Spivey, Benjamin (ExxonMobil Upstream Research Company) | Bailey, Jeffrey (ExxonMobil Upstream Integrated Solutions) | Pokluda, Joshua (ExxonMobil Upstream Integrated Solutions) | Page, Coby (ExxonMobil Upstream Integrated Solutions)
Drillpipe and casing can become stuck while running into a wellbore or pulling out of a wellbore which can lead to significant non-productive time (NPT) events and cause costly remediation and lost production. These sticking events can be caused by wellbore quality issues, also called mechanical sticking, or differential sticking due to the pressure differential between the wellbore and formation pressure which creates normal force on the tubular. Pipe running friction is traditionally monitored by viewing trends for average hookload and torque values while picking up, slacking off, or rotating off-bottom. These average hookload and torque values only apply to dynamic friction factors and do not differentiate the cause of friction.
Static friction correlates with differential sticking whereas dynamic friction can be caused by mechanical sticking. This paper discusses a sticking advisory system for detecting both static and dynamic sticking as friction increases while running into or pulling out of the wellbore. The system uses one-second surface drilling data to identify friction indicators in the axial and rotational directions based on data-driven analysis of time series data trends. The system can create these trends in near-real-time to enable a team to monitor ongoing operations and make proactive decisions. The system trends may also be used in hindcasting mode to identify thresholds for friction indicators based on historical events. The friction indicators are validated by comparing data trends from wells with and without known sticking events and correlating trends with operational activities. Two indicators, Static Sticking Index and Dynamic Sticking Index, are introduced to provide useful surveillance parameters during pipe running operations.
Characterization of key parameters in unconventional assets continues to be challenging due to the geologic heterogeneity of such resources and the uncertainty associated with fracture geometry in stimulated rock. Limited data and the accelerating pace of asset development in plays like the Permian present an increasing need for an efficient and robust assisted history matching methodology that produces better insights for asset development planning decisions, e.g. well spacing.
A multi-scenario approach is presented to build an ensemble of history matched models that take into account existing uncertainty in reservoir description and well completions. We discuss parametrization of key uncertainties in the reservoir rock, fluid properties, fracture geometry and the effective permeability of stimulated rock. Ensemble-based assisted history matching algorithms are utilized to reduce and characterize the uncertainties in the model parameters by honoring various types of data including field dynamic data and measurements. We discuss the implementation of automated schemes for weighting of various types of data in the ensemble-based history matching algorithms. These schemes are introduced to define the history matching objective functions from various types of data including bottomhole pressure data, and the oil, water and gas productions rates. The computational results show that our adaptive scheme obtains better history match solutions.
The presented multi-scenario approach, coupled with the ability to efficiently run a high number of scenarios, enables better understanding of reservoir and fracture properties and shortens the learning curve for new development in unconventional assets. The shown case study illustrates a comprehensive analysis, using thousands of simulation cases, to obtain multiple history match solutions. Given the non-uniqueness of reservoir history matched models presented in the scenarios, this workflow improves forecasting ability and enables robust business decision makings under uncertainty.
Manchanda, Ripudaman (The University of Texas at Austin) | Zheng, Shuang (The University of Texas at Austin) | Gala, Deepen (ExxonMobil Upstream Research Company) | Sharma, Mukul (The University of Texas at Austin)
Horizontal well fracturing is an established practice to improve the recovery of hydrocarbons from oil and gas reservoirs. To simulate fracture propagation, fracture closure during production and fracture reopening during fluid re-injection, it is essential to combine three important aspects of the problem: multiphase flow, geomechanics and fracture propagation. Current simulation software utilize separate models for these processes. Our objective in this paper is to present a streamlined workflow that we have developed to integrate these highly coupled processes into a single computationally efficient simulation model.
A fully coupled 3-D geomechanical reservoir simulator has been developed to perform multi-cluster hydraulic fracturing and reservoir simulations. The model (Multi-Frac-Res) uses coupled fluid and proppant transport in the fracture with multi-phase reservoir flow and reservoir stresses, in one system of equations. It also accurately models fluid and proppant distribution between multiple perforation clusters in the wellbore. Fracture closure during shut-in or production requires the use of implicit contact models and these models account for the impact of proppant embedment on fracture conductivity. The coupled system allows for seamless transition between fracture propagation, fracture closure, reservoir production and re-injection. This is done in one streamlined workflow without the need for inefficient transfer of information between different simulation software.
An effective hydraulic fracturing treatment aims at maximizing the EUR while maintaining high hydrocarbon production rates. The integrated model allows us to directly evaluate the impact of cluster spacing, frac fluid injection rate, proppant volume, and drawdown on the effectiveness of a hydraulic fracturing treatment. Simulation results are presented that show the relative importance of all the above parameters during the lifecycle of a typical horizontal well. We show how smaller cluster spacing can cause more interference between fractures and hamper the EUR. Larger proppant volume is shown to improve the conductivity of the created fractures and improve the productivity. Faster drawdown is shown to cause faster depletion and faster closure of the fracture but also helps in producing more fluid. Changes in the stress field around the fracture are presented and are shown to impact the growth of fractures in in-fill wells as well as the performance of refracturing treatments. These poroelastic effects are also shown to play a very important role in the growth and reorientation of fractures in injection wells during waterflooding.
Current simulation software utilize separate models for these processes leading to inefficient data transfer between several models that can cause loss of data. This study showcases an integrated model that can simulate the lifecycle of hydraulically fractured wells all the way from creation of the hydraulic fractures to production and reinjection and allows for a holistic comparison between scenarios by comparing productivity numbers and EUR estimates.
Ma, Xiang (ExxonMobil Upstream Research Company) | Zhou, Fuping (ExxonMobil Upstream Research Company) | Ortega Andrade, Jose Alberto (ExxonMobil Upstream Research Company) | Gosavi, Shekhar (ExxonMobil Upstream Research Company) | Burch, Damian (ExxonMobil Upstream Research Company)
Hydraulic fracturing has revolutionized shale oil/gas production in the last decades, but further robust understanding is needed as to what happens during fracture placement. At the end of the pumping process, the sudden change in pump rate leads to a pressure fluctuation in the wellbore referred to as water hammer. These pressure pulses could contain useful information about the generated fractures and their geometry. Our study focused on a detailed look of water hammer signatures on the time-domain as a diagnostic tool for hydraulic fracture geometry. Our water hammer model extends on current industry known formulations. The model solves the fluid transient equation and treats the hydraulic fracture as the boundary condition. We propose a new way to derive the fracture boundary condition based on an improved description of the physics downhole. The model includes pressure-dependent leak-off and perforation friction which are key to determine fracture growth and near wellbore tortuosity. The boundary condition is derived through the fracture entry friction equation instead of previous attempts to use an electrical-circuit analogous system. By changing the input fracture dimensions of the model, we are able to simulate different pressure fluctuations at the wellhead. A novel workflow was also developed to link fracture dimensions to the observed field data. It utilizes an iterative ensemble smoother algorithm to solve the resulting optimization problem. The results and trends obtained with the new water hammer model were validated using well data from the Permian basin. The model validation effort showed and characterized the non-uniqueness of simulation outcomes. We considered several cases of input parameters in order to probe the parameter space of water hammer interpretations. The range of fracture geometry predictions for a particular stage was shown to be broad despite reasonable matches of the water hammer waveform. It proved challenging to find patterns for refining input ranges without converging to widely different fracture geometries. The simulated pressure losses associated the fracture geometry prediction were not consistent in some instances with engineering understanding. A potential source for the characterized non-uniqueness of water hammer simulation outcomes is the inability of time-domain methodologies to generate an appropriate number of physical relations to resolve the physically meaningful variables of interest (fracture length/height/width). In addition, the use of low-frequency water hammer waveforms may not embody the information necessary to resolve hydraulic fracture features. Based on our results and observations, the non-uniqueness of the solution space does not allow us to effectively use the time-domain water hammer interpretation as a diagnostic tool for hydraulic fracture geometry.
Chhatre, Shreerang (ExxonMobil Upstream Research Company) | Chen, Amy (ExxonMobil Upstream Research Company) | Al-Rukabi, Muhammed (ExxonMobil Upstream Research Company) | Berry, Daniel (ExxonMobil Upstream Research Company) | Longoria, Robert (ExxonMobil Upstream Research Company) | Guice, Kyle (ExxonMobil Upstream Research Company) | Maloney, Daniel (ExxonMobil Upstream Research Company)
Relative permeability is a significant source of uncertainty in current modeling practices for performance prediction of unconventional reservoirs. Due to the lack of reliable measurements or representative analogs, relative permeability is often used as an unconstrained history matching parameter for tight/shale rock formations. To date, reliable laboratory measurements of gas-oil relative permeability have been limited to rocks with permeability on the order of hundreds of microDarcies or greater. This work describes laboratory measurements on rock with permeability of hundreds of nanoDarcies, and the use of that data to reduce uncertainty in modeling and performance prediction.
Laboratory measurements of full gas-oil relative permeability curves were made on an unconventional core sample from a tight oil producing interval from the Permian Basin with permeability of hundreds of nanoDarcies. These difficult measurements were achieved through novel experiment design, equipment, and technique. In addition, these measurements were made using a combination of steady-state and unsteady-state techniques that resulted in direct measurement of the relative permeability curves over a broad range of saturations.
The measured steady-state gas-oil relative permeability curves were used to constrain Corey exponents and endpoint saturation values for gas-oil relative permeability curves in history matching simulation models and reducing uncertainty in performance predictions for tight/shale formations. Examples will be discussed.
This work describes the first known successful laboratory measurement of full gas-oil relative permeability curves on rocks with permeability on the order of hundreds of nanoDarcies (~1,000 times tighter than previous measurements). Measured laboratory data assists in constraining parameters used for history matching simulation models and significantly reduces the uncertainty in performance predictions.
Oil and gas production from tight/shale formations has increased from a small value a decade ago to 59% of total U.S. crude oil production in 2018.  Hydrocarbon production from such tight rocks has been commercially viable due to large improvements in horizontal drilling and hydraulic fracturing technologies. Characterization of the tight/shale rock matrix, however, remains an open challenge given the extremely low permeability of the rock matrix and relatively small production history. Intervals like the Spraberry, Bone Spring, and Wolfcamp in the Permian Basin region in West Texas and Southeast New Mexico account for a dominant share (~ 4.1 million barrels/day oil and ~14 Bcf/day associated gas), based on a recent EIA estimate of production from tight/shale rocks. 
We present a novel framework for generating reduced-order models that combines agglomeration of cells from existing high-fidelity reservoir models and flow-based upscaling. The framework employs a hierarchical grid-coarsening strategy that enables accurate preservation of geological structures from the underlying model. One can also use flow information to distinguish regions of high or low flow, and use this division, or other geological or user-defined quantities, to select and adapt the model resolution differently throughout the reservoir. Altogether, the framework provides a wide variety of coarsening strategies that allow the user to adapt the reduced model to important geology and explore and identify the features that most impact flow patterns and well communication. By preserving these features, while aggressively coarsening others, the user can develop reduced models that closely match an underlying high-fidelity model. Various types of simple flow diagnostics based on time-of-flight and volumetric well communication are used to predict the accuracy of the resulting reduced models.
In this paper, we systematically apply this framework to the Great White Field, but also present results from other real or synthetic models, to demonstrate the asymptotic scaling of accuracy metrics with coarsening levels. Our aim is to identify and illustrate best practices when designing and improving coarsening strategies that can guide future applications of the framework to other reservoir models. We also discuss practical limitations when applying the framework to new simulation models where flow regimes or geologic features may differ.
Recent mooring research indicates that fiber ropes with higher strength and higher stiffness would benefit floating offshore platforms in water depths beyond 2000 meters in terms of reduced offset and reduced weight in comparison with polyester rope mooring. More advanced fibers with high strength and high stiffness are also entering into market. The industry has used high strength and high stiffness ropes for temporary moorings and mobile offshore drilling unit moorings. However, high strength and high stiffness fiber ropes have not yet been used for permanent moorings. This paper summarizes studies conducted by the industry on the high strength and high stiffness fiber ropes. An overview is provided for the existing research results, testing conducted, application guidelines, rope qualification processes, project experience, lessons learnt and the challenges of using high strength and high stiffness ropes for permanent moorings. Based on the industry experience of using polyester rope for permanent mooring and knowledge gained on high strength and high stiffness rope, this paper provides recommended assessments that could facilitate the application of high strength and high stiffness ropes for permanent deepwater moorings.
Morrow, Timothy (ADNOC-Offshore) | Al-Daghar, Tariq (ADNOC-Offshore) | Troshko, Andrey (ExxonMobil Upstream Research Company) | Schell, Caroline (University of Tulsa) | Keller, Michael (University of Tulsa) | Shirazi, Siamack (University of Tulsa) | Roberts, Kenneth (University of Tulsa)
The long-term development plan for a giant oil field offshore Abu Dhabi calls for new extended reach wells drilled from artificial islands. The existing wells in this field have historically suffered from inorganic sulfate-based scale deposition in the production tubing which is mitigated by periodic scale inhibition squeeze treatments. The new extended reach wells will have more sophisticated lower completions, including limited-entry liners (LELs) and inflow control devices (ICDs) with external debris barriers. It is currently planned to mitigate inorganic scale in these wells with periodic coiled tubing or bullhead scale inhibition squeeze treatments, which are anticipated to be more challenging and costly due to the extended reach. It is unknown as to whether these types of completion equipment are susceptible to scale deposition or how much scale deposition can be tolerated before well productivity is impacted. Knowledge of the rate of scale buildup on ICDs and LELs versus the volume of water produced through the devices is an important factor for choosing the optimum frequency for scale inhibition squeeze treatments to mitigate scale in these completions while keeping operational costs down. A two-phase laboratory study is currently underway to assess the susceptibility of ICDs to scale deposition. The first phase of the study will focus on the potential for strontium sulfate scale deposition on the debris barrier upstream of the ICD. This paper reports the experimental design and results of laboratory scale deposition experiments on a series of debris barrier test coupons with the goal of estimating the rate of scale buildup on the full-size ICD debris barriers, and the volume of scaling brine that can be produced through the ICD debris barrier (in the absence of any scale inhibitor chemical) without risking significant plugging.
Xiao, Feng (ExxonMobil Upstream Research Company) | Long, Ted A. (ExxonMobil Upstream Research Company) | Velamur Asokan, Badri (ExxonMobil Upstream Research Company) | El-Bakry, Amr (ExxonMobil Upstream Research Company) | Dani, Neeraj R. (ExxonMobil Upstream Research Company) | Holub, Curtis (ExxonMobil Upstream Research Company) | Poole, Stephanie (ExxonMobil Upstream Research Company)
Examples will be presented to show the successful implementation of quantitative slug scanning, gas lift multi-pointing diagnostics, and production surveillance for unstable flows. The slug scanning module supplements the pattern recognition and machine learning features in GLOWTM [
The developed system combines data-driven and physics-based approaches for online advisories, and utilizes a novel history matching procedure to enhance the model's predictive capability for complex well behaviors. In summary, the system provides a comprehensive, robust, and integrated set of tools that meets the surveillance and optimization needs of asset staff.
Chen, Peng (ADNOC Offshore) | Willingham, Thomas (ADNOC Offshore) | Al Sowaidi, Alunood (ADNOC Offshore) | Stojkovic, Dragan (ExxonMobil Upstream Research Company) | Brown, James (ExxonMobil Upstream Research Company)
In the oil industry, oil and gas are usually accompanied with water when they are produced from the subsurface. How to tackle water is one of the major concerns for the field development, especially as fields mature and water production increases. Produced water reinjection (PWRI) has been considered an environmentally friendly way to handle large amounts of waste fluid, though it needs to be carefully designed. In this paper we present a lab study conducted to determine the water specification requirements for reinjecting produced water back into the subject carbonate reservoirs.
The objective of this study is to assess the required produced water quality to maintain matrix injection into the targeted reservoirs. The assessment includes (1) evaluation of the inorganic scaling potential of water sources (fluid compatibility), (2) core flood tests to quantify the impact of various oil content concentrations of produced water on reservoir performance, and (3) a solids loading core flood test to evaluate the injectivity impact of different filtration sizes and different suspended solid concentrations in the produced water. While the previously published paper (
Produced water (PW) collected from the field was utilized in all stages of this study. Analysis of the composition of the suspended solids in the collected produced water revealed a large amount of iron in the PW’s suspended solids, most likely a corrosion product from the long-distance pipeline between the subject field and the current water treatment and separation facilities. Consequently, the collected produced water’s particle size distribution is inadequate to represent the future reinjected produced water which will come from artificial island wells without going through the pipeline. To replicate the anticipated particle size distribution, filtered produced water was mixed with synthetic solid micro particles according to the particle size distribution measured at the well head and the solids loading specification from the skimmer design to mimic the ‘outlet water’ from the skimmer. The skimmer ‘outlet water’ was then filtered to different sizes, starting with 2μm and relaxing the filtration requirements with each step. To replicate oil carryover, 300 ppm of the field’s oil was added to the sequential filtration stages of the skimmer ‘outlet water’ and was flowed through a preserved core plug of the field’s dominant rock type.
Coreflood results suggest that for particle concentrations which represent the solids loading coming from the designed skimmer (TSS=33mg/L), a surface/external filter cake may form with no significant particle penetration into the rock matrix when filtration size is larger than 2µm. More specifically, particles smaller than 2µm did not contribute to the permeability decline, and most of the permeability decline was caused by a filter cake composed of particles in the 5-10µm range. Particles larger than 10µm do not have a significant effect on the permeability decline, most likely due to their low concentration.