of Petroleum Engineering, China University of Petroleum (East China))
• 18:30
Study on the Ultrasonic Propagation Law in Gas-Liquid Two-Phase Flow of Deep-Water Riser Annulus 15m

In order to achieve gas kick early detection outside the riser, the ultrasonic propagation law in gas-liquid two-phase flow of riser annulus need to be analyzed. Therefore this paper aims to explore the ultrasonic propagation law in gas-liquid two-phase flow of riser annulus under different conditions and to establish a quantitative relationship between the ultrasonic signals and gas void fraction.
1.The theoretical model of ultrasonic propagation in gas-liquid two-phase flow of riser annulus was established, which was used to analyze the ultrasonic propagation law preliminarily.
2.The experimental device of ultrasonic gas kick monitoring outside the riser was designed, and the ultrasonic signal responses under different conditions (such as probe installation method, void fraction, drill pipe eccentric degree, etc.) were analyzed.
3.The ultrasonic propagation law in gas-liquid two-phase flow of riser annulus was analyzed by using numerical simulation method. The reliability of numerical simulation method established was confirmed by comparing with the experimental results.
Through comparing the theoretical analysis, experimental and numerical simulation results, the reliability of experimental methods and numerical simulation model established in this paper were verified. Therefore the experimental device and numerical model in this paper could be used to study the ultrasonic propagation law in gas-liquid two-phase flow of deep-water riser annulus. On this basis, the mounting position and transmission frequency of ultrasonic probes were optimized. Meanwhile, the gas void fraction sensitive parameters of ultrasonic such as sound amplitude and speed were selected. Through further studies, the calculation model of comprehensive sensitive parameters was founded, and the quantitative relationship between ultrasonic characteristic parameters and gas void fraction under different conditions was established based on a large number of experiments and numerical simulations.
In this paper, the quantitative relationship between ultrasonic characteristic parameters and gas void fraction under different conditions was established. It provided a theoretical and technical support for the gas kick quantitative monitoring in deep-water riser by using ultrasound. On this basis, the riser gas kick early detection equipment (R-GED) is being developed, which is expected to achieve gas kick quantitative monitoring and even accurate determination of the degree of bottom hole gas kick.

Speaker: Yuqiang Xu (China university of petroleum (East China))
• 17:15 18:45
Poster 2: Poster 2-G
• 17:15
Semi-analytically derived flow-rate/pressure drop relationships for the flow of yield stress fluids through rectilinear pipes of non-circular cross-sections. 15m

For the study of yield stress fluids flow in porous media, the complex pore-scale structure has been extensively idealized as bundles of capillaries. A bundle of rectilinear capillaries of circular cross-section has been used over the past decade for the development of a new method of pore-size distribution (PSD) characterization based on the injection of a yield stress fluid. The main idea of this method, considered as an alternative to the mercury injection porosimetry, is that since those fluids cannot flow below a certain stress threshold, by measuring the evolution of the flow rate versus the pressure drop across a rock sample, its PSD can be retrieved through some “inversion” procedure. The numerical inversion techniques allowing to derive the PSD from experimental data and the experimental feasibility of this innovative and non toxic technique have been presented in recent works (e.g. Rodríguez de Castro et al., 2014). In order to be more representative of real pore cross sections, the flow rate/pressure drop relationship has been further investigated using detailed numerical simulations in bundles of capillaries with square or triangular cross-sections (Malvault et al, 2017).
In this study we propose a set of semi-empirical formulas that relate flow rate and pressure drop for yield stress fluids flowing in rectilinear capillaries of square, triangular, and rectangular cross sections. This approach is based on two main notions: the Critical Bingham number beyond which a yield stress fluid cannot flow anymore in a pipe of a given cross section; and a Shape Factor, whose physical meaning is the deviation of the considered shape of cross section (square, triangular, rectangular), from the ideal circular one.
This approach was initially proposed by Saramito and Roquet (2001) for the case of Bingham fluids flowing in pipes of square cross section. By following the same reasoning, a generalization for the flow of Herschel-Bulkley fluids in pipes of square, rectangular and triangular cross sections is developed.
The results obtained using the semi-analytical formulas developed in this work are shown to be in very good agreement with those derived from detailed numerical simulations. They can therefore be used for modeling flow in bundles of capillaries in a much more efficient and rapid manner.

Speaker: Prof. Azita AHMADI-SENICHAULT (Professor, University of Bordeaux-FRANCE)
• 17:30
Nanoparticle heterogeneous adsorption in porous media 15m

Colloidal particles released by various chemical and industrial processes penetrate soils and groundwater, and transport themselves other contaminants like heavy metals or PCBs. Thus, an accurate description of the transport and retention of these particles is required to prevent and manage environmental contamination, like the pollution of drinking water supplies. Literature stands that colloid sorption (also referred to as attachment) is the main process controlling transport in the case of strong particle-medium interactions and absence of straining events. Sorption mechanisms have been widely studied using micromodels (2D systems or pore scale studies [2,3]) or indirect observation (Breakthrough curves [6]), completed by simulations to account for the complexity of real media like soils [7]. We propose here to tackle the usually neglected impact of the system’s heterogeneity, in direct 3D visualizations at local and global scales.

Glass beads at random close packing are chosen as a model porous medium for their large pore throat size distribution (down to 0 at contact points). We inject in the system suspensions of iron nanoparticles (NPs), with a strong attraction interaction with the medium, which leads to heterogeneous adsorption and coverage of the beads. Coverage is followed by confocal imaging in time and at multiple scales; from local singular geometries (pore scale, tens of µm) to global average behaviours (sample’s scale, cm).

We show that for a given NPs-medium affinity, the dynamics of adsorption strongly depend on the accessibility of the surface, in accordance with the flow streamlines. We identify configurations (regions of highest confinement, i.e. contact points between beads) where the adsorption is fully diffusion limited. At early stages of adsorption, these regions can account for up to 10% of the surface of the medium and therefore impact the supposedly known dynamics at global scale.

To go further, we link local observations and global description of the dynamics to discuss the impact of the pore size heterogeneity, at various flow rates (Darcy velocities in the range 0.05 to 9 mm/s). We show the impact of the flow on the heterogeneity of the deposition of the NPs through 1D coverage profiles over time and over the whole sample. Finally, we propose a simple model describing the full beads coverage dynamics over space and time. We can define and decouple a geometric (characteristic deposition length) and a kinetic (characteristic exploration time) components for a complete understanding of the adsorption mechanisms.

Speaker: Gaétan Gerber (ENPC - Harvard SEAS)
• 17:45
Analysis of flow behavior for a well with a vertical fracture at an arbitrary azimuth in a rectangular anisotropic reservoir 15m

Permeability anisotropy is a common feature of hydrocarbon reservoirs. In practice, hydraulic fracturing has been an effective technique for enhancing the productivity of wells in low permeability reservoirs. However, most study about hydraulic fractures focuses on isotropic reservoirs and few literatures discussed the effect of permeability anisotropy and azimuth angle of fractures on the producer’s transient flow. Therefore, we attempt to develop an analytical model for a fracture with an arbitrary azimuth angle in an anisotropic reservoir. First, we derive an analytical solution by spatial integration of point-source along the direction of the fracture. Then, we do Cartesian-coordinate transformation to transform the original problem with anisotropic permeability into an equivalent problem with isotropic permeability. Finally, we obtain an analytical solution for the proposed model. Several general conclusions drawn from this model are as follows: 1) The solutions obtained by the conductivity influence function are validated with Chen and Raghavan’s results. Five flow regimes can be observed: bilinear flow, linear flow, early radial flow, compound linear flow, and pseudo-steady-state flow. 2) Horizontal permeability anisotropy has a strong effect on transient responses, which can be extended to pseudo-steady-state flow regime. Effect of the azimuth angle of a fracture mainly focuses on the bilinear and linear flow regimes. The more the fracture deviates from the direction of maximum horizontal permeability, the less the drawdown needed to maintain constant flow rate. Therefore, the optimum orientation for the hydraulic fracture is perpendicular to the direction of maximum horizontal permeability. 3) Outer boundary size dominates pseudo-steady-state flow and aspect ratio mainly affects the periods of radial flow and compound linear flow. The larger the outer boundary size, the longer the radial flow regime and the later pseudo-steady-state flow appears. The existence and duration of bilinear flow and linear flow are mainly determined by dimensionless fracture conductivity. This model provides a theoretical basis for well pattern deployment and fracture configuration in anisotropic reservoirs.

Speaker: Mr. Guoqiang Xing (Research Institute of Petroleum Exploration and Development, PetroChina)
• 18:00
Multiscale model reduction for thermoporoelasticity problem in fractured media. 15m

Keywords:
Generalized Multiscale Finite Element Methods, heat and mass transfer, elasticity, thermoporoelasticity, fractured media.

Abstract:
In this work, we consider heat and mass transfer in the geothermal reservoir and present a multiscale model reduction technique using a Generalized Multiscale Finite Element Methods (GMsFEM). The solution techniques for these problems on the fine grid require high resolution. In particular, the discretization needs to honor fracture distribution on a fine grid. This gives rise to a fine-scale problems with many degrees of freedom, which can be very expensive to solve. We investigate multiscale approaches that attempt to solve such problems on a coarse grid by constructing offline and online multiscale basis functions in each coarse grid block. In our work, we follow Generalized Multiscale Finite Element Method and develop a multiscale procedure where we identify multiscale basis functions in each coarse block using snapshot space and local spectral problems. To illustrate the performance of our method, we present numerical results for thermo-poroelasticity problem in domain with complex fracture distribution. We present the details of mathematical model, multiscale methods, and the numerical analysis.

Speaker: Valentin Alekseev (NEFU an internationally competitive Research Laboratory on Multiscale Model Reduction (MMR))
• 18:15
Nonlinear finite-volume schemes for complex flow processes and challenging grids 15m

The numerical simulation of subsurface processes requires efficient and robust methods due to the large scales and the complex geometries involved. To resolve such complex geometries, corner-point grids are the industry standard to spatially discretize geological formations. Such grids include non-planar, non-matching and degenerated faces. The standard scheme used in industrial codes is the cell-centered finite-volume scheme with two-point flux (TPFA) approximation, an efficient scheme that produces unconditionally monotone solutions. However, large errors in face fluxes are introduced on unstructured grids. The authors present a nonlinear finite-volume scheme applicable to corner-point grids, which maintains the monotonicity property, but has superior qualities with respect to face-flux accuracy. The scheme is compared to linear ones for complex flow simulations in realistic geological formations [1,2]. In addition, we present recent developments regarding convergence analysis for a family of nonlinear finite-volume schemes [3].

Speaker: Mr. Martin Schneider (University of Stuttgart)
• 18:30
Numerical Simulation Study on Pore Scale Seepage of Porous Media Based on Finite Volume Method 15m

Abstract: The numerical simulation of pore scale seepage in porous media is of great significance for the development of new energy sources, such as shale gas and geothermal energy. In the past few years, many scholars have developed the numerical method including finite volume method (FVM), Lattice Boltzmann Method (LBM) and molecular dynamics to achieve this simulation in different levels. It is easy for us to understand the basic idea of FVM, and then we can give a direct physical explanation for the above phenomena. FVM has stronger advantages in CFD (computational fluid dynamics) since the conservation law is meet in the whole computational domain. Based on the FVM, a large number of models on multi-phase flow and non-Newtonian fluid have been developed and it has showed more advantages in the aspect of solving complex pore flow problems. However, there are two shortcomings in the process of using FVM: (1) Due to the complexity of the pore geometry in the porous media and irregularly curved pore walls, the quality of the generated grid is not high; (2) It is relatively difficult to deal with a large number of pore boundaries in the porous media. To cure the above problems, a simple processing method is proposed as follows in this paper: On the one hand, in porous medium, both the penetrating space and solid frame are regarded as the fluid region, while, to describe solid boundary approximately, the viscosity of the solid skeleton structure is set to infinite. On the other hand, the whole region is discretized through the Cartesian orthogonal grid and the Immersed Boundary Method(IBM) is adopted to describe the smooth and complexed boundaries between the penetrating space and solid frame. In this paper, the proposed method stated above is conducted on OpenFOAM by the finite volume method in which the convection term is discreted in the QUICK scheme. After that, the corresponding systematic evaluations of the proposed method are performed and results illustrate that the method is of great accuracy and robustness. It can be used to study the flow in complex porous media.

Speaker: Ms. Shuo Wen (Beijing Institute of Petrochemical Engineering)
• 18:30
Optimization Method Research of Multi-stage Polymer Flooding 15m

Polymer flooding is an effective way to enhance the recovery rate of heavy oil reservoirs in high water-cut stage. To overcome the commonly existing defects of general polymer injection such as high costs and high transmissibility pathways. Multi-stage concentration polymer flooding and optimization methods are researched to clear out its influence on heterogeneous reservoir recovery enhancement. Different polymer injection experiments are performed on heterogeneous cores to study the polymer flooding injection order, permeability ratio, oil-water viscosity ratio on the influence of recovery. Numerical simulation experiments are also carried out on different heterogeneous models to research the influencing factors of polymer flooding to probe into the mechanism of polymer flooding enhancing oil recovery.
Results show that permeability ratio is an important factor affecting polymer flooding recovery. In low permeability ratio situations, the displacement effect is well, and as the permeability ratio increases the EOR effect goes down. Two stage concentration slug polymer flooding is better than single slug and three stage slug injecting in low and middle permeability ratio formations. Three stage concentration slug polymer flooding demonstrate a good EOR effect on high permeability ratio formations. When the oil viscosity and permeability ratio is low, three stage ploymer flooding effect is better than that of two stage polymer flooding, and as oil viscosity and permeability ratio increases, two stage polymer flooding exceeds the three stage flooding modes. Numerical simulation results show that the three stage concentration slug polymer flooding is better than single stage and two stage, which is accordance with the experimental rules.
Keywords: polymer flooding, multi-stage concentration slug, permeability ratio, viscosity, EOR

Speaker: Shuaishi Fu
• 17:15 18:45
Poster 2: Poster 2-H
• 17:15
Quantifying Dual Porosity Flow and Contaminant Transport Processes Using an Integrated Pore-Scale Network Modeling Approach 15m

Many tropic soils exhibit double- or even triple-porosity features as reflected by heterogeneous pore- and/or particle-size distributions. While the amount of clay in tropical soils is generally relatively high, cementation of the finer particles into larger grains make field soil often behave macroscopically more like coarse-textured media. Natural aggregation may further enhance preferential flow paths for water and contaminant transport. The same can be observed in some sedimentary rocks (coquinas) composed partially or entirely of transported, abraded and/or mechanically-sorted fragments of the shells of mollusks, trilobites, brachiopods or other invertebrates. Such coquinas often contain large interconnected pore networks that directly or indirectly influence fluid flow and contaminant transport processes. Analysis of their pore properties (such as their size, shape and connectivity) as estimated from three-dimensional images (3D) provides a way to link microscale pore structures with their macroscopic functioning, and how all this may affect overall fluid flow processes in such media. We evaluated the applicability of an integrated characterization approach involving 3D microtomography, measurements of the pore-size distribution (PSD) using 3D images of two tropical soils, evaluating the double porosity nature of undisturbed soil samples, and modelling flow and transport using the PoreFlow pore network model. The procedure combines the use of commercial software such as Avizo, with in-house software developed at Utrecht University (Netherlands) and the University of Rio de Janeiro (Brazil). Images were obtained with a benchtop X-ray microtomography system at spatial resolutions of 3.6, 6, 12, and 30 microns by varying the size of the samples from 0.5 cm to about 4 cm in diameter. Pore-size distributions obtained at each image resolution were fitted with lognormal distribution functions. Results showed that large pore sizes are better represented in low resolution images of relatively large samples, while proper characterization of the smaller pores require higher resolutions of by necessity smaller samples. We also analyzed the PSDs obtained after skeletonization of the samples having different sizes. The image-based approach was found to correlate well with PSDs measured using mercury intrusion porosimetry. Saturated hydraulic conductivity estimates obtained with the pore network model were further compared with theoretical values based on soil texture.

Speaker: Enno de Vries (Department of Earth Sciences, Utrecht University, Netherlands)
• 17:30
Characterization of modified nanoscale zero-valent iron particles transport through sandstones by nuclear magnetic resonance 15m

With the increasing application of nanomaterials for environmental remediation, modified nanoscale zero-valent iron (nZVI) particles have been extensively examined in terms of their enhanced mobility in porous media as compared to bare nZVI particles. However, the monitoring of nZVI particle transport processes in low permeability media is still a challenge. To quantify the particle transport behavior, low-field nuclear magnetic resonance (LF-NMR) was employed to image the modified nZVI particles through tight artificial sandstones, which is preferable in the laboratory research of fluids in rock with the limited influence of internal gradients compared to high-field NMR. The spin-echo single point imaging (SE-SPI) sequence of NMR was applied to monitor the fluid flow processes in porous media in terms of the changing site-specific transverse relaxation time (T2), which was available to capture the transient effects at a time scale of seconds. More importantly, the SE-SPI sequence allows the high resolution detecting of nanoparticle transports through nanoscale porous materials. Quantitative concentration profiles converted from T2 distribution mapping profiles were analyzed by CXTFIT software to estimate the transport parameters. A comparison of the parameters calculated by various points along the length of the sandstone at different time intervals indicated that the dispersion coefﬁcients, deposition rate constant, and collision efﬁciency decreased with time, whereas the fast deposition rate constant and average particle velocity increased with time. Meanwhile, the three-dimensional structure of sandstone, which was built using ORS Visual software with micro computed tomography (micro-CT) images, indicated the possibility of observing nanoparticles clogging in the pore throats. Accordingly, the modified nZVI particles exhibited better migration through porous media, which may result in their widespread commercial applications in the environment remediation.

Speaker: Qian Zhang
• 17:45
Fluid Flow Property Estimation Using a Pore Network Modeling Approach 15m

Pore network modeling is a technique that has been booming in recent years, and several authors have used it to obtain properties as absolute permeability, relative permeabilities and capillary pressures, which are common obtained from laboratory tests and/or experimental correlations. The scope of this work is to model flow and immiscible displacement and estimate fluid flow properties such as, absolute permeability and capillary pressure curve, using a systematic methodology. With this in mind, the workflow begins choosing an open access carbonate’s micro-CT image and its extracted network. The digital sample is similar to a rock sample used in a primary drainage test. From the available information, a statistical analysis to explore the network’s topological properties and the medium’s geometric properties is performed. This analysis will allow us to identify and propose spatial dependencies between some properties of the network elements. Then, through multiple realizations, equivalent networks are generated using OpenPNM, an open source pore-network modeling project. Subsequently, to simulate flow and primary drainage same conditions as those of the laboratory test are taken into account. The Hagen-Poiseuille model and invasion percolation with trapping are considered, for flow and primary drainage respectively. In primary drainage process, statistically equivalent pore network realizations give rise to a family of capillary pressure curves that comprises a reliability window, i.e. a value range that capillary pressure can take for the medium under study. Finally, a capillary pressure model is fitted and the confidence intervals are validated.
This methodology is validated in a case of study for pore network modeling of a carbonate rock sample.

Speaker: Edgar G. Martínez-Mendoza (Universidad Nacional Autónoma de México)
• 18:00
Reactive Transport Modelling on the Drill Core Scale, Parameterized by GeoPET/µCT Process Tomography 15m

In-situ leaching of ores is considered as an economic and environmentally friendly production method. However, the leaching process is complicated by its dependence on the material’s heterogeneity and by retroactive effects over large scales. We developed an experimental procedure which is based on positron emission tomography (PET) during transport experiments with radiotracers (Kulenkampff et al. 2016), supported by µCT, to derive flow and diffusion parameters with molecular sensitivity (picomolar) and with reasonable resolution (1 mm) on samples with drill core dimensions. The procedure directly yields the spatial distributions of flow velocity and effective volume from flow experiments, the diffusion coefficient, and the real geometry of the sample. This approach is validated by using a core sample (L = 10 cm, D= 6 cm) with an induced fracture, obtained from Permian Kupferschiefer sandstone. After the tomographic measurements the core sample was leached stepwise with (1) fresh water to remove salt minerals, (2) acidic solution (H2SO4, pH 1.5) to reduce the carbonate content, and (3) acidic solution with added ferric iron to dissolve the Cu-sulfidic ore.
The measured hydrodynamic and structural parameters from PET and µCT were then directly imported into a reactive transport model using interface Comsol Multiphysics with Phreeqc (iCP, Nardi et al. 2014) to simulate core sample leaching. The geochemical conditions of the simulation were considered consistent to the laboratory leaching experiment. The reactive transport simulation is based on the real geometry of the sample and the observed flow fields on the continuum scale and does not require a high performance computation of flow simulations on pore scale. The results of these economical simulations are compared to the results of the laboratory leaching experiments.

References
J. Kulenkampff, M. Gründig, A. Zakhnini and J. Lippmann-Pipke. Geoscientific process monitoring with positron emission tomography (GeoPET), Solid Earth, 7, 1217-1231, ( 2016).
A. Nardi, A. Idiart, P. Trinchero, L.M. de Vries and J. Molinero. Interface COMSOL-PHREEQC (iCP), an efficient numerical framework for the solution of coupled multiphysics and geochemistry. Computers & Geosciences 69, 10-21, (2014)

Speaker: Cornelius Fischer (HZDR)
• 18:15
Unsaturated porous medium effective diffusion coefficient calculation through lattice Boltzmann method 15m

In the context of radioactive waste management, deep geological repository in indurated argillaceous media appears as a potential solution. Radionuclides transport through argillaceous media is then of concern and in particular via diffusion processes in pore water. The Excavated Damaged Zone (EDZ) surrounding the storage cells and the galleries of the repository presents a complex network of micro- and macro-fractures where radionuclides transport may be enhanced with respect to the diffusion in intact host rock. Especially, the micro-fractures are supposed to form a connected pathway and thus the diffusion of solutes in this network is of high interest in the context of the storage safety assessment.
During the transitory period after the excavation of the repository, the fractures undergo several desaturation-resaturation cycles resulting from desaturation process during the excavation phase or from hydrogen flow where hydrogen is produced from corrosion of the waste metallic canisters after repository closure. Their state can vary from fully saturated with water to almost complete desaturation depending on their location, aperture and on the instant in the history of the repository. The saturation state of porous media has an important impact on the diffusion processes as it modifies the connectivity and the tortuosity of residual pore water.
In order to simulate diffusion process inside porous geometry, we chose to use a Lattice Boltzmann Model (LBM) that allows to i) to easily take into account the geometry as available from X-ray tomographic images or reconstructed through mathematical models, ii) simulate water-gas distribution inside the porous media for different saturation level, and iii) simulate diffusion inside the resulting connected water pathway.
If the presented work was initiated focusing on radionuclide diffusion through unsaturated micro-fractures, diffusion process in unsaturated porous media is of concern in many other fields like pesticides or carbonated nutriments behavior in soils or solute migration in reservoirs. We present here the application of our technique in several kind of porous media, namely on an argillite micro-fracture, a soil macro-pore and a generic reservoir representative elementary volume.
The LBM we used in this work is based on a Two-Relaxation-Time (TRT) collision operator. Water-gas distribution was simulated using the LBM described in [1] which follow the Shen-Chen approach through a particular source term. At initial time, a mean density was imposed inside the void space of the porous media. The presence of the wetting walls induces spontaneous phase separation leading to a low density phase (gas) and a higher density phase (water). On the basis of the calculated liquid-gas distributions inside the pore space, a basic thresholding algorithm allowed to extract the connected liquid phase. For the diffusion computations, we considered a non-volatile tracer and only took into account diffusion in the liquid phase. We conducted diffusion simulations for different saturation in the connected liquid phase. Computations were performed starting from an initial Dirac imposed concentration on the symmetry line of the media. The mean concentration was calculated as a function of distance to the central line and then fitted with an analytical diffusion equation candidate curve. Effective diffusion coefficient was deduced as a function of saturation.

Speaker: Dr. Alain Genty (CEA)
• 18:30
A New Dynamic Single-Pressure Network Model: Experimental Comparisons and Calibrations. 15m

We present a simple single-pressure dynamic network simulator for two-phase flow in porous media with a focus on exploring the limits of a single-pressure network model.

Our work builds upon the work of Aker et al.[1] and Knudsen et al.[2], where we aim to more accurately describe and model the interactions of ganglia moving through the porous material. Whereas the previous models assumed constant cross-section pipes, we instead account for a varying cross-section along a pipe. This is important as the size of trapped wetting bubbles is determined by both the throat size and the pore size, and are over-predicted for straight pipes.

Another focus is on accurately modelling the interaction of ganglions across adjacent pipes. This is the dominant mechanism that determines the ganglion size distribution in the network - a ganglion cannot split below the size of a pore for sufficiently slow flow rates.

Our model is of the single-pressure kind, where only one phase can occupy a cross-section at a time. Consequently we neglect wetting fluids in films and corners. One aim of our study is to investigate the limits of a simple single-pressure network model without resorting to a much computationally heavier two-pressure model[3]. This excludes us from modeling strong imbibition invasions where the main transport mechanism is dominated by film and corner flow.

We will present direct comparisons to drainage and mixed-wet invasion studies by Zhu et al.[4], and to experiments by Moura et al.[5,6] studying the burst dynamics of an invasion process. In addition we will present some qualitative comparisons to the experimental work of Avraam and Payatakes[7] on the dynamics of ganglions.

Speaker: Mr. Morten Vassvik (Department of Physics, Norwegian University of Science and Technology)
• 18:30
A random connection model for pore network modeling 15m

Pore network models have been applied for predicting petrophysical properties at pore scale. From a geometry point of view, basically a pore network and pore and throat size distributions are required for pore network modeling. Although different pore network models have been constructed using data extracted mainly from images, it is not always possible to count on the necessary information, and working with a unique extracted network could generate non-representative results. Therefore, a statistical analysis of the data offers the advantage of generating different realizations of the network and its geometry. Building statistically representative networks require analysis of the image to extract size distributions of pores, throats and their connectivity. The last is relevant for percolation properties of the pore system.
In this work a random connection model (RCM) for network modeling is proposed. In a random connection model the critical density is a connection function g. A connection function defined as g: Rd→ [0, 1] is chosen, where a pair of points (x,y) are connected with probability g(|x - y|), independently of all other pairs of points, here |.| denotes Euclidean distance in Rd. In general, if unbounded connected components arise, we say that percolation occurs.
The RCM is applied to a case of study for pore network modeling of a carbonate rock sample. In particular, from multiple realizations of the RCM the effect on fluid flow properties is investigated.

Speaker: Ana T. Mendoza-Rosas (CONACYT-Centro de Ingeniería y Desarrollo Industrial)
• 18:30
Droplet Flow Regimes in a T-Section Microchannel: Assessment of Volume of Fluid Formulations 15m

Keywords:
Droplet flow regimes, Interfacial dynamics, Volume of Fluid method, Parasitic currents, Direct Numerical Simulations

Abstract:
Multiphase flow in microfluidic devices that produce identical droplets/ bubbles at a known frequency has gained attention due to its use in biomedical, chemical and engineering applications.

In this numerical study, we investigate the flow dynamics in a 2D ‘T’ shaped microfluidic channel where two fluids (wetting and non-wetting) are injected orthogonally and meet at a junction. We use Direct Numerical Simulations (DNS) with the Volume of Fluid (VoF) method where the fluids are distinguished based on a volumetric colour function. DNS-VoF captures the topological changes (ex: pinch-off of an interface) automatically, unlike other methods (ex: Front-tracking) which require manual interventions.
However, Parasitic currents (PC) around the interface potentially occur due to inaccurate computations of the capillary force $\textbf{F}_\sigma$. The Continuum Surface Force (CSF) [1] expresses $\textbf{F}_\sigma$ in a simplistic manner. However, the capillary terms are imprecise and show larger PC. To reduce PC, several formulations are available, such as Sharp Surface Force (SSF) [2, 3] and Filtered Surface Force (FSF) [3]. These formulations advect the interface numerically while trying to retain a sharp interface. Alternatively, Piecewise Linear Interface Construction (PLIC) [4], geometrically constructs and advects the interface. Height functions are used to increase the order of accuracy in computing $\textbf{F}_\sigma$.

Though a number of DNS-VoF validations with experiments and analytical models are available [5], a systematic investigation of the potential impact of PC on the droplet formation (possibilities of initiating a premature break-up or skipping pinch-off) and on the droplet length for a wide range of capillary number are still not available. Indeed, our numerical simulations show different flow regimes in the microchannel using the above mentioned VoF formulations for the same flow conditions. Further, when there is a pinch-off, the length of the droplet formed also varies for different VoF formulations. To validate the accuracy of the numerical solvers we make use of a solution that can be derived using geometrical and force balancing techniques.

Speaker: Dr. Florian Doster (Heriot-Watt University)
• 18:30
Evaluation of a Proposed Workflow for Digital Petrophysics of Coquinas Involving Experimental Data and 3D Digital Models Using PNM- and FEM-Based Simulations 15m

Heterogeneous carbonate rocks, especially coquinas, present several challenges regarding their morphological characterization, petrophysical parameterization, and obtained a more complete understanding of their fluid flow properties. Within this context, a workflow is proposed and analyzed for digital petrophysics using a combination of experimental data and numerical 3D fluid flow simulations. The proposed workflow is applied to coquinas (Brazilian Pre-Salt analogue carbonates) in efforts to evaluate their petrophysical properties.

The workflow involves basic petrophysics as well as state-of-the-art approaches in digital analyses, to characterize coquina samples. Outcrop coquinas samples were for this purpose selected from the Morro do Chaves Formation in Northeastern Brazil, which has been studied as analogue rocks of Santos Basin Pre-Salt carbonates. Important steps in the workflow are the geological and morphological analysis of the selected coquinas samples (including basic petrography), acquisition of data through X-ray computed microtomography using different spatial resolutions, three-dimensional reconstruction and modeling of the coquinas and their pore systems, and numerical simulation of fluid flow in the porous media. Experimental data were obtained using permeametry, porosimetry, NMR, and SEM in support of several key steps of the workflow, such as segmentation procedures. Numerical simulations were used to estimate absolute and relative permeabilities using PNM- and FEM-based approaches. We used for this purpose commercial and in-house codes developed at both Utrecht University (Netherlands) and the University of Rio de Janeiro (Brazil).

All digital and numerical steps within the proposed workflow are being validated against experimental data to understand the limitations and uncertainties of the various steps. We highlight the main challenges and uncertainties encountered during each step, including challenges related to proper estimation of the microporosity of coquinas, and appropriate definition of the REV (Representative Elementary Volumes) based on mathematical and statistical analyses. The workflow identified several important issues related to the digital petrophysics experiments of the coquina samples, including major challenges to properly characterize the porosity and related PNM- and FEM-based fluid flow simulations. Conclusions from this study will enable improved tailoring and optimizations of digital petrophysics predictions of Brazilian Pre-Salt carbonate reservoirs, based on a well-structured workflow that is self-correcting when combined with a range of experimental data.

Speaker: Enno de Vries (Department of Earth Sciences, Utrecht University)
• 18:30
Multiscale Modeling of Particle Transport and Retention in Porous Media 15m

Multiscale Modeling of Particle Transport and Retention in Porous Media

John Blears and Karsten Thompson

Craft & Hawkins Department of Petroleum Engineering
Louisiana State University

December 4, 2017

Abstract
Modeling particle transport and retention in porous media is important in fields such as hydrocarbon extraction, groundwater filtration, and membrane separation. While the continuum-scale is usually of practical interest, pore-scale dynamics govern the movement and retention of particles. Therefore, accurate modeling of continuum-scale behavior requires an effective incorporation of pore-scale dynamics. Due to current computational limits however, the large spatial and temporal discrepancies of these scales prohibit modeling an entire continuum-scale system as a single pore-scale model. Even if a pore-scale model could incorporate every pore contained in a continuum-scale system, an upscaling scheme that coupled pore-scale and continuum-scale models would likely be more efficient and achieve acceptable accuracy. This study focuses on initial work performed to develop a numerical continuum-scale model for particle transport and retention that uses a novel pore-scale coupling technique. In the proposed model, continuum-scale grid blocks are embedded with pore network models (PNMs). Simulations progress sequentially as continuum-scale grid blocks provide their embedded PNMs with pressure gradient boundary conditions and influent particle concentration information. The PNMs then use a Lagrangian particle tracking method to identify a continuum-scale particle retention coefficient and determine any changes in permeability and porosity due to particle trapping. This information from the PNMs can be used to update the petrophysical properties of the continuum-scale grid blocks. Continuum-scale flow may then be simulated using the updated information. Coupling in this manner has the potential to allow for an understanding of how particle induced changes at the pore-scale will impact continuum-scale behavior spatially and evolve temporally.

Keywords: Particle transport and retention, multiscale modeling, numerical modeling, pore network modeling

Speaker: Jack Blears (Louisiana State University)
• 18:30
Toward direct pore-scale reactive modelling of low-salinity flooding in 2D/3D porous media images 15m

We present a two-phase multicomponent reactive pore-scale model based on Direct Numerical Simulation of the Navier-Stokes equations using the Volume-Of-Fluid and the Continuous Species Transfer methods. In order to study wettability change during low-salinity flooding, simple upscaling rules from the nano-scale to the pore-scale are integrated to the model to describe (1) change of local contact angle with change of electrical double layer potential, (2) effect of ionic transport and electrical field in the water thin film and (3) ganglion remobilization due to build-up of osmotic pressure in the thin film. We show that our model can be fitted to match experimental results of oil recovery by low-salinity effect. We then investigate the impact of each mechanism on the increase of oil recovery during low-salinity flooding in 2D/3D porous media images.

Speaker: Julien Maes (Heriot-Watt University)
• Wednesday, 16 May
• 09:30 10:43
Parallel 6-A
• 09:32
Pore shape evolution in different transport regimes: Single-pore scale simulations in carbonate rocks 15m

Calcite is the main mineral found in carbonate rocks, which form significant hydrocarbon reservoirs and subsurface repositories for CO2 sequestration. Major processes that occur upon injection of CO2 in carbonates are the dissolution of calcite and modification of the pore-space structures. A number of pore network models acknowledge this dissolution-induced geometry change in the form of a relation between pore throat radius and the changing pore volume (Nogues et. al., 2013). However, this relation is considered to be independent of the transport and reaction regimes, in spite of observations at the continuum of different dissolution features being controlled by flow velocity and reaction rates (Luquot et. al., 2014).

To simulate the processes accompanied by CO2 injection in a single pore of calcite, we used COMSOL Multiphysics. The fluid velocity was within the range of creeping flow, so we solved Stokes equation. The chemistry was defined by coupled equilibrium and kinetic reactions. The reactions related to CO2 dissolution in water, the formation of carbonic acid and dissociation of carbonic acid are assumed to be in the bulk and fast enough to be in equilibrium at each time step. These solution species diffuse towards the surface of calcite and cause calcite dissolution. This dissolution results in a moving boundary of calcite. The arbitrary-Lagrangian Eulerian (ALE) approach is used to track the solid-fluid interface at each time step. The transport of bulk species is solved by the advection-diffusion-reaction equation.

We used three average flow velocities as boundary condition, to cover diffusion dominated and advection-dominated transport regimes. Each flow velocity yields a typical pore shape during the initial, transient stage of the simulation. We observed that this pore shape is a result of the competition between advection of reaction products along the pore length and diffusion-led transversal mixing. In the transient stage, these processes define a flow-regime specific pore shape that is subsequently observed to be quasi-steady state. The diffusion-dominated flow is witnessed to produce the non-uniform shapes because the fluid becomes less reactive along the pore length. However, for the advection-dominated flow, the fast velocity keeps the fluid relatively more reactive towards the end of the pore length, thus yielding uniform pore shapes. To summarize, different transport regimes affect pore-shape evolution differently during the initial, transient stage of fluid flow. The uniform or non-uniform pore shapes are inherited, remaining unaffected during quasi-steady state fluid flow. Since uniformity of a pore impacts the pressure drop profile along the pore, our observations at the transient-stage suggest that pore-shape and conductivity evolution cannot be assumed to be independent of flow regime. Future work involves validating these observations experimentally and implementing the conductivity and volume change in such a way that our shape-factor information can be included in pore network modelling.

Speaker: Ms. Priyanka Agrawal (Department of Earth Sciences, Utrecht University, Utrecht, Netherlands, )
• 09:50
Upscaling of two-phase flow in porous media with free boundary at the pore scale. 15m

Reactive flows and transport models through the porous medium are important for a wide range of scientific and industrial processes. Examples in this sense are groundwater remediation, oil recovery from reservoirs, CO2 sequestration etc. The main goal of the research is to develop mathematical models that describe such processes at the pore scale and to derive effective models at the macro (Darcy) scale through analytical upscaling. A particular feature of the models addressed here is the occurrence of freely moving interfaces separating different phases, like fluid-fluid (two-phase/unsaturated flow) or solid-fluid (one-phase flow with dissolution/precipitation) at the pore scale. The focus of the research is to give a rational derivation of the upscaled models, which are not only less complex to describe, but also very efficient to simulate. To this aim, homogenization theory based on asymptotic expansion is applied.

Speaker: Mrs. Sohely Sharmin
• 10:08
Upscaling of coupled geomechanics, flow, and heat, in a poroelastic medium in the quasi-static situation 15m

Motivated by geothermal energy storage in the subsurface, we undertake a formal derivation of a linear poro-thermo-elastic system within the quasi-static framework. This derivation is based upon the well known derivation of the quasi-static poroelastic equations (also known as the Biot consolidation model) from the micro structure, except that we now include energy conservation equations in the micro-scale model. These are coupled to the fluid/structure model by using linear thermo-elasticity for the solid structure instead of the usual linear elasticity. The resulting upscaled system is similar to the linear poro-elastic equations, but with an added conservation of energy equation, fully coupled to the momentum and mass conservation equations. We start at the pore scale, and apply the technique of homogenization to derive the upscaled model in the case of periodically distributed pores. Assuming the homogenization ansatz holds true, we obtain a fully coupled system of equations on the macro-scale accounting for the effects of geomechanics, heat transfer, and fluid flow within a fully saturated porous material.

Speaker: Mats Brun (University of Bergen)
• 10:26
Coupled Electro-hydrodynamic Transport in a Geological Fracture 15m

Fractures are very common features in subsurface crystalline rocks, where they are organized in networks of interconnected elements. A number of essential mechanical properties of the rock formations, such as their mechanical strength and their transport properties (hydraulic and electrical conductivities), are dictated by the behavior of the fracture networks. Within these networks, individual geological fractures are the basic structural unit controlling the flow of fluids and the transport of solute chemical species. Fracture wall roughness is responsible for flow channeling (and therefore, heterogeneity) within the fracture plane, which, at the fracture scale, impacts the fracture’s transmissivitty [1,2]. The characteristic length scale $L_c$ at which the two fracture walls are matched [3], plays a crucial role as it is the upper limit scale for flow heterogeneities [2].

The most prevalent way of computing the transport properties and transmissivitty of a rough fracture in an efficient way and without resorting to a full three-dimensional flow simulation, is to use the lubrication approximation, which leads to a Darcy flow type equation for the pressure, the Reynolds equation [1]. This method has been used extensively to simulate the flow, as well as the electric current (without flow) through a rough fracture [5]. However, the effect of the electrical properties of the fracture walls on the transport properties of a fracture still remains an open question, to the best of our knowledge. Since dissolved minerals and salts are ever present in the fluids inside the fracture, Electrical Double Layers (EDL) almost inevitably form at the fluid-solid interface [6], and their strength depends on the chemical properties of the rock and ionic strength of the fluid. Therefore, the ocurrence, at the fracture scale, of externally-imposed or naturally-occurring gradients in electrical potential and/or ionic concentration, can lead to significant changes in the fluid motion through the fracture, as compared to flows driven primarily by hydraulic head differences.

In this work, we attempt to explore the flow dynamics that result from such coupled electro-hydrodynamic forcings. To this end, we generalize the standard lubrication theory for flow to include a description of the coupled transport of fluid mass, solutes, and electrical current under application of fixed differences in hydraulic head (or pressure), electrical potential and concentration across the fracture. By invoking the requirement of conservation of volumetric flow rate, ions and electrical charge fluxes, a coupled system of equations can be derived, which governs the spatial distribution of electrical potential, pressure and concentration in the bulk fluid within the fracture. This system of equations is the generalization of the Reynolds equation to the coupled transport of fluid mass, solutes, and electrical charges. It is solved using an iterative Finite Volume Method to gain insight into the dynamics of the coupled transport processes, in geological fractures with a realistic aperture field. We investigate in particular the role of the characteristic length scale $L_c$.

Speaker: Dr. Uddipta Ghosh (ITT Gandhinagar, Departement of Mechanical Engineering)
• 09:30 10:43
Parallel 6-B
• 09:30
Highspeed Imaging of Frictional Fracturing in Deformable Porous Media 15m

The invasion of gas via fractures into a fluid-saturated deformable porous media is a process that occurs in numerous geoengineering and natural circumstances: fracturing for enhanced contaminant remediation, stimulation of hydrocarbon reservoirs, methane venting in sediments, and outgassing of crystal-rich magmas for example [1-6]. However, it is challenging to characterise and predict complex flows through deformable fractured media. We use image analysis on a simple experimental system that produces intriguing maze-like fracture patterns to illuminate these processes and develop a model to predict pattern properties.

A wet unconsolidated granular packing confined in a Hele-Shaw cell is subject to a slow invasion of gas, causing fractures. We study the influence of granular properties (size/shape), and gas injection rate on the system. We image the complete fracture pattern as it forms, using photography, highspeed imaging, and PIV (particle image velocimetry), allowing us to investigate the dynamics of its growth and the properties of the resulting pattern.

We find that fractures are triggered by initial pore invasion into a compacted front, followed by fluidization of previously undisturbed grains resulting in fracture growth. The fractures grow in an intermittent, stick-slip manner, interspersed with inactive periods. Growth is impeded by friction from local compaction fronts that form around the growing fracture branches, causing a spatial frequency in the pattern as it fills the space of the cell by a process of local cooperative fracturing. We describe a simple analytical model that predicts the fracture density from basic granular medium properties and demonstrate a limit of maximum flow rate after which the fracture dynamics transition to viscous fingering.

Speaker: Deren Ozturk (Swansea University)
• 09:48
Pattern formation of frictional fingers in a gravitational potential 15m

Aligned finger structures, with a characteristic width, emerge during the slow drainage of a liquid/granular mixture in a tilted Hele-Shaw cell. A transition from vertical to horizontal alignment of the finger structures is observed as the tilting angle and the granular density are varied. The dynamics is reproduced in simulations. We also show how the system may explains patterns observed in nature, created during the early stages of a dyke formation.

Speaker: Prof. Knut Jørgen Måløy (PoreLab, Department of Physics, University of Oslo)
• 10:06
How to sink: impact of fluids in soil liquefaction during earthquakes, computation of critical acceleration. 15m

Soil liquefaction is a significant natural hazard associated with earthquakes. Some of its devastating effects include tilting and sinking of buildings and bridges, and destruction of pipelines. Conventional geotechnical engineering assumes liquefaction occurs via elevated pore pressure. This assumption guides construction for seismically hazardous locations, yet evidence suggests that liquefaction strikes also under currently unpredicted conditions. We show, using theory, simulations and experiments, another mechanism for liquefaction in saturated soils, without high pore fluid pressure and without special soils, whereby liquefaction is controlled by buoyancy forces. This new mechanism enlarges the window of conditions under which liquefaction is predicted to occur, and may explain previously not
understood cases such as liquefaction in well-compacted soils, under drained conditions, repeated liquefaction cases, far-field liquefaction and the basics of sinking in quicksand. We next introduce viscous forces between grains and fluids, and examine how they modify the dynamics once liquefaction has been triggered.
These results may greatly impact hazard assessment and mitigation in seismically active areas.

Speaker: Prof. Renaud Toussaint (CNRS, IPGS UMR 7516, University of Strasbourg)
• 10:24
Hydrodynamic dispersion in porous media: Macroscopic description derived from statistical physics. 15m

At the pore level hydrodynamic dispersion is simply the transport of a miscible substance due
to advection and molecular diffusion. At the level of the representative elementary volume, however,
the dispersion is described by a dispersion tensor which may give a flux in one direction due to a concentration gradient in a different direction. By applying Onsager theory we show that this tensor is symmetric under reversal of the background velocity field in the porous medium. This symmetry is general and we demonstrate its relevance by means of simple lattice Boltzmann simulations. In doing this we show how the cross-coupling between different spatial directions result from heterogeneities in the flow field and from this we introduce a generalization of the conventional expression
for the dispersion tensor. The cornerstone of the theoretical development is the observation that the macroscopic dynamics derives from an underlying meso-scale dynamics, that is time-reversible. This principle is generally applicable to more complex situations too, in particular those where the fluids couple to a deformable medium where the deformations are reversible, such as in elastic systems.

Speaker: Prof. Eirik Flekkøy (PoreLab, University of Oslo, Norway)
• 11:17 12:44
Parallel 7-A
• 11:17
Impact of surface complexation and electrostatic interactions during pH fronts propagation in silica porous media: Experiments and model based interpretation 15m

Propagation of proton fronts exerts a fundamental control on geochemical processes and contaminant transport in subsurface systems (Muniruzzaman & Rolle, 2015). Protons are of key importance in pore water solutions since they affect the sorption behavior of charged contaminants by setting the surface charge at the mineral-liquid interface. Therefore, it is of primary importance to enhance our understanding regarding proton transport and complexation on mineral surfaces in order to accurately describe and predict reactive transport processes in porous media. Although significant effort, based on experimental studies and model development, has been dedicated to investigate pH sorption and surface complexation behavior in batch systems, only a few studies addressed the impact of these processes under flow-through conditions (McNeece & Hesse, 2016).

In this work, we systematically investigate the transport and sorption behavior of proton under flow-through conditions by means of laboratory experiments and reactive transport modeling. We performed experiments both in batch (titration) and in flow-through setups packed with saturated silica media (i.e. quartz sand and glass beads) with different grain sizes as porous matrix. Batch experiments were focused on characterizing and obtaining insights on the pH sorption capacities of the porous matrix and on the electrostatic interactions at the mineral pore water interface. Successively, we performed flow-through experiments in laboratory columns by applying continuous injections of acidic solutions. Breakthrough curves of pH and all other major ions were measured at the outlet of the columns. In order to identify the importance of ionic strengths on charged species at the liquid-solid interface and on the multicomponent ionic transport, we performed parallel sets of analogous experiments (both in batch and column setups) with different background electrolyte solutions.

To quantitatively interpret the experimental results, we used a reactive transport model explicitly taking into account the triple-layer charge distribution multisite surface complexation (CD-MUSIC) model (Hiemstra & Van Riemsdijk, 1996) together with the Donnan approximation (Appelo & Wersin, 2007), and the cross-coupling of dispersive fluxes due to the Coulumbic interactions between aqueous charged species (Rolle, et al., 2013). The modeling approach was based on a coupling between the geochemical code PHREEQC with MATLAB using IPhreeqc module (Muniruzzaman & Rolle, 2016). The relative affinity constants of the surface complexation reactions were obtained through inverse modeling of the batch-titration and flow-through experiments. Experimental and modeling results suggest that the sorption behavior of protons differs in the considered silica porous media and are differently affected by the ionic strength of the background solutions.

Speaker: Lucien Stolze
• 11:35
Effects of Groundwater Circulation Well to contaminant Back-Diffusion from low-permeability layers: investigation by laboratory test and numerical simulations. 15m

The presence of contaminants in low permeability zones of aquifer can represent a real limitation for a complete and effective groundwater restoration. When dissolved plume encounter low permeability layers, concentration gradient between low and high-permeability zones determines storage of dissolved pollutants into the lower permeability layers by molecular diffusion (Forward-Diffusion). After the end of the plume passage there is an inversion of the gradient direction that leads to a slow re-distribution of contaminant from the lower permeability zones back to the higher permeability zones (Back-Diffusion). The low-permeability zones become therefore secondary contamination sources that cause long plume tail.
The aim of this study was to evaluate by laboratory test and by numerical simulations the suitability of Groundwater Circulation Well (GCW) to restore contaminated low permeability zones of aquifer. GCW is a well characterized by a different number of screens with extraction induced from one screen and injection from another. The use of GCW develops a circulating flow field near the well, increasing the vertical component of groundwater flow.
A sand aquifer with two low-permeability lenses was reconstructed inside a tank containing a GCW model. The lenses were saturated with a known quantity of tracer and a circulating flow field was generated inside the aquifer injecting clean water from the upper GCW screen and extracting contaminated water from the lower screen. During the test, images of the box model were acquired and using an image analysis procedure the tracer mass released by the two lenses was estimated.
A numerical model was developed to reproduce the Back-Diffusion process and to investigate effects of pumping technologies to contaminant redistribution process from low to high permeability zones of aquifer. The model was validated comparing the numerical results with those obtained experimentally by laboratory test. Numerical simulations were carried out to evaluate effects of innovative GCW technology and traditional Pump and Treat system on the Back-Diffusion process. To achieve the goal, numerical tests were performed considering various injection/extraction water flow rates and different features of polluted low permeability layers.
Results demonstrate the more suitability of the GCW technology to restore contaminated low permeability zones than the traditional Pump and Treat system. However, the efficiency of GCW appear to depend on features of low-permeability layers, as their geometry and their position inside the aquifer.

Speaker: Dr. Fabio Tatti (Department of Civil, Building and Environmental Engineering (DICEA), University of Rome “La Sapienza”, Rome, Italy)
• 11:53
Colloidal transport in a microfluidic porous medium with surface charge heterogeneity 15m

Colloidal facilitated transport of contaminants is a major concern for transport of low solubility chemicals in ground water flows. Compared with themselves alone, contaminants travel much further after adsorbing to natural colloidal particles, as far as kilometers over years. Therefore, understanding the transport behavior of colloids would provide insightful knowledge for environmental protection and remedy. In the past most lab-scale experiments focused on studying the interactions between particles and homogeneous or physically heterogeneous porous media. In reality, porous media exhibit different chemical properties at the grain scale. To shed some light in this aspect, we built a family of pseudo-three-dimensional microfluidic porous media models packed by 10 µm polystyrene (PS) beads with opposite surface charges. Each porous medium contains more than 6000 beads and has a size of 1000 µm in length, 600 µm in width and 15 µm in height. An on-chip cross channel linked with an off-chip 4-way valve was applied to generate a step input of concentrated colloidal suspension, which also greatly reduced the influence of Taylor dispersion of colloidal particles. We further studied the effect of chemical heterogeneity on the transport and retention of colloidal particles. The breakthrough of the particles was recorded at the downstream of the porous medium. The colloidal retention profile was mapped using a confocal microscope. Beads with different surface properties are distinguished by their different fluorescent labels. Therefore we also obtained colloidal retention data for beads with different surface properties. With our platform, we developed a correlation between the overall collector efficiency and its constitutive collector efficiency under different salt conditions.

Speaker: Yang Guo (Colorado School of Mines)
• 12:11
A New High-fidelity Mesh Model for Simulation of Transport Process in a Fixed Bed Reactor 15m

The traditional fixed-bed reactor design is ususally based on empirical correlations of plug flow pattern. This empirical method is usually not suitable for the low tube-to-particle diameter ratios (N=D/d< 4) where the local phenomena of channeling near the wall and the backflow in the bed are dominant. The recent “solid particle” method1 is too complicated for mesh generation, especially for large random packed beds, which seriously hinders its development. In this work, a novel mesh model is proposed and used for the simulation of fixed bed reactors by combining discrete element method with the user defined subroutine. The mesh generation process is simple and easy to be implemented, which ingeniously avoids handling the complex “contact point” problem. In this study the packed beds with spherical and cylindrical particles are investigated, the local flow in the bed can be high fidelity. The predictions of the pressure drop across the fixed bed and the heat transfer of the single particle are in good agreement with the corresponding empirical relations.

Acknowledgement
Financial support from the National Key Research and Development Program (2016YFB0301702), National Natural Science Foundation of China (21490584, 91534105, Major National Scientific Instrument Development Project (21427814) and Key Research Program of Frontier Sciences of CAS (QYZDJ-SSW-JSC030) is gratefully acknowledged.

Speaker: Mr. Bing Yuan (College of Chemical Engineering, Sichuan University, Chengdu 610065, China;CAS Key Laboratory of Green Process and Engineering, Institute of Process Engineering, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Beijing 100190, China)
• 12:29
Simulation of particle straining in porous media using a coupled pore-network and CFD-DEM model 15m

Geometrical straining of particles in porous media is of critical importance in a broad range of natural and industrial settings, such as the contaminants transport in aquifers and the permeability decline due to pore plugging in oil reservoirs. Despite its importance, relatively few studies have been performed on particle straining under fluid-driven flows in porous media. Pore-network modeling is attractive option for predicting particle straining. However, network models often lack predictive capability due to simplification of pore-geometry and require adjustable parameters. This is especially true for particle straining. Direct numerical simulation using computational fluid dynamics (CFD) to model the fluid phase, coupled with the discrete element method (DEM) to model the particle transport is a more rigorous approach but computationally expensive. The CFD-DEM simulations are limited to systems with particle number less than 105.
Our current research has coupled a pore network model with a CFD-DEM model in a computationally-efficient framework. Particle jamming is a matter of probability. We perform CFD-DEM simulations on particle filtration by a single layer of grains and formulate the jamming probability as a function of particle/pore size ratio, particle concentration, pore throat geometry etc. The upscaled results are then implemented into the pore-network model. During a time step, the jamming probability of each pore throat is calculated based on the local particle concentration, flow rate etc. The hydraulic conductivities of pore throats are updated dynamically. The numerical results are compared to direct CFD-DEM and experiment results and good agreement is achieved.

Speaker: Mr. Hongtao Yang (The University of Texas at Austin)
• 11:17 12:47
Parallel 7-B
• 11:17
Thermal conductivity predictions for porous materials via effective medium approximations and cross-property relations 15m

The effective thermal conductivity of porous materials is determined by all details of their microstructure. Since lower bounds (both Wiener and Hashin-Shtrikman bounds) are not available for porous materials (with vacuous voids), all predictions based on the porosity alone are necessarily model-based and thus tentative. In this contribution we first recall the exact solution of the single-inclusion problem for spherical and spheroidal pores [1], give a comprehensive summary of admissible nonlinear model relations (Maxwell-Eucken relation, Coble-Kingery relation, power-law relation, our exponential relation) [2] and explicitly exclude those model relations that are either redundant (self-consistent / Landauer-Bruggemann model), non-admissible (Spriggs‘ exponential relation) or useless (minimum solid area models) [3]. Further it is shown how the exact solution for spheroidal pores (oblate or prolate) is to be implemented into the admissible nonlinear effective medium approximations [1]. In the second part of this contribution we show that the problem of characterizing microstructural details and implementing microstructural information beyond volume fractions can be circumvented via cross-property relations (CPRs). In particular, the knowledge of the relative tensile modulus (Young’s modulus) can be used to predict the relative thermal conductivity of porous materials. The CPRs currently available for this purpose are recalled, including the Sevostianov-Kováčik-Simančík CPR [4], our CPR for isometric pores [5] and the recently proposed generalized version of the latter for anisometric pores (spheroidal-prolate and spheroidal-oblate) [6]. Using numerical modeling on a wide range of different computer-generated digital microstructures (convex pores, concave pores, spheroidal pores, foams) it is shown that our CPR provide the best thermal conductivity predictions currently available.

Speaker: Willi Pabst (University of Chemistry and Technology, Prague)
• 11:35
Influence of RVE generation method on effective heat conduction modeling in open-cell ceramic foams: Review and recent advances 15m

Ceramic foams play a crucial role in thermal engineering, as their porosity confers interesting properties which, combined with their high-temperature stability, allows them to meet both heat exchange and heat insulation demands. To facilitate optimization of the pore-scale morphology, numerous techniques have been developed to generate periodic representative volume elements (RVEs) of the foam mesostructure. As a fascinating variety of morphologies has been observed in ceramic foams, the choice of RVE generation method remains an open problem. In the present work, a review of available methods is performed to identify those most capable of reproducing the morphologies of real foams. Through a case study of conductive heat transfer modeling in an open-cell ceramic foam, the influence of RVE generation method is then analyzed.

In the context of heat transfer modeling, developments in RVE generation techniques are driven mainly by studies on radiative or convective transfer, where the effect of morphology is pronounced. Two main approaches to generate RVEs are distinguishable from the literature:

• The first treats the foam as a subtraction of elementary objects, usually spheres, from a solid phase. Early works assume regular packing (such as in face-centered cubic lattices), while recent developments introduce dispersity and contact laws to simulate bubble physics.
• The second approach partitions the volume into cells with minimum surface energy, then grows the solid phase from the cell faces (walls) and edges (struts). As with the first approach, regular partitions (such as the Kelvin cell) are increasingly being superseded by Voronoi-based structures that reflect the disorder in real foams.

In both approaches, physics-based techniques (bubble simulation or Voronoi-based tessellation) currently give the most realistic foam representations, often with predetermined cell size distributions as input. For open-cell structures, the key differences in the resulting geometries lie in the void phase connectivity and the range of pore and strut shapes possible. The Voronoi-based approach seems well suited to distinctively polyhedral pores. However, when the pores appear spherical (as is the case for many ceramic foams), both approaches have been successfully used to reproduce key morphological parameters of real foams. It is thus interesting to quantify the influence of the choice of RVE generation approach on the modeled effective thermal conductivity.

A case study is performed on an open-cell alumina foam with 80% porosity, with morphological parameters extracted from micro-tomographic images. Two periodic RVEs are then generated with state-of-the-art approaches, starting with random packing of non-overlapping, polydisperse spheres in a cubic volume. The first inflates the spheres before subtracting them from a solid matrix. The second uses the spheres as seeding points for a Voronoi-Laguerre diagram, to which polygonal struts of non-constant section are added. Both RVEs are geometrically validated with the morphological parameters of the real foam. Finite element modeling is used to obtain the effective thermal conductivity of the real foam mesostructure and the two RVEs. The RVEs may then be used to perform parametric studies on the foam morphology. The results highlight the influence of RVE generation technique, and provide modeling guidelines for future work.

Speaker: Zi Kang Low (Université de Lyon, INSA Lyon, LaMCoS CNRS UMR5259 ; Saint-Gobain C.R.E.E.)
• 11:53
Prediction of the thermal conductivity of porous building materials with nanoscale pore size distributions 15m

Porous materials find frequent use in numerous thermal applications, offering a strong reduction of the total heat flow. Typical applications can be found in the automotive industry and aerospace engineering for the protection of sensitive components, but also in the building industry where heat losses through the opaque building components still account for a large part of the needed heating energy. Therefore, recent research activities have focused on the development of several new types of porous materials, i.e. vacuum insulation and aerogels, showing strongly reduced thermal conductivities compared to conventional insulating materials. Their improved performance is mainly attributable to their microstructure, composed of pores and matrix walls with nanoscale characteristic dimensions, rendering the classic Fourier diffusion heat law no longer valid. Indeed, at decreasing length scales and pressures, the heat flow behavior is known to shift to the Knudsen diffusion regime, resulting in reduced conductive heat transfer through the gaseous and the matrix phase. A more thorough understanding of the impact of these microstructural parameters on the total heat flow through the material could hence lead to a significant optimization of insulating materials. However, current numerical studies are often based on simplified analytical models and microstructures or focus on only one form of heat transfer.
In this study, a recently implemented pore-scale model for studying heat transfer in conventional cellular porous materials is expanded for studying materials with nanoscale features or reduced gas pressures. The model is based on 3D voxel images of the microstructure, thus incorporating the true microstructural parameters. The nanoscale ballistic behavior of the energy carriers in the gaseous phase is modelled by defining local thermal conductivities based on the kinetic theory. The reduced local effective mean free path is calculated using several geometrical parameters, considering both the shape and the size of the pore structure. Hence the model allows for an efficient simulation of the heat flow through the material, while implementing the Knudsen diffusion locally at the pore scale. The model is verified using existing analytical models and simple microstructures, showing good agreement. Finally, the model is used to make a preliminary study on the relative impact of several microscale parameters, showing the potential improvement of these new materials compared to conventional materials.

Speaker: Wouter Van De Walle (KU Leuven, Building Physics Section)
• 12:11
Determination of solid-phase conduction shape factor for spherical-void-phase REVs generated by a random discrete element model 15m

This article describes the direct comparison between pore-level computations on a spherical-void-phase (SVP) porous material and volume-averaged computations done for the same domain. Pore- level simulations are conducted on random SVP geometries generated using the Discrete Element Modelling (DEM) approach developed by Dyck & Straatman [1] over a range of flow and heating conditions. Pore-level simulations for cases of constant wall temperature require nearly 17,000,000 tetrahedral elements and provide data that are used to establish the fluid permeability and interstitial convection coefficient of the REVs under consideration. Additional pore-level simulations are then conducted on the same REVs except treating it as being heated on one side by a constant-temperature substrate. These simulations require discretization of both the fluid and solid phases of the REV and require nearly 32,000,000 tetrahedral elements to produce grid-independent solutions of forced convection heat transfer. Analogous conjugate heat transfer simulations are done using a volume-averaged solver on the same domain using the closure parameters established in the pore-level simulations. These simulations require only 5800 hexahedral elements to achieve grid convergence to better than 1%. Comparison of the conjugate results from the pore-level and volume-averaged solvers provides data to establish a solid-phase conduction shape factor that is necessary to modify the diffusion term in the volume-averaged solid-phase energy equation. The solid-phase conduction shape factor is dependent only on the geometric structure of the porous domain and accounts for the additional resistances in the conduction path due to elongation and area changes along the path [2-5]. Even though there is no analytical means for establishing the conduction shape factor for complex geometric models, it is shown that the factor can be established for a particular geometric model on the basis of comparisons for a single heating condition. Subsequent simulations comparing the pore-level and volume-averaged results show that the accuracy of volume-averaged computations is enhanced by as much as 20% with the proper characterization of the solid-phase conduction shape factor.

Speaker: Prof. Anthony Straatman (Western University)
• 12:29
Microstructural Modeling and Simulation of Heat Transfer in Wood Fiber based Insulating Materials 15m

Wood fiber based materials are of high interest in building insulation. Their application is desirable due to the sustainability of renewable resources. Furthermore, wood fiber based materials outmatch petrochemical based materials with respect to health aspects during process and application.

The insulation properties of such fiber based materials are often characterized experimentally, which has certain disadvantages. For instance, the experimental characterization requires a high effort to characterize only few produced material variants. The connection between the properties of the fibrous microstructure and the effective thermal conductivity is hardly enlightened.

To overcome these disadvantages in the current presentation a microstructural modeling and simulation approach for the determination of the effective heat transfer is presented. The development of imaging procedures and powerful computer simulations allow the characterization of structure property relationships even for highly complex fiber networks.

The following work flow is applied.
First of all, the single fibers which form he compund are geometrically analyzed. Furthermore, highly resolved three-dimensional computer tomography (µCT) images of wood fiber based insulating materials are generated. From these geometrical characterizations the fiber network and the pore volume distribution are evaluated.
In a second step based on these characterizations virtual realizations of the materials are generated.
Subsequently, the microstructural simulation of the heat transfer in these virtual representation is carried out and compared to experiments. An advantage of this microstructural simulation technique is that as input only the conductivity of the wood fibers is required.
Studies on the influence of different process parameters as fiber length distribution, fiber orientation and raw density are possible by generation of appropriate virtual microstructures.
Furthermore, effective (global) as well as local quantities are evaluated.
Therefore, this virtual material testing approach allows the prediction and optimization of insulating material without extensive trial and error approaches.

The presented simulations are prepared by using the commercial software tool Geodict [1] and the fast microstructure Solver FeelMath [2].
This efficient solver directly operates on voxel images and thus no effortful generation and storage of meshes is required. Thus, very large microstructures, either virtually generated or directly obtained from µCT images, can be simulated fast and memory efficient.

[1] www.math2market.com
[2] www.itwm.fraunhofer.de/en/departments/sms/products-services/feelmath.html

Speaker: Sarah Staub (Fraunhofer ITWM)
• 14:35 16:49
Parallel 8-A
• 14:37
Investigating Treatment Techniques for Stimulated Ureolytic Microbially-Induced Calcite Precipitation at Field Scale Treatment Depths 15m

Microbially Induced Calcite Precipitation (MICP) is a bio-mediated soil improvement process that can improve the engineering properties of granular soils through the precipitation of calcium carbonate on soil particle surfaces and contacts. Although bio-cementation has been investigated extensively in laboratory experiments (DeJong et al. 2006, Martinez et al. 2013, and others) and successfully up-scaled in several meter-scale applications (van Paassen 2011; Gomez et al. 2015), most studies have relied on the injection of laboratory-cultured bacterial strains, or bio-augmentation, to enable the bio-cementation process. Bio-stimulation offers the transformative ability to use selective environmental conditions and substrates to enrich native microbial populations in-situ to obtain desired metabolic and enzymatic capabilities. The enrichment of native ureolytic bacteria for bio-cementation may enable significant reductions in process treatment financial costs and detrimental environmental impacts.

Although past studies have shown that ureolytic stimulation is feasible in surficial soils and groundwater samples (Fujita et al. 2000; Tobler et al. 2011; Burbank et al. 2011; Gat et al. 2014; Gomez et al. 2014; 2016), many engineering applications may require stimulation of native ureolytic microorganisms at much greater treatment depths near 15 meters. At the same time, microbial abundances and activities are anticipated to decrease with increasing soil depth due to reductions in nutrient availability, soil temperature, and changes in other environmental factors (Fierer et al. 2003; Eilers et al. 2012). If bio-cementation via stimulated microbial ureolysis is to be used for deeper subsurface applications, the effect of soil depth on ureolytic enrichment in natural soils must be better understood.

In this study, batch and soil column experiments were performed using alluvial sand and gravel samples obtained aseptically from a recently exposed cut slope and geotechnical boring at a field site (Gomez et al. 2018) at depths up to 12 meters. Experiments were completed to investigate the effect of soil depth on the enrichment of native ureolytic microorganisms, the bio-cementation process, and the performance of several different stimulation solution techniques. During tests, solution samples were obtained over time to monitor changes in aqueous cell densities, urea degradation, and solution pH occurring during stimulation and bio-cementation treatment phases.

Results suggest that significant biological differences may exist between soil samples with increasing soil depth and that different stimulation techniques than previously used on surficial soils may be required to achieve successful bio-cementation in deeper materials. Despite most specimens achieving similar cell growth and solution pH increases over time, significant differences in urea degradation were observed, suggesting that solution pH monitoring alone may not be an effective indicator of urea hydrolysis and that total cell counts may omit important information about differences in enriched microbial populations. Increases in stimulation solution yeast extract and ammonium chloride concentrations and adjustment of solution pH to more alkaline values was shown to improve ureolytic enrichment through both increases in obtained total cell densities and increased selective pressure for alkaliphilic ammonium-tolerant microorganisms. The achieved results suggest that stimulated ureolytic MICP is possible in deeper subsurface locations at treatment depths near 12 meters.

Speaker: Dr. Michael G. Gomez (Department of Civil and Environmental Engineering, University of Washington, Seattle)
• 14:55
MICP in the Field: Enhancement of Wellbore Cement Integrity and Permeability Modification 15m

Microbially-induced calcite precipitation (MICP) is being widely researched as an emerging technology for subsurface engineering applications including sealing defects in wellbore cement and modifying the permeability of rock formations [1]. Our study team, including Montana State University’s Energy Research Institute and Center for Biofilm Engineering, The University of Stuttgart, Montana Emergent Technologies, and Schlumberger Carbon Services, has conducted two successful MICP-based field trials. The first field test successfully used MICP to seal a horizontal “pancake fracture” in tight sandstone [2]. The following test resulted in MICP sealing of compromised wellbore cement. Both tests were carried out in the Gorgas well in Alabama and used the ureolytic bacteria, Sporosarcina pasteurii, to promote calcium carbonate precipitation. A third field test is now underway targeting an existing oil well, Rexing #4, in Indiana. The Rexing #4 well was previously used as an injection well to sweep residual oil to production wells. Several years ago, injection pressure was lost, and the well was removed from service. Subsequent well logging measurements suggested that, rather than entering the target injection formation, injectate was traveling up the wellbore through defects in the well cement to a sandstone thief zone approximately 50 feet above the target formation. The goal of the field demonstration project at the Rexing #4 well is to use MICP to reduce permeability in the thief zone and cement defect until the injection pressure reaches the pre-breakthrough condition necessary to return Rexing #4 to service as an injection well. The field demonstration in Indiana, funded by the US Department of Energy, will be conducted the first week of December 2017 using a custom mobile laboratory designed for field applications of MICP. Results presented at the conference will include a rationale for the injection strategy employed, as well as its success at sealing the leakage pathway.

Speaker: Dr. Catherine M. Kirkland (Montana State University)
• 15:13
Modeling porous medium modification through induced calcium carbonate precipitation 15m

Fluid storage in the subsurface is important to reduce climate change (sequestration of CO2) or for energy storage (CH4, H2) to cope with the intermittent, unpredictable production of renewable sources like wind and solar. However, the fluids have the potential to leak through damaged cap rocks or wellbores. One method to remediate these problems is inducing calcium carbonate precipitation (ICP). Currently, most applications of ICP rely on urea hydrolysis by microbes (MICP) to promote precipitation within the porous media. However, precipitation may also be induced by injection of extracted or plant-based sources of the enzyme urease (EICP) or at elevated temperatures (TICP). The applicability of a certain method of ICP is largely determined by the depth below ground surface and the local geothermal gradient. MICP has been demonstrated to have immense potential to seal leakage pathways, even at field scale [1] but is only effective within a limited temperature range, as it relies on the activity of living bacterial cells. As a consequence, the other ICP methods EICP and TICP have to be developed and demonstrated in the field. To assist experimental investigations on EICP and TICP, a previously developed numerical model for MICP [2,3] is generalized and adapted for the new precipitation-inducing processes. As the models are intended for the use in predicting the leakage mitigation for subsurface gas storage, they account for two-phase flow. Additionally, a variety of different components and processes are necessary to describe ICP, the specific number of components and processes being dependent on the precipitation-inducing process. All models are implemented in the open-source simulator DuMuX [4]. The primary variables solved are the aqueous-phase pressure, mole fractions of each component in the water phase, temperature, and, for the solid phases, the volume fractions. The mass balance equations are solved fully implicitly and are coupled through the source and sink terms due to the reactions. The new kinetic rate equations for the developed EICP and TICP models were fitted to experimental data obtained from batch experiments at Montana State University. The porosity and permeability reduction due to calcium carbonate precipitation and accumulation of biomass or enzyme are accounted for by updating the porosity using the volume fractions of the solid phases and for the permeability by using the updated porosity and a Verma-Pruess-type relation [5]. The new models for EICP and TICP will be calibrated and validated using column experiment data, similarly to the procedure outlined in [3], which will then be used to determine the optimal mineralization method and injection strategy for given boundary conditions.

Speaker: Johannes Hommel (University of Stuttgart)
• 15:40
Calcium carbonate precipitation and strontium co-precipitation in porous media flow reactors 15m

Strontium-90 (Sr-90), a toxic and carcinogenic radionuclide, is the product of uranium fission and is found in soil and groundwater at numerous DOE sites.1 Sr can also enter the environment through mine tailings leachate, produced water ponds (oil and gas extraction) or can occur naturally in rock formations.
A potential technology for groundwater remediation is the use of subsurface microorganisms to induce calcium carbonate (CaCO3) precipitation resulting in the partitioning of metal contaminants into CaCO3 precipitates for long-term sequestration. Precipitation can be facilitated by an increase in alkalinity as a result of urea hydrolysis, which can be induced by microbes in a process called microbially induced CaCO3 precipitation (MICP). Thus, Sr co-precipitation studies have the potential to provide insight into Sr-90 partitioning into CaCO3.
Our studies in three flow systems have aimed to characterize and control CaCO3 distribution and Sr co-precipitation through MICP by manipulating fluid flow and CaCO3 saturation conditions. First, the effects of flow rate and Ca-concentration on the strontium partition coefficient (DSr) were determined in porous media flow cells (FC), similar to those described in Lauchnor et al.1 Second, spatio-temporal analysis of strontium partitioning was performed using a novel modified flow cell; this spatially-sampled flow cell (SFC) enabled fluid sampling from different locations within the porous medium during the experiments. Finally, a laboratory-scale reactor, wherein glass beads formed a packed porous bed, was used to investigate MICP under radial flow conditions.
Two flow rates and two calcium concentrations were studied in the FC. The calcium precipitation rate in FC experiments suggested that under the conditions chosen, MICP was limited by calcium transport. The low calcium concentration and low flow rate experiment led to the highest MICP efficiency and Sr co-precipitation.
SFC experiments revealed that calcium and strontium gradients did not remain constant over time. Spatially collected samples aided in DSr calculations to study the spatio-temporal behavior of strontium co-precipitation. The calcium precipitation rate decreased with time in all three replicates. A decrease in strontium partitioning with distance into the SFC was coupled with an increase in the size of the precipitates.
For effective field employment of MICP, it is important to control MICP specifically under radial flow conditions relevant in near-well environments. Porous media radial flow systems were utilized to evaluate spatial distribution of CaCO3. MICP experiments were performed at three fluid flow rates and three calcium concentrations. MICP efficiency showed an inverse relationship to flow rate and the greatest precipitation efficiency was observed at the lowest calcium concentration. Preferential flow paths developed due to precipitates formed via MICP, affecting fluid flow.
These results allow for predicting MICP distribution and the effect of MICP on porous media flow properties and lend insight to potential MICP strategies for remediating Sr-contaminated groundwater.

1. Lauchnor, E. G.; Schultz, L. N.; Bugni, S.; Mitchell, A. C.; Cunningham, A. B.; Gerlach, R., Bacterially Induced Calcium Carbonate Precipitation and Strontium Coprecipitation in a Porous Media Flow System. Environmental Science & Technology 2013, 47 (3), 1557-1564.
Speaker: Neerja Zambare (Montana State University)
• 15:58
Low Field Nuclear Magnetic Resonance to Monitor Bio Mineralization Processes in Porous Media 15m

Low-field nuclear magnetic resonance (NMR) is a non-invasive measuring technique and an excellent tool for determining properties of materials with high magnetic susceptibilities such as rock cores and natural sediments. NMR is sensitive to parameters such as pore size, pore fluid changes, and permeability that are of interest to engineering applications such as subsurface fracture sealing and CO2 sequestration by bio mineralization. Using a 2MHz Rock Core Analyzer, NMR relaxation and diffusion measurements are utilized to monitor changes in porous media during biofilm growth and bio mineralization. T2 relaxation measurements are used to monitor changes in pore size distributions as well as chemical changes at the surface of pore walls due to microbially-induced mineral precipitation. T2 relaxation measurements are also used to track changes in the total porosity of the system. The spatial distribution of relaxation rates throughout the length of the sample are determined using T2 profiles, and signal intensity profiles provide a means to characterize the spatial heterogeneity as well as a way to monitor the progress of the bio mineralization process as the signal intensity decreases due to calcite precipitation. Diffusion measurements are performed to detect restricted diffusion as calcite begins to occupy the pore space. The application of bio mineralization as a fracture sealing agent as well as the ability of calcite to form in the pore space of a porous media is investigated using the above measurement techniques.

Speaker: Linn W. Thrane (Mechanical and Industrial Engineering, Montana State University )
• 16:16
Permeability Reduction Caused by Multiple Treatments of Biomineral Precipitation in Homogeneous Porous Media: Experimental Study and Pore Scale Modelling 15m

Several biochemical processes have been investigated to modify engineering properties of soil. Biomineral precipitation can increase strength and stiffness, and reduce porosity and permeability. Enzymatically induced calcium carbonate precipitation (EICP) is a biochemical process in which urea is hydrolyzed into ammonium and inorganic carbon, which promotes carbonate mineral precipitation. Different morphologies and patterns of carbonate mineral precipitation, such as particle surface coating, pore filling, and soil particle contact bonding have been observed in the previous studies, in which the mineral structure and distribution has been evaluated after the completion of the treatment using SEM (Scanning Electron Microscope) imaging and XRD (X-ray Diffractometer) structural analysis. As the hydro-mechanical properties of soils can be significantly affected by the distribution of precipitated mineral in the pore space it is important to investigate these properties at pore scale. In this research, an EICP reaction medium is injected into a microfluidic channel to observe the entire process of carbonate mineral precipitation through several cycles of EICP treatment in the porous medium. Mineral phase changes and concomitant porosity and permeability reduction throughout the system are evaluated by image processing. A two-dimensional simulation model involving urea hydrolysis and mineral precipitation is developed, and the results are analyzed in comparison to the experimental results.

Speaker: Daehyun Kim (Arizona State University)
• 16:34
Pore space sealing using microbially mediated calcite precipitation: a lab to field scale study 15m

Microbially driven calcite precipitation (via ureolysis) has shown great potential in a wide range of applications, including solid-phase capture, concrete crack remediation, soil stabilisation and carbon sequestration. Here, this process is investigated as a means of reducing the primary porosity and/or secondary fracture porosity of host rocks surrounding nuclear waste repositories in order to control or prevent radionuclide transport. To determine a suitable field injection approach, a series of bench scale experiments were undertaken in the laboratory. First, batch experiments focussed on the kinetics of calcite precipitation as a function of bacterial mass, urea and Ca2+ concentration and anaerobic vs aerobic conditions. Results showed that the ureolytic bacteria performed equally well under both oxygen poor and oxygen rich conditions. In the next stage, flow-through experiments in various media (sand columns, rock cores) were carried out to examine the homogeneity and extent of the pore space fill along the column / core as a function of injection strategies. It emerged that a staged injection strategy, where we alternate between bacterial and reactant injection, yields the most homogeneous calcite fill, reducing overall porosity by up to 45 %. Ultimately, this approach was tested at the field scale, led by University of Birmingham, to seal a fractured rock (dacite) at ~28 m depth, in a quarry in Leicestershire, UK. Within few injection cycles, the single fracture was substantially plugged by calcite, yielding a significant transmissivity decrease over several meters.

Speaker: Dr. Dominique J. Tobler (Nano-Science Center, Department of Chemistry, University of Copenhagen, Denmark)
• 14:37 16:49
Parallel 8-B
• 14:37
Flow and Mechanics in Fractured Media as Mixed-Dimensional PDEs 15m

In the mixed-dimensional representation of fractured media, the fractures are considered as lower-dimensional manifolds. This concept is successively applied to the lines and points at the intersections between fractures leading to a hierarchical geometry of manifolds of codimension one. By imposing these modelling assumptions a priori in the continuous setting, a basis is formed for the introduction of coupled PDEs on the mixed-dimensional geometry, which we refer to as mixed-dimensional PDEs.

In this work, we consider Darcy flow combined with linear elasticity on mixed-dimensional representations of fracture networks. Since the associated, governing equations are fully coupled, the systems of equations are presented using mixed-dimensional differential operators which map between the different dimensions. In turn, the resulting system of equations is considered as mixed-dimensional and can be analysed as such, before the introduction of the discretization scheme.

We present theoretical results related to the structure of mixed-dimensional elliptic partial differential equations from which multiple conforming discretization schemes arise using dimensionally hierarchical finite elements.

Keeping later purposes such as transport problems and fracture propagation in mind, our main interest lies in obtaining accurate flux fields and stress states which respect physical conservation laws. Therefore, we employ mixed finite elements which allow for a local preservation of such laws. The symmetry of the stress tensor is imposed in a weak sense, thus leading to the use of familiar, conforming, finite elements with relatively few degrees of freedom.

Results concerning convergence and stability of the mixed finite element schemes are shown. These are supported by numerical examples in two- and three-dimensional domains in which the lower-dimensional inclusions, intersection lines, and points have significantly different material properties compared to the surroundings.

Speaker: Wietse Boon (University of Bergen)
• 14:55
Linking geometric characteristics of tumor vasculature and fractional diffusion 15m

Tumor vasculature is characterized by enhanced permeability, increased
tortuosity and higher fractal dimension \cite{baish2000fractals, gazit1997fractal}
leading to a higher degree of heterogeneity between cancerous and healthy tissue.
Diffusion processes in strongly heterogeneous media have been
successfully predicted by fractional diffusion models, e.g. signal decay
in diffusion weighted MR imaging \cite{magin2008anomalous}. In MRI the
fractionality of the diffusion operator can be established from the measured intensities
and the value can then be used to infer fractal dimension of the tissue, e.g.
gray and white matter \cite{hall2008diffusion}. Relation between the structural
heterogeneities, which may be typical for certain pathologies, and a measurable
quantity, such as the exponent of the diffusion operator, can thus be
exploited for diagnostics.

In this work we use computer-generated vasculature networks with different
structural complexities and employ a coupled 3$d$-1$d$ model of \cite{d2008coupling} to
characterize the diffusion properties of the tissue. The predictions are
compared with those of a simple 3$d$ fractional diffusion model in order
to establish the connection between the value of the exponent and the
geometric characteristics of the network.

Speaker: Miroslav Kuchta (Department of Mathematics, Division of Mechanics, University of Oslo)
• 15:13
Simulation of Metabolic Processes in Plant Cells 15m

The interior of living cells can be considered as a porous medium consisting of three compartments: cytosol, chloroplasts and mitochondria. Diffusion and reactions take place inside the cytosol, inside the chloroplasts and on the surfaces of the mitochondria. Furthermore, biochemical species can be exchanged between the compartments. Consequently we need to solve a system of fully dimensional partial differential equations coupled with partial differential equations on lower dimensional manifolds. The nonlinear coupling terms result from reversible enzyme kinetics. We present an efficient algorithm for solving this system on parallel machines and investigate the role of enzyme localization on the mitochondria surfaces.

Speaker: Mr. Tobias Elbinger (Friedrich-Alexander-Universität Erlangen-Nürnberg)
• 15:40
A new simulation framework for plant-scale soil-root interaction including evaporation from soil, growth and transport processes 15m

We present a general model concept and a flexible software framework for the description of plant-scale soil-root interaction processes including the essential fluid mechanical processes in the vadose zone. The model is developed in the framework of non-isothermal, multi-phase, multi-component flow and transport in porous media. The software is an extension of the open-source porous media flow and transport simulator \dumux to embedded mixed-dimensional coupled schemes. Our coupling concept allows us to describe all processes in a strongly coupled form and adapt the complexity of the governing equations in favor of either accuracy or computational efficiency. We present the necessary numerical tools to solve the arising strongly coupled non-linear PDE systems with a locally mass conservative numerical scheme even in the context of evolving root architectures. We demonstrate the model concept and its features discussing a virtual hydraulic lift experiment including evaporation, root tracer uptake on a locally refined grid, the simultaneous simulation of root growth and root water uptake, and an irrigation scenario comparing different models for flow in unsaturated soil. We analyze the impact of evaporation from soil on the soil water distribution around a single plant's root system. Further, we show that locally refined grids around the root sytem increase computational efficiency while maintaining accuracy. Finally, we demonstrate that the assumptions behind Richards equation may be violated under certain conditions.

Speaker: Katharina Heck (University Stuttgart)
• 15:58
Mixed-dimension models for network flow in biological systems 15m

Hierarchical network structures occur in many biological systems, since they are responsible for the transport of fluids, nutrients or oxygen. Such a network structure is, for example, the blood vessel network supplying organs with oxygenated blood or removing metabolic waste from the tissue. A further example is the root network of a plant, ensuring the plant‘s water supply.

One way to obtain a realistic mathematical model for such processes is based on a decomposition approach. Thereby, the network structure is separated from the surrounding medium and different models are assigned to both domains. Often, the surrounding medium (e.g. tissue or soil) is considered a three-dimensional (3D) porous medium. To decrease computational costs while maintaining a certain degree of accuracy, flow and transport processes within the networks are modeled by one-dimensional (1D) PDE-systems based on cross-section averaged quantities. A coupling of the network and the porous medium model is achieved by first averaging the 3D quantities and projecting them onto the 1D network structure. As a next step, a transfer function based on the difference of the averaged 3D and 1D quantities is computed and incorporated into the source/sink terms of the corresponding models. The source term of the 3D problem exhibits a Dirac measure concentrated on the 1D network.

In this talk we present several application areas for this kind of coupling concepts. Furthermore, we are concerned with the numerical analysis of PDE systems arising in the context of this model concept. In particular, it is investigated how the Dirac source terms and averaging operators affect the convergence behavior of standard finite element methods. Therefore, elliptic model problems with Dirac source terms and averaging operators are investigated. Our theoretical results are confirmed by numerical tests.

Speaker: Tobias Koeppl (University of Stuttgart)
• 16:16
Water management of plant tissues during frost-thaw cycles 15m

Plant tissues have developed several strategies to cope with multiple cycles of freezing and thawing events without being damaged. Some of these strategies are of physiological nature, others arise from structural properties. Understanding the involved strategies and mechanisms of plants exposed to frost conditions is of high interest, as they could potentially be used for the development of bio-inspired construction materials with optimised properties in terms of frost resistance, thermal isolation and guided water/moisture transport.

A decisive factor with regard to frost resistance in plants is the dehydration of the tissue cells leading to an increase of mobile water in the intercellular space. In contrast, freezing within the cells threatens the survivability of the plant. The intercellular water is then transported to species-specific and tissue-specific locations in the plant where freezing is not critical. For this water management, properties of the plant's microstructure are crucial, which are arising from the arrangement of the tissue cells. This arrangement results in highly heterogeneous and anisotropic conditions. Particularly the role of these structural effects is of interest for a transfer to construction materials.

Since the involved thermo-hygro-mechanical processes in plant tissues upon freezing and thawing cycles are strongly coupled, a modelling approach based on the Theory of Porous Media (TPM) is applied, which describes the multiphasic and multicomponent aggregate on the macroscale. In particular, a quaternary model is introduced with a solid skeleton (composed of the tissues cells) and two fluids in the intercellular space, namely, gaseous air and liquid water, which may turn into solid ice. The phase transition of water occurs at a singular surface, which is characterised by a jump in physical quantities, such as the density. The mass transfer can be formulated by using the energy jump at this interface leading to a thermodynamically consistent formulation. In this approach, the interfacial area between liquid and solid water in the partially saturated plant tissues needs to be considered. The freezing of the pore water leads to the necessity to consider the so-called compaction point in the material description, as the bulk material undergoes a transition from a porous material to a solid material. Furthermore, the cell dehydration is included by a production term in the mass balance of the solid skeleton using a Darcy-type law for the cell-wall perfusion. A similar Darcy-type approach was chosen for the fluid flow within the intercellular space with spatially varying anisotropic permeability conditions enabling, thereby, a description of the water management. All these effects are illustrated by selected numerical examples.

Speaker: Mr. Lukas Eurich (University of Stuttgart)
• 16:34
Proliferation rate estimation using a continuum-mechanical cancer cell model 15m

Once lung-cancer cells have invaded the brain tissue via the blood-vessel system, the cells proliferate and migrate in the tissue. The nutrients in the interstitial fluid ensure the proliferation and the basal reactions of the cancer cells. Over time, the cancer cells proliferate and form metastases. In experiments, multicellular lung-cancer spheroids are grown under fully-nutrient-supplied conditions, which allow the comparison of the experiment to the early stage of the formation of lung-cancer metastases in a continuum-mechanical model. Moreover, the experiments enable the adaptation of relevant model parameters via maximum likelihood estimation.

In this contribution, the model is focused on the cancer-cell proliferation starting from an initial cancer cell amount within the tissue. Furthermore, the cells can spread within the brain tissue and, thus, increase the affected region.
As a consequence of the time series of the experimental data, the model is simulated for the same observation period, allowing for the comparison of the overall cancer-cell amount in space and time.

In this contribution, the continuum-mechanical model is based on the
Theory of Porous Media (TPM). Therein, the microscopic structure is volumetrically homogenised over a representative elementary volume leading to a macroscopic multi-phase model with interacting continua. In particular, the constituents are given by an elastic solid skeleton (brain cells) and two immiscible pore liquids (blood and interstitial fluid). Herein, the latter is a real mixture of the proliferating cancer cells and nutrients. The proliferation and consumption themselves are described via mass-production terms.

Numerically, the weak forms of the overall momentum balance and the adapted mass-balance relations are solved for their related primary variables. Thereby, the primary variables are the solid deformation, the pressures of the liquids as well as the concentrations of the cancer cells and the nutrients. The resulting coupled system of equations is solved monolithically applying the finite-element tool PANDAS. In this procedure, the spatial discretisation is realised via Taylor-Hood elements and the time is resolved on the basis of an Euler time-integration scheme.

Finally, the results are compared to the volumetric lung-cancer-cell data obtained from experiments. Thereby, an optimisation of the proliferation parameters is performed using the maximum likelihood estimation. This procedure leads to repeated executions of the numerical model and consequently to a better estimation of the proliferation rates.

Speaker: Mr. Patrick Schröder (Institute of Applied Mechanics (CE), University of Stuttgart)
• 17:15 18:45
Poster 3: Poster 3-A
• 17:15
Effect of salinity on the transport of heavy metals and radionuclides in reactive porous media 15m

Hydraulic fracturing (or fracking) is a well stimulation technique for unconventional oil and gas extraction [1]. Around 8-38 million cubic meters of fracking fluid containing water, chemicals, and sand are injected into the shale every day [2]. High-pressure injection of fracking fluids allows to create fractures and mobilize the gas and the oil towards the surface. Together with the gas, a hypersaline brine (i.e., the flowback and the produced water) is extracted which contains heavy metals and radionuclides, such as barium, strontium, and radium. Spills of flowback and produced water into the subsurface may occur during operation, handling, and storage with potential negative impact on potable aquifers. However, the behavior of the heavy metals and radionuclides at the condition of hypersaline brine has not been studied, yet.
Here, we present an experimental and modeling work to describe the effect of salinity on the transport of the major cations in the produced water. We selected goethite-coated silica beads as reactive material. We performed column-flood experiments using both produced water (obtained from MSEEL, WV) and synthesized produced water. A triple layer model (TLM) [3] was developed and implemented in PHREEQC to simulate the metal transport behavior. Preliminary experiments were performed by injecting a pH 8 solution containing barium at various salinity into a column filled with goethite-coated beads. Results show that the retardation of barium significantly decreases with salinity due to its decreasing free ion activities. This suggests that in a case of a spill of produced water within an aquifer containing iron-oxide mineral, a fast migration of the major cations may occur with a potential negative impact on aquifer water quality.

Speaker: Ms. Zi Ye
• 17:30
Chelation for filter regeneration: wave structures and process optimization 15m

Treatment of water for human consumption often involves filtration through reactive porous media, such as activated carbon, or metal oxides. The sorption of contaminants inevitably leads to a depletion of available free sites and therefore a decrease in efficiency. A possible technique to regenerate filtration systems is the use of chelation agents. Such compounds, i.e. ethylenediaminetetraacetic acid (EDTA), readily form aqueous complexes with heavy metals. The introduction of a chelating agent into a filtration system thus forces the contaminant off the surface and into the mobile aqueous phase. Here we investigate the effect of chelation agents on the transport behavior of the chromium-iron oxide-proton-EDTA system through chromatographic theory. The analysis leads to a regime diagram in the phase plane illustrating the fundamental wave structures. A unique feature of this system is a Cr desorption shock which travels at the fluid velocity. This is in contrast to classic chromatographic theory, where shocks are brought about by sorption and are often retarded relative to the fluid velocity. The filter material can therefore be theoretically regenerated with one pore volume, a marked improvement from classic competitive sorption alone. We go on to optimize the filter regeneration process by engineering feed water composition (pH-EDTA) to minimize flush water volume and chelate mass usage.

Speaker: Colin McNeece
• 17:45
A novel pore structure reconstruction procedure facilitating simulative analysis of multiphase displacement processes in porous media 15m

With the increasing demand for enhancing production of oil and gas all over the world, improving oil recovery technologies aiming at developing low permeable and tight oil reservoirs are becoming the focus of current research. To help understanding the mechanism of multiphase displacement in underground reservoir, the microcosmic characterization of porous media must be explored. In recent years, Digital rock physics (DRP) technology for simulative analysis of rock physical properties based on extracted CT scanning images plays more and more important roles in geosciences, soil science, petroleum engineering and many other fields.
In this paper, A novel procedure is proposed for accurate reconstruction the 3-D pore internal structures, based on which the pore volume is successfully meshed for further detailed numerical investigations of the multiphase displacement process in porous media. The first step is to determine the REV of the studied rock after processing the all the CT images with the software of ImageJ®, with which the image noise reduction , threshold segmentation and binarize processes are carried out. The second step is to employ Matlab® software to transform the binary black and white series of rock images into a digital 3-D matrix consists of 0 and 1. Then reconstruction of the pore void volume is performed through employing the software of ProE®, which could supply the compatible format for the fluid dynamics analysis software of Fluent®.
Single phase and two-phase displacement processes are numerically investigated with the software of Fluent® in the reconstructed pore structures. Numerically obtained gas permeability is 2.02D, which reasonably lies in the region of 1.2~3D of the Bentheimer core samples. It is concluded the proposed novel rock internal structure reconstruction procedure can ensure the simplicity and accuracy on predicting the multiphase transport characteristics in porous media.

Speaker: Dongxing DU (Qingdao University of Science and Technology)
• 18:00
Generalized Multiscale Method for flows in pore space with inhomogeneous boundary conditions 15m

Keywords:
Generalized Multiscale Finite Element Methods, porous media, perforated domains, reactive flows
Abstract:
In this work, we develop a generalized multiscale methods for reactive flows in porous media. The flow equations are formulated in the pore space and Robyn type boundary conditions are used to represent the reaction terms.The solution techniques for these problems require high resolution and leads to the large system in the fine-scale approximation that computationally expensive. On the other hand, the problem often admits a homogenization in the scale separation case. In this poster, we are interested when there is no scale separation and present a multiscale method that attempts to solve such problems on a coarse grid by constructing multiscale basis functions in each coarse grid block, where the coarse grid can contain many perforations. In our work, we follow Generalized Multiscale Finite Element Method (GMsFEM) and develop a multiscale procedure where we identify multiscale basis functions in each coarse block using snapshot space and local spectral problems. For handling of the inhomogeneous boundary condition on perforations, we present an additional multiscale basis functions that enrich our multiscale space. To illustrate the performance of our method, we present numerical results with both small and large perforations in 2D and 3D.

Speaker: Mr. Denis Spridonov (Multiscale model reduction Laboratory, North-Eastern Federal University; )
• 18:15
Increasing complexity of sediment flux models to quantify the effect of marsh sediment and pore water sorption processes 15m

Wetland sediments and pore waters are observably dynamic systems for organic matter transport and transformation. Particulate and dissolved organic matter constantly flux throughout these media, affected by soil characteristics, composition, and water quality parameters. Current sediment flux models do not explicitly include most organic matter transformations due to interaction intricacies that are not yet fully understood. The sediment flux model in this study originally encompassed only hydrolysis, the biotic degradation of particulates within the sediments, and remineralization of the dissolved organic matter pools. This research incorporates sorption processes between particulate and dissolved organic matter into the sediment flux model. Specifically, new versions of the model were constructed with increasing levels of complexity of sorption processes. These models were evaluated using new organic matter reaction rate experiments and sediment observations. The new model formulations were also run under various forcing conditions and compared to thirteen-year base model runs without the additional complexity. Results indicate that including sorption processes between organic matter pools both increases model sensitivity to input values and dampens nutrient fluxes out of the sediment. This research will, ultimately, provide insight into the ways to formulate more accurate model representations of sediment/water interactions and exchange, validated with marsh observations, of organic matter transport and transformation within wetland sediments and pore water.

Speaker: Ms. Hannah Morrissette (University of Maryland Center for Environmental Science)
• 18:30
Experimental Study of Membrane Properties of Niobrara Shale to Hydrocarbon Mixtures 15m

In nanoporous rocks, fluid-surface interactions and potential size exclusion at pore throats may change the rock into a semi-permeable membrane that hinders the transport of certain molecules through yet lets other components pass through freely. In this work, we present an experimental study designed to evaluate whether membrane effect exists for hydrocarbon molecules traveling through Niobrara shale. The experimental setup uses an inline filter as a mini core holder. Rock samples from the Niobrara formation were dry cored and machined to 0.5 inch diameter and 0.8 inch length. The annulus between the sample and the core holder was sealed by epoxies. Hydrocarbon mixtures with known compositions were injected into the shale sample. The injection pressure was controlled by a regulator and kept constant. The filtered fluids were collected periodically from the outlet, and analyzed for compositions by using Gas Chromatography.

Speaker: Ziming Zhu (Colorado School of Mines)
• 18:30
Experiments of Microbially Induced CarbonatePrecipitation in Calcareous Sand by Mixing Method 15m

Calcareous sand is oftenused as the filling materialin marine geotechnical engineering, and it is to be reinforced in practice. The technology of Microbial Induced Carbonate Precipitation（MICP）provides a new cementway to reinforce the calcareous sand.In this paper, we study how to use MICP by the mixing method to reinforce the calcareous sand.The mixing method is very different from the grouting method often used in previous studies.Experiments of direct shear tests and oedometer tests of calcareous sand sampleswere carried out. The different particle sizescomposition and the different reaction solution concentrations were considered. Thetest results showed that high concentration of reaction solution led to more precipitation of calcium carbonate, and the shear strength and the stiffness of calcareous sand were improved. It is shown thatthe mixingmethod of MICP technologyhas certain feasibilityinthe dredged filltreatmentof calcareous sand.

Speaker: Hongxian Guo
• 18:30
Impact of wellbore treatment fluids on calcium carbonate attachment in MICP grouted sands 15m

The interest in developing microbial induced carbonate precipitation as a cement alternative in oil and gas wells stems from the fact that this biotechnology can penetrate pore networks that conventional cement grouts cannot due to their high viscosity. Currently MICP is under investigation as a potential wellbore barrier technology to mitigate hydrocarbon leakage through a) micro-channels in the cement matrix b) micro-annuli as a communication pathway between the casing and the wellbore cement sheath c) and compromised wellbore cement sheaths. A major concern of this grouting technology is how to guarantee long-term well integrity. In particular, wellbore treatment fluids and multiphase subsurface fluids may interfere with carbonate production and/or attachment, thus compromising the integrity of the wellbore repair. In this research, we investigate the impact and influence of hydrocarbons and of currently available state-of-the-art wellbore treatment fluids and determine how these affect the performance of our “biomineral-seal” in the subsurface.

Biologically mediated calcium carbonate precipitation on grain surfaces within a sandy porous media, is known to cause several effects. Carbonates that precipitates in pore throats can form bridges in between the individual sand gains which results in increased strength of the porous media and some reduction in permeability. Carbonate biominerals that precipitate on grain boundaries within the pore space provide additional grain roughness, which leads to an increase in nucleation sites for further precipitation of calcium carbonate crystal polymorphs. Continued biomineral treatment results in a decrease in rock porosity.

The impact of hydrocarbons and other well-bore treatment fluids on biomineralization processes at the microstructural scale, including bacterial attachment and subsequent carbonate precipitation are largely unknown. In this experimental study, we explore the effect of biomineralization on the microstructural and physical properties of a series of batch cemented sand samples at elevated temperatures and pressures in the presence of hydrocarbons. We also explore, for the first time, the effect of wellbore treatments fluids.

We use Micro X-ray computer tomography to compute pore-structural properties (porosity, permeability, specific surface area and volume fraction of calcite attached to the sand grains). To investigate the attachment of precipitated carbonate and bacterial cells (entombed in calcium carbonate) to grain boundaries or calcium carbonate (free-floating) precipitated within pores, Scanning Electron Microscopy with Energy Dispersive X-Ray Spectroscopy (SEM-EDS) are deployed. In parallel, calcium carbonate crystal polymorphology is evaluated by X-ray diffraction (XRD) and, in case of hybrid crystal polymorphology, by Transmission Electron Microscopy (TEM).

Ongoing experiments focus on 1) investigating the survival of S. pasteurii in particular wellbore environments with well-known drilling fluids, corrosion inhibitors and hydrocarbons 2); identifying fluids that might negatively impact on attachment of calcium carbonate 3); the influence of key fluids on crystal polymorphology.

Speaker: Mr. Fabian Steinacher (University of Strathclyde - Department of Civil and Environmental Engineering)
• 18:30
Modelling approach and benchmark experiments for Nernst-Plank based transport, Coulombic interactions and geochemical reactions in saturated porous media 15m

The transport of electrolytes in porous media is affected by physical, chemical and electrochemical processes. Coulombic interactions significantly influence the behavior of electrolyte plumes at different scales, not only in diffusion-dominated conditions but also in advection-dominated flow regimes [1-3]. To model the spatial behavior of charge-induced interactions in multi-dimensional homogeneous and heterogeneous domains, we propose a Nernst-Planck based modeling approach for conservative and reactive transport. The model is based on a coupling between COMSOL Multiphysics® and PhreeqcRM [4]. Important features of the proposed approach include transport of chemical species and not of chemical components, the definition of physically and chemically heterogeneous domains and the implementation of electromigration as well as of the flux components arising from the activity coefficient gradients. The model has been benchmarked with numerical simulations in PHREEQC [5], analytical solutions and high-resolution experimental datasets in homogeneous and heterogenous setups for steady-steady state and transient conditions in different dimensions (1D, 2D and 3D). Fully three-dimensional experiments on multicomponent ionic transport were also performed in this study and were used to validate the proposed modeling approach. Simulations show an excellent agreement with the experimental and modeling benchmark problems, thus highlighting the potential of the Nernst-Planck based model for the evaluation of conservative and reactive multicomponent ionic transport.

Speaker: Riccardo Sprocati (Department of Environmental Engineering, Technical University of Denmark, Miljøvej, Building 115, 2800 Kgs. Lyngby, Denmark)
• 18:30
Surface complexation modeling of arsenic mobilization from goethite: Interpretation of in-situ experiments in a sedimentary basin of Inner Mongolia, China 15m

Sorption competition onto Fe-(oxyhydr-)oxides surfaces is a well-known mechanism controlling the release and mobility of arsenic (As) in subsurface (Dixit & Hering, 2006). Over the last decades, surface complexation models (SCMs) have been implemented to model interactions between sorbants and mineral-oxides surfaces by considering the thermodynamic properties underlying complexation and electrostatic interactions (Goldberg, 1992). However, SCMs development are typically based and/or applied on well-controlled laboratory experiments with simple aqueous systems rather than complex environmental groundwater conditions.

In this study, we present and compare conceptual and numerical modeling approaches developed to quantitatively interpret in-situ experiments that consisted in monitoring the temporal change of adsorbed-As concentration by incubating As-loaded goethite coated sand in the groundwater (Zhang, et al., 2017). Reactive transport models were developed using the Iphreeqc model coupling the geochemical code PHREEQC and MATLAB (Muniruzzaman & Rolle, 2016). The two surface complexation modeling approaches available in PHREEQC, the diffuse double layer (DDL) and the charge-distribution multisite complexation (CD-MUSIC) models (Hiemstra & Van Riemsdijk, 1996), were applied to simulate sorption competition assumed to be the only geochemical process leading to the release of As from goethite. Model parameters were calibrated through inverse modeling in order to simulate experimental results. Whereas a satisfying agreement with the measured As-adsorbed concentrations was obtained, the role of the aqueous species in the As desorption significantly differs between the predictions of the DDL and the CD-MUSIC models.

Speaker: Lucien Stolze
• 18:30
Visualizing and Quantifying Biomineralization in a Wellbore Analog Reactor 15m

Subsurface fluid injection is a proposed method for the storage of hydrocarbon fuels and the mitigation of fossil fuel emissions. Concerns about leakage exist when storing fluids in the subsurface given their potential to damage functional groundwater aquifers or be emitted to the atmosphere. Defects detrimental to the integrity of subsurface storage systems can occur in and around the wellbore, thus fluid storage systems are heavily dependent on the cement surrounding the wellbore to maintain a seal.

A method proposed to seal defects in the subsurface is Microbially Induced Calcium Carbonate Precipitation (MICP). MICP is a technique that uses low viscosity fluids and microorganisms (~2 µm diameter) to seal defects troublesome to subsurface fluid storage. In the MICP process, microorganisms such as Sporosarcina pasteurii that contain the enzyme urease catalyze the hydrolysis of urea to produce ammonium and carbonate species. When this process occurs in the presence of dissolved calcium, calcium carbonate may precipitate.

To study MICP in defects common to the wellbore, two reactors systems were created. The first was constructed to mimic the geometry of the wellbore and allowed the visual observation of MICP formation. The second quantified MICP in a cement channel defect using X-ray computed microtomography. A reduction in apparent permeability and void fraction was observed in both systems, demonstrating the ability of MICP to restrict fluid flow in defects common to the wellbore. Observations made during these experiments will aid in improving the safety and efficacy of subsurface fluid storage systems.

Speaker: Adrienne Phillips (Montana State University)
• 17:15 18:45
Poster 3: Poster 3-B
• 17:15
Porous media deformation and self-structuring under capillary bulldozing 15m

An experimental observation of the structuring of a granular suspension under the progress of a gas/liquid meniscus in a narrow tube is reported here. The granular material is moved and compactifies as a growing accumulation front. The frictional interaction with the confining walls increases until the pore capillary entry pressure is reached. The gas then penetrates the clogged granular packing and a further accumulation front is formed at the far side of the plug. This cyclic process continues until the gas/liquid interface reaches the tube’s outlet, leaving a trail of plugs in the tube. Such 1D pattern formation belongs to a larger family of patterning dynamics observed in 2D Hele-Shaw geometry. The cylindrical geometry considered here provides an ideal case for a theoretical modelling for forced granular matter oscillating between a long frictional phase and a sudden viscous fluidization.

Speaker: Dr. Guillaume Dumazer (University of Oslo, Porous Media Laboratory)
• 17:30
Diffusive processes across frictional patterns 15m

The progress of the interface between an invasive fluid and a defending mixture of granular material together with a second fluid, immiscible with the invading phase, has recently been used to form complex patterns in both Hele-Shaw cells or millifluidic confinements. These are the result of the deformation of a confined porous material, made of an homogeneous granular phase, into an heterogeneous structure exhibiting various sizes and permeabilities at different scales. A labyrinthine pattern can then be obtained after the withdraw of a liquid phase containing glass beads and confined in an horizontal Hele-Shaw cell, see Knudsen et al., PRE 77, 021301 (2008). What are then the transport properties within such a geometrical structure? The geometrical constraints with dead-ends are limiting the transport of diffusing species from a central entry point towards the edges of the labyrinth. A transient subdiffusive transport can be obtained over a time scale determined by the wavelength of the pattern.

Speaker: Dr. Guillaume Dumazer (University of Oslo, Porous Media Laboratory)
• 17:45
Mixed-dimensional modeling of the brain’s waterscape 15m

The clearance of the metabolic waste in the body is handled by the lymphatic system. Except in the brain, which appears to be the only organ devoid of lymphatic channels. Indeed, the mechanisms underlying the clearance processes of the brain are still unknown, and the topic sparks debate and controversy. What is clear however, is that dysfunction of cerebral metabolic waste clearance is associated with neurodegenerative disorders such as Alzheimer’s disease.

The term the brain’s waterscape refers to the circulation, flow and exchange of tissue fluid and transport of solutes through the brain. While these processes are not fully understood yet, most hypotheses point out the major role of the cerebral blood vessels and possibly paravascular spaces. We propose to investigate these processes using a mathematical approach based on coupled mixed-dimensional models mimicking the vasculature and paravasculature as topologically one-dimensional structures embedded in a three-dimensional porous medium.

This poster presents a mixed-dimensional model dedicated to the mesoscale-macroscale interaction between the brain tissue, the vasculature and the paravasculature, aiming at gaining new insight into the waste clearance process of the brain. The formulation and well-posedness of this model will rely on non-standard techniques such as weighted Sobolev spaces and non-local averaging operators used in [1].

Speaker: Cécile Daversin-Catty (Simula Research Laboratory)
• 18:00
Mixed methods for coupled 1D-3D flow models in porous media 15m

The physical processes governing flow and transport in porous media span a wide array of spatial scales. Furthermore, in many applications it would be computationally intractable to resolve each scale of interest. To still capture the effects of the smaller-scale processes, one option is to couple together models of different dimensionality. In e.g. vascularized tissue and root networks, the arteries and roots have negligible radius compared to their own lengths. They can therefore be viewed as 1D inclusions embedded in the 3D domain, and the system modeled using a coupled 1D-3D flow model.

This poster presents an analysis of the mixed finite element method applied to coupled 1D-3D porous media flow problems. Due to the high dimensional gap, the solutions will be singular around the inclusion. As a result, the problem eludes the standard $H^1$ framework, and weighted Sobolev spaces must be employed instead. We prove existence of a solution to the continuous and discrete model, obtain weighted error rates for the method, and present simulation results.

Speaker: Ingeborg Gåseby Gjerde (University of Bergen)
• Thursday, 17 May
• 08:30 10:17
Parallel 9-A
• 08:32
Phase Equilibria in CO2-Multicomponent Hydrocarbon Systems in Shale Organic Nanopores: A Coarse Grained Molecular Simulation Study 15m

Studying the phase behavior of complex hydrocarbon and hydrocarbon/CO2 mixtures in kerogen structures is extremely important for understanding the mechanisms involved in enhanced gas recovery, storage, and production of hydrocarbons from shale. The objective of this work is to determine the phase behavior of a number of binary, ternary, and multicomponent CO2/hydrocarbon systems using molecular dynamics simulation in three dimensional kerogen structure of type IIA. These systems include CO2/n-Hexadecane, CO2/n-Eicosane, CO2/n-Butane/n-Decane, and CO2/synthetic oil.
The kerogen molecule is prepared based on detailed structural analysis of NMR experiments. In order to build a representative solid state model of kerogen, eight kerogen molecules are placed in a periodic cubic cell. Once the initial configuration of kerogen molecules is prepared, constant-temperature constant-volume (NVT) simulations and then constant-temperature constant-pressure (NPT) simulations are performed to obtain the final structure. The SAFT-γ coarse graining methodology is used to develop force fields for the fluid-phase behavior of hydrocarbon/CO2 mixtures comprising CO2 and n-alkanes. The densities, phase compositions, and volume of each phase in kerogen are determined at different overall compositions and pressures and compared with unconfined case.
For the final kerogen structure, density values are calculated and compared with the reported density range for kerogen density. A good agreement is found between the experimental densities and our kerogen molecular structure density. Calculated phase equilibria for the unconfined system (where experimental measurements are available) are in fair agreement with experimental data. The liquid phase density did not change significantly in confined case (in kerogen) compared with that of unconfined (or bulk). However, the heavier components tend to vaporize in kerogen structure compared with the unconfined case. For example, n-decane have much higher composition in vapor phase in kerogen compared to bulk condition. Furthermore, as the pore diameter increases, the phase envelopes approach the bulk conditions.
Experimental measurements of the phase behavior of hydrocarbon mixtures in confined systems are extremely difficult, if not impossible, with current technologies. This work is one of the few in-depth investigations of the phase behavior in organic matters of shale. The results of this study can potentially modify the equation of state for shale reservoirs and help in understanding the transport mechanism in nanoscale pores.

Speakers: Dr. Ali Takbiri-Borujeni (West Virginia University) , Dr. Mohammad Kazemi (Kansas University)
• 08:50
Gas transport in porous geological media with contract of properties, and irregular distribution of pores. 15m

In this work, we perform multiscale modeling of gas transport through the heterogeneous solid having irregular pore structure and contrast of properties on different spatial scales. We assume that the solid consists of inorganic material (clay, sand) with organic (kegogen) inclusions imbedded into it. There exist a contrast of properties and spatial scales between the matrix and inclusions. The pore sizes vary from micro to nanometers, permeability and diffusivity can differ by several orders of magnitude. We consider filtration and molecular diffusion as mechanisms for the free gas transport, and surface diffusion as a main mechanism of gas transport through nanoporous organic inclusions. The irregularities of porous structure we characterize by their deviations from the regular (periodic) distribution.
Multiscale analysis is applied to mass balance equations, the equation of state for free gas, and an isotherm of adsorption. We focus on the upscaling from pore-scale to the core-scale and then from core-scale to reservoir scale.
As a result of upscaling, we get a macroscopic equation describing gas transport through an effective medium. It turns out that macroscale parameters characterizing gas transport depend on diffusivity, permeability, and porosity of components of the system, the amount of inclusions and their spatial distribution. We investigate sensitivity of macroscale parameters to deviations in pore distributions from their averaged values. We also determine the distribution of gas concentration through the production time and evaluate the production rate as a function of time.

Speaker: Dr. Viktoria Savatorova (UNLV)
• 09:08
Analytical Investigation of the Stability and Universal Scaling of the Transition from Spontaneous to Forced Imbibition in Porous Media 15m

Spontaneous imbibition is an important recovery mechanism in naturally fractured reservoirs as capillary forces control the movement of fluid between matrix and fracture. Imbibition is also important in unconventional reservoirs as the capillary pressure will increase when permeability decreases, impacting fracture fluid imbibition during the fracturing process but also during the soaking period before the initiation of production. However, the classic self-similar solution to spontaneous imbibition is limited in representing physical boundary conditions, as during most physical multiphase flow conditions flow occurs with contributions from both capillary and viscous forces. In this research, we present a theoretical semi-analytic approach to analyze the transient imbibition process where both capillary and viscous forces exist, and compare it with the self-similar solution. Unlike previous analyses that assume purely counter-current or co-current flow, this research proves that for a more general situation, strict self-similarity no longer exists, although a new universal relationship of imbibition rate versus time is obtained.
The transient imbibition boundary conditions we examine are easily achievable in the laboratory and are also comparable to those that will exist in a reservoir. We consider a model where the wetting phase is maintained in contact with the inlet, and hydrocarbon production is allowed on both ends. The hydrocarbon phase is produced from the outlet at a constant imposed rate. Initially, counter-current spontaneous imbibition caused by capillarity at the inlet dominates. As the flood front propagates, co-current flow gradually increases in importance as does the viscous force. The imbibition rate at the inlet will drop to be equal to a prescribed injection flow rate, after which the forced imbibition state is reached and viscous pressure drop totally controls the flow. Traditional Buckley-Leverett theory can then be applied to analyze the subsequent forced imbibition process.
The analytic solution for the transient imbibition process utilizes a fractional flow concept. In the current study, the fractional flow changes with time, as does the ratio of co-current and counter-current fluxes. It will be shown that not all choices of boundary conditions are stable, but that the wetting phase imbibition rate must increase at early time beyond any imposed injection rate to reach a limit of stability. The result of the analysis includes a dimensionless parameter that describes the relative magnitude of capillary and viscous forces at the continuum scale. A universal scaling envelope exists for the limit of stability which may be expressed in terms of this parameter and the dimensionless ratio of imposed fluxes at both ends of the system. Above the envelope, the flow is unstable as capillary pressure will cause the imbibition rate to increase and the dimensionless ratio to decrease. Any point below the envelope is stable and is subject to forced imbibition. The boundary of the envelope is the limit of stability, which describes the overall mechanism of transient imbibition and the relative magnitude of capillary and viscous forces at the continuum scale. This stability limit is different from the result obtained by the assumption of a self-similar solution.

Speaker: Mr. Lichi Deng (Texas A&M University)
• 09:26
Sorption Stresses in Organic-Rich Rock Formations: Fundamental Processes and Reservoir Scale Implications 15m

Hydrocarbon sorption in unconventional formations accounts for a significant amount of gas in place depending on rock mineralogy and organic content. Traditionally, sorption amount experiments have been performed on powders or crushed porous solids that ignore changes in pore structure.
Dilatometry experiments, however, show that changes of volume during sorption can result in dramatic changes to porosity structure that alter the organic-rich nanoporous solid. The maximum linear swelling strains in natural nanoporous media such as coal and shale caused by sorption of light hydrocarbons and CO2 vary from 0.05% to 1% and differ with respect to the orientation to deposition bedding.
An organic-rich nanoporous solid exposed to a sorbate and constrained to do not deform, develops sorption stresses rather than sorption strains. The sorption stress was measured in coal with values as large as ~40 MPa at a CO2 bulk pressure of 5 MPa. Direct measurements of desorption stress have also been done in coal with CH4.
In this work we show a comprehensive review of sorption-induced strains and stresses in organic-rich rocks. We use these laboratory measurements in an up-scaled reservoir model that permits quantifying the effects of sorption-induced stresses on fracture permeability and on production rates. We compare cases of coal and shale natural gas reservoirs.
The results show that gas desorption-induced stresses have a significant effect on the evolution of stresses in the reservoir and therefore on the permeability of natural and induced fractures.

Speaker: Prof. D. Nicolas Espinoza (The University of Texas at Austin)
• 09:44
Mineral precipitation in fractures: The role of aperture and mineral heterogeneity on the evolution of transport properties 15m

Fractures act as dominant pathways for fluid flow in low-permeability rocks. However, in many subsurface environments, fluid rock reactions can lead to mineral precipitation, which alters fracture surface geometry and reduces fracture permeability. In natural fractures, surface mineralogy and roughness are often heterogeneous, leading to variations in both velocity and reactive surface area. The combined effects of surface roughness and heterogeneous mineralogy can lead to large variations in local precipitation rates. The resulting alteration of transport properties defy description by currently existing continuum models. We present results from an integrated experimental and computational study aimed at quantifying the relative importance of aperture variability and mineral heterogeneity on local mineral precipitation rates.

Experiments were carried out in transparent analog fractures with one of the glass fracture surfaces initially seeded with small (~0.01 mm$^2$) localized CaCO$_3$ regions uniformly distributed over ~12% of the fracture surface. We considered two fractures, a smooth-walled fracture and a variable-aperture fracture, both with a mean aperture of about 0.1 mm. During flow experiments, a CaCl$_2$-NaHCO$_3$ solution that was supersaturated with respect to CaCO$_3$ (log($\Omega_{CaCO_3}$) = 1.44) flowed through the fracture at a constant flow rate of 0.5 mL/min. At weekly intervals during the months-long experiments, we paused flow of the reactive fluid and measured fracture aperture, tracer transport through the fracture, and reaction site distribution at high spatial resolution (83 x 83 $\mu$m) using light transmission techniques. Results showed that conductive pathways persisted for much longer times than predicted when assuming uniform surface reactivity. Furthermore, although the reactive surface area increased during both experiments, increased flow channeling led to a significant reduction in measured fracture-scale reaction rates.

To simulate reactive transport we used a quasi-steady-state 2D model that uses a depth-averaged mass-transfer relationship to describe dissolved mineral transport through the fracture and local precipitation reactions. Both aperture variability and mineral heterogeneity are explicitly represented in the model. Mineral-precipitation-induced changes to fracture surface geometry are accounted for using two different approaches: (1) by only allowing vertical growth at reactive minerals, and (2) by allowing three-dimensional mineral growth at reaction sites. Results from simulations using (1) suggest that as precipitation reduces local aperture, flow becomes increasingly focused into thin preferential flow paths. This flow focusing causes a reduction in the fracture-scale precipitation rate, and precipitation ceases when the reaction zone extends the entire length of the fracture. This approach reproduces experimental observations at early time reasonably well, but as precipitation proceeds, reaction sites can grow laterally along the fracture surfaces, which is not predicted by (1). To account for three-dimensional mineral growth (2), we have incorporated a level-set-method based approach for tracking the mineral interfaces in three dimensions. This provides a mechanistic approach for simulating the dynamics of the formation and eventual closing of preferential flow channels by precipitation-induced aperture alteration that do not occur using (1).

Speaker: Prof. Russell Detwiler (University of California, Irvine)
• 10:02
Sensitivity studies of different scenarios of polymer injection applied to Ainsa Quarry1 outcrop 15m

The presented work deals with polymer injection in an oil reservoir of which the geological organization is obtained thanks to an outcrop located close to the Ainsa town in southern Pyrenees, Spain. In this study the permeability distributions are not fixed. We address the following question: what is the impact of permeability distribution on the oil recovery considering a polymer slug injection. This question makes sense because the outcrop may be an analogue of different types of reservoirs. Indeed the outcrop helps to characterize the facies distribution but has a different geological story compared to a subsurface reservoir. Three different reservoirs are thus considered to be associated to three permeability distributions, respectively corresponding to a non-altered reservoir, a fractured reservoir and an unconsolidated reservoir. The second permeability model results from an upscaling step in order to take into account the presence of fractures. The large permeability values of the third permeability model are due to unconsolidated facies. Models are respectively called “permeable”, “fractured” and “unconsolidated”.

A water injection is modeled for 4 years with a pressure constraint of 500 bar. For each simulation, after a year of water flood, a polymer injection is carried out for two years. Finally, a water post-flush is injected until the end of the simulation. Polymer performance is tested against a water flood simulation whose final cumulated oil production equals to 0.2 hm3.

• Permeable model
By injecting 0.21 hm3 of water, 0.21 hm3 of oil is produced. Due to weak facies permeabilities pressure reaches quickly the imposed limit. As a consequence, the water injection rate decreases during the water flood and, even more strongly when polymer is added to the solution. Therefore polymer injection does not help to better produce oil in the permeable model.

• Fractured model
0.28 hm3 of oil and 0.2 hm3 of water are produced by injecting 0.65 hm3 of water. The produced water volume ratio (with and without polymer) is reduced by ~70%. Polymer injection is quite appropriate in this case since enhanced oil recovery reaches 40%.

• Unconsolidated sandstone
0.23 hm3 of oil is produced after injecting 0.87 hm3 of water. The high permeability allowed the injection of larger volumes of water and polymer solution. However, the oil recovery volume is decreased in comparison to the fractured case. An earlier water breakthrough and a higher produced water volume contribute to this bad performance.

As a conclusion, the polymer impact critically depends on the permeability distributions. If the reservoir permeability is too weak it will be difficult to inject polymer without damaging the reservoir or the injection wells. If the reservoir is too permeable, an earlier water breakthrough is observed which has a negative impact on oil production.

Speaker: André Fourno (IFPEN)
• 08:30 10:17
Parallel 9-B
• 08:32
Ink flow in fibrous layer: direct pore-scale modeling and experimental observation 15m

In this study, the detail flow of ink though the printing paper is simulated using pore-scale formulations. The exact 3D topology of an uncoated paper was obtained though micro-tomography imaging to reconstruct the domain with a resolution of 0.9 µm. Afterwards, the reconstructed domain was used for running direct numerical simulation of ink flow in paper. Confocal microscopy was applied to determine the spreading of ink and ASA measurements were used to determine the penetration depth. The simulation results showed a good agreement with the experimental observations. After validation, the impact of contact angle (CA) on ink spreading and penetration was studied using three values of 0, 60, and 120 ⸰. CA0 provided the maximum penetration depth while CA60 and CA120 caused movement of droplet and its deviation from the jetted location which is not favorable during printing. Water-based ink’s properties were applied to study the effect of ink additives on its spreading and penetration. The results have shown a slower spreading and penetration compared with using the pure water as the ink liquid.

• 08:50
Print curl of paper as a time-scale dependant process 15m

Paper curl due to wetting and drying is known to be determined by the degree of fiber swelling, the paper structure and material inner tensions from the paper production process which are released due to wetting. Apart from these well documented processes we have found that the development of paper curl is govererned by different mechanisms depending on the observed time domain.

Our investigation shows that different curl mechanisms are taking place during initial wetting, immediately after paper drying and during the first 48 hours after printing. Accordingle we are defining three different types of curl: initial wetting curl, short term print curl and long term print curl. Initial wetting curl is defined to take place in the first few seconds after contact with the printing ink. Short term print curl is defined as the paper curl occuring directly after printing and drying of the printed paper. Long term printing curl is the paper curl occuring 24 hours after printing and drying of the paper.

An analysis of the reasons for the development of paper curl in the different time scales revealed that initial wetting curl seems to be related to the fiber swelling, short term printing curl is related to the structure of the paper and long term curl is triggered by re-conditioning the paper after printing.

Speaker: Ulrich Hirn (Graz University of Technology)
• 09:08
Droplet Impact on Fabric 15m

textile materials is still not fully understood. Compared to a solid surface, on
textile the liquid can penetrate the holes in the fabric but also spontaneously flow through the porous networks inside the fabric (wicking), making droplet
spreading more complex compared to smooth surfaces. Understanding droplet
spreading on textile materials is important for applications in the textile industry and forensic research.

We study droplet impact on thin mono?lament polyester fabric as a function
of the fabric pore size and its wettability. First, the difference between droplet
spreading on a smooth surface (stainless steel) and the fabric is investigated
where the fabric is either placed on a substrate or suspended in the air. We
show that a droplet spreads less on the fabric compared to the smooth surface.
fabric with and without substrate due to the liquid penetrating the fabric. Sec-
ondly, droplet fragmentation of the penetrating liquid is investigated. Using
simulations, we determine the physical processes behind droplet spreading and
the subsequent fragmentation.

Speaker: Mr. Thijs de Goede (University of Amsterdam)
• 09:26
Transport processes and water based ink – paper interactions 15m

The inkjet technology fuels the rapidly evolving world of printing. This printing technology delivers good print quality using the flexibility of digital printing at a breakthrough cost price. The R&D department of Océ Technologies, a Canon company, is a major player in the development of inkjet technologies for many different applications.

Liquid spreading, evaporation and imbibition into porous material are physical processes that describe the interactions of aqueous ink with paper. Understanding them is vital to have prints of high quality; and this is the aim of this work. The influence of the liquid physical properties as well as of the paper characteristics will be considered. Experimental studies based on optical spectroscopy, microscopy, Scanning Electron Microscopy (SEM), Nuclear Magnetic Resonance (NMR) and Automatic Scanning Absorptometer (ASA) are presented within this work revealing the today level of understanding the transport of complex liquids into porous media. For each method of investigation we will present the main models, including their strengths and limitations.

Speaker: Nicolae Tomozeiu (Océ-Technologies B.V.)
• 09:44
Exploring the limits of macro-homogeneous models of carbon-fiber papers 15m

Carbon-fiber papers (CFPs) are an integral component of many energy-conversion and energy-storage technologies, including gas diffusion layers (GDLs) in polymer electrolyte membrane fuel cells (PEM fuel cells), cathode GDL in PEM electrolyzers and metal-air batteries, and as electrodes in redox flow batteries (RFBs). CFPs must fulfill several functions such as providing adequate mechanical support to the membrane electrode assembly, a transport pathway for reactants/products through its pore volume, and electrical and thermal conductivity through its solid fibrous structure. In RFBs, they have the added functionality of providing an active catalytic surface area. Understanding of the transport processes that occur in CFPs is necessary to help predict cell performance and durability, optimize materials and diagnose problems. The most common technique used to model these thin porous media is the macro-homogeneous approximation based upon the existence of a representative elementary volume (REV). However, the applicability of the continuum approach to CFPs has been questioned many times, and the error introduced in the predictions is certainly unclear.

In this work, the limitations of macro-homogeneous descriptions of CFPs are explored for various single-phase transport processes: diffusion, convection and electrical/thermal conduction. Multiple sub-domains with different widths and thicknesses are examined by combining the lattice Boltzmann method with X-ray tomography images of uncompressed CFPs. The results show that a REV cannot be defined due to the lack of a well-defined separation between pore and volume-averaged scales in these inherently thin heterogeneous materials. The representative size in the material plane is in the order of 1 mm, which is comparable to the rib/channel width used in the above-mentioned devices. As for the through-plane direction, no representative length scale smaller than the thickness can be identified. In particular, it is found that the highly porous surface region (amounting up to 20% of the material) can significantly reduce the through-plane electrical/thermal conductivity.

In the second step, fuel cell performance predictions using a pore-scale model are compared to a traditional macro-homogeneous model. The results show that the overall mass, heat and charge fluxes predicted by both models are similar, provided that the macro-homogeneous model is equipped with suitable effective properties. However, the spatial distributions can be significantly different due to the lack of separation of scales. Specifically, the pore-scale simulations reveal the presence of inhomogeneities in the transport of species and heat that are not accounted for by the macro-homogeneous model. These deviations could have a substantial impact when modeling degradation phenomena, which are sensitive to local conditions.

• 10:02
On modeling partially-saturated flow of a liquid in multilayered thin swelling porous media 15m

Understanding fluid flow and deformation processes in thin swelling porous media is critical for developing superior consumer absorbent hygiene products such as wipes, paper towels, feminine pads and diapers [1-4]. Fluid-flow models have proven very valuable for the development of these products and have led to the development of fundamental understandings in transport mechanisms, numerical simulation tools, computation infrastructure and lab methods for both characterizing absorbent materials as well as validation of flow and deformation models.
In this study we developed a quasi -2D averaged macroscopic mass balance model, based on the volume averaging approach [5-17], for modeling partially-saturated flow of a liquid in multilayered thin, absorbing swelling porous media. In order to describe the absorbency process in [18], fast and accurate simulations with this model are carried out to predict the time and spatial behavior of variables such as piezometric head, saturation, porosity, and layer thickness, and to understand the flow and storage of a liquid in conjunction with the layer deformation. This model enormously improved the computational speeds, allowing to develop a fast and reasonably accurate simulation of the unsaturated flow at lower cost. The numerical results of the simulations predicted well the flow fields of both liquid and solid phases and were in good agreement with the experimental and previous numerical results.

Speaker: Dr. Ahmed Kaffel
• 10:45 12:17
Parallel 10-A
• 10:47
Pore-scale simulation of mass transfer across scCO2-water interface using phase-field method 15m

Interfacial mass transfer between scCO2water in porous media is a key process for dissolution and mineral trapping of CO2 during geological storage of CO2. Recently, both core- and pore-scale drainage and imbibition experimental studies have shown non-equilibrium dissolution of scCO2 and an extended depletion of residual scCO2 (Chang et. Al. 2016, 2017). For better understanding and quantifying the dissolution process of we need models to capture the rate-limited dissolution of scCO2 at the pore level. In this work, we develop a simulation model to capture two-phase flow and interfacial mass transfer at the pore-level. We used a continuum approach for interfacial mass transfer along with a phase-field method as an interface-capturing technique for the two-phase flow system. The model has been applied to a single pore for validation and subsequently to numerical simulation of 2D homogeneous micromodel experiments. The preliminary results show that the model can capture the mass transfer between the two phases and at the same time the interface adjustment.

Speaker: Farzad Basirat (Department of Earth Sciences, Uppsala University)
• 11:05
Solutal convection in bead packs 15m

I will, present recent experiments and numerical simulations of solutal convection in bead packs. The experiments utilise an analog system of Methanol-Ethylene-Glycol (MEG) and water and the numerical simulations use high-order spectral methods. We study the effect of mechanical dispersion of the solute on the convective pattern and the quasi-steady solute flux. Our results show that mechanical dispersion is the dominant dissipative process in bead packs, where It controls the pattern and affects the solute flux.

Speaker: Marc Hesse (The University of Texas at Austin)
• 11:23
The impact of horizontal groundwater flow on the dissolution of CO2 in saline aquifers 15m

The dissolution of supercritical CO2 in aquifer brine is one of the most important trapping mechanisms in CO2 geological storage. As supercritical CO2 is less dense than the ambient groundwater, the injected CO2 is susceptible to leakage in case that the sealing layer is not perfectly impermeable. However, when CO2 is dissolved in water it is not buoyant anymore. In fact, CO2-saturated water is slightly heavier than CO2-free water. This situation where CO2-free water is overlaid by heavier CO2-rich water, leads to a hydrodynamic instability in which fingers of dense CO2-rich water are formed and propagate downwards, causing the CO2-free water to move upwards [1,2]. This convective process accelerates the dissolution rate of CO2 into the aquifer water.
The majority of previous studies assumed there is no natural groundwater flow in the aquifer and neglected the associated hydrodynamic dispersion and therefore assumed there is no effect on the dissolution dynamics. However, it was found that in some of the saline aquifers considered for CO2 storage groundwater flow rate, although small, is not zero [3]. A few studies investigated numerically the effect of groundwater flow and dispersion on dissolution dynamics [4,5] but no experimental evidence was provided yet.
In this research, we study the effect of groundwater flow on dissolution trapping by performing laboratory experiments and conducting numerical simulations. Experiments were performed in a physical aquifer model using a mixture of methanol and ethylene-glycol (MEG) as a CO2 analog while varying the water horizontal flow rate. Simulations were then carried out to reproduce experimental results. We found that water horizontal flow has a significant effect on the dynamic of the instability and the fingers morphology. As the horizontal flow increases, the number of fingers, their wavenumber and their propagation rate decrease. In high water flow rates, no fingers were developed and the dissolution process was driven by diffusion and dispersion alone. While the classic dissolution behavior, consisting of a diffusive regime followed by a convective regime was clearly observed, the effect of water flow on the dissolution rate did not show a clear picture. When increasing the horizontal flow rate, the convective dissolution flux slightly decreased and then increased. It seems that when horizontal flow rate increases, there is a tradeoff between the decay of instability which suppresses dissolution and the increase in dispersive flux and fresh water inflow which enhances dissolution.

Speaker: Ravid Rosenzweig (Geological Survey of Israel)
• 11:41
Effect of heterogeneity on the mixing of fluids under convective flow 15m

Mixing in the presence of convective instabilities in an homogeneous porous media is governed by the behavior of stagnation points where the fluid interface is stretched and compressed. It has been shown that an interface compression model is able to predict the behavior of the scalar dissipation rate. The mixing regimes experienced by these kind of systems are linked to the dependency of compression on diffusion and the interaction between stagnation points and the correlation structure of the velocity field. We apply this approach to an heterogeneous porous media in which the variations in heterogeneity distorts the convection patterns and the way the fluid interface is compressed. We consider a Rayleigh-Bénard instability in which the stagnation points are at a fixed interface and Rayleigh-Taylor instability in which the interface is mobile. Using a stochastic approach we perform a series of numerical simulations using randomly generated conductivity fields realizations with varying statistical properties. Numerical results are used to analyze the impact of variance and spatial correlation of conductivity fields on the way the fluid interface is compressed and on the mixing behavior of the system. The flow structures are visualized by the strain rate and characterized by their correlation length.

Speaker: Juan J. Hidalgo (IDAEA-CSIC)
• 10:45 12:17
Parallel 10-B
• 10:47
Mathematical modeling of microstructured membrane filters: A stochastic approach 15m

Membrane filters have been widely used in industrial applications to remove contaminants and undesired impurities from the solvent. During the filtration process the membrane internal void area becomes fouled with impurities and as a consequence the filter performance deteriorates, which indeed depends on filter internal structure, particles concentration and flow. The complexity of membrane internal morphology and stochasticity of particles flow make the filtration process and fouling mechanisms a mysterious phenomenon and hard to study. Therefore, mathematical modeling can play a key role in investigating filter fouling and discovering efficient filtration process. So far various mathematical models have been proposed to describe the complexity of membrane structure and stochasticity of particles flow individually but very few focus on both together. In this work, we present an idealized mathematical model, in which a membrane consists of a series of bifurcating pores, which decrease in size as the membrane is traversed and particles are removed from the feed by adsorption within pores (which shrinks them) and stochastic sieving (blocking by large particles). We discuss how filtration efficiency depends on the characteristics of the branching structure.

Speaker: Dr. Pejman Sanaei (CIMS-NYU)
• 11:05
On the Pressure Generation anRelaxation in a Porous Media under a Spherical Loading Surface 15m

The phenomenon of pressure generation and relaxation inside a porous media is widely observed in biological systems. For example, the pressurization inside the cartilage plays the key in the load bearing and lubrication of the knee joints. In this paper, we report a biomimetic study to examine the transient pressure distribution inside a soft porous layer when a spherical loaded surface suddenly impacts on it. A novel experimental setup was developed that includes a fully instrumented spherical piston with supporting structures, and a soft fibrous porous layer underneath. The materials were precisely characterized on their porosities, pore sizes, fiber stiffnesses and permeability. Extensive experimental studies were performed with different porous materials, different loadings and different sized loading surfaces. The pore pressure generation and the motion of the loading surface were recorded by pressure transducers and laser displacement sensors, respectively. A novel theoretical model was developed to characterize the process of the pore pressure generation and relaxation underneath the loading surface and inside the undeformed surrounding materials. Excellent agreement was observed between the experimental results and the theoretical predictions. It clearly demonstrated that the hydrodynamic similarity between porous media on different scales are governed by the Brinkman parameter, $\alpha=h/{K_p^{0.5}}$, where $h$ is the porous layer thickness, and $K_p$ is the undeformed Darcy permeability. The study significantly improves our understanding of the dynamic response of soft porous media under rapid compression, which has broadd impact on the study of transient load bearing phenomenon in biological systems and industry applications.

Speaker: Dr. Rungun Nathan (Penn State Berks)
• 11:23
From Red Cells to Soft Porous Lubrication 15m

In this paper, we report a novel experimental study to examine the lubrication theory for highly compressible porous media (Feng & Weinbaum, JFM, 422, 282, 2000), which was applied to the frictionless motion of red cells over the endothelial surface layer (ESL). The experimental setup consists of a running conveyer belt covered with a soft porous sheet, and an upper planar board, i.e. planing surface. The pore pressure generation was captured when the planing surface glides over the porous sheet. If the lateral leakage was eliminated, we found that the overall pore pressure’s contribution to the total lift, $f_{air}$≈80%, and the friction coefficient $η=0.0981$, when $U=5$ m/s, $L=0.381$ m, $λ=h_2/h_0=1$ and $k=h_2/h_1=3$, where $U$ is the relative velocity of the conveyor belt; $L$ is the length of the planing surface; $h_0$, $h_1$ and $h_2$ are the undeformed, leading edge and trailing edge porous layer thickness, respectively. $f_{air}$ increases with the increase in $U$, $λ$ and $L$, while decreases with the increase in $k$. $η$ decreases with the increase in $f_{air}$. If later pressure leakage exists, the pore pressure generation is reduced by nearly 90%. All the experimental results agreed well with the theoretical predictions. The study presented herein lays the foundation for applying soft porous media for new type of bearing with significantly reduced friction.

Speaker: Qianhong Wu (Villanova University)
• 12:15 13:45
Poster 4: Poster 4-A
• 12:15
Experimental measurement of CO2 diffusion coefficient in water based nanofluids 15m

The objective of this paper is experimental measurement of CO2 diffusion coefficient in different nanofluids including (SiO2), aluminum oxide (Al2O3) and titanium oxide (TiO2) nanofluids. Nanofluids in concentrations of 0.05, 0.1 and 0.2 wt% were used in experiments. Different factors such as temperature, weight percentage of nanoparticles, as well as the effect of particle size were investigated on CO2 diffusivity. The results clearly show that at 25˚C, the highest CO2 diffusivity occurred for 0.2% concentration of TiO2 nanofluid which is 6.5×10-9 m2/s. Also, at 30˚C, the highest CO2 diffusivity was obtained for 0.05% concentration of SiO2 nanofluid which is 7.8×10-9 m2/s. In addition, the highest CO2 diffusivity for 0.1% concentration of TiO2 nanofluid at 35˚C was 14.7×10-9 m2/s. Finally, a new correlation was developed for the effective diffusion coefficient of CO2 in nanofluid based on the nanofluids Reynolds (Re) number, nanoparticles Reynolds (Renp) number, Schmidt (Sc) number and nanoparticles volume fraction with an average relative error percent (REP) of 9.58% and R2 of 0.945.

Speaker: Prof. Reza Azin (Faculty of Petroleum, Gas and Petrochemical Engineering (FPGPE), Persian Gulf University (PGU), P.O. Box 75169-13817, Bushehr, Iran )
• 12:30
Pattern Formation and Mixing Dynamics in Three-Dimensional Non-Boussinesq Solutal Convection 15m

Motivated by the process of convective mixing in porous media, here we study the pattern-formation and coarsening dynamics arising from dissolution of CO2 in a single-aqueous phase during three-dimensional (3D) Rayleigh-Benard-Darcy convection. Our focus is on comparing the pattern-formation aspects of solutal convection between conditions of constant concentration and constant flux prescribed at top boundary. In the constant-flux case, a low (constant) injection-rate of CO2 is considered, such that all CO2 dissolves and the system remains indefinitely in single phase. We adopt non-Boussinesq, compressible formulation, whereby the nonlinear phase behavior and density variation of mixture as a function of pressure, temperature, and solute concentration are fully accounted for. We perform high-resolution finite element simulations of 3D convective mixing to examine the interplay between the density-driven hydrodynamic instability and the pattern formation at top boundary as well as various cross sections. We find that gravitational instabilities---triggered at the boundary layer by the local increase in density following CO2 dissolution---further develop into columnar plumes of CO2-rich brine because of self-organization of concentration field in the diffusive boundary layer as a cellular network structure irrespective of top boundary type. While such pattern formation occurs at top boundary for the constant-flux case, it can be captured only below the top boundary for the constant-concentration case. By studying the statistics of the cellular network and concentration field, we identify various regimes of finger coarsening over time for both model types, and show the existence of a quasi-steady state where the average cell size at top boundary, or below top for the constant-concentration case, remains constant due to an equilibrium between cell division and merging in correlation with a constant-mixing rate regime. The morphology of mixing patterns in three dimensions for different types of boundary conditions has implications to numerous density-driven problems in fluid mechanics as well as in environmental and geological settings.

Speaker: Amin Amooie (The Ohio State University)
• 12:45
A numerical simulation study on the hydraulic fracture propagation in heavy oil reservoir with the THM coupling 15m

For further research on the effect of heavy oil viscosity on the fracture geometry, this paper establishes heavy oil fracturing model and conventional fracturing model based on thermal-hydraulic-mechanical (THM) coupled theory, Walther viscosity model and K-D-R temperature model. We take viscosity and density within heavy oil fracturing model as functions of pressure and temperature, while that as constants within conventional fracturing model. A heavy oil production well is set as an example to analyze the differences between the two models in account of thermo-poro-elastic effect. The results show that temperature has the greatest influence on heavy oil viscosity, while pressure presents the least influence; A cooling area which has 0～1meters width and varied length generates near the fracture. Heavy oil viscosity increases sharply in this area, presenting an area of viscosity increment. Heavy oil viscosity increases faster closer to wellbore, and a large viscosity increment will reduce the mobility of heavy oil and prevent fracturing fluid from bumping into reservoir; The special viscosity distribution results in significant differences of pore pressure, oil saturation and changing trend between these two models; In heavy oil reservoir fracturing model, the thermal effect completely exceeds the influence of pore elasticity, and the values of the fracture length, width and static pressure are larger than those calculated in conventional fracturing model. Thus the change of heavy oil viscosity plays a dominant role in influencing the expansion of hydraulic fracture.

Speaker: Qiang Wang
• 13:00
Response of Relative Permeability to Coal Surface Chemistry through Steady-State Core Flooding Measurements using X-ray CT Scanner and Packed Bed Samples 15m

The relative permeability behaviours of gas and water in coal are primary factors in the productivity of a coal seam gas reservoir, and it is dependent on many factors including fluid saturations and pressure, cleat geometry and network, and wettability (surface chemistry). In this study, we performed steady-state relative permeability measurements using X-ray CT scanner on packed beds of coal particles to allow systematic investigation of coal surface chemistry on permeability behaviour. Packed bed approach provides a homogeneous, isotropic coal sample with controllable coal surface properties; pore size and channel geometry (by control of particle size), and also removes the natural cleat geometry effects from the experimental measurement.
In this paper, we compare packed cores made with coals from two different locations within the Bowen Basin, Queensland, Australia. Both samples have the same rank (1% random reflectance), but with different maceral composition. The sample from Broadmeadow (BRDM) is vitrinite rich (76% mineral-matter free basis) and the sample from Isaac Plains (IP) is inertinite rich (61% mmfb) – both petrographic and proximate analyses were carried out. Samples were constructed using coal particles with size between 53 – 212 μm, resulting in a packed bed with porosity around 19%.
Core flooding experiments were conducted at a constant effective pressure of 22 bar followed by a sensitivity analysis using a hypothetical coal seam gas reservoir simulation model. The CT images were used to calculate water saturation based on the grey-scale without relying on the dead-end volumes of the core flooding system. The relative permeability curves suggest that the vitrinite rich coal (BRDM) had a more water-wet behaviour than the inertinite rich coal (IP), where the crossover point occurs at (krw=krg=0.45; Sw = 0.62), for the BRDM, and at (krw=krg=0.32; Sw = 0.44) for the IP. Hypothetical simulation using Eclipse shows that the cumulative water production for the IP can be 40% higher than the BRDM over a period of 30 years of gas production, due to its water wet behaviour.

Speaker: Mr. Fabio Terzini Soares (The University of Queensland)
• 12:15 13:45
Poster 4: Poster 4-B
• 12:15
Modeling of liquid infiltration in stack of thin fibrous layers and their interfaces 15m

Given the fact that thin layer is not satisfying REV length scale requirement, it is essential to better understand how to model the fluid flow process in such a medium. Recently, Qin and Hassanizadeh [1,2] introduced a new thermodynamically approach for modeling multiphase fluid in a stack of thin porous layers which is called: “Reduced continua model (RCM)”. All equations in this model are derived in terms of thickness average properties. The main objective of this study is to compare RCM to the traditional Darcy based models and improves it in order to better predict the liquid infiltration in stack of hydrophilic thin porous layers and their interfaces. Furthermore, water dynamic in imbibition process in stack of two thin fibrous layers has been simulated considering a layer-layer mass transfer coefficient. Final results of numerical simulation using RCM model show reasonable agreement compared to laboratory data. Additionally, the computational effort is decreased by one order of magnitude comparing to traditional models.

References:
1.Qin, C.Z. , Hassanizadeh, S.M. A new approach to modelling water flooding in a polymer electrolyte fuel cell, international journal of hydrogen energy 40 (2015) 3348-