IOR NORWAY 2018Smart solutions for future IOR

Table of ContentsThe Speakers:Tor Bjørnstad 4Bergit Brattekås 4Mojdeh Delshad 5Dario Grana 5Alfred Hanssen 6Jarle Haukås 6Aksel Hiorth 7Aojie Hong 7Matthew D. Jackson 8Morten Jakobsen 8Lesley James 9Signe Kjelstrup 9Gunnar Hjelmtveit Lille 10Danielle Morel 10Ann Muggeridge 11William R. Rossen 12Hans Christen Rønnevik 12Sigmund Stokka 13Ana Todosijevic 13Arvid Østhus 13

The Posters:Pål Østebø Andersen 14Mirza Baig 15Dhruvit Berawala 16Alberto Bila 17Tine Bredal 17Harry Collini 18Aditya Dixit 18Samuel Erzuah 19Katherine Esquivel 20Hossein Fazeli 20Siri Gloppen Gjersdal 21Kun Guo 21Edgar Hernandez 22Anastasia Ivanova 22Shaghayegh Javadi 23Emanuela Kallesten 23Emilie Aasen Kavli 24Mona Wetrhus Minde 25Remya Nair 25Jan Inge Nygård 26Oddbjørn Nødland 26Karen Synnøve Ohm 27Aruoture Omekeh 28Eystein Opsahl 29Oleg Pichugin 30Irene Ringen 30Jaspreet S. Sachdeva 31Dmitry Shogin 31Mario Silva 32Saeed Sofla 33Edison Sripal 33Rowena Su Wen Shi Thu 34Odilla Vilhena 35Tijana Voake 36Yiteng Zhang 37Siv Marie Åsen 37

Programme 38

The Workshop 39

Partners & Observers 41

THE SPEAKERS

Experiments on the core scale can be used to test ideas and develop EOR methods. Numerical modelling can validate proposed hypotheses and explain the mechanisms and flow processes observed in experiments. Although often investiga-ted separately, the combination of experiments and modelling is necessary to accurately predict EOR on larger scales: by implementing new knowledge from core scale experiments into numerical simulator tools. We have used special core scale experiments and the newly developed simulator tool, IORCoreSim for this purpose, and present a multidisciplinary collaboration between several institutions, where synergy between experiments and modelling allowed improvement of both. Numerical modelling was used to identify weaknesses in the experimental setup, allowing adjustments of the experi-mental design and procedures. Likewise: experiments were designed to measure specific parameters, necessary for furt-her mathematical description and modelling. The goal of the ongoing collaboration is to accurately include EOR methods and process mechanisms in numerical simulators; first on core scale, thereafter on reservoir grid and field scale. Two subjects are specially investigated: spontaneous imbibition and polymer gel behavior.Spontaneous imbibition, where the wetting phase enters the matrix by capillary forces to spontaneously displace the non-wetting phase, is an important oil recovery mechanism in naturally fractured reservoirs. Recently, spontaneous imbibi-tion of solvent (brine bound in gel) from formed polymer gel into an adjacent, oil-saturated porous medium was observed in core scale experiments. Polymer gel can significantly reduce

fracture conductivity, and its efficiency determines the success of EOR chase floods. Dehydration and shrinkage of gel caused by spontaneous solvent imbibition may cause

gel treatments to be less efficient than expected in water-wet reservoirs; because the gel will no longer fill the entire fracture volume. Polymer gel is an inherently complex fluid. Experimen-tal observations and numerical description and quantification of gel properties and behavior on the core scale was, however, necessary to predict the performance of gel placed in an oil field, and required multidisciplinary efforts. An original model-ling approach was developed by the team and implemented into a newly developed core scale simulator. Simulations of spontaneous imbibition from gel was performed, and compa-red to free spontaneous imbibition of water, achieving a good overall match between experiments and simulations. New field developments also require knowledge of sponta-neous imbibition mechanisms in unconsolidated media. We have used high permeability sand packs to perform imbibition experiments with combined co-/counter-current imbibition and interpreted the results using the core scale simulation software IORCoreSim, in particular incorporating the capilla-ry back pressure. The simulations supported and explained experimental observations of inlet/outlet artifacts in the sand packs, and several necessary adjustments in the experimental design were proposed.

Core scale EME for IOR: experiment- modelling- experiment

Bergit Brattekås is currently a postdoc and project leader at The Na-tional IOR Centre of Norway. She holds MSc and PhD degrees in reservoir physics from the Univer-sity of Bergen, during which she also worked as a visiting researcher at New Mexico Tech. Research interests: spontaneous imbibition, mobility con-trol by foams, polymers and gels, CO2 for EOR and in-situ investigations of pressure and saturations during multiphase flow in fractured and porous media using MRI, CT and recently PET-CT.

Nanoparticles as oil detectives?Nanoparticles (NPs) come in different shadings. Some NP types may be applicable to trace mass flow. The mass flow may be aqueous or organic liquids and solids of different kind. The flow may take place in confined pipelines or vessels, in open conduit in the geosphere (rivers, lakes, oceans) or in subsurfa-ce geological structures like ground water flow or fluid flow in oil reservoirs or in geothermal systems.The NP types include simple metals, metal oxides, quantum dots like ZnS or CdSe, core/shell and core/shell/shell structu-res, carbon quantum dots (C-dots), dendrimers, nanogels, functionalized «smart» NPs with metallic or metal oxide core surrounded by silica and attached fluorophores for ease of detection or even more complex “hybrid” NPs with special functionality (hydrophilicity, lipophilicity or surfactant-like). Some nanoparticles may be based on a solid core of various metals or metallic compounds surrounded by various functio-nalization shells, and may even be superparamagnetic. This presentation will briefly review NPs of potential interest for industrial mass flow studies, but especially discuss whether

or not they have a potential as tracer components for quanti-tative measurement of remaining (residual) oil saturation in a reservoir after secondary recovery. Can they be implemented for interwell use or single-well/near-well use? Which are the perceived and/or experienced limitations or challenges? Which answers do we have today, and which prospects are possible for further NP technology development to meet the challen-ges? What does it take to convert the question mark in the title of this presentation to an exclamation mark?In the end of the presentation I have taken the liberty to spe-culate somewhat on these subjects.

Tor Bjørnstad is at present Special Advisor at IFE within reservoir technology, and Prof. Em. in nuclear chemistry at University of Oslo. Main interests: Tracer technology, IOR and flow as-surance. He holds a PhD (Dr. Philos.) in Nuclear Chemistry from UiO.

4

Recent Advances in Chemistry and Design Expand Applicability of Polymers to High Temperature/High Salinity

in Sandstone and Carbonate ReservoirsPolymer flooding is still one of the most widely used chemical EOR methods due to its operational simplicity and low cost and its applicability to onshore and offshore reservoirs. High salinity-high temperature reservoirs are favourable targets for polymer flooding. Field studies show that polymer can produce incremental recovery up to 20% of OOIP in poorly swept reser-voirs, decrease water cut, and accelerate production. However, polymer flooding is currently limited due to a combination of reservoir temperature, make-up brine salinity and hardness, since synthetic polymers can chemically and thermally degrade at these conditions. There are a wide variety of polymer pro-ducts with different performance characteristics which make the selection of an appropriate product challenging. HPAM type synthetic polymers are today industry standard due to low cost and large-scale availability. However, they suf-fer instability at high temperature and salinities. They require large doses to be effective as, increasing salinity decreases the viscosity. They suffer from mechanical degradation in valves and chokes. Finally hydrating these polymers requires instal-lation of large facilities which are often impractical, especially offshore. On the other hand biopolymers such as Scleroglucan (SG) and Schizophyllan (SPG) are more robust than synthetic polymers, but have been limited by production scale concerns and high retardation in low permeability reservoirs. These EOR grade biopolymers are stable at temperatures up to 120°C, in brines of virtually any salinity/hardness. We directly compared the SG in terms of rheology, viscosifying power, and thermal stability to HPAM/AMPS copolymers at various conditions. We described various transport characte-

ristics and limitations for both determined from core and field studies.There have also been major developments in advancing capa-bilities of polymer property models such as (1) viscoelastic rhe-ology and impact of elasticity on remaining oil saturation using an empirical relaxation time correlation, (2) polymer hydrolysis and degradation mechanisms, (3) polymer injectivity and scale up to the field, (4) effective finger models for polymer flooding of viscous oils. These models are supported by carefully desig-ned laboratory experiments.To assess these polymer potentials in EOR applications, we de-veloped simulation scenarios based a real-world reservoir. We evaluated SG and HPAM/AMPS for injectivity and oil recovery performance. We discussed the impact of viscosity, rheology, adsorption, and stability on their performance. The simulation results show that polymer flooding can have wider applicability when it is well designed and the differences between polymer classes can have a significant impact on EOR performance and should be carefully considered before pursuing a field project.

Mojdeh Delshad is the president/CEO of Ultimate EOR Services and a research professor in the department of Petrole-um and Geosystems Engineering at the University of Texas-Austin. Her experience and research interests are in modeling petrophysical proper-ties, modeling and simulation of chemical, gas, conformance control, and low salinity enhanced oil recovery methods. She has 80 refereed journal publications and co-authored 125 conference papers all about EOR processes.

Stochastic approaches to seismic reservoir characterization for improved modeling and prediction

D. Grana (University of Wyoming), T. Bhakta (IRIS), R. Lorentzen (IRIS), X. Luo (IRIS), R. Valestrand (IRIS), G. Naevdal (IRIS)

Seismic reservoir characterization generally focuses on the prediction of reservoir properties, including porosity, mineral volumes, pre pressure, and fluid saturations, from seismic data. Fluid flow simulation predictions strongly depend on the static reservoir model, in terms of average properties such as total pore volume and hydrocarbon in place as well as the spatial distribution of the reservoir properties. In particular, different spatial features of porosity and permeability models can lead to different predictions of hydrocarbon production, water cut, and bottom hole pressure. Porosity models are generally estimated from seismic data, whereas the associated permeability models are computed from the porosity using deterministic or statistical relationships. The spatial distribu-tion of porosity and permeability can then be updated during the production phase, every time new measured data become available, including production data at well locations as well as time-lapse seismic surveys.In the first part of this work, we discuss recent advances in seismic reservoir characterization. Several stochastic inversi-on methods have been proposed for the estimation of static reservoir properties. In this work, we focus on a new inversion approach based on ensemble methods and data order redu-

ction techniques. This method can generate multiple realiza-tions of porosity, as well as other rock and fluid properties, conditioned by the measured seismic data. In the second part of the talk, we investigate recent developments in the field of seismic history matching to update the initial porosity and permeability model by simultaneously assimilating producti-on and time-lapse seismic data. The advantage of stochastic optimization methods lies on the ability of capturing the uncertainty in the model predictions by either estimating the posterior probability distribution of the model parameters of the inversion problem or generating an ensemble of stochastic realizations of the model variables. Examples of applications to Norwegians fields will be shown to illustrate the proposed methodologies.

Dario Grana is assistant professor at Department of Geology and Geophysics at University of Wy-oming. He is a geophysicist with a background in mathematics and statistics. His research inte-rests: rock physics, seismic reservoir characteri-zation, inverse problems, geostatistics, and time-lapse reservoir modeling. Grana got the EAGE’s Young Professional Award 2017. 5

Seismics in the Arctic: ice and bubble generated noise

To increase the recovery factor beyond that of natural depleti-on, injection support is required. An important challenge in op-timizing the injection sweep is the presence of zones of enhan-ced permeability, sometimes referred to as thief zones. These zones can be laterally continuous stratigraphic features such as channels and / or conductive fault and fracture networks. To assess the impact on fluid flow and to optimize the injection support, thief zones must be characterized and represented in the reservoir simulation model with sufficient detail.In this paper, we consider a naturally fractured chalk reservoir. We focus our study on permeability characterization of condu-ctive faults and fracture corridors which are visible as discon-tinuities in the seismic data. 4D seismic can also reveal these features due to fluid changes and changes in pore pressure. The mapped seismic discontinuities are characterized in terms of (1) confidence level, (2) azimuth, (3) planarity (how well a plane would represent the discontinuity) and (4) associated 4D changes, using automated seismic characterization tools. The objective is to establish a relationship between these characte-ristics and the permeability field that controls the fluid flow.The reservoir simulation model used in the study is a modified version of the actual simulation model for the field. The fluid model is black-oil, thermal effects are not included and the compaction model only considers pressure effects. Further-

more, local grid refinements are introduced around the most important fault and fracture zones in the southern part of the field, close to relevant wells. The purpose of the local grid refinements is to discretize seismic

discontinuities mapped with 12.5 meters lateral sampling. The refinement level is selected such that the fault and fracture connectivity is preserved while keeping simulation run times at acceptable levels. Finally, an ensemble based history matching framework is invoked to explore the relationship between the seismic discontinuity characteristics and the permeability field within the local grid refinements. As part of the process, the model realizations are evaluated against available production data and 4D seismic data.The presented work highlights the value of automation in reservoir characterization and history matching, enabled by tight integration between 4D seismic data and locally refined reservoir simulation models.

Analysis of enhanced permeability using 4D seismic data and locally refined simulation models

J. Haukås1*, W. Athmer1, J. Ø. H. Bakke1, Q. D. Boersma1,2, A. Bounaim1, M. Etchebes1, P. G. Folstad3, B. H. Fotland1, R. Moe3, C. Pacheco3 and E. Tolstukhin3

1Schlumberger Stavanger Research, 2TU Delft, 3ConocoPhillips

Jarle Haukås is Reservoir Management and IOR Program Manager at Schlumberger Stavanger Research. He has a PhD in Applied Mathematics from the University of Bergen (2006), where he did his thesis on compo-sitional reservoir simulation. After 6 months as a PostDoc, he joined Schlumberger Stavanger Rese-arch in 2006. In Schlumberger Stavanger Resear-ch he is involved in 4D seismic analysis and history matching, reservoir simulation, geomechanics and salt body extraction from seismic data.

Alfred Hanssen, ARCEx, Dept. of Geosciences, Univ. of Tromsø – The Arctic University of Norway

Seismic sources in the vicinity of sea ice can generate a multi-tude of wave modes in the ice sheet. These waves are often of large amplitude, and in general, they overlap in time, frequ-ency, space, and wavenumber with the weaker information carrying seismic reflection signal. In this presentation, we will discuss methods for identification and subsequent reduction of ice generated noise in seismics. For seismic arrays, the interference of source pressure waves reflected from the sea surface may generate high-frequency noise due to ghost cavitation and subsequent collapse of the vapor bubbles. We will describe the highly nonlinear and inter-mittent dynamics of the bubbles, and characterize the associa-ted acoustic noise.

Alfred Hanssen is a professor of geophysics at the Dept. of Geosci-ences at the University of Tromsø – The Arctic University of Norway (UiT). He serves as the Director of ARCEx (The Research Centre for Arctic Petroleum Exploration), and he is the Vice Dean for Innovation at the Faculty of Science and Technology, UiT. Alfred holds a PhD in

mathematical physics from UiT, and he has held research positions at the Max-Planck-Institute für Aeronomie, at the Los Alamos National Laboratory, at the European Commission Joint Research Centre, at Colorado State University, at the University of Oslo, and at the University of Tromsø. He has worked in research based in-dustry in several periods, and he prefers to work in multi- and cross-disciplinary environments.

6

Simulation Tools for Predicting IOR Potential on the Norwegian Continental Shelf

What would be the best time for IOR? – Fast analysis in a decision analysis framework

Complex production models such as grid-based reservoir simulation models are useful but may not be tractable for decision analysis. Simple or proxy production models are po-tentially very useful and tractable because they are computati-onally attractive whilst still providing insight to the decision at hand. Useful and tractable models are required for supporting high-quality decisions in uncertain, complex, and computatio-nally demanding contexts.A key decision for development planning is: What is the opti-mal time to initiate an Improved-Oil-Recovery (IOR) process? This presentation illustrates the implementation and applica-tion of a useful and tractable approach for the analysis of the optimal IOR switch time using a two-factor production model and Least-Squares Monte Carlo (LSM) simulation.The two-factor production model contains only two parame-ters for each recovery phase. One parameter describes how much recovery efficiency a recovery mechanism can ultimately achieve, and the other describes how fast the recovery effici-ency increases. The simplicity of the model makes it compu-tationally attractive. The LSM algorithm is an approximate dynamic programming approach, which allows for learning over time. It provides a near-optimal decision policy for the IOR switch time problem.Closed-Loop Reservoir Management (CLRM) is considered to be a state-of-the-art approach to solving for the optimal IOR switch time. However, this approach can produce a subop-timal solution as the CLRM approach considers only uncer-

tainties and actions reflecting currently available information but not those uncertainties and actions arising from future information. Our Sequential Reservoir Decision Making (SRDM) approach achieved by using the LSM algorithm considers both the impact of the information obtained before a decision is made and the impact of the information that might be obtai-ned to support future decisions. We conclude that compared to CLRM, SRDM can significantly improve both the timing and value of decisions, leading to a significant increase in a field’s economic performance. Further-more, the two-factor model combined with the LSM algorithm is tractable and provides useful insight into the IOR switch time problem.

Aojie Hong holds a PhD degree in Ensemble-based History Matching, Roboust Production Optimization and Value-of-Information. He is

currently working as a Researcher in Reservoir Engineering & Decision Analysis at University of Stavanger (UiS). Hong was the first student to defend his doctoral degree at The National IOR Centre of Norway. Despite his young age (29 years), he has already published several scientific articles. He has also contributed gre-atly to the cooperation between the academic communities at the University of Stavanger and the University of Texas, UT Austin.

The Norwegian Petroleum Department (NPD) estimates a technical potential of the order of 320-860 MMSm3 for En-hanced Oil Recovery (EOR) on the Norwegian continental shelf (http://ressursrapport2017.npd.no/). Out of 13 EOR methods investigated, the NPD concludes that polymer combined with low salinity water injection is the most promising method for enhanced oil recovery (EOR). In this talk we will give an overview of the work in the IOR Centre that is being performed in order to reduce the risk of applying these methods offshore. The first part of the talk will focus on polymeric liquids. Polymeric liquids decreases the mobility of water and can therefore improve the sweep of water in the reservoir and improve the recovery of oil. Due to the non-Newtonian behavior of these liquids, they are extre-mely challenging to model. We have developed models that are capable of describing the most commonly observed flow regimes in porous media: (i) Newtonian, (ii) Shear thinning, (iii) Shear thickening, and (iv) Mechanical degradation. The time constants for the shear thinning and shear thickening behavior are related to variations in reservoir properties and conditions, thus making it possible to translate lab results to larger scale without introducing new fitting parameters. These models can be used to suggest practical injection protocols for various polymer molecules.

The second part of the talk focus on low salinity water/smart water flooding, and on the importance of proper interpretati-on of core scale experiments. The simulation work in the area of low salinity and smart water is clearly lagging behind the experimental investigations. Significant effort has been done in the IOR Centre to include most of the important chemical ef-fects into simulation models, such as adsorption, ion exchange, precipitation/dissolution, surface complexation (surface poten-tial), CO2 dissolved in the oil phase. We are currently working on lattice Boltzmann methods on the pore scale, and Darcy scale models on the core and field scale, and in this talk we will outline our latest progress, and ideas in these areas.

Aksel Hiorth is Chief research scientist within enhanced oil recovery (EOR) at IRIS and Professor within reservoir technology at the Univer-

sity of Stavanger. Currently he is research dire-ctor at The National IOR Centre of Norway. He has a PhD within theoretical physics from Univer-sity of Oslo, and has been principal investigator within several large research projects supported by the industry and the Research Council of Norway. In the last decade he has mainly worked with developing simulation models that can des-cribe the physical and chemical processes taking place during multiphase flow in porous rocks.

7

Time-lapse full-waveform inversion as a smart monitoring tool for future IORWith the development of new innovative geophysical met-hods, it should be possible to further increase oil recovery from existing fields. This makes sense from an environmental as well as an economic perspective. Seismic methods generally allow one to image sedimentary basins and reservoirs using similar principles as in medical imaging. The term time-lap-se seismic refers to repeated seismic acquisition over time. Time-lapse seismic methods capture the dynamic behavior of the reservoir and aid reservoir management. The main value of time-lapse seismic methods is the additional information to update a model of the reservoir, and the localization of undrai-ned oil and flow barriers, which is important for well planning.In the geophysical community, seismic full-waveform inversion (FWI) has recently emerged as the final and ultimate solution to the Earth’s resolution and imaging objective. Seismic FWI can potentially include a complete computational simulation of the true field experience, in which the observed data is repro-duced using the inverted model of the subsurface. However, seismic FWI faces many challenges, including the sensitivity of the FWI results to the starting model and it’s huge computati-onal cost.In this talk, the speaker shall address the main challenges of FWI in general and time-lapse FWI in particular. This includes a discussion about different strategies for dealing with time-lap-se waveform data as well as the use of modern domain de-

composition and renormalization methods for reduction of the computational cost and improved convergence, respectively. Special attention will be given to the use of

Bayesian methods for time-lapse full-waveform inversion that provides information about uncertainties as well as the most likely values of the elastic parameters in the underground. Such uncertainty information is required in existing systems for quantitative integration of seismic waveform and production data designed to give an improved reservoir management.The talk will close with an illustration of analogies between seismic full-waveform and electromagnetic inversion methods. This is interesting since electromagnetic and seismic waveform data complement each other and we have experienced great synergies when developing inversion methods at the interse-ction between these two geophysical domains. In any case, the integration of different data types is essential for any attempt to develop a really smart solution for future IOR; and time-lap-se waveform inversion represents the state-of-the-art and the future when it comes to monitoring of fluid movements in reservoirs under production.

Morten Jakobsen(1,2,3), Kjersti S. Eikrem(2,3) and Geir Nævdal(2,3)

(1) Department of Earth Science, University of Bergen.(2) International Research Institute of Stavanger

(3) The National IOR Centre of Norway

Morten Jakobsen is professor at the Department of Earth Science of the University of Bergen. He also has a 20 percent position at IRIS under The National IOR Centre of Norway to strengthen the Centre’s competen-ce within mathematically oriented petroleum geo-physics. Professor Jakobsen has a strong track record in rock physics, seismic wave-form inversion and integrated petroleum research. He also has a degree in physics that make him very flexible when it comes to initiation of interdiscipli-nary projects.

Zeta potential changes at mineral-brine and oil-brine inter-faces control improved oil recovery during smart waterflooding

It is well known than oil recovery can be increased by modify-ing the composition of injected brine in a process sometimes termed ‘smart’ waterflooding. However, the pore- to mine-ral-surface scale mechanisms responsible for improved oil recovery (IOR) have not yet been identified and there is no method to predict the optimum injection brine composition for a given crude-oil-brine-rock system. Recent experimental work has demonstrated that rock wet-tability and improved oil recovery (IOR) during smart water-flooding are strongly correlated to changes in zeta potential at both the mineral-water and oil-water interfaces. In these experiments, rock wettability correlated with changes in zeta potential after aging in formation brine. Moreover, IOR during smart waterflooding occurred only when the change in brine composition induced a repulsive electrostatic force between the oil-brine and mineral-brine interfaces. The results suggest that the polarity of the zeta potential at both interfaces must be determined when designing the optimum injection brine composition. Moreover, they also suggest that the zeta po-tential at the oil-water interface may be positive at conditions relevant to many hydrocarbon reservoirs. A key challenge for any model of smart waterflooding is to explain why IOR is not always observed. We suggest that failures using the conventional (dilution) approach to smart waterflooding may have been caused by a positively charged

oil-water interface that had not been identified. Further expe-rimental data supported by modelling is required to under-stand the dependence of the zeta potential at the oil-brine interface on parameters such as oil composition, brine compo-sition and temperature. The zeta potential of the oil-brine interface is usually assumed to be negative at reservoir conditi-ons but this is not the case for many crude oils and brines.

Matthew D. Jackson is Professor in Geological Fluid Mechanics and Director of Research in the Department of Earth Science and Engi-neering, Imperial College London. His research interests include de-velopment and application of new reservoir modelling and simulation methods, development and application of new reservoir monitoring and surveillance techniques for production optimization, and novel technology for improved and enhanced oil recovery. Jackson leads the Novel Reservoir Modeling and Simulation (NORMS)

and the Smart Wells and Reservoir Monitoring (SWARM) groups at Imperial College. He also teaches on the Department’s MSc courses in Pe-troleum Engineering and Petroleum Geoscience. He has published over 100 scientific and technical papers across a broad range of journals and con-ferences. He holds a B.S. degree in physics from Imperial College and a Ph.D. in geological fluid mechanics from the University of Liverpool.

8

The Impact of Digitalization of SCAL on Field Development

Energy efficiency in the process industry: Learning from natureElisa Magnanelli1 Mari Voldsund2 and Signe Kjelstrup31 Department of Chemistry, NTNU, Trondheim, Norway

2 SINTEF Energy Research AS, Trondheim, Norway3 PoreLab, Department of Chemistry, NTNU, Trondheim, Norway

We have found that the heat and mass exchange in the reindeer nose serves the animal well; the construction of the nose mucosa is such that it regains body heat and water much better than simpler nose constructions [1]. Minimum entropy production (minimum lost work or lost exergy) can mean ma-ximum energy efficiency in terms of the second law of thermo-dynamics [2]. When this applies - one may find that the path of operation is characterized by uniform entropy production [2].The exergy value of the hydrocarbon stream that enters or leaves a production platform is huge. For some North Sea platforms now in operation [3], the order of magnitude is 104 MW. The specific exergy destructed on a platform itself in the upstream processing part is also large, 757 MJ/Sm3 oil equiva-lents in one of the platforms mentioned. Using thermodynamic optimization procedures, one may find a state of operation with smaller losses (minimum entropy production) [2]. A new membrane process for CO2 separation will be used to illustrate this [4]. We conclude with some general guidelines for energy efficient design.Acknowledgement: SK is grateful to the Research Council of Norway through its Centres of Excellence funding scheme, project no. 262644, PoreLab. EM and SK acknowledge financial support from the Research Council of Norway and user partners of HighEFF, a Centre for Environ-ment-friendly Energy Research, project no. 257632/E20.[1] E. Magnanelli, Ø. Wilhelmsen, M. Acquarone, L. P. Folkow, S.

Kjelstrup. The efficiency of a reindeer nose: The Nasal Geometry of the Reindeer Gives Energy-Efficient Respiration, Non-Eq. Thermodyn. 42 (2016) 59-78 doi: 10.1515/jnet-2016-0038[2] S. Kjelstrup, D: Bedeaux, E. Johannessen, J. Gross, Non-Equilibrium Thermodynamics for Engineers, World Scientific, 2.ed., 2017[3] M. Voldsund, T.-V. Nguyen, B. Elmegaard, I. S. Ertesvåg, A. Rø-sjorde, K. Jøssang, S. Kjelstrup, Exergy destruction and losses on four North Sea offshore platforms: A comparative study of the oil and gas processing plants, Energy, 74 (2014) 45-58.[4] E. Magnanelli, E. Johannessen, S. Kjelstrup, Entropy production minimization as design principle for membrane systems: Comparing equipartition results to numerical optima, Ind. Eng. Chem. Res. 56 (2017) 4856–4866.

Signe Kjelstrup is a physical chemist. She is a professor at the Depart-ment of Chemistry, Norway’s Technical and Natural Sciences Univer-sity. She graduated as a civil engineer at the Norwegian Technical

College in 1971, wrote a licentiate dissertation in 1974 and a doctorate in 1982. She was appointed Professor at NTH in 1986 and has since 2005 also held a part-time position as a professor at the Technical University of Delft. Her scientific work has been of great importance in irreversible thermodynamics with and espe-cially electrochemical cells, membrane systems and energy optimization for process-chemical applications. 9

Digitalization and the automation of extractive based indus-tries, such as mining and oil and gas, has re-emerged from the 90’s lean manufacturing. High power computing, artificial intelligence, and high resolution sensors, i.e. sub-micron ima-ging have advanced in leaps and bounds in the last 20 years. These advancements have and are being integrated into all oil and gas operations to increase productivity, lower costs, and shorten timelines. Digitalization and automation are having an impact on core analysis and specialized core analysis (SCAL) as well. In this session, we will explore the impact of digitalization on core analysis and SCAL. Can we get data and make faster/better decisions regarding formation evaluation using digital rocks/digitalization? What is the cost/benefit analysis of digi-talized core analysis? How are digital rocks, machine learning, and automation being used and what is the future potential in core analysis and SCAL? The technology is changing, as are the work flows, but so too are the tasks/roles and education requ-ired by the people performing digital SCAL. Ultimately, we will question the role of core analysis and SCAL in field evaluation and EOR screening.

Dr. Lesley James is an Assistant Professor and Chevron Chair in Petro-leum Engineering in the Faculty of Engineering and Applied Science at Memorial University. Her research focuses on enhanced and improved oil recovery and her time is split between research, teaching under-graduate and graduate level courses in process engineering and oil and gas, and various committees and outreach.Dr. James received her bachelor of applied science in chemical engineering and a diploma in technology management and entrepre-neurship (1997) from the University of New Brunswick as well as a master’s (2003) and PhD (2009) in chemical engineering, both from the University of Waterloo. Between degrees, Dr. James worked as a management consultant with Accenture for more than four years with work assignments in Canada, the United States and Europe.

Dr. James’ research interests focus on sustainable oil production by increasing oil recovery rates through enhanced oil recovery (EOR). Current-ly, her focus is on maximizing recovery from offshore Newfoundland and Labrador oil and gas fields through understanding the fluid-fluid and rock-fluid interactions and particularly examining miscible/near-miscible fluid injection and optimal EOR strategies for offshore production.

The promise of technology to unlock value on the NCSA rapidly growing global population needs affordable energy with reduced emissions to avoid unacceptable climate chan-ges. Norway is well positioned to supply energy to the global market as more than 50% of the resources on the Norwegian Continental Shelf (NCS) remain to be produced. However, the NCS production will have to compete in a market where the future petroleum demand is uncertain and where low costs and low emissions are essential.OG21 is a collaboration between universities, research institutes, industry enterprises and governmental bodies to develop robust technology strategies that secure the competi-tiveness of the NCS as well as of the Norwegian supplier indus-try. It addresses technology to overcome physical limitations, as well as technology needed to improve cost-efficiency and productivity, and reduce emissions. In addition, the OG21-strategy describes barriers to technology development and implementation, such as competence and technology externalities, risk aversion and inefficient business and contract models. The OG21-strategy puts forward recommendations on me-asures and actions to address technology needs and reduce

technology barriers, e.g.: sufficient public support of R&D and joint industry initiatives to reduce externalities, and changes to industry business models to engage suppliers earlier and improve risk/reward-distributions between oil companies and technology providers.The potential value related to the development and implemen-tation of new technologies is vast. In addition to reducing GHG emissions substantially, increased production on the NCS to a value of up to 5000 billion NOK could be realized over the pe-riod until year 2050, according to estimates in the OG21-stra-tegy. A continued high emphasis on R&D and technology implementation would be required to achieve such results.

Gunnar Hjelmtveit Lille is Director for the Division for Energy, Resources and the Environment in Department for Petroleum Research in The Re-search Council of Norway. He is also director for OG21, the Norwegian oil and gas strategi.

Innovation and Standardization for an agile EOR deploymentThe oil price has always been unstable, and highs and lows have even been more severe, in this century with a blooming of EOR teams and projects early 2000’ when the oil price ste-adily increased up to 100S/bbl., and the drastic cost exercise that was recently implemented when it went down to 50 and even lower. Such an environment is particularly detrimental to a long term EOR strategy. The situation may also be alarming when we look at the energy mix changes in the 20 years to come; Fossil fuels will still be a significant component of that mix, but with a strong decrease for oil contribution, and an in-creased contribution of gas, including shale gas. In this compe-titive environment the future of EOR is to be more innovative, and still implement some standardization for cost efficiency, and agile deployment. Innovation is to be found in multiple domains. One domain is to push the existing technologies to untouched domains. Some examples arose recently for example in the Middle East where single well surfactant polymer pilots have proven successful in high salinity high temperature carbonates, thanks to the development of new chemicals. Hybrid mechanisms are also to be piloted, that could unlock massive stakes.Innovation is also a matter of boldness, and whereas for deca-des the only offshore EOR was gas injection, other options are increasingly deployed and polymer EOR is now to play a major role offshore.Last but not least innovation is compulsory to develop new concepts for a more sustainable EOR, and multiple examples are illustrating how our industry is improving. Synergies bet-ween CO2 capture and storage, and CO2 EOR are there. Water management is a key concern as water is a resource, just as oil

is a resource. Intensive R&D and piloting is achieved to recycle produced water, and minimize disposal. Solar energy is now used, when applicable, to generate steam

for thermal EOR projects, saving gas consumption and minimi-zing greenhouse gas emissions.Cost is one of the major limitation to an extensive deployment of EOR. Standardization by the development of ready to go and versatile equipment is a smart way to go faster and cheaper to field deployment. Keeping in mind that the sooner the better, this approach is also a way to bring more incremental barrels on stream. Even if our industry suffers from a highly unstable oil price environment, there is a future for EOR. It is in our hands, in our brains, and most importantly in the way we will keep attracting young talents.

Danielle Morel has a doctorate in engineering and is a graduate of the École nationale supérieure des ingénieurs en arts chimiques et techn-ologiques (ENSIACET). She wrote her thesis in thermodynamics on the modeling of gas condensate at the IFPEN (IFP Énergies nouvelles) at Paris VI University.After ten years at the IFPEN as a research engineer, Danielle joined Total in 1991, working with the Laboratory of Miscible and Compo-sitional Processes. In 1996, she joined the Reservoir Evaluation and Management division, where she became a specialist in modeling

gas injection. In terms of research, Danielle was responsible for the subsurface aspect of the major project on acid gas reinjection and carbon sequestration in 2000.In 2003, E&P asked her to rebuild an EOR team (Enhanced Oil Recovery: improved hydrocarbon recovery). In 2010, Danielle became Total’s EOR expert, working with subsidiaries, as well as with R&D and Management.

10

Screening for EOR and Estimating of Potential Incremental Oil Recovery on the Norwegian Continental Shelf

P.C. Smalley, A.H. Muggeridge (Imperial College London), M. Dalland, O. S. Helvig, E. J. Høgnesen, M. Hetland, A. Østhus (Norwegian Petroleum Directorate)

We present an improved screening tool to assess the potential of different EOR processes in candidate fields and estimate the associated incremental oil recovery. The tool can be applied to groups of fields to allow ranking of those fields with the greatest opportunity as well as identifying EOR processes that show the greatest promise across the portfolio. Effort can then be focussed on detailed studies of the most promising fields, exploring synergies between groups of fields or further resear-ch and development of specific EOR processes. The tool uses up-to-date screening criteria combined with an improved screening scoring system and a better estimation of incremental oil recoveries compared to previous published methods. The EOR processes screened for are: hydrocarbon miscible WAG, hydrocarbon immiscible WAG, CO2 miscible and immiscible WAG, alkaline, polymer, surfactant, surfactant/polymer, low salinity, low salinity polymer, thermally activa-ted polymer gels and conventional near well gel treatments. Sliding-scale scores are used for individual screening criteria, weighted for importance and then combined to derive the overall screening score. It is possible to allocate non-zero sco-res when non-critical criteria are outside their desired range to avoid the problem of processes being ruled out completely if rock or fluid properties are only marginally outside the thres-hold of applicability.The methodology calculates the expected increment (and uncertainty range) for each EOR process based on data from existing field implementations of those processes, capped by theoretical maximum recovery factors. These maximum incre-ments are calculated from theoretical/laboratory values for

microscopic displacement efficiency and macroscopic sweep efficiency.The new tool was used to estimate the potential EOR oppor-tunity for 53 reservoirs from 27 fields on the Norwegian Continental Shelf (NCS). The average recovery factor for this area is already much higher (47%) than most other petrole-um provinces. The results indicate an EOR technical potential of 592 million standard cubic metres (scm) with a range of 320-860 million scm. The most promising processes are low salinity with polymer, surfactant with polymer, and hydrocar-bon and CO2 gas injection. Some field clusters were identified that could provide economies of scale for such processes.These results have enabled the Norwegian Petroleum Dire-ctorate to advocate further EOR-technology studies, including pilots, in specific regions or fields. Pilots are very important as they narrow the uncertainty range for the EOR potential of a given process in that field as well as exploring feasibility. Testing, qualifying and deploying new and existing IOR met-hods within a reasonable time will help prevent substantial oil volumes on the NCS remaining unrecovered.

Ann Muggeridge is a Professor of Reservoir Physics and EOR, Dept. Of Earth Science and Engineering, Imperial College London. Professor Muggeridge’s research focuses on methods for improving oil recovery. Following her DPhil she worked at the then BP Research Centre, follo-wed by a service company (SSI (UK) Ltd) before joining Imperial College in 1995. From 2006-08 she was a Technology Fellow at BP.

11

A Laboratory Study of Foam for EOR in Naturally Fractured Reservoirs

W. R. RossenDept. of Geoscience and Engineering, Delft University of Technology

Naturally fractured reservoirs contain large reserves of hydro-carbons worldwide. These reservoirs have very poor sweep efficiency in gas-injection EOR processes, however. Foam is one possible means of improving gas sweep efficiency, but foam generation and propagation in natural fractures is not well understood.Our goal is to characterize foam generation and propagation in natural fractures in a way that extends to a wide variety of fracture apertures and other properties. To that end we examined foam generation and propagation in model fractures formed by two glass plates, one roughened and one smooth. This allowed direct observation of pore-level mechanisms, together with pressure measurements, over relatively long distances (40 cm), with a wide variety of fracture geometries. These model fractures can be characterized as 2D networks of pore bodies (locations of wide apertures) and throats (conne-ctions between bodies). An initial study demonstrated a new definition of capillary number for mobilization of trapped bubbles or droplets of nonwetting phase in fractures, that fits a variety of fracture geometries better than the conventional definition. This defini-tion may help in understanding the effect of fracture geometry on foam properties. We studied foam generation and propagation in nine diffe-rent model fractures of widely different apertures and scales of roughness. Bubble generation was observed by the same mechanisms as reported for other porous media. The relative importance of snap-off and lamella division to foam generation depends on the geometry of pore throats and bodies and on gas fractional flow. As in previous studies, pressure gradient correlates inversely with the average size of the bubbles.

In some cases, we find in our model fractures the same two foam flow regimes central to the understanding of foam in 3D porous media: a low-quality (wet) regime, where pressure gradient is independent of the liquid velocity, and a high-qua-lity (dry) regime, where the pressure gradient is independent of the gas velocity. The mechanisms behind these regimes are different in our fractures from those thought to control the same regimes in 3D porous media, however. In the low-quality regime, bubbles are not fixed at size larger than pores, becau-se diffusion is relatively slow to combine bubbles in this 2D geometry. The high-quality regime is controlled, not by foam stability or capillary pressure, but by fluctuating foam-generati-on rate. Pressure gradient is very small in model fractures with large aperture, as expected.

William R. Rossen is Professor In Reservoir Engineering, Delft Uni-versity Of Technology. He was formerly Professor at The University of Texas at Austin, and before that a research engineer at Chevron

Oil Field Research Co. He has more than 90 peer-reviewed journal publications. Professor Rossen’s current research concerns use of foams for diverting fluid flow in porous media, modeling complex transport processes in networks, and understanding flow in naturally fractured geological formations. In 2012 he was named an IOR Pioneer at the SPE/DOE Symposium on Improved Oil Recovery, Tulsa, OK, and he is a Distinguished Member of SPE.

Most fields on the NCS were impossible before proven other-wise by the drillbit.Several of the breakthroughs are results of commitment wells. Major breakthroughs have occured decades after the areas have been considered mature and assumed understood. Continued adaption to reality through delineation, develop-ment and production and feedback to exploration improve the knowledging process. The difference between companies is in the relations between humans involved. A teleological culture is needed.The technology and data will always be Incomplete and prediction of reality should be scenario based. The quality of models are depending on the conceptual input. The technology breakthroughs occured after the exploration success. The technology has to be mastered by acknowledging

the limitations.

Unfolding the reality from plays to prospects and fields is continued learning based on awareness of conceptual,

factual and technological IncompletenessHans Christen Rønnevik - Career history1945: Born in Haugesund Norway1971: University of Bergen cand.real (MSc. degree in geology)1971: Research assistant in marine geology at UiB1972/73: Petroleum geologist Ministry of Industry1973-83: Petroleum geologist in various positions in the Norwegian Petroleum Directorate. 1983/1984: Senior petroleum geologist in Norske Shell1984-99: VP Exploration in Saga Petroleum implementing an organic

growth strategy 2000-04: Exploration Manager and geologist in a revitalized DNO in Norway. 2004-15: Exploration Manager and geologist in Lundin Norway 2015- : Senior G&G advisor in Lundin Norway

Member of Norwegian Academy of Technolo-gical Sciences – NTVA. NTVA honorable price 2013, AAPG Norman Forster Award 2015

12

DrillWell – Drilling and Well Centre for Improved RecoveryThe Centre for Research-based Innovation DrillWell has obtai-ned results within Drilling Process Control and Optimisation, Geo-steering, Life-cycle Well Integrity, Plugging and Abandon-ment of wells, and Slot Recovery. Results are contributing to cost reduction in exploration, field development, field opera-tion and decommissioning. Cost efficient wells give a valuable contribution to improved recovery of oil and gas by connecting additional reservoir segments, allowing marginal fields near existing infrastructure to be developed, as well as extending the life of producing fields. Low cost wells could be necessary for realising Enhanced Oil Recovery projects.DrillWell is financed jointly by the Research Council of Norway and the industry group AkerBP, ConocoPhillips, Statoil and Wintershall until summer 2019. It is a joint R&D effort with IRIS

(host), SINTEF, University of Stavanger and NTNU. Extension of DrillWell beyond 2019 is being discussed with the industry.

Sigmund Stokka is Manager at DrillWell & Senior Vice President at IRIS Energy AS. He is a civil engineer (NTH in 1977) in technical physics, and Dr. Ing. from the same place in 1981. The doctorate is in the subject of structural phase transitions in crystal systems of type oxides (KMnF3, SrTiO3). A new theoretical-ly predicted critical crossover exponent for the transition between two phases was determined for the first time. As part of the doctoral thesis, a new automated calorimetric laboratory was built for measurement of thermal capacity.

EOR Competence along the Value ChainAna Todosijevic, Head of EOR at Wintershall will present about the EOR Competence along the Value Chain: to successfully execute EOR project, any operator needs to ensure the presen-ce of the EOR competence along the value chain, starting in the assessment phase of cEOR Suitability, Subsurface Evaluat-ion, selection of the suitable chemistry, Upscaling, Field Pilots to Logistics and the impact on the oil and water processing. Throughout the value chain many different partners are invol-ved, excellent communication, integration and close collabora-tion is crucial for any successful project. Ana is going highlight key factors that can influence the success of any (EOR) project and share the recommendations for faster delivery, based on EOR project experience of Wintershall EOR team.

Ana Todosijevic graduated as Engineer of Chemistry and Techno-logy at the University of Applied Sciences in Monchengladbach and joined Wintershall in 2013. As Head of the Enhanced Oil Recovery

(EOR) Program, Ana manages the worldwide EOR activities of Wintershall since 2016. In her previous position, Ana managed the technical EOR cooperation with partners, including the biopolymer Schizophyllan field application. The decade before, Ana worked for major compa-nies in the chemical industry with special focus on Oilfield / EOR products in several positions such as Senior Business Development Manager, Marketing Manager, Product Line Manager and Sales Manager.

The Norwegian Continental Shelf – Even more to gainSignificant value remains to be extracted from the Norwegi-an continental shelf (NCS). The NCS resource accounts as of 31.12.2017 will be presented together with status of reserves and expectations for future reserves. The Norwegian Petrole-um Directorate (NPD) plays an important role in communica-ting facts about and knowledge of Norway’s petroleum sector. The big remaining resources in discoveries and fields, and the opportunities these offer for a continued high level of value creation to the benefit of society over many years to come will be highlighted.NPD believes it is important that the companies continue to take a long-term approach to their work on the NCS. They must think beyond their own production licenses, take an integrated view of a wider area and collaborate across license and company boundaries. The authorities expect all resources which can create value for society to be produced, not only the «easy barrels». In addition to the volumes in the resource accounts the NPD has identified a technical potential for substantial quantities of oil and gas. These could be accessed by adopting enhanced oil recovery (EOR) methods and taking opportunities for efficient production of tight reservoirs. Such petroleum would not be

produced under current plans, but could be recovered by app-lying new technology.Norway’s petroleum industry has been a global front-runner in developing offshore technology. The NCS has been a labora-tory for testing new technology. Norway must now become a front-runner in adopting the solutions which have been de-veloped. Norway has a strong offshore technology community. It needs to ensure that this is also maintained in the future. The need for pilots and field tests will be highlighted.

Arvid Østhus is Assistant Director, Development and Operations – Northern North Sea at Norwegian Petroleum Directorate.

Education:MSc. Reservoir Technology – University of Stavan-ger (1989)Experience:Arvid Østhus has been with the NPD since 2014, and joined the management team in January 2017. From 1990 to 2014, he held both management and technical positions at Phillips Petroleum (now ConocoPhillips). 13

THE POSTERSCapillary Pressure Correction of the EOR Potential during

Low Salinity and Smart Water FloodingPål Østebø Andersen, Dep. of Energy Resources, University of Stavanger / The National IOR Centre of Norway

Many experimental works have investigated smart water and low salinity water flooding and observed significant incremen-tal oil recovery following changes in the injected brine compo-sition. The commonway approach to model such EOR me-chanisms is by shifting the input relative permeability curves, particularly including a reduction of the residual oil saturation. The capillary forces trapping the oil are assumed to release oil from the smaller pores due to a favorable shift in wetting state following the chemical alteration with the injected brine. Counter-intuitively, this is not modelled using the capillary pressure function, which is further often ignored in modelling of such tests. Cores that originally display oil-wetness will re-tain much oil at the outlet of the flooded core due to capillary pressure being zero at a high oil saturation (see the Figure). This issue is difficult to overcome in high permeable cores at typical lab rates. Injecting a brine that changes the wetting state to less oil-wet conditions (represented by zero capillary pressure at a lower oil saturation) will lead to a release of oil previously trapped at the outlet. Although this is chemically induced incremental oil, it represents a reduction of remaining

oil saturation, not necessarily of residual oil saturation.This paper illustrates the mentioned issues of interpreting the difference in remaining and residual oil saturation during che-mical EOR and hence the evaluation of potential smart water effects. We present a mathematical model representing core flooding accounting for wettability changes due to changes in the injected composition. For purpose of illustration, this is performed in terms of adsorption of a wettability alteration component coupled to shifting of relative permeability and capillary pressure curves. The model is parameterized in accor-dance with experimental data by (a) applying sets of consistent saturation functions from the same wetting conditions (b) further comparing the model to experiments where wettability alteration takes place dynamically due to changing one chemi-cal component.By use of simple techniques the smart water effect can be determined very precisely and distinguished from the capillary end effects. Also, an a priori estimate of the magnitude of the end effects is proposed.

14

Petrography Properties Evaluation by Digital Image Processing of Thin Sections

M.H. Baig1, A. Nermoen1,2,4, P.H. Nadeau1, P.Ø. Andersen1,2, I. Austvoll3 1Department of Energy Resources, University of Stavanger, Norway

2The National IOR Centre of Norway, Stavanger, Norway3Department of Electrical Engineering and Computer Science, IDE, University of Stavanger

4International Research Institute of Stavanger, Norway

Economic feasibility of any field development largely depends upon its reservoir storage and flow capacity. Porosity, satura-tion and permeability are important parameters to determine the type and volume of hydrocarbons in place and to estimate recoverable reserves. They are also key parameters in plan-ning, modelling and development of a reservoir. The porosity and saturation of the reservoir can be determined with reasonable certainty through interpretation of petrophy-sical logs or through analysis of physical core samples. Perme-ability evaluation is challenging, because the definition and the scale of this measurement varies across its sources. Logs provide an empirically derived absolute value, cores provide both scalar (Kair) and vector (Kv, Kh) permeabilities, while the reservoir volume investigated by a well test is quite large as compared to logs or cores. Hence, porosity, saturation and permeability are often compared between its sources, and calibrated as needed. While logs based interpretation is a fast interpretation technique, getting core results can take signi-ficant time. Also, not all wells are cored, or sometimes core samples are too small to carry out the analysis, leading to a missing link between core-log integration. To improve some of the inaccuracies and limitations, ‘digital image analysis’ on core or on drill cuttings ‘thin sections’ can be a useful technique in estimating reservoir properties. Digital image analysis can provide porosity, pore size distribution, flow path tortuosity (permeability), irreducible water saturation and mineralogy of

the samples. Pore space area and perimeter is also determined and can be used in studying chemical reactions in the pore wall area for Improved Oil Recovery (IOR). This paper aims to develop novel automated digital image ana-lysis methods for Petrophysical Evaluation, and thus overco-mes some of the limitations of the traditional manual techni-ques. To analyze porosity of thin sections, a threshold value is required to separate pores response from the matrix. A set of rules have been developed to remove human subjectivity in selecting this threshold value. Red and green external boun-daries are marked on the analyzed image to provide a human assisted quality control for the results of captured pores (Figu-re-1). Correlations have been developed for permeability and tortuosity evaluation to understand reservoir flow potential. Petrographic thin section samples of reservoir rocks from two wells in the Barents Sea are studied. The thin section images are digitalized and analyzed using MatLab functions. Petrophy-sical properties, namely porosity, permeability and irreducible water saturation are quantified. In addition, some features of the pore space are quantified, including area, perimeter, major & minor axis of the pore area and orientation of the pores. The results from digital image analysis are compared against re-sults from conventional core analysis and well logging methods to establish the validity and limitations of thin section image interpretation technique.

Figure 1: Digital image processing for petrophysical properties and pore size distribution.

15

Modelling of CO2 Injection in Shale Gas Reservoirs: Improved Recovery and CCS

Dhruvit S. Berawala1, Pål Ø. Andersen2,3, Jann Rune Ursin11 Department of Energy and Petroleum Technology, University of Stavanger, Norway

2 Department of Energy Resources, University of Stavanger, Norway3 The National IOR Centre of Norway

With the current technology, only 3-10 % of gas from tight shale is recovered economically through natural depletion, demonstrating that there is a significant potential for enhan-ced recovery. Experimental results have shown that injecting CO2 into shales allows simultaneous desorption of methane and adsorption of CO2 through a substitution process, resulting in significantly increased recovery up to 20-50 %. This mam-moth quantity of gas can be used to replace coal worldwide for electricity generation reducing carbon emissions. This recovery strategy also provides a smart solution for subsurface perma-nent storage of carbon dioxide and can, conceptually, lead to carbon neutral energy. This paper presents a simple mathe-matical model capturing CO2 injection in shale as an enhanced recovery method and other flow mechanisms using a single well-induced fracture-matrix geometry.Experimental studies have demonstrated that shale kerogen/organic matter has higher affinity for CO2 than methane, CH4. CO2 is preferentially adsorbed over CH4 with a ratio up to 5:1. Moreover, at subsurface conditions, CO2 exhibits liquid-like density and gas-like viscosity, which are favorable properties for improving volumetric and pore scale sweep efficiency. This phase and flow behavior is built into the model equations which are solved using the numerical reservoir simulator CMG.

The production scenario we consider is a 2D geometry where a single well-induced fracture is surrounded symmetrically by shale matrix on both sides.During primary depletion gas is produced from the shale to the fracture by pressure driven flow and gas desorption from the shale. This process tends to give low recovery results. Stopping production and then injecting CO2 through the fracture and into the shale leads to a greater release of methane due to the substitution process. Restarting production from the well then allows the released gas to be produced. We present a systematic investigation of the impact of fracture-matrix di-mensions and characteristics (porosity and permeability), total organic content, operating parameters such as CO2 injection volume and injection time and substitution kinetic parameters. Through a series of reservoir simulations, the most amenable characteristics of the system are identified with emphasis on the critical parameters that control the success of CO2 injection as an enhanced shale gas recovery process.This work thus provides a model for intuitive interpretation of complex shale gas production and CO2 injection systems. It enables the industry to identify ideal candidates for enhanced gas recovery by CO2 injection and screen available gas reser-voirs for CO2 storage.

Figure 1 Schematic of flow dynamics of CO2 and CH4 (left), system geometry (right).

16

Micro- and nanoanalyses of fracture-filling after flooding on-shore chalk with different IOR fluids

Tine Bredal1, 2, Mona Wetrhus Minde1, 2, Udo Zimmermann1, 2, Merete Vadla Madland1, 2, Reidar Inge Korsnes1, 2, Caroline Ruud21)The National IOR Centre of Norway, 2)The University of Stavanger, Department of Energy Resources, Stavanger, Norway

Water injection into the Ekofisk-reservoir was introduced in 1987 to enhance oil recovery and create pressure support to prevent compaction. Previous studies have detected disso-lution and precipitation by exposing chalk to various brines at reservoir temperatures, which further increased deformation of chalk matrix. This deformation affected mechanical pro-perties of the rock and is referred to as water weakening of chalk. Attention has previously been given to intact chalk cores flooded with various brines. This particular thesis will focus on minerals precipitated in fractured chalk. The studied chalk originates from the Mons basin in Belgium. Its composition and characteristics are similar to chalk in Ekofisk. Three hollow cylinder tests on chalk cores were flooded by magnesium chloride (MgCl2), sodium chloride (NaCl) and synthetic seawater (SSW) in 2013 by Abubeker and Geitle as part of their unpublished master theses. The authors compa-red these three flow-through experiments with two intact cores flooded by MgCl2 and SSW. They found that there is no significant difference in yield strength, yet the average yield strength is slightly higher for intact cores. They also suggested that minerals were mainly formed in the fracture region. Effluent water during the creep phase was collected in the flooding process, hence this study will involve an investigation

of the minerals which were exchanged during the flooding of the three various brines. The main purpose will be to detect and recognize precipitated minerals and to search for a possi-ble pattern in distribution of those minerals. Methods which will be used to achieve desired goals are:1. Secondary electron microscopy (SEM) combined with

energy dispersive spectrometry and use of back-scattered electron microscopy at the University of Stavanger

2. Mineral liberation analysis (MLA) at TU Bergakademie Freiberg (Germany)

3. Transmission electron microscopy (TEM) on Focused ion beam (FIB-SEM) samples at the University of Stavanger

The first method will try to describe and to image the fractu-red areas. MLA studies (second method) will map those as an entire entity, while finally TEM analysis (third method) will image and identify the new grown phases down to nano-scale. The distribution of new grown minerals might occupy mainly fractures, but possibly clog and cement pore throats as well. When this is proven, it is of importance to clarify if will this influence further fluid flow within the reservoir? Will a fractu-red reservoir increase oil recovery more efficient than a nonfractured one, when exposed to fluid injection). And if so, which brine could give a more expedient result?

Screening of Nanofluids for Enhanced Oil Recovery by Micro-fluidics and Core Flood: An Experimental Investigation of

Oil Recovery and Nano-Eor MechanismsAlberto Bila, Erik Carlsen Kjørslevik, and Ole Torsæter

Norwegian University of Science and Technology (NTNU)

The increase in world energy consumption has increased the need of developing new techniques to recover as much oil as possible from the reservoir. Nanotechnology named from the scale it operates (1 – 100 nm) has found acceptance among researchers in many fields of science. In petroleum industry, the mechanism of oil recovery by nanotechnology is not well understood.Aurand (2017) found that the commercially available unmo-dified-silica nanoparticles produced by Evonik Industries are unstable in synthetic seawater (SSW) with total dissolved solids content of 35 000 ppm, and stable when dispersed in 3 wt.% NaCl (Li & Torsaeter, 2015). This finding prompted the company to develop surface-modified nanoparticles. However, the surface modifications targeting stability did not result in expected additional oil production (Aurand, 2017).With these key learnings, the company improved the surfa-ce-modification of the particles, which are stable in SSW and have EOR potential in a specific rock and crude oil.In this work, we investigated the potential of twenty-five improved silica nanoparticles to produce oil and possible EOR mechanisms. A North Sea crude oil and nanofluids prepared with concentration of 0.1 wt.% in SSW were used in this work. The screening procedure consisted of 1) saturating the glass micromodel with crude oil until irreducible water saturation (Swir) followed by nanofluid injection until no oil production. The best performing nanofluids with respect to oil recovery

were tested in water-wet Berea sandstone cores. 2) Core flood tests were carried out by injecting nanofluid (secondary reco-very) at 0.2 ml/min until no oil production. 3) The nano-floo-ded cores were subsequently aged in the nanofluid at 40°C for 10 days to determine how they react with an oil reservoir and whether it is beneficial for oil recovery. 4) Then, tertiary water injection was applied to the aged cores.Interfacial tension (IFT), dynamic contact angle on glass plates and wettability index measurements, and pressure drop were used to investigate nanofluids’ EOR mechanisms.The results of pore and core scale displacements indicate that some nanofluids are promising for EOR purposes. In micromo-del experiment, we observed a deformation and breaking up of the larger oil droplets into smaller ones, which easily were produced. Significant oil recovery was observed even after nanofluid breakthrough. The core-flood recovery factors varied from 49 to 72 % OOIP. Tertiary water injection added an avera-ge oil recovery of about 3% OOIP.Contact angle and Amott wettability index measurements are still ongoing; however, addition of the modified silica nano-particles to SSW resulted in minor IFT reduction. The pressure gradient during the core flood indicated no particle agglome-ration or pore blockage; hence fluid flow diversion may not be the primary mechanism. The micromodel results are only applicable for screening purposes, since the recoveries are much larger than in actual rock samples.

17

Interpretation of steady state relative permeability measurements on composite cores with end effects

Aditya Dixit1,2, Pål Østebø Andersen1,3, Dag Chun Standnes1, Alexander Shapiro2, Svein Magne Skjæveland1,3, Kenny Walrond11Department of Energy Resources, University of Stavanger

2Department of Chemical and Biochemical Engineering, Technical University of Denmark3The National IOR Centre of Norway, Stavanger, Norway

This paper will interpret steady state relative permeability measurements from several composite cores. Each composite core consisted of three-four smaller cores (see Figure). The first core will have highest permeability and be placed at the inlet and the core with lowest permeability will be placed at the outlet. Steady state is difficult to achieve as it takes long time to establish and the costs associated with the tests incre-ase. Steady state signifies that there is a stable pressure drop across the core and the flow rates of injected and produced fluids are the same. Typically, history matching compensates for not reaching steady state. This paper will in particular investigate the role of capillary pressure effects during steady state relative permeability measurement on composite cores. These effects complicate the interpretation of relative permeability curves, but must

be accounted for. To mitigate these complications, CT scan measurements were made on the studied cores before every change of the injected fractions of oil and water. This allowed evaluation of the in situ saturation profiles.For reference, comparisons will also be made with homogeno-us cores. It is a standard assumption in core scale simulation software like Sendra that the capillary end effects occur only at the outlet. This is valid for a homogenous core. On the other hand, for heterogeneous or composite cores the capillary end effects occur between the boundaries of the small cores. We will develop a wider framework for history matching that allows accounting for capillary pressure end effects throughout the core and determine if significant or improved interpretati-on can be obtained compared to the standard software used in industrial application, e.g. Sendra.

Injection of Oil and Water at different ratios in a hetero-geneous core.

18

Anomalous Zeta Potential Trends in Natural Clay-Rich Sand-stone & Implications for Low-Salinity Waterflooding

Harry Collini, MEngPostgraduate Researcher

Department of Earth Science & EngineeringImperial College London

It is widely accepted that the zeta potential in natural sand-stones is negative and increases in magnitude with decreasing ionic strength. However, the impact of clay minerals remains poorly understood. The zeta potential is a measure of the electrostatic forces present at an interface. Changes in the zeta potential at the mineral-brine and oil-brine interfaces have been suggested to cause an increase in oil recovery during low salinity waterflooding. Here we report measurements of the zeta potential on intact, clay-rich Berea sandstone cores saturated with NaCl and CaCl2 electrolytes of varying ionic strength, obtained using the streaming potential method.Initial dilution of a NaCl electrolyte caused the zeta potential to increase in magnitude (become more negative) until ~65mM, as has been previously reported. However, with further ele-

ctrolyte dilution, the zeta potential decreased in magnitude and became less negative. Dilution of CaCl2 electrolyte yielded no change in the zeta potential. Trace Ca2+ ions were detected in the effluent when flooding with NaCl brines. At high ionic strength this is insignificant however, at low ionic strength, it was shown that these calcium ions dominate surface adsorp-tion, thus creating a more positive surface potential. It is suggested that the presence of Ca2+ ions is due to the clay-rich mineralogy of Berea sandstone. These results show that the change in zeta potential may be suppressed by the presence of Ca2+ sourced from the sample and that dilution may not always yield a more negative zeta potential as has previously been assumed.

Wettability Estimation via Oil AdhesionSamuel Erzuah, National IOR Center of Norway, University of Stavanger; Ingebret Fjelde, The National IOR Center of Norway, IRIS and University of Stavanger;

Aruoture Voke Omekeh, IRIS National IOR Center of Norway

Wettability is an inevitable parameter in multiphase flow in oil reservoirs due to its pronounced effect on oil recovery. It is defined as the ability of a fluid to adhere to a solid surface in the presence of other immiscible fluids. Most often than not, wettability estimation is time consuming. Hence, the need for a fast, cheap and accurate wettability characterization tool. This study aims at estimating the wettability by determining the amount of oil adhere onto reservoir mineral surfaces using the flotation test, Quartz Crystal Microbalance with Dissipati-on (QCM-D), and the Surface Complexation Modelling (SCM) techniques. To better understand the wetting preference of the reservoir rock, oil adhesion tendencies for minerals were mimicked using both flotation and QCM-D techniques. The flotation test relies on the affinity of the reservoir minerals to a fluid. The QCM-D technique also depends on the changes in the frequ-ency of the resonating crystal to determine the amount of oil adhesion. The mineral-brine and the oil-brine interactions that led to the retention of oil on the minerals surfaces during the flotation and QCM-D tests were also assessed via the SCM. Two dominate minerals in sandstone reservoir rocks namely quartz and kaolinite were considered for this study. In additi-on, formation water (FW) and stock tank oil (STO) were used to represent the reservoir fluid phases. The materials used

in both the flotation tests and the QCM-D experiments were used as input into the SCM. In addition, the STO was converted to their respective acidic and basic component to be used in the SCM.Both the flotation test and the QCM-D results show that kaolinite is less hydrophilic than quartz. This was depicted in flotation tests by the higher concentration of oil-wet particles for kaolinite than quartz. The QCM-D results also portrayed higher attenuation of the frequency signal during the STO inje-ction with the kaolinite sensors than with the quartz sensors. This reduction in the frequency was as a resulted of adsor-bed oil (Figure 1). The SCM results also confirm that quartz is strongly hydrophilic compared to kaolinite. This was depicted by the relatively high total bond product (TBP) which expresses the tendency of oil to be adsorbed onto a surface. In addition, the SCM results revealed that the most dominant mechanism for oil adhesion onto both quartz and kaolinite surfaces is by divalent ion bridging such as Ca2+ and Mg2+.When both flotation test and QCM-D techniques are coupled to the SCM, it provides a promising technique for characteri-zing the wettability by capitalizing on the oil adhesion tenden-cies of the minerals. It can be concluded that quartz is strongly water-wet compared to kaolinite.

Figure 1: Result of the QCM-D experiment depicting oil adhesion during STO injection

19

Effect of small scale mineral heterogeneity on fracture geometry evolution during CO2 injection

Hossein Fazeli1, Helge Hellevang11 Department of Geosciences, University of Oslo, Pb. 1047, Blindern, Oslo, Norway

To sequester CO2 more safely, it is needed to have some predictive modeling tools that can show the behavior of CO2 when it is flowing through fractured caprocks. Hydrodynamical properties of fractures will change due to chemical disequilibri-um resulting from interactions between injected fluid and rock minerals. Experiments have shown that small scale hetero-geneities (such as mineral spatial heterogeneity) may have significant impacts on fracture geometry evolutions. Different continuum scale numerical models have been developed to predict these evolutions, these small scale heterogeneities are better described using pore scale models that can capture the processes at the scale they are happening i.e. pore scale. Very recently, Lattice Boltzmann Method (LBM) has attracted much attention and been used for simulation of reactive transport processes involving the geometry change of porous media due to chemical reactions.In this study, a new LBM based model have been developed to model the evolution of a fracture geometry subject to the injection of CO2-rich brine. Two different initial mineral spatial distributions (banded and mixed structures) are considered for the rock matrix around the fracture meaning that the rock

matrix is composed of two different minerals with diffe-rent reactivities (such as Carbonate and Clay minerals). 2D and 3D simulations are run to investigate how poro-

sity and permeability of the fractured media will change during dissolution of the rock matrix.Simulation results show that where the initial mineral distribu-tion is mixed, due to the difference in dissolution rates of diffe-rent minerals, degraded zones will form after dissolution. Also, when the initial mineral distributions are banded, comb-tooth structure will be created. Permeability values calculated for both cases will show that permeability reaches a steady state condition after a specific time. This behavior is attributed to the tortuosity increase (after dissolution) in case of mixed stru-ctures. When tortuosity increases, the flow regime in the more tortuous regions is more diffusive which causes the reactants not to flow fast in the degraded zone and therefore most of dissolution will happen in the main flow channel. In case of banded structures, the vertical flow barrier will create more resistance to flow hence similar to the previous case, most of dissolution takes place inside the main fracture and dissolution in the areas further away from the main channel does not play an important role in the permeability increase.Results show that when different mineral spatial distribution exist in the rock matrix with mineral having different dissoluti-on rates, the porosity-permeability relation does not follow a power law (or cubic law) behavior which is usually used in the continuum scale models.

Interpretation of reactive flow and its impact on compaction in chalkKatherine Esquivel1, Pål Østebø Andersen1,2, Dhruvit Berawala3

1Department of Energy Resources, University of Stavanger2The National IOR Centre of Norway

3Department of Energy and Petroleum Technology, University of Stavanger

Chalk oil reservoirs in the North Sea (e.g. Ekofisk and Valhall) have been successfully flooded with seawater for decades. Flooding chalk core plugs with seawater or modified seawa-ter at high temperatures has displayed increased chemical interaction with corresponding water weakening effects as seen by higher compaction rates. These mechanisms and brine dependent wettability alteration explain the high oil reco-very seen in the field and lab experiments. The link between compaction and chemical reactions in the brine-dependent recovery should therefore be interpreted for further beneficial application.A mathematical model will be considered to describe single phase core flooding experiments from the literature perfor-med with outcrop chalk cores and seawater-like brines at reservoir conditions. The model initially considers mechanisms such as advection, dispersion, mineral precipitation/disso-lution, and surface chemistry based on the most prominent ions and minerals involved in these setups. The model is then extended to understand the coupling with enhanced compacti-on. This study will mainly focus on the dissolution/precipitati-on of minerals occurring when injecting reactive seawater-like brines and compaction. Being able to capture the chemical response due to effluent measurements and mineralogical profiles and compaction response can allow further under-standing of the combined effect of the chemical reactions and

deformation of chalk cores. Numerical and analytical solutions are proposed.Reviewed flooding experiments present chemical effects in reactive flow induced by brine injection – such as precipitati-on of Mg2+-bearing minerals and dissolution of Ca2+-bearing minerals. This in turn enriches the Ca2+-ion concentration in the effluent – correlates with the mechanical strength of chalk under stress at high temperatures – thus enhanced dissoluti-on in rock-fluid chemistry affects permeability reduction and compaction rate. Measured ion concentrations in effluent fo-und from chemical analysis, mechanical tests results and creep rate from such experiments will be analyzed and interpreted using the proposed model.

20

Polymer-injection for IOR purposes at the Norwegian Continental shelf – micro- and nanoanalytical approach for the un-

derstanding of phase-formation and its implication for upscalingSiri Gloppen Gjersdal1, 2, Mona Wetrhus Minde1, 2, Udo Zimmermann1, 2, Merete Vadla Madland1, 2, Aksel Hiorth1, 2, Nils

Harald Giske1, 3, Arne Stavland1, 31)The National IOR Centre of Norway, 2)The University of Stavanger, Department of Energy Resources, Stavanger, Norway,

3)International Research Institute of Stavanger, IRIS AS

Improved Oil Recovery (IOR) is a “hot topic” in the Oil and Gas industry. Sodium silicate (SS) is a polymer which can be injected into a reservoir as a fluid and turns under high tem-peratures into a semi rigid gel. However, the knowledge about the polymer on a micron- and nano-scale is restricted. The method was successfully tested in the Snorre field in the North Sea to increase the sweep efficiency by blocking the high permeable zones, forcing injected water and hydrocarbons through the surrounding layers. The objective for this project is to investigate and quantify the polymer distribution and its properties on a micron- and nano-scale, including its chemistry and the mineralogical and chemical alterations in the host rock after injection. In this project the polymer composition will be mapped, and particularly the contact with the rock will be examined in detail. Increasing the oil recovery factor using environmentally friend-ly and sustainable methods represents a great value. Sodium silicate (SS) is categorized as a green chemical (PLONOR). Expe-rimental work shows that SS can be used for in-depth water di-version to increase sweep efficiency. When the solution forms a blocking gel in the high-permeable zones in the reservoir, the injected water is forced through the low- or non-swept zones of the reservoir, increasing oil production. To be able to improve the field-scale operations, one needs to understand

the pore-scale changes in the host-rock. The pore-scale results cannot automatically be upscaled to field-scale, but to under-stand chemical alterations in the host rock is paramount. The polymer will be observed at a nano-scale using scanning ele-ctron microscope (SEM), and the energy-dispersive x-ray spe-ctroscopy will give first data on the chemical composition on a micron-scale. Transmission electron microscope (TEM) analysis will be able to determine the chemistry on a nano-scale using focused ion beam-SEM sections of selected areas. It will also define relations to the host-rock on micron- and nano-scale. Geochemistry and x-ray diffraction might give an indication to whether the chemistry has been altered, comparing the flooded and the unflooded sample. To map the distribution of the gel within the core, mineral liberation analyser (MLA) will be used. MLA results in images of thin sections as a map of the different minerals. Petrography will display how the gel appears in the pore spaces visible at low magnifications in thin sections. Raman spectroscopy will offer information about the type of substance after gelation.For this study, Berea sandstone cores were flooded with floo-ding parameters (pressure, temperature, saline pre-flush, etc) chosen in accordance to models of gelation time for sodium silicate.

Microwave-Assisted Hyperthermia Using Magnetic Nano-particles for In-Situ Upgrading and Recovery of Heavy Oil

Kun Guo1,2* Zhixin Yu1,21 The National IOR Centre of Norway, University of Stavanger, 4036 Stavanger, Norway

2 Department of Energy and Petroleum Engineering, University of Stavanger, 4036 Stavanger, Norway*E-mail: kun.guo@uis.no

The recovery and upgrading of heavy crude oil (HCO), which accounts for ca. 70% of the total world oil reserves, have been a challenging task due mainly to its high viscosity or poor mobility. Based on the heat-induced viscosity reduction, thermal injection methods are being prevailingly used for the in-situ upgrading and recovery of HCO. However, key issues in the steam generation, heat loss and heat transfer remain to be addressed. Lately, microwave irradiation has been proposed as an alternative heat source owing to the fast and efficient heating at molecular level. When hyperthermic nanoparticles (NPs) are introduced in the oil phase, they can adsorb the microwave energy and induce surface localized heat gradients. As such, these NPs function as independent heat emitters, allowing a uniform and sufficient heating of the surrounding oil phase.This study focuses on the design, synthesis, and characterizati-on of hyperthermic NPs that are defined by their capability to emit heat under external electromagnetic irradiation. When purposely functionalized, including for magnetic susceptibi-lity, NP stability and recyclability, the proposed NPs endow an

improved heating efficiency, thus potentially advancing the heavy crude oil cracking and upgrading. Towards these aims, we will systematically study the effect of critical structural parameters of NPs, including size, shape, composition, and support, on the NP aggregation, heating efficiency, upgrading reactions, and magnetic-based recovery/recyclability. Based on our readily available knowledge of the controllable synthesis of monodispersed NPs, iron oxide (Fe3O4) and manganese oxide (MnO2) NPs will be tailor made and tested in microwave-as-sisted oil upgrading in the laboratory. The heating temperature will be monitored and advanced product analysis will enable the understanding of key upgrading mechanism. The main objective is to evaluate the applicability of microwave-assisted hyperthermia using magnetic NPs for the in-situ upgrading and recovery of HCOs, particularly the fundamental structure-(re)activity relationships. Despite clear advantages, optimized use of magnetic, hyperthermic NPs remains scarce and no specific structure-activity relationships have been reported for advanced HCO processing.

21

Optimization of Polymer Flooding for an Extra Heavy Oil Reservoir by using Low Salinity Water

Hernández Edgar, Valero Emil

Polymer flooding has been foreseen as an EOR option for Heavy and Extra Heavy Oil Reservoirs in the cases when an acceptable technical/economical water-oil mobility ratio can be achieved. Under this premise, the design of a Polymer Flooding Pilot Project to be carried out in the Orinoco Belt Ve-nezuela was subjected to optimization by using lower salinity water for polymer solution mixing in order to reduce polymer consumption, but an additional effect on water-oil displace-ment efficiency was also measured at lab tests. The present work describes: the composition of the former and current wa-ter sources for the polymer solution, the reservoir conditions tested at the lab as well as rheology and displacement tests results, and their impact on the Pilot Test production forecast. The first design of this Pilot Test considered a mixing-injection water salinity of 1000 ppm, whereas the optimization study used an ultralow water salinity of 160 ppm. The polymer used is a partially-hydrolyzed-polyacrylamide added to the water to reach a target viscosity of 80 cP and a water/oil mobility ratio lower than 10. As expected, rheology tests showed that a lower amount of polymer (50% less) is required when using lower salinity water. Displacement tests used live extra heavy oil of API 8.7 with high content of basic and acidic compounds (6.27mgKOH/g and 2.3 mgKOH/g respectively) and a 2000 cP

viscosity, while the reservoir core is an unconsolidated sand with 8% of clay (90% Kaolinite) and a high permeability of 8-15 Darcies, formation water salinity is around 40000 ppm. Displacement tests were carried out under two scenarios: wa-terflooding followed by polymer flooding (tertiary mode) and only polymer flooding (secondary mode). The main findings in-dicate that water fingering also occurs at core level during wa-terflood, leading to an apparent Sor of 60%, reduced by both secondary and tertiary polymer flood to 25% with the higher salinity injection water, and to 16% when using ultra low water salinity. This final improvement could be attributed to further wetting conditions alterations, taking place during ultra-low salinity polymer flooding. Previous studies suggest that oils containing high amounts of basic and acidic compounds, which is the case for most heavy and extra-heavy oils, have a natu-ral tendency to wet reservoir rocks when clays are present. These compounds could be detached from the rock during low salinity water floods, explaining the low Sor measured by labo-ratory tests. This synergy of lower polymer consumption and better displacement efficiency was incorporated by reservoir simulation studies, resulting in more attractive oil recovery and economic indicators for the Heavy Oil Polymer flooding Pilot.

Effects of high salinity and high temperature on nano-surfactant formulations properties

Ivanova A.A., Miturev N.A., Cheremisin A.N., Spasennykh M.Y.Skolkovo Institute of Science and Technology, Moscow

Due to the growth of world energy demand, it is become especially important to continue improving oil recovery (IOR). The one of IOR technology is chemical oil recovery method, which based on surfactants, polymers and alkaline (ASP) injection into wellbore either together or separately. The main idea is that injected fluids react with reservoir rocks, alter their properties and help to gain additional oil from reservoir. Surfactant solutions are usually added to decrease interfacial tension (IFT) between two immiscible fluids– oil and brine. As a result, trapped in pore throats oil beco-mes free and mobile. This aid to recover about 80% of origin oil in place. However, IFT should be reduced up to 10-1 – 10-2 mN/m to improve oil recovery. This is called ultra-low value and not every surfactants can cause such significant reduction, but those who can are difficult to produce and expensive. Th-erefore, should be found the modifications of already existed surfactant solutions.In the present work, we investigated the nanoparticles (silicon dioxide) influence on cationic surfactant solutions properties under different temperature and salinity conditions with the purpose to use these mixtures in chemical flooding. First, in this work, the interfacial tension measurements were perfor-med with surfactant solutions containing salt (180 g/ml) and nanoparticles (0-0.5 wt.%). It was shown, that small amount of nanoparticles (0.1 wt.%) reduces IFT between brine sur-

factant solutions and oil to ultra-low value, but further nanoparticles concentration increasing leads to IFT increment (see Fig.1). It can be due to the fact that

nanoparticles preferably associate with each other when their concentration is high. Moreover, in this work we experimental-ly found the minimum effective surfactant concentration (0.05 wt.%) to enhance oil recovery.Interfacial tension properties dependence on temperature was also observed in this work. It was found that addition of nan-oparticles has slight effect on IFT reduction in solutions with small surfactant concentration at 70 °С. This result is opposi-te to the one obtained under ambient conditions. However, when surfactant concentration is high IFT maintains low at 70 °С. It can be explained by taking into account that surfactant molecules start to break down under high temperature and remained nanoparticles in solutions cannot reduce IFT.To summarize all above mentioned it can be concluded, that the addition of nanoparticles to surfactant solutions aids in IFT reduction between brine surfactant solutions and oils. Howe-ver, under high temperature, more surfactant concentration is needed to maintain the low interfacial tension.Figure 1. Nanoparticle concentrations influence on IFT between brine surfactant solutions and oil.

22

Permeability and Stress StateKallesten, E.I.1,2, Madland, M.V.1,2, Korsnes, I.R. 1,2, Zimmermann, U. 1,2

1). University of Stavanger, 2). The National IOR Centre of Norway

Changes in the pressure state of a petroleum reservoir, closely linked to the different production stages, have a direct effect on the porosity and permeability of the reservoir and consequ-ently on the oil recovery. Besides compaction, rock-fluid inte-ractions further enhance mechanical changes of the reservoir, as the injected fluids lead to chemical reactions that can alter the rock’s mineralogy, structure and strength. It is therefore essential to understand how the reservoir stress state affects rock permeability, what the chemical-mineralogical alterations during core flooding are and the extent of mechanical de-formation of the rock during testing.Mechanical tests performed in a triaxial cell simulate different reservoir conditions, allowing systematic variations of the parameters assumed to influence rock permeability, like pres-sure, temperature, injecting fluid. This study presents results from experiments highlighting the effects of confining pressure on permeability of water-saturated cores from outcrop chalk (Kansas) with porosities varying between 32% and 39%.Deviatoric loading above yield at low confining pressure (1.2MPa), 50˚C and equilibrium sodium chloride brine as injecting fluid, resulted in shear failure with a steeply dipping fracture of the core and a concomitant increase in permea-bility. Further, during creep and unloading, the permeability

changes were negligible, such that end permeability remained higher than the initial values.Under the same thermal and chemical conditions, but higher confining pressure (3.0MPa), the core experienced reduced axial and radial strain rate during loading, compared to the previous experiment. The result was a more ductile deforma-tion of the core, with a low-angle dipping shear fracture and permeability decline throughout the test.The core’s initial porosity may have influenced the strength of the core, but had minor effects on permeability behaviour (Figure 1).Further investigations regarding the interplay of various parameters on rock permeability are ongoing. The results, combined with experimental data from actual reservoir rocks, will not only enhance the understanding of the impact of typical water-related IOR techniques, but it will also improve the accuracy of reservoir predictions, and contribute to finding smarter solutions for future IOR.

Synthetic CaCO3 surfaces in aqueous solutions AFM and Surface Force Apparatus (SFA) measurements

S. Javadi1 2 3, J. Dziadkowiec1, A. Ryne1, J. E. Bratvold4 and A. Hiorth2 3 51Department of Physics, Oslo University, Oslo, Norway.

2The National IOR Center of Norway, Stavanger N-4036, Norway.3Petroleum engineering department, University of Stavanger, Norway.

4Department of Chemistry, Oslo University, Oslo, Norway.5IRIS, P.O. Box 8046, Stavanger N-4068, Norway

Fluid injection in chalk reservoirs is a high risk associated process due to chalk compaction in the water-saturated re-gions. To evaluate such risk, it is necessary to understand the mechanisms that underlie the chalk compaction. It has been shown that changes in the pore fluid chemistry can influencethe mechanical behavior of chalk [1]; such response has been attributed to the nanoscale forces acting between mineral grains [1, 2]. Interfacial forces, attractive and repulsive, betwe-en calcite surfaces separated by a nanometer-thick film of aqu-eous solutions, are influenced by the change in properties ofthe conned fluid that aect the dissolution, mass transport and reactivity of the surfaces. In this work, those interactions are measured using the Atomic Force Microscope (AFM) and the Surface Force Apparatus (SFA). The SFA force measuring technique provides an opportunity of in situ observations of the possible changes in the surfaces via light interferometry technique while measuring the interfacial interactions. The contact area in the SFA force measurements (50 m < contact diameter < 150 m) is larger than in the AFM, at the scale that

surface roughness is an important parameter in interactionforces between two surfaces. In this scale, interactions between contacting asperities dene the nature of interfacial forces, repulsive or attractive. This study shows how the crystal growth, dissolution and changes in surface roughness affect these interaction forces. We use the synthetic CaCO3 lms (pre-pared by Atomic Layer Chemical Vapor Deposition (ALCVD) method [3]) for the SFA. Synthetic calcites are made by growing thin polycrystalline lms (100nm < thickness < 200nm) of CaCO3 on gold and mica substrates. We also show the re-sults of our investigation on the topographical evolution, dis-solution and crystallization processes of these lms by the AFM in air and CaCO3 pre-saturated solutions. We show that the experimental parameters, like temperature and the base sub-strate, characterize the deposited lms that are mostly bound by f104g planes. The overall purpose of this study is tounderstand the eect of fluid chemistry, surface recrystallization and asperity deformation on interaction forces between two rough surfaces.References[1] Risnes, R., M. Madland, M. Hole, and N. Kwabiah (2005). doi:10.1016/j.petrol.2005.04.004[2] A.Ryne, K.N.Dalby, & T.Hassenkam (2015). doi: 10.1002/2015GL064365.[3] Nilsen, O.; H., F.; Kjekshus, A., Thin Solid Films 2004, 450, 240-247.

23

Figure 1: Permeability response to confining pressure during deviatoric loading at 50˚C. The graphs show similar behaviour of permeability evolution in both lower (black) and higher (red) porosity cores, abruptly amplified as a response to brittle deformation (right) and decreasing throughout the ductile deformation of the core (left). Notice the scale difference between the two graphs, as the higher the confining pressure, the higher axial stress required to overcome the core’s shear strength and the more axial strain.

Evaluation of calcareous nannofossils in flooded and unflooded chalk used in NIOR experiments

Emilie Aasen Kavli1, 2, Mona Wetrhus Minde1, 2, Udo Zimmermann1, 2, Elisabetta Erba3, Merete Vadla Madland1, 2, Reidar Inge Korsnes1, 21)The National IOR Centre of Norway, 2)The University of Stavanger, Department of Energy Resources, Stavanger, Norway, 3)University of Milano, Dipartimento di

Scienze della Terra ”A. Desio”, Via Mangiagalli 34, 20133 Milano -Italy

Carbonate reservoirs represent a significant proportion of the hydrocarbon reserves in the world. In its World Energy Outlook 2006, the International Energy Agency estimated that more than 60 % of the world’s oil and 40 % of the world’s gas reser-ves are trapped in carbonates. In the North Sea, an important amount of the oil production comes from chalk reservoirs, which are continuous subjects of experimentation in enhanced oil recovery research to improve oil recovery (IOR). Chalk is a highly fossiliferous rock, with calcareous nannofossils being the main constituent. It is well known that the type, the amount and the form of paleontological components do affect signi-ficantly the flow of fluids in chalk. The three parameter are solely dependent on the facies mostly unknown or not tackled in chalk-related IOR-research. To study the facies meticulous profiling is paramount and extraordinary time consuming. Ano-ther approach is to determine and quantify the paleontological material to understand the paleoecological situation at time of deposition. These calcareous nannofossils are less than 30 μm in diameter and include three forms: coccolithophores, disco-asters and nannoconids. As chalk is almost entirely composed of calcareous nannofossils, it is of uttermost importance to characterize this paleontological material. Indeed, the quan-tity of different taxa, their shapes, orientations and degree

of diagenesis determine the porosity and density of chalk reservoirs and affect fluid flow. By combining light microscopy with scanning electron microscopy (SEM)

investigations, this study aims to characterize the nannofossils, by describing the nannofacies and reconstructing a 3D nan-nofossil composition of both onshore outcrop- and offshore reservoir-chalks. We furthermore like to gain: (1) Biostratigrap-hy using calcareous nannofossils in smear slides, (2) Calcareous nannofossil preservation in smear slides and SEM, (3) Descri-bing calcareous nannofossil assemblages and their implications in paleoecology and paleoceanography, (4) Simple relative abundance of calcareous nannofossil genera and species in thin sections giving quantitative data, (5) Calcareous nanno-fossil biometrics, (6) Calcareous nannofossil orientation using SEM, and (7) Magnesium overgrowth distribution on nann-ofossil surfaces and if certain genera or species are favoured for magnesium overgrowth and mineral replacement during flooding IOR-experiments. The paleontological nannofacies’ investigations might quite often get underestimated in chalk reservoir characterization. This is surprising, as the assemblage of calcareous nannofossils, their sizes, compositions, orienta-tion and preservation will heavily influence the properties of the reservoir. They are the basic building blocks of chalk, and even though they are small, the control they have on reservoir qualities might be severe. This thesis aims to contribute to IOR chalk-research by going back to its most basic constituent: the nannoplankton that deposited on the ancient ocean floors and formed the chalk.

24

Mineral alterations in water-wet and mixed-wet chalk due to flooding of seawater-like brines

Mona Wetrhus Minde1, 2, Jaspreet Singh Sachdeva1, 2, Udo Zimmermann1, 2, Merete Vadla Madland1, 2, Reidar Inge Korsnes1, 2, Anders Nermoen1, 2, 3

1)The National IOR Centre of Norway, 2)The University of Stavanger, 3)International Research Institute of Stavanger (IRIS)

After decades of hydrocarbon-production and injection of seawater into chalk reservoirs, combined with systematic chalk mechanical flow-through experiments, it is well understood that flooding of non-equilibrium brines affects the geo-me-chanical properties of chalk. Seawater, or modified seawater, is an important injection fluid for improved and enhanced oil recovery. To further optimize the injection fluid, close interacti-on between experimental studies and modelling is needed.To simplify the system, many experiments are performed wit-hout the presence of oil, i.e. in water-wet systems. This ena-bles a greater variation in methods of investigation of minera-logical alterations and eases analyses of the effluent water and sample rock-material. Other experiments are carried out on chalk aged in oil to alter wettability towards a mix-wet system.We have performed field emission gun scanning electron microscopy on cores from outcrop chalk collected from the Upper Cretaceous Niobrara Formation in Kansas and compared how high-temperature flooding with MgCl2 alters the mine-ralogy of the rock in both water-wet and mixed-wet sam-

ples. Scanning electron microscopy, together with elemental analyses by energy dispersive X-ray spectroscopy, provide an efficient method to image and detect precipitation of secon-dary minerals, alterations of primary mineralogy, and where these alterations preferentially take place. Our analyses show that the mineralogical alterations of wa-ter-wet and mixed-wet cores are similar, matching the similari-ties found in changes of geo-mechanical stiffness and strength parameters as well as density and specific surface area, as magnesium bearing mineral phases precipitates on the expen-se of calcite. The main observations are dissolution of calcite and precipitation of magnesite in both systems. Studies of water-wet cores have shown that outcrop chalk from Kansas undergoes a more intense alteration with large morphological differences (grain shapes and size) at the inlet of the flooded cores, compared to other outcrop chalks. This is also the case for mixed-wet Kansas cores; thus demonstrating the useful-ness of studying such simplified water-wet systems.

25

Smart Water Production by Membranes for Enhanced Oil Recovery in Carbonate and Sandstone Reservoirs

Remya Ravindran Nair, Evgenia Protasova, Skule Strand, Torleiv BilstadUniversity of Stavanger, 4036, Norway

Membranes are commercial market-competitive technologies for desalination and water treatment onshore and offshore. Membranes used offshore are not only for desalination but also for scale prevention by sulfate removal from seawater prior to injecting into petroleum reservoirs. In this research, nanofiltration (NF) and reverse osmosis (RO) are the two main membrane types involved in Smart Water production. Both seawater and produced water were used as feed for NF membranes. The performance of NF membranes at different feed pH values and ionic compositions at various pressures were evaluated. Adjustment of ionic composition in feed water is successfully accomplished by membranes with defined pore sizes. Proper choice of membrane material and membrane surface area similarly provide optimal water reco-very and ion separation.

Offshore platforms often include RO membranes for desali-nation for potable water production. With RO infrastructure already in place, installation of NF membranes for Smart Water production is cost effective. Normal operating pressure for RO in seawater is 60 bar. For NF the operating pressure is 3-30 bar without the use of chemicals. Membranes producing Smart Water minimizes freshwater consumption, reduce water-hand-ling costs and power consumption.

Keywords: Nanofiltration membranes, Reverse Osmosis, En-hanced Oil Recovery, Smart Water

Core scale modelling of EOR transport mechanisms

Polymer flooding has long been considered a promising EOR method. Nevertheless, despite half a century of studies, applications to the field scale have been relatively rare, and mainly limited to onshore reservoirs. One reason for this is the complicated nature of polymeric liquids, which makes them very challenging to model. To reduce uncertainty in forecasts of polymer flood performance, a central aim of this PhD work was to improve upon existing simulation models by incorpora-ting more realistic physics, and hence less free input parame-ters, into the models.The basic idea behind polymer flooding is that by adding large macromolecules to the injected brine, the mobility ratio between water and oil is lowered, leading to a more efficient sweep of the reservoir. This is complicated by the fact that polymer solutions are non-Newtonian fluids, with a non-linear relation between shear stress and shear rate. At realistic flow rates deep within a reservoir, polymer solutions tend to exhibit an approximately Newtonian or shear thinning fluid rheo-logy. However, the behaviour inside a porous rock may differ significantly from that seen in a bulk solution. This is especially the case for large molecular weight synthetic polymers, at high flow rate conditions. One worry has been the apparent shear thickening behaviour frequently observed in the laboratory. If significant pressure build-up occurs during polymer injection at

the field, well injectivity will be greatly reduced and, in the worst-case scenario, make polymer injection econo-mically infeasible. Another potential show stopper is the

issue of polymer mechanical degradation, that is, the break-down of chemical bonds along the polymer chain backbone because of excessive applied force.In the PhD work, we tested a newly developed simulation model for polymer flooding, implemented in the software IORCoreSim. In one case, the model was applied to core floods involving four different HPAM polymers, of differing initial molecular weight (5 to 20 MDa), flooded through sandstones with more than one order of magnitude variation in both permeability and flow rate. The simulator was able to repro-duce experimental, steady-state mobility reduction factors and sampled effluent viscosities, using mostly the same set of input parameters.Using history matched input parameters derived from these experiments, we subsequently applied the model to a synthe-tic case in radial geometry, to investigate well injectivity and mechanical degradation near an injection well. We derived several approximate scaling relationships for how the mo-delled degradation is affected by flow rate and permeability, which may prove useful when evaluating the risk of degrada-tion at the field. An excessive amount of grid cells was needed to compute accurate solutions to the governing equations. However, the full model is not needed for all grid blocks in a full field simulation and, in future work, our approach can hopefully be extended to create an accurate well model for polymer flooding.

Oddbjørn Nødland, PhD student (Dep. of Energy Resources, University of Stavanger; The National IOR Centre of Norway / IRIS)

26

Controlling parameters during WAG injection – A simulation studyJan Inge Nygård1, Kenny Walrond1, Pål Østebø Andersen1,2

1Department of Energy Resources, University of Stavanger, Norway2The National IOR Centre of Norway

Mobility control is one of the smart features behind efficient oil production, which can be achieved by applying the EOR process known as Water-Alternating-Gas (WAG). By the use of alternating injections of gas and water, better volumetric sweep is achieved due to the presence of water, which limits gravity segregation effects by controlling the gas mobility. Se-condly, better microscopic displacement efficiency is achieved due to the lower interfacial tension (IFT) that exists between the gas and oil phases, compared to that between water and oil. With the combined effects of being able to reach more of the unswept zones and mobilize oil previously trapped by capillary forces, the potential for upstream production of oil increases. Due to the degree of complexity arising from the choice of injection fluids, and how fast and how long they should be injected, the resulting three-phase interactions in the reservoir make it difficult to properly characterize the performance with WAG. We simplify and build an understanding of WAG in a simulation study by adding complexity / mechanisms one step at a time. The goal is to predict WAG performance in a hetero-geneously layered model with gravity segregation and unfa-vorable displacement using as few characteristic parameters as possible while incorporating the main mechanisms and input

parameters. This may ultimately lead to better optimization procedures for implementation of WAG in field practice. This study focuses on non-miscible injection of WAG. A black-oil model is used, which is solved using ECLIPSE 100. We base our approach on running a full simulation model and seeing how well our theory derived from a simplified model predicts the full model performance. First we consider a 1D setting. For this case Buckley-Leverett flow theory serves as the basis of our methodology and is extended to cover injection of water or gas in different WAG ratios. The performance of the model is illustrated in the figure, where we see that the recovery with WAG is accurately described from the theoretical points lying on the red line. As we move to the 2D reservoir model, we include gravity segregation and heterogeneity which contri-bute to a less efficient recovery process in general, but where WAG is more optimal than single phase injection alone. We show that the recovery of WAG can be predicted by the use of a few dimensionless numbers which incorporate important parameters such as WAG ratio, mobility ratio, gravity number, viscosities, Corey exponents and relative permeability end-points. The future goal is to extend our approach to 3D, and include effects such as capillary pressure and relative permea-bility hysteresis.

Uncertainty estimate of elastic full waveform inversionKaren Synnøve Ohm

The National IOR Centre of NorwayDepartment of Energy Resources, University of Stavanger

This PhD project will try to make an estimate of the uncertain-ty of the elastic full waveform inversion (FWI). The elastic FWI will be able to record small changes in the elastic parameters. Combined with an uncertainty estimation the elastic FWI isan important tool to monitor the oil and gas reservoirs more accurately for enhanced oil recovery.

It is becoming increasingly important with well-resolved and accurate estimates of the subsurface parameters from seis-mic data as exploration moves towards increasingly difficult geology where traps are becoming harder to find. At the same time, good estimates of the subsurface parameters are also becoming more important with regards to increased recovery of oil and gas reserves, as the ability to detect small changes in elastic parameters due to fluid substitution will be useful. By using elastic ocean bottom seismic one is able to record the p-wave and the s-wave. Using these recorded data one is able to estimate the physical parameters in the subsurface using full waveform inversion (FWI). The elastic equation is as follows,

ρ(x)∂t2ui (x, t) − ∂jcijkl (x)∂kul (x, t) = Fi (x, t),

where ui represent ith component of the displacement, ρ(x) is the volume density of the material with variations in x directi-on. The component cijkl is the stiffness tensor, which will give information about the elastic parameters in the subsurface.To find the best model given the observed data, the misfit equation is

S = dobs − dpred(m),where dobs is the observed data and dpred(m) is the predicted

data given the model m. If the misfit is too large, the model mi is updated iterative by

mi+1 = mi − αigiwhere mi+1 is the updated model of mi, and ai is the step length and gi is the derivative of the misfit function. This will loop un-til the misfit function reaches a minimum. By using time lapse data, it will be then be possible to study thechanges in the elastic parameters. Time lapse inversion works as one can see in figure 1, by inverting two seismic data sets from the same area taken at a different time. The difference in the models is found by

€ = l(m0, dmon) − l(m0, dbase),where m0 is the initial model and dmon and dbase are the data at time i + 1 and i respectfully. Given the change in the elastic parameters one is able to monitor the effect of production and injection on the oil and gas reservoirs, either forenhanced oil recovery or CO2 storage.However, in seismic tomography the uncertainty in the model is a challenge. By finding a statistical method that can qualify the result of the full waveform inversion, it is believed that bet-ter decisions will be made regarding the oil and gas reservoirs. As both full waveform inversion and certain statistical methodsrequires a lot of computer power, it is important to optimize the inversion and use an cost efficient statistical methods.In this project the aim is to find a uncertainty estimate of the full waveform inversion, that can be used in time lapse data for oil and gas recovery.

Figure 1: Figure of time lapse data method. The top of the figure shows the seismic data sets beinginverted independently, before the difference in the models is found.

27

Particle growth Kinetics: A physio-chemical modelling approachOmekeh, A. V.1, 3 and Hiorth, A.1, 2, 3 1 The National IOR Centre of Norway

2 Dep. of Energy Resources, University of Stavanger3 IRIS

The kinetics of particle growth from supersaturated solutions or smaller nano-sized particles is of importance in various fields of application, ranging from aerosols responsible for cloud formation to life science. In reservoir engineering, a thorough understanding and prediction of the growth kinetics is crucial for the design of in-depth silicate placement for wa-ter diversion, stability or retention of nano-particle transport during injection in reservoirs, scale formation, etc. The growth of the particles is a function of the agglomeration between particles. As a result, the particle size distribution (PSD) and the surface properties of the particles play a vital role in the growth kinetics especially for metal oxides where the surface properties are dominant.In this study we model the growth kinetics with the aid of the quadrature method of moments (Q-MOM). This allows us to approximate the particle size distributions into groups of fewer classes and then follow the change of the moments of the dis-tributions as particle growth by agglomeration takes place. We model the agglomeration by implementing the Einstein-Smo-luchowski equation that describes the collusion of two smaller particles to form a larger particle and a breakage model for possible disintegration of particles. The surface properties of the particle were modelled as a function of pH, brine composition, temperature and pressure using an extended triple layer surface complexation model (SCM). Such approach allows us to distinguish between the

properties at the surface and the properties at the outer stern layer and diffused layer. This distinction is vital in describing the surface behavior of metal surfaces (e.g. silica) at high pH in the presence of divalent cations.The model developed in this study is illustrated for the growth of barite (BaSO4) particles from supersaturated solutions. It is observed that the surface potential and the supersaturation are the main factors affecting the growth kinetics. The model is also applied in the growth of silica -particles from supersatura-ted sodium silicate solutions. The critical factors determining the growth kinetics are the initial PSD, temperature and po-tential at the stern layer. The surface potential as determined from the SCM is a complex function of mainly brine compositi-on and pH. At the high pH of the sodium silicate solution, the concentration of the divalent ion is critical for determining the stern potential and in turn growth rate of the particles. Finally, the performance of the model is evaluated by comparing the average size of the particles with time as predicted by the model to turbidity evolution of sodium-silicate under test con-ditions. Turbidity is the scattering of light as it passes through the test sample and it is proportional to the particle size and concentration. The particle size as predicted by the model correlates with turbidity measurements.

Figure 1: Illustration of the agglomeration of two particles of radii R1 and R2 to form larger particle with radius R3 such that the volume of the particle is conserved.

28

Using Rainbow trout gill cells (RT-W1) to screen for environ-mental toxicity of produced polymer in polymer flooding

There is an considerable risk that produced water from poly-mer floods will be discharged to the marine environment in offshore enhanced oil recovery scenarios. Marine pollution brought on by discharges related to Norwegian offshore petro-leum extraction is an ongoing threat to the North Atlantic ma-rine habitats. The standardized ecotoxicological testing requ-ired through the activities regulations aims to limit the use of harmful chemicals offshore, but it does not provide knowledge to the public domain nor detailed enough knowledge in itself to assume that environmental risk has been fully assessed. At the same time there is a strong push towards the eliminati-on of live animals in ecological toxicity testing. One approach is to employ cell cultures isolated from the specific target organs of the organism of interest. However, further evaluation of replicability, predictive capacity and applicability is needed for in vitro animal alternatives. In addition, there is also the con-tinuous need for rapid, reliable and affordable environmental toxicity screening/monitoring tools for whole effluents.Because of the growing use of non-standard in vitro methods, OECD has released a guide on conducting such studies to make the abundant, and often very specific data, useable for environmental risk assessors to adjust for the new challenges with the introduction of animal alternatives. Existing ecotoxicological studies on similar polymers fails to provide information on polymer structure, residual monomer content, and absolute molecular weight distributions. Para-meters critical to the risk assessor and the understanding of adverse outcome pathways. Acharya et al. 2010 is one such example, where no information is disclosed reducing the value of the research to third parties. A few studies have included

some of these elements, but again lacks in resolution. Kerr et al. 2014 takes another approach and only reports trade names for commercial products. The studies referred to in the second to last paragraph all suffers from either of these. With this study, the criteria above are met by using the met-hod described in Dayeh et al. 2013 together with reporting the residual monomer content, absolute molecular weight distribution, fully disclosing structures, and in a format that is in accordance with the OECD guideline. It is generally well accepted that EOR-polymers at some point undergoes hyd-rolyzation and depolymerziation, either during use or in the environment. Smaller polymers with varying degrees of hyd-rolyzation has therefore been included in this study.Residual monomer concentrations were determined using reverse phase HPLC-DAD, molecular weight distributions were characterized by size exclusion chromatography (Agilent 1260) and multi angle laser light scattering (Heleos II). Acute toxicity towards rainbow trout gill W1 cells was assessed by using three distinct fluorophores according to the procedure in Dayeh et al. 2013.Our findings strongly (Figure 1) support that the anionic polye-lectrolytes employed in EOR does not exhibit basal cytotoxicity within environmentally relevant concentrations. These findings are in agreement with what has been reported previously for similar compounds in in vivo studies using live fish. It is therefore likely that toxic effects observed at lower levels in ecotoxicological studies using microorganisms, e.g. Harford et al. 2011, is due to mechanical effects which is also what is concluded with in those paper.

Eystein Opsahl and Roald KommedalUniversity of Stavanger / The National IOR Centre of Norway

Figure 1: Relating molar mass distribution and structure to toxicity. PAM = polyacrylamide, HPAM = 35 % hydrolyzed polyacrylamide, PAC = Po-lyacrylic acid, BSA = Bovine serum, SDS = Sodium dodecyl sulfate, PDI = polydispersity index, NOEL = No observed effects level, AM = Acrylamide, AC = Acrylic acid.

Polymer Residual AM (mg/g)

Residual AC(mg/g) Mw kDa PDI NOEL (mg/L)

HPAM 0,65 ± 10 % 5,2 ± 2 % 3012 ± 3.8% 1.6 ± 4.80% 407

PAM 0,29 ± 15 % - 602 ± 1.7% 1.8 ± 8.7% 2340

PAC - 1,2 ± 3 % 667 ± 1.3% 2.2 ± 10.6% 53

SDS (control) - - - - 8

BSA (control) - - 64 ± 1.9% 1 ± 2.7% -

29

The Technology To Select Wells For Cyclic Water InjectionRodionov S.P.1,2,3, Pichugin O.N.1, Kosyakov V.P.1,2,3, Shirshov Y.V.1,2,3, Pyatkov A.A.1,2,3

1 OJSC ”CONCORD”2 Tyumen branch of the Institute of theoretic and applied mechanics SO RAN - TyumF IPTM of S.A.Khristianovich SO RAN

Cyclic water injection is the most cost effective IOR technology. It has been proposed and developed in a number of publica-tions [1-3]. Methodology for design of cyclic water injection projects, described in these publications, are based on 1D so-lutions and do not take into consideration real relative location of existing wells. Cyclic injection project based on 3D geologic models and reservoir simulation is presented in [4]. Design of cyclic injection on reservoir simulator with basic two-phase flow equations is much more time consuming than regular injection design.The main reason of troubles with basic two-phase flow equ-ations in reservoir simulators is the limitations in time step. If the time step is exceeded, the calculation becomes incorrect and unreliable. Time step is limited not only by the conditi-on of sustainability of the difference scheme or condition of finiteness of changes for specific value per one step. Time step has to be much less than halfperiod of a cycle and than time required to balance pressure between adjacent layers, which sometimes reach several days or even hours. In this case using basic two-phase flow equations will be very time consuming. Therefore, flow equations have to be modified so that they could be solved during acceptable time (i.e. during time for simulation of regular injection). One of the ways to reach this target is to average two-phase flow equations by time cycle [2, 3] that do not have metioned above limitations of time step.The research in the above mentioned publications [1-3] is limi-ted by analysis of cyclic injection using analytical solutions of

1D problems, i.e. with substantial assumptions, which results in decreased accuracy of the results. On the other hand, ap-proach based on analytical solutions allows for almost instan-taneous result. The authors propouse the technology that will consolidate strengths of both abovementioned approaches: on one hand – speed (analytical solutions), on the other hand – accuracy (averaged equations). This publication is devoted to development of the techn-ology for designing cyclic injection based on averaging of flow equations [1-3] during adequate time. This publication presents averaged flow equations and expression for intensity of overflow between the layers in case of periodic (sinusoidal) law of change in parameters on the wells. Analytical solution to 1D problem of cyclic injection is obtained., This publication proposes a technology based on this analytical solution for evaluation of cyclic injection effectiveness and relevant ranking of field areas.

References1. Boxerman A.A., Shalimov B.V. On cyclic water injection on oil layers with double porosity // Изв. AN USSR, MZGHG, 1967, №2.2. Surguchev M.L., Tsynkova O.E., Sharbatova I.N.. Cyclic injection in oil layers. M, Izd. VNIIOENG, 1977. 3. Sharbatova I.N., Surguchev M.L., Cyclic injection in heterogeneous layers. M, Nedra, 1988, 121p. 4. Langdalen H. Cyclic Water Injection (A Simulation Study). MS thesis, Norwegian University of Science and Technology, June 2014.

Polymer adsorption at different brine salinitiesIrene Ringen1,2, Markus Lindanger1, Magnus K Raaholt1, Aksel Hiorth1,2,3, Arne Stavland2,3

(1) University of Stavanger, Norway,(2)The National IOR Center of Norway,

(3)IRIS AS – International Research Institute of Stavanger, Norway

Polymer flooding is one of the most promising methods for en-hancing oil recovery from reservoirs. The success of injecting polymer depends greatly on the speed of the polymer front compared to the injection water. Polymer adsorption will tend to slow down the front and inaccessible pore volume will spe-ed up the polymer, and both may be affected by water salinity. Which water is most suitable for a polymer flood? Combining polymer flooding with low salinity flooding is of great interest since it has the potential for improving both macroscopic and microscopic sweep. In addition, a low salinity polymer injecti-on could lead to cost reductions and environmental benefits since less polymer is required to obtain the target viscosity, the thermal stability is increased, and polymer adsorption is reduced.In this work, we investigate how polymer adsorption is affec-ted by different salinities. We set up a system where we can design polymer solutions with the same target viscosity by va-rying brine salinity and polymer concentration. The controlling factor for the polymer viscosity is the product of the intrinsic viscosity and the polymer concentration, Χ=[η]c, which implies that polymer solutions with low intrinsic viscosity must be compensated by a high polymer concentration to achieve

the same viscosity as a polymer solution with higher intrinsic viscosity. Based on this relation we created a master type curve, illustrated in Figure 1, and thus,

we designed 4 different polymer solutions, all with a target viscosity of 10 mPas, that we flooded in sandstone cores to evaluate the polymer adsorption. The sandstone was Bent-heimer cores and the polymer solution was a synthetic ATBS polymer. The polymer concentration was monitored on line by measuring the pressure drop across the porous media and by using a capillary tube viscometer during flooding. The effluent normalized concentration profiles were then used to estima-te the polymer adsorption. The flooding experiments show that polymer adsorption increases with increasing salinity. An important consequence of this is that the polymer front moves with nearly the same speed whether it is mixed in high or low salinity water, which is important to take into account when designing a polymer flood.

Figure 1 Master curve describing the relationship between polymer concentration and modified salinity, c_MIS, for a target polymer viscosity of η = 10 mPas.

30

Modelling polymer solutions for improved oil recovery with tensor differential non-Newtonian fluid models

In this report we present a numerical analysis of rheologi-cal measurements done for several polymers typically used in oil recovery operations. We test several dierential tensor non-Newtonian fluid models to determine which of them give a better prediction of the non-Newtonian viscosities and the first normal stress coecients of the considered polymers. The following non-Newtonian fluid models are included in this research: (i) FENE-P dumbbell model; (ii) FENE-2P model; (iii)

Linear and Exponential Phan-Thien-Tanner models; and (iv) FENE-P bead-spring-chain model with and without hydrodyna-mic interaction.Finally, we have a closer look at how the scalar parameters in the models depend on factors such as polymer concentration in the solution, polymer molecular weight, and salinity of the solvent. The obtained dependencies are compared to thepredictions based on microscopic physics.

Dr. Dmitry Shogin is physicists at The National IOR Centre of Norway. His research is presently focused on advanced tensor models of non-Newtonian fluids, including polymers for oil recovery applications.

Shogin holds a Ph. D. degree in Applied Mathematics and Physics (2015) from University of Stavanger and M. Sc. (2010) from Moscow Institute for Physics and Technology.

Effect of presence of oil and water on chalk mechanicsJaspreet S. Sachdeva1,2, Anders Nermoen1,2,3, Reidar I. Korsnes1,2, Merete V. Madland1,2

1The National IOR Centre of Norway2University of Stavanger, Norway

3International Research Institute of Stavanger, Norway

Research carried out on water-wet chalk has shown how seawater and other simplified brines alter stiffness, strength and time-dependent mechanical parameters. This study deals with the mechanical effects due to changes in the pore fluid composition when oil is also present in the pores. The extent to which flow rate of reactive fluids play a role on volumetric deformation with time is also studied.Outcrop chalk from the Upper Cretaceous Niobrara Formation in Kansas was used in this study. The wetting state of chalk cores was altered to a mixed wet state by flooding 60-40 % Heidrun – heptane oil mixture and aging for three weeks at 90°C. The wettability estimates were based on studies per-formed on 11 cores (seven water-wet and four mixed-wet). They were flooded by a brine that contained both sulfate and tracer ions. The sulfate concentration of the effluent is delayed compared to the non-affine tracer due to sulfate adsorption on charged calcite surfaces. The area between sulfate and tracer concentration versus pore volumes injected is proportional to the calcite surface area in contact with water. The ratio of the area for mixed wet cores to water wet cores gave the wettabi-lity index.Four other chalk samples from the same block were tested for their mechanical properties under isotropic conditions. Two samples were completely water-wet and were saturated by 1.1 M NaCl-brine, and two samples were mixed-wet that were ini-tially saturated by the same brine and then wettability altered in the same way as the four wettability altered cores above. All samples were hydrostatically loaded to a stress level 1.5 times the yield stress at 130°C and left to creep over time. No NaCl-brine was flooded through the cores during loading and initial creep phase. After a period of creep with stagnant 1.1 M NaCl-brine, 0.219 M MgCl2–brine was flooded through all cores with different flow rates while the strain and oil production were monitored.The results show that during loading with stagnant fluids insi-de the pores, the observed elastic stiffness and plastic strength are affected by the water wetness of the core. During creep,

the axial deformation with time for all samples overlapped irrespectively of their wettability and water saturation (Figure 1). Remark that the creep stresses were 1.5 times the yield stresses. This imply that to predict creep rate with time the onset of yield is the controlling parameter, while wettability controls the elastic stiffness and yield strength. Injecting MgCl2 into chalk has shown to induce chemical reactions leading to additional creep rates. When MgCl2 is injected at 0.010 ml/min rate, the strain curve went from having a negative second deri-vative to a straight line with a constant strain rate. Quadrupling the flow rate increased the linear creep rate by a factor of 2.5-4. This show how chemical reactions drive deformation by reducing the solid volume via a reduction in the overall mass and increasing of the mineral density. These chemical reacti-ons, as monitored by the ion chromatography, are insensitive to the initial wettability and hence the saturation of oil and water in the core.

Figure 1. Observed creep strain with time for cores K1 to K4 (K1 and K2 were water wet cores, and K3 and K4 were mixed wet cores). Creep stresses for core (a) K1: 17.91 MPa, (b) K2: 17.94 MPa, (c) K3: 26.08 MPa, and (d) K4: 20.83 MPa.

31

Flooding experiments of phase-partitioning tracer candidates for the inter-well region

Mario Silva1,2,3, Helge Stray3, Mahmoud Ould Metidji1,3, Tor Bjørnstad1,3 1The National IOR Centre of Norway, 4036 Stavanger Norway

2University of Stavanger, Department of Energy Resources, 4036 Stavanger, Norway3Institute for Energy Technology, Tracer Department, 2007 Kjeller, Norway

EOR projects have a growing importance in assuring sufficient oil production to satisfy the demand as the number of mature oil fields increases worldwide together with the fact that most of the remaining large unexplored basins rich in hydrocarbons are located in highly environmentally sensitive and/or remote regions. EOR methods can be costly and the deepest possi-ble understanding of the reservoir is crucial in ensuring the success of such projects. Residual oil saturation (SOR) in the inter-well region of oil reservoirs is an important parameter for its characterization which can be directly used to identify EOR targets, assess efficiency of EOR operations, and the efficiency of volumetric sweep between wells. SOR can be determined through a partitioning inter-well tracer test (PITT) which con-sists of the simultaneous injection of one or more passive and phase-partitioning tracers. This type of tracer test is primarily conceived for mature water-flooded oil fields. To date, only a small number of phase-partitioning tracers for the inter-well region have been developed and several field tests were unsu-ccessful due to insufficient knowledge about the behavior of the used tracers under typical reservoir conditions. The selection of new phase-partitioning tracer compounds for oil reservoirs is focused on a few critical physicochemical properties such as water solubility, pKa, octanol-water partiti-on coefficient, boiling point, etc. When a promising candidate compound is identified, a series of tests must be performed to qualify it.

Such tests may be summed as follows:1. Development of analytical methods for its identification

and quantification in the ppt/ppb level in production water from oil fields;

2. Static experiments to ensure its thermal, chemical, and biological stability under typical reservoir conditions, and also absence of significant interactions with reservoir rock;

3. Assessment of the variation of its partition coefficient as a function of and influence of relevant parameters on it, such as oil composition, water salinity, temperature, and pH. The partition coefficient should also be independent from the compound’s initial concentration.

4. Core/column flooding experiments to assess its dynamic behavior and possible influence of velocity in its charac-teristics or diffusivity in the porous material. The partition coefficient should be constant regardless of the flooding conditions and be sufficiently high to ensure the test’s accuracy. These experiments also contribute for further characterization of the interaction between the candidate and the rock material, which must be absent or negligible.

In this poster, we propose to present our methodology and findings from the core flooding experiments carried out on 7 phase-partitioning tracer candidates already successfully te-sted in the above points 1 to 3. These experiments performed with sandstone and carbonate rock, at different injection flow rates, different temperatures, and with and without residual oil saturation in the cores used.

32

Wettability alteration potential of nanoparticles-Experimental fault or practical opportunity?

Saeed Jafari Daghlian Sofla, Lesley Anne James , Yahui ZhangMemorial University of Newfoundland

Numerous coreflooding and micromodel experiments have examined the potential of nanoparticles to increase oil reco-very. Diverse types of nanoparticles have been tested and, in most studies, a significant enhance in oil recovery is repor-ted. Oil-water interfacial tension (IFT) reduction, wettability alteration, and in-situ emulsification are proposed as a main enhanced oil recovery (EOR) mechanism of nanoparticles. Among these three proposed mechanisms, researchers agree regarding wettability alteration potential of nanoparticles due to significant change in the contact angle measurements in the presence of nanoparticles. However, in the contact angle measurements, the substrates are either aged with (immersed in) nanoparticles-fluid before conducting the experiments or it contacted with nanoparticles-fluid before attachment of the oil droplet on the rock surface. Hence, in both cases, as shown in figure 1, before initiating the contact angle measurements, some nanoparticles have already been trapped at the rock-oil interface and can affect the ultimate results. However, in the practical condition there are no nanoparticles at the oil-rock interface before nanoparticles-fluid injection. Hence, these experiments cannot replicate the real condition. The trapped nanoparticles at the fluid-rock interface may causes to overes-

timate the influence of nanoparticles on wettability alteration. We modified the contact angle measurements in a way to have more consistency with the practical EOR conditions and avoid the mentioned limitation of contact angle measurements. We conducted two series of experiments; a) simple system and b) complex system to evaluate the effect of trapped nanoparticles at oil-rock interface on the wettability alteration results. In the simple system, the substrates are chosen from pure minerals such as calcite, dolomite, and quartz, and decane is used as an oil phase. Furthermore, the silica nanoparticles are dispersed in 1 wt% NaCl. In the complex system, the nanoparticles are dispersed in seawater using our H+ protected method; Hiber-nia crude oil; and real reservoir rocks are used. The results revealed that a large portion of reported wettability alteration using silica nanoparticles is due to mistakes in the experiments (trapped nanoparticles at oil-rock interface). We observed that silica nanoparticles can slightly alter the oil-water-rock contact angle only when the initial wettability of rocks are water-wet and silica nanoparticles has no effect on the wettability altera-tion of oil-wet rocks. We have found that any significant chan-ge in the wettability of the rock surface due to the presence of nanoparticles is likely due to the experimental methodology.

An Alternate Method for Wettability Assessment Using SEM-MLA Edison Sripal1, Lesley James1, Dave Grant2

1Department of Process Engineering, Memorial University of Newfoundland2CREAIT Network - Memorial University of NewfoundlandCorresponding Author: Dr. Lesley James (ljames@mun.ca)

Wettability in rocks have been well documented to influence rock fluid interactions in multi-phase flow systems. Restoration of wettability is inevitable for reliable Special Core Analysis experiments. In the laboratory wettability is often estimated via USBM method, Amott Harvey method or contact angle measurement. As in the case with Special Core analysis experi-ment, wettability measurements are not only time consuming and expensive but also limits the number of samples that can be analyzed. The main objective of this study is to establish a new method for estimating wettability on restored core samples using SEM-MLA method. Previous studies have used SEM-MLA method to understand mineralogical changes at the pore scale, but only after the sample has been treated there-

by compromising the sample integrity. In a seminal method we have used aged samples that are saturated with brine and oil under untreated (unpolished or coated) condition and assessed wettability. Using a FEI Quanta 650 Field Emission Gun (FEG) SEM under very low vacuum conditions a suite of sandstone, carbonate and chalk samples were analyzed and the results were successfully compared with wettability of native state samples. The mineralogy was determined using GXMAP measurement mode within FEI Mineral Liberation Ana-lyzerTM software method was also successfully compared with wettability results from USBM-Amott tests and contact angle measurement.

Figure 1: Nanoparticles at oil-rock interface during contact angle measurement

33

Systematic specific surface area analyses on rocks to implement as a necessary, quick and informative method to understand geomechanical parameter in IOR experiments

Rowena Su Wen Shi Thu1, 2, Mona Wetrhus Minde1, 2, Udo Zimmermann1, 2, Merete Vadla Madland1, 2, Reidar Inge Kornes1,2, Dori Yosef Kalai1,31)The National IOR Centre of Norway, 2)The University of Stavanger, Department of Energy Resources, Stavanger,

3) Department of Energy and Petroleum Engineering, Stavanger

The purpose of this study is to calculate and understand the Specific Surface Area (SSA) of various rock core samples from around the world. The rock samples also include a variety of chalk samples that were collected for various IOR-experiments. Furthermore, several types of rock samples will be tested to compile a thorough database for rock samples using the method at University of Stavanger (UiS) for SSA measurements (Brunauer-Emmett-Teller theory (N2)). The benefits of knowing the generic SSA of a particular rock type is of great significan-ce in understanding the main characteristics of reservoir and non-reservoir rocks. What is more important is understanding the variation of SSA measurements across various samples of the same lithology due to mineralogical or textural variations. However, there is a lack of published content focused around a compilation of SSA measurements and its relationship with mineralogical or texture contributing factors for even the most common rock types. Thus, the goal of the study is to measure, evaluate, observe, and compile SSA analyses and its results for several rock samples in a published form as a quick reference guide. Furthermore, the SSA will be used to calculate permea-bilities of samples and compared with measured values.Specific Surface Area analysis is an important rock characteris-tic in understanding fluid sensitivity and movement through a

porous rock. The specific surface of porous material is defined as the interstitial surface area of the voids and pores either per unit mass or per unit bulk volume of

the porous material. It is obvious that the fine grain materials will exhibit much greater specific surface than coarse grain ma-terials. Some fine porous material contains an enormous high value for its SSA. The SSA of porous material is affected by its porosity, mode of packing, grain size and shape of the grains. SSA analysis is a commonly deployed technique and has been used in the industry for evaluating reservoir characteristics of rocks but the technique has not been deployed across a larger number of samples under similar laboratory conditions (both environment and equipment). The analysis focuses mainly on chalk samples due to the thesis as being related to specific is-sues in relation to Improved Oil Recovery (IOR) research at UiS. The scope of this study can be summarized by the following objectives: • To perform a preliminary unaided analysis of the core

samples to determine lithology and important geologic features such as color, texture, grain size, packing, sorting, apparent porosity, etc.

• To prepare subsamples for analysis in both powder and pellet form through mechanical processes in the lab

• To compile the SSA results versus rock sample types in graphical or tabulated form for cross validation and comparisons

• To compare the calculated permeability using SSA with measured permeability values of different rock types

Figure 1. Berea sample saturated with brine and crude oil (a) BSEM image (b) Mineral map

34

Experimental Investigation of Induced Fracture Effects in Oil Recovery Independent of the Matrix Permeability

Odilla Vilhena, Amir Farzaneh, and Mehran SohrabiCentre for Enhanced Oil Recovery and CO2 Solutions, Heriot-Watt University (HWU)

A deep understanding of effects of fractures is fundamental to minimizing errors in any evaluation, prediction and plan-ning of fractured reservoirs and for the development of new solutions for Enhanced Oil Recovery (EOR) for these reservoirs. This work shows the results of laboratory experiments where two Indiana limestone cores were compared; one of them was unfractured with permeability of 60mD and the second one had an induced fractured with unknown matrix permeability but the total permeability (fracture + matrix) similar to the unfractured core. One objective of this work is to develop a better understanding of oil displacement processes and esti-mating how fractures can affect oil recovery independent of the matrix permeability and how much error could be involved in prediction of reservoir performance when taking into acco-unt only the total permeability and neglecting fractures.Spontaneous imbibition is considered as one of the most important recovery mechanisms in fractured reservoirs. Amott cells were used to perform brine imbibition tests at a tempe-rature of 58ºC simulating the conditions found in Presalt re-servoirs offshore Brazil. Values of oil recovery were monitored during approximately 50 days in both cores, i.e. the unfractu-red and fractured cores. The induced fracture was created along the length of the core by applying a controlled stress dividing the core in two parts. In order to keep the fracture open, the core was fixed with aluminium tape at both ends as illustrated in Figure 1. The results of the spontaneous imbibi-

tion experiments show almost the same value of oil recovery during the first two days of the tests for both cores. Then, the oil recovery in the fractured core starts to increase gradually reaching a value higher than 11% above what had been reco-vered for the unfractured core.The analysis of these results leads to the conclusion that the displacement mechanism in fractured porous media is more complex and the interaction between matrix-fracture needs to be considered in order to avoid error caused by the impact of fractures in a reservoir. The next steps of this project is to perform a series of coreflood experiments in fractured and non-fractured cores focusing on WAG injection. Lack of expe-rimental data for well-defined fractured porous media and the necessity of generating reliable data for improvement of techniques to manage and optimise EOR strategies motivates this study in gathering a systematic experimental data bank for fractured porous medium.

AcknowledgementsThis work was carried out in association with the ongoing Project registered as “BG-14 Fomento à Formação de Recursos Humanos visando aumentar a compreensão dos Reservatórios do Pré-Sal Brasileiro – “EOR”” (UFRJ/HWU/BG Brasil/ANP/CNPq) funded by CNPq through Science Without Borders Programme and BG Brasil under the ANP R&D levy as “Compromisso de Investimentos com Pesquisa e Desenvolvimento”.

Figure 1 – Fractured Indiana Limestone core and the aluminium tape strips fixing the core at both ends.

35

Temperature cycling and its effect on stress-strain relationships in high porosity chalks

T. Voake1, 2, A. Nermoen2, 3, R. I. Korsnes 1, 2, and I. L. Fabricius 1, 41University of Stavanger, Stavanger, Norway

2The National IOR Center of Norway3IRIS AS - International Research Institute of Stavanger, Stavanger, Norway

4Technical University of Denmark (DTU), Copenhagen, Denmark

It is hypothesized that due to the calcite anisotropic thermal expansion, repeated heating and cooling during oil production resulting from varying cold water injection rates into a warm chalk reservoirs, may destabilise and potentially cause perma-nent damage. In order to test this hypothesis a series of tests were performed on two outcrop chalk types of different burial history, from Kansas (USA) and Mons (Belgium). These chalks were used because of their different induration and the degree of contact cement, where Kansas chalk has a higher induration and is hence more cemented than Mons chalk. In addition, to investigate to what degree the pore fluid composition plays a role in the observed mechanical behavior for the two chalk types, tests series were performed with equilibrium calcitic water or with Isopar-H as saturating fluids. Both fluids were selected to minimize the impact of chemical reactions and thereby to focus the study on rock-fluid interactions related to ion adsorption on mineral surfaces. For each chalk type and saturation fluid, we compared the amount of irreversible strain that accumulated during a series of hydrostatic stress cycles (within the elastic domain) from 1.2 to 5.2 and back to 1.2 MPa. All stress cycles were performed at 30°C. Two test procedures were followed: a) samples held at constant 30°C temperature, and b) samples exposed to a temperature cycle

(from 30°C to 130°C and back to 30°C) following each hydro-static stress cycle. For each sample, ten stress and temperature cycles were performed.Results show that when a temperature cycle was performed between each stress cycle, the two chalk types behaved dif-ferently. The Kansas chalk (more cemented than Mons chalk), accumulated significantly more irreversible strain within each stress cycle regardless of the pore fluid compared to Mons chalk. In addition, the water saturated Kansas chalk samples exposed to temperature cycles accumulated more than double total irreversible deformation than samples tested at the con-stant temperatures (Figure 1a,b). Please observe that all stress cycles were performed at 30°C, such that the difference in the observed strain is a consequence of the heat cycle. The effect of temperature cycling was less pronounced for the Mons samples. This may indicate that more competent chalk types with lower Biot coefficients, in which the overall stiffness and strength is more dependent on cement bonds holding neigh-boring particles together rather than electrostatic forces, are more sensitive to anisotropic thermal expansion effects when exposed to temperature cycles, than weaker and less indura-ted chalk types with a higher Biot coefficients.

Figure 1. Hydrostatic loading/unloading and volumetric strain for water saturated Kansas chalk samples tested at 30°C. a) Constant temperatu-re, and b) temperature cycle between each stress-cycle.

36

An Experimental Investigation of Polymer Mechanical Degradation at cm and m Scale

Siv Marie Åsen UiS, IRIS and The National IOR Centre of Norway; Arne Stavland, Daniel Strand, IRIS and The National IOR Centre of Norway;

Aksel Hiorth, UiS, IRIS and The National IOR Centre of NorwayCopyright 2018, Society of Petroleum Engineers

In this work, we challenge the common understanding that mechanical degradation takes place at the rock surface or wit-hin the first few mm. The effect of core length on mechanical degradation of synthetic EOR polymers was investigated. We constructed a novel experimental set-up for studying mechani-cal degradation at different flow rates as a function of distan-ces travelled. The set-up enabled us to evaluate degradation in serial mounted core segments of 3, 5, 8 and 13 cm individually or combined. By recycling we could also evaluate degradation at effective distances up to 20 m. By low rate reinjecting of polymers previously degraded at higher rates, we simulated the effect of radial flow on degradation.Experiments were performed with two different polymers (high molecular weight HPAM and low molecular weight ATBS) in two different brines (0.5% NaCl and synthetic seawater).In linear flow at high shear rates, we observed a decline in degradation rate with distance travelled, but a plateau was not observed. Even after 20 m there was still some degradation taking place. The molecular weight (MW) of the degraded po-lymer could be matched with a power law dependency, MWD ~L-x, where x for the HPAM was 0.07 and x for ATBS was 0.03.

We conclude that in linear flow, the mechanical degradation depends on the core length. However, in radial flow where the velocity decreases by length, the mechanical degradation reaches equilibrium with no further degradation deeper into the formation.For the experiments where we evaluated degradation over large distances at high shear rates, we observed a decline in degradation rate with distance travelled, but we could not con-clude that we reached a plateau. Even after 20 m there is still some degradation taking place. It is important to consider this knowledge when interpreting core scale experiments. Howe-ver, the observed degradation is associated with high-pressure gradients, in the order of 100 bar/meter, which at field scale is not realistic.We confirmed previous findings; degradation depends on salinity and molecular weight. Results show that in all experi-ments with significant degradation, most of the degradation takes place in the first core segment. Moreover, the higher the shear rate and degradation, the higher is the fraction of degra-dation that occurs in the first core segment.

Production Optimization in Reservoir Management Through the Ensemble-Based Trust-Region Method

Yiteng Zhang, University of Stavanger / The National IOR Centre of Norway

Our interest in production optimization is natural. Normally a large amount of residual oil left in the reservoir after primary production leaves the average recovery factor somewherebetween 20% and 40%. Even though improved oil recovery and enhanced oil recovery technologies can possibly achieve a higher recovery factor between 50% and 70%, the remainingreserves can be trapped merely due to a lack of knowledge of reservoirs, such as geological uncertainties, fluid properties, rock properties, etc. The essence of production optimization,in a mathematical sense, pertains to the category of numerical optimization, and more particularly, to seeking values of the constrained variables that optimize the objective.The concept of the ensemble-based optimization has matured over the last several years and is rooted in the field of reser-voir-model-based production optimization. Usually, abacktracking line search scheme is used, but this approach often leads to inefficient searching path direction (e.g. zig-zag patterns). As such, it has some redundancies in finding an optimal solution effectively, increasing the computational overhead with greater risk of problematic numerical instability. Here, a trust-region conjugate gradient method is introduced embedding in EnOpt, motivated by the general applicability and the success in practice. The mathematics (or statistics) of ensemble-based optimization with several mathematical treatments is studied. Conjugate gradient Steihaug method in solving the sub-problem of ensemble-based optimization is carried out, aiming at delivering a faster convergence approach

for reservoir management. The conjugate gradient approach is known for its prescriptive convergence theory, in which the progress can be observed at each iteration for a quadratic programming.For comparison purposes, the Rosenbrock function and five-spot waterflooding are tested for both ensemble-based backtracking method and ensemble-based trust-regionmethod. With numerical experiments, we illustrate that trust-region method is competitive against line search method in ensemble-based production optimization. The proposed approach is then tested on a large scale synthetic model developed by TNO, namely the Brugge benchmark model. The optimization strategy is to control maximum allowed water cut in each connection at which control valves (ICVs) are closed, and the objective is to maximize the net-present-value (NPV). Injectors are controlled using voidage replacement. The methodology performs well on the case considered here. In particular, the entire ensemble of controls is used to adapt the covariance matrix. As such, the gradient estimate of EnOpt is statistically equivalent to that of a Gaussian Mutation Optimi-zation (GMO) algorithm.In general, the method is a viable option to the traditional EnOpt. The method is fast, accurate, and flexible in regard to two-phase field optimization. With the characteristics ofthe trust region method and the conjugate gradient approach, the method can be used as a powerful engine for a variety of optimization tasks.

37

09:00 Registration

10:00 Marit Boyesen, Rector, UiS10:15 Merete Vadla Madland, The National IOR Centre of Norway/UiS and Arne Graue, NFiP10:30 Ingvil Smines Tybring-Gjedde, Ministry of Petroleum and Energy IOR in a government perspective

#1: IOR potential of the Norwegian Continental Shelf10:50 Arvid Østhus, NPD NCS - Even more to gain11:10 Gunnar Hjelmtveit Lille, OG21 The promise of technology to unlock value on the NCS11:30 Questions and PhD stand up 112:00 Lunch

#2: Simulation and modeling of IOR processes13:00 Mojdeh Delshad, University of Texas, Austin Recent advances in chemistry and design expand applicability of polymers to high temperature/high salinity in low permeability sandstone and carbonate reservoirs13:20 Bergit Brattekås, The National IOR Centre of Norway/UiS/UiB Core scale EME for IOR: experiment- modelling- experiment13:40 Aksel Hiorth, The National IOR Centre of Norway/UiS/IRIS Simulation Tools for Predicting IOR Potential on the Norwegian Continental Shelf14:00 Questions and PhD stand up 214:25 Coffee break

#3: Using the data information for IOR15:00 Amos Nur, Stanford University The Future Impact of Digital Rock Physics15:20 Dario Grana, University of Wyoming Stochastic approaches to seismic reservoir characterization for improved modeling and prediction15:40 Morten Jakobsen, The National IOR Centre/UiB/IRIS Time-lapse full-waveform inversion as a smart monitoring tool for future IOR16:00 Jarle Haukås, The National IOR Centre/Schlumberger Analysis of enhanced permeability using 4D seismic data and locally refined simulation models16:20 Alfred Hanssen, ARCEx Seismics in the Arctic: Ice and Bubble Generated Noise16:40 Debate17:00 End of scientific programme day one19:00 Conference dinner

#4: Joint work for future ior09:00 Greetings from County Mayor Solveig Ege Tengesdal09:10 Sigmund Stokka, IRIS/DrillWell DrillWell – Drilling and Well Centre for Improved Recovery09:30 William Rossen, TU Delft A Laboratory Study of Foam for EOR in Naturally Fractured Reservoirs09:50 Questions and PhD stand up 310:15 Coffee break

#5: EOR methods for the future10:30 Matthew Jackson, Imperial College, London Zeta potential changes at mineral-brine and oil-brine interfaces control improved oil recovery during smart waterflooding10:50 Lesley James, Memorial University of Newfoundland Smart SCAL for Smart IOR Solutions11:10 Ana Todosijevic, Wintershall EOR Competence along the Value Chain 11:30 Tor Bjørnstad, The National IOR Centre of Norway/IFE Nanoparticles as oil detectives11:50 Questions12:00 Lunch

#6: Reservoir characterization and decision making13:00 PhD stand up 413:10 Ann Muggeridge, Imperial College, London Screening for EOR and Estimating of Potential

24 April 25 April

Incremental Oil Recovery on the Norwegian Continental Shelf13:30 Danielle Morel, TOTAL Innovation and Standardization for an agile EOR deployment13:50 Aojie Hong, The National IOR Centre of Norway/ UiS What would be the best time for IOR? – Fast analysis in a decision analysis framework14:10 Questions14:20 Coffee break

#7: Sustainability of IOR14:45 Hans Christen Rønnevik, Lundin Unfolding the reality from plays to prospects and fields is continued learning based on awareness of conceptual, factual and technological Incompleteness15:05 Signe Kjelstrup, NTNU Energy efficiency in the process industry: Learning from nature15:25 Questions15:35 Skjæveland Award / summing up16:00 End of conference 38

THE WORKSHOP

Reidar B. Bratvold is a Professor of Petroleum Investment and Decision Analysis at the University of Stavanger, Norway and the Norwegian Institute of Technology where he is teaching and supervising graduate students doing research in uncertainty assessment, risk management, investment analysis, decision sciences, and market based valuation.

Reidar Bratvold, The National IOR Centre of Norway / UiS

Chief Scientist at IRIS in Bergen. Holds a secondary position at the Nansen Center in Ber-gen. Previously he has worked 11 years at the Nansen Center and 13 years in Statoil. He is an expert on data assimilation and has developed and introduced the Ensemble Kalman Fil-ter (EnKF) which is now the most used method for data assimilation in the Earth sciences.

Geir Evensen, The National IOR Centre of Norway / IRIS

Research Director at Schlumberger, in charge of defining the long-term strategy of the rese-arch teams and ensuring development and business impact of new disruptive technology. Holds a Master’s degree in Telecommunications Engineering from Politecnico di Milano.

Senior research scientist at IRIS, where he has worked since 2012. His current research inte-rests include data assimilation, optimization, signal/image processing, and statistical/machi-ne learning. He has more than 30 peer-reviewed publications and 50 conference presentati-ons. He holds a DPhil degree in applied mathematics from the University of Oxford, UK.

Massimo Virgilio, Schlumberger

Xiaodong Luo, The National IOR Centre of Norway / IRIS

Big Data, Little Thinking

Using iterative ensemble smoothers in the presence of model errors

Integration of disciplines and automation of processes for digital production optimization

Big data assimilation and uncertainty quanti-fication in 4D seismic history matching

39

Knut Uleberg, Statoil (Chair)Knut Uleberg works as a researcher at Statoil. He is also one of the members of the technical committee at The National IOR Centre of Norway. Uleberg holds a PhD degree in Reservoir Technology from NTNU, Trondheim. Uleberg is th chair of the workshop.

Alf Birger Rustad, Statoil

Martin A. Fernø, The National IOR Centre of Norway / UiB

Lawrence Cathles III, Cornell University / The National IOR Centre of NorwayMahmoud Ould Metidji, The National IOR Centre of Norway / IFE

Amos Nur, Standford University

Lesley James, Memorial University of Newfoundland

Alf Birger Rustad holds a PhD in Mathematics from the Norwegian University of Science and Te-chnology. He is a specialist in reservoir technology at Statoil, where he has worked with the Open Porous Media project since its early start.

Martin A. Fernø is Professor at the Department of Physics and Technology, University of Bergen, Norway. He has been a visiting researcher at Stanford University (USA), University of Bordeaux (France) and University of Leoben (Austria), and his research focuses on CO2 injection for combined oil recovery and subsurface CO2 storage. He has more than 100 scientific publications.

After his PhD degree from Princeton, Cathles spent 7 years at Kennecott Copper Corporati-on’s Ledgemont Laboratory. He worked at Pennsylvania State University (1978-82) and Chev-ron Oil Field Research Laboratory (1982-86). Since 1986 he has been at Cornell University.Mahmoud Ould Metidji is a Post-Doctoral Researcher between the Tracer Technology Department of IFE and the National IOR Center of Norway. He is currently focused on the use of nanoparticles and fluorescent tracers for oil-reservoir characterization.

Applies rock physics results to the understanding of tectonophysical processes in the Earth’s crust and lithosphere, a major thrust of which is the role of fluids in crustal processes and in energy resources. Nur pioneered the use of seismic velocity measurements to characterize the changing state of oil and gas reservoirs as the volume of fluid in the rock changed during pumping.

Dr. Lesley James is an Assistant Professor and Chevron Chair in Petroleum Engineering in the Faculty of Engineering and Applied Science at Memorial University. Her research focuses on enhanced and improved oil recovery and her time is split between research, teaching undergraduate and graduate level courses in process engineering and oil and gas, and various committees and outreach.

OPM for IOR projects

How to upscale from lab to field: lessons learned from an ongoing CO2-foam field pilot project

Multicomponent tracer methods to asses fracture controlled flow tens of meters from the wellbore & a proposed field test

Spatial heterogeneities: Seismic velocities, moduli, and perm; Fault system complexity; Upscaling schemes; and Diagenesis

The impact of digitalisation of SCAL on field development

40

Olwijn Leeuwenburgh is a Senior Research Associate (part-time) at TU Delft and a Senior Scientist at TNO. He holds a PhD in Geophysics, University of Copenhagen. His main resear-ch interests are conditioning of models to data (data assimilation), model simulation, and model-based optimization, with a particular interest in ensemble-based methods.

Olwijn Leeuwenburgh, TNO / TU DelftEnsemble-based closed-loop field development

THE 2018 PARTNERS

ConocoPhillips

OBSERVERS

www.uis.no/ior

CONTACTCentre Director:Merete Vadla Madlandmerete.v.madland@uis.no+ 47 913 86 589

Administrative Coordinator:Iren Lobekkiren.lobekk@uis.no+ 47 922 55 285

Communications Advisor:Kjersti Riiberkjersti.riiber@uis.no+ 47 995 00 701