Peer Reviewed
SPE Disciplines
Geologic Time
Journal
Conference
Publisher
Author
Concept Tag
Country
Layer | Fill | Outline |
---|
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Gandhi, Ankur (Occidental) | McConkey, Sara L. (Occidental) | Kimbrough, Jeremy (Occidental) | Bolingbroke, Hannah F. (Occidental) | Kapoor, Yogesh (Occidental) | Walker, Thor J. (Occidental) | Buquet, Brandon (Occidental) | Beecher, Richard E. (Occidental) | Tryon, Benjamin R. (Occidental) | Rodrigues, Neil (Occidental) | Kalich, Kevin (Arion)
Successful identification, evaluation, and management of bottlenecks in a complex, offshore production processing systemโthough challengingโcan significantly increase daily production for the system owner. Historically, such optimization plans were developed in relative isolation of the entire production system from wellhead to export pipeline. That approach benefits simplistic systems with sufficient ullage and in which discrete changes do not affect other flow system components. However, the Constitution platform in the Green Canyon area of the Gulf of Mexico, which was commissioned in 2006 with a nameplate capacity of 70,000 BOPD, is a complex system with four fields in varying stages of development. These fields have both dry and wet tree wells with varying fluid compositions and pressures flowing through the facility, which necessitates varying process requirements, making it challenging to manage. Such a system requires a holistic and focused approach by all technical and commercial disciplines. This paper focuses on a multidisciplinary process developed to identify, evaluate, and eliminate interdependent bottlenecks on the Constitution platform and its flowline network during a 16-month period. A multidisciplinary study was kicked off in 2017 to address these complex bottlenecking issues, and the resulting project achieved a 30% improvement in deliverability of the process system.
Smart completions enable physical measurements over space and time, which provides large volumes of information at unprecedented rates. However, optimizing inflow control valve (ICV) settings of smart multilateral wells is a challenging task. Traditionally, ICV field tests, evaluating well performance at different ICV settings, are conducted to observe flow behavior and configure ICV's, however this is often suboptimal. This study investigated a surrogate-based optimization algorithm that minimizes the number of ICV field tests required, predicts well performance of all unseen combination of ICV settings, and determines the optimal ICV setting and net present value (NPV).
A numerical model of a real offshore field in Saudi Arabia was used to generate scenarios involving a two-phase (oil and water) reservoir with trilateral producers. Multiple scenarios were examined with variations in design parameters, mainly well count, placement and configuration. Eight discrete settings were assumed to match the commonly installed ICV technology, where all possible scenarios were simulated to establish ground truth. The investigation considered three major algorithmic components: sampling, machine learning, and optimization. The sampling strategy compared physics-based initialization, space-filling sampling, and triangulation-based adaptive sampling. A cross-validated neural network was used to fit a surrogate dynamically, while enumeration was adopted for optimization to avoid errors arising from using common optimizers.
This study evaluated two sampling techniques: space-filling and adaptive sampling. The latter was found superior in capturing reservoir behavior with the smallest number of simulation runs, i.e. ICV field tests. Algorithm performance was evaluated based on the number of ICV field tests required to: 1) surpass an R2 threshold of 0.9 on all unseen scenarios, and 2) match the optimal ICV settings and NPV. Surface and downhole flow profile prediction and optimization were achieved successfully using this approach. To determine the diminishing value of additional ICV field tests, the triangulation sampling loss was used as a stoppage criterion. When running the algorithm on a single producer for both surface and downhole oil and water flow prediction, the algorithm required six and 11 ICV field tests only to achieve 80% and 90% R2 across the different cases of this real reservoir model. Fishbone wellbore configurations were found to pose a more challenging task as changes in any ICV pressure drop affects multiple laterals simultaneously, which increases the level of interdependence. The resultant surrogate was used to decide on the optimal settings of ICV devices and also predict the NPV effectively. Further improvement was accomplished through adaptively sampling and fitting surrogate to rather predict NPV explicitly where NPV predictions were generated with nearly 95% R2 given only ten ICV field tests.
Using adaptive sampling and machine learning proved effective in the prediction of surface and downhole flow profiles, and optimization of smart wells. The method further allows for dynamically optimizing field strategy in a reinforcement learning setting where production data are used continuously to further improve the prediction performance.
The need for petroleum engineers to provide energy for the world population in the upcoming decades and the role of SPE in this regard are discussed in this paper. There have been recent papers suggesting that there is a diminished or no need for petroleum engineering (e.g., SPE 194764, The End of Petroleum Engineering as We Know It; SPE 195908, Petroleum Engineering Enrollment: Past, Present and Future). This is simply inaccurate and the proliferation of their message is alarming in that it may deprive the industry from having the required talent to produce the energy the world needs from natural gas and oil for decades to come.
Engineering disciplines, including petroleum, have been transforming all the time. The practices we use today have little resemblance to what we did 20 years ago. They will continue to evolve, not stop, due to three major points: Although the percentage of the energy from oil and natural gas the world needs is expected to decline from about 53% to about 48% over the next 30 years, the amount of production is actually expected to increase by 20% for oil and 50% for natural gas. This is due mainly to expected increase in world population (from 7.5 to 9 billion) and improvements in the economic conditions of developing countries. Considerable advances and efforts in all aspects of petroleum engineering (from drilling under more challenging conditions to completion and production from deeper wells to increasing recovery from heavy oil and through EOR) are required to meet those higher production levels. We know that even maintaining production at the current level is a challenge. Climate change is real. As engineers we do not debate the science, and we should transform our industry in every possible way to minimize any adverse effects of our work on the environment and to comply with state regulations. One manifestation of this has been the measured increase in historical natural gas production and simultaneous decrease in CO2 emissions. Another example as one of the major factors in reducing Green House Gases is to inject and store CO2 in underground formations. Petroleum engineers will be the ones tasked with finding answers to the challenge of injecting in already fluid-saturated formations. With the recent developments in Data Science and Engineering Analytics, there is a greater need for petroleum engineers with an understanding of physics to take advantage of these improvements and optimize the processes we use (Anadarko's SPE 187222 Creating Value by Implementing an Integrated Production Surveillance and Optimization System โ An Operator's Perspective and Chevron's SPE 181437 Application of Machine Learning in Transient Surveillance in a Deep-Water Oil Field, are good examples).
Required and expected efforts by various stake holders of SPE; educational institutions; national and international operators, service companies, and regulatory bodies to provide needed petroleum engineers are discussed.
Soni, Kishan (Petroleum Affairs Division, Department of Communications, Climate Action and Environment, Ireland/ iCRAG, School of Earth Sciences, University College Dublin) | Manzocchi, Tom (iCRAG, School of Earth Sciences, University College Dublin) | Haughton, Peter (iCRAG, School of Earth Sciences, University College Dublin) | Carneiro, Marcus (iCRAG, School of Earth Sciences, University College Dublin)
Oil reservoirs hosted in deep-water slope channel deposits are a challenge to manage and model. A six-level hierarchical arrangement of depositional elements within slope channel deposits has been widely recognized, and dimensional (width and thickness) and stacking (amalgamation ratio and volume fraction) data have been acquired from published studies to establish parameters for a representative slope channel system. A new static modelling workflow has been developed for building models of channel complexes based on a simplified hierarchical scheme using industry-standard object-based modelling methods and a new plugin applying the compression algorithm. Object-based modelling using the compression algorithm allows for independent input of volume fractions and amalgamation ratios for channel and sheet objects within a hierarchical modelling workflow. A base-case channel complex model is built at the resolution of individual sandstone beds, conditioned to representative dimensional and stacking characteristics of natural systems. Inclusion of explicit channel axis and margin regions within the channels governs bed placement and controls inter-channel connectivity where channels are amalgamated. The distribution of porosity and permeability within these beds mimics grain-size trends of fining in the vertical and lateral directions. The influence of various geological parameters and modelling choices on reservoir performance have been assessed through water-flood flow simulation modelling. Omission of the compression method in the modelling workflow results in a three-fold increase in oil recovery at water-breakthrough, because the resultant unnaturally high amalgamation ratios result in overly-connected flow units at all hierarchical levels. Omission in the modelling of either the bed-scale hierarchical level, or of the axial and marginal constraints on the bed placement in models that do include this level, results in a two-fold increase in oil recovery at water-breakthrough relative to the base-case, because in these cases the channel-channel connections are too permissive.
Learn more about training courses being offered. Learn more about training courses being offered. This course covers the fundamental principles concerning how hydraulic fracturing treatments can be used to stimulate oil and gas wells. It includes discussions on how to select wells for stimulation, what controls fracture propagation, fracture width, etc., how to develop data sets, and how to calculate fracture dimensions. The course also covers information concerning fracturing fluids, propping agents, and how to design and pump successful fracturing treatments. Learn more about training courses being offered. Current and future SPE Section and Student Chapter leaders are invited to engage and share.
Smart completions enable physical measurements over space and time, which provides large volumes of information at unprecedented rates. However, optimizing inflow control valve (ICV) settings of smart multilateral wells is a challenging task. Traditionally, ICV field tests, evaluating well performance at different ICV settings, are conducted to observe flow behavior and configure ICVs; however, this is often suboptimal. This study investigated a surrogate-based optimization algorithm that minimizes the number of ICV field tests required, predicts well performance of all unseen combination of ICV settings, and determines the optimal ICV setting and net present value (NPV).
A numerical model of a real offshore field in Saudi Arabia was used to generate scenarios involving a two-phase (oil and water) reservoir with trilateral producers. Multiple scenarios were examined with variations in design parameters, mainly well count, placement, and configuration. Eight discrete settings were assumed to match the commonly installed ICV technology, where all possible scenarios were simulated to establish ground truth. The investigation considered three major algorithmic components: sampling, machine learning, and optimization. The sampling strategy compared physics-based initialization, space-filling sampling, and triangulation-based adaptive sampling. A cross-validated neural network was used to fit a surrogate (in this case, machine learning algorithm) dynamically, whereas enumeration was adopted for optimization to avoid errors arising from using common optimizers.
This study evaluated two sampling techniques: space-filling and adaptive sampling. The latter was found superior in capturing reservoir behavior with the smallest number of simulation runs (i.e., ICV field tests). Algorithm performance was evaluated based on the number of ICV field tests required to exceed an R2 threshold of 90% on all unseen scenarios and match the optimal ICV settings and NPV. Surface and downhole flow profile prediction and optimization were achieved successfully using this approach. To determine the diminishing value of additional ICV field tests, the triangulation sampling loss was used as a stoppage criterion. When running the algorithm on a single producer for both surface and downhole oil and water flow prediction, the algorithm required only 6 and 11 ICV field tests to achieve 80% and 90% R2 across the different cases of this real reservoir model. Fishbone wellbore configurations were found to pose a more challenging task because changes in any ICV pressure decrease affects multiple laterals simultaneously, which increases the level of interdependence. The resultant surrogate was used to decide on the optimal settings of ICV devices and effectively predict the NPV. Surrogates, in this approach, are statistical proxies of the targeted ground-truth production function. Further improvement was accomplished through adaptively sampling and fitting surrogates to predict NPV explicitly, where NPV predictions were generated with nearly 95% R2 given only 10 ICV field tests.
Using adaptive sampling and machine learning proved effective in the prediction of surface and downhole flow profiles and optimization of smart wells. The method further allows for dynamically optimizing field strategy in a reinforcement learning setting where production data are used continuously to further improve the prediction performance.