Learning from examples to improve code completion systems

In many of today's completions, sliding sleeves have become the preferred device for allowing zone selectivity, flow shut-off, commingling of zonal production, and throttling of production. Sliding sleeves have evolved from very simple devices to an assortment of designs that meet specific applications. This paper traces the development and early field applications of intelligent completions centered on new hydraulically-actuated, sliding-sleeve technology. Today, because of the desire to reduce interventions and their associated risk and costs, enhanced equipment is needed to meet the more specialized applications. This paper shows how hydraulically actuated sliding sleeves, with the full-bore access they permit, provide an economic and reliable alternative to more complex integrated Intelligent Completion systems when open/close functionality is desired. A typical system, its functions, mechanical contingencies and plumbing challenges are described along with various solutions to reservoir management made possible by these systems. Performance testing and installation case histories confirm the ability of the system to accurately interpret reservoir performance, and to monitor and control production to shorten detection and response times in the face of changing downhole conditions. Introduction Many wellbores penetrate more than one reservoir. By commingling the flow, a more economic model can be achieved. Within these wellbores, for zonal isolation, zonal selectivity, flow shut-off, commingling production, and transient testing, sliding sleeves have become the preferred flow control devices. A sliding sleeve is a device that provides a means of opening or closing a flow path, and is typically placed in the tubing string. Whether commingling flow in production tubing or separating laterals in a multilateral well, the generally accepted method for mechanically isolating zones is with packers. Therefore, to take advantage of a remotely operated sliding sleeve and a packer's isolating capabilities, that will allow separate or commingled production from many zones, based on real time data, new types of completion designs and techniques have been developed. Hydraulic Based Intelligent Completion System Traditional use of conventional sliding sleeves for zone isolation, zone selectivity and flow shut-off, transient testing and commingling production requires intervention to mechanically shift the sleeve insert. The industry has for some time now been looking forward to a variety of intelligent completion systems that would allow the remote actuation of flow control devices based on real time data from downhole monitoring. This paper specifically addresses intelligent completion systems based on hydraulically-actuated flow control devices. By combining real-time data monitoring with a surface controlled hydraulic sliding sleeve, an operator will be able to reduce or eliminate interventions, reduce production down time, reduce well costs, and reduce the risks associated with well intervention. This is especially the case in high angle, extended reach and multilateral wells. In addition, a surface controlled, hydraulic sliding sleeve will allow an operator access to zones below an Electric Submersible Pump. A hydraulic based intelligent completion system primarily consists of hydraulic sliding sleeves and feedthrough isolation packers. The hydraulic sliding sleeve is remotely operated from the surface via dedicated hydraulic control lines.

Advances in drilling and well construction technologies have enabled extended-reach horizontal and multilateral wells to increase reservoir contact and improve recovery. Completion systems with passive inflow control devices (ICD) or intelligent wells are used to optimize and control inflow from the compartments or laterals. ICD completions are passive systems with torturous paths or orifices designed to choke back inflow from dominant zones and equalize influx along the wellbore. ICDs combined with openhole packers provide very fine compartmentalization of the open hole and it is common to have completions with 40 inflow compartments in some extended-reach wells. ICD completions require well intervention to close zones with high water or gas. Intelligent completions, on the other hand, have downhole valves that can be remotely operated from surface. The requirement for downhole control lines and penetrations at the wellhead and packers has limited the number of downhole valves to a maximum of six in a well. This limitation excludes a significant number of applications where greater compartmentalization can yield superior well performance. The industry has recently developed innovative completion systems that increase the number of surface-controlled downhole valves while minimizing the number of downhole lines. These technologies improve completion design while overcoming limitations like the number of control lines, feedthrough ports on packers and wellhead. Saudi Aramco has recently installed an all-electric intelligent completion with 8 downhole active flow-control devices (AFCD) controlled via a single downhole control line. This installation is currently the world record for the highest number of inflow control valves installed for an intelligent completion. This paper presents details of technology development, evaluation, completion design, installation and in-well testing of the all-electric intelligent completion system in a multilateral well. The paper presents the novel completion design with redundancies, system components, advantages of the chosen technologies and the completion method. The section about the planning phase of installation highlights steps taken to ensure successful execution of the job while minimizing operational risks associated with the deployment of the new technology. Finally, the execution phase of the installation is presented with special emphasis on procedures and actions during installation to improve operational efficiency and minimize health and safety (HSE) risks. Thorough design, planning and execution of the job yielded the successful installation of an advanced and complex intelligent completion system in an efficient and safe manner. All-electric intelligent completion systems capable of controlling multiple downhole valves are anticipated to increase well life and recovery through improved downhole monitoring and control. This system can control as many as 27 AFCDs on a single electric control line.

This paper describes the design, testing, installation, and performance of the first ‘fully-completed’ well using an intelligent inner completion inside an un-cemented liner with openhole packers for zonal isolation. The well design concept evolved from technical challenges associated with completing long cased and cemented laterals in the mature Ekofisk waterflood. The term ‘fully-completed’ implies full reservoir access along the pay length for production and high rate matrix acid stimulation using limited entry for fluid diversion within well segments. The paper covers the development and qualification of custom openhole 7⅝ in. liner components that can handle high differential pressures and extreme temperature fluctuations, the marriage of this complex liner with a five zone intelligent completion system, and results from a year of system integration testing. The paper also discusses the strategic placement of both mechanical openhole and inner string packers based on caliper and drilling logs; challenges met and overcome during installation; and a remarkable collection of down-hole gauge data that confirms the performance of each component before, during, and after the stimulation. The Ekofisk field waterflood began in 1987 and continues to date, exceeding expectations for improved oil recovery while mitigating reservoir compaction. As the waterflood matures, new wells are more often found partially water-swept. Limited infrastructure for lifting and handling the high water production has led to increased emphasis on isolating these water-swept intervals. Cased, cemented and perforated completions have traditionally been used for this service. It has become increasingly difficult to execute a successful cement job in longer horizontals with 4,000 ft to 8,000 ft laterals where rotation of the liner is impossible and high effective circulating densities (ECDs) limit rates during cementing. Wide variations in reservoir pore pressures, often in excess of 2,000 psi difference along the lateral, are typical of the Ekofisk chalk and exacerbate the difficulties of cementing. As a result, a new method for zonal isolation has been developed to ensure the success of future infill drilling campaigns. The design and careful planning that went into the fully-completed openhole un-cemented liner strategy resulted in a successful field trial and has proven this solution to be an effective alternative to cemented reservoir liners in long horizontals where zonal isolation is critical. Use of the intelligent well system (IWS) allowed offline acid stimulation without rig, coiled-tubing, or wireline intervention. What would have traditionally been a significant water producer, with three water-swept zones totaling nearly 2,000 ft across a 4,000 ft reservoir section, has turned out to be one of the best oil producers in the field with nearly zero water cut. Production results show high productivity with highly negative acidized completion skins. With large investments in intelligent completions to provide zone-specific inflow control and water shut-off, isolation outside the liner becomes much more important. Over recent years, the Ekofisk wells have illustrated the difficulty of achieving effective cement along lengthy reservoir targets. The openhole fully-completed solution combining an accessorized un-cemented liner with an inner intelligent completion string will allow operators to push the limits in terms of lateral length while maintaining full control over producing and non-producing zones.

Description of the Proposed Paper Describe the background and history of a producing well leading up to a workover decision and installation of a multi-zone intelligent well completion system. State the guidelines used to design and complete an offshore horizontal five zone well with intelligent completion equipment. State the success and/or problems seen during the installation. Highlight the lessons learned during design and installation and show benefits of such a completion during the production life of the well. Application: Offshore completions in development wells targeting multiple stratigraphic units located over highly compartmentalized fault blocks with distinctly separate fluid contacts and pressure regimes. Control of water breakthrough and GOR management on multi-zone oil producers. Safe completion of complex horizontal wells. Results, Observations, and Conclusions: Installation was done successfully and safely. The defined completion successfully worked as designed. Significance of Subject Matter: The use of multi-zone intelligent completion equipment to optimize the reservoirs is critical to the economic success of this type of field. The technology used provided the capability to produce each zone alone or commingle production without subsequent well intervention to shut-off zones after water breakthrough or high GOR production. These capabilities will be shown to improve the overall reservoir recovery. Introduction The oil and gas industry is the primary contributor to Brunei's economy. The exploration and production of Brunei's hydrocarbon reserves is undertaken with great care to extend the income from this natural asset for future generations. To this effect, Brunei Shell Petroleum (BSP) is an early adopter of new technology in this region, to solve technical challenges and improve economics and reserve recovery. Examples of this include the use of innovative directional drilling technology1, the integration of intelligent completion equipment with expandable sand control technology2 and use of solid expandable tubulars3. A high level of local experience in Brunei contributes to the development of the Iron Duke field and neighbouring field such as Ampa and Champion. The Iron Duke field is located offshore, 50 km from the Brunei coast. Iron Duke came online in 1992 with reserves estimated at 17 million m3 of oil (105 MMbbl) and 27 billion m3 of gas (950 bcf) in the Upper Miocene shallow marine sandstones. Current production from 7 producers is 3000 m3/day liquids and 2.5 million m3/day gas. In the Iron Duke area, 25 wells have been drilled during a major development drilling campaign. The Iron Duke structure is an anticlinal dome dissected by a large number of collapse faults. A long strike of the collapse fault structure, the phenomena of "fault flipping" is observed. The faults are often sealing and define a large number of blocks within the field. Vertical sealing capacity may vary due to the sand shale sequence. The drive mechanism in Iron Duke is predominately gas cap expansion with limited aquifer support; thus wells are targeted close to the oil water contact (OWC) to delay gas breakthrough. Well ID-19 is one of several development wells targeting five main strategraphic units located over several fault blocks as shown in Figure 1.

Recent empirical studies that compared static and dynamic type systems on API usability showed a positive impact of static type systems on developer productivity in most cases. Nevertheless, it is unclear how large this effect is in comparison to other factors. One obvious factor in programming is tooling: It is commonly accepted that modern IDEs have a large positive impact on developers, although it is not clear which parts of modern IDEs are responsible for that. One possible---and for most developers obvious candidate---is code completion. This paper describes a 2x2 randomized trial that compares JavaScript and Microsoft's statically typed alternative TypeScript with and without code completion in MS Visual Studio. While the experiment shows (in correspondence to previous experiments) a large positive effect of the statically typed language TypeScript, the code completion effect is not only marginal, but also just approaching statistical significance. This seems to be an indicator that the effect of static type systems is larger than often assumed, at least in comparison to code completion.

Intelligent completion injectivity refers to controlling and monitoring fluids injected into reservoirs to enhance oil and gas production at nearby producers. Traditional methods often lack the precision and adaptability required to optimize multi zonal injectivity effectively, particularly in the activation of flow control valves (FCV) to optimize production or injection. In conventional operations, these valves are activated manually on-site by a delayed chain of instruction decided on in town and communicated to an on-site technician. This must be done physically on unmanned platforms, thus adding operational complexity and high operating cost. The intelligent completion system deployed in this work leveraged advanced sensors and actuators to enable real-time adjustments, allowing for a more nuanced and efficient injection process. To enable fully compliant remotely controlled intelligent completions with smart edge computation, the interfaces were managed from design to execution to achieve a fit-for-purpose solution. The system integration test was essential to verify smooth execution with improved injectivity interface panel (i.e., onsite GUI) of zonal rate measurement visualization. Downhole equipment comprised permanent downhole gauges (PDG) and FCV. Surface equipment such as a hydraulic pump system with engineered logics of rate calculation were connected to the downhole flow control valves measurements. Meanwhile surface acquisition units (SAU) were connected to PDG systems. Data from PDG SAU and the hydraulic control system are transmitted from offshore to office via satellite communication or internet. Smart algorithm embedded software was engineered to enable real-time data monitoring and control remotely from offshore. With the automated system, the end users can confidently actuate the FCV on the offshore platform from a device located on offshore or office. The process from receiving information to making a decision to take action, which was previously done in months, can now be done in minutes. This allows water/gas injection optimization by maximizing areal and vertical sweep, enables remote injection control, and provides on-site gas and water single-phase injection rate computation. This means the monitoring rate profiling is now directly provided from site acquisition capability, in contrast to the conventional method of post-processing done in town to first analyze the data and then generate the profile. The intelligent system consists of rate computation algorithms to help the user estimate the water/gas injection rate and back allocation when flow is commingled from multiple zones through FCV. The algorithm uses real-time data that are acquired through the PDGs (i.e., annulus and tubing gauges), with the pressure drop across the FCV, FCV position information, and reservoir properties input into the system. Subsequently, the estimated flow rate is used to calculate the injectivity index, which is applied in optimization to overcome challenges such as uneven zonal split injection and unfavorable zonal water-front propagation due to over- or under injection. Intelligent completions are chosen for wells known to have potential for injection or production optimization from multiple zones. Operational complexity, cost, and resource challenges are the bottlenecks to operations that realize the wells’ full potential. Embedding engineered real-time data monitoring and control has proven to simplify operations, reduce OPEX, and minimize HSE risk. The adoption of these technologies has enabled acceleration in digital transformation program to deliver well and resource productivity, efficiency safety, and performance.

Extended reach wells with thousands of feet of open hole reservoir contact and multilateral wells with 8-10 laterals are becoming common practice as the industry develops tight and difficult reservoirs. Over the years, drilling and well construction technologies have made significant advances in the field of geosteering, multilateral junctions, etc. to make these complicated wells a reality and commonplace. The industry has recently introduced several new well completion technologies for the downhole monitoring and control inflow of these extended reach and multilateral wells. Over the last 10 years Saudi Aramco has drilled and completed hundreds of multilateral wells with intelligent completions for real-time downhole monitoring and remote valve operation to control inflow from each laterals. These technologies have improved well performance and reduced well intervention. The paper presents the case study for the design, planning, installation, and operation of Saudi Aramco's first intelligent completion that is operated with combined electro-hydraulic control system. New technologies like these are expected to improve the operation and inflow from the new generation multi-compartment wells. Conventional intelligent completion systems uses downhole valves that are operated by hydraulic pressure and the pressure is supplied from surface through dedicated hydraulic lines to each of the downhole valves. The wellhead and operations equipment has a maximum limit of eight downhole lines and this limits the number of downhole valves in a completion to five. The system in this case study uses a electro hydraulic control module at each of the downhole valves and can control of up to 12 intelligent completion tools with the use of only two hydraulic lines and one electrical line from surface. Technologies like these allow inflow control from each of the laterals and enhance the performance of the completion.

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 209965, “System Engineering Recommendations To Improve Intelligent Completions,” by Ron Nelson, SPE, Consultant; Eric Handley, MacDermid Offshore Solutions; and Jonathon Joubran, SPE, Halliburton, et al. The paper has not been peer reviewed. _ The complete paper compiles recommendations from a broad range of sources into a single document to aid future intelligent-completion installations and operations. While emphasis is placed on subsea completions, these recommendations can be applied readily to differing well types. Important differences between production wells and injection wells also are addressed. Introduction Intelligent-completion interfaces span drilling, completions, facilities, production operations, and subsea-systems-engineering disciplines. Affected system components include the tree and tubing hanger, control system, control fluid, control lines, well fluids, well casing, downhole sensors, feed-through packers, and interval control valves (ICVs). Example diagrams are provided in Appendix B of the complete paper. While a successful intelligent completion can enhance a well’s value significantly, a failed intelligent completion can severely diminish that value, possibly rendering the well inoperable. Proposed Intelligent-Well-Design Process Control-Fluid Selection. The design process should start with the control fluid, which provides hydraulic force to open and close the ICV. It traverses the entirety of the system and will affect most system components. A successful control fluid will accomplish the following goals, which are further detailed in the complete paper: - Provide adequate lubricity - Maintain required properties across all design temperatures - Be compatible with all materials - Be compatible with completion brines - Provide adequate hydrate inhibition - Minimize environmental impact If the intelligent completion is to be installed in an existing development already using a fluid that cannot be used for the ICV, it may be possible to fill or “stack” the ICV lines with another control fluid that meets the downhole requirements and also is fully compatible with the in-service control fluid. The fluid’s original equipment manufacturer (OEM) should be consulted when a mixed-fluid system is considered. Assessing Thermally Induced Pressure Gains. Once a preferred control fluid has been identified, the next recommended step is to assess worst-case thermally induced pressure changes in the control lines. Thermal effects directly affect numerous intelligent-completion components. API SPEC 17D recommends keeping the tree-mounted ICV control-line isolation valves [tree-isolation valves (TIVs)] closed when not shifting the ICVs. At the other end of the control line, the ICV’s seals isolate the control fluids from the well fluids. Each downhole control line is a closed system when the TIVs are closed. A closed control line will behave like a casing annulus in that internal pressure will rise when warmed and fall when cooled. The pressure changes in the lines, however, can be far greater than those seen in the casing annuli and can affect the intelligent completion system’s pressure requirements profoundly.

Code completion is an important feature in Integrated Development Environments (IDEs). These years, researchers have been making efforts for intelligent code completion. However, existing work on intelligent code completion either only considered production code, or did not distinguish between production code and test code. It is unclear how effective existing completion models are for test code completion, nor whether we can further improve it. In this work, we focus on the completion of test code. We first find through experiments that completion models for production code are suboptimal for test code completion. Then we analyze the specific characteristics of test code, and observe that test code has inter- and intra-project similarities, and a strong relationship with its focal class and other production classes depending on the focal class (i.e., focal-related code). By incorporating test code from other projects to fine-tune existing models, we leverage the inter-project similarity of test code to improve the completion of tokens specific to test code. By introducing a local component and constructing existing test code as well as the focal-related code in the project as references, we enhance existing code completion models with the intra-project similarity and the focal-related code of test code. Experiments show that each characteristic of test code we exploit can bring substantial improvement to test code completion and our integrated framework outperforms other baseline frameworks. Compared to the base completion model, on token-level completion, our optimal model for test code completion relatively improves all-token and identifier completion accuracy by 7.68% and 19.96%, respectively; on line-level completion, it relatively improves edit-distance similarity and exact-match metrics by 8.89% and 22.82%, respectively. Moreover, we perform error analysis and point out potential directions for future work.

This article is a synopsis of paper SPE 50587, "Intelligent Completion Systems - The Reservoir Rationale," by Y. Jalali, SPE, Schlumberger Wireline & Testing; T. Bussear, SPE, Baker Oil Tools; and S. Sharma, SPE, Schlumberger Wireline & Testing, prepared for presentation at the 1998 SPE European Petroleum Conference, The Hague, The Netherlands, 20-22 October.

Code completion has become an essential component of integrated development environments. Contemporary code completion methods rely on the abstract syntax tree (AST) to generate syntactically correct code. However, they cannot fully capture the sequential and repetitive patterns of writing code and the structural information of the AST. To alleviate these problems, we propose a new code completion approach named CCAG, which models the flattened sequence of a partial AST as an AST graph. CCAG uses our proposed AST Graph Attention Block to capture different dependencies in the AST graph for representation learning in code completion. The sub-tasks of code completion are optimized via multi-task learning in CCAG, and the task balance is automatically achieved using uncertainty without the need to tune task weights. The experimental results show that CCAG has superior performance than state-of-the-art approaches and it is able to provide intelligent code completion.

Intelligent code completion has become an area of interest to accelerate software development. In this study, novel methods for code completion are proposed and the performance of these methods on code completion is demonstrated. Deep Learning models are applied over large datasets to achieve code completion in dynamically written programming languages. In this study, an LSTM model that performs next word prediction is trained on the JavaScript programming language. The code completion system consists of two phases. In the first phase, multiple completion suggestions are generated from the LSTM model. In the second phase, these suggestions are checked by the JavaScript Parser and the best results are achieved. This system is an alternative solution where the model can be trained without preprocessing the training set with a Parser. The proposed method has been tested on the test dataset and the success of the code completion is shown in a quantitative way.

This paper describes the challenges faced on the deployment of intelligent well completion (IWC) systems in some of the wells built in Buzios field, mostly related to heavy fluid losses that occurred during the well construction. It also presents the solutions used to overcome them. This kind of event affects not only drilling and casing cementing operations, but may also prevent a safe and efficient installation of the completion system as initially designed. The IWC design typically used in Brazilian pre-salt areas comprises cased hole wells. Perforation operations must be performed before installing the integral completion system, as it does not include a separation between upper and lower completion. Therefore, the reservoir remains communicated to the wellbore during the whole completion installation process, frequently requiring prior fluid loss control as to allow safe deployment. Rock characteristics found in this field make it difficult to effectively control losses in some of the wells, requiring the use of different well construction practices that led to the development of some new well designs. The well engineering team developed a new well concept, where a separated lower completion system is installed in open hole, delivering temporary reservoir isolation. This new well architecture not only delivers reduced drilling and completion duration and costs, but also provides the IWC features in wells with major fluid losses. This is possible by the use of multiple managed pressure drilling (MPD) techniques when required, which were considered since the initial design phase. Safe and effective construction of some wells in pre-salt fields was considered not feasible before the adoption of MPD solutions, both for drilling and completions. Other important aspects considered on the new well design are the large thickness and high productivity of Buzios field reservoirs, as well as the need of some flexibility to deal with uncertainties. Finally, the new completion project was also designed to improve performance and safety on future challenging heavy workover interventions. The well construction area has gradually obtained improved performance in Buzios field with the adoption of the new practices and well design presented in this paper. The new solutions developed for Buzios field have set a new drilling and completion philosophy for pre-salt wells, setting the grounds for future projects. The improved performance is essential to keep these deepwater projects competitive, especially in challenging oil price scenarios. One of the groundbreaking solutions used is the possibility of installing the lower completion using managed pressure drilling techniques.

Technology Focus On occasion, there is a tendency to equate new advances in automation and intelligent systems with an inevitable workforce reduction, to an extent that people can begin to resist their application. However, as we all know from our own oilfield experiences, intelligent systems require well-prepared and creative engineers to deliver a successful and economic outcome and a synergetic relationship enhances the business and opportunity set as a whole. These technology advances offer a newer dimension to those options for qualified engineers to select, develop, and apply effective applications to ever-more-complex issues. As the resources that the industry develops become increasingly challenging, access to a broader range of options and new and developing approaches allows more-efficient and -effective recovery. This month, I have selected a suite of papers that considers intelligent systems, all clearly demonstrating that these approaches are becoming increasingly mainstream in their selection and use. All the papers presented and suggested here demonstrate that the equipment, the techniques, or the systems alone cannot deliver the solutions. Instead, the systematic and innovative ways considered in their applications are responsible. In time, the use of intelligent and automated systems will mature further into just another set of tools that the engineering community uses with increasing frequency to deliver the most-efficient development solutions. When that occurs, the industry preoccupation will have moved on to the next wave of innovations, technologies, and advancements, which no doubt will raise new and specific challenges. Recommended additional reading at OnePetro: www.onepetro.org. OTC 28103 Intelligent Completions Used During Extended Well Test of Exploratory Wells in Brazil by L. Costa, Halliburton, et al. OTC 27748 The Qualification Methodology Behind an Integrated Multizone Frac-Pack and Intelligent Completion for the Lower Tertiary by R. Jannise, Halliburton, et al. SPE 187071 Real-Time Downhole Measurements During Deepwater Frac-Pack Completions in the Gulf of Mexico by A. Hawthorn, XACT Downhole Telemetry, et al.

Developers work in the IDE, but search online resources in the web browser. The separation of the working and search context often cause the ignorance of the working context during online search. Several tools have been proposed to integrate the web browser into the IDE so that developers can search and use online resources directly in the IDE. These tools enable only the shallow integration of the web browser and the IDE. Some tools allow the developer to augment search queries with program entities in the current snapshot of the code. In this paper, we present an in-IDE ambient search agent to bridge the separation of the developer's working context and search context. Our approach considers the developers' working context in the IDE as a time-series stream of programming event observed from the developer's interaction with the IDE over time. It supports the deeper integration of the working context in the entire search process from query formulation, custom search, to search results refinement and representation. We have implemented our ambient search agent and integrate it into the Eclipse IDE. We conducted a user study to evaluate our approach and the tool support. Our evaluation shows that our ambient search agent can better aid developers in searching and using online programming resources while working in the IDE.