Search

Summary An analysis is presented that quantifies the value-adding brought to a well by installing downhole flow control valves (FCVs) in a reservoir with uncertainties represented by multiple equiprobable realizations, two aquifer strengths, and three oil/water contacts (OWCs). The complexity of the model makes nonintuitive its response to FCV adjustment. From an openhole (OH) baseline net present value (NPV), it was possible to extract specific value-adding for the following, nonadjustable valves, the ability to adjust the valve, and how future information can add further value. The issue of measurement bin size is discussed, and the implication of making them too large, or too small, is described. Finally value erosion due to valve failure is also addressed. The solution presented used dynamic programming (DP), which was validated against a full enumeration of the problem using reservoir simulation. Approaches to proxies to traverse the (potentially large) decision space are also discussed.

After 9 years of contributing to JPT, it’s time to hang up my spurs. I’ve had the privilege of being exposed to a broad range of topics, technologies, and papers (good, bad, and indifferent). The technological advances I have observed, and the changes in our way of working, have been startling. I wanted to use this opportunity, however, to raise a couple of issues, namely the effect and utility of machine learning (ML) and the usage of point-and-click engineering software. Back in June 2018, I wrote an editorial outlining concerns I held regarding the promise of ML/artificial intelligence to resolve complex physics-based engineering problems. Nothing has changed my opinion since then. There is no doubt that ML has utility in advancing our understanding of certain problems, such as monitoring of large numbers of wells and preempting run-time problems. To claim that ML will provide an accurate forecast of, say, a new multi-phase network flow model, however, would be—let’s be diplomatic here—problematic. In my opinion, ML needs to be applied with surgical precision, not as a blunderbuss shot in the rough direction of a problem with hopes that a pellet will hit a target. It’s still too soon to draw conclusions on the broad utility of ML, but it is unlikely to prove to be the universal cure-all that some have promoted. The best way to summarize the what, where, and how, to apply ML is: It depends. Time will tell, and I remain hopeful but cautious. The other issue is a need to maintain an appreciation of the underlying physics of the problem being modeled. Continuously evolving software has enabled greater overview, visualization, and intricate analysis of ever-more-complex systems. This is a definite boon and provides a level of performance insight hitherto unheard of. Such software, however, still requires a pilot to operate it, and key inputs blunderbuss shot in the rough and physics may be masked, or even hidden, beneath a dazzling interface. Optimization of such systems is also nontrivial, yet this nontrivial task is now made almost pedestrian by a mere click of the mouse. It is so easy to accept software-generated solutions, even though they may turn out to be suboptimal (or even wrong) because of some errant input or a poorly defined parameter. I am a strong proponent of the amazing advances made in software, democratizing and seamlessly unifying other-wise specialist tools and disciplines. But no substitute exists for digging into the problem physics/mechanisms themselves and applying simple pen-and-paper (“back-of-the-envelope”) checks. So, the next time that gas bill arrives, I might suggest keeping the envelope it came in; one never knows when it may prove useful. Recommended additional reading at OnePetro: www.onepetro.org. SPE 203139 - Reduce Cost and Schedule by Developing an Optimum Well Surface Facility Program by Cathy Farina, DyCat Solutions, et al. SPE 206181 - Development of Methods for Top-Down Methane Emission Measurements of Oil and Gas Facilities in an Offshore Environment Using a Miniature Methane Spectrometer and Long-Endurance UAS by Brendan Smith, SeekOps, et al. OTC 30292 - Development of an Assessment Procedure To Predict Local Buckling Behavior Using Probabilistic Method With Finite-Element Techniques for Reeled Pipeline With Additional Features by Sheralia Ufairah Abdullah, McDermott International, et al.

The notion of reducing our environmental footprint, minimizing leaks and spillages, and identifying operational efficiencies is nothing new. We have been addressing these issues for years. Sustainability, however, has gained a higher profile recently, especially since the 2015 United Nations Framework Convention on Climate Change accord and the evolution of alternative energies. It came, therefore, as a pleasant surprise to review an extraordinary wealth of well-written papers relating not only to this topic, but to all manner of fascinating engineering issues. Let me back up a bit. For 6 years, I have provided commentary concerning matters relating to reservoir simulation. This was often illuminating, but my primary interest has always been in full-field modeling: coupling our models for flow in porous media, wellbore, network, and to facilities. I firmly believe that unified full-field modeling can furnish benefits both operationally and economically. In accepting the challenge to write this, however, I did not expect to encounter the extraordinary breadth of topics covered. Saying that, one clear thread did emerge from the articles reviewed - namely reliability, emissions control, and energy saving and monitoring, all of which relate to sustainability in one way or another. Novel methods for valve- failure prediction, pipeline coatings, floating storage tank roof optimization, and the use of drones (from spray painting to surveillance) along with robots for unmanned installations were all topics represented in the papers provided me. Paper SPE 198165 is an informative overview of alternative fuels and how these may affect Middle East production; however, this should appeal to a wider readership because it reviews alternative energies nicely. Paper IPTC 19775 provides a means to better quantify separator shrinkage, which can result in errors in stock-tank rates up to 20% (essential if one is to quantify unit energy consumption accurately). Paper SPE 197759 proposes methods for energy savings. These papers represent but a small sample of a broader swath of articles from the past year relating to this catch-all term “sustainability.” Recommended additional reading at OnePetro: www.onepetro.org. SPE 197753 Parametric Study on Wind Fatigue Life of Long Span Flare-Boom Based on Directional Input Parameters to Harris Wind Spectra and Selection of Wind Blocks by Sachin Samant, National Petroleum Construction Company, et al. SPE 198133 Internal Corrosion Severity Ranking of Crude Oil Pipelines by Amer Jaragh, Kuwait Oil Company, et al. SPE 197620 Ground Robotics Enabler for Normally Unattended Installations by Jean-Michel Munoz, Total, et al.

Technology Focus The original realm of this Technology Focus was reservoir simulation, but the scope has been expanded to include all simulation. This is, indeed, a broad sweep, and I fear that I may not be able to do full justice to technical domains in which I am not fully conversant. Nonetheless, muddling through, I have selected three papers with reasonably broad coverage. The one I particularly like describes an open-source 3D-printing micromodel tool kit. This highlights the need to validate simulation through experimental observation, and the work provides a practical means to do so. This will be my last editorial (it has been 6 years now), and I leave you with two main thoughts, one positive and one less so. My optimistic remark concerns the advances I have observed in the field of simulation of complex reservoirs, especially fracturing of tight reservoirs. While the topic remains challenging, the advances made have been quite significant over the past 5 years. Nonetheless, in my view, we still do not possess a full understanding of oil production in unconventional fractured reservoirs. Our ability to forecast such assets remains elusive, even with copious amounts of analytics, mountains of data, and an arsenal of machine-learning tools. We still cannot ascribe the level of confidence to such assets as we wish would be possible. More fundamental experimental investigation is necessary here, and, while we are gradually increasing our understanding, the journey has some way to go. My final comment concerns buttons. Specifically, I refer to these so-called big green “simulate” buttons: the ones that entice a user to blindly “press it, and for-get it” (with apologies to Ron Popeil). Well-crafted, user-experience-optimized, appealingly designed interfaces are now standard. Nothing new in that. Nonetheless, I cannot help but feel that, rather than assisting the engineer, such interfaces form a metaphorical barrier between the user and the simulation engine. I am of the generation that was quite happy navigating large keyword-driven ASCII files with the “vi” editor (remember that?). While these were awkward, slow, and often excruciatingly painful to operate, being forced to work directly with keywords and ASCII files yields one very significant advantage: an unavoidable and direct connection with the data. One had no choice but to become acquainted with all aspects of an important keyword and its input requirements. This ensured consistency of data input and facilitated a closer bond between user and simulator (greater transparency of what was going on under the hood). Being unashamedly old school, I feel that “optimized user-interface (UI) dashboards” often cast a misty veil over human/machine connectedness and sometimes may even impede the pathway to understanding of simulation behavior and the solution itself. My point here is this: Do not hesitate to dive into the files typically generated by these UIs and be unafraid to be old school, even if only for a few moments. The insight this affords is well worth the effort. Saying this, I am clearly showing my age, so it’s time for a fresh face to take over this editorial. I thank you for your patience over the past few years. Meanwhile, I feel an overwhelming urge to write another technical paper (that no one will read), written in TeX and coded in FORTRAN77, using my trusty “vi” editor—happiness awaits. Recommended additional reading at OnePetro: www.onepetro.org. SPE 191213 Application of Memory Formalism and Fractional Derivative in Reservoir Simulation by Mahamudul Hashan, Memorial University of Newfoundland, et al. SPE 193880 A Massively Parallel Algebraic Multiscale Solver for Reservoir Simulation on the GPU Architecture by A.M. Manea, Saudi Aramco, et al. SPE 193844 A Bayesian Sampling Framework With Seismic Priors for Data Assimilation and Uncertainty Quantification by Siavash Nejadi, University of Calgary, et al.

Technology Focus A radical digital revolution is happening all around us (or so we are told). This digital domain is moving so rapidly that new acronyms are being added to our lexicon almost weekly (BD, DA, DNN, GAN, and ML, to name but a few). So how will this affect reservoir simulation? Speaking for myself, I really do not know. However, I came across an interesting article from the Harvard Business Review that examined the effect of innovative information technology (IIT) on some large US-based consumer businesses, and, while specific benefits of IIT were clearly stated, the authors opined, “In times of radical technological change, there’s a lot of figuring out to do. [We] have to understand what new technologies can do.” Applying this to reservoir simulation, we apparently need to understand better when and, more importantly, when not to use such technology—to appreciate its bounds, its limitations, its range of validity, and so on. Deep neural networks (DNNs), for example, are excellent at labeling images (cat, dog, axolotl, etc.). However, some futurists have an impression that IIT can solve almost any kind of problem, sometimes leading to extravagant claims about its potential utility. Futurists are optimists in nature (the ones I have seen certainly are). However, such buoyant proclamations require some counterbalance through honest questions, healthy discussion, and a reluctance to accept such bold assertions as dogma. Without wishing to appear too skeptical, I came across an article from The Royal Society that is well worth quoting directly: “No matter their ‘depth’ and the sophistication of data-driven methods, in the end, they merely fit curves to existing data. Not only do these methods invariably require far larger quantities of data than anticipated … they can also fail in circumstances beyond the range of the data used to train them because they are not designed to model the structural characteristics of the underlying system.” While this article focuses on biology and medicine, the authors apply their keen insight to any multiscale and complex system, which covers our domain of interest. I feel that a tempered, modestly restrained approach to IIT is indeed wise (as applied to reservoir simulation), at least until we have had time to fathom where this digital revolution is leading. Recommended additional reading at OnePetro: www.onepetro.org. SPE 186079 Implicit Modeling for Permeability Enhancement in Carbonate Reservoirs: A Novel Approach To Bridge Data Gaps for Honoring Dynamic Observations by Arthur P.C. Lavenu, Abu Dhabi Marine Operating Company, et al. SPE 187046 Integrated Reservoir-Network Simulation Improves Modeling and Selection of Subsea Boosting Systems for a Deepwater Development by Gaurav Seth, Chevron, et al. SPE 187453 Assessing Single EOS Predictability Using PVT Properties of a Wet-Gas Reservoir on a Compositional Simulator by Bander N. Al Ghamdi, Saudi Aramco, et al.

Technology Focus Benchmark cases are a good thing. They provide a systematic means to compare processes, technologies, and performance that benefits companies, industry, and research organizations. One notable benchmark was deliberated at a 2008 SPE Applied Technology Workshop in the picturesque Belgium town of Bruges. This particular challenge involved history matching followed by forecast optimization. The goal was to maximize a 20-year net present value (NPV). The results were published as SPE-119094-PA in 2010. The Bruges study has since become the basis of numerous articles and simulation-based studies and continues to be influential to this day. Now, almost 10 years later, a new benchmark challenge has been set: Olympus. This has different, but no less challenging, objectives, and the goal is, once again, to maximize a 20-year NPV subject to some stated constraints. Olympus has been set by Integrated System Approaches for Petroleum Production 2 (ISAPP-2), a research partnership between Delft University of Technology and the Netherlands Organization for Applied Scientific Research (TNO), who originated the original Bruges challenge, along with Eni, Statoil, and Petrobras as industrial partners. For this, we owe a debt of thanks. Issues identified in the original Bruges exercise have been addressed, and a more streamlined benchmarking procedure has been defined. The ambition is no less broad and can be concisely stated as “field-development optimization under uncertainty.” While Olympus has no history-matching component (which formed part of the Bruges challenge), we are now faced with geological uncertainties related to faulting, fault throws, barriers, and channeling. The ensemble of 50 Olympus reservoir models provides such geological uncertainties that we must face when dealing with real field-development planning studies. The participant must decide how many producers and injectors to drill and where to drill them, along with their operational management during a 20-year production forecast. All this, of course, is addressed in the presence of geological uncertainty. One can imagine the results will vary considerably in formulation, ambition, and approach. I hope that this will not be considered a winner-take-all competition because that goes against the spirit of such a benchmark. The reward will ultimately turn out to be increased insight and appreciation of new approaches and techniques that will be shared throughout the industry. So, I humbly request that you consider participating in this challenge. Clear and concise online material is available from the ISAPP-2 website (www.isapp2. com/optimization-challenge), which also instructs you on how to download model files. The gauntlet has been thrown down, and I know that we, as an industry, have the wherewithal to pick it up. Recommended additional reading at OnePetro: www.onepetro.org. SPE 182637 Probabilistic Uncertainty Quantification of a Complex Field Using Advanced Proxy-Based Methods and GPU-Based Reservoir Simulation by N. Goodwin, Essence Products and Services, et al. SPE 181686 Effects of Confined Space on Production From Tight Reservoirs by Brian C. Stimpson, Texas A&M University, et al. SPE 182636 Dynamic Mesh Adaptivity for Immiscible Viscous Fingering by A. Adam, Imperial College London, et al. SPE 182718 A Massively Parallel Semicoarsening Multigrid Linear Solver on Multicore and Multi-GPU Architectures by A.M. Manea, Stanford University, et al. IPTC 18955 Building More-Realistic 3D Facies-Indicator Models by Thomas Le Blévec, Imperial College, et al.

Overview Reservoir-simulation-model inputs are numerous, and uncertainty is pervasive— before, during, and after development. On top of that, there is always pressure to deliver quality results as quickly as possible. This gives rise to a simple question, one that has yet to find a simple answer: How refined is refined enough and how coarse is too coarse? I run the risk of oversimplification here, but it seems we are faced with a classic dichotomy, one that is exacerbated by the pull of advances in high-performance computing that permit ever-greater model refinement, while, simultaneously, we have the push of (possibly stochastic) sampling of uncertainty, thereby encouraging the development of simplified (or surrogate) models that can run hundreds, even thousands, of times. The question is one of striking the right balance between two apparently contradictory approaches to simulation. The adage “horses for courses” is not particularly helpful in itself, even though it is probably appropriate. Does one have two distinct models with roughly commensurate scaling, or can we build a single all-purpose model with different scales within it that is both fast and accurate with multiscale grid? Dimensional scale represents just one aspect of the term “multiscale,” which I have mistakenly taken to mean just the juxtaposition of, essentially, geometrical scale within a single model (such as that found in coupling a simulation grid and the wellbore). The large ratio associated with domain size, and the resolution of the geological data, is usually managed by upscaling. However, so-called multiscale methods represent a new avenue of research, one that may provide a bridge between the aforementioned push and pull of refinement resulting from the needs of different decision makers. Multiscale, which has been the subject of ongoing study over the past decade, knits together geometrical quantities (dimensional scale) with tailored computational schemas (numerical scale). This multifaceted multiscale concept may offer a means to construct an accurate coarser-scaled model, one honoring the attributes of the fine-scale heterogeneous geological data from both numerical and spatial standpoints. This method class computes local basis functions for the solution variables, to construct a smaller (coarse) system for computing an approximate solution on the original simulation grid. While it is too early to say whether this broader notion of multiscale (numerical and geometrical) will provide a single, unifying, model for engineers, it is possible that this, or some other such method, may strike that elusive balance between refinement (accuracy) and surrogacy (speed). For those interested in reading up on this topic, the peer-reviewed SPE papers SPE 119183 and SPE 163649 provide more detail and clarify the status of some ongoing research. JPT Recommended additional reading at OnePetro: www.onepetro.org. SPE 169063 Application of Multiple- Mixing-Cell Method To Improve Speed and Robustness of Compositional Simulation by Mohsen Rezaveisi, The University of Texas at Austin, et al. SPE 177634 Multiscale Geomechanics: How Much Model Complexity Is Enough? by Gerco Hoedeman, Baker Hughes SPE 174905 Experimental Design or Monte Carlo Simulation? Strategies for Building Robust Surrogate Models by Jared Schuetter, Battelle Memorial Institute, et al. SPE 169357 Reduced-Order Modeling in Reservoir Simulation Using the Bilinear Approximation Techniques by Mohammadreza Ghasemi, Texas A&M University, et al.

Technology Focus Coupled, full-field, integrated, holistic, or asset: Call it what you will, this class of modeling is becoming increasingly commonplace. The ability to interrogate the response of a system, from the reservoir to fiscal meter, is quite potent and encourages integration in all senses of the word—technically, operationally, and organizationally. Such end-to-end models may reveal bottlenecks, constraints, and potential flow-assurance issues. While the benefits are manifest, two related issues are worthy of consideration: complexity and piloting. The complexity of any individual component (such as the reservoir model) refers to the engineering judgment required to ensure that it is fit for purpose. For example, should one use a simple model (computationally rapid but approximate) or should a more complex model (computationally slower but accurate) be adopted? The former is better suited for rapid scenario screening and sensitivity analysis, while the latter is better suited for detailed engineering, with a spectrum of models in between. Piloting is another consideration. The degree of technical insight required to competently steer any single component simulator is often quite high. Having several simulators under a single, seamless platform places an extra burden on the pilots involved because the added layer of software (the integration platform) may mask possible shortcomings in a model, thus rendering potentially incorrect results and even suboptimal capital decisions. These considerations notwithstanding, I am genuinely optimistic about the potential these full-field models will bring and the insights they can deliver. More articles are certain to appear detailing predictions vs. observations and how planning, operations, and economics benefited from such integrated studies. I respectfully submit that the story of fully coupled, full-field, integrated asset modeling has only just begun. JPT

Technology Focus As the development of shale oil and gas becomes increasingly significant, so does the need for modeling tools for their accurate and timely forecasting. The basic question then arises: Are the simulation tools that we have suitable for the job? Can they explain, and ultimately forecast, the contribution to flow from natural or hydraulically induced fractures? Ultimately, we need to be confident that the results from such models are sufficiently reliable for decision-making purposes. As Alexandre Emerick stated in JPT (April 2014), “The goal is to generate models for production forecasting aiding the decision-making process involved in the development and management of petroleum reservoirs.” The ability to model this increasingly important asset class is one of the primary issues facing reservoir and simulation engineers today. Our existing arsenal of simulation tools is certainly time-tested for conventional assets, but can they deliver reliable forecasts for this emergent class? Recent literature is rich with relevant studies ranging from fundamental laboratory experiments to entirely empirical, data-driven, approaches for predictive reservoir management. These articles raise numerous questions, including the possible oversimplification of a complex problem, the very nature of the production mechanisms, and the role played by any pre-existing discrete fracture networks. Additional concerns relating to accuracy and reliability of the underlying data used to populate such models have also been raised. Keeping an open mind on this topic is, I feel, appropriate. One needs to balance the very real need for such predictive tools today with the recognition that future evidence may disrupt certain preconceptions. To this end, four articles have been selected that span a spectrum of ideas. One considers practical applications and solutions, while another covers the emerging technique of molecular simulation. Another eloquently probes various notions in our current state of understanding of shale assets and examines issues in interpreting the forecasts furnished by our simulation models. The fourth article demonstrates some creative thinking by representing a planar fracture by use of a proxy comprising a coupled flowing network model, a concept that may prove flexible and warrants further investigation. Conventional modeling tools used to predict flow in tight, fractured shales (e.g., dual-porosity models, local grid refinements, tartan grids) may, indeed, provide reliable forecasting. Analytic and data-driven approaches are also acceptable in some circumstances. However, as our knowledge in exploiting these assets deepens with new theory, laboratory experiments, and field and operational experience, we should be prepared to reconsider some of today’s axioms in the future.