Abstract A Canadian oil & gas operator has been setting new benchmarks drilling the vertical and tangent section of Montney horizontal wells in the Placid field of Northern Alberta. Initially, the operator drilled vertical wells to kick off point (KOP) with polycrystalline diamond compacts (PDC) and conventional mud motors. As a result of increasing well density, however, the well plans consistently required a 15° to 30° tangent section. With PDC drilling, toolface and build up rates were problematic and the sliding rate of penetration (ROP) was slow. A Rotary Steerable System (RSS) was introduced, but despite the improved performance, the technology came at a premium cost and the severity of drilling dysfunctions generated an increase in tool failures. With falling oil prices, a more cost effective solution was required. Hybrid bit technology, which combines the cutting mechanism of both fixed cutter and roller-cone bits, has been extensively utilized in Canada to drill build sections, providing outstanding results. They have not, however, been commonly used to drill the vertical (drill-out) and tangent sections. The operator combined a state-of-the-art hybrid bit with a mud motor to drill the interval with an 85% success rate. The combination of the hybrid bit and conventional motor, compared to PDC and RSS, resulted in a 30% cost savings to complete the interval. The present case study outlines how hybrid bit technology development, driven by field data in a continuous improvement cycle, identifies performance opportunities, which have a significant impact on drilling time and cost savings in drill out sections. The overall objective of this current case study is to highlight the results and lessons learned throughout the implementation process.

Abstract Verifying that conventional coiled tubing (CT) has entered the correct lateral of a multilateral well can be a time-consuming and tedious challenge. While passing the kickoff point (KOP), surface pressure sensors may not clearly identify that a hydraulic knuckle joint has entered the lateral. To verify which lateral has been entered, the CT must be run-in-hole (RIH) to tag the unique total depth (TD) of the lateral. It can be difficult to prove that the correct lateral has been entered if the TDs for two laterals do not differ significantly. If the CT tags before or after the expected TD is reached, then the operator must pull out of hole (POOH) and repeat this process. This paper presents an improved method for detecting which lateral has been entered shortly after passing the KOP by using a real-time, fiber optic cable (RTFO) bottomhole assembly (BHA). The RTFO BHA is equipped with sensors for gamma radiation (GR), tool inclination, tool face, casing collar locator (CCL), internal and external pressure, and temperature. These readings enable the operator to use the CCL and gamma detector to correlate to the desired lateral’s pipe joints using an existing CCL/GR log. The tool face reading theoretically predicts which angle the knuckle joint should be indexed. The internal pressure sensor provides clear indication of pressure changes when a hydraulic knuckle joint has entered the lateral and the inclination sensor verifies that the inclination of the BHA matches the inclination of the desired lateral. This paper discusses two dual-lateral water injection wells; the TDs of the laterals in the first well were equal, with no trait that conventional CT would be able to distinguish. In addition, the inclination of each lateral was very close to the other, making the difference of pipe weight, read from the surface, not significantly different enough to verify which lateral had been entered. However, with the RTFO BHA, the correct lateral was easily entered and verified, significantly reducing time, risk, fluids, and CT pipe fatigue, while providing assurance that the stimulation fluid was accurately placed. This paper describes the first time that a flow-through gamma, inclination, tool face sensor module was deployed to accurately enter, identify, and stimulate a casedhole multilateral well without cycling the CT at the KOP, and without relying on tagging TD to confirm that the BHA is inside the desired lateral. The new process proved to be a better, more cost effective, and efficient way to stimulate multilateral wells. This solution can be used by operators to extend the life of their mature fields.

Abstract A key challenge in developing brown fields is identifying a strategy that enables placement of horizontal wells in a field riddled with existing, depleted wells. These wells have drained multiple reservoirs in proximity to current target intervals, resulting in altered in-situ pressures that may impose additional technical and economic drilling risks. This work presents a new technique which optimizes well path design by combining hybrid nature-based metaheuristics with spline curvature to navigate around depleted zones. The proposed method is validated by testing on synthetic and actual well cases. The nature-based metaheuristic method employed is a modified firefly algorithm with hybrid implementations of mutation and annealing. It considers a potential well's starting coordinates, target coordinates, possible obstructions, subsurface stress distribution, an RSS tool's dogleg limitation range, kick off depth limitation, and required length of lateral section to optimize overall wellbore length, all of which can directly be linked to the economics behind drilling a well. The functionality of the designed algorithm is examined with both synthetic data and publically available field data. Further complexity is added in the model by including geomechanical stresses in the model when available. Comparisons of overall wellbore length, wellbore orientation, and wellbore profile energy are also provided for a case in the Wattenberg basin derived from public data. Using sparse information, the algorithm was able to automatically design entire well paths in a relatively short period for all cases and the final solutions resembled industry solutions based on minimum design constraint. The uniqueness of the work is highlighted by the algorithm's ability to converge towards optimal solutions which can help the operator shift work load from well design to more critical tasks.

Abstract This paper describes the procedures and methods used to successfully drill an extended reach horizontal well in Permian Shale Source Rock Reservoir, in the Cooper Basin, South Australia. A previous attempt at a similar well design did not reach the objective. An independent review of this well's performance was used as a starting point to design an improved drilling program for the second well. This included the following: Closer liaison and collaboration was encouraged between the operator and the service providers. Improvements were planned for the second well to eliminate or reduce the severity of hole problems that occurred in the previous well. Written procedures were established to ensure that hole quality was maintained throughout the drilling phase. Technologies that could assist in improving drilling performance were identified, evaluated, and used. As a result of these initiatives, the second well was considerably more successful than the 1. To compare and contrast the results of both wells as an indicator of improvement: –The 1 well achieved only 37% of the objective. The 2 well achieved 170% of the objective. –The 1 well exposed 589m of lateral wellbore. The 2 well exposed 1014m of lateral wellbore –The 1 well required 38 days of drilling. The 2 well was drilled in 11, despite being twice as long –The lateral in 1 well required 7 bits. The lateral in the 2 well was drilled with 1 bit, again despite being twice as long –The 1 well observed 485 hours of NPT. The 2 well was drilled with ZERO NPT. The technologies and improved procedures used in the second well facilitated its success and these initiatives will be refined for future similar wells. The major contributors to difference between these two wells are: Improved drilling fluid including the use of mud lubricant The use of drillstring torque-reducing subs improved bottomhole assembly (BHA) performance and reduced wear on downhole equipment. A new bit designed to match the rotary steerable system (RSS) tool and formation drilled the entire extended lateral and could have achieved even more.

The placement of cement plugs in extended-reach drilling (ERD) wells is a challenging task. Poor hole cleaning, insufficient mud displacement, incomplete centralization, and tight ECD management are some of the more common risks. Mud displacement efficiency and accurate cement placement is notably more complicated in ERD wells due high deviation angles and asymmetrical fluid velocities; problems which are compounded in presence of Synthetic Based Mud (SBM). In fact, the industry standard for setting and successfully testing cement plugs in highly deviated wells is 2.4x per successful attempt. For deepwater projects, these risks must be mitigated to ensure timely operations as the financial implications of failed cement plugs are vast. Adopting best cementing practices in job planning and execution are beneficial elements to overcome the challenges; however even with excellent operational guidelines cement plug failures are common. A properly designed job should consider appropriate selection of equipment and materials, tailored properties of slurry and spacer systems, accurate down-hole pressure and temperature simulation, extensive laboratory testing, and identification and management of key risks. This paper will discuss the various operational, engineering, and unique design techniques for setting horizontal cement plugs in a SBM environment on a project in deep waters offshore Indonesia. To date, four kick off plugs have been set successfully in the first attempt on four consecutive ERD wells. Conservative cost savings for achieving first attempt kick-off is 1.2 MM USD, applying 12 hours per well (48 hours total) with a daily rig rate of 600K USD. The achievement is owing to application of best cementing practices in job design and execution, along with use of a unique cement system and innovative spacer. Techniques used in placing challenging kick off plugs are discussed followed by discussion on how the trouble-free operation has created value for both the client and service provider. Multiple case histories will be presented, along with specifics of the unique slurry and spacer system.

Abstract Field testing was performed on plunger lift systems in horizontal Marcellus shale wells. Plunger lift software from Echometer Company, Total Well Management, or TWM was used on certain wells to monitor acoustic trace, casing pressure, and tubing pressure throughout the plunger cycle. The plunger lift testing program addresses the challenges presented in shifting from vertical to horizontal well plunger lifting. The plunger lift testing addresses the feasibility of running plunger lift on wells with X and XN-profiles. These nipples are used to provide safe snubbing operations. The most valuable insight from the testing involves the possibility of running plunger lift to deviation angles of 70° and greater. The results of the plunger lift testing yielded several conclusions. An important result was that all plungers successfully passed through the X-profile, contrary to industry doubt. Another conclusion was that all plungers fell down to their bottom deviation angle with the deepest falling of 70°. Testing also showed the importance of performing TWM investigations on every plunger lift well to optimize plunger cycles. The most notable result was the phenomenon that plungers fell faster below kickoff point. This phenomenon suggests the possibility of plungers travelling to 90°. The observed trends of 53 TWM traces suggest that current plungers in horizontal Marcellus shale wells can theoretically reach a final deviation angle of 74°. Plunger lift systems have been used across the United States for decades to unload small amounts of liquid (water, oil, and/or condensate) as gas rates fall below the minimum critical rate to continuously unload fluids. The Marcellus shale is a vital asset to natural gas production in the United States. Development has transitioned from a vertical well science program to a proven horizontal well development program. Plunger lift has become common among some operators in the Marcellus shale for wells which have fallen below their critical gas rate. Our plunger lift program shows promise to maintain production after wells reach unstable flow up the tubing. Marcellus shale production can potentially benefit significantly if plungers eventually reach deviation angles of 90° since it will allow for production of fluid left in the horizontal lateral.

Abstract Branches in Level 1 and 2 lateral extensions traditionally have offered limited access. Coiled tubing (CT) intervention was limited to the main hole only, and thus stimulation of main reservoir targets was not possible. A new system designed with a 2 1/8-in. outside diameter (OD) enables rigless intervention through the completion restriction. The system includes a reentry bottomhole assembly (BHA), which consists of a surface-controlled orienting tool and a controllable bent sub (CBS). The Discovery MLT* multilateral tool identifies the window of the selected lateral before attempting reentry, and confirms successful identification. The entry is visible at the surface through a software-displayed pressure log. The technique does not require a wired coiled tubing string; the corrosion-resistant reentry tool is operated solely by flow and is conveyed with standard CT equipment. As proved by the success of the new reentry tool in Upper Zakum field selective matrix stimulation and nitrogen kickoff operations detailed in this paper, the need to bullhead huge acid treatments that result in major openhole washouts has been eliminated. These washouts can cause both poor acid distribution and prevention of future access to the lateral for CT logging operations. This paper discusses the successful deployment of a selective multilateral entry tool that was utilized to access multilateral wells in Upper Zakum field, two case histories, and the method's potential in Upper Zakum field. The two case studies presented demonstrate the success of this system in reentering a desired lateral on the first attempt for matrix stimulation operations. Current developments that enlarge the scope of application to wellbore logging are under study. Introduction Zakum Development Company (ZADCO) is operating three fields offshore Abu Dhabi in the United Arab Emirates. Upper Zakum is the main field with more than 400 wells, mostly dual completions, producing from different reservoirs and 80 offshore platforms. H-I, H-II, and H-III are the main reservoirs in a formation of mudstone / wackstone and dolomite. Multilateral technology implementation in ZADCO is evolving rapidly to provide new methods for reservoir management and increase production potential. No significant rigless well intervention operations had been conducted previously in Upper Zakum field, apart from wireline operations. In the past accessibility into horizontal multilateral wells was limited to the mother hole; accordingly acid stimulation and laterals assessments were difficult to achieve. Therefore a system that offered the means to overcome such a problem was thoroughly investigated. There are limited options available for well intervention with harsh offshore logistics. The solution technique was a rigless well intervention unit, which includes the Discovery MLT tool, a CT unit, pumping equipment, and a neutralization package equipped with zero flaring facilities, placed on a self-propelled barge (See Figs. 1 and 2). This provides relevant services without delay. Furthermore a self-propelled barge dramatically minimized in and out time. The operation was executed in combination with other rigless activities. The success of the multilateral intervention on the two wells case histories will encourage future multilateral intervention in Upper Zakum field toallow access to upper cased lateral and lower barefoot lateral on demand for dry oil production in an optimum layer of upper reservoir achieve effective stimulation that increases production capacity of wells with up to 12 laterals currently existing in Upper Zakum field leave future options for reservoir monitoring and extending the well life by applying selective water shutoff in selected laterals.

Abstract

In prolific gas reservoirs with strong aquifer drive, water coning is often the determining factor in well productivity and ultimately, gas recovery factor. One of the main drivers in optimising the field development in these cases is to reduce the drawdown of the wells and consequently, a horizontal well design is often used. It is the case, however, that in such highly permeable gas fields extending the horizontal length beyond 1000-2000 ft often does not help to improve well productivity any further, as well bore friction becomes the constraining factor. This paper presents a cost effective and simple solution to increase well productivity by 60% compared to single horizontals at a marginal cost increase. A simple level-1 dual lateral with large bore casing-flow well design has shown to be able to deliver the additional productivity at only a 5-10% incremental increase in cost. Furthermore, we present the results of the numerical simulation that helped to justify the well design and to deal with interference between the branches.

Introduction

Horizontal wells have become common place in oil and gas fields to increase well productivity. In the Central Luconia region 250 km offshore Sarawak, Malaysia, horizontal wells are a key element of the field development approach for a large number of gas bearing carbonate build-ups that have prolific reservoir quality (200 to >1000 mD) and a strong aquifer drive. As a consequence of the field characteristics, (fast) water encroachment and water coning play a dominant role in the field behaviour. The horizontal wells aim to reduce the drawdown and thus the effect of coning, whilst at the same time maximise the offtake per well, and the ultimate recovery.

However, in the prolific gas fields, extending the horizontal well length beyond 1000-2000 ft does not reduce drawdown any further as the well bore friction becomes the constraining factor, and the toe of the well does not contribute to the well's production. To further enhance the productivity of our horizontal wells, a cost-effective dual lateral well design was identified to potentially increase the well productivity by some 60% compared to a single horizontal.

This paper will present the well design, the details of a numerical simulation study to justify and optimise the design concept, and finally a discussion of the simulated versus the actual well results.

Field summary

The field is a prolific carbonate build-up that has been in production since Dec 1996 at an average field offtake rate of several hundred MMscf/d. The average field permeability is in the range of several hundred mD, and the original gas column height was some 300 ft. To date the field has produced 74% of the ultimate recovery and has been experiencing water breakthrough in its deviated producer wells since mid-1998. The aquifer has risen on average some 150 to 200 ft, leaving only a relatively small gas column unswept. To maintain the field capacity and to further increase the field ultimate recovery it was decided to drill two horizontal infill wells, one according to the newly-identified well design, and one conventional 7.5/8-in tubing-flow single lateral horizontal well. In this field, coning and cusping would be a major constraint in maximising ultimate recovery as well as reducing the well capacity upon water breakthrough.

Summary

We propose a new method to estimate the Dykstra-Parsons coefficient that leads to a more statistically reliable indication of the true heterogeneity level. The new method extracts more information from the data to produce an estimate that gives half the error of the traditional approach. Several cases are considered to demonstrate the effects of more reliable Dykstra-Parsons coefficients on predicted reservoir performance.

Introduction

The heterogeneity of a petroleum reservoir is a vital factor when the performance of an enhanced recovery project is considered. Numerous studies have shown that permeability variations in the reservoir can be important in determining the amount of petroleum recovered. These variations are also influential in determining how the petroleum is recovered; performance factors such as time to breakthrough and peak hydrocarbon production have important economic implications for a recovery process. Unfortunately, the influence of heterogeneity on process performance is usually quite complex; detailed studies involving substantial amounts of data and extensive analyses are needed to predict hydrocarbon recovery accurately. The petroleum engineer, however, may have to make a preliminary assessment of the reservoir performance when extensive data sets are unavailable or detailed analyses are not justified. In such cases, one is forced to use a simple statistic (e.g.. the Dykstra-Parsons coefficient or Koval's heterogeneity factor) to represent the level of heterogeneity. Such situations can arise, for example, during screening studies. Basic models may then be applied to estimate the recovery behavior and to determine whether a detailed study is justified. The use of simple heterogeneity measures can also arise when results from different field trials"' or different process models are compared. Thus. despite the complex nature of the heterogeneity/performance relationship, a need for simple, representative, heterogeneity measures still exists. In the past, several simple statistical measures of reservoir heterogeneity have been used. The most popular appears to be the Dykstra-Parsons coefficient, KDP. This coefficient has been found to be a good indicator of the level of heterogeneity and has been used in a variety of enhanced recovery studies. Jensen and Lake, however, have shown that estimates of the true reservoir KDP can suffer from substantial statistical errors because the estimated value of KDP, (KDP)est, is computed on the basis of a limited number of samples. These statistical variations in (KDP)est can lead to significant errors in performance predictions. To make the most of this useful heterogeneity measure, the statistical variations of (KDP)est should be reduced as much as possible. Three factors influence the variability of KDP estimates (assuming random sampling). Two factors are the size of the data set and the true value of KDP for the reservoir. 12 Only the first of these two can be changed for a given reservoir. Not surprisingly, as the number of samples increases, the variations in (KDP)est decrease. Typically, a four-fold increase in the number of samples will halve the error. On this basis, obtaining an acceptably accurate (KDP)est could prove rather expensive in terms of the number of permeability measurements required. Also. we already observed that there are times when predictions must be made with small data sets. The remaining factor that influences (KDP)est. errors is the method used to compute (KDP)est. A prudent choice in the Dykstra-Parsons coefficient estimator could also reduce (KDP)est variability. The method currently used in the industry to compute (KDP)est has not substantially changed from the original description by Dykstra and Parsons. Lambert examined six estimators of KDP by comparing, on a well-by-well basis, the values of (KDP)est for data from five different fields. Her study compared the estimators for agreement of values (bias). She did not, however, examine the statistical variability of the estimates (efficiency). The study by Jensen and Lake proposed a new heterogeneity measure, U, that is related to KDP and gives estimates 30% less variable than (KDP)est. This paper proposes a method, based on the maximum likelihood technique, that improves on the traditional method of calculating (KDP)est. We show that, by obtaining more information from the data at hand, we can halve the variability of (KDP)est compared with the traditional method of estimation. This performance gives a further 30% improvement over that achieved by Jensen and Lake. 12 The implications of this improvement on performance predictions are considered in several examples. Because these results depend on estimating the underlying probability density function (PDF) of the permeability data, we discuss and compare some methods to do this efficiently. The results of a Monte Carlo study suggest two equally efficient methods.