May 2016

Marine Seismic Imaging

Microseismic Fracture Mapping

Corrosion Review

Slide Drilling—Farther and Faster

Oilfield Review

Join us on Facebook. Oilfield Review is now on Facebook. Join to learn when new issues are published and defining series articles added and improve your knowledge through definitions selected from Oilfield Glossary.

For users of the Oilfield Review app, learn when new issues are available for download.

Download the app by visiting the Apple or Google Play online store and searching for Schlumberger Oilfield Review.

Oilfield ReviewAuthoritative. Relevant. Informative.

16-OR-0002

In the early 1920s, Jim Abercrombie and Harry Cameron designed and built the world’s first blowout preventer (BOP) in their shop in Houston. In France, brothers Conrad and Marcel Schlumberger were developing methods to explore the Earth’s subsurface using electronic sensors. The Cameron Iron Works Company was established in Houston in 1920, just six years before the Schlumberger brothers registered their company—Société de Prospection Électrique—in France.

Although the two companies are of similar vintage, they pursued different paths to success and now are one. Cameron built its reputation predominantly on the surface, providing the equipment necessary for operators to manage drilling and production operations. Schlumberger success is anchored by decades developing and refining methods to remotely explore, understand and produce from reservoirs below the surface. As is often the case when two companies join forces, the greatest benefit arises from a unique com-bination of similarities and differences.

Cameron has constantly defined and refined the surface technologies and services essential for operators to drill, complete and produce hydrocarbons. Schlumberger tech-nologies help operators find, access, quantify and optimally recover hydrocarbons. Both companies have forged much of their reputations in the most challenging surface and subsurface environments.

Cameron provides the means and methods to confidently manage and control flow in all environments, pressure regimes and temperature ranges throughout the world. Schlumberger uses technology to help operators understand and optimize their assets through seismic surveying, down-hole logging, testing, laboratory analysis and software plat-forms. For Cameron and Schlumberger, the meeting place is the wellhead.

Managed pressure drilling (MPD) is one illustration of the Schlumberger and Cameron fit. Today, a small but growing group of engineers is coming to view MPD not as a specialty method reserved for wells that are otherwise technically impossible to drill but as a conventional drill-ing technique. Schlumberger will leverage its research, software and engineering analytics capabilities and down-hole expertise together with Cameron manufacturing, flow control and automation technology to enable MPD to fulfill its potential as a standard well construction technique.

Schlumberger and Cameron also share a culture of excel-lence. They employ award-winning engineering fellows, holders of PhDs and recognized industry experts. Both companies invest substantially in research and development;

1

as a consequence, they are recognized as innovators in their respective industry sectors.

Cameron is adopting wireless technologies to track assets and monitor equipment condition and is investigating the advantages of nanotechnology and composite materials. And both companies will combine existing research pro-grams such as the use of 3D printing to print metal parts using high-end metals such as Inconel† alloy. This tech-nology has the potential to allow engineers to design highly complex structures that would otherwise be impossible to manufacture.

Schlumberger state-of-the-art mechanical, electronics and sensing technologies, combined with the rigorous application of materials science, have led to the develop-ment of high-performance downhole drilling systems. Joining these systems with the Cameron drilling equip-ment portfolio and automation capabilities will create a fully integrated land drilling system to create a step change in operational efficiency.

The integration of Schlumberger and Cameron is a story of combining expertise and innovative technologies. The union creates a company that can function as a partnership internally and with its clients. In the larger view, the inte-gration of surface and subsurface experts can enhance the performance of drilling and production systems and serve as a model of the collaboration and commercially beneficial alignment between operator and service company.

Justin RounceVice President Marketing & Technology, Cameron GroupHouston, Texas, USA

Justin Rounce returns to Schlumberger as the Vice President Marketing & Technology, Cameron Group from the position of Cameron vice president of marketing and chief technology officer. Prior to joining Cameron, Justin was vice president of marketing and technology for OneSubsea, a joint venture with Schlumberger and vice president and director, mergers and acquisitions for Schlumberger. Justin joined Schlumberger in 1987 as a testing engineer in the North Sea region; he moved to the Wireline division in 1992 and then into various roles within the company, including field operations, operations man-agement, product management and new technology development. In 2003, he joined the Schlumberger Information Solution Segment as vice president, software products. In 2007, he was promoted to vice president, software governance, and was responsible for all internal and commercial software in Schlumberger. In 2009, he became vice president, marketing and technology for the Schlumberger Production Group.

EDITORIAL

Schlumberger and Cameron: A Meeting of Minds and Technologies

The integration of Schlumberger and Cameron is a story of combining expertise and innovative technologies.

† Inconel is a registered trademark of Special Metals Corporation.

May 2016 Volume 28, Number 2ISSN 0923-1730www.slb.com/oilfieldreview

Articles4 Marine Imaging in Three Dimensions: Viewing Complex StructuresSeismic imaging may fail to resolve potential exploration targets beneath shallow rock layers. New developments in marine seismic acquisition and imaging are helping to reduce this uncertainty.

16 Hydraulic Fracturing Insights from Microseismic MonitoringMicroseismic monitoring is used for evaluat-ing hydraulic stimulation operations in unconventional reservoirs. Advanced tech-niques and new technologies are allowing opera-tors to use microseismic data for effective well placement and manage stimulations in real time.

34 Corrosion—The Longest WarOil and gas operations often provide ideal envi-ronments for corrosion to develop and grow. In spite of the challenges created by corrosion, anti-corrosion practices in use today help operators maintain equipment integrity and safely produce hydrocarbons.

50 Slide Drilling—Farther and FasterA surface-mounted torque-oscillation system provides consistent weight transfer to the bit, improving rate of penetration and directional control during sliding operations.

Executive EditorCharlie Cosad

Senior EditorsTony SmithsonMatt VarhaugRick von Flatern

EditorsIrene FærgestadRichard Nolen-Hoeksema

Contributing EditorsDavid AllenH. David LeslieGinger Oppenheimer

Design/ProductionHerring Design

Illustration Herring DesignPennebakerGeorge Stewart

PrintingRR Donnelley—Wetmore Plant

Advisory PanelHani Elshahawi Shell Exploration and Production Houston, Texas, USA

Gretchen M. Gillis Aramco Services Company Houston, Texas

Roland Hamp Woodside Energy Ltd. Perth, Australia

Dilip M. Kale ONGC Energy Centre Delhi, India

George King Apache Corporation Houston, Texas

Michael Oristaglio Yale Climate & Energy Institute New Haven, Connecticut, USA

John ThorogoodDrilling Global Consultant LLPAberdeenshire, UK

PublishingOilfield Review is published and printed in the USA.

© 2016 Schlumberger. All rights reserved.

2

Departments

1 EditorialSchlumberger and Cameron: A Meeting of Minds and Technologies

57 Looking BackBirth of La Pros: The 90th Anniversary of the first Schlumberger Company

60 The Defining SeriesSelected from the Defining Series online: Subsea Infrastructure and Geophysics

64 Looking BackOrigins of the Technique of Wireline Logging

Oilfield Review

On the cover: Cameron specialists with a seven-cavity TL* offshore ram-type blowout preventer (BOP) stack at the Cameron facility in Berwick, Louisiana, USA. Cameron became part of Schlumberger in March 2016. Although the company is most well-known for BOPs, the first designed in 1922 by company founder Harry Cameron, the portfolio includes a complete range of rig, wellhead, production and process equipment and services as well as valves; OneSubsea provides subsea production and processing systems.

An asterisk (*) denotes a mark of Schlumberger.

Marine Imaging in Three Dimensions: Viewing Complex Structures

Prospective reservoirs are typically located beneath complex rock layers that act as reflective barriers to seismic signals and thus distort images of the subsurface. The chal-lenge to geophysicists is to peer through the overburden to the reser-voir beneath it.

To do so, geoscientists use seismic survey data, which undergo advanced processing, imaging, inversion and interpretation workflows. The results enable exploration teams to make immediate decisions about a pros-pect’s value.

In challenging offshore frontier environments, operators are using a recently developed seismic technol-ogy and inversion process to provide 3D, full-bandwidth imaging of fine-scale structures. Full waveform inversion results in a model of seis-mic velocities that can be used with the seismic data to form an image of the geology from the surface to the targets of interest. Page 4.

Hydraulic Fracturing Insights from Microseismic Monitoring

Horizontal drilling and hydraulic fracturing revolutionized the exploi-tation of tight and unconventional oil and gas reservoirs. Engineers can track the progress of hydraulic fractures through a formation using microseismic monitoring.

Armed with new techniques and technologies, operators are able to develop unconventional reservoirs with reduced risk and greater under-standing and certainty than ever before. By mapping the spatial and temporal patterns produced by micro-seismic events, they can refine reser-voir models and create field development strategies.

Microseismic monitoring, intro-duced in the 1980s, evolved as the application of hydraulic fracturing of horizontal wells increased since that time. Using new methods and tech-niques, geoscientists are improving the answers that can be derived from microseismic data. Engineers can even adjust stimulation programs on the fly using real-time data to respond to downhole events as they occur. Modeling software can predict antici-pated downhole responses and then update the production predictions based on the data they acquire from downhole. Page 16.

Corrosion—The Longest War

Corrosion has brought down bridges, downed aircraft, leveled chemical plants, parted drillpipe and ruptured pipelines. Given sufficient time, this phenomenon has the potential to degrade any material. In certain environments, the unchecked effects of corrosion can come swiftly, and the consequences of failure to man-age corrosion can be costly.

One North Sea operator reported that outlays for corrosion prevention and control averaged about 8% of its total project capital expenditures. Direct costs associated with those expenses include replacement of cor-roded equipment and lost production and contamination; indirect costs arise from health, safety and environ-ment concerns.

The environments that host oil and gas operations often provide ideal conditions for corrosion. Ongoing research and advances in coatings, cathodic protection, nondestructive testing, corrosion analysis and inhibi-tors allow operators to safely produce oil and gas in these corrosive environ-ments. Page 34.

Slide Drilling—Farther and Faster

Directional wells allow operators to efficiently access reservoirs and maximize wellbore exposure to pro-ductive zones. In many such wells, directional drillers use steerable mud motors to kick off the well, build angle, drill tangent sections and maintain trajectory necessary to hit target zones.

When using mud motors, drillers alternate between rotating and sliding modes of drilling. In rotating mode, in addition to the downhole motor, the drilling rig’s rotary table or topdrive rotates the entire drillstring to trans-mit power to the bit. During slide drilling, the drillstring does not rotate, and the bit is turned by the mud motor alone. As a consequence, less weight is transferred to the bit, and slide drilling is less efficient than rotary drilling.

An automated torque control sys-tem alternates torque direction to rock the drillstring, which increases ROP through better transfer of weight to the bit while in sliding mode. This weight transfer also helps control tool-face orientation. In addition, it helps minimize the number of downhole motor stalls and increases bit life by preventing weight from being trans-ferred to the bit suddenly. Page 50.

Article Summaries

3

Oilfield Review onlineVisit www.slb.com/oilfieldreview for the current and all previous editions, videos associated with certain articles, the defining series, as well as links to the apps and Oilfield Glossary.

www.slb.com/oilfieldreview

Oilfield Review appDownload the free app by visiting the Apple or Google Play online store and searching for Schlumberger Oilfield Review. Previous issues and video content are available.

About Oilfield ReviewOilfield Review is the Schlumberger flagship journal of technology, innovation and the science of E&P. Contributors to articles are industry professionals and experts from around the world. Those listed with only geographic locations are employees of Schlumberger.

Reproductions without permission are strictly prohibited.

CorrespondenceOilfield Review 5599 San FelipeHouston, TX 77056United States(1) 713-513-3760E-mail: oilfieldreview@slb.com

4 Oilfield Review

Marine Imaging in Three Dimensions: Viewing Complex Structures

Recent developments in multimeasurement marine seismic acquisition and full

waveform imaging enable geophysicists to compensate for distortions caused by

shallow geology and sharpen images of deep targets to reduce the uncertainty of

seismic information.

Anatoly AseevMoscow, Russia

Sandeep Kumar ChandolaLow Cheng FooPETRONAS Carigali Sdn BhdKuala Lumpur, Malaysia

Chris CunnellMalcolm FrancisShruti GuptaPeter WattersonGatwick, England

Michelle ThamKuala Lumpur, Malaysia

Oilfield Review 28, no. 2 (May 2016).Copyright © 2016 Schlumberger.For help in preparation of this article, thanks to Thomas Ajewole, M. Nabil El Kady, M. Faizal Idris, Satyabrata Nayak and M. Iqbal Supardy, PETRONAS Carigali Sdn Bhd, Kuala Lumpur, Malaysia; and Richard Coates, Houston, Texas, USA.Dynel 2D, IsoMetrix and Q-Marine are marks of Schlumberger.

Hydrocarbon exploration requires that geoscien-tists understand the geology of prospective reser-voirs often located beneath complex rock layers. From the geophysicist’s perspective, the overbur-den acts as a defective lens, distorting seismic images of deeper geologic structures. As a result, targets appear indistinct, distorted, out of place or, in extreme cases, completely obscured. The geo-physicist’s challenge has been to devise methods for peering through the overburden and bringing the underlying geology into focus.

Make-or-break decisions on project viability often hinge on how well prospective reservoirs can be imaged, a key factor determining exploration risk. Operators need accurate images of reservoirs to help them place exploration wells where they effectively test the prospect, conduct field planning and place development wells. In addition to imaging reservoirs, geophysicists must correctly image the overburden—the layers above the reservoir—to reduce drilling risks from operational challenges such as maintaining a stable wellbore and controlling formation pressure.

The value that seismic data adds to the exploration process depends on the quality of the image produced and the cost incurred in acquiring such data. Cost-effective seismic acquisition requires surveying large areas quickly without compromising data quality and while minimizing operational and environmental exposure. Fast acquisition helps shorten the time frame between the decision to evaluate a play and the decision to drill.

High-quality data enable exploration teams to attain a clear understanding of the geology from the seafloor to the target prospect and then to decide whether to test and appraise the prospect. The acquired data must also be suitable for use in advanced processing, imaging, inversion and

interpretation workflows. These workflows pro-vide vital inputs for geomechanical, reservoir and basin models.

IsoMetrix marine isometric seismic technology and full waveform inversion processing are enabling imaging of complex structures in frontier areas. The IsoMetrix technology allows for full-bandwidth imaging of fine-scale structures in the subsurface in all directions—inline, crossline and vertical—for detailed imaging from seabed to reservoir. Full waveform inversion results in a model of seismic velocities, which is used with the seismic data to form an image of the geology from the surface to the targets of interest.

This article describes surveys acquired using IsoMetrix technology in offshore Malaysia and the North Sea. The survey results demonstrate the benefits of IsoMetrix technology for over-coming a challenging acquisition environment and increasing spatial bandwidth and of applying full waveform inversion for determining over-burden and reservoir properties, specifically seis-mic velocities.

Improving Data and Image QualityGood seismic imaging requires a chain of factors: a good acquisition system, optimal survey geom-etry and accurate processing algorithms and workflows. More than 15 years ago, Schlumberger geophysicists embarked on a program to move from conventional seismic acquisition toward dis-crete sensor technology. The technology includes improvements in receiver sensitivity and position-ing accuracy, steerable streamers, increased source control and point-receiver acquisition, which records traces from individual receivers to provide consistently repeatable high-quality data.1 These capabilities are evolving. New measure-ments of the crossline and vertical gradients—

1. Christie P, Nichols D, Özbek A, Curtis T, Larsen L, Strudley A, Davis R and Svendsen M: “Raising the Standards of Seismic Data Quality,” Oilfield Review 13, no. 2 (Summer 2001): 16–31.

2. Robertsson JOA, Moore I, Vassallo M, Özdemir K, van Manen D-J and Özbek A: “On the Use of Multicomponent Streamer Recordings for Reconstruction of Pressure Wavefields in the Crossline Direction,” Geophysics 73, no. 5 (September–October 2008): A45–A49.

3. For more on full-azimuth seismic surveying and imaging: Brice T, Buia M, Cooke A, Hill D, Palmer E, Khaled N, Tchikanha S, Zamboni E, Kotochigov E and Moldoveanu N: “Developments in Full Azimuth Marine Seismic Imaging,” Oilfield Review 25, no. 1 (Spring 2013): 42–55.

4. Inline is in the direction the seismic vessel travels and acquires data; crossline is in the direction perpendicular to vessel travel.

5. For more on the Q-Marine system: Christie et al, reference 1. For more on the 3C seismic MEMS accelerometer unit:

Paulson H, Husom VA and Goujon N: “A MEMS Accelerometer for Multicomponent Streamers,” paper We P6 06, presented at the 77th European Association of Geoscientists and Engineers Conference and Exhibition, Madrid, Spain, June 1–4, 2015.

May 2016 55

variations with distance—of the pressure wave-field enable the signals received from a marine seismic shot to be processed as a full 3D wave-field rather than as a collection of 2D profiles.2 In addition, a newly developed, calibrated, broad-band marine seismic source provides improved low-frequency signal content; no source notches, or missing frequencies, below 150 Hz for all direc-tions within a 20° cone from the vertical; and cancellation of the source ghost—a delayed reflection of the source from the sea surface.

These acquisition improvements have been complemented by innovations in marine surveying geometries—for example, multivessel shooting and full-azimuth source-receiver configurations. Together, these technologies make it possible to illuminate targets of interest previously obscured by folded or faulted sediment, overlying salt layers or other complex geologic bodies.3

Seismic acquisition and survey geometry are only the starting points for seismic imaging. Accompanied by onboard processing capabilities, data reliability has vastly improved. In addition, the application of robust seismic inversion and imaging techniques, such as full waveform inver-sion and reverse time migration, allow geophysi-cists to deliver sharper images and estimate rock properties for explorationists and reservoir

engineers who develop static and dynamic models of the reservoir. These models are based on the seismic results—images, velocities and horizons—that are integrated with well data. Before drilling, explorationists use the models to predict the petroleum systems present within the seismically imaged volume, define plays and locate prospects for drilling. Reservoir engineers use refinements of these models to plan field development and, later, manage hydrocarbon recovery operations.

Imaging Between StreamersThe purpose of IsoMetrix technology is to pro-vide a densely sampled representation of the wavefield in all directions. An idealized seismic acquisition system would be able to record the seismic signals from everywhere below the sur-face. This capability would maximize the oppor-tunities for separating the signal from unwanted noise and imaging the reflectors in the subsur-face. However, conventional seismic data are recorded along only a small number of long streamers towed behind a vessel. Thus, although conventional seismic data are well sampled in recording time and along the streamer (inline), they are not recorded between the streamers (crossline), which may be separated by large dis-tances of 50, 75 or 100 m [164, 246 and 328 ft].4 As a result, any waves propagating in the

crossline direction may be aliased, or inade-quately sampled.

Often, the focus of marine seismic imaging is to thoroughly sample the wavefield in the reservoir. However, good sampling of the wavefield in the overburden is also important because these depths must be imaged correctly to enable the geophysi-cist to see clearly into the reservoir. Sampling the seabed or other interfaces that generate multiple reflections is important because such reflections interfere with primary reflections. Shallow depths are important because of possible seabed and shallow subsurface hazards to drilling.

Typical marine seismic receivers are hydro-phones that record the pressure wavefield only. Reconstruction of the pressure field between streamers requires interpolation between known pressures at each streamer location and results in crossline pressure fields becoming aliased and incorrect.

The IsoMetrix technology is based on the Q-Marine point-receiver marine seismic system and combines hydrophones for measuring the seis-mic wavefield pressure with a three-component (3C) microelectromechanical systems (MEMS) unit.5 The 3C MEMS unit contains three orthogo-nal accelerometers for measuring the full 3D vec-torial motion—magnitude and direction—of the recorded wavefield (Figure 1).

Figure 1. Streamer element. An element of the IsoMetrix streamer system (left) combines a hydrophone (inset) that measures pressure (P ) with a calibrated, triaxial microelectromechanical (MEMS) accelerometer that measures the axial, or inline (x), radial, or crossline (y), and vertical (z) accelerations. The IsoMetrix technology facilitates interpolation between streamers. Using hydrophone streamers (top right), only the amplitude of

the wavefield (blue) can be measured at each streamer location (black dots). Therefore, the reconstructed wavefield (red) between streamers is aliased and incorrect. Using multisensor streamers (bottom right), the wavefield amplitude and gradient (cyan) can be measured at each streamer location. Consequently, using both attributes of the wavefield, geophysicists can reconstruct the wavefield accurately between streamers.

x

yz

P

10 mm

Crossline direction

Multichannel Reconstruction

Single Channel Interpolation

Crossline direction

6 Oilfield Review

By adding 3C accelerometers, the marine receivers record the variation of acceleration, which is proportional to the pressure gradient, or the spatial derivative of pressure with direction. In an acoustic material such as water, hydro-phones measure the pressure (P) fluctuations caused by the seismic wave. Three-component accelerometers measure the accelerations in three orthogonal directions (ax, ay and az). Newton’s Second Law specifies the force that results from a difference in pressure; the force is directed from high to low pressure. The relation-ship between the difference in pressure with direction—the spatial derivative of P—and the acceleration, for example in the x direction, is ρ × ax = −∂P/∂x, where ρ is the material density, and the direction of force is opposite, or negative to, that of the pressure gradient. This type of rela-tionship holds for each spatial direction (x, y and z) and allows the calculation of the spatial derivative of pressure directly from the acceleration mea-surement. Consequently, knowing the pressure gradients, geophysicists can reconstruct the una-liased pressure field in all directions. Therefore, geophysicists can estimate the 3D wavefield around the streamers using the same spacing in all directions—inline, crossline and vertical.

Reconstructing the WavefieldThe ability to measure the crossline wavefield gradient enables geophysicists to acquire marine

seismic data using streamers spaced farther apart than those in conventional surveys and to reconstruct the 3D wavefield on a dense grid at points between streamers (Figure 2). For exam-ple, if the actual recordings were accomplished using eight streamers spaced 75 m apart, provid-ing a streamer spread that is 525 m [1,720 ft] wide, the wavefield may be reconstructed as if it were recorded using virtual streamers spaced a tenth of the distance—7.5 m [24.6 ft] apart. When wide streamer spacing is used, areas of exploration can be surveyed faster and more effi-ciently using fewer sail lines, thereby reducing survey duration, acquisition cost, operational complexity and exposure to adverse environmen-tal conditions.

Recording the vertical wavefield component improves the geophysicist’s ability to remove noise, particularly ghost reflections, which are always present in marine seismic survey record-ings. Ghosts are generated when the upward trav-eling primary signal is reflected downward by the sea-air interface. This downward traveling ghost is detected by the seismic receivers and, if uncor-rected, causes a frequency dependent blurring of the final image. Using the vertical acceleration measurements, the geophysicist can separate the upgoing and downgoing components of the wave-field, thereby facilitating removal of ghost reflec-tions. The ability to remove the ghosts also allows IsoMetrix streamers to be towed deeper than

hydrophone-only streamers; towing deep often reduces other sources of noise such as those caused by ocean waves and by the motion of the streamer through the water.

Generalized matching pursuit (GMP) is a pro-cessing method that can take advantage of the multimeasurement data delivered by the IsoMetrix technology.6 The GMP process operates on components of the seismic wavefield that are not confined to traveling straight from the source to the receiver but instead have a significant degree of propagation across the streamer spread. These components may include seismic reflections, diffractions, multiples or other noise modes, and, if not treated correctly, can generate spurious effects in the final images. For example, any energy arriving from the crossline direction, which had been spatially aliased previously in conventional datasets, can now be sampled appropriately using GMP spatially and tempo-rally by taking advantage of the crossline and vertical gradient measurements.

The GMP process is data driven and has proved that it can interpolate the pressure wavefield accurately in the crossline direction, even in adverse situations in which the results from conventional processing would be highly aliased. The output from the GMP process is a grid of data channels spaced 6.25 m [20.5 ft] apart in the inline direction along virtual streamers, which are nominally separated by 6.25 m in the crossline direction.

The ability to image in 3D enables geophysi-cists to consider seismic survey acquisition designs that depart from common practice, as one operator learned when faced with data acquisition challenges.

Challenging Acquisition ConditionsTo clearly define prospects in the South China Sea, geophysicists at PETRONAS Carigali Sdn Bhd acquired a broadband 3D seismic survey off-shore Malaysia. The survey area is an elongated rectangle oriented NW–SE. A major N–S striking fault crosses the survey area, and structural dips

6. For more on generalized matching pursuit: Özbek A, Vassallo M, Özdemir K, van Manen D-J and Eggenberger K: “Crossline Wavefield Reconstruction from Multicomponent Streamer Data: Part 2—Joint Interpolation and 3D Up/Down Separation by Generalized Matching Pursuit,” Geophysics 75, no. 6 (November–December 2010): WB69–WB85.

7. Chandola SK, Foo LC, El Kaldy MN, Ajewole TO, Nayak S, Idris MF, Supardy MI, Tham M, Bayly M, Hydal S, Seymour N and Chowdhury B: “Dip or Strike?—Complementing Geophysical Sampling Requirements and Acquisition Efficiency,” Expanded Abstracts, 85th SEG Annual International Meeting and Exhibition, New Orleans (October 18–23, 2015): 110–114.

Figure 2. Marine seismic acquisition via conventional versus IsoMetrix technology. Conventional surveys (left ) are acquired using streamers of closely spaced hydrophones. The resultant seismic dataset consists of a set of parallel vertical sections. Surveys acquired with IsoMetrix technology (right ) use streamers of closely spaced multisensor receiver units. The multiple components enable interpolation between streamers, and the resultant dataset is a true 3D grid.

Survey vessel

Streamers

Seismic dataset

Crossline

Inline

May 2016 7

as high as 50° occur in the area along a W–E trend (Figure 3). The area is bounded on the west by a no-access zone that survey vessels are not permitted to enter.7

Typically, optimal seismic acquisition geometry for conventional 3D surveys requires shooting parallel to the predominant structural dip direction. This inline direction facilitates close-spaced sampling of the seismic wavefield in the dip direction, in this case W–E, in which the geology has the most variation. In addition, the typical conventional seismic bin, or survey subdivision, into which geophysicists sort seismic traces, is asymmetric and elongated in the structural strike direction, which is the crossline direction.

The no-access zone prohibited the vessel from obtaining full subsurface coverage at the western edge of the survey and presented an acquisition challenge to geophysicists, who considered two options (Figure 4). In the first scenario, they could acquire most of the survey by shooting short lines, spaced 100 m apart, parallel to dip to avoid the no-access area. Then, complete the sur-vey using long lines, spaced 50 m apart, sailing parallel to strike adjacent to the no-access zone boundary; the close line spacing of these strike-parallel lines ensured adequate sampling of the structural dip. Alternatively, they could acquire the entire survey using exclusively strike-parallel sail lines.

The first option was inefficient because of the two acquisition directions, which required nonpro-ductive time during the many turns and while the streamers were repositioned for close spacing. The second option was more efficient for acquiring data but risked degrading the seismic information if acquired using conventional technology. According to conventional wisdom, the strike-par-allel survey direction, which had typical line spac-ing and sampling of the seismic wavefield in the dip direction, was not ideal for imaging the subsur-face and meeting the objectives of company geolo-gists and geophysicists.

The company used IsoMetrix technology, which enabled symmetric, isometric, or equidis-tant, sampling of the wavefield in the inline and crossline directions, to acquire the survey paral-lel to the structural strike. In addition, the com-pany acquired a smaller swath of data in the direction of the dominant structural dip, which would allow comparison and validation of the integrity of survey shooting in strike.

The data were acquired using ten 8-km [5-mi] long streamers spaced 100 m apart. The stream-ers were towed at a water depth of 18 m [59 ft] to minimize noise from variable currents and inclement weather during the survey campaign.

Figure 3. Geologic structure. In the time structure map of the horizon of interest (left ), the contour interval is 100 ms two-way traveltime. The black area is a major fault surface that dips to the east. The white quadrilateral is the survey area, and a no-access area is west of it. The fault is 5 to 8 km [3 to 5 mi] wide, has a N–S strike and a throw of about 2.5 s two-way traveltime. The horizon map on the right—the surface area at the prospective reservoir level—shows structural dips that have been estimated from legacy seismic data. The dips are aligned along a W–E trend.

Structural dip, E–W directionMajor faults, N–S direction

Geologic dip of surface of interest

No-access area

30˚

60˚

0˚

N

Average dip~7˚

Average dip~40˚ to 50˚

Average dip~25˚

Fault

Figure 4. Acquisition options. The survey was restricted by a no-access area on its western boundary. The company geophysicists considered two options for acquiring the seismic data. In the first option (left ), the acquisition vessel would sail the main survey area in the dip direction and then reconfigure the streamers and sail the patch survey area, adjacent to the no-access boundary, in the strike direction. In the second option (right ), the entire survey area would be acquired by sailing in the direction of geologic strike and would parallel the no-access boundary. The company chose the second option and elected to use IsoMetrix technology, which allows for reconstruction of the wavefield sampled equally in both inline and crossline directions, to acquire the data.

Estimated Survey Duration:Scenario 1 = 2 × Scenario 2.

˜ 50 km

No-access area

˜ 50 km

No-access area

Entire Survey Area

˜ 16 km

5 km˜ 16 km

5 km

Patch Survey AreaScenario 2Scenario 1Entire Survey AreaN–S shooting10 streamers, 8,000 m long, towed 100 m apartDuring seismic processing, the streamer data are processed and output to a 6.25-m × 6.25-m grid that is 8,000 m long and 950 m wide.

Main Survey AreaE–W shooting10 streamers, 8,000 m long, towed 100 m apartPatch Survey AreaN–S shooting10 streamers, 8,000 m long, towed 50 m apart

Strike

Dip

Main Survey Area

Main Fault Zone

Strike

Dip

8 Oilfield Review

After acquisition, the data were preprocessed and then the full 3D wavefield was calculated using simultaneous interpolation and deghosting by means of the GMP method. Next, the upgoing pressure wavefield (P-wave) was output on a 6.25-m by 6.25-m grid for each shot record for further processing and imaging.

The data proved to be of high quality. For example, a map of the seafloor surface showed sand banks similar to those observed in bathym-etry data obtained using a high-resolution multi-beam echo sounder (Figure 5).

Upon comparing the dataset acquired in the strike direction with that acquired in the dip

direction, geophysicists judged the datasets to be similar (Figure 6). The fine spatial sampling of the wavefield in the inline and crossline direc-tions obtained with IsoMetrix technology enabled the company to accomplish its geologic and geo-physical objectives and achieve acquisition oper-ational efficiency.

In addition to freeing up constraints on seismic survey acquisition design, uniform inline and crossline data wavefield estimation facilitates the increase in spatial resolution and bandwidth required to compensate for distortions caused by shallow overburden layers and to sharpen images of deeper targets. These improvements in resolu-tion and bandwidth helped reduce the uncertainty of seismic information across the operator’s drill-ing prospects.

Broadband in 3DOil discoveries at three locations in the south-west Barents Sea have generated significant interest in exploration of the region. The discov-eries offshore northern Norway at the Gohta prospect in 2013 and at the Alta prospect in 2014 were both by Lundin Norway AS; those at the Wisting Central prospect in 2013 were by OMV (Norge) AS. The Gohta and Alta discoveries were west of the Loppa High, a roughly 150 km [90 mi] long and 100 km [60 mi] wide tilted fault block that has been affected by a series of events in the North Atlantic Ocean that include:• Paleozoic rifting• Mesozoic opening of the North Atlantic Ocean

and of the Greenland and Norwegian seas• Quaternary glaciation.

Figure 5. Seafloor features. Bathymetry data (left) were acquired using a multibeam echo sounder; the black arrows indicate features such as sand dunes, waveforms, mounds and pockmarks on the seafloor. A map of the seafloor surface (right) from the seismic data, which were acquired using IsoMetrix technology, showed similar features.

5 km

10 × 100 m spread6.25 × 6.25 m binning

Bathymetry from IsoMetrix Technology

5 km

5 × 5 m sampling

Bathymetry from Multibeam Echo Sounder

Figure 6. Comparing acquisition directions. Both seismic sections (left and right ) are from identical locations but resulted from perpendicular acquisition directions. The section on the left was from the control swath acquired in the dip, or crossline, direction and stacked in the strike, or inline, direction. The section on the right was from the production volume and acquired in the strike direction but stacked in the dip direction. Except for subtle differences, the sections show similar results and indicate that the IsoMetrix technology yields similar quality data regardless of the acquisition direction. The magenta ovals indicate structures, or features, that appear different from one another as a result of acquisition in the strike direction rather than in the dip direction.

5 s

1 s

18 km 18 km

5 s

1 s

Dip shooting direction, inline stack Strike shooting direction, crossline stack

Production VolumeControl Swath

May 2016 9

8. In this context, phase refers to a wave of a single frequency in a wave train. The phase could be of a compressional (P) wave, shear (S) wave, other waves or their associated reflections and refractions; the wave’s velocity is the phase velocity.

9. Henriksen E, Ryseth AE, Larssen GB, Heide T, Rønning K, Sollid K and Stoupakova AV: “Tectonostratigraphy of the Greater Barents Sea: Implications for Petroleum Systems,” in Spencer AM, Embry AF, Gautier DL, Stoupakova AV and Sørensen K (eds): Arctic Petroleum Geology. London: The Geological Society, Memoir 35 (August 9, 2011): 163–195.

Gernigon L, Brönner M, Roberts D, Olesen O, Nasuti A and Yamasaki T: “Crustal and Basin Evolution of the Southwestern Barents Sea: From Caledonian Orogeny to Continental Breakup,” Tectonics 33, no. 4 (April 2014): 347–373.

The WesternGeco seismic vessel Western Trident acquired the East Loppa Ridge survey in 2014. The survey covered 4,777 km2 [1,844 mi2] and is part of the Schlumberger Multiclient Barents Sea program. The program used IsoMetrix technology to record wide spatial bandwidth data—the recorded wavefield con-tains the fine-scale detail necessary to represent subsurface geology accurately.

In conventional 3D seismic surveys, a com-mon objective is to acquire broadband surveys of high temporal—traveltime—bandwidth and resolution. The ideal broadband survey has a wide band, or range, of frequencies and is acquired at a high sample rate. The objective for maximizing temporal bandwidth is primarily to maximize resolution in depth—to image thin beds and small faults.

Geology is best understood by observations in three dimensions, which requires maximizing

spatial bandwidth in all directions. In the spatial domain, the wavenumber (k) is the spatial frequency, or the number of wavelengths—wave-cycle lengths—λ per unit distance. The wavenumber is analogous to the temporal frequency (f) or the number of wave periods—wave-cycle times—T per unit time. Wavenumber in the space domain and frequency in the time domain are related through the phase velocity (vp), which is equivalent to wavelength divided by period (vp = λ /T), frequency divided by wavenumber (vp = f/k) or wavelength times fre-quency (vp = λ × f ).8 Consequently, for 3D seis-mic imaging of geology, the notion of broadband must be expanded to include 3D spatial band-width and resolution.

The East Loppa Ridge survey was acquired using 12 streamers that were 7 km [4.3 mi] long, spaced 75 m apart and towed at a constant depth of 25 m [82 ft]. After acquisition, the datasets

were preprocessed and then simultaneously spa-tially dealiased and receiver-deghosted in 3D by means of the GMP method.

The tectonic, stratigraphic and petroleum systems geology of the southwest Barents Sea region is complex.9 The structural setting resulted from several tectonic events that established a dense mosaic of fault systems (Figure 7). The

Figure 7. Fault system. This seismic time slice (left ) at 1,100 ms is through the Loppa High; the seismic attribute is displayed to emphasize the variance in seismic reflectivity—areas of high variance values are colored from black to red and yellow. Three major fault systems, which show up as areas of high variance, affected the Loppa High. The W–E striking Asterias fault complex crosses the Loppa structure in the south; the southern portion of the SW–NE striking Hoop fault complex cuts across and forms the narrow

Loppa High graben and the Bjørnøyrenna fault complex separates the Loppa High from the Bjørnøya basin (not shown) on the west. Sections A and B (right top and bottom) display grabens associated with the fault systems. The northern portion of the Loppa structure is in the center of Section A. Section B shows graben structures in the north associated with the Hoop fault complex and in the south associated with the Asterias fault complex, which separates the Loppa High from the Hammerfest basin.

Variance

Low High

Reflection amplitude

– 0 +

W

A

B

A

B

E

N S

5.0 s

0.5 s

0.5 s

5.0 s

Bjørnøyrennafault complex

Asterias fault complex

Loppa High

Hoopfault complex

Bjørnøyrennafault complex

Asterias fault complex

Hoopfault complex

10 Oilfield Review

Loppa High area contains three major fault com-plexes.10 The Asterias fault complex forms the southern boundary, which separates the Loppa High from the Hammerfest basin to its south. The southern portion of the Hoop fault complex strikes SW–NE and cuts across the Loppa struc-ture as a narrow graben. The Bjørnøyrenna fault complex separates the Loppa High from the Bjørnøya basin on the west. Broadband seismic images make it possible to delineate the fault patterns and establish the regional structural framework within the East Loppa Ridge survey area. The structural framework influences local petroleum systems.

The Gohta and Alta oil discoveries were in reservoirs located in carbonates of the Gipsdalen Group, which were deposited in warm, shallow marine environments during the Late Carboniferous to Permian periods and, since then, have been altered by dolomitization and karstification (Figure 8).11 Additional petroleum systems elements in the Loppa High area include reservoir prospects in Triassic sandstones, source rocks in Carboniferous synrift and postrift sedi-ments and in Permian and Triassic sediments and seals formed by Triassic and Cretaceous shales.12 The broadband East Loppa Ridge seis-mic dataset offers an opportunity for detailed interpretation of the complex geology in the Loppa High area.

The upper Paleozoic carbonates have been the most promising stratigraphic level for Loppa High exploration (Figure 9). Broadband seismic data facilitate detailed mapping, analysis and interpretation of the carbonate morphology, which has polygonal ridges characteristic of mod-ern carbonate platforms.

Oil has been discovered in the upper Triassic Snadd Formation but at locations with low res-ervoir quality. Within the Snadd Formation, the broadband seismic data reveal the fluvial sys-tem and aid automated mapping, which should reduce the uncertainty of locating higher qual-ity reservoir sands. The data show complex

10. Gabrielsen RH, Færseth RB, Jensen LN, Kalheim JE and Riis F: “Structural Elements of the Norwegian Continental Shelf. Part I: The Barents Sea Region,” Stavanger, Norway: Norwegian Petroleum Directorate, NPD Bulletin no. 6, May 1990.

11. For more on the Gipsdalen Group: Larssen GB, Elvebakk G, Henriksen LB, Kristensen S-E, Nilsson I, Samuelsberg TJ, Svånå TA, Stemmerik L and Worsley D: Upper Paleozoic Lithostratigraphy of the Southern Norwegian Barents Sea. Stavanger, Norway: Norwegian Petroleum Directorate (2002).

“Barents Sea—Carboniferous to Permian Plays,” Norwegian Petroleum Directorate, http://www.npd.no/en/Topics/Geology/Geological-plays/Barents-Sea/Carboniferous-to-Permian/ (accessed August 29, 2015).

Figure 8. East Loppa regional seismic section. The seismic section (top) runs from the Loppa Ridge in the west toward the Ottar basin in the east. The interpretation (bottom) is a balanced section, which was modeled using the Dynel 2D restoration and forward modeling tool. This section shows extensional rifting and synrift and postrift sediment deposition during the Carboniferous period. During the Late Carboniferous to Permian, a carbonate platform developed and evaporate was deposited. During the lower to middle Triassic uplifting and tilting of the Loppa High, karstification of the carbonates and sedimentation of shales occurred. The upper Triassic and Jurassic periods were characterized by high clastic sedimentation rates and floodplain development from rivers and deltas. Rifting occurred again during the upper Jurassic to lower Cretaceous; the base Cretaceous unconformity (BCU) defines the transition from synrift to postrift sedimentation. Finally, Tertiary sediments occur above the BCU to the seafloor.

Reflection amplitude

– 0 +

BCU

Upper Triassic-Jurassic

Evaporites

Tertiary

Seafloor

Lower-Middle Triassic

Carboniferous postrift

Carboniferous synrift

Late Carboniferous toPermian carbonates

Pre-Carboniferousbasement

5.0 s

0.5 sW E

5 km

12. For more on petroleum systems: Al-Hajeri MM, Al Saeed M, Derks J, Fuchs T, Hantschel T, Kauerauf A, Neumaier M, Schenk O, Swientek O, Tessen N, Welte D, Wygrala B, Kornpihl D and Peters K: “Basin and Petroleum System Modeling,” Oilfield Review 21, no. 2 (Summer 2009): 14–29.

13. For more on the Snadd Formation: Klausen TG, Ryseth AE, Helland-Hansen W, Gawthorpe R and Laursen I: “Regional Development and Sequence Stratigraphy of the Middle to Late Triassic Snadd Formation, Norwegian Barents Sea,” Marine and Petroleum Geology 62 (April 2015): 102–122.

14. Houbiers M, Wiarda E, Mispel J, Nikolenko D, Vigh D, Knudsen B-E, Thompson M and Hill D: “3D Full-Waveform Inversion at Mariner—A Shallow North Sea Reservoir,” Expanded Abstracts, 82nd SEG Annual

International Meeting and Exhibition, Las Vegas, Nevada, USA (November 4–9, 2012).

Houbiers M, Mispel J, Knudsen BE and Amundsen L: “FWI with OBC Data from the Mariner Field, UK—The Impact on Mapping Sands at Reservoir Level,” paper We 11 05, presented at the 75th European Association of Geoscientists and Engineers Conference and Exhibition, London, June 10–13, 2013.

Østmo S, McFadzean P, Silcock S, Spjuth C, Sundvor E, Letki LP and Clark D: “Improved Reservoir Characterisation by Multisensor Towed Streamer Seismic Data at the Mariner Field,” paper We P03 12, presented at the 76th European Association of Geoscientists and Engineers Conference and Exhibition, Amsterdam, June 16–19, 2014.

May 2016 11

fluvial and floodplain geology and reveal that the channel system is associated with flood-plain development (Figure 10).13 The data reveal a variety of fluvial features, including point-bar systems, clustered channel fill com-plexes and ribbon-channel sandstone bodies; the ribbon channels were at depths greater than 1,000 m [3,280 ft] and estimated to be less than 100 m wide.

The East Loppa Ridge survey demonstrates the imaging power of acquiring true 3D, broad-band seismic data. High spatial resolution in all directions facilitates and improves imaging of complicated 3D geology such as fault networks, anastomosing fluvial channel complexes and car-bonate platform deposition and karstification. The increased detail offered by broadband images promotes improved understanding of petroleum system geology and better discrimination of lithol-ogies and their rock properties.

Full Waveform InversionGeophysicists use full waveform inversion (FWI) for calculating horizontal and vertical seismic wave velocities of geology from the surface to tar-gets of interest. The result is a velocity image in depth that reveals the sought-after structural and depositional information.

Traditional migration produces an image of the subsurface by attempting to reposition, or migrate, seismic data reflection points to their correct loca-tions in 3D space. A velocity model is almost always an input to migration; and a refined veloc-ity model may be a byproduct of migration.

Unlike conventional migration, FWI is a method for building a velocity model by attempting to match the complete recorded wavefield that

results as seismic waves travel through the Earth and encounter changing properties in the subsurface geology. The starting point for FWI is an approximate model of velocities. Geophysicists use this velocity model to simulate the recorded wavefield. They then subtract the simulated wavefield from the observed wavefield to obtain the residual wavefield. The residual wavefield is then backward propagated—extrapolated downward in space or backward in time—through the velocity model to obtain a dataset of velocity gradients. These gradients inform where to increase or decrease velocities but not by how much. To calculate a velocity model update, the gradients are multiplied by a step length, which scales the gradients. The velocity updates are added to the current velocity model to create a new velocity model, and the process is repeated. The iterations continue until the residual wavefield is acceptably small, meaning that the modeled wavefield closely approximates the observed wavefield. The final model of seismic velocities can be used as an input to migration to produce an image that better represents subsur-face rock characteristics or may be used directly to interpret rock and fluid properties.

This technique was used in Mariner field, dis-covered in 1981 and located about 150 km [93 mi] east of the Shetland Islands on the UK Continental Shelf in the North Sea. The field is under develop-ment by operator Statoil UK Limited with partners JX Nippon Exploration and Production (UK) Limited and Dyas UK Limited. The field consists of two reservoirs. The shallow reservoir contains heavy oil of 12.1 API gravity and is about 1,200 m [3,940 ft] below sea level in sands of the Heimdal member of the Middle to Late Paleocene Lista

Formation, composed predominantly of shale. The deeper reservoir in the Maureen sandstone member contains heavy oil of 14.2 API gravity and is at the base of the Early Paleocene Våle Formation at depths of 1,400 to 1,500 m [4,590 to 4,920 ft] below sea level.

The Mariner field presents various challenges for seismic imaging.14 The shallow overburden above the reservoirs contains channel sands that have higher seismic velocities than those of sur-rounding geologic units. These sands can be mapped easily, but their presence causes distor-tions in the images of the reservoir zones beneath them. For example, shallow, high-velocity chan-nel sands cause pull-ups of, or apparent struc-tural high spots in, underlying reflectors. The Heimdal reservoir sands consist of complex chan-nel sands as well as sand injectites, or sand intru-sions; these sands are difficult to image because of their low impedance contrast with the shales that host them. The Maureen sandstone contains small-scale faults and calcite layers that are important for developing production from the

Figure 9. Carbonate reservoir. The seismic time horizon of the top surface of the Late Carboniferous– to Permian-age Gipsdalen Group is shown in map (left ) and perspective (right ) views. These views show a seismic attribute that emphasizes edges on the surface. The surface in both views displays polygonal ridges that are reminiscent of polygonal oceanic reef systems observed in modern carbonate platform environments (inset ).

Figure 10. Floodplain channels. This seismic time slice at 1,100 ms is at the depth of the upper Triassic Snadd Formation. The time slice shows the variance in reflectivity. The dark linear and curvilinear features are faults. The lighter gray, sinuous and interweaving features are networks of fluvial channels crisscrossing a floodplain.

Variance

Low High

12 Oilfield Review

sandstone but are below the detection capabilities of traditional seismic techniques. The imaging challenges presented by the reservoirs may be mitigated by full waveform processing techniques that enable removal of the distortions caused by the high-velocity channel sands in the shallow overburden.

In 2012, the operator acquired a broadband seismic survey at the Mariner field using the WesternGeco IsoMetrix technology. The survey data were acquired using eight streamers, each

3 km [1.9 mi] long, spaced 75 m apart and towed at a constant depth of 18 m. After acquisition, the data were preconditioned and then simultane-ously interpolated and deghosted using the GMP method. The upgoing pressure wavefield was then output on a 6.25-m by 6.25 m grid for subse-quent processing and imaging.15

Initial inspection of the dataset showed it to be richer in high frequencies than in two conven-tional 3D seismic datasets and richer in low fre-quencies than in an earlier ocean bottom cable (OBC) survey. Both qualities are important for resolving subsurface geology and velocities through inversion of seismic data. High frequen-cies enable resolution of relative velocities between small stratigraphic and structural details. Low frequencies facilitate determination of absolute velocities, which are calibrated against borehole data.

The data underwent fast-track processing, using prestack time migration, which demon-strated the Heimdal member sands could be imaged more reliably using the broadband data than the earlier data.16 The operator’s geoscien-tists were able to establish the relationship between seismic reflectors and geologic horizons with improved confidence.17 Encouraged by these results, WesternGeco geophysicists applied FWI to the broadband dataset.18

The starting point for FWI is a velocity model (Figure 11). The geophysicists began with a simple model, using seismic velocities interpreted from sonic logs from wells in the area of the Mariner field, which were then interpolated laterally between the wells along layers bounded by known geologic horizons. Based on previous processing studies, the overburden formations were assumed to be anisotropic; the P-wave anisotropy parame-ters epsilon (ε) and delta (δ) were initially defined as linearly increasing from the seafloor to the base Cretaceous unconformity but were subsequently updated using a multiparameter inversion step in the FWI workflow.19

The geophysicists wanted to know whether the results of FWI would isolate the velocities in shallow channel sands within the overburden. As a test, one of the known channels delineated from legacy 3D seismic data was inserted into the initial velocity model and given a higher velocity than its host units. If successful, the FWI method would sharpen the velocities within this control channel but also pick out other channels in the area.

To compensate for velocity imprecisions introduced by interpolation, the geophysicists applied one iteration of common image point (CIP) tomography to the interpolated velocity model. Common image point tomography is an iterative method of inverting for seismic veloci-ties using seismic reflections. During an itera-tion, the amount of residual moveout—depth variation—along reflections in prestack depth-migrated (PSDM) CIP gathers is used to deter-mine adjustments in the velocity model to bring the subsequent version of the PSDM image into better focus.20 After one iteration of CIP tomogra-phy, the velocity model was smoothed and ready for input to the FWI process.

Next, the geophysicists started the FWI pro-cess, which, beginning with the initial earth model of velocities, iteratively models the observed seismic wavefield and adjusts the veloc-ities in the earth model until there is an accept-able match between the modeled wavefield and the recorded wavefield.21 The observed wavefield was the upgoing P-wave wavefield that had been isolated at an early stage of processing from the broadband dataset. The criterion for conver-gence to an acceptable match between synthetic and observed wavefields is to minimize a misfit function that quantifies the difference between the modeled and measured data. To ensure that the FWI process converges on the global, or true, minimum rather than a localized minimum, the geophysicists conduct FWI in stages. First, they find an acceptable fit of the low-frequency wavefield. They then add and fit to successively

15. Özbek et al, reference 5.16. Migration is a seismic processing step in which

reflections in seismic data are moved to their correct locations. Time migration locates reflections in two-way traveltime—from the surface to the reflector and back as measured along the image ray. Depth migration locates reflectors in depth. Mathematically, migration is performed by various solutions to the wave equation that describe the passage of seismic waves through rock. Kirchhoff migration is a ray-based approximation founded on the integral solution to the wave equation derived by 19th-century German physicist Gustav Kirchhoff.

For more on migration and imaging: Albertin U, Kapoor J, Randall R, Smith M, Brown G, Soufleris C, Whitfield P, Dewey F, Farnsworth J, Grubitz G and Kemme M: “The Time for Depth Imaging,” Oilfield Review 14, no. 1 (Spring 2002): 2–15.

17. Østmo et al, reference 14.18. Gupta S, Cunnell C, Cooke A and Zarkhidze A:

“High-Resolution Model Building and Imaging Workflow

Figure 11. Workflow for full waveform inversion.

Final high-frequency FWI iterations to enhancethe resolution of velocities

Build initial velocity model

Control test for FWI: Insert known channel into velocity model

One iteration of common image point (CIP)tomography to smooth velocity model

Low-frequency FWI iterations usingfrequency band, 1.5 to 7 Hz

peak frequency, 2.5 Hz

Intermediate-frequency FWI iterations usingfrequency band, 1.5 to 13 Hz

peak frequency, 5 Hz

One iteration of CIP tomography to refinevelocities in deepest intervals of the model

Multiparameter FWI to refine anisotropic parameters

Using Multimeasurement Towed Streamer Data: North Sea Case Study,” Expanded Abstracts, 85th SEG Annual International Meeting and Exhibition, New Orleans (October 18–23, 2015): 1049–1053.

19. The base Cretaceous uniformity is the term applied to a strong seismic reflection surface that is mappable over much of the continental shelf in the North Sea. The reflector is an unconformity that is located close to the bottom of Cretaceous-age rocks and separates sediments deposited before rifting of the North Sea from sediments deposited after rifting.

Anisotropy is the variation of a physical property, such as P- or S-wave velocity, with the direction of its measurement. For more on elastic anisotropy: Armstrong P, Ireson D, Chmela B, Dodds K, Esmeroy C, Miller D, Hornby B, Sayers C, Schoenberg M, Leaney S and Lynn H: “The Promise of Elastic Anisotropy,” Oilfield Review 6, no. 4 (October 1994): 36–47.

Epsilon (ε) and delta (ε) are P-wave parameters that describe vertical transverse isotropy. Epsilon is the P-wave anisotropy parameter and equal to half the ratio

of the difference between the horizontal and vertical P-wave velocities squared divided by the vertical P-wave velocity squared. Delta is describes near-vertical P-wave velocity anisotropy and the difference between the vertical and small-offset moveout velocity of P-waves. For more on seismic anisotropy parameters: Thomsen L: “Weak Elastic Anisotropy,” Geophysics 51, no. 10 (October 1986): 1954–1966.

20. For more on CIP tomography: Woodward M, Nichols D, Zdraveva O, Whitfield P and Johns T: “A Decade of Tomography,” Geophysics 73, no. 5 (September– October 2008): VE5–VE11.

21. Vigh D, Starr EW and Kapoor J: “Developing Earth Models with Full Waveform Inversion,” The Leading Edge 28, no. 4 (April 2009): 432–435.

22. For more on multiscale inversion: Bunks C, Saleck FM, Zaleski S and Chavent G: “Multiscale Seismic Waveform Inversion,” Geophysics 60, no. 5 (September–October 1995): 1457–1473.

23. Reference 16.

May 2016 13

higher frequency bands until there is an acceptable fit of the full-frequency wavefield. This sequential FWI procedure stabilizes the inversion algorithm and ensures that the process converges to a global minimum.22

Application of FWI to the broadband dataset collected at Mariner field showed that a seismic

dataset acquired using IsoMetrix technology can be inverted for a geologically relevant seismic velocity model that is capable of sharpening the focus of seismic images. After FWI processing, the velocity model was input into two prestack depth migration algorithms: a Kirchhoff depth migration (KDM) to compare directly against legacy data volumes and a

high-frequency reverse time migration (RTM) per-formed directly in the natural shot domain after GMP.23 The velocity model from FWI sharpened the image of the control channel embedded into the overburden of the initial velocity model and highlighted additional channels (Figure 12).

Figure 12. Comparing models before and after full waveform inversion (FWI). Both seismic sections (left top and bottom) show the same geology to a depth of 1,200 m [3,940 ft] below sea level. The depth sections are the result of Kirchhoff depth migration (KDM); the sections are overlain by the velocity model (colors) that was used as input to KDM. The top section resulted from KDM using the initial velocity model. The control channel is in the top center and was given a higher velocity than its surroundings. The bottom

section resulted after using the velocities output after completion of FWI. The control channel is in better focus, and the velocities of other channels are evident. The velocities of the overburden units have become more defined. The images on the right are depth slices at 158, 278 and 844 m [518, 912 and 2,770 ft] below sea level. Compared with the before FWI processing results, geologic features (yellow arrows) have become better defined after FWI processing.

Before FWI Processing

After FWI Processing

2,250 m/s

1,700 m/s

2,025 m/s

1,625 m/s

2,025 m/s

2,225 m/s

2,250 m/s

1,700 m/s

2,025 m/s

1,625 m/s

2,025 m/s

2,225 m/s

2,500 m/s

1,500 m/s

2,500 m/s

1,500 m/s

Depth slice, 158 m

Depth slice, 278 m

Depth slice, 844 m

Depth slice, 158 m

Depth slice, 278 m

Depth slice, 844 m

14 Oilfield Review

Figure 13. Modeling results. A clear progression of improvement occurs from the legacy velocity model and Kirchhoff depth migration (KDM, top) to the revised KDM using FWI model (middle) to the high-resolution reverse time migration (RTM) also using the FWI model (bottom ). The progression demonstrates improved imaging of the steep dips and signal-to-noise characteristics in the reservoir section, discriminating the Heimdal injectite, or intrusion, features (circled) from the background Lista shales. The inset shows conceptual sketches of how the sandstone bodies or intrusions might have become incorporated into the Lista shales above the Heimdal formation. (Inset adapted from Huuse et al, reference 24.)

Legacy KDM Model

KDM using FWI Model

RTM using FWI Model

Reflection amplitude

– 0 +

Concept 1: isolated depositional sandstones

1,500

1,600

1,700

1,800

1,900

2,000

2,100Heimdal sandstone

Wing-like intrusion

Onlap Onlap

Heimdal sandstone

Two-

way

trav

eltim

e, m

s

Two-

way

trav

eltim

e, m

s

Sandstone intrusions crosscut theBalder and Frigg Formation mudstones

Isolated depositional sandstonesin the Frigg Formation mudstones

1,500

1,600

1,700

1,800

1,900

2,000

2,100

Concept 2: sandstone intrusions crosscut mudstones

Heimdal sandstone

Heimdal sandstone

Heimdal sandstone

Top of the Frigg sandstoneTop of the Balder sandstoneTop of the Sele FormationTop of the Heimdal sandstone

1 km1 km

Crestal intrusion fringe

The velocities in the shallow layers became more clearly defined. Below them, the reservoir zones of interest were less distorted. Cross sections through the KDM image volume showed that the velocities from FWI made a demonstrable differ-ence in the focusing and positioning of overburden formations, while the RTM image volume gave the best resolution and signal-to-noise discrimination of Heimdal injectites against the background Lista shales (Figure 13).24

The IsoMetrix marine isometric seismic tech-nology and full waveform imaging are enabling and complementary technologies for increasing the qualitative and quantitative accuracy of seismic information. The IsoMetrix technology allows deghosting and interpolation of the recorded wavefield to produce unaliased seismic records. In turn, FWI provides geologically relevant velocities at scales that can be used to bring the overburden into focus. Together, these techniques enable geo-physicists to image reservoir targets more clearly (Figure 14).

Advances in the sequence of steps from seis-mic data acquisition to final imaging are helping operators characterize the subsurface more distinctly. Measurements of the pressure wavefield and its gradients using IsoMetrix technology represent a significant development

24. For more on injectites: Braccini E, de Boer W, Hurst A, Huuse M, Vigorito M and Templeton G: “Sand Injectites,” Oilfield Review 20, no. 2 (Summer 2008): 34–49.

Huuse M, Cartwright J, Hurst A and Steinsland N: “Seismic Characterization of Large-Scale Sandstone Intrusions,” in Hurst A and Cartwright J (eds): Sand Injectites: Implications for Hydrocarbon Exploration and Production, AAPG Memoir 87. Tulsa: AAPG (2007): 21–35.

May 2016 15

worked in technical, service and marketing managerial positions in the UK, US and Egypt. Before joining the team for IsoMetrix technology, Chris was based in Cairo and managed advanced imaging services, including full waveform inversion and Seismic Guided Drilling* service, across the Middle East and North Africa. He received an MBA degree from the Rotterdam School of Management at Erasmus University, the Netherlands.

Low Cheng Foo is Custodian of Geophysical Acquisition for PETRONAS Carigali Sdn Bhd in Kuala Lumpur, where he is involved with new technology projects such as broadband, multicomponent, multiazimuth and full azimuth seismic data acquisition. He has 35 years of experience with the company. Previously, he was head of acquisition after serving as an acquisition and processing geophysicist. He has been involved in land, marine and transition-zone seismic acquisition programs in various countries in Southeast Asia, the Middle East, Suriname and Cuba. Low earned a BSc (Hons) degree in physics, majoring in geophysics, from the University of Science Malaysia in Penang.

Malcolm Francis is a Schlumberger Advisor and the Eastern Hemisphere Exploration Services Manager for WesternGeco in Gatwick. Before his current role, he held technical and management positions as the Eastern Hemisphere multiclient chief geophysicist, global manager of geology and interpretation and senior manager E&P solutions. Earlier in his career, Malcolm managed the special processing and interpretation departments. He began in the industry in 1980 with Western Geophysical, where he undertook collaborative research with Saudi Aramco. He obtained a bachelor’s degree in geology from the University of Manchester, England, and MSc and PhD degrees in geophysics from Imperial College London. Malcolm is a member of the European Association of Geoscientists and Engineers, the SPE, SEG and Petroleum Exploration Society of Great Britain and is a Fellow of the Geological Society of London.

Anatoly Aseev, based in Moscow, was a Seismic Interpreter for Schlumberger Multiclient seismic projects from 2014 to 2016, with focus on the Norwegian Continental Shelf area. He began his career in 2006 as a geologist with Rosneft in Krasnodar, Russia, and worked on exploration projects in the Ciscaucasia basin. He joined Schlumberger PetroTechnical Services (PTS) in 2011 and served as a geologist and then senior geologist working on exploration projects in the Timan-Pechora, West Siberia, Barents Sea and West Black Sea basins. Anatoly holds an MSc degree in petroleum geology from the North-Caucasus Federal University, Stavropol, Russia. He is pursuing a PhD degree in regional geology from Lomonosov Moscow State University.

Sandeep Kumar Chandola is a Custodian of Geophysics with PETRONAS Carigali Sdn Bhd in Kuala Lumpur. He served with Oil and Natural Gas Corporation, the Indian national oil company, for more than 20 years before joining Petronas Carigali in 2005. His work has supported the design of 3D acquisition geometries and the introduction of new geophysical technologies to the company. He has a master’s degree in physics from Hemvati Nandan Bahuguna Garhwal University, Sringar, Uttarakhand, India, and a specialized diploma in petroleum geophysics from the Indian Institute of Technology (IIT) Roorkee, Uttarakhand. He is a member of the SEG, the European Association of Geoscientists and Engineers and the Society of Petroleum Geophysicists (India), an SEG Honorary Lecturer and an adjunct lecturer at Universiti Teknologi PETRONAS, Malaysia. Sandeep has authored more than 50 publications and is a recipient of the National Petroleum Management Programme Award for Excellence from the government of India.

Chris Cunnell leads Technical Sales and Marketing of IsoMetrix* for WesternGeco in Gatwick, England. Chris, who has more than 20 years of geophysics experience, joined Schlumberger in 1997 and has

Shruti Gupta is an Area Geophysicist for Schlumberger in Gatwick, England, where she provides technical support for time and depth processing of marine and ocean bottom cable (OBC) seismic data. Shruti, who has more than seven years of experience in the oil and gas industry, started her career with Schlumberger as a field geophysicist on a land and transition zone seismic acquisition crew in Egypt. She then worked with the depth imaging group in Houston. She has an MSc degree in applied geology from the IIT Kharagpur, West Bengal.

Michelle Tham is the Technical Support Manager for WesternGeco in the Asia Pacific region as well as the Petrotechnical Expertise Discipline Career Manager for the Schlumberger Asia region; she is based in Kuala Lumpur. She began her career with Schlumberger in Calgary and has worked in the US, Myanmar, Indonesia, Australia, Nigeria, UAE and Malaysia. Before her current position, she served as a seismic data processing geophysicist, data processing supervisor, staff geophysicist, area geophysicist, seismic survey design and modeling manager and geophysics global discipline career manager. Michelle holds a BS degree in geophysics from the University of Calgary.

Peter Watterson is the Manager of the Marine Geosolutions Technology Commercialization group for WesternGeco in Gatwick. His focus is on research, engineering and marketing of various marine seismic acquisition and processing technologies. Pete began his career in the geophysics industry with Western Geophysical in 1991 in London. He has held positions in seismic data processing and technology management in the UK, Venezuela, US, Trinidad and Brazil and worked for several years as the regional geophysicist for WesternGeco for South America. He received a BSc degree in physics from the University of Leeds, England.

Contributors

in marine seismic data acquisition. The development of circle shooting, simultaneous firing of sources and full-azimuth source-receiver configurations embody advances in marine seismic survey geometry and design. Full waveform inversion, along with reverse time migration, is advancing geophysicists’ capability to develop data-driven velocity models. The converging improvements on all three fronts—acquisition, survey design and processing—provide the means for imaging complex geologic structures, forecasting drilling hazards and illuminating reservoir targets. —RCNH

Figure 14. Comparing images from ocean bottom cable (OBC) and IsoMetrix technologies. Both images are seismic depth sections to a depth of 1,700 m [5,600 ft] below sea level. They show the same geology extracted from datasets that have been processed using similar workflows through FWI and prestack depth migration; in each case, the color overlay is the P-wave velocity model that results after processing. For the 2008 OBC survey (left), the FWI processing was completed to a peak frequency of 10 Hz before migration using KDM. For the 2012 survey using IsoMetrix technology (right), the FWI processing was completed to a peak frequency of 5 Hz, followed by migration using high-resolution RTM. Despite some differences in the two workflows, both used a 2.5-Hz peak frequency for the first FWI updates. After processing, the velocity model result from IsoMetrix technology has the same, or better, resolution in the shallow overburden as the model result from the OBC survey.

Ocean Bottom Cable Survey

1.7 km

IsoMetrix Survey

An asterisk (*) denotes a mark of Schlumberger.

16 Oilfield Review

Hydraulic Fracturing Insights from Microseismic Monitoring

Horizontal drilling and hydraulic fracturing revolutionized the exploitation of tight and

unconventional oil and gas reservoirs. Microseismic monitoring provides operators

with crucial information to improve these operations and helps reservoir engineers

with modeling and making decisions on well placement, completion design and

stimulation operations.

Joël Le CalvezRaj Malpani Jian XuHouston, Texas, USA

Jerry StokesMid-Continent Geological, Inc.Fort Worth, Texas

Michael WilliamsCambridge, England

Oilfield Review 28, no. 2 (May 2016).Copyright © 2016 Schlumberger.For help in preparation of this article, thanks to Julian Drew, Perth, Western Australia, Australia; Tony Probert and Ian Bradford, Cambridge, England; and Nancy Zakhour, Callon Petroleum, Houston.CMM, ECLIPSE, Mangrove, MS Recon, NetMod, Petrel, ThruBit, UFM, VISAGE, VSI and VSI-40 are marks of Schlumberger.

Operators producing from unconventional reser-voir plays face many challenges. Fluid flow through unconventional reservoir rocks is lim-ited by matrix permeability, which is generally several orders of magnitude smaller than that of conventional reservoir rocks. Preexisting faults and fracture networks often provide pathways for the flow of hydrocarbons and play an impor-tant role in increasing reservoir drainage vol-umes. Hydraulic fracture stimulation treatments can often connect the wellbore to existing natu-ral fracture networks; however, effective stimu-lation requires knowledge of the distribution of those networks.

Well completion engineers use geomechani-cal and fracture models to plan where to initiate hydraulic fractures and predict their propagation through the reservoir. These models require cali-bration and validation. Microseismic monitoring has proved to be a viable means for calibrating the models and for providing empirical data about the effectiveness of stimulation operations.

Microseismic monitoring is a technique that records and locates microseismic events—col-lectively referred to as microseismicity—which are small bursts of seismic wave energy gener-ated by minute rock movements in response to changes of the in situ stresses and rock volume such as those that occur during fracture stimula-tion operations.1 During these operations, frac-tures are created by injecting fluid at high pressure. These fractures propagate and are then held open using a solid proppant. Mapping the spatial and temporal pattern of these events has proved successful for monitoring the progress of

hydraulic fractures as they advance through and alter a formation.

Engineers may employ several techniques to determine the effectiveness of hydraulic stimula-tion operations.2 For instance, during stimulation operations, microseismic (MS) monitoring and tiltmeter measurements can indicate mechanical changes in the subsurface that occur over a wide area centered on the treatment well.3 Afterward, engineers have used radioactive and chemical tracers, temperature tools and production logs to provide complementary indications of changes in fluid pathways resulting from the stimulation.

Geophysical service companies often acquire MS data, which they interpret and integrate with other measurements to provide oil and gas opera-tors with an understanding of hydraulically induced fracture systems. The primary data used for evaluating MS events are waveform measure-ments acquired from a network of receivers placed either downhole or at the surface. Geoscientists use these data to map the extent and evolution of MS events. These maps provide valuable information related to strain and stress variations in the reservoir and surrounding for-mations and are used to guide stimulation deci-sions during job execution. If MS events indicate undesired fracture growth or fault activation, operators may choose to terminate stage pump-ing early, use diverter technology or skip stimula-tion stages.

Microseismic monitoring also provides infor-mation about the nature of the physical pro-cesses—induced fracturing of the rock or slippage on preexisting fractures—that occur at the

1. Seismic waves convey energy by means of the particle motion of solid materials.

2. For more on fracture diagnostic techniques: Bennett L, Le Calvez J, Sarver DR, Tanner K, Birk WS, Waters G, Drew J, Michaud G, Primiero P, Eisner L, Jones R, Leslie D, Williams MJ, Govenlock J, Klem RC and Tezuka K: “The Source for Hydraulic Fracture Characterization,” Oilfield Review 17, no. 4 (Winter 2005): 42–57.

3. A tiltmeter measures minute rotations—changes of inclination—of the ground in which it is embedded.

1717May 2016

18 Oilfield Review

location of the MS sources. Characterization of the population of MS sources helps quantify the magnitudes and directions of stress and displace-ment variations in the affected reservoir volume during the stimulation. To describe the magnitude and direction of the rock movements at each source location, geophysicists process MS wave-form recordings, account for propagation effects, determine the radiation pattern of the acoustic emission and invert for source properties—rock movements and energy released.4 Reservoir engi-neers then combine the space-time evolution of source characteristics with additional informa-tion to determine the state of stress and fluid flow paths in the reservoir. From this information, they make productivity predictions, which help opera-tors develop and manage their reservoirs.

In this article, we review the acquisition, pro-cessing and interpretation of MS monitoring data. Advances in these areas are described, and workflows that integrate MS data into geome-chanical modeling and reduce interpretation uncertainty are presented. A case study from an unconventional reservoir in Arkansas, USA, illus-trates performance trade-offs for surface and downhole acquisition geometries. Case studies from Texas, USA, illustrate how MS monitoring has added value to stimulation operations by helping geoscientists identify fault interactions, fracture growth and stage-to-stage variability of stimulation responses.

Typical Monitoring SystemsMicroseismic monitoring (MSM) is the detec-tion of signals generated by small seismic—or microseismic—events. Engineers began using this technique during hydraulic fracturing in oil and gas operations as early as the 1980s.5 The Cotton Valley Consortium—a research group studying hydraulic fracturing of the Cotton Valley Formation play in Texas and Louisiana, USA—used microseismic monitoring to understand fluid flow in a Cotton Valley reservoir in 1997.6 Operators also successfully applied MSM in evalu-ating fracture stimulations in the Barnett Shale in Texas, which helped improve their understanding of the fracture network development during stim-ulations, avoid geohazards and enhance produc-tion.7 These early MSM operations incorporated arrays of three-component (3C) geophones or accelerometers deployed near reservoir depth in a nearby vertical monitor well (Figure 1).8

Figure 1. Hydraulic fracture monitoring from a vertical well. Multicomponent sensors in a vertical monitoring borehole record microseismic events caused by hydraulic fracturing (top). Event locations determined from data processing allow engineers to monitor the progress of stimulation operations. To acquire high-fidelity seismic data, the VSI versatile seismic imager (bottom) uses three-axis (x, y and z) geophone accelerometers (inset ) that are acoustically isolated from the tool body by isolation springs. The VSI service is mechanically coupled to the casing or formation by a hydraulically powered anchoring arm. The acquisition engineer can test the coupling quality by activating an internal shaker before operations begin. The VSI-40 40-shuttle versatile seismic imager allows up to 40 sensor packages, or shuttles, to be linked together; however, 12 shuttles are typically used in hydraulic fracture monitoring operations.

Oilfield Review MAY 16Microseismic Fig 1ORMAY 16 MCSMC 1

Anchoringarm

xz

y

Shaker Couplingcontacts

Threecomponent

accelerometers

Isolationspring

Shaker

xy

z

Reservoir

SensorsMicroseismic event

Stimulated volume

Treatment well Monitoring well

May 2016 19

Extensive hydraulic fracturing of horizontal wells began after 1997 as a result of Mitchell Energy’s successful application of the method in the Barnett Shale; MSM from adjacent horizontal boreholes soon followed. The use of sensor arrays in horizontal wells next led to the evaluation of MSM performance.

The effectiveness of the sensor array geometry depends on the layout of the monitoring and treatment wells. Monitoring from vertical wells close to treatment stages results in improved location accuracy for in-zone and out-of-zone microseismicity. Monitoring from nearby horizon-tal wells often provides coverage along the length of a stimulated lateral well; similar coverage may be unavailable from surface arrays or vertical monitor wells. The recording geometry may require cost trade-offs between monitoring the entire treatment well and detecting MS events that may occur outside the target interval.9 Microseismic events that are detected outside the targeted interval can indicate unintended consequences of the stimulation program such as breaching the reservoir seal or activating exist-ing faults.10

Knowledge of MS measurement accuracy is crucial for understanding the validity of MS data interpretations.11 Survey designers have devel-oped modeling software that predicts the mini-mum detectable event magnitude with respect to the distance of the monitoring array from source locations. The software also outputs estimates of the associated uncertainty for locating and char-acterizing MS events.12 Accuracy of the estimated event hypocenters—3D locations using easting and northing geographic Cartesian coordinates along with the depth of event initiation points—is affected by the monitoring geometry and the accuracy of the velocity model that is used to transform waveform arrival times at recording instruments to distances of the instruments from the events.

The precision of hypocenter estimates depends on the geophone array geometry and data errors, which influence the determination of event arrival time and the direction of arrivals at the receivers.13 During stimulation operations, extraneous, high-amplitude noise sources are numerous. As a conse-quence, the low signal-to-noise ratio (S/N) is one of the greatest challenges in the acquisition and processing of MS data.

Early MSM from a single monitoring well pro-vided valuable information, although it had shortcomings. Microseismic monitoring from single wellbores imposes the requirement that all multicomponent sensors have the same vector

fidelity—accuracy for measuring signal magni-tude and direction—because accurate waveform polarization information is crucial for determin-ing the direction to each event hypocenter.14 In addition, seismic tools must record incoming MS signals with the same spectral fidelity—accu-racy for measuring frequency content—within the typical signal bandwidth of 10 to 1,000 Hz used in these operations. When monitored from single wells during a multistage stimulation, some stages may be too distant from the sensors for reliable event detection or characterization. Sensor geometries along a single linear array are insufficient to determine the source mecha-nisms—size, direction, orientation and duration of 3D rock movements—associated with MS events; thus, microseismic engineers seek to record seismic waveforms from multiple observa-tion points and azimuths.

Valid interpretation of MS data also requires careful signal analysis. The processing of MS data is preceded by the construction and calibration of a model of seismic P-wave and S-wave velocities extending from the planned stimulated volume

of interest to the receiver array. Geophysicists calibrate the velocity model using perforation shot, string shot, checkshot or VSP survey data.15 Analysts originally performed event detection and localization processing using P-wave and S-wave arrival time picks and polarization from 3C MS waveforms.16 Today, event localization algorithms, such as the CMM coalescence microseismic mapping procedure, use an automatic scanning and grid search algorithm that correlates signal traveltimes and waveform polarizations to locate hypocenters.17 Multicomponent waveforms are processed to assess how well the observed tim-ing and polarization of arrival phases across the receiver array match the modeled values associated with potential hypocenter locations in the volume of interest. Arrival time picking can also be performed automatically and then refined manually.

Analysts interpret MS locations to show induced fracture extent—length, height and azimuth. However, stimulations in tight and unconventional reservoirs often produce com-plex, nonplanar hydraulic fracture geometries.

4. The radiation pattern is a description in 3D space of the amplitude and sense of initial motion of P and S wavefronts as they propagate away from the initiation position of a microseismic event. For more on seismic sources and their radiation patterns: Lay T and Wallace TC: Modern Global Seismology. San Diego, California, USA: Academic Press, 1995.

5. For more on the context for microseismic monitoring: Maxwell SC, Rutledge J, Jones R and Fehler M: “Petroleum Reservoir Characterization Using Downhole Microseismic Monitoring,” Geophysics 75, no. 5 (September–October 2010): 75A129–75A137.

6. The Cotton Valley formation is a Cretaceous-age tight sandstone that stretches from Texas to northern Florida, USA. The main play produces mostly natural gas and is located in north Louisiana and northeast Texas. For more on the Cotton Valley Consortium project: Rutledge JT, Phillips WS and Mayerhofer MJ: “Faulting Induced by Forced Fluid Injection and Fluid Flow Forced by Faulting: An Interpretation of Hydraulic-Fracture Microseismicity, Carthage Cotton Valley Gas Field, Texas,” Bulletin of the Seismological Society of America 94, no. 5 (October 2004): 1817–1830.

7. For more on the use of microseismic monitoring in the Barnett Shale: Maxwell S: “Microseismic: Growth Born from Success,” The Leading Edge 29, no. 3 (March 2010): 338–343.

8. Seismic data acquired from three-component (3C) geophones use three orthogonally oriented geophones or accelerometers. The early Schlumberger microseismic acquisition system typically included a VSI tool with eight 3C geophones.

9. For more on the accuracy of hypocenter estimates: Maxwell S and Le Calvez J: “Horizontal vs. Vertical Borehole-Based Microseismic Monitoring: Which is Better?,” paper SPE 131780, presented at the SPE Unconventional Gas Conference, Pittsburgh, Pennsylvania, USA, February 23–25, 2010.

10. Induced seismicity refers to earthquakes that are attributable to human activities, which may alter the local stresses and strains in the Earth’s crust and cause rock movements that generate earthquakes.

11. Maxwell, reference 7.

12. To predict sensor network performance, Schlumberger engineers used the NetMod microseismic survey design and evaluation software. For more on microseismic survey design: Raymer DG and Leslie HD: “Microseismic Network Design—Estimating Event Detection,” presented at the 73rd EAGE Conference and Exhibition, Vienna, Austria, May 23–26, 2011.

13. The hypocenter, or focus, is the point within the Earth at which rupture starts during an earthquake or microseismic event. The point directly above it on the Earth’s surface is the epicenter.

14. For more on vector fidelity: Berg EW, Rykkelid, Woje G and Svendsen Ø: “Vector Fidelity in Ocean Bottom Seismic Systems,” paper OTC 14114, presented at the Offshore Technology Conference, Houston, May 6–9, 2002.

15. In a checkshot survey, seismic specialists measure the traveltime of seismic waves, usually P-waves, from the surface to known receiver depths. A vertical seismic profile (VSP) is a more extensive survey in which geophones are placed at regular, closely spaced positions in the borehole. Both surveys use a seismic source positioned on the surface. Perforation shots and explosive string shots serve as seismic sources in the treatment well, and traveltimes are measured downhole or at the surface. In all cases, the source and receiver locations are known and, from the observed traveltimes, velocity may be calculated.

16. The P-waves used for seismic processing are elastic body waves, or sound waves, in which particles oscillate in the direction the wave propagates. The S-waves are elastic waves in which particles oscillate perpendicular to the direction in which the wave propagates.

17. Drew J, Bennett L, Le Calvez J and Neilson K: “Challenges in Acoustic Emission Detection and Analysis for Hydraulic Fracture Monitoring,” paper presented at the 17th International Acoustic Emission Symposium, Kyoto, Japan, November 9–12, 2004.

Drew J, Leslie D, Armstrong P and Michaud G: “Automated Microseismic Event Detection and Location by Continuous Spatial Mapping,” paper SPE 95513, presented at the SPE Annual Technical Conference and Exhibition, Dallas, October 9–12, 2005.

20 Oilfield Review

Consequently, geophysicists compute the effec-tive stimulated volume (ESV) as a measure of MS activity. The size, shape and extent of the ESV is based on a distribution of event locations and their uncertainties (Figure 2).18 The ESV provides information on the complexity of the hydraulic fracture network. A long, narrow ESV is probably dominated by a single planar through-going frac-ture, whereas a short, wide ESV probably consists of a complex, multibranching fracture network.

Taking the Broad View Engineers may conduct MSM from a single well, multiple wells, grids of shallow wells, surface arrays or networks of surface sensor patches. To meet acquisition goals, they may also combine a variety of designs (Figure 3).19 Typically, ana-lysts employ numerical simulation techniques that account for signal frequency content and attenuation and that make use of source, geol-ogy and noise models. They may use statistical analysis to predict the number of detectable events for given monitoring geometries. Analysts also recognize the importance of accounting for anisotropy in the velocity models. Seismic velocity and attenuation tomography based on crosswell surveys can be used to constrain these models.20 During perforating, data acquired from surface sensors can provide calibrated P-wave traveltimes.21

Monitoring from surface and near-surface positions offers a potentially larger field of view than that from monitoring wells alone, and it eliminates the need for providing dedicated deep monitoring wells.22 Surface monitoring enables long treatment laterals to be monitored along their entire lengths. However, because the S/N is often low, locating and characterizing MS events using data recorded at the surface may be diffi-cult. To overcome low S/N and detection uncer-tainty, survey designers use receiver arrays containing many hundreds to thousands of sen-sors. Data from multiple points can be processed to reduce noise and accentuate the true signal. Aided by recent improvements in signal process-ing, geophysicists can use these monitoring arrays to map microseismic events more completely over extended stimulations than is possible from an array placed in a single monitoring well.

When nearby observation wells are available, downhole monitoring offers proximity to treat-ment well stages and ensures higher S/N than that offered by surface monitoring. Broad bandwidth signals recorded by downhole arrays often retain more high-frequency content than do surface arrays. This high-frequency content is useful for MS event characterization. Recording P-wave and S-wave arrivals downhole using 3C sensors also improves localization accuracy compared with that from surface recordings of P-waves alone.

Monitoring from multiple wells provides observations of the source position from multiple directions and enables more-complete source characterization compared with that from sin-gle-well monitoring. Downhole monitoring requires correct velocity models to reduce event localization uncertainty, and it requires precise well deviation surveys to determine exact receiver positions. The models must also contain accurate values for Qp and Qs, the quality factors related to P-wave and S-wave attenuation during propagation. These factors are used to deter-mine wave amplitudes at the MS hypocenters and reduce uncertainty in the inversion for the source mechanism.23

Modern signal processors use nonlinear mathematical methods for the detection and localization of MS events. In combination with CMM processing, these mathematical methods have the potential to detect and locate weak MS events automatically without prior knowledge of the source mechanism and its radiation pattern.24 Analysts have also extended event localization algorithms to use the full MS waveforms. Early event localization methods used traveltimes and polarizations of the P-wave and S-wave direct arrivals only. However, geoscientists can use waveform synthetics to model the source time functions, the principal features of the waveform time series recorded by each instrument in the

Figure 2. Microseismicity and effective stimulated volume. Located events (circles, color-coded according to time) were generated during the stimulation of a horizontal well (red line) in the Barnett Shale in Denton County, Texas. Analysts built 3D cells in a model of the monitored volume of the reservoir. They counted the number of events that exceeded a predetermined threshold in each cell and calculated the resulting stimulated volume within the cells. The green envelope gives the effective stimulated volume, estimated here at 180 million ft3 [5 million m3]. Yellow and blue disks on the horizontal well mark perforation clusters for stimulation Stages 1 and 2, respectively. Isolated events shown outside the green envelope are not considered hydraulically connected to the stimulated volume. (Adapted from Le Calvez et al, reference 59.)

Oilfield Review MAY 16Microseismic Fig 2ORMAY 16 MCSMC 2

Y-axis S–N, ftSouth North

Top of Barnett Shale

Bottom of Barnett Shale

Z-ax

isDe

pth,

ft

7,130,0007,120,000

Tim

e, h

r:min

17:00

18:00

19:00

20:00

21:00

22:00

7,200

7,600

8,000

8,400

May 2016 21

sensor arrays. Analysts then extract arrival times of direct, refracted and reflected P-wave and S-wave arrivals. An extension of the CMM method uses these additional arrivals to identify their energy for event detection and characterization.25

Tracking Microseismicity to the SurfaceIn 2011, Schlumberger and an independent operator acquired a comprehensive MS dataset while monitoring hydraulic fracturing operations in the Fayetteville Shale in Arkansas. Completion engineers stimulated two horizontal wells using a zipper fracture method, in which hydraulic fracturing is conducted sequentially in side-by-side wells.26 Concurrently, MS survey engineers conducted a test to assess and quantify the event detection capabilities, accuracy and resolution of surface, near-surface and downhole acquisi-tion systems. Sixteen stimulation stages were

Figure 3. Sensor network deployment options. Sensors for MS monitoring of hydraulic fracturing may be deployed in vertical (1), horizontal (2) or deviated monitoring boreholes. Survey engineers may use a grid of shallow wells (3) containing arrays of multicomponent sensors. On the surface, they may deploy single component or multicomponent geophones in 2D patches or in extensive linear arrays (4). Sensor networks that record MS waveform data over a broad area provide data that can be used to characterize MS

event compression (red ellipsoids) and dilation (blue ellipsoids) radiation patterns (5) and estimate source mechanisms. Schlumberger engineers use the MS Recon high-fidelity microseismic surface acquisition system to acquire MS data at the surface. The system incorporates proprietary geophone accelerometers (6) (inset ), ultralow-noise electronics and a nodal-based wireless acquisition technology (7). (Adapted from Le Calvez et al, reference 19.)

Oilfield Review MAY 16Microseismic Fig 3ORMAY 16 MCSMC 3

1

2

5

43

6

7

18. Effective stimulated volume (ESV), also referred to as stimulated reservoir volume, is an estimate of the total rock volume affected by the hydraulic fracture stimulation.

19. For more on survey design: Le Calvez J, Underhill B, Raymer D and Guerra K: “Designing Microseismic Surface, Grid, Shallow and Downhole Surveys,” Expanded Abstracts, 85th SEG Annual International Meeting and Exposition, New Orleans (October 18–23, 2015): 2645–2649.

20. For more on the use of crosswell surveys: Le Calvez J, Marion B, Hogarth L, Kolb C, Hanson-Hedgecock S, Puckett M and Bryans B: “Integration of Multi-Scale, Multi-Domain Datasets to Enhance Microseismic Data Processing and Evaluation,” paper BG08, presented at the 3rd EAGE Workshop on Borehole Geophysics, Athens, April 19–22, 2015.

21. For more on the use of perforation shots to build velocity models: Probert T, Raymer D and Bradford I: “Comparing Near-Surface and Deep-Well Microseismic Data and Methods for Hydraulic Fracture Monitoring,” paper PS07, presented at the 4th EAGE Passive Seismic Workshop, Amsterdam, March 17–20, 2013.

22. For more on surface microseismic monitoring: Duncan PM and Eisner L: “Reservoir Characterization Using Surface Microseismic Monitoring,” Geophysics 75, no. 5 (September–October 2010): 75A139–75A146.

23. For more on uncertainty in microseismic monitoring: Eisner L, Thornton M and Griffin J: “Challenges for Microseismic Monitoring,” Expanded Abstracts, 81st SEG Annual International Meeting and Exhibition, San Antonio, Texas, USA (September 18–23, 2011): 1519–1523.

24. For more on nonlinear processing methods: Özbek A, Probert T, Raymer D and Drew J: “Nonlinear Processing Methods for Detection and Location of Microseismic Events,” paper Tu 06 06, presented at the 75th EAGE Conference and Exhibition, London, June 10–13, 2013.

25. For more on full MS waveform processing: Williams MJ, Le Calvez JH and Gendrin A: “Using Surface and Downhole Data to Drive Developments in Event Detection Algorithms,” Extended Abstracts, 76th EAGE Conference and Exhibition, Amsterdam, June 16–19, 2014.

26. Zipper fracture, or “simul-frac,” is a technique in which two or more parallel wells are drilled, perforated and stimulated via an alternating sequence of stages. This stimulation method results in a high-density network of fractures between the wells that increases production in both wells.

For more on advances in fracturing technology: Rafiee M, Soliman MY and Pirayesh E: “Hydraulic Fracturing Design and Optimization: A Modification to Zipper Frac,” paper SPE 159786, presented at the SPE Eastern Regional Meeting, Lexington, Kentucky, USA, October 3–5, 2012.

22 Oilfield Review

monitored across a reservoir interval at 3,600 ft [1,100 m] TVD.

Survey engineers deployed a wide-aperture borehole seismic array that extended from the reservoir to the surface (Figure 4). This array acquired an MS dataset at the reservoir level as well as data that revealed how signals propagated and how noise levels varied between the reser-voir and the surface. Engineers acquired addi-tional MS data from a deep horizontal well and from five shallow vertical wells, each containing a seismic array. They also recorded MS data using an extensive surface seismic array that consisted of five radial lines fanning and emanating from the treatment wellhead, two parallel lines that crossed the radial lines and three 2D surface patches that were located about 3,000, 5,000 and

Figure 4. Fayetteville Shale MSM operation. The map view (top) shows the sensor network layout for a Fayetteville Shale stimulation. The 4,100-channel surface seismic array consisted of five radial lines (red, Lines 1 through 5) offset and emanating from the treatment wellhead, two crosslines (red, Lines 6 and 7) and three areal 2D patches (green squares). Data were also acquired using sensors deployed in a deep horizontal borehole (yellow), in one deep monitoring well (green) and in five vertical shallow wells (blue circles).The vertical section view (bottom) shows well trajectories used for the MS monitoring. The trajectories of treatment Wells 1H and 2H are shown in gray and yellow, respectively. Well M (green) is shown along with the vertical monitoring wells (blue). (Adapted from Schilke et al, reference 28.)

Oilfield Review MAY 16Microseismic Fig 4ORMAY 16 MCSMC 4

Dept

h, ft

1,000 ft

Well 1H

Well 2H

Well M

N

Line 4

Line 5Line 6

Line 7

Line 3

Line 2

Line 1

Well 2H

Well 1H

Well 3

Well M

Well 2

Well 5

Well 1

Well 4

Treatment well

Downhole array

N Shallow well2D patch

1,000 ft

8,000 ft [915, 1,520 and 2,440 m] from the treat-ment wellhead. Time synchronization between all recording systems ensured that the same MS events could be identified on all monitoring sys-tems (Figure 5).27

Data from this comprehensive test allowed analysts to compare the effectiveness of near-surface and downhole hydraulic fracture moni-toring. Analysts observed that surface array line segments can mitigate surface-wave noise from a known source, such as treatment wellhead pumps, but are less effective against distributed or moving sources of noise, which may be domi-nant in areas covered by the array.28 Surface patches—2D arrays of closely spaced sensors—can effectively remove noise coming from multi-

ple directions but cannot cover the same distances as can linear arrays.29 Sensor arrays in shallow wells are less sensitive to noise propagat-ing along the surface, but signal processing that discriminates against noise is hampered by the small number of sensors available in these arrays.30 Surface and near-surface array designs may be adapted to known noise conditions but are constrained by land access, environmental effects and cost concerns.

The results of the Fayetteville survey showed that a downhole array could detect MS events from nearby stage treatments better than from other array geometries. But the downhole array suffered reduced sensitivity and increased loca-tion uncertainty for distant stage treatments and events. Surface and near-surface monitoring,

May 2016 23

27. For more on the Fayetteville Shale MS test: Maxwell SC, Raymer D, Williams M and Primiero P: “Tracking Microseismic Signals from the Reservoir to Surface,” The Leading Edge 31, no. 11 (November 2012): 1300–1308.

Peyret O, Drew J, Mack M, Brook K, Maxwell S and Cipolla C: “Subsurface to Surface Microseismic Monitoring for Hydraulic Fracturing,” paper SPE 159670, presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, October 8–10, 2012.

28. For more on surface array performance: Schilke S, Probert T, Bradford I, Özbek A and Robertsson JOA: “Use of Surface Seismic Patches for Hydraulic Fracture Monitoring,” paper We E103 04, presented at the 76th EAGE Conference and Exhibition, Amsterdam, June 16–19, 2014.

although less sensitive to recording deep signals than downhole monitoring, offers more uniform sensitivity over a wider area. Surface patches are easily deployed over wide areas and may ulti-mately become the preferred surface monitoring configuration. However, their successful use will require the recording of sufficient signal and the effective application of noise attenuation meth-ods. The lessons learned in the test provide the planners of future MSM systems with a clearer understanding of the trade-offs involved when specifying acquisition equipment layouts.

Figure 5. Data record from one microseismic event. A microseismic event was detected across the downhole, near-surface and surface sensor arrays during a test in the Fayetteville Shale. The modeled waveforms (top, purple) from the event are shown along with sensor positions (green). Waveforms propagate from the event’s hypocenter, the location of which was estimated from the data. Recorded waveforms from the event are shown from the vertical seismic array (middle left ), the array in the horizontal monitoring well (bottom left ) and the 3C arrays in the five shallow vertical MSM wells (middle right ). Five stacked traces from the vertical component waveforms recorded on the five surface radial lines are also shown (lower right ). (Adapted from Peyret et al, reference 27.)

Oilfield Review MAY 16Microseismic Fig 5ORMAY 16 MCSMC 5

MSM surface linesS N

Horizontal array

Verticalarray

MSM wells

Vertical Seismic Array

Horizontal Seismic Array

MSM Wells

MSM Stacked Data

29. For more on signal processing approaches applied to surface patch data: Petrochilos N and Drew J: “Noise Reduction on Microseismic Data Acquired Using a Patch Monitoring Configuration: A Fayetteville Formation Example,” Expanded Abstracts, SEG 84th Annual International Meeting and Exposition, Denver (October 26–31, 2014): 2314–2318.

30. The amplitude of seismic surface waves decreases as its position away from the surface increases. For more on noise sources in the Fayetteville Shale test: Drew J, Primiero P, Brook K, Raymer D, Probert T, Kim A and Leslie D: “Microseismic Monitoring Field Test Using Surface, Shallow Grid and Downhole Arrays,” paper SEG 2012 0910, presented at the 82nd SEG Annual International Meeting and Exposition, Las Vegas, Nevada, USA, November 4–9, 2012.

24 Oilfield Review

Making Sense of the Microseismic Cloud Geophysicists studying earthquakes use charac-teristics such as the seismic moment, moment magnitude, stress drop, stress change and source dimensions to describe the physical processes occurring at earthquake hypocenters.31 In earth-quake seismology, a typical earthquake is caused by shear displacement—surface parallel slip—along a preexisting fault plane. Earthquake intensity is related to the seismic moment, MO, which can be determined by measuring the amplitudes of seismic waves generated during the event (Figure 6).32

In 1977, Japanese seismologist Hiroo Kanamori used the relationship between seismic moment and energy to introduce the moment magnitude (Mw) scale. Today, geophysicists in the oil and gas industry are applying earthquake seismology concepts to analyze MS data; moment magnitude is routinely used to characterize the size of MS events. Individual source dimensions such as incremental fracture surface areas and lengths can be estimated for MS events from their waveform spectra and source models.

The distribution of MS source locations pro-vides an indication of the rock volume affected by the hydraulic stimulation. Reservoir engineers initially related, with some success, well produc-tivity to MS activity using the ESV as a measure of the volumetric extent of reservoir stimulation (Figure 7).33 Pinpointing the location of MS events, however, is sometimes insufficient for accurately predicting reservoir performance. This insufficiency may result from inaccurate

estimates of the event density and the actual extent of the ESV.

When determining ESV, geoscientists must take into account the spatial density of intercon-nected fractures within the stimulated volume and their surface area in contact with the reservoir. The set of MS event locations—the microseismic cloud—may include stress-induced events in nonhydraulically connected areas. Thus,

the active volume of microseismicity may be an overestimation of the hydraulically connected vol-ume. To reduce this uncertainty, some analysts consider how many events occur in the neighbor-hood of each event location when computing ESV. For MSM that has limited array coverage, distant low-amplitude events may not be detected. This phenomenon, referred to as monitoring bias, may also reduce the computed ESV.

Figure 6. Seismic moment, energy and magnitude equations. The seismic moment, Mo, of an earthquake source is defined as the product of the shear modulus (µ) of the host rock that the fault cuts through, the average shear displacement (D) along that surface and the affected fault surface area (A). The amplitudes of emitted seismic waves are directly proportional to the seismic moment. Seismologists have also related seismic moment to the seismic energy (Es), which is radiated when a fault slips, resulting in a change in the static shear stress (Δσs) along the fault. The moment magnitude (M w) is computed from Mo and is a logarithmic measure of the energy released during a seismic event. In this equation, Mo is expressed in units of N.m. When expressed in units of dyne.cm, 10.73 is used instead of 6.06; for units of lbf.ft, the constant is 5.97.

Oilfield Review MAY 16Microseismic Fig 6ORMAY 16 MCSMC 6

Mo μ D A.

Mw 2/3 (log Mo) – 6.06.

Es = 0.5 (Δσs / μ) Mo.

Figure 7. Eagle Ford Shale effective stimulation. The map view (bottom) shows MS events (colored dots), effective stimulated volume (ESVs, colored opaque envelopes) and production log results (dashed red lines) for the stimulation of a horizontal well (dashed blue line) in the Eagle Ford Shale in Texas. The ESVs were calculated based on the density and magnitude of MS events for each perforated interval. The length of the bisecting red lines from the production data are related to the contribution of individual perforation clusters to the total flow of hydrocarbons. Engineers observed a definite correlation (top) between production contribution from individual perforated intervals (red circles) and the ESV derived from the hydraulic fracture model analysis. For the plotted data, R2 is a linear regression measurement related to the quality of curve fitting. A value of 0.80 indicates a good fit. (Adapted from Inamdar et al, reference 33.)

Oilfield Review MAY 16Microseismic Fig 7ORMAY 16 MCSMC 7

876,000

1,538,000 1,539,000 1,540,000 1,541,000

877,000

878,000

Y-ax

is, f

t

X-axis, ft

879,000

0

2

4

400 500 600 700 800 900 1,000

6

8

10

12

14

16

18

ESV, 1,000 ft3

R2 = 0.80

Prod

uctio

n, %

Monitoring well

Productiondata

May 2016 25

Microseismic events are produced when rapid deformation occurs within the reservoir or sur-rounding formations in response to stress changes arising from increased pressure during fracture stimulation operations. The deformation consists of slip of unknown length along failure planes of unknown area and orientation.34 Analysts estimate the seismic moment of indi-vidual events using the amplitudes and frequency content of received seismic waveforms.35 Geophysicists can use the seismic moment to enhance the interpretation of MS data.36 By sum-ming these moment values over time for all events within distinct spatial volumes or grid cells, analysts obtain the cumulative moment as a function of time and space.

Engineers may compare time series of cumu-lative moment and stimulation treatment data to better understand the stimulation process (Figure 8). An increase in cumulative moment over time indicates progressive deformation. Maps of final values of cumulative moment indi-cate the spatial distribution of seismic deforma-tion observed during the stimulation. Because large, detectable events contribute significantly more moment than numerous, small, undetect-able events, cumulative moment provides a mea-sure of stimulation response that is less sensitive to monitoring bias than is an ESV based on event locations alone. Using 3D mapping of seismic moment or cumulative moment provides insight

into fracture behavior during stimulations; such insights can be used to calibrate complex hydrau-lic fracture models.

Borrowing concepts from earthquake seis-mology, analysts use statistical measures such as b-values and D-values to further describe groups of detected MS events.37 The relative frequency of occurrence of earthquakes over a range of magni-tudes is described by b-values. Many more small magnitude events tend to occur than do large ones, and the b-value quantifies this tendency. The statistics of the distances separating earth-quake hypocenters is described by D-values. Populations of events occurring on the same frac-ture and fault planes tend to present characteris-tic distributions of spatial separation.

31. For more on source parameters: Shearer PM: Introduction to Seismology, 2nd ed. Cambridge, England: Cambridge University Press, 2009.

32. Seismic moments range from 105 N.m [105 lbf.ft] in the case of the smallest detectable microearthquakes to 1023 N.m [1023 lbf.ft] in the case of great earthquakes.

33. For more on ESV versus reservoir production: Inamdar A, Malpani R, Atwood K, Brook K, Erwemi A, Ogundare T and Purcell D: “Evaluation of Stimulation Techniques Using Microseismic Mapping in the Eagle Ford Shale,” paper SPE 136873, presented at

Figure 8. Cumulative seismic moment. Engineers plotted cumulative seismic moment (black) along with the MS event rate (gray vertical bars) and pumping parameters to help them understand fracture stimulation job performance during the treatment of a well in the Eagle Ford Shale. The pump rate (blue), surface pressure (red) and proppant concentration (green) are shown. Analysts use these plots to identify the time-dependent response of MS events to the stimulation. An abrupt increase in cumulative seismic moment indicated that deformation increased significantly about halfway through the planned pumping schedule. By comparing multiple treatments, engineers can determine how microseismicity changes in response to adjustments to the pumping schedule and whether it is consistent across stages. (Adapted from Downie et al, reference 34.)

Oilfield Review MAY 16Microseismic Fig 8ORMAY 16 MCSMC 8

00

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

5

10

15

20

25

30

35

40

45

50

55

60

6,000

8,000

6,200

8,200

8,400

6,400

6,600

6,800

7,000

7,200

7,400

7,600

7,800

1 am 2 am 3 am 4 am 5 am

Time of day

Pressure

Trea

tmen

t pre

ssur

e, p

si

Pum

p ra

te, b

bl/m

in

Prop

pant

con

cent

ratio

n m

easu

red,

lbm

/gal

US

1.3

1.2

1.1

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

00

5

10

15

20

25

30

35

40

45

50

55

Cum

ulat

ive

mom

ent,

N.m

× 1

06

Even

t rat

e, a

ccep

ted

even

ts p

er 2

min

60

Pump rate

Proppant concentration

Event rate

Cumulative moment

the SPE Tight Gas Completions Conference, San Antonio, Texas, November 2–3, 2010.

34. Downie R, Xu J, Grant D, Malpani R and Viswanathan A: “Utilization of Microseismic Event Source Parameters for the Calibration of Complex Hydraulic Fracture Models,” paper SPE 163873, presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, February 4–6, 2013.

35. Seismic sources are formally treated as “displacement discontinuities” to describe the difference in motion of material on opposing faces of fracture surfaces. This motion need not be parallel to the surfaces.

36. Analysts use source models, formation material properties and measured frequency spectra to estimate the fracture surface area of individual events. Slip lengths can then be inferred from seismic moment estimates.

37. For more on b-values and D-values: Grob M and van der Baan M: “Inferring In-Situ Stress Changes by Statistical Analysis of Microseismic Event Characteristics,” The Leading Edge 30, no. 11 (November 2011): 1296–1301.

26 Oilfield Review

In seismology, the Gutenberg-Richter empiri-cal law relates the magnitude of earthquakes, M, to their frequency of occurrence, N (Figure 9). In microseismic studies, geoscientists have substi-tuted moment magnitude, Mw, in this relation and have explored how hydraulic fracture param-eters affect the values of a and b, the slope and

intercept, respectively, of the log10 N versus Mw relationship. Some evidence suggests a possible relationship between the value of a—the num-ber of events at the intercept at Mw equaling 1—and the pump rates used in hydraulic fracturing, for example, when the cumulative volume of pumped fluid influences the microseismic event

rate.38 Global seismology data often show the slope b to be about 1 for tectonic earthquakes. In some geologic settings, MS interpreters use b-values to distinguish failure along naturally occurring faults from that along hydraulically induced fractures.

Determination of b-values may also provide completion engineers with an indication of stress changes over the course of multistage stimula-tions. During hydraulic fracturing treatments, b-values greater than 1 are typically observed, whereas b at about 1 has been observed dur-ing MS episodes dominated by movement along faults.39 Scientists have observed relationships between the b-value and local stress conditions.40 Some studies of MS data have shown variations in b-values within regional shale formations being stimulated.41 Other studies have shown that b-values may be time dependent and vary as stress changes throughout the stimulation process.42

Seismologists use D-values to convey the spa-tial statistics of earthquake hypocenter occur-rence. Computed from event locations, D-values may be used to summarize interevent distance statistics. If the cloud of events maps onto a point, D is expected to equal 0. A D-value of 1 is expected if the geometric distribution is linear, 2 if planar and 3 if dispersed.43 The distribution of MS hypocenters has the potential to reveal the location of interconnected fracture surfaces, and analysts have developed a variety of techniques to extract linear and planar features from the microseismic cloud.44

In one method, D-values were computed sepa-rately around each detected MS event location (Figure 10). Closely spaced events are more likely to show some random scatter due to processing effects (in which D equals approximately 3). Events occurring on the same fracture and fault plane will tend to align (in which D equals approx-imately 2). Interpreters identified linear and pla-nar structures in the data by selecting only events for which the D-value is less than or equal to 2. Analysts have used changes in both b-values and D-values to infer stress changes in the reservoir.

Applying additional concepts from earth-quake seismology, geophysicists measure P-wave and S-wave amplitudes across broad receiver net-works to determine radiation patterns and then invert them to estimate seismic moment tensors, which describe the orientation, magnitude and slip of individual MS events.45 For MSM, geophysi-cists use moment tensor inversion (MTI), which is an advanced seismic processing technique, to provide information about the mechanism of fail-ure at fracture sites.46 They then decompose each moment tensor into its constituents to estimate

Figure 9. Plot of b-values. In seismology, the Gutenberg-Richter empirical law relates the magnitude (M ) of earthquakes to their frequency of occurrence, N, where N is the number of events of magnitude M or greater, and a and b are constants. Events observed during a stimulation stage in the Barnett Shale in Denton County, Texas (red dots), show that the cumulative frequency is equal to N divided by 336, the total number of detected events. A statistical method has been used to estimate b, which accounts for the limited ability to detect low-magnitude events. Because data from small magnitude events cannot be recorded reliably, events that have magnitudes smaller than the magnitude of completeness, Mc (dashed line), are not used in the calculation because the S/N is too low. (Adapted from Williams et al, reference 61.)

Oilfield Review MAY 16Microseismic Fig 9ORMAY 16 MCSMC 9

Cum

ulat

ive

frequ

ency

Mag

nitu

de o

f com

plet

enes

s, M

c

Moment magnitude

Events

Gutenberg-Richter relationshiplog10 N = a – b × M.

0.01

–2.8 –2.6 –2.4 –2.2 –2.0 –1.8 –1.6 –1.4 –1.2 –1.0

0.1

1.0

Figure 10. Plot of D-values. The D-value is estimated from the slope of the line of event locations during a stimulation in the Barnett Shale in Denton County, Texas. The cumulative distribution of interevent distances, C(l), is the total number of MS events (purple circles) that have hypocenters separated by a distance of l meters or less. The circles represent measured data and show a good match with the modeled data (dashed black line), which corresponds to D = 1.86. This D-value indicates a near-planar distribution of event locations for a range of intermediate distances around the event. The closer and more distant events trend away from straight-line behavior. (Adapted from Williams et al, reference 61.)

Oilfield Review MAY 16Microseismic Fig 10ORMAY 16 MCSMC 10

–2.0

–3.0

–4.0

–5.01.2 1.4 1.6 1.8 2.0 2.2 2.4

–2.5

–3.5

–4.5

log 1

0 (C

(l)) log10 (C) D log10 (l)

log10 (l)

May 2016 27

38. Shapiro SA, Dinske C, Langenbruch C and Wenzel F: “Seismogenic Index and Magnitude Probability of Earthquakes Induced During Reservoir Fluid Stimulations,” The Leading Edge 29, no. 3 (March 2010): 304–309.

39. Cipolla C, Maxwell S and Mack M: “Engineering Guide to the Application of Microseismic Interpretations,” paper SPE 152165, presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, February 6–8, 2012.

40. Schorlemmer D, Wiemer S and Wyss M: “Variations in Earthquake-Size Distribution Across Different Stress Regimes,” Nature 437 (September 22, 2005): 539–542.

Downie RC, Kronenberger E and Maxwell SC: “Using Microseismic Source Parameters to Evaluate the Influence of Faults on Fracture Treatments—A Geophysical Approach to Interpretation,” paper SPE 134772, presented at the SPE Annual Technical Conference and Exhibition, Florence, Italy, September 19–22, 2010.

41. Boroumand N: “Hydraulic Fracture b -Value from Microseismic Events in Different Regions,” presented at the GeoConvention 2014, Calgary, May 12–16, 2014.

42. Zorn EV, Hammack R and Harbert W: “Time Dependent b and D-values, Scalar Hydraulic Diffusivity, and Seismic Energy From Microseismic Analysis in the Marcellus Shale: Connection to Pumping Behavior During Hydraulic Fracturing,” paper SPE 168647, presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, February 4–6, 2014.

43. Grob and van der Baan, reference 37.44. Williams MJ, Khadhraoui B and Bradford I: “Quantitative

Interpretation of Major Planes from Microseismic Event Locations with Application in Production Prediction,” Expanded Abstracts, 80th SEG Annual International Meeting and Exposition, Denver (October 17–22, 2010): 2085–2089.

the relative proportion of each failure mode—such as shear slip, tensile opening, expansion or other process or a combination of them—and the orientation of local fracture planes and the direc-tion of shear slip. Recently, mathematicians have developed theoretical extensions of MTI in terms of potency tensors; such extensions display unique fracture planes and displacement vectors (Figure 11).47

Although the deformation represented by MS signals constitutes a small fraction of the total deformation and fracture volume created dur-ing stimulation, MTI processing holds promise to provide insights into natural fracture characteris-tics and local stress fields. Geophysicists extract planar features for input to construct discrete fracture networks (DFNs), which represent the distribution, orientation, shape, connectedness and fluid flow properties of a population of frac-tures. The MTI results may provide constraints to help build these networks and calibrate hydraulic fracture models.48

Integrating MS Data and Geomechanical Modeling Engineers integrate MS data with geologic mod-els, mechanical earth models (MEMs), formation imaging logs and production logs to characterize reservoirs and aid their understanding of micro-seismicity. In the past, engineers predicted future reservoir production based on correlations between poststimulation production and ESV. Analysts now use geomechanical modeling to enhance the forecasting.

Geomechanics can aid in the design of hydraulic stimulations to maximize the hydraulic fracture surface area exposed to the reservoir and to the system of natural fractures within. For planning wells and determining in situ stress states, engineers perform stress simulations using static and time-lapsed 3D MEMs that are integrated with results from reservoir simulation models.49 Geomechanical modeling can provide insight into the extent of fracture-to-fracture interference between fracture stages in a treat-ment well, nearby treated wells and the natural fracture system.

Knowledge of the regional stress state and the characteristics and distribution of natural frac-tures in the reservoir is important for predicting the effectiveness of reservoir stimulations. During stimulation operations, hydraulic frac-tures interact with preexisting natural fractures. Slip along natural fractures generally increases the permeability of the stimulated fractures. The density and orientation of the natural fracture population are significant factors that influence

Figure 11. Source mechanism expansion, opening and slip. The estimated moment tensor for each event detected during a well stimulation stage (top) has been decomposed into expansion, opening and slip components and displayed as glyphs. For reference, the well (magenta) is shown with perforation clusters (red disks). A glyph is composed of two disks and a wireframe sphere superimposed over them (bottom). The wireframe sphere represents expansion if red or contraction if blue. The thickness of the disks represents opening, and their relative displacement represents the degree of slip. The glyph’s central plane, which is parallel to the disks’ planar surfaces, is oriented with respect to the strike and dip of the fracture plane activated or caused by the event. (Adapted from Leaney et al, reference 47.)

Oilfield Review MAY 16Microseismic Fig 11ORMAY 16 MCSMC 11

Expansion Opening Slip

Perforationcluster

Well

45. By analyzing the amplitudes of waveforms received at an array of recording sensors, geophysicists can determine the location, shape, size and orientation of the motions of the causative event. Geophysicists then use the amplitude data to invert for the moment tensor, a system of point-force couples, which is the best-fit seismic radiation pattern equivalent to that observed from seismic event displacement discontinuities. For more on radiation patterns and moment tensors: Lay and Wallace, reference 4.

46. Moment tensor inversion (MTI) is now integrated with the Petrel software platform and Mangrove reservoir-centric stimulation design software workflows.

47. For more on the theory, decomposition and display of moment tensors: Leaney S, Chapman C and Yu X: “Anisotropic Moment Tensor Inversion, Decomposition and Visualization,” Expanded Abstracts, 84th SEG Annual International Meeting and Exposition, Denver (October 26–31, 2014): 2250–2255.

48. For more on deriving DFN from MS data: Yu X, Rutledge JT, Leaney SW and Maxwell S: “Discrete-Fracture-Network Generation from Microseismic Data by Use of Moment-Tensor- and Event-Location-Constrained Hough Transforms,” paper SPE 168582, presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, February 4–6, 2014.

49. Schlumberger reservoir engineers couple ECLIPSE 3D simulations with the VISAGE finite-element geomechanics simulator to create dynamic, time-lapse models of stress and production history of single and multiple wells and fields. For more on integrated modeling: Alexander T, Baihly J, Boyer C, Clark B, Waters G, Jochen V, Le Calvez J, Lewis R, Miller CK, Thaeler J and Toelle BE: “Shale Gas Revolution,” Oilfield Review 23, no. 3 (Autumn 2011): 40–55.

28 Oilfield Review

the stimulated fracture network development and control reservoir productivity.50

Hydraulic fracturing creates a tensile frac-ture that opens slowly, and most of the rock deformation occurs aseismically at much lower frequencies than the typical MS signal band.51 Shear deformation occurs in the process zone around the fracture tip, in the vicinity of the frac-ture face as a result of leakoff into preexisting natural fractures and at doglegs and other geo-metric deflections. In contrast to the aseismic nature of tensile dilation, shear deformation often emits sudden, high-frequency, audible seis-mic energy.

Reservoir engineers have developed a variety of approaches to characterize poststimulation fracture networks (Figure 12).52 In one approach, analysts use information from seismic reflection surveys, well logs and cores to build a DFN, which they combine with a set of earth models to describe the reservoir and surrounding forma-tions.53 Fracture information may often be derived from resistivity or ultrasonic image logs (Figure 13). Hydraulic fracture models such as

the UFM unconventional fracture model can then be used to predict the fracture geometries that result from the stimulation.54

Modelers can use treatment data such as pump pressure, fluid volumes and proppant loadings as inputs to the numerical calculations and then use MS data to constrain the results. These UFM simu-lations yield predictions of stimulated fracture geometry and conductivity distributions. Analysts iteratively calibrate the model by adjusting the input parameters to the UFM simulation to achieve a match between fracture geometry predictions and observed MS event locations and deformation, or seismic moments (Figure 14). Adjustable parameters include horizontal stresses and prop-erties of the DFN and fracturing fluid. Analysts use volumes of fluids pumped and stresses to constrain the fracture model to reduce the set of possible solutions. They conduct sensitivity analyses to determine how much the answer changes as input parameters are varied and to ensure that non-unique multiple solutions give results that are reasonable in terms of predicted production.

Reconciling the UFM modeling results with the MS event patterns requires the evaluation of multiple DFN realizations and may not result in an exact match but rather a statistically probable match to the MS pattern. After the UFM simula-tions are calibrated, the fluid flow characteristics predicted by the UFM technique can be incorpo-rated in a reservoir simulation. In this process, the hydraulic fractures are explicitly gridded in the reservoir model to honor the 3D hydraulic fracture geometry and proppant distribution.

In addition to the reservoir grid, a fluid model, a set of relative permeabilities, stress-dependent hydraulic fracture conductivity profiles, histori-cal production rates and bottomhole pressure are input into the reservoir simulator. After the res-ervoir models are calibrated, analysts can use them to forecast hydrocarbon recovery and per-form sensitivity analyses. The analysts can vary completion parameters, including the number of stimulation stages, the number of perforation clusters per stage and reservoir parameters such as permeability, porosity and saturations to maxi-mize future stimulation operations.

50. For more on fracture network development: Johri M and Zoback MD: “The Evolution of Stimulated Reservoir Volume During Hydraulic Stimulation of Shale Gas Formations,” paper SPE 168701/URTeC 1575434, presented at the Unconventional Resources Technology Conference, Denver, August 12–14, 2013.

51. Maxwell SC and Cipolla C: “What Does Microseismic Tell Us About Hydraulic Fracturing?,“ paper SPE 146932, presented at the SPE Annual Technical Conference and Exhibition, Denver, October 30–November 2, 2011.

52. For more on an early, computationally efficient geomechanical modeling approach: Xu W, Le Calvez J and Thiercelin M: “Characterization of Hydraulically-Induced Fracture Network Using Treatment and Microseismic Data on a Tight-Gas Formation: A Geomechanical Approach,” paper SPE 125237, presented at the SPE Tight Gas Completions Conference, San Antonio, Texas, June 15–17, 2009.

53. For more on the construction of DFNs: Will R, Archer R and Dershowitz B: “Integration of Seismic Anisotropy and Reservoir Performance Data for Characterization of Naturally Fractured Reservoirs Using Discrete Feature Network Models,” paper SPE 84412, presented at the SPE Annual Technical Conference and Exhibition, Denver, October 5–8, 2003.

Offenberger R, Ball N, Kanneganti K and Oussoltsev D: “Integration of Natural and Hydraulic Fracture Network Modeling with Reservoir Simulation for an Eagle Ford Well,” paper SPE 168683/URTeC 1563066, presented at the Unconventional Resources Technology Conference, Denver, August 12–14, 2013.

54. UFM processing is embedded in the Mangrove software platform and uses the output of the VISAGE simulator. For more on UFM processing: Weng X, Kresse O, Cohen C, Wu R and Gu H: “Modeling of Hydraulic-Fracture-Network Propagation in a Naturally Fractured Formation,” SPE Production and Operations 26, no. 4 (November 2011): 368–380.

For more on UFM modeling: Cipolla C, Weng X, Mack M, Ganguly U, Gu H, Kresse O and Cohen C: “Integrating Microseismic Mapping and Complex Fracture Modeling to Characterize Fracture Complexity,” paper SPE 140185, presented at the SPE Hydraulic Fracturing Technology Conference and Exhibition, The Woodlands, Texas, January 24–26, 2011.

Figure 12. Workflow for completion and stimulation design and field development applications of microseismic mapping. Using seismic data, well logs and core data along with treatment and production data (left ), engineers build a series of earth models and discrete fracture network (DFN) models (middle). They use these models to generate hydraulic fracture predictions for the fracture geometry and conductivity distribution resulting from stimulation operations. They also use MS event locations and deformation to calibrate the earth and fracture models. Reservoir engineers are then able to predict fracture network geometry and conductivity and generate reservoir performance simulations (right ). (Adapted from Cipolla et al, reference 39.)

•Event processing•Interpretation•Visualization

Microseismic Monitoring

Hydraulic Fracture Models

•Fracture model calibration

Completion and Stimulation

•Modify completion strategy

ReservoirSimulation Models

•Reservoir model calibration

Earth Models

•Velocity model•1D or 3D MEM•Geologic model•Reservoir model•DFN

Hydraulic Fracture Model

•Fracture geometry and conductivity distribution

Real-time TreatmentModification

•Modify pump rate•Modify proppant schedule•Continue or abort job

Reservoir Simulation

•Production profile•Hydrocarbon recovery•Drainage architecture

Completion andStimulation

•Perforation locations•Stage sequence•Treatment data

Flow

•Production logs•Production data

Product

Model

Data

Seismic

•Structure•Natural fractures and faults•Rock properties•Stress variations

Logs, Cores and Petrophysics

•Natural fractures•Rock properties•Stress profile and anisotropy•Reservoir properties

Oilfield Review MAY 16Microseismic Fig 12ORMAY 16 MCSMC 12

May 2016 29

Figure 14. Flowchart for calibrating UFM processing and DFN simulations. Analysts build a DFN model (top right, light gray lines) based on geologic, geophysical, well log and core data. Preexisting discrete fractures directly affect the hydraulic fracture system. The UFM simulations predict fracture geometry (middle right, heavy blue lines) based on treatment parameters and an earth model that includes the estimated stress field. The stress field can be calculated with a 3D geomechanical simulator using wellbore measurements as calibration points. Analysts compared the fracture geometry predicted from UFM processing with maps of the observed MS event pattern (bottom left, red dots), taking into consideration the deformation represented by the seismic moment obtained from MS monitoring. The engineers then executed an iterative calibration loop, adjusting UFM processing inputs for multiple DFN realizations (bottom right ) to arrive at the best overall agreement between the modeled fracture geometries and the observed deformations. (Adapted from Cipolla et al, reference 54.)

Oilfield Review MAY 16Microseismic Fig 15ORMAY 16 MCSMC 15

Microseismic Events

DFN Realization

Observations: Logs,seismic data and natural fractures

Fracture treatment data andmechanical earth model (MEM)

UFM Prediction

Comparison

Match?

Outputfracturemodel

No

Yes

Start

Figure 13. Construction of a discrete fracture network. For a fracture stimulation in the Eagle Ford Shale, analysts processed logging data obtained from a well drilled with oil-base mud using an imaging tool and were able to detect fractures and determine their orientations from a rose plot (top). The red dots indicate dip azimuth and inclination angle of the poles of the primary fracture planes; blue dots are dip azimuth and inclination angle of the poles of the secondary fracture planes. From the inside to the outside edge of the rose plot, the dip varies from 0° to 90°. The black lines represent fracture strike orientation and the length is related to the abundance along a given direction. Analysts identified a primary and a secondary fracture set. The intensity of fracture occurrence along the wellbore correlated with formation curvature determined using reflection seismic surveys. Co-kriged—a geostatistical technique for data interpolation—fracture intensity and curvature were used to populate the 3D volume of interest with fractures and build the DFN (bottom). The SW–NE trending band of fracture intensity corresponds to an area of high formation curvature. The well trajectory (blue line) is superimposed on the DFN. (Adapted from Offenberger et al, reference 53.)

Oilfield Review MAY 16Microseismic Fig 13ORMAY 16 MCSMC 13

0°

30°

60°

90°

Primary fracture setn = 775Orientation:approximatelyN54°E

Secondary fracture setn = 175Orientation:approximatelyN60°W

120°

150°

180°

210°

240°

270°

300°

330°

Composite Fracture Set

1,00

0 ft

30 Oilfield Review

Engineered CompletionsThe Eagle Ford Formation, an upper Cretaceous marl in South Texas, is a target for oil and gas development. Because the formation is highly laminated and has ultralow permeability, effec-tive completion designs are required to maximize production. In September 2013, an operator tested MSM as a method for guiding hydraulic fracture operations and evaluating fault interac-tions during stimulation.

The operator drilled two horizontal lateral wells in the gas-condensate window of the Eagle Ford Shale trend in Karnes County, Texas. These wells, which were drilled parallel to each other and approximately 330 ft [100 m] apart, crossed a major fault. After the operator drilled the wells, measurements of petrophysical and geomechani-cal properties were acquired using the ThruBit through-the-bit logging services.55 Using the Mangrove engineered stimulation design module in the Petrel platform, engineers developed a completion design for one of the two laterals.56

Figure 15. Vertical section view of MS events recorded during the stimulation of two horizontal wells in the Eagle Ford Shale. Microseismic event hypocenters (spheres, color-coded by stage number) and the trajectories of Wells A (orange), B (blue) and C (yellow) are shown in conjunction with the formation tops (horizontal tan and light blue) and faults (vertical gray and green). Treatment Wells A and B were stimulated in a 21-stage zipper fracture operation. Engineers used a 12-level geophone array positioned in the vertical portion of Well C to monitor 16 of the 21 stages

of the stimulation operation. Microseismic events were largely confined to the Eagle Ford Formation, bounded above by the Austin Chalk and below by the Buda Limestone. During stimulation Stage 8 (dark blue) of Well A, the MS events aligned with the adjacent major fault and propagated downward into the Buda Limestone. This observation suggested the fault affected fracture growth during the stimulation. A decision was made to abort Stages 9 (teal) and 10 (crimson) and proceed with Stage 11 (orange) on the other side of the fault. (Adapted from Zakhour et al, reference 58.)

Oilfield Review MAY 16Microseismic Fig 16ORMAY 16 MCSMC 16

Monitor Well CGeophones

Well plug

Austin Chalk

Eagle FordShale

BudaLimestone

Treatment WellsA and B

Stimulation stage

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Figure 16. Trajectories of three wells to be stimulated in the Barnett Shale in relation to a fault system mapped from 3D surface seismic data. Proximity of faults can influence the local stress field, affecting induced fracture propagation and associated microseismicity. The well stimulation plans for these operations included a buffer zone containing no perforations in Wells 1H, 2H and 3H near a major fault (aqua surface). The colored disks represent the perforation intervals for the stimulation stages; no stimulation stages were attempted south of the fault in Wells 1H and 3H. (Adapted from Le Calvez et al, reference 59.)

Oilfield Review MAY 16Microseismic Fig 17ORMAY 16 MCSMC 17

1H3H

2HN

6,000

6,400

6,800

7,200

7,600

8,000

8,400

8,800

400 ft

Dept

h, ft Major fault

May 2016 31

Stimulation design engineers used reservoir and completion quality parameters derived from the Mangrove software to aid in optimizing stage intervals and the placement of perforations along the lateral.57 Flow measurements in some previ-ous wells that had geometric—evenly spaced—completions had shown unequal contribution to production across perforations. The engineered stages grouped perforation clusters in regions of the lateral that had similar horizontal stress. Completion engineers anticipated that all perfo-ration clusters in a stage would break down and initiate fractures simultaneously because each cluster had similar stress characteristics. Measurements made during stimulation con-firmed that lower than average treating pres-sures were required for the engineered completion than had been used for the lateral that was stimulated using a geometric model.

Survey engineers acquired MS data during the stimulation using a 12-receiver downhole array in a nearby vertical well. Using these MS data, the operator monitored hydraulic fracture development near a fault system, identified fault interaction and adjusted the completion design to avoid the fault. Engineers later studied the mechanisms of fault interaction through postjob integration of MS data with treatment data.

Completion designers normally try to avoid the interaction of hydraulic fractures with large faults. Avoiding faults can prevent the loss of treatment fluid and proppant to thief zones along the fault. In this project, the completion design excluded stimulation stages that were within 250 ft [76 m] of the identified major fault.

Monitoring revealed that MS events were generally well bounded within the target Eagle Ford Formation and the overlying Austin Chalk (Figure 15). However, for some stages, MS activ-ity and treating pressure records indicated unex-pected fracture bridging and potential proppant screenout. Analysis of MS event clusters and b-values alerted the engineers that the hydrau-lic fractures had encountered the nearby fault, which blocked and limited fracture develop-ment and led to premature stage terminations. Real-time interpretation allowed modification of the completion strategy, and several stimulation stages planned near the fault were abandoned. Recommendations were made for future comple-tion designs to increase buffer zones from 250 ft to about 400 ft [120 m] on either side of major faults to minimize the risk of pumping nonpro-ductive stages.58

Maximizing RecoveryIn 2011, engineers and geoscientists with Teleo Operating, LLC and Eagleridge Energy, LLC con-ducted a multistage, multilateral stimulation in the Barnett Shale in Denton County, Texas. The operators drilled three parallel horizontal wells into the lower Barnett Shale. Well trajectories were about 500 ft [150 m] apart; the central well was landed about 80 ft [25 m] shallower than the outside laterals. Because of lease boundary con-straints, the wells had to be placed in the vicinity of several large faults. The lateral sections of the wells were drilled away from the major fault and through a smaller fault (Figure 16). Fault throws

55. For more on ThruBit services: Aivalis J, Meszaros T, Porter R, Reischman R, Ridley R, Wells P, Crouch BW, Reid TL and Simpson GA: “Logging Through the Bit,” Oilfield Review 24, no. 2 (Summer 2012): 44–53.

56. For more on the Mangrove service: Ajayi B, Aso II, Terry IJ Jr, Walker K, Wutherich K, Caplan J, Gerdom DW, Clark BD, Ganguly U, Li X, Xu Y, Yang H, Liu H, Luo Y and Waters G: “Stimulation Design for Unconventional Resources,” Oilfield Review 25, no. 2 (Summer 2013): 34–46.

57. For more on determining reservoir and completion quality: Slocombe R, Acock A, Fisher K, Viswanathan A,

varied from 20 to 100 ft [7 to 30 m]. Engineers used a zipper fracture stimulation on the central Well 1H and eastern-most lateral 3H. The west-ern-most lateral 2H was stimulated later.

Microseismic survey engineers monitored stimulations on Wells 1H and 3H using a 3C-accelerometer array that was placed in hori-zontal Well 2H using a tractor and repositioned according to the well and stage to be monitored. Operations on Well 2H were later monitored using an array deployed in a vertical section below the 3H wellhead. During monitoring, engi-neers observed MS activity across all pumped stages (Figure 17).

Figure 17. Microseismic events detected during stimulation of three horizontal wells in the Barnett Shale. Fault traces (cyan) are mapped at the depth of laterals 1H (red), 2H (green) and 3H (yellow). The zipper fracturing performed on Wells 1H and 3H was monitored from Well 2H. Engineers also monitored four stimulation stages performed on Well 2H using sensors in Well 3H. Color coding and symbols are used to represent stages and wells. Microseismicity for stages closer to the fault system tended to be compact, an observation that was explained later by fracture modeling as stimulation-fault interaction. Longer hydraulic fracture wings occurred in the stages that were executed closer to the toe of the well than those executed close to the heel. Microseismicity overlap observed between successive stages indicates insufficient fracture isolation. (Adapted from Le Calvez et al, reference 59.)

Oilfield Review MAY 16Microseismic Fig 18ORMAY 16 MCSMC 18

400 ft

400

ft

X-axis, ft

Y-ax

is, f

t

Stage 4Stage 3Stage 2Stage 1Well 2H

Stage 4Stage 3Stage 2Stage 1Well 1H

Stage 4Stage 3Stage 2Stage 1Well 3H

Stage 5

1H 3H2H

N

Chadwick C, Reischman R and Wigger E: “Eagle Ford Completion Optimization Using Horizontal Log Data,” paper SPE 166242, presented at the SPE Annual Technical Conference and Exhibition, New Orleans, September 30–October 2, 2013.

58. For more on horizontal completion optimization across a major fault: Zakhour N, Sunwall M, Benavidez R, Hogarth L and Xu J: “Real-Time Use of Microseismic Monitoring for Horizontal Completion Optimization Across a Major Fault in the Eagle Ford Formation,” paper SPE 173353, presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, February 3–5, 2015.

32 Oilfield Review

During zipper fracturing of Wells 1H and 3H, interpretation of MS event location trends indi-cated the hydraulic fracture azimuth, N50–55°E, was consistent with the expected maximum hori-zontal stress direction; however, the extent of the microseismicity varied from stage to stage. The MS locations for the early stages, near the toe of the wells, extended farther from the wellbore than those observed during later stages, toward the heel of the wells. These later stages were closer to the main fault and displayed a more compact microseismicity pattern and shorter fracture wings away from the wellbore.

The MS locations observed in these later stages overlapped with those observed in some early stages. Several later stages near the heels of the wells displayed microseismicity that was out

of the target depth interval. Observations revealed downward fracture growth and alerted the operator to stop pumping to avoid fracturing into the water-bearing Viola Limestone below the zone of interest. During another stage, engineers recognized that the planar alignment of MS events indicated slip along a fault and were able to stop pumping for that stage and bypass a faulted zone before resuming stimulation. Microseismic monitoring allowed the operator to modify the stimulation program during the ongo-ing job operation.59

Postsurvey modeling and data integration provided a more complete explanation of the stimulation and MS responses.60 Analysts con-structed a complete history of the treatment

using a range of analytical techniques and incor-porated summary statistics of MS event attri-butes within the workflow. Event attributes included seismic moment and moment magni-tude. From these parameters, analysts deter-mined summary b-value statistics and inferred relative stress magnitudes for each stimulation stage.61 They used D-value estimates to extract fracture planes from the clouds of microseismic events—D-values near two indicated planar alignments—and then used the planes for DFN construction. Monitoring geometries that can provide full MTI offer an additional means to con-strain DFN construction and modeling but were not available in this study. Additional input to the analysis included earth model parameters such as the rock mechanical properties of layers and natural fracture geometries along with stimula-tion data such as well geometry, flow rates, fluid types and surface pumping pressures.

Engineers combined the Mangrove, UFM and VISAGE software to model the hydraulic fractur-ing process, the interaction of induced and natu-ral fractures and the stress field.62 These models predicted the chronological development of the interconnected fracture network and were cali-brated iteratively using the observed evolution of MS activity. They tested fracture propagation sce-narios by matching time-distance relationships within the microseismicity pattern and then iteratively improved the interpretation by updat-ing the location and properties of the natural fractures. Engineers constrained the simulation using material balance, which reconciled the fracture volume opened during stimulation with the volumes of pumped fluid and proppant and the volume of fluid estimated to have leaked off into the formation. The simulation provided a description of proppant placement together with a prediction of which natural fractures might

59. Le Calvez J, Xu W, Williams M, Stokes J, Moros H, Maxwell S and Conners S: “Unconventional Approaches for an Unconventional Faulted Reservoir—From Target Selection to Post-Stimulation Analysis,” paper P336, presented at the 73rd EAGE Conference and Exhibition, Vienna, Austria (May 23–26, 2011).

60. Williams MJ, Le Calvez JH and Stokes J: “Towards Self-Consistent Microseismic-Based Interpretation of Hydraulic Stimulation,” paper Th 01 15, presented at the 75th EAGE Conference and Exhibition, London, June 10–13, 2013.

61. Williams MJ, Le Calvez JH, Conners S, Xu W: “Integrated Microseismic and Geomechanical Study in the Barnett Shale Formation,” Geophysics 81, no. 3 (May–June 2016): 1–13.

62. In the Barnett Shale case study, Schlumberger engineers used Mangrove software with the UFM complex fracture simulator to model fracture interaction and the VISAGE simulator to model stress.

63. For more on modeling of faulted rock masses: Pande GN, Beer G and Williams JR: Numerical Methods in Rock Mechanics. Chichester, New York, USA: John Wiley and Sons Ltd., 1990.

Figure 18. Geomechanical modeling using finite element analysis. The stimulation of the Barnett Shale Well 3H, Stage 5, was near a major fault. The fault plane projections (olive green) are shown intersecting Wells 1H (blue), 2H (green) and 3H (yellow). Simulations were run for materials near the fault that had equivalent high stiffness (top) and low stiffness (bottom). The fault properties affect the extent of the region perturbed by fracturing in the simulations. Analysts consider the region perturbed by fracturing to be similar to the region where microseismicity occurs. In this case, failure and microseismicity are expected to be limited in the region near the fault. The colored volumes show only the simulation elements where the stresses are perturbed toward failure by the stimulation. The colors correspond to the minimum in situ effective stress; purple is low compression, showing the region where tensile failure is most likely to occur, and red is high compression. (Adapted from Williams et al, reference 61.)

Oilfield Review MAY 16Microseismic Fig 19ORMAY 16 MCSMC 19

Effe

ctiv

e st

ress

, psi

0

10

20

30

40

50

60

70

80

90

100

Fault

2H

3H

1H

0

10

20

30

40

50

60

70

80

90

100

Effe

ctiv

e st

ress

, psi

2H

1H

3H

Fault

May 2016 33

open and which ones behaved as barriers that promoted vertical or asymmetric fracture growth.

Modeling for this Barnett Shale stimulation helped analysts understand the downward growth of microseismicity into the Viola Limestone. The complex fracture simulator reproduced observa-tions of hydraulic fracture interactions with natural fractures, which acted as barriers to prop-agation in the Barnett Shale and in the underlying Viola Limestone. By tuning the fault properties within the finite-element geomechanical model, analysts were able to match results with the observed distribution of MS events.

Analysts used finite-element geomechanical modeling to study how fault properties influ-enced the zone that a hydraulic fracture perturbs toward failure (Figure 18). In this type of model-ing, the rock mass, including its embedded faults and fractures, is described as an equivalent medium that has standard mechanical properties such as elastic modulus and strength.63 The pres-ence of an active fault may alter the regional stress field in its vicinity, pushing outward, nor-mal to its plane, and increasing the stress on frac-ture interfaces. The simulations revealed that the compact distribution of MS events in stages pumped close to the fault was consistent with

large values of fault-related stiffness. By simulat-ing all stages and varying fault-related stiffness, modelers demonstrated that the interaction with the fault was consistent with microseismicity that was more compact in the stages toward the heel than in those toward the toe. The under-standing of fault interaction gained from this case study should benefit reservoir engineers planning future refracturing operations in these wells or stimulation treatments in nearby wells.

Challenges and the FutureAs operators develop unconventional resources, determining the optimum well spacing and com-pletion strategies that maximize ultimate recov-ery is critical. To help operators achieve these objectives, MSM provides key data for constrain-ing and calibrating models used to help geoscien-tists with data interpretation and integration. Microseismicity is induced as the reservoir and adjoining formations respond to stimulation treatments. Models used to predict how these for-mations should respond to stimulations must simulate and faithfully reproduce the observed microseismicity. Challenges for geoscientists are the accurate measurement of MS events and extracting of maximal information from them.

Advances in acquisition, processing, interpre-tation and integration of MS data are providing unique insights into and increased understand-ing of stimulated reservoir behavior. Models help engineers interpret and constrain MS data, and geomechanical earth models help them charac-terize the variability of reservoir properties. Fracture network modeling facilitates predic-tions of the interactions between hydraulic frac-tures and rock fabric. Reservoir simulations assist in predicting field drainage patterns, and productivity may be validated through produc-tion history matching. Monitoring microseismic-ity offers valuable data for validating these models and simulations.

Methods such as MSM give operators insight into reservoir dynamics that far exceeds what was possible even a few years ago. Success in develop-ing unconventional reservoirs owes much to the pioneers working in plays such as the Barnett Shale. Recent advances in tools and technolo-gies are allowing operators to develop unconven-tional reservoirs with greater certainty, reduced risk and deeper understanding of the nature of these formations. —HDL

Joël Le Calvez is a Schlumberger Geophysics Advisor and Microseismic Domain Expert in Houston. He works on development and commercialization of microseismic and borehole seismic products while heading a team of geophysicists, geologists and stimulation engineers working on various plays around the world. He has man-aged the Microseismic Services Answer Product Center and the borehole seismic processing and crosswell seis-mic groups in Houston since 2014. His main responsibili-ties are the processing and interpretation of data for geologic, geophysical and geomechanical applications. He also works with product centers defining and testing software programs and with research centers on defin-ing and testing of algorithms. Joël joined Schlumberger in 2001, and after several years in the field acquiring and processing seismic data, he led the microseismic processing and interpretation team in Dallas from 2008 until 2011. He then moved to Houston to manage the North America microseismic processing and interpreta-tion center. He earned a BSc degree in mathematics and physics and an MSc degree in geology and geophysics, both from the Université de Nice Sophia Antipolis, France; a Diplôme d’Etudes Approfondies in tectono-physics from the Université Pierre et Marie Curie, Paris; and a PhD degree in geology from The University of Texas at Austin, USA.

Raj Malpani is a Senior Completions and Production Engineer with Schlumberger Technology Corporation in Houston. For the past 10 years, he has been a part of integrated teams that address technical challenges pertaining to unconventional reservoirs. His interests

include hydraulic fracture treatment design and evalu-ation, production data analysis, reservoir simulation, geomechanics, microseismic monitoring, restimulation, multiwell pad development and weak interface model-ing. Raj holds a BTech degree in petrochemical engi-neering from Dr. Babasaheb Ambedkar Technological University, Lonere, Maharashtra, India, and an MS degree in petroleum engineering from Texas A&M University, College Station.

Jerry Stokes is the President and Owner of Mid-Continent Geological, Inc. in Fort Worth, Texas. He has been a certified petroleum geologist with the AAPG for more than 35 years. Since 1987, he has been involved in oil and gas exploration, geologic consulting and sales and marketing of geologic projects throughout Texas and nearby states. As a geologist for Panhandle Eastern Pipeline, Jerry was responsible for the early develop-ment of underground gas storage fields in Kansas, Louisiana, Illinois and Michigan, USA. He then worked for Rust Oil Corporation as the exploration manager for the Permian basin. He is a member of the Society of Independent Professional Earth Scientists and the Fort Worth Wildcatters. Jerry has a BSc degree in geology and geophysics from Texas Tech University, Lubbock.

Michael Williams is a Principal Reservoir Engineer in Geophysics at Schlumberger Gould Research (SGR), Cambridge, England. Since 2008, he has worked in the area of interpretation of microseismicity, specifically the accurate recovery of statistical information from detection-limited microseismic data, and the application of microseismic interpretation to reservoir simulation

and geomechanical modeling. He joined Schlumberger GeoQuest in 1997 as a commercialization software engineer and worked as project leader and team leader in Abingdon, England. In 2002, he was a team leader in Sugar Land, Texas, developing the first hydraulic frac-ture monitoring software to support the interpretation of microseismic information in the context of fracture stimulation. He joined SGR as a senior research scien-tist in 2004, where he worked in applied reservoir engi-neering, fluid measurements (as program manager) and well test interpretation. Michael received a BS degree in physics and an MS degree in geophysics, both from Imperial College of Science, Technology and Medicine, University of London. He also has a PhD degree in phys-ics from the University of Wales, Aberystwyth.

Jian Xu is a Senior Microseismic Services Engineer in Houston. He focuses on microseismic data interpreta-tion, hydraulic fracture monitoring and stimulation program evaluation in unconventional plays. He joined Schlumberger in 2008 as a field engineer in Bryan, Texas. He held various positions, including access field engineer, production stimulation engineer and microseismic services engineer working on several unconventional plays in the US, all while located in Houston. Before his current assignment, Jian was a senior production stimulation engineer at the Production Technology Integration Center in Houston. He obtained BS and MS degrees in electrical engineer-ing from Tianjin University, China, and a PhD degree in petroleum engineering from Texas A&M University, College Station.

Contributors

34 Oilfield Review

Corrosion—The Longest War

Locations that host oil and gas operations often provide ideal conditions for corrosion.

Ongoing research and advances in coatings, cathodic protection, nondestructive

testing, corrosion analysis and inhibitors allow operators to safely produce oil and

gas in these corrosive environments.

Nausha AsrarBruce MacKaySugar Land, Texas, USA

Øystein BirketveitMarko StipanicevBergen, Norway

Joshua E. JacksonG2MT Laboratories, LLCHouston, Texas

Alyn JenkinsAberdeen, Scotland

Denis MélotTotalParis, France

Jan ScheieStavanger, Norway

Jean VittonatoTotalPau, France

Oilfield Review 28, no. 2 (May 2016).Copyright © 2016 Schlumberger.DS-1617 is mark of M-I LLC.Hastelloy is a registered trademark of Haynes International, Inc.Inconel and Monel are trademarks of Special Metals Corporation.

Corrosion validates the universal law of entropy; everything trends toward a state of greater chaos and disorder. The flecks of rust on an iron bar or the green patina on a copper fixture are evidence of the insidious effects of corrosion. These examples may be regarded as an annoyance, but taken to the extreme, the results of corrosion can lead to cata-strophic outcomes.

Corrosion has brought down bridges, downed aircraft, leveled chemical plants, parted drill-pipe and ruptured pipelines. Given sufficient time, this adversary has the potential to degrade any material. In certain environments, the unchecked effects of corrosion can come swiftly, and the consequences of failure to manage corro-sion can be costly.

Oilfield Review MAY 16Corrosion Fig OpenerORMAY 16 CRSSN Opener

May 2016 3535

According to the US Federal Highway Admin-istration, the approximate annual direct cost of corrosion for the US in 2015 was an estimated US$ 500 billion, representing around 3.1% of the nation’s gross domestic product.1 This figure amounts to six times the average annual cost of weather-related disasters for the US, which was about US$ 87 billion in 2011.2 Unlike weather events, corrosion can be controlled or at least managed; scientists estimate that 25% to 30% of corrosion costs could be avoided if good corrosion management practices and preventive strategies were employed.3

Throughout the ages, and despite an early lack of understanding concerning the fundamen-tal mechanisms involved, humans have attempted to control corrosion. In ancient times, corrosion resistance was sometimes imparted to materials as a matter of circumstance rather than design (Figure 1).4 Early corrosion control methods included the use of bitumen and lead-based paints by the Romans in the first century. Around 500 BCE, Chinese sword makers used copper sulfide coatings to inhibit corrosion on bronze swords. Centuries later, the copper sheathing used on British sailing vessels to reduce biofoul-ing—fouling of underwater surfaces by organisms such as barnacles and algae—and increase speed accelerated the corrosion of nails that held the ships together.5

Michael Faraday was one of the most impor-tant contributors to the early understanding of corrosion; in the early 1800s, he established a quantitative relationship between the chemi-cal action of corrosion and electric current.6 Although much more is known about the subject today, scientists continue to study the mecha-

nisms of corrosion and search for methods to manage and control it.

Combating corrosion is a significant source of expenditures for the oil and gas industry (Figure 2). British Petroleum (BP) conducted a study of its operations in the North Sea in 1995.7 The company found that outlays for corrosion

1. Koch GH, Brongers MPH, Thompson NG, Virmani YP and Payer JH: “Corrosion Costs and Preventive Strategies in the United States,” Washington, DC: US Department of Transportation Federal Highway Administration, Publication FHWA-RD-01-156, March 2002.

Jackson JE: “Corrosion Will Cost the US Economy over $1 Trillion in 2015,” G2MT Laboratories, http:// www.g2mtlabs.com/corrosion/cost-of-corrosion/ (accessed January 6, 2016).

Papavinasam S: Corrosion Control in the Oil and Gas Industry. Waltham, Massachusetts, USA: Gulf Professional Publishing, 2014.

2. The US$ 87 billion cost of weather-related disasters in 2011 was the highest on record. The average annual cost has been closer to US$ 10 billion in recent years. For more on the cost of weather-related disasters: Smith AB and Katz RW: “U.S. Billion-Dollar Weather and Climate Disasters: Data Sources, Trends, Accuracy and Biases,” Natural Hazards 67, no. 2 (June 2013): 387–410.

3. Chillingar GV, Mourhatch R and Al-Qahtani GD: The Fundamentals of Corrosion and Scaling for Petroleum and Environmental Engineers. Houston: Gulf Publishing Company, 2008.

4. Kumar AVR and Balasubramaniam R: “Corrosion Product Analysis of Corrosion Resistant Ancient Indian Iron,” Corrosion Science 40, no. 7 (July 1, 1998): 1169–1178.

Balasubramaniam R: Story of the Delhi Iron Pillar. Delhi, India: Foundation Books Pvt. Ltd, Cambridge House, 2005.

5. Groysman A: Corrosion for Everybody. Dordrecht, The Netherlands: Springer Science+Business Media, 2010.

6. Ahmad Z: Principles of Corrosion Engineering and Control, 1st ed. Burlington, Massachusetts: Butterworth-Heinemann, 2006.

7. Kermani MB and Harrop D: “The Impact of Corrosion on the Oil and Gas Industry,” SPE Production & Facilities 11, no. 3 (August 1996).

Figure 1. Delhi pillar. This iron pillar is located in the Qutub Complex in New Delhi, Delhi, India (inset). It is about 9.1 m [30 ft] tall and weighs approximately 6,000 kg [13,200 lbm]. Erected in 400 CE, the pillar is essentially free of the typical rusting that would be expected to take place over 1,600 years of exposure. Reasons for the lack of corrosion include New Delhi’s low humidity but are primarily attributed to the high concentration of phosphorus in the iron.

Oilfield Review MAY 16Corrosion Fig 1ORMAY 16 CRSSN 1

I N D I A

P A K I S

T A N

NEPAL

C H I N A

BANGLADESH

SRI LANKA

Mumbai

New Delhi

km

miles0

0 500

500

Oilfield Review MAY 16Corrosion Fig 2ORMAY 16 CRSSN 2

Production1.4

Pipelines7.0Tankers

2.7

Storage7.0

Refining3.7

Distribution5.0

US Oil and Gas Corrosion Expenditures, US$ billion/year

Figure 2. Corrosion expenditures. Corrosion expenditures in the US oil and gas industry are about US$ 26.8 billion/year. The downstream segment of the industry—production, pipelines and tankers—accounts for 41% of the total, or US$ 11 billion/year. (Adapted from Koch et al, reference 1.)

36 Oilfield Review

prevention and control averaged about 8% of the total capital expenditure for its projects. On the UK Continental Shelf, 25% to 30% of BP’s operating costs were related to the control and management of corrosion. Costs associated with replacing corroded equipment, lost production and corrosion-related contamination contributed to overall expenditures. In addition to the direct costs, the company found that corrosion had a significant indirect cost on health, safety and environmental concerns.

This article focuses on descriptions of cor-rosion, management techniques and advances in corrosion abatement technologies. Field examples from Gabon, deepwater Nigeria and the North Sea illustrate the ongoing battle waged against corrosion by oil and gas operators.

The Corrosion ProcessScientists and engineers today have a better understanding of corrosion processes than did the ancient Romans and Chinese. Fighting cor-rosion requires an understanding of the principal elements that cause and contribute to the corro-sion. There are several categories of corrosion; for the oil and gas industry, common types include exposure to carbon dioxide [CO2, sweet corro-sion], hydrogen sulfide [H2S, sour corrosion], oxygen [O2] and corrosion causing microbes, referred to as microbiologically influenced cor-rosion (MIC).8

Some forms of metal corrosion are related to stability; for example, galvanic corrosion is an electrochemical process associated with the movement of electrons between areas that have different electrochemical potentials. The corro-sion cell schematically describes oxidizing corro-sion, which is analogous to a battery in which two dissimilar metals are connected by an electrolyte (Figure 3).9 A metal that has a higher corrosion rate—more unstable—represents the negative part of the cell and acts as the anode; a second metal that has a lower corrosion rate—more stable—acts as the positive part of the cell, the cathode.10

During the galvanic corrosion process, metal oxides are formed as electrons flow from the anode to the cathode through the electrolyte—the fluid in contact with the anode and cathode. A simplified version of iron oxidation can be used to illustrate the galvanic corrosion process—the actual process is more complex. The presence of water [H2O] on the surface of the iron [Fe or Fe0] releases electrons to form ferrous iron [Fe+2] and ferric iron [Fe+3] ions, which act as the anode in our battery analogy. The liberated electrons flow to the cathode, where, in the presence of oxygen [O2], ferrous oxide [FeO] and ferric oxide [Fe2O3] form as scales of rust or precipitates. A byproduct of the reaction at the cathode is hydroxyl ions [OH–] from the reduction of oxygenated water.

Iron can also react with CO2 to form iron carbon-ate [FeCO3] and with H2S to form iron sulfides [FexSx]. In the absence of O2 but the presence of CO2 and H2S, the cathodic reaction can generate hydrogen gas.

These reactions can occur rapidly, but if the reaction rate can be reduced, the overall corro-sion rate will also be reduced. Many factors influ-ence the reaction rate. These include the type and quality of metal, electrolyte compositions, pH, temperature, pressure, presence of dissolved gases, liquid velocity, water salinity, applica-tion of cathodic protection and the presence of microbes.11 To manage corrosion and corrosion rate, knowledge of the metallurgy of the mate-rials to be used and the environments in which they will operate is important.

If CO2 comes into contact with water in the producing or transportation system of an oil and gas operation, areas typically affected include well internals, gathering lines and pipelines. In CO2 corrosion of iron, the products of reaction are carbonic acid, iron carbonate [FeCO3] and hydrogen gas [H2].12 For CO2 corrosion to occur, the partial pressure of the gas can be as low as 21 kPa [3 psi]. To prevent this type of corrosion, operators commonly use organic films that act as barriers and inhibitors that neutralize the acidity of the carbonic acid generated in the corrosion process. Operators may also use corrosion resis-tant alloys (CRAs), which are resistant to general and localized corrosion, in environments that are corrosive to carbon and low-alloy steels.

Hydrogen sulfide is often found in produced fluids or as a result of MIC.13 Although H2S is not corrosive, it becomes corrosive in the presence of water.14 Sour corrosion from H2S can affect any part of the producing system, including well internals and oil and gas gathering lines. Oilfield fluids are considered sour if the produced gas con-tains more than 5.7 mg of H2S per m3 [4 parts per million (ppm)] of natural gas or produced water has greater than 5 ppm H2S.15 At the anode, the H2S reacts with the iron to form several vari-ants of iron sulfide [FexS] such as mackinawite [(Fe,Ni)(1 + x)S], pyrrhotite [Fe(1 - x)S] and troilite [FeS].16 These iron sulfide species precipitate and can form localized microgalvanic corrosion cells.

The corrosion cells formed during sour corro-sion cause pitting, sulfide stress cracking (SSC) and hydrogen embrittlement.17 Stress corrosion cracking is a result of tensile stress combined with a wet environment and often causes shal-low, round pits that have etched bottoms accom-panied by branching cracks that can lead to rapid failure. Hydrogen embrittlement occurs when H2S and H2 diffuse into metal, recombine with

Figure 3. Corrosion cell. When steel in water rusts, several reactions take place simultaneously. At the anode, steel [Fe0] goes readily into solution to form ferrous iron [Fe2+] and ferric iron [Fe3+] (not shown) ions, and electrons move to the cathode. Electrons at the cathode react with water [H2O] to form oxygen [O2] and hydroxyl [OH–] ions. The OH– ions combine with the solubilized Fe2+ to form iron hydroxide [Fe(OH)2].

Oilfield Review MAY 16Corrosion Fig 3ORMAY 16 CRSSN 3

Fe0Fe0

Fe0Fe0

Anode

Water

Electron flow

Steel

Cathode

Fe2+ Fe(OH)2Fe(OH)2

OH–

OH–

OH–

Anodic reaction Fe0 Fe2+ + 2e–

Cathodic reaction H2O + 2e– 0.5 O2 + 2OH–

Fe2+ + 2OH– Fe(OH)2

May 2016 37

other molecules and create pressure within the metal matrix; byproducts of cathodic protection, galvanic corrosion and other mechanisms may lead to hydrogen embrittlement.

The failure mode during hydrogen embrittle-ment depends on the steel type; for example, low-strength steels exhibit blistering. The failure mode of high-strength steels can be catastrophic when the pressure of the trapped gas exceeds the tensile strength of the metal. To control sour cor-rosion, operators use organic film formers, H2S scavengers, metals resistant to SSC, flowline pig-ging, nitrate treatments and biocides that reduce the growth of microbes that cause MIC.18

Oxygen-related corrosion in oil and gas producing environments is often much more aggressive than corrosion caused by CO2 or H2S (Figure 4).19 Corrosion by oxygen is directly pro-portional to the concentration of the dissolved gas. If chlorides, CO2 or H2S are present, the cor-rosion rate can increase significantly.

Oxygen has the ability to induce corrosion throughout producing systems. Inhibition of oxy-gen corrosion is difficult, and corrosion reduction efforts for production and water handling facili-ties have usually been directed toward exclusion of oxygen from the system and the use of oxy-gen scavengers. Typical oxygen scavengers are ammonium bisulfite [NH4HSO3], sodium sulfite [Na2SO3] and sodium bisulfite [NaHSO3].20 In addition to scavenger stripping, vacuum deaera-tors are sometimes used to control the corrosive effects of oxygen on metals.

Exposure to oxygen is also a major source of drillpipe corrosion. While it is being run in and out of the well, drillpipe is exposed to atmo-spheric oxygen. During drilling, drillpipe comes into contact with oxygen in the mud system. Both instances can induce corrosion. The usual expression of oxygen-related corrosion is pitting. Pitting can even develop under mud left on and inside drillpipe, where pipe storage racks contact the pipe and at crevices. Deep corrosion pits in drillpipe can lead to the onset of fatigue failure. Drillpipe may be coated with epoxies or resins to stop corrosion, but the harsh downhole envi-ronment often quickly removes these protective coatings. Pipe dope, lubricating grease applied to threaded connections, may help prevent cor-rosion of these connections.

Corrosion Form and Appearance The word corrode comes from the Latin corrodere meaning to gnaw; it can carry the additional meaning of eat or wear away gradually.21 Corrosion typically leaves a visible signature that is characteristic of the agent and mechanism

8. Popoola LT, Grema AS, Latinwo GK, Gutti B and Balogun AS: “Corrosion Problems During Oil and Gas Production and Its Mitigation,” International Journal of Industrial Chemistry 4, no. 1 (2013).

Chillingar et al, reference 3. 9. Stansbury EE and Buchanan RA: Fundamentals of

Electrochemical Corrosion. Materials Park, Ohio, USA: ASM International, 2000.

Brondel D, Edwards R, Hayman A, Hill D, Mehta S and Semerad T: “Corrosion in the Oil Industry,” Oilfield Review 6, no. 2 (April 1994): 4–18.

10. Although the battery analogy is acceptable for explaining corrosion involving two dissimilar metals, corrosion processes also take place on single metals. In single metals, the mechanism for corrosion consists of small crystals with slightly different compositions. The anode and the cathode are located on different areas of the metal surface and, depending on the conditions, may be close to each other or far apart.

11. Heidersbach R: Metallurgy and Corrosion Control in Oil and Gas Production. Hoboken, New Jersey, USA: John Wiley & Sons, Inc., 2011.

12. At elevated temperatures, magnetite [Fe3O4] may also form.

13. In MIC, the H2S is produced as a byproduct of the activities of sulfate reducing bacteria (SRB).

14. Chillingar et al, reference 3.

Figure 4. Corrosion rates. The relative rates of corrosion in milli-inches/year (mpy) of carbon steel show pronounced differences when the steel is exposed to varying concentrations of O2, CO2 and H2S. At a concentration of 5 ppm, O2 is almost three times more corrosive than is H2S and 30% more corrosive than is CO2. Photographs near each curve show the effects of these corrosion agents on metal surfaces.

Oilfield Review MAY 16Corrosion Fig 4ORMAY 16 CRSSN 4

Corro

sion

rate

of c

arbo

n st

eel,

mpy

30

20

10

0

Oxygen1 2 3 4 5 6 7 8 9

O2

CO2

H2S5

15

25

Hydrogen sulfideGas concentration in water phase, parts per million

100 200 300 400 500 600 700 800 900

Carbon dioxide50 100 150 200 250 300 350 400 450

15. Stewart M and Arnold K: Gas Sweetening and Processing Field Manual. Waltham, Massachusetts. Gulf Professional Publishing, 2011.

16. Ning J, Zheng Y, Young D, Brown B and Nesic S: “A Thermodynamic Study of Hydrogen Sulfide Corrosion of Mild Steel,” paper NACE 2462, presented at the NACE Corrosion 2013 Conference and Exhibition, Orlando, Florida, USA, March 17–21, 2013.

17. Kvarekval J: “Morphology of Localized Corrosion Attacks in Sour Environments,” paper NACE 07659, presented at the NACE Corrosion 2007 Conference and Exposition, Nashville, Tennessee, USA, March 11–15, 2007.

18. Pipeline operators send mechanical devices called pigs through pipelines to clean the inner surface. This can be done without halting flow, and the flow stream pushes the pig through the piping.

19. Popoola et al, reference 8. Chillingar et al, reference 3.20. Chillingar et al, reference 3. Care must be taken when using NH4HSO3 as an oxygen

scavenger. This compound is corrosive in itself and can also act as a food source for bacteria, thereby potentially encouraging MIC.

21. Davis JR: Corrosion—Understanding the Basics. Materials Park, Ohio: ASM International, 2000.

38 Oilfield Review

that caused it. Although not an exclusive list, cor-rosion usually falls into one or more of the fol-lowing categories: general or uniform, localized, galvanic, erosion or flow induced, crevice, pitting, under deposit, cavitation, intergranular, stress cracking and corrosion fatigue (Figure 5). Other types of corrosion include environmental, top-of-line and microbial. Based on the observed char-acteristics of the corrosion, engineers can adopt appropriate preventive and mitigation measures.

Uniform corrosion is typical of low-alloy steels and may be observed over an entire exposed area. Initial evidence of uniform corro-sion is surface roughness. The metal becomes thinner as the corrosion progresses, and it will eventually fail from internal pressure or external forces. Because this type of corrosion is linked to surface exposure, it may be prevented by prop-erly protecting the surface. Uniform corrosion may occur in equipment used for oilfield opera-tions such as hydraulic stimulation and acidizing.

Localized corrosion occurs at specific sites rather than over a generalized area and may be more dangerous than some other types of cor-rosion because of its unpredictable nature and

the potential for rapid growth. Localized corro-sion, of which even CRAs such as stainless steels are susceptible, can be subdivided into pitting, crevice and under deposit corrosion. Pitting ulti-mately can cause holes in metal components and is one of the primary causes of failure in oilfield hardware, including tubing, casing, sucker rods and surface equipment.

Crevice corrosion occurs in constricted areas, wherein the metal at the crevice becomes anodic and the rest of the metal serves as the cathode. The crevice can form where two dissimilar metals come into contact or be created by microgalvanic cells that may occur in certain steel alloys.

Pitting corrosion rates are often much higher than those of other types of corrosion. Inhibitors may be applied to the surface to prevent initia-tion, but once a pit has formed the inhibitors are often unable to slow its growth.

Under deposit corrosion occurs when sand, corrosives or porous solids adhere to the metal surface. Although the area underneath the deposit is resistant to inhibitors and can corrode quickly, this type of corrosion can often be man-

aged by cleaning internal piping surfaces, for example, with the use of pipeline pigs.

Galvanic corrosion can be a problem when two dissimilar metals are in contact. The metal that has the least resistance to corrosion acts as the anode and the more resistant metal serves as the cathode. The anode typically corrodes pref-erentially. This form of corrosion is frequently observed in offshore platforms and pipelines. The galvanic series, which orders metals accord-ing to their anodic or cathodic tendencies, is a good predictor of corrosion severity (Figure 6). Galvanic corrosion is controlled and mitigated by use of the following:• good engineering design—to ensure that cor-

rosively active components present larger sur-face area than do less active components

• material selection—to avoid metals far apart in the galvanic series

• isolation—to provide pipelines coming from the sea with sacrificial anodes and protect those going into land with impressed current systems

• inhibitors and coatings—to control initiation of corrosion, although this method may be inef-fective once corrosion forms.

Figure 5. Generalized categories of corrosion. Corrosion can be categorized by appearance and the agent of causation. These eight corrosion types cover most of the observed corrosion mechanisms for metals.

Oilfield Review MAY 16Corrosion Fig 5ORMAY 16 CRSSN 5

Steel

General or Uniform

Pitting IntergranularStress

Corrosion Cracking Corrosion Fatigue

Galvanic Erosion or Flow Induced Crevice

Anode

Water

Water

Flow

Cathode

ForceSteel

May 2016 39

Flow-induced corrosion occurs when liquid flow accelerates corrosion. Wellheads and pumps are susceptible to this form of corrosion, which may occur as erosion or cavitation. Erosion cor-rosion results when fluid flow removes the pro-tective film that forms naturally or has been applied externally. Because of their abrasive properties, suspended solids will accelerate the process. Damage can be seen as grooves in the piping that correspond to the flow direction. Proper engineering design that allows for suf-ficient pipe diameter and removing solids from flow streams can minimize this type of corrosion. Inhibitors may be applied to replace protective films stripped away by the flowing fluids.

Cavitation is caused by collapsing bubbles that occur when the pressure changes rapidly in flowing liquids. Over time, cavitation may cause deep pits to form in areas of turbulent flow, espe-cially in pump impellers. Low-carbon steels are susceptible; stainless steels are more resilient.22

Intergranular corrosion results from cor-rosive attacks at metal grain boundaries in the form of cracks. The grain boundaries can become anodic with reference to the cathodic surround-ing surface, typically due to formation of chro-mium carbides or nitrides. Metal impurities can increase the effect, as can precipitates in the metal that form during heat treatments. When chromium combines with nitrogen or carbon, less free chrome is available locally for corro-sion protection, and cracks can form along the grain boundaries. Quenching—the rapid cool-ing after heat treatments—may be effective in reducing or eliminating intergranular corrosion. Material selection—avoiding metals that are susceptible to this condition—is the most reli-able method to preclude intergranular corrosion. Tests such as ASTM A262 can be used to evaluate susceptibility of materials to this mechanism.23

Environmental cracking occurs when cor-rosion coincides with tensile stress. It may be manifested as the following:

• hydrogen embrittlement—hydrogen enters the metal matrix and weakens it

• stress corrosion cracking—cracks form after corrosion has attacked a surface

• sulfide stress cracking—a failure of the metal caused by H2S.

Material selection—opting for materials that are resistant to hydrogen embrittlement and sulfide cracking—is the primary avoidance technique. Low-stress design practices and stress relief by heat treatment are also commonly used, and pre-venting corrosion in components subject to stress is another method.

Pipelines are subject to top-of-line corrosion (Figure 7). Water condenses at the top of the pipe as the fluid inside cools. The corrosion rate depends on the condensation rate and concen-tration of organic acids. Generally, this type of corrosion is controlled with inhibitors and pipe-line insulation that reduces condensation.

22. Port RD: “Flow Accelerated Corrosion,” paper NACE 721, presented at the NACE Corrosion 98 Annual Conference, San Diego, California, USA, March 22–27, 1998.

23. ASTM International: “Standard Practices for Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels,” West Conshohocken, Pennsylvania, USA: ASTM International A262-15, 2015.

Oilfield Review MAY 16Corrosion Fig 6ORMAY 16 CRSSN 6

Magnesium

Anodic

Zinc

Cadmium

Aluminum

Steel

Chromium steel

Stainless steel

Lead

Tin

Nickel

Inconel

Hastelloy

Brasses

Copper

Bronzes

Monel

Chromium steel

Silver

Titanium

Graphite

Gold

Platinum

Cathodic

Figure 6. Galvanic series. Metals (not all shown) can be described by their anodic or cathodic tendencies arranged in a galvanic series. When dissimilar metals are connected electrically and submerged in an electrolyte, the anodic metal, rather than the cathodic metal, will preferentially corrode. The rate of corrosion is a function of the separation between the paired metals in the galvanic series. The series shown here is for seawater; the order may change based on the electrolyte.

Figure 7. Top-of-line pipeline corrosion. Top-of-line corrosion can result from the stratified multiphase flow of wet gas in horizontal pipelines. Liquids—including condensate and inhibitors such as monoethylene glycol—settle to the bottom of the pipe. Wet gas fills the pipe above the liquid line. If either CO2 or H2S are present in the gas, along with water, corrosive byproducts form at the top of the pipe and may not be controlled if the inhibitor remains at the bottom of the pipe.

Condensate

Wet gas

Pipe

Monoethylene glycol

Oilfield Review MAY 16Corrosion Fig 7ORMAY 16 CRSSN 7

40 Oilfield Review

Microbiologically Influenced CorrosionFrom the moment of immersion in a nonster-ile fluid that supports microbial development, microorganisms begin to attach to the material surface (Figure 8). Planktonic microorganisms become fixed at the fluid boundary, typically a pipe wall, or onto porous media such as preexist-ing corrosion.24

Attachment of microorganisms leads to the formation of biofilms—microbial communities

embedded in or on an attaching surface.25 Wher-ever biofilms are found in producing systems, MIC can occur, including inside production tub-ing, gravity and hydrocyclone separators, storage tanks, pipelines and water injection systems. Depending on the microbial species, the corro-sion mechanisms can take various forms.

Biofilms can trap ions and create localized electrochemical potentials analogous to a gal-vanic corrosion cell or may contribute to corro-

sion by taking a role in the formation of cathodic and conductive corrosion products on the metal surface (Figure 9). Sulfate producing prokaryotes (SPPs) are the chief offenders. Prokaryotes are microbes that have no cell nucleus or membrane-bound organelles. The most prominent group of SPPs are sulfate reducing bacteria (SRB) and sulfate reducing archaea (SRA). They contribute to corrosion by various means, for example, by taking a role in the formation of cathodic ferrous sulfide corrosion products and the formation of galvanic cells. The production of H2S by SPPs can also lead to sour corrosion.

Biofilms may develop local concentration cells that are created by oxygen depletion or may attach to a metal surface. Microbes can contrib-ute to corrosion by the direct effects of meta-bolic waste products such as organic acids that are capable of altering the local pH and forming pH cells. Some microbes are anaerobic and can tolerate extremes of pressure, temperature, pH and fluid salinity. These include methanogens—microbes that produce methane as a metabolic byproduct in anoxic conditions.

Regardless of the source of MIC, prevention measures in most cases attack planktonic and sessile populations.26 Methods include biocides to kill the microorganisms, coatings to inhibit biofilm formation, removal of nutrients from the flow stream to control microbial populations and mechanical removal of an established biomass via pigging.

Corrosion Control MethodsMetallurgical solutions can be effective deter-rents of corrosion, but their costs may be beyond the economic limit of many oilfield projects. Building every structure and tubular from irid-ium—the most corrosion resistant element—might win the battle against corrosion, but incur unsustainable expenses, and that would be assuming a sufficient supply of iridium exists in the world to attempt such a task. Aluminum is a corrosion-resistant metal used in many oilfield applications; however, it is unsuitable for high-pressure and high-temperature operations. In addition, although aluminum is considered cor-rosion resistant in seawater, the mechanism for resistance relies on the formation of a thin film of aluminum oxide on the surface of the metal. In environments that have high levels of acidity (low pH) or alkalinity (high pH), the aluminum oxide can become unstable and thus nonprotective. In many cases, steel alloys and CRAs are required to meet both strength and cost requirements.

Figure 8. Microbiologically influenced corrosion products. Softscale corrosion, referred to as schmoo (right ), can form in production systems if microbes are not controlled. The photograph shows a mixture of iron sulfide [FeS], asphaltenes and biomass that was collected at the sidestream outlet of a separator tank (top left ). Corrosion inhibitors form protective films around iron sulfide particles (bottom left ) inside the separator and prevent softscale formation in produced waters.

Water

Oil outletWater outlet

Sidestream Emulsion

Separator Tank

Oil layer

Iron sulfideparticle

Corrosioninhibitor film

Oilfield Review MAY 16Corrosion Fig 9ORMAY 16 CRSSN 9

Figure 9. Microbiologically influenced corrosion from byproducts. Biofilm on the surface of this metal piece produced H2S that damaged the piece and led to premature failure of the equipment.

Oilfield Review MAY 16Corrosion Fig 8ORMAY 16 CRSSN 8

May 2016 41

24. Stipanicev M: “Improved Decision Support Within Biocorrosion Management for Oil and Gas Water Injection Systems,” PhD thesis, Institut National Polytechnique de Toulouse, France (2013).

25. Madigan MT, Martinko JM, Bender KS, Buckley DH, Stahl DA and Brock T: Brock Biology of Microorganisms, 14th ed. San Francisco: Benjamin Cummings, 2014.

26. Sessile refers to fixed or immobile organisms.27. Lehmann JA: “Cathodic Protection of Offshore

Structures,” paper OTC 1041, presented at the First Annual Offshore Technology Conference, Houston, May 18–21, 1969.

28. Davy H: “On the Corrosion of Copper Sheeting by Sea Water, and on Methods of Preventing This Effect; And on Their Application to Ships of War and Other Ships,” Philosophical Transactions of the Royal Society of London 114 (January 1, 1824): 151–158.

29. von Baeckmann W: “The History of Corrosion Protection,” in von Baeckmann W, Schwenk W and Prinz W (eds): Handbook of Cathodic Corrosion Protection—Theory and Practice of Electrochemical Protection Processes, 3rd ed. Houston: Gulf Professional Publishing (1997): 1–26.

30. Amani M and Hjeij D: “A Comprehensive Review of Corrosion and Its Inhibition in the Oil and Gas Industry,” paper SPE 175337, presented at the SPE Kuwait Oil and Gas Show and Conference, Mishref, Kuwait, October 11–14, 2015.

Although materials selection is a major part of the corrosion control process, once the equip-ment is deployed, oilfield operations generally follow three methodologies to battle corrosion. Operators and service companies rely on surface coatings to protect susceptible metals, cathodic protection for active protection and inhibitors as a low-cost treatment option.

Surface coatings provide chemical and mechanical resistance. They may also offer ther-mal protection. For surface coatings to provide maximum effectiveness, good adhesion to the target surface is required. Coatings are available in organic and inorganic types. Organic coatings include epoxies, phenolic resins, polyurethanes, polyethylenes and polyesters. Metals applied as suspensions and electroplating are examples of inorganic coatings; inorganic ceramics may also be applied to protect surfaces. Although not normally an advanced-technology solution, the cement placed in the annulus between the wellbore casing and the formation can act as an inorganic coating that prevents corrosion.

Cathodic protection (CP) consists of two pri-mary forms: passive and active (Figure 10). In either form, it relies on a movement of electrons (current) from an external anode to the equip-ment being protected, which acts as a cathode. Both the cathode and anode must be in the same electrolyte and electrically connected. The most common uses of CP are protecting large struc-tures, piping, casing and equipment exposed to the elements. It may also be installed inside or outside tanks and pressure vessels.

Operators often use sacrificial anodes with CP to protect structures in areas where electrical power sources are not readily available such as in remote operations or on offshore structures. If the structure can be made to serve as the cathode in relation to an anode, the disposable sacrificial anode will corrode while the cathode remains unscathed. This type of CP has been referred to as fighting corrosion with corrosion.27

The first use of CP is attributed to Sir Humphry Davy, who described the process in a series of articles to the Royal Society of London in 1824.28 The technique was used in an attempt to prevent the corrosion of nails used in wooden oceangoing vessels. Accelerated corrosion of the nails occurred when copper cladding—used to prevent biofouling—was applied to the outside of vessels. Davy found that sacrificial anodes protected the iron nails. The actual processes were not well understood at that time, but it is recognized today that the contact of dissimilar metals—the copper cladding and the iron of the

nails—led to the corrosion. Davy and his assis-tant carried out a number of experiments on cor-rosion prevention techniques; that assistant was Michael Faraday, who would later establish the relationship between the chemical action of cor-rosion and electric current.

In the oil field, CP was first applied to land-based pipelines, and the first documented use was by Robert J. Kuhn in 1928.29 He established a negative 850 mV potential between the steel pipe of a pipeline and a copper-sulfate electrode. This example became the foundation of modern CP technology, although for many years the effec-tiveness was met with scientific skepticism.

Today, CP uses sacrificial elements made from aluminum, zinc and magnesium to protect the steel of large structures and piping. These dissimilar metals create the galvanic coupling that establishes a current path between the anode and the cathode, and, over time, the sacrificial anode rather than the protected structure experiences metal loss. Appropriate placement and distribution of the anodes is cru-cial to ensure that all parts of the structure are sufficiently protected.30

Because the direct current (DC) is exter-nally applied, this type of corrosion manage-ment is referred to as impressed cathodic protection. It is most frequently used for cases in which the electrolyte resistance is high, such as in soil or freshwater, and where a constant

Figure 10. Cathodic protection circuit. Cathodic protection methods may use naturally occurring galvanic current or employ a direct current (DC) source (impressed current) when the electrolyte is resistive. The protected element—a pipeline is shown—is the cathode. The sacrificial element, located some distance from the cathode, serves as the anode. The DC source may be batteries or solar panels in remote pipeline applications.

Cathode

Anode

Backfill

Galvanic DC current

Electrolyte

Oilfield Review MAY 16Corrosion Fig 10ORMAY 16 CRSSN 10

milliamp

42 Oilfield Review

source of current is readily available. The use of solar panels in remote locations has greatly increased the potential applications of impressed cathodic protection.

In the impressed CP technique, current of several amps from a low-voltage rectifier passes, or is impressed, from an inert anode (for exam-ple, graphite or iron) to the structure being pro-tected, which acts as the cathode. The anode is attached to the positive terminal of the DC source, and the cathode is attached to the nega-tive terminal. The anode and cathode are often some distance from each other, separated by an electrolyte.

To counteract corrosion, sufficient current density must be supplied to all parts of the pro-tected structure and the current density must always exceed what would be the measured cor-rosion rate under the same conditions. If the corrosion rate increases, the impressed current density must be increased.31 Although the initial equipment cost may be higher for impressed CP than it is for sacrificial protection, this technique may be less expensive over the long term because sacrificial anodes do not need to be replaced. Impressed CP also has the advantage of provid-ing information to the operator about the extent of corrosion over time.

Corrosion InhibitorsAnother line of defense against corrosion is inhibitors, of which there are a variety of types and applications. The primary goal of inhibitors

is to interrupt the electrochemical process by which the corrosion cell forms between the metal and the liquids in and around the equipment. Inhibitors can be a flexible and cost-effective method of fighting corrosion, and the inhibi-tor application can be altered when conditions change. Although acquiring and delivering the inhibitor incur an ongoing cost, the lower costs associated with using less corrosion resistant low-carbon steels usually more than make up the difference.

Inhibitors fall into four main categories: scav-engers, reactive agents, vapor phase and film formers. Oxygen scavengers are frequently used in operations in which oxygen poses a corrosive threat. These agents not only reduce oxidizing corrosion, but also control the growth of microbes that require oxygen to thrive. Examples of oxygen scavengers used in the oil and gas industry are sodium sulfite, sulfur dioxide, sodium bisulfite, sodium metabisulfite and ammonium bisulfite. Ammonium bisulfite and sodium bisulfite are commonly used in seawater injection systems. To speed reaction rates, a catalyst may be included in the chemical.

Hydrogen sulfide scavengers reduce the level of H2S in the flow stream. Examples of H2S scavengers are amines, aldehydes and zinc car-boxylates. Common forms of amines are mono-ethanolamine (MEA) and monomethylamine (MMA) triazine. In some situations, operators may be able to regenerate MEA and MMA for reinjection and reuse.

Reactive inhibition operates at the cathode level for the corrosion cell. The cations of the inhibitor react with the cathodic anions to form insoluble films, which adhere to the surface of the metals and prevent O2 from coming into contact with the metal. These films also prevent the evolution of H2, a byproduct of the corrosion cell. Examples are forms of calcium carbonate, magnesium carbonate and iron oxides. Reactive inhibitors can also serve as poisons to the corro-sion cell process by interfering with the forma-tion of H2 and reducing the reaction rates at both the cathode and anode.

Vapor phase inhibitors are primarily used for combating CO2 corrosion. These inhibitors neu-tralize CO2 and block the formation of carbonic acid [H2CO3]. They are transported via vapor phase in wet gas lines. To protect against future corrosion, they may also be used during hydro-static testing of components with water, espe-cially when the components are to be stored after fitness testing. Examples of these types of inhibi-tors include morpholine and ethylenediamine.

Film FormersFilm formers are the most widely used corrosion inhibitors in the oil and gas industry. They cre-ate a continuous layer between the metal and the reactive fluids, thus reducing the attack of corro-sive elements (Figure 11). They may also attach to the surface of corroded metal, altering it and reducing the corrosion rate. Although they are effective in reducing CO2 and H2S corrosion, film formers are not effective against O2 corrosion.

Film formers are available in oil-soluble, water-soluble and oil soluble–water dispersible forms. Oil-soluble inhibitors are used to treat oil- and gas-producing wells. Water-soluble inhibi-tors are used in high water-cut flow streams, including those found in producing wells, trans-mission lines and separators. Oil soluble–water dispersible inhibitors are used in oil and gas wells that are also producing water.

Film-forming inhibitors take various chemical forms but are typically composed of long carbon chains with nitrogen, phosphate esters or anhy-drides. Inhibitors may adhere to or be adsorbed on the metal surface, which prevents the corro-sives from attacking the metal. The most effec-tive film-forming inhibitors create a molecular bond at the metal surface in a process of charge sharing or charge transfer. For effective inhibi-tion, the surface of the metal being protected must be fully covered; injection of the proper concentrations of the inhibitor are crucial. After they interact with the corrosive elements, some

Figure 11. Film formers. Although they vary in composition and avenue of protection, film formers create barriers between corrosive elements (water and oil, top) and metal surfaces. Inhibitors may be adsorbed on the surface (alkyl chains, middle) or form a strong bond by sharing charges with the metal (polar head group, bottom). When molecules of the polar head group of film formers attach to the surface of the metal, a portion of the molecule extends into the fluid. This usually oil-soluble tail is hydrophobic, repelling water away from the metal surface.

Oilfield Review MAY 16Corrosion Fig 11ORMAY 16 CRSSN 11

Water

Oil

Alkylchain

Polar headgroup

Metal surface

Water

Oil

CH3

CnH2n

N+

May 2016 43

inhibitors are gradually removed from the metal surface and must be continuously replenished with new inhibitor.

In the petroleum industry, organic inhibitors are frequently used because they can form protec-tive layers even in the presence of hydrocarbons. Amides and imidazolines are examples of organic film-forming inhibitors that are effective over a wide range of conditions, especially in sweet (CO2) and sour (H2S) gas corrosion environments. They can be water or oil soluble. Amines, which are also organic inhibitors, are effective for sweet and sour corrosion but may exhibit biologic toxic-ity and are thus not as environmentally friendly as are amides.

Quaternary ammonium salt, or quaternary amine, inhibitors are effective against sour corro-sion.32 The corrosive element formed by sour gas is iron sulfide on the metal surface. Quaternary ammonium cations, or quats, are positively charged, and when they are adsorbed on the surface of the material to be protected, they dis-rupt the normal corrosion cell charge. However, at least one study indicated that quaternary ammonium inhibitors may actually increase the corrosion rate of sweet corrosion in the presence of brine.33 The biocide properties of quaternary ammonium salts may also prevent MIC.

Many additional film formers are used in the oil and gas industry, including phosphate esters, ester quats, dimer and trimer acids and alkyl pyr-idine quaternary compounds. Most film-forming applications include multiple inhibitors; labora-tory testing is used to establish optimum concen-trations, fluid tolerances, stability, effectiveness and persistency of the film. Inhibitor selection can be a complicated process and typically must be adjusted over time to meet the demands of changing fluid conditions.

Inhibitor SelectionLaboratory evaluation is the key to developing an effective program in inhibitor selection for corrosion control. Technicians begin the process using fluid samples that replicate field condi-tions—actual produced fluids are best if avail-able. Simulated and synthetic fluids are used when produced fluids cannot be obtained. From laboratory tests, corrosion rates can be mea-sured and predictions can be made for large- scale operations (Figure 12). Methods for test-ing corrosion inhibitors include the following tests: wheel, kettle (also called linear polariza-tion resistance (LPR) tests), rotating cylinder

31. Schweitzer PA: Corrosion of Linings and Coatings: Cathodic and Inhibitor Protection and Corrosion Monitoring. Boca Raton, Florida: CRC Press, 2006.

32. Binks BP, Fletcher PDI, Hicks JT, Durnie WH and Horsup DI: “Comparison of the Effects of Air, Carbon Dioxide and Hydrogen Sulphide on Corrosion of a Low Oilfield Review

MAY 16Corrosion Fig 12ORMAY 16 CRSSN 12

Use of corrosioninhibitor recommended

Predict corrosion ratefrom field data

Develop and executetesting program

Rotating CylinderElectrode Test

Sidestream Test

Kettle Test Autoclave Test

Recommend and implement corrosion

inhibitor addition

Conduct field trial

Analyze fieldtrial results

Make finalrecommendation

Figure 12. Laboratory testing of corrosion inhibitors. Operators usually develop corrosion control plans and then test inhibitors using conditions expected from the field. This flowchart follows a testing sequence. Three common testing methods are the rotating cylinder electrode, kettle and autoclave tests (middle). Even after laboratory testing, field trials should be conducted to validate the effectiveness of the program. A sidestream test (lower left ) acquires samples for analysis. If the proposed method provides acceptable results, the method is adopted, although the corrosion inhibition program must be reevaluated during the life of the well.

Carbon Steel under Water and Its Inhibition by a Quaternary Ammonium Salt,” paper NACE 05307, presented at the NACE Corrosion 2005 Conference and Exhibition, Houston, April 3–7, 2005.

33. Binks et al, reference 32.

44 Oilfield Review

electrode, autoclave, jet impingement and flow loop (Figure 13). The most common are the wheel and kettle tests.34

The wheel test measures the loss of metal during a specified period of exposure to corro-sive liquids. Corrosives include produced fluids, brines and refined oils. The test fixture includes a rotating wheel inside a sealed box that keeps the specimen, usually strips of metal or coupons, in constant motion. Temperature can be main-tained at a constant value or varied to simulate field conditions. The samples are tested with and without inhibitor and the results are compared.

The kettle test, or LPR test, measures corro-sion rates electrochemically. Metal electrodes are placed in the test vessel, which is heated while the corrosive fluid is continously agitated. Agitation attempts to replicate field conditions—mild agitation is similar to flow of two distinct sources, and high agitation replicates turbulent fluid flow that has dispersed hydrocarbons. To simulate the presence of gases, CO2 and H2S can be bubbled through the liquid in the vessel in a process referred to as sparging.35

To establish a control corrosion rate, the test is run with the electrodes exposed to the fluids in aqueous phase without an inhibitor and then followed by a series of tests on solutions that have increased inhibitor volumes. Linear polarization is performed by controlling the voltage potential and measuring the current then controlling the

Figure 13. Autoclave corrosion testing. A high-temperature autoclave is used to test the effectiveness of corrosion inhibitors on metal coupons (inset ). Hydrostatic pressure and temperature can be applied to simulate downhole conditions.

Rack

Pressure source(hydraulic pump)

High-temperature autoclave

Metal coupon container

Figure 14. Kettle test. To perform kettle tests, or linear polarization resistance tests, technicians use a test fixture (left ) and control the pressure and temperature. They submerge electrodes inside the fixture into the fluids expected downhole and then measure electrical properties of the electrodes. The tests are performed by controlling the voltage potential

and measuring the current then controlling the current and measuring the voltage. The electrolyte can be agitated using the stir bar. Gas can be injected into the test fixture, a process referred to as sparging. From the slope of the polarization resistance curve (right ), the corrosion rate can be computed.

Current, mA

Volta

ge p

oten

tial,

mV

Steel electrode

Heating mantel

Stir bar

Thermometer

Gas sparge

Corrosion rateVoltageCurrent

May 2016 45

34. NACE Task Group T-1D-34 on Laboratory Corrosion Inhibitor Test Parameters: “Laboratory Test Methods for Evaluating Oilfield Corrosion Inhibitors,” Houston, NACE International, NACE Publication 1D196, December 1996.

35. NACE Task Group T-1D-34 on Laboratory Corrosion Inhibitor Test Parameters, reference 34.

36. Efird KD: “Jet Impingement Testing for Flow Accelerated Corrosion,” paper NACE 00052, presented at NACE Corrosion 2000 Conference and Exhibition, Orlando, Florida, March 26–31, 2000.

37. Melot D, Paugam G and Roche M: “Disbondments of Pipeline Coatings and Their Effects on Corrosion Risks,” Journal of Protective Coatings and Linings 26, no. 9 (September 2009): 67–76.

current and measuring the voltage potential. The data are plotted, and the slope of the line is the polarization resistance, which is inversely proportional to the corrosion rate (Figure 14). This technique provides corrosion rate evalua-tion from external measurements, whereas other methods require technicians to physically mea-sure and evaluate corrosion.

The effectiveness of inhibitors is dependent on fluid velocity. For fluids containing little or no solid particles, high flow rates can lead to flow-accelerated corrosion. If the flow stream contains solid particles, the accelerated corro-sion is termed erosion corrosion. Several test methods have been developed to model corrosion in high-flow conditions and determine a film’s persistence, especially where turbulent flow is present.36 Test methods include jet impingement, rotating cylinder electrodes and flow loop testing.

The testing of inhibitors should determine the following:

• thermal stability• emulsification tendency• foaming tendency• metal compatibility• elastomer compatibility• compatibility with other chemicals used in the

same stream.Application methods should be evaluated as well. Injection may be continuous, batch or squeeze. The rate of film removal is a key concern when determining the optimal application mode.

Corrosion in the Oil FieldA recent example of pipeline corrosion from Gabon illustrates the need for thorough testing and understanding of the corrosion process.37 A pipeline transports oil from the Rabi field to Cape Lopez—a distance of approximately 234 km [145 mi]. The 18-in. pipeline comprises three sections: Section 1 from Rabi to Batanga, 105 km [65 mi]; Section 2 from Batanga to Tchengué,

100 km [62 mi] and then Section 3 from Tchengué to Cape Lopez, 29 km [18 mi] (Figure 15).

The inlet pressure at Rabi was about 40 bar [580 psi], and the flowing temperature was 60°C [140°F] at the inlet. Beyond the inlet, the line oper-ates at about 35°C [95°F]. Impressed cathodic pro-tection is used for the pipeline, which has sections that have solar cells to provide current. The pipeline was coated with three-layer polyethylene; each joint was brush cleaned and wrapped with heat-shrink

Figure 15. Corrosion in a pipeline from the Rabi field to Cape Lopez. A three-section, 18-in. pipeline carries oil from the inland Rabi field in Gabon to Cape Lopez on the coast. Cathodic protection stations are located along the pipeline. Because the incoming oil is hot (around 60°C), Section 1 of the pipeline (red and dark blue) is exposed to a higher temperature than is the remainder of the pipeline. The elevated temperature led to the disbondment

of the protective outer covering of the pipeline. Engineers concluded that corrosion observed in Section 1 resulted from a combination of the disbonding of the protective covering and ineffective cathodic protection. Although the pipeline’s safety was not compromised, the operator implemented new procedures to prevent the corrosion from recurring.

Cape LopezCathodic protection station

TchéngueCathodic protection station

BatangaCathodic protection station

Rabi FieldCathodic protection station

Input temperature 60˚ C

Cathodic protection stationspowered by two solar cells

Section 1

Section 2Section 3

GABONATLANTIC OCEAN

~ ~

46 Oilfield Review

sleeves and hot-melt adhesive that overlapped the polyethylene.38 The pipe was buried in wet, com-pacted sand that had a pH of approximately 5.4.

Problems began to develop in the first 15 km [9.3 mi] of pipe. Routine inspections found dis-bondment at the sleeves where they overlapped the polyethylene coating in the Rabi section. Disbondment allowed water under the protective coating, which negated the cathodic protection and allowed corrosion to develop (Figure 16). The remainder of the pipeline did not experience the same level of corrosion, although significant

disbondment occurred in the sleeves. The main difference between the sections that had differ-ing corrosion levels was that the temperature in the more corroded sections was higher. Further testing of pipe Sections 2 and 3 found no evidence of similar levels of disbondment or corrosion.

After a thorough examination, engineers recom-mended abrasive blast cleaning prior to applying heat-shrink sleeves for future installations rather than the standard brush cleaning of connections. Another possible solution was liquid polyurethane or epoxy applied at the joints. The disbonded coating

may have prevented cathodic current from reaching and protecting the surface of the exposed steel. Although engineers discovered corrosion as a result of disbonding of coatings, based on ASME standards, the degree of corrosion was deemed not mechani-cally dangerous. They also concluded that as long as coatings remain bonded to the steel and cathodic protection is correctly applied, monitored and main-tained, no corrosion risk existed for this pipeline.

Corrosion Inhibitors in DeepwaterDeepwater projects can pose unique challenges for corrosion control because the completions are usually located at the seafloor and flowlines must come to the surface or back to shore. A deepwater field located in the southern Niger delta demonstrates the use of inhibitors to com-bat CO2-induced corrosion (Figure 17).39

The production path for deepwater wells passes through cold water, which can subject the originally hot fluids in the flow stream to rapid cooling. Conversely, inhibitor injection is often through long umbilicals that are subjected to temperature contrasts between the surface and subsea wellheads. Injection of inhibitors is fur-ther complicated by the normally high flowing pressures associated with deepwater production.

Temperature extremes, pressure extremes and long umbilicals combine to affect inhibitor

Figure 16. An example of pipeline corrosion. After the protective coating disbonded on a pipeline in Gabon, corrosion formed as pitting (inset ).

Figure 17. Niger delta subsea operations and a floating production, storage, and offloading (FPSO) vessel. Production from subsea wellheads (yellow) at a field in the Niger delta off the coast of Nigeria (inset) is sent to an FPSO. Oil is transferred to tankers, and natural gas is piped directly to the mainland.

Niger Deltafield

NIGERIA

CAMEROON

BURKINA FASO

COTED’IVOIRE

GHANA

BENIN

GABON

FPSO vessel

Flowlines andumbilicals

Subseawellheads

Gulf of Guinea

May 2016 47

stability, performance and properties. Thorough testing of the inhibitors is required to ensure corrosion controlling properties are maintained, that the injected chemicals remain stable and that the inhibitors can be reliably delivered via the umbilicals into the flow stream.

Another risk in deepwater production is the formation of hydrates—ice-like solids of water and gas that form above the normal freezing point of water—that can plug flowlines. To ensure cor-rosion controlling properties are maintained, inhibitors must be thoroughly tested to confirm that the injected chemicals remain stable and that the inhibitors can be reliably delivered via umbilicals into the flow stream.

These conditions were faced by an operator of a deepwater production platform in Nigeria. The platform served nine wells drilled in water depth of 1,030 m [3,380 ft]. The operator used subsea completions that included five manifolds and eight production flowlines and risers. The flowlines were connected to a floating produc-tion, storage, and offloading (FPSO) vessel that had 320,000 m3 [2 million bbl] of onsite storage capacity. Produced oil flowed to the FPSO for transfer to tankers. Produced gas was directed to shore via pipelines.

The pipelines used to transport the oil and gas were constructed of carbon steel. The flow-ing pressure from the wells averaged 80 bar [1,160 psi], and the average temperature was 85°C [185°F]. The water cut was 45% and the natural gas contained about 1.4% CO2. The com-bination of produced water (brine) and CO2 presented a high corrosion-rate potential for the low-carbon steel. In wet gas pipelines such as these, produced water has a tendency to con-dense at the top of the pipe, allowing top-of-line corrosion; the presence of both water and CO2 accelerates corrosion.

Engineers installed chemical umbilicals of 1 to 20 km [0.6 to 12.4 mi] to inject corrosion inhibitor into the deepwater production flow-lines. As the project progressed, engineers at M-I SWACO, a Schlumberger company, reevalu-ated the initial inhibitor used in the project, which also protected the topside piping and storage vessels, and deemed it to be insufficient.

Engineers developed the DS-1617 deepwater corrosion inhibitor to meet the challenges of this facility.

To qualify this inhibitor, they tested the chemicals in accordance with the API TR 17TR6 standard, which requires replicating the temper-atures and pressures experienced by the inhibitor during deployment through the umbilicals.40 The evaluation included high-pressure flow-loop sta-bility tests. The engineers conducted additional tests to look at resistance to hydrate formation, thermal aging and compatibility with seawater. Because the operator was concerned about foam-ing in the glycol regeneration unit, the inhibitor was tested for foaming tendency.

Laboratory technicians performed kettle tests using the DS-1617 inhibitor at 20 ppm, which is a relatively low dosage; the corrosion rate was reduced by 99% (Figure 18). They also performed high-temperature autoclave testing on carbon steel coupons. The samples were sub-jected to test fluids that had corrosion inhibitor

and to pressurized CO2 heated to 120°C [248°F]. The results indicated a 98% reduction in the corrosion rate.41 The 20 ppm concentration yielded corrosion rates of about 0.00016 in./year [0.004 mm/year]. For corrosion rate, the stan-dard industry units are milli-inches/year, or mpy. For this test, the corrosion rate was equivalent to 0.16 mpy. Test technicians reported no foaming problems associated with the inhibitor.

The operator adopted the use of the DS-1617 inhibitor and monitored corrosion at six loca-tions on the FPSO. No corrosion monitoring was installed on the deepwater flowlines. The DS-1617 inhibitor was injected at a 100-ppm rate, which is a lower rate than the initial inhibi-tor that was deemed insufficient. Criteria for corrosion protection established by the opera-tor was a rate below 0.05 mpy [0.0013 mm/year]. Testing at all six locations indicated corrosion rates below the target rate (Figure 19). Based on the testing, the operator implemented the use of the DS-1617 inhibitor.

38. Roche M: “The Problematic of Disbonding of Coatings and Corrosion with Buried Pipelines Cathodically Protected,” presented at the 10th European Federation of Corrosion, Nice, France, September 12–16, 2004.

39. Jenkins A: “Corrosion Mitigation in a Deepwater Oilfield Case Study,” paper IBP1194_15, presented at the Rio Pipeline Conference and Exposition, Rio de Janeiro, September 22–24, 2015.

40. API: “Attributes of Production Chemicals in Subsea Production Systems,” Washington, DC: API, API Technical Report 17TR6, 2012.

41. Jenkins, reference 39.

Figure 18. Corrosion testing of the DS-1617 inhibitor. Technicians conducted kettle tests with fluids representative of field conditions to evaluate the effectiveness of the DS-1617 inhibitor (top). They also performed autoclave testing at high temperature (bottom). The corrosion rate is in milli-inches of penetration/year (mpy).

Inhibitor Dose Rate, ppm Uninhibited Corrosion Rate, mpy

Inhibited Corrosion Rate, mpy

Protection, %

DS-1617 inhibitor 10 173.01 4.18 97.58

DS-1617 inhibitor 20 156.43 0.98 99.37

Inhibitor Dose Rate, ppm Corrosion Rate, mpy Protection, %

None — 71.04 —

DS-1617 inhibitor 20 1.16 98.37

Figure 19. Corrosion monitoring at a production facility. The operator injected DS-1617 inhibitor into the flowlines of producing wells using underwater umbilicals. The corrosion rate of the flowlines was monitored at the low-pressure separator A (blue), low-pressure separator B (red) and the bulk oil treater (green) as well as at three other locations (not shown). The corrosion rate dropped below the target level (black) established by the operator. Corrosion rates remained below the threshold at all test sites for the duration of the testing period.

6.0Low-pressure separator A

Low-pressure separator B

Bulk oil treater

Target corrosion rate

0.5 1.00

0.0

1.0

2.0

3.0

4.0

5.0

1.5 2.0 2.5 3.0Time, hr

Corro

sion

rate

, mpy

3.5 4.0 4.5 5.0 6.0

48 Oilfield Review

North Sea Cathodic ProtectionNorth Sea production platforms routinely use cathodic protection. On one platform, the opera-tor installed 10 sacrificial anodes below the sur-face of the water and left them in place for eight years. The anodes were composed of zinc, silver and silver chloride and were located at various depths and locations on the plaform.42 The system was designed to protect the structure for a mini-mum of 20 years. Engineers monitored the output current from three of the anodes over the period. The anodes were removed and inspected at the end of eight years.

After retrieval, the sacrificial anodes were cleaned and weighed (Figure 20). Technicians

determined changes in physical dimensions and measured electrical properties. Four anodes were analyzed for the study. The reduction of anodes that had been placed in deeper water was greater than that of those placed in shallower water. Some of the anodes were so corroded that visual inspection was difficult (Figure 21).

The original 20-year design projected that at eight years, the anodes should be reduced by 40%; however, the average weight loss of the anodes was only 24%. The engineers concluded that the original design, although conservative, would protect the structure for at least 20 years. Based on the results of the study, a model was estab-lished for periodic inspections to be performed.

New Developments in Corrosion ControlControlling corrosion has been an ongoing battle between humans and nature for millen-nia. Since scientists such as Sir Humphry Davy and Michael Faraday discovered some of the underlying physics that explained corrosion, various methodologies have been adopted and adapted. Modern scientific understanding and new technologies are combining to improve the tools available to fight the unending battle with corrosion.

One area of emerging materials science is nanoparticles and nanostructures.43 Having sur-face thickness of 1 to 100 nm, these coating mate-rials have unique properties that may make them almost impervious to corrosion. Nanoparticles and nanostructures may be deposited on metal surfaces as films, similar to film-forming tech-niques, but because of nanoparticles’ greater persistence, reapplying them is unnecessary. The surfaces also become super-slick—exhibiting low friction coefficients—which reduces wear and increases durability. Such surfaces are also less likely to experience biofouling.44

The battle against corrosion will never be won; entropy will eventually win the war. Humans will, however, continue to search for effective means to combat this nemesis. The costs of ignoring the problem are too great and the consequences of failure can be potentially catastrophic. At least in the oil field, operators are armed with knowledge, science and effective tools that allow them to actively manage or miti-gate the effects of corrosion. —DEA/TS

42. Roche M: “Offshore Cathodic Protection: The Lessons of Long-Term Experience,” paper OMC-2005-020, presented at the 7th Offshore Mediterranean Conference and Exhibition, Ravenna, Italy, March 16–18, 2005.

43. El-Meligi AA: “Nanostructure of Materials and Corrosion Resistance,” in Aliofkhazraei M (ed): Developments in Corrosion Protection. Rijeka, Croatia: InTech (2014): 3–23.

Figure 21. Cathodic protection on a North Sea platform. Anodes were recovered after eight years of service from a North Sea platform. After the anodes were cleaned and weighed, technicians were able to determine the effectiveness of the anodes at protecting the structure.

Figure 20. Anode corrosion after eight years of service in the North Sea.

Anode Water Depth, m Weight Loss, %

1 13 13

2 73 31

3 116 25

4 116 39

44. Tesler AB, Kim P, Kolle S, Howell C, Ahanotu O and Aizenberg J: “Extremely Durable Biofouling-Resistant Surfaces Based on Electrodeposited Nanoporous Tungstite Films on Steel,” Nature Communications 6, no. 8649 (October 20, 2015).

May 2016 49

Nausha Asrar is the Manager for Materials Support and Failure Analysis at the Schlumberger Houston Pressure and Sampling and Formation Evaluation Centers in Sugar Land, Texas, USA. He began his career with Schlumberger in 2005 as a senior materi-als scientist. He previously worked for Shell Global Solutions in the US, the Saudi Basic Industries Corporation Technology Center and Saline Water Conversion Corporation, both in Saudi Arabia, and as principal corrosion engineer at the Research and Development Center for Iron and Steel for the Steel Authority of India, Ltd. A NACE certified material selection and design specialist, Nausha is a member of NACE, ASM and SPE as well as a life member of the Indian Institute of Metals; he is the author of more than 60 technical papers and reviews on corrosion, phase diagrams, composite materials and failure cases. He received an MS degree in chemistry from Aligarh Muslim University, Uttar Pradesh, India, and a PhD degree in materials science and engineering from the Moscow State University.

Øystein Birketveit is Technical Manager for Production Technologies for M-I SWACO, a Schlumberger company, in Bergen, Norway. For the past 18 years, he has specialized in the field of corrosion. Prior to joining M-I SWACO, Øystein worked for Statoil and for Det Norske Veritas. He earned his MSc degree in materials and electrochemistry from the Norwegian University of Science and Technology, Trondheim.

Joshua E. Jackson is the CEO of G2MT LLC as well as the cofounder of G2MT Laboratories, LLC in Houston. G2MT Labs is a metallurgical consulting and analysis company that performs nondestructive materials char-acterization to evaluate residual stress mechanical properties and other critical parameters including the effects of corrosion. His scientific focus areas include corrosion analysis, high-temperature materials, hydro-gen absorption effects, failure analysis and statistics. Joshua is the coauthor of numerous papers in the field of materials science covering subjects including non-destructive testing, metallurgy, welding, corrosion and hydrogen. He obtained BS degrees in both mathemat-ics and physics from the Massachusetts Institute of Technology, Cambridge, USA, and MS and PhD degrees in metallurgical and materials engineering from the Colorado School of Mines, Golden, USA.

Alyn Jenkins, based in Aberdeen, serves as the Global Asset Integrity Manager for Schlumberger Production Technologies. He manages asset integrity product lines that include corrosion inhibitors, biocides, H2S scavengers and oxygen scavengers and is responsible for research and development projects related to corrosion. He began his career in 1998 with Clariant Oil Services in Aberdeen and then worked for Baker Hughes in Liverpool, England. Alyn joined M-I SWACO in 2005 as a corrosion specialist in Stavanger and then served as lead integrity management specialist. Alyn holds BS and MS degrees, both in chemistry, from the University of Wales, Bangor.

Bruce MacKay is Client Support Manager for the Schlumberger North America fracturing and cement-ing operations in Sugar Land, Texas. He has worked as a chemical problem solver in various capacities for Schlumberger for 10 years, spanning the R&D spectrum from research to product development to technology implementation. He has authored 12 peer-reviewed scientific journal articles and five SPE papers and has been granted several patents on chemical technologies related to oilfield applications. He has been a speaker on the importance of chemistry in oil-field development to a variety of audiences, including the US National Academy of Sciences, the American Chemical Society and the National Aeronautics and Space Administration Jet Propulsion Laboratory. Bruce was a Natural Sciences and Engineering Research Council of Canada postdoctoral research scholar at the California Institute of Technology, Pasadena, USA. He earned a BS degree in chemistry and a PhD degree in inorganic chemistry from the University of British Columbia, Vancouver, Canada.

Denis Mélot is a Nonmetallic Materials Expert with the technology department of Total Upstream, in Paris. Using his foundation of studies in polymer science, his focus is on nonmetallic materials and corrosion. Prior to beginning work with Total in 2003, he was a researcher in the R&D department of Elf Atochem, which is now Arkema, in Serquigny, France. He then spent six years as the technical manager for pipe coating products with the company. Denis chaired the ISO 12736 working group on wet thermal insula-tion systems, was a member of pipeline coatings work group ISO 21809 and holds certifications from the Association pour la Certification et la Qualification en Peinture Anticorrosion and Faglig Råd for Opplæring og Sertifisering av Inspektører innen Overflatebehandling . He holds numerous patents in his field and has coauthored several papers on the subject of coatings and corrosion. Denis has a degree in materials science from the École Universitaire D’Ingénieurs de Lille, France, and received his PhD degree in polymer science from Université de Lille.

Jan Scheie is a Project Leader and Account Manager for Production Technologies (PT) in Schlumberger Norge A/S in Stavanger, where he serves customers in Scandinavia. He has also been an account manager for production technologies, an international sales manager and an area manager for production chemi-cals in Stavanger. He has worked for M-I SWACO in market development for the eastern hemisphere, as technical manager in the Middle East and CIS, as sales manager in South Asia and as principal engineer for developing sales strategy in mainland Europe. He is a member of TEKNA, the Norwegian Society of Graduate Technical and Scientific Professionals, the SPE and the National Association of Corrosion Engineers. He received an MSc degree in chemical engineering from the Norwegian University of Science and Technology, Trondheim, Norway, and an MBA degree from Thunderbird School of Global Management, Glendale, Arizona, USA.

Marko Stipanicev is Corrosion Discipline Lead for Schlumberger Production Technologies in Bergen, Norway. Upon graduation from the University of Zagreb, Croatia, he worked as an external consultant on industry related projects at the Faculty of Chemical Engineering and Technology, in Croatia. Beginning in 2010, he worked as a research engineer for Det Norske Veritas in Bergen, investigating corrosion-based failures and performing root cause analysis studies. He joined M-I SWACO in 2013 as a corrosion specialist, working in Bergen, and in 2015, he was named the corrosion discipline lead. Marko is responsible for Schlumberger corrosion products, which include inhibitors, biocides, scavengers and nutrients. He has authored and coau-thored numerous papers and publications related to corrosion and corrosion management. He holds an MSc degree in chemical engineering and technology from the University of Zagreb and a PhD degree in environ-mental process and biocorrosion management from the Université de Toulouse, France.

Jean Vittonato, is Head of the Total E&P Technology Division corrosion department in Pau, France. He is responsible for the headquarters’ corrosion team and provides technical assistance to projects and operating subsidiaries worldwide. He started work in 1999 focusing on cathodic protection with COREXCO, an engineering cathodic protection company, where he was in charge of designing cathodic protection systems for both onshore and offshore and for installa-tion, monitoring and maintenance follow-up. In 2006, he joined Total as a corrosion specialist and was in charge of cathodic protection activities. He provided support for projects for both Total E&P and operating subsidiaries and was in charge of research projects related to cathodic protection. He spent three years in Republic of the Congo as the head of the Total cor-rosion department, where he supervised all projects related to corrosion. Jean is a certified Cathodic Protection Specialist with the National Association of Corrosion Engineers and with the Centre Français de la Protection Cathodique and is chair of the ISO TC 67 SC2 GW11 working group on cathodic protection of pipelines. He obtained an engineer-ing degree from Institut National Polytechnique de Grenoble, France.

Contributors

50 Oilfield Review

Slide Drilling—Farther and Faster

For decades, the mud motor and bent housing assembly have played a critical role in

directional drilling; however, the technique used to drill a lateral section resulted in

slow drilling rates. A surface-mounted torque control system is helping drillers reach

farther while improving rates of penetration and toolface control.

Steven DuplantisHouston, Texas, USA

Oilfield Review 28, no. 2 (May 2016).Copyright © 2016 Schlumberger.For help in preparation of this article, thanks to Edina Halilagic and Brandon Mills, Houston.Slider is a mark of Schlumberger.

Directional wells have been a boon to oil and gas production, particularly in unconventional plays, where horizontal and extended-reach wells maximize wellbore exposure through pro-ductive zones. In many of these wells, steerable mud motors have been crucial to achieving the well trajectory necessary to hit operators’ target zones. Directional drillers use a downhole mud motor when they kick off the well, build angle, drill tangent sections and maintain trajectory.

A bend in the motor bearing housing is key to steering the bit toward its target. The surface-adjustable bend can be set between 0° and 3°.

This slight bend is sufficient for pointing the bit in a given direction yet is small enough to permit rotation of the entire mud motor assembly during rotary drilling. This seemingly minor deflection determines the rate at which the motor builds angle to establish a new wellbore trajectory. By orienting that bend in a specific direction, called its toolface angle, the driller can change the inclination and azimuth of the well path.

To maintain the orientation of that bend and thus change wellbore trajectory, the drillstring must not be allowed to rotate, and this is where the mud motor comes into play. A mud motor is

1. Maidla E and Haci M: “Understanding Torque: The Key to Slide-Drilling Directional Wells,” paper IADC/SPE 87162, presented at the IADC/SPE Drilling Conference, Dallas, March 2–4, 2004.

Figure 1. Typical mud motor. The bent housing of the mud motor (left ) is the key to building wellbore deviation and controlling wellbore trajectory while the rotor turns the bit. The bend in the housing is dialed in at the drill floor when the drilling crew makes up the bottomhole assembly; here, the bend has been set at 2.89 degrees (middle). By selecting a larger bend, the driller is able to obtain curve having a smaller radius. The motor, installed immediately above the bit, consists of an eccentric rotor within an elastomer stator (right ). As drilling mud flows through the stator, it displaces the helical rotor shaft, causing the shaft to rotate within the stator’s protective housing, which turns the bit.

Wear pad

Rotation

Mud flow

Rotor

Effectivebend

3.00

2.89

2.602.77

2.383.00

2.77

Stator

2.89

Protectivehousing

May 2016 5151

a type of positive displacement motor powered by drilling fluid. An eccentric helical rotor and stator assembly drive the mud motor (Figure 1). As it is pumped downhole, drilling fluid flows through the stator and turns the rotor. The mud motor converts hydraulic power to mechanical power to turn a drive shaft that causes the bit to rotate.

Using mud motors, directional drillers alter-nate between rotating and sliding modes of drill-ing. In rotating mode, the drilling rig’s rotary table or topdrive rotates the entire drillstring to transmit power to the bit. This rotation enables the bend in the motor bearing housing to point equally in all directions and thus maintain a straight drilling path (Figure 2). In most opera-tions today, measurement-while-drilling (MWD) tools provide real-time inclination and azimuth measurements that alert the driller to any devia-

tions from the intended course. To correct for those deviations or to alter the wellbore trajec-tory, the driller switches from rotating to sliding mode. In sliding mode, the drillstring does not rotate; instead, the downhole motor turns the bit and the hole is drilled in the direction the bit is pointing, which is controlled by toolface orienta-tion. Upon correcting course and reestablishing the wellbore trajectory needed to hit the target, the driller may then switch back to rotating mode. (Figure 3).

Of the two modes, slide drilling is less effi-cient; lateral reach usually comes at the expense of penetration rate. The rate of penetration (ROP) achieved using conventional sliding methods typically averages 10% to 25% of that attained in rotating mode.1 Conversely, by rotating the entire drillstring, drillers gain a substantial advantage in ROP. This article describes an automated sys-tem that helps drillers achieve significant gains in horizontal reach with noticeably faster rates of penetration. Field experience in Colorado, USA, illustrates how a torque-oscillation system can help operators exploit unconventional plays.

Slide Drilling Challenges To initiate a slide, the driller must first orient the bit to drill in alignment with the trajectory pro-posed in the well plan. This requires the driller to stop drilling, pull the bit off-bottom and recipro-cate the drillpipe to release any torque that has built up within the drillstring. The driller then orients the downhole mud motor using real-time MWD toolface measurements to ensure the speci-fied wellbore deviation is obtained. Following this time-consuming orientation process, the driller sets the topdrive brake to prevent further rotation from the surface. The slide begins as the driller eases off the drawworks brake to control the hook load, which, in turn, affects the mag-nitude of weight imposed at the bit. Minor right and left torque adjustments are applied manually to steer the bit as needed to keep the trajectory on course.

As the depth or lateral reach increases, the drillstring is subjected to greater friction and drag. These forces, in turn, affect the driller’s ability to transfer weight to the bit and control toolface ori-entation while sliding, making it difficult to attain

Figure 2. Drilling using a bent housing. In rotating mode, the bit carves a straight path parallel to the axis of the drillstring, which is also rotating. Because the bent housing forces the bit to tilt outward by a few degrees, the bit drills a hole that is slightly larger than the diameter of the bit. When the driller switches to sliding mode, only the bit rotates. The resulting hole is in gauge and follows the axis of the BHA below the bent housing.

Bent housing

Rotarymode

Increased diametercaused by outward

tilt of bit

Slidingmode

Kickoff

Build section

Gauge hole Figure 3. Directional drilling trajectory. After the well is drilled vertically to the specified kickoff point, the mud motor is used to build angle while slide drilling. When the target angle is achieved, a straight-line tangent section may be drilled in rotary mode. While the BHA maintains inclination and azimuth, the driller resorts to sliding mode only when the drilling direction deviates from the planned trajectory. In some fields, a stratigraphic marker bed above the reservoir can be detected by LWD tools, prompting the driller to initiate a second slide section to land the well horizontally within the reservoir.

Reservoir

Kickoff 1

Kickoff 2

Tangent section

Build section, 3°/100 ft

Build section, 3°/300 ft Lateral section

Marker bed

52 Oilfield Review

sufficient ROP and maintain trajectory to the tar-get. Such problems frequently result in increased drilling time, which may adversely impact project economics and ultimately limit the length of a lat-eral section.

The capability to transfer weight to the bit affects several aspects of directional drilling. The driller transfers weight to the bit by easing, or slacking off, the brake; this transfers some of the hook load, or drillstring weight, to the bit.2 The difference between the weight imposed at the bit and the amount of weight made available by eas-ing the brake at the surface is primarily caused by drag. As the horizontal departure of a wellbore increases, so does the longitudinal drag of the drillpipe along the wellbore.

Controlling weight at the bit throughout the sliding mode is made even more difficult by drillstring elasticity, which permits the pipe to move nonproportionally. This elasticity can cause one segment of drillstring to move while other segments remain stationary or move at different velocities.3

Poor hole cleaning may also affect weight transfer. In sliding mode, hole cleaning is less effi-cient because there is no pipe rotation to facilitate turbulent flow; this condition reduces the drilling fluid’s ability to carry solids. Instead, the solids

accumulate on the low side of the hole in cut-tings beds that increase friction on the drillpipe, making it difficult to maintain constant weight on bit (WOB).

Differences in frictional forces between the drillpipe inside of casing versus that in open hole can cause weight to be released suddenly, as can hang-ups caused by key seats and ledges. A sud-den transfer of weight to the bit that exceeds the downhole motor’s capacity may cause bit rotation to abruptly halt and the motor to stall. Frequent stalling can damage the stator compo-nent of the motor, depending on the amount of the weight transferred. The driller must operate the motor within a narrow load range to maintain an acceptable ROP without stalling.4

At the driller’s console, an impending stall might be indicated by an increase in WOB but with no corresponding upsurge in downhole pres-sure to signal that an increase in downhole WOB has actually occurred. At some point, the WOB indicator will show an abrupt decrease, indicat-ing a sudden transfer of force from the drillstring to the bit.5

Increases in drag impede a driller’s ability to remove torque downhole, making it more dif-ficult to set and maintain toolface orientation.6 Toolface orientation is affected by torque and

WOB. When weight is applied to the bit, torque at the bit increases. Torque is transmitted downhole through the drillstring, which turns to the right, in a clockwise direction. As weight is applied to the bit, reactive torque, acting in the opposite direction, also develops. This left-hand torque is transferred upward from the bit to the lower part of the drillstring. Reactive torque builds as weight is increased, reaching its maximum value when the motor stalls. This reactive torque also affects the orientation of the motor.

Reactive torque must be taken into account as the driller tries to orient the drilling motor from the surface. In practice, the driller can make minor shifts in toolface orientation by changing downhole WOB, which alters the reac-tive torque. To produce larger changes, the driller can lift the bit off-bottom and reorient the tool-face. Even after the specified toolface orienta-tion is achieved, maintaining that orientation can be challenging. Longitudinal drag increases with lateral reach, and weight transfer to the bit becomes more erratic along the length of the horizontal section, thus allowing reactive torque to build and consequently change the toolface angle.7 The effort and time spent on orienting the toolface can adversely impact productive time on the rig.

Figure 4. Torque versus friction. Longitudinal drag along the drillstring can be reduced from the surface down to a maximum rocking depth, at which friction and imposed torque are in balance. By manipulating the surface torque oscillations, this point can be moved deep enough to produce a significant reduction in drag. Similarly, reactive torque from the bit creates vibrations that propagate back uphole, breaking friction

and longitudinal drag across the bottom section of the drillstring up to a point of interference, where the torque is balanced by static friction. An intermediate zone remains unaffected by surface rocking torque or by reactive torque. By continuously monitoring torque, WOB and ROP while sliding, the Slider system helps minimize the length of this intermediate zone and thus reduces longitudinal drag.

Topdrive

Rotatingonly

Sliding plus topdrive torque

Dynamic friction of rotation

Maximumsurface-applied

torque

Maximum rocking depth Point of interference Bit

Minimum surface-applied torque Minimum reactive torque Maximum reactive torque

Mud motor

Static friction Dynamic friction

Sliding only Sliding plus reactive torque

May 2016 53

2. The hook load includes more than simply the weight of the drillstring in air; it is the total force pulling down on the hook as it hangs beneath the derrick traveling block. This total force includes the weight of the drillstring, drill collars and ancillary equipment reduced by any forces that tend to decrease the weight. These forces might include friction along the wellbore wall and buoyant forces on the drillstring caused by its immersion in drilling fluid.

The Slider SystemManually correcting and maintaining toolface orientation can be a difficult process. Drilling efficiency is largely dependent on the driller’s ability to:• transfer weight to the bit without stalling the

mud motor• reduce longitudinal drag sufficiently to achieve

and maintain a desired toolface angle• attain acceptable ROP.

The Slider automated surface rotation control system was developed to help operators regain some of the drilling performance of a conven-tionally rotating drillstring. The Slider interface interacts with the topdrive control system to rotate the drillstring back and forth. This torque rocking technique reduces longitudinal drag along part of the drillstring while slide drilling. Rocking back and forth subjects the upper drill-string to near-constant tangential motion, pro-ducing a dynamic friction coefficient, which is lower than a static friction coefficient created by

nonrotating pipe. Rocking can also help reduce axial friction along the drillstring. However, this motion is not necessarily transmitted all the way to the bit—other processes are at work.

Torque from the topdrive rotates the drill-string from the surface down to a maximum rocking depth, where friction against the side of the hole prevents the pipe from turning. At the same time, as the mud motor turns the bit, it gen-erates a reactive torque in the opposite direction. This torque is transmitted a short distance up the drillstring until it is overcome by friction at some point between the bottom of the wellbore and the BHA, referred to as the point of interfer-ence (Figure 4). Throughout the interval between the bit and the point of interference, the velocity component of reactive torque imposes a reduction in longitudinal drag along the lower part of the drillstring and possibly a change in toolface orien-tation. Between the depth where surface torque is overcome by friction and the point where reac-tive torque is overcome by friction, the pipe does

not rotate. This section of drillstring, which has no tangential motion, moves by sliding only and is subject to static friction, which is greater than the dynamic friction of pipe in motion.

The location of the point of interference varies with changes in the amount of reactive torque. To efficiently minimize the sliding inter-val between the depth of rocking and the point of interference while keeping the maximum rocking depth relatively constant, an automated control system must be used.

The amount of surface torque supplied by the topdrive dictates in large part how far downhole the rocking motion will be transmitted. This rela-tionship between torque and rocking depth can be modeled using conventional torque and drag programs (Figure 5). However, these programs are not needed when using the Slider system. Using inputs from surface hook load and stand-pipe pressure as well as downhole MWD toolface angle, the Slider system automatically deter-mines the amount of surface torque needed to

Figure 5. Torque versus depth plot. Surface-applied torque will tend to twist the drillstring to a certain depth depending on the drag encountered over the length of the pipe and on pipe thickness and weight. In this model, 2,000 ft-lbf of torque applied at the surface will cause the pipe to twist to a depth of 6,400 ft. (Adapted from Maidla et al, reference 5.)

0

2,000

4,000

6,000

8,000

10,000

12,000

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000

Dept

h of

twis

t, ft

Surface applied torque, ft-lbf

No friction reduction

Reactive torque

3. Maidla and Haci, reference 1.4. Tello Kragjcek RH, Al-Dossary A, Kotb W and Al Gamal A:

“Automated Technology Improved the Efficiency of Directional Drilling in Extended Reach Wells in Saudi Arabia,” paper SPE 149108, presented at the SPE/DGS Saudi Arabia Section Technical Symposium and Exhibition, Al-Khobar, Saudi Arabia, May 15–18, 2011.

5. Maidla E, Haci M, Jones S, Cluchey M, Alexander M and Warren T: “Field Proof of the New Sliding Technology for Directional Drilling,” paper SPE/IADC 92558, presented at the SPE/IADC Drilling Conference, Amsterdam, February 23–25, 2005.

6. Maidla and Haci, reference 1.7. Maidla and Haci, reference 1.

54 Oilfield Review

transfer weight downhole to the bit, thus elimi-nating the need to come off-bottom to make toolface corrections (Figure 6). This results in an efficient drilling operation and reduced wear on downhole equipment.

System HardwareSlider system hardware consists of a compact package that houses the circuitry and sensors needed to interact with the rig’s topdrive con-trol system. An interface plug is installed on the control panel for the topdrive, and the system is mounted at the driller’s console. Installation typically takes less than two hours with no interruption to the drilling process. The Slider system’s connections require no alterations to the drilling contractor’s topdrive mechanism or modifications to the drilling rig. The system is entirely surface mounted and has no downhole equipment that might become lost in the hole. To ensure operational safety, the system is designed to allow manual intervention at any time.

The directional driller’s interface consists of a ruggedized notebook computer with a display configured to enable the driller to command the Slider system while monitoring surface and downhole parameters (Figure 7). The Slider system takes input such as MWD toolface angle, surface torque and standpipe pressure from mea-surements already available at the rig. The MWD toolface measurement is used to determine the amount of correction needed to restore the tool-face to the angle needed to drill the prescribed trajectory. Surface standpipe pressure provides an indicator of reactive torque. The Slider software processes these inputs to determine whether additional torque should be applied to the drillstring to maintain the toolface angle and ROP.

To begin slide drilling, the driller can acti-vate the Slider system and initiate the automatic rocking action, which alternately applies torque to the right and the left. The transfer of weight is controlled by varying surface torque to compen-sate for changes in reactive torque. Corrections in toolface angle are achieved through addi-tional torque pulses during the rocking cycles.8 For every torque cycle to the left or right, a cor-responding differential pressure peak occurs, indicating that the weight is being transferred to the bit. To adjust the toolface orientation, the driller can control the magnitude and frequency of torque pulses during a rocking cycle.

Figure 6. Comparison of rotating and sliding drilling parameters. Rate of penetration (ROP) and toolface control depend largely on the driller’s ability to transfer weight to the bit and counter the effects of torque and drag between rotating and sliding modes. The best ROP is achieved while rotating (top); however, toolface varies drastically, as there is no attempt to control it (Track 3). Hook load (Track 2) and weight on bit (WOB) remain fairly constant while differential pressure (Track 1) shows a slight increase as depth increases. To begin manual sliding (middle), the driller pulls off-bottom to release trapped torque; during this time, WOB (Track 1) decreases while hook load (Track 2) increases. As drilling proceeds, inconsistencies in differential pressure—the difference between pressures when the bit is on-bottom versus off-bottom—indicate poor transfer of weight to the bit (Track 1). Spikes of rotary torque indicate the directional driller’s efforts to orient and maintain toolface orientation (Track 2). Toolface control is poor because of trouble transferring weight to bit, which is also reflected by poor ROP (Track 3). Using the automated Slider system (bottom), the directional driller quickly gained toolface orientation. When the WOB increased, differential pressure was consistent, demonstrating good weight transfer (Track 1). Weight on bit during a Slider operation is lower than during a manual sliding operation. Left-right oscillation of the drillpipe is constant through the slide (Track 2). Average ROP is substantially higher than that attained during the manual slide, and toolface orientation is more consistent (Track 3).

Differential Pressurepsi 8000

Weight on Bit1,000 lbf 1000

Hook Load1,000 lbm 3000

Rotary Speedrpm 1500

Rotary Torquelbf 20,0000

Rate of Penetrationft/h 500

Toolfacedegree 3600

Rotating Mode

Manual Mode

Slider Mode

Rotating Mode

May 2016 55

Field ExperienceSlider technology has been instrumental in developing unconventional plays throughout North America. Wattenberg field, one of the more prolific fields in the Denver-Julesburg basin, is located in Weld County, Colorado. There, a lead-ing operator in the area used the Slider system to drill horizontal wells in the Cretaceous Niobrara gas play (Figure 8).

One of those wells, spudded in February 2016, was drilled vertically to its kickoff point, then drilled in a westerly direction to its land-

Figure 7. The Slider system graphical display. Downhole performance parameters are monitored and controlled via a notebook computer interface between the topdrive and Slider system. The directional driller can configure this display to show various key parameters such as toolface (dial, center) and torque and differential pressure (chart, bottom right ). Torque curves show higher values for righ-hand torque (yellow) than for

left-hand torque (orange). Up-and-down keys allow the driller to set values for left and right torque (upper left ). Brief torque increases above set values can be added for one oscillation cycle by bumping left or right (middle left ). The driller can immediately override the system by hitting the disable button (upper right ).

Figure 8. Wattenberg field. The prolific Wattenberg field lies in north-central Colorado, USA, within the Denver-Julesburg basin.

USA

Colorado

Nebraska

Kansas

Wyoming

Denver-JulesburgbasinWattenberg

field

8. “Slider: New Level of Efficiency to Directional Drilling,” Drilling Contractor 11, no. 4 (July–August 2004): 28–31.

56 Oilfield Review

ing point in the Niobrara pay zone. Beginning at X,X25 ft, the directional driller manually con-trolled the slide drilling while averaging 38 ft/h [11.6 m/h] ROP. Upon engaging the Slider auto-mated system, the driller reported an average ROP of 51 ft/h [15.5 m/h], for an improvement of 34% in ROP compared with that of manually controlled sliding (Figure 9). The Slider system was engaged several times during the course of drilling this well. Each time the trajectory began to drift beyond specified tolerances, the direc-tional driller switched from rotating to sliding modes to bring the wellbore back on course. Comparisons between manually controlled slid-ing and automated sliding with the Slider system consistently showed significant gains in ROP over the manual approach.

Faster and FartherBy sensing the amount of surface torque required to transfer weight to the bit and by eliminating the need to pull off-bottom to make toolface cor-rections, the Slider automated surface rotation control system enables substantial increases in ROP and lateral reach for directional wells. Rocking, or oscillating, the drillstring back and forth helps the driller overcome friction and thus reduce drag on the drillstring. Along with reduc-ing drag, operators can decrease the amount of mud additives normally used for lubrication. The Slider automated system typically applies less WOB to maintain toolface control and has mark-edly fewer motor stalls than are experienced while manually slide drilling. By achieving con-sistent toolface control, this automated torque

Steven Duplantis is an Operations Manager for Schlumberger in Houston, where he oversees day-to-day operations, sales and development for the Slider auto-mated surface rotation control system. Steven began his oilfield career in 1994 and has held positions with MD Totco, Noble Engineering and Epoch/Optidrill. Since 2006, he has been focused on the Slider system, working as a field coordinator and engineer involved with Slider system testing and field trials along with implementation and ongoing development of advanced features for the Slider product line.

Figure 9. ROP improvement. A drilling parameter display indicates that from X,X25 ft to X,X35 ft, the directional driller was manually controlling the slide drilling operations. After taking 16 min to drill that interval (averaging 38 ft/h ROP), the driller activated the Slider automated system. Drilling from X,X35 to X,X47 ft in 14 min (averaging 51 ft/h ROP), the driller achieved a 34% improvement in ROP.

Weight on Bit1,000 lbm

Time,h:min 1000

Hook Load1,000 lbm 3000

Rate of Penetrationft/h 3000

Rotary Speedrpm 800

Rotary Torquelbf 34.390

Pump 1strokes/min 2000

Toolfacedegree 3600

Differential Pressurepsi 1,0000

Standpipe Pressurepsi 12,0000

07:12

Depth,ft

X,X25

X,X35

X,X47

07:14

07:16

07:18

07:20

07:22

07:24

07:26

07:28

07:30

07:32

07:34

07:36

07:38

07:40

07:42

07:44

07:46

07:48

07:50

rocking system facilitates a longer horizontal sec-tion, which has less tortuosity, ultimately leading to increased production. —MV

Contributor

LOOKING BACK

May 2016 57

Though difficult to fathom, large companies that have tens or hundreds of thousands of employees were once startups employing a dozen or fewer people. Some statistics suggest that, today, nine out of ten startups fail within the first three years. Investigating the critical elements that help a young organization survive those early years—in this case, Schlumberger—is compel-ling. Studying a company’s beginnings can be essential for understanding the corporate image and characteristics that eventually emerge.

Many successful global companies that have familiar names started small. The inventor of xerography, Chester Carlson, conducted early experiments in his apartment kitchen, where occasional explosions or, perhaps, malodorous results occurred because his methods involved hydrogen sulfide. The Paul Allen and Bill Gates partnership, which led to the creation of Traf-O-Data and then of Microsoft, is, perhaps the ulti-mate example of humble beginnings to mega- corporation stories.

The Schlumberger beginnings, forged in the crucible of a global search for hydrocarbons, may be considered, in many ways, more adventurous and exciting.

Between the early experiments by Conrad Schlumberger—the electri-cal resistivity mapping of the lawn of the family property at Val-Richer in Normandy, France, in the summer of 1912—and the official registration of Société de Prospection Électrique—or SPE, and soon nicknamed La Pros—in July 1926, a long gestation period occurred.

The primary reason for the delay in the company’s full formation was World War I. Engineers, scientists and inventors in Europe did not have the freedom to choose not to spend time in the army of fighting in the war. The Schlumberger brothers were no exception. Conrad Schlumberger joined the French artillery as an officer and his younger brother Marcel joined the cavalry. Consequently, the Schlumberger startup history was put on hold in 1914.

Conrad was intimately involved with and horrified by the war. He began to question his future and considered giving up his work to become a pacifist. While in the trenches at Verdun, he said of the family textile business: “I’m ashamed of belonging to a family of textile manufacturers whose employees got up at dawn, walked for miles to reach the factory, and worked twelve or fourteen hours a day. Not one of those workers had a life fit to live.”1

But Paul Schlumberger, Conrad and Marcel’s visionary father, who had another four children—Jean, Daniel, Maurice and Pauline—with his wife Marguerite de Witt, redirected Conrad from factory life and tied him to his brother’s future.

Paul wrote, “The scientific interest in research must take precedence over financial interests. Marcel will bring to Conrad his remarkable competence as an engineer and his common sense. Conrad will be the wise physicist. I will support them.” He offered a sum of 500,000 francs for the endeavor.2

In 1919, 500,000 francs represented 160 kg [5,100 troy oz] of gold. Unfortu-nately, the franc value was cut by a factor of five from 1914 to 1926, and La Pros, which had engineers on four continents, soon had to learn how to manage currency variations. The business invested in trucks and equipment and sent people to other countries, some of which were not hospitable. These early processes meant the company had essential cash flow needs.

The survival of the young company was based on three components: seed money, innovation and sound management of intellectual property and people. Success came about because the company built an invincible team and instilled exceptional values.

Birth of La Pros: The 90th Anniversary

Oilfield Review 28, no. 2 (May 2016).

Copyright © 2016 Schlumberger.

Microsoft is a registered mark of the Microsoft Corporation.

Xerox is a trademark of the Xerox Corporation.

Philippe TheysHouston, Texas, USA

1. Schlumberger AG: The Schlumberger Adventure: Two Brothers Who Pioneered in Petroleum Technology. New York City: Arco Publishing, Inc., 1982.

2. Schlumberger, reference 1.

Oilfield Review58

LOOKING BACK (continued)

Figure 1. Patents filed in France, Brazil, Mexico and Australia. The first patent was filed in France on September 27, 1912 (top left); the illustration (top right) is from the first patent. Conrad Schlumberger filed patents in, among other countries, Mexico and Brazil (middle) although SPE did not

work in these countries until 1936 and 1945, respectively. The patent at the bottom was filed in Australia. (Documents courtesy of the Schlumberger museum in Crevecoeur, France, and Collections École Polytechnique, Palaiseau, France.)

May 2016 59

Philippe Theys was hired by Société de Prospection Électrique in 1972, 46 years after its creation. Much like the first Schlumberger employees who worked in many parts of the world, he worked in the oil field in France, Sweden, Germany, Austria, Western and Eastern Australia and Taiwan as well as in Louisiana and Texas in the US before a 26-year career in marketing, research and engineering and quality. He graduated from École Centrale Paris in 1971, 64 years after Marcel Schlumberger, alumnus of the same école d’ingénieurs. Philippe retired from Schlumberger in 2004 and now consults for oil and gas companies.

Figure 2. Société de Prospection Électrique founders and early employees. From left to right, Paul, Conrad and Marcel Schlumberger; Henri-Georges Doll; and Eugene Leonardon, the first engineer hired by the brothers. Other early employees include Jacques Gallois and Roger Jost. (Photographs courtesy of the Schlumberger museum in Crevecoeur, France, and Collections École Polytechnique, Palaiseau, France.)

Innovation and Intellectual Property Innovation is the backbone of startups, but new techniques and technolo-gies often fall in the public domain and quickly become available to com-petitors and to clients who may take advantage of this availability without rewarding the original inventor. Chester Carlson, for example, working in the 1930s and 1940s, wisely registered every development of his research for the company that would eventually become Xerox Corporation.

Conrad Schlumberger filed his first patent, Procédé pour la détermina-tion du sous-sol au moyen de l’électricité (Subsurface characterization by electrical methods) on September 27, 1912, in France. He then took care to file patents in what seemed like unusual locations, including Mexico, Czechoslovakia, Brazil and the Belgian Congo. By 1926, he had filed patents in 18 countries, which would become the core of the intellectual property of SPE (Figure 1).

The Team Reinforced by their father’s monetary donation, Conrad and Marcel consti-tuted an unbreakable partnership. “Question the one without the other and the answer would be: ‘I’ll talk about it with my brother Marcel… .’ [and] ‘I’ll ask my brother Conrad.’ ” 3

Eugène Léonardon was the first full-time engineer hired by the brothers. He began working with Conrad in 1913 but then served in the French artil-lery during World War I; he was rehired in 1919. Additional field hands were cautiously recruited. The ads in Le Journal des Mines called for ingénieurs sportifs, athletic engineers. After Léonardon came a series of intrepid engi-neers (Figure 2). Jules Carré had worked with Conrad as early as 1913. Gilbert Deschâtre and Raymond Sauvage participated in the first survey for Shell in 1928. Jacques Gallois and Pierre Baron constituted the first pros-pecting crew in Freeport, Texas, USA. Felicien Mailly joined in 1925 to man-age the electrical research division.

Not all of these new recruits were graduate engineers. Roger Jost was the nephew of the régisseur, or stage manager, for events at Val-Richer. The brothers also looked for help within the family. Nephew Marc Schlumberger was sent to the US to scout for business. Henri-Georges Doll, who was Pauline’s nephew, married Conrad’s daughter Annette and joined the com-pany in January 1926 as a full-time employee. Not all of the new hires were French. Sherwin F. Kelly, an American who graduated in France, started in 1921. Swiss geologist Edouard Poldini joined in 1922. By 1926, the company had a staff of 17 field engineers working in Romania, Serbia, Canada, South Africa and the Belgian Congo. Because of the distances and lack of com-munication, these people were given high levels of independence, initiative,

responsibility and trust that are difficult to imagine in our era of instant texting and emailing.

All the right components were finally available; Société de Prospection Électrique was thus incorporated in Paris on July 1, 1926, with 200,000 francs in capital divided in 2,000 shares. SPE was now the organization overseeing all operations, personnel and premises that were previously held loosely by the brothers. Conrad was president. Maurice and Albert Doll, who was Pauline’s husband and Henri-George’s uncle, were scrutateurs, or trea-surers. Jean, cofounder of La Nouvelle Revue Française or NRF, and Maurice—cofounder of the bank Neuflize Schlumberger Mallet—were named minority shareholders. The headquarters, at 30 rue Fabert in Paris, consisted of a small, five-room apartment close to Place des Invalides.

In September 1926, Conrad left France for the US to further develop that market. In October of that year, Paul Schlumberger passed away on one of those drizzling autumn days well known to those living in Normandy. Conrad, Marcel and Henri-Georges—who contributed to the success of the mission Paul had dreamed of—joined the family patriarch in the cemetery of the vil-lage of Saint-Ouen-le-Pin, France, in 1936, 1953 and 1991, respectively.

Of the many Schlumberger companies, SPE is the oldest. In the US, the Schlumberger Electrical Prospecting Method, defunct as the result of the Great Depression, was replaced by Schlumberger Well Services Corporation in 1934 and by Schlumberger Technology Corporation in 1984. In 2016, SPE still exists and recruits young engineers to send to the far corners of the globe.

Schlumberger currently operates in more than 85 countries and has about 100,000 employees. Wherever the drill bit is turning in the search for hydrocarbons, the Schlumberger name can be found. The startup com-pany formed by a few industrious and adventurous engineers has become one of the most recognized names in the oil and gas industry. Could Paul, Conrad or Marcel have ever imagined the enterprise that today bears their name? Schlumberger stresses its guiding values of people, technology and profits. Although they differ from the original principles, they echo the “seed money, innovation and team” concepts that formed the company that would become Schlumberger; together, these ideas continue to guide the company into the future.

3. Schlumberger, reference 1.

Oilfield Review60

THE DEFINING SERIES

To replace dwindling reserves from onshore fields or from offshore wells drilled in shallow waters, many E&P companies are turning to their deepwa-ter prospects. Exploration or production in deep and ultradeep waters is carried out at water depths of 300 to 3,050 m [1,000 to 10,000 ft] or greater. These depths dictate that most wells be completed subsea, with wellheads, pressure-control equipment and production equipment placed at the seafloor.

From deepwater and ultradeepwater completions, produced fluids are sent to a processing facility by way of a subsea production system. A sub-sea production system consists of the subsea infrastructure used to pro-duce oil and gas from offshore reservoirs. It encompasses one or more subsea wells and the subsystems necessary to deliver hydrocarbons to a fixed, floating, subsea or onshore processing facility. These subsystems can be divided into subsea trees, production controls, manifolds, jumpers, flow-lines, risers, umbilicals and processing components. Injection of water or gas back into subsea wells is also a function of the subsea production sys-tem. Generally, oil, gas and water produced from the reservoir will flow from a wellbore to a subsea tree and through a jumper to a manifold and subsea flowline. Today, many operators route the flowline to a booster pump to energize the flow as it travels between the seafloor and a riser that carries it to the surface for processing.

Seabed EquipmentThe subsea wellhead, installed at the beginning of the drilling phase, pro-vides the structural foundation for the well. The wellhead is also where the subsea tree is mounted. In some configurations, the tree contains the pro-duction tubing hanger and accommodates hydraulic and electrical lines used for managing downhole safety valves, completion valves and pressure or temperature sensors. The function of the subsea tree is to control and manage pressure and flow over the life of the well and enable any necessary intervention. The tree is the primary mechanism for shutting in the well at the seabed and serves as the interface for well reentry operations. A subsea control module (SCM) attached to the tree contains the instrumentation, electronics and hydraulics connections needed for safe operation of the subsea tree valves, chokes and downhole valves.

Sections of pipe, known as jumpers, run between subsea structures to serve as links through which fluids are transmitted. These pipe segments range in length from a few meters to hundreds of meters. A jumper is often installed to carry production downstream from the tree. The produced fluid may be routed through a multiphase flowmeter to measure production rates and volumes.

Where multiple wells produce in a subsea development, flowline jumpers from individual wells send produced fluids to a subsea production manifold (Figure 1). By routing produced fluids from multiple wells to the production manifold, the operator can reduce the number of flowlines that must be accommodated at the next step in the production chain. Upon reaching the

Subsea Infrastructure

Oilfield Review 28, no. 2 (May 2016).

Copyright © 2016 Schlumberger.

Matt VarhaugSenior Editor

Typical subsea field layout. Subsea trees positioned atop four wells contain pressure-control valves and chemical injection ports. A jumper carries produced fluids from each tree to a subsea manifold, which commingles production from the wells before sending it through flowline jumpers to a subsea boosting pump station. The pump provides energy to send the produced fluids through two pipeline end terminations (PLETs) and then through flowlines and up the risers to the production deck of the floating production, storage, and offloading vessel (FPSO). An integrated umbilical (green) from the FPSO supplies electric and hydraulic power for subsea tree or manifold control functions along with chemicals to suppress the formation of scale and hydrates in the production stream. On the seabed, the umbilical termination assembly (UTA) routes the chemical and hydraulic fluids (white jumpers) to the manifold, which sends them to each tree. The UTA also sends electric power to a distribution system which routes power lines (black leads) to the manifold, boosting pump and trees.

UTA

PLET

Umbilical riserProduction risers

Hydraulic fluid jumper

Electric power leadsElectric power

distribution unit

Injection fluid jumper

Boosting pump station

Jumper

Subsea tree

Manifold

FPSO

May 2016 61

manifold, produced fluids from the various wells are commingled before they are directed to a flowline that leads to the production platform. Injection manifolds function similarly and are used to manage the distribution of injected water, gas and chemicals to one or more subsea wells.

Some subsea installations have a pipeline end manifold (PLEM), which connects a flowline with another subsea structure or joins a main pipeline with a branch pipeline. The PLEM can incorporate tie-in points with other components such as isolation valves, diverter valves and sensor arrays. Some PLEM designs incorporate facilities for launching pipeline pigs—devices used to clean or monitor the inside of a pipeline.

When a reservoir does not have sufficient energy to produce the fluids from one subsea component to the next, a subsea boosting pump may be installed. Boosting pumps function as a seafloor artificial lift system, increas-ing both flow rate and recovery by reducing backpressure on the reservoir.

Other recent advances in subsea processing are used to enhance field economics. Seafloor separation and reinjection of produced water can alle-viate constrained topside water handling capacity while supplementing res-ervoir energy through waterdrive. Subsea gas compression, including wet gas compression, can improve viability of certain marginal developments.

Flow to the SurfaceFlowlines tie one or more fields back to a production facility—a shore-based processing facility or fixed production platform in shallower waters—but in deeper waters, a semisubmersible, spar and floating production, storage, and offloading vessel (FPSO) is used. The flowlines do not necessar-ily trace a straight course from wellhead to platform but may bend to avoid obstacles such as existing subsea infrastructure or natural obstructions such as underwater seamounts or canyons. As it follows the topography of the seafloor, the flowline climbs gradually from the colder, deeper reaches of the field upward through relatively warmer, shallower waters before reaching the production facility.

Water depth affects temperature, which can adversely impact flow between the subsea tree and the production facility. Upon exiting the well-head, warm produced fluids may encounter deepwater temperatures approaching 2°C [36°F] at the seafloor. Heat transfer between the produced fluid in the pipeline and the surrounding seawater can cool the fluid to the

Figure 1. Subsea manifold. This manifold, hoisted in preparation for installation on the seafloor, will take produced fluids from several wells and route them to a flowline running to a production platform.

Figure 2. Cross section of an umbilical. Umbilicals supply electric power, hydraulic fluid, chemicals and fiber-optic communications to the subsea production system. Separate hoses, cables, injection lines and other conduits are bundled together and enclosed within an armored outer ring designed to withstand harsh subsea conditions.

Chemical injectionfluid line

Electric cable

Hydraulic fluid line Wire cable armor

Fiber-optic cable

Outer plastic sheath

High-voltage cable

point that gas hydrates start to form. The change in fluid temperature beyond the tree influences the operator’s thermal management strategy. At some fields, chemicals such as methanol [CH3OH] or monoethylene glycol [C2H6O2] are injected into the system to keep the wellstream flowing then recovered on the surface and reused. Some operators use electrically heated flowlines; others use foam-insulated pipe. Some operators bury the flowline beneath the seafloor for insulation, but flowlines at certain fields require no additional heat or insulation at all. The chemistry and rheology of the pro-duced fluids ultimately dictate which methodology is adopted.

Production flowlines run from the manifold to structures that are linked to risers that direct the flow to the production facility. Risers transport pro-duced fluids from the seafloor to the surface production facility. Like flow-lines, many risers are insulated against cold seawater temperatures. They offer a measure of flexibility to withstand subsurface water currents or movement of the floating facility.

Surface LifelineThe surface processing facility provides power, control, communication and chemical injection services back to the subsea production system. These services are transmitted through a subsea distribution system using umbil-icals. Multiple steel and thermoplastic conduits are often bundled together with hydraulic lines, chemical injection lines, power conductors and fiber-optic cables to form a single integrated umbilical (Figure 2). These flexible conduits require sophisticated materials and manufacturing techniques to withstand deep-ocean currents, pressures and temperatures. Power con-ductors provide electricity for subsea equipment and system sensors. Hydraulic lines are used to open and close subsea valves. Fiber-optic lines instantly relay sensor information and control commands between the sea-floor and the surface. Some umbilical lines pump chemicals into the produc-tion stream. Umbilicals directly or indirectly service nearly every component in the subsea production system and are critical to operating the field. The lines typically run from the surface processing facility to an umbilical ter-mination assembly (UTA) on the seafloor, from which services are distrib-uted throughout the field.

Upon reaching the surface, the produced fluids are separated and treated by the processing facility. From there, an export pipeline transports the product to a storage and offloading installation or to an onshore refinery for further processing and distribution.

Oilfield Review62

THE DEFINING SERIES

Oilfield Review 28, no. 2 (May 2016).

Copyright © 2016 Schlumberger.

Geophysics is the study of the physics of the Earth, the propagation of elas-tic waves within it as well as its electrical, gravitational and magnetic fields. Although the origins of geophysics can be traced to ancient times, it was not until the early 20th century that scientists began applying geophysical con-cepts and techniques to the search for hydrocarbons and minerals and to evaluate geothermal energy resources. Now, geophysics plays a critical role in the petroleum industry because geophysical data are used by exploration and development personnel to make predictions about the presence, nature and size of subsurface hydrocarbon accumulations.

Unifying CharacteristicsThe challenges usually encountered in geophysics are posed in the form of an inverse problem, such that a set of measurements and known physical laws will permit a geophysicist to determine the Earth’s structure and char-acteristics that are consistent with those measurements. In geophysics, the answers are almost always nonunique, meaning there is more than one pos-sible solution that satisfies the measurements. Geophysicists attempt to resolve this ambiguity by integrating complementary data acquired from dissimilar methods or by adding supplemental knowledge such as wellbore measurements to determine which solution is correct.

In addition to nonuniqueness, all geophysical methods exhibit a decrease in the resolving power with distance from the measuring equip-ment. This concept is analogous to the difficulty of distinguishing objects by sight at increasing distances. The characteristic is more pronounced for some measurement methods than for others, but the result is that the deeper the subterranean structures, the less precise are the images of such structures.

Seismic SurveyingIn the oil field, the dominant geophysical data acquisition method is the seismic survey, whose history dates from the early 1920s. Seismic surveying employs a source—typically an airgun or a vibrating truck—to generate vibrations, or seismic waves, that propagate into the Earth. The seismic waves are refracted and reflected by subterranean strata and structures (Figure 1). Some of the energy returns to the surface, where it is recorded by sensors such as hydrophones or geophones. The distances between source and sensor can exceed 15 km [9.3 mi].

Geophysicists process the survey data to form an image and to estimate the physical characteristics of the subsurface. This requires two steps: develop a 3D velocity dataset, or volume, to produce a smooth estimate of the spatially varying velocity with which the seismic waves propagate in the Earth—a process called tomography—then, with the help of this velocity dataset, locate the subsurface layers from which the seismic waves were reflected, a process called migration.

The resulting 3D representation of the Earth is called a structural image, or volume. The reflecting surfaces are interpreted as the interfaces between rock layers, some of which may have been folded, cracked, faulted or eroded over geologic time. It can be sliced vertically to obtain a cross section or horizontally to map the depths of the rock layers beneath the survey area. The operator can use these interpretations to help determine suitable drilling targets. Modern seismic surveys routinely produce detailed 3D images of these reflecting surfaces to depths of 10 km [6 mi].

Additional information about the characteristics of the rocks can be extracted from seismic data. For example, by studying the size, or ampli-tude, of the reflections and how the amplitude changes with the angle at which the seismic waves hit the reflectors, geophysicists may be able to determine whether the pores within the rocks contain gas, oil or water. This step, known as amplitude versus offset (AVO), often has a higher level of uncertainty than does structural imaging.

Although most seismic work uses active sources designed to create seis-mic waves, the detection of weak seismic waves generated during hydraulic fracturing is of increasing interest. These faint signals are used to deter-mine the locations of microseismic events, which can indicate the position and extent of the hydraulic fractures.

GeophysicsRichard CoatesResearch Manager and Scientific Advisor

Figure 1. Marine seismic acquisition. An airgun array (top) produces pulses of seismic energy (green) that penetrate the subsurface and are reflected back (red) from the seabed and interfaces between rock layers. These reflections are detected by hydrophone arrays towed behind the seismic vessel. Geophysicists invert the recorded data to construct a 3D image of the subterranean layers. Modern seismic acquisition methods (bottom) illuminate 3D swaths of the Earth from a variety of angles. In one common geometry, four ships cruise in line abreast approximately 1.2 km [0.75 mi] apart. Each ship tows an airgun source array (red rectangles) a short distance behind. The outermost ships also tow streamers (black lines) typically 10 km [6 mi] in length, which record the reflections from below the seabed and within a rock volume (tan) beneath and between the two sets of streamers.

Streamer with hydrophone sensor arrays, 6 to 12 m deepSea surface

Seismic vessel

Airgun array

Seabed

Sedimentary layers

May 2016 63

Electromagnetic MethodsTo reduce the interpretational uncertainty remaining after seismic survey-ing, geophysicists can choose from several techniques. The most common are electromagnetic (EM) methods, which leverage the fact that some important subterranean formations have strong EM signatures. For exam-ple, rocks saturated with hydrocarbons often have much higher electrical resistivity than those containing water and is the basis of wireline resistivity logging. Salt deposits have both a high seismic velocity and a high electrical resistivity. Their high velocity makes seismic imaging beneath them prob-lematic, but their high electrical resistivity makes them easy to detect using EM surveys.

Geophysicists have two distinct methods for acquiring information about the electrical characteristics of rocks at depth. They can use either a high-powered EM source or fluctuations in the Earth’s magnetic field induced by the solar wind as a natural EM source. In both cases, the response of the Earth is detected via an array of receivers deployed on, or near, the surface. The first technique is called controlled-source EM (CSEM) and was developed in the 1980s. It is most commonly used in marine settings, where anthropogenic noise, for example, radio signals or power line noise, is less problematic than on land. The second EM tech-nique, magnetotellurics (MT), was introduced in the 1950s. Some modern systems can acquire CSEM as well as MT data when the controlled source is not active (Figure 2).

Because of the frequency of the EM signal and the acquisition geometry, MT surveys are best suited for basin-scale studies, while CSEM surveys are more appropriate for detailed reservoir-scale targets and high-resistivity anomalies. Consequently, the CSEM method is typically used to investigate potential hydrocarbon reservoirs previously suggested by seismic images.

Magnetic SurveyingMagnetic surveying is another type of subsurface prospecting. Unlike EM methods, which rely on fields that fluctuate rapidly in time, magnetic survey-ing depends on the permanent magnetic properties of rocks, whose strength and orientation are fixed at the time of their deposition and may be in con-trast with those of the surrounding rock. Measuring these subtle anomalies can help geophysicists map subsurface formations over large areas.

The advantage of magnetic surveying is that data can be collected from aircraft or satellites as well as from land or by ship. Consequently, magnetic surveys can inexpensively cover large geographic areas as well as sites that are otherwise difficult to access. Because the strongest anomalies are pro-duced by volcanic or metamorphic formations, magnetic surveys are widely used for mineral exploration.

Gravimetry SurveyingGravity measurements have been applied in the oil field since the 1920s. The technique is based on recording spatial variations in the Earth’s gravi-tational field, caused by differences in the density of rocks below the survey location. The size of these variations is typically less than 1/100,000th of Earth’s gravitational field’s nominal value of about 9.81 m/s2 [32.2 ft/s2].

Detecting such small variations requires extremely sensitive instru-ments and the application of multiple corrections. For example, the Bouguer correction accounts for variations in gravity caused by local topography and corrects for the influence of latitude and measurement altitude that might otherwise mask the signal. Because the low density of salt generates a large gravity anomaly, the most common oilfield applica-tion of gravity surveying is to help delineate salt domes. Gravity data are

frequently acquired using aircraft and satellites; taking measurements by ship is also common.

Variations and Value Geophysical methods are applied in various ways. For example, seismic receivers are sometimes deployed in boreholes to generate detailed images of small portions of the Earth. In addition, certain niche techniques, such as the hyperspectral imaging, spontaneous potential and electrokinetic—seismoelectric and electroseismic—methods, are available but are not widely used. Of all geophysical techniques, seismic surveying is by far the most widespread. Because of this dominance, “seismics” and “geophysics” are often used interchangeably in the oil industry, although for purists, this is wrong. Nevertheless, the integrated use of complementary geophysical methods provides critical information about the subsurface. This informa-tion is used by exploration and development personnel to make decisions about where and how to drill.

Figure 2. Marine magnetotellurics (MT) and controlled-source EM (CSEM) acquisition. For marine MT studies (top), electric and magnetic recorders on the seafloor make time series measurements of the Earth’s varying magnetic field and the induced electric field, which can be interpreted to infer deep geologic structures. Towing a CSEM transmitter, which has receivers behind it, close to the seafloor allows geophysicists to map shallow structures, including thin, resistive features such as hydrocarbon reservoirs. Joint evaluation of multiple geophysical measurements (bottom) enables geophysicists to obtain a consistent interpretation of the base of salt. The best interpretation based solely on seismic data showed a thick section of salt to the right of middle, whose base is indicated by the white line. Adding MT resistivity data (colors) provides significant new information. Combining seismic and MT data improves the previous interpretations of the base of salt and gives interpreters greater confidence in their result (yellow dashed line).

Air (resistive) Natural-source magnetotelluric fields

Seafloor(variable conductivity)

Oil and gas (resistive)

Seawater (conductive)CSEM transmitter

10

1

Resi

stiv

ity, o

hm.m

Distance

Dept

h

Joint imaging base salt

Seismic base salt 1 km

1 km

Salt, carbonatesand volcanics

(resistive)

Electric and magnetic field recorders

64

LOOKING BACK

On the occasion of 90 years following the July 1926 official registration of Société de Prospection Électrique (abbreviated to SPE and soon nicknamed La Pros), this Looking Back article examines the origins of the first logs and a contributed article by Philippe Theys (“Birth of La Pros,” page 57) discusses those first engineers who established an industry.

In the late 1920s, an alternative to the time-consuming process of obtaining and analyzing rock cores was born in the wine region of Alsace, France. This new technique became known as “carottage électrique,” or electrical coring, and eventually developed into the wireline logging recognized today.1

Oil had been produced near the small village of Pechelbronn, France, since the 1740s. By the 1920s, the well count was at 3,000 and increasing. The Pechelbronn Oil Company had opened a new refinery that could handle 80,000 metric tons, approximately 11,000 bbl, of oil annually and needed to know that adequate reserves were available to feed the refinery. In early 1927, company personnel discussed with Conrad Schlumberger the idea of making resistivity measurements in the borehole to help company geologists better understand the oil-bearing formations.

Marcel Schlumberger had already tested the technique in 1921, taking resistivity measurements over about a meter [several feet] at the bottom of a 760-m [2,500-ft] hole in Molières-sur-Cèze in south-ern France. The results were inconclusive, but the feasibility of a downhole resistivity measurement had been proved. The geophysical community remained skeptical, however.

On September 5, 1927, Henri-Georges Doll, Conrad’s son-in-law, and two colleagues, Roger Jost and Charles Scheibli, conducted the first electrical logging operation in a 500-m [1,600-ft] Pechelbronn well named Diefenbach 2905. The team logged an interval of 140 m [460 ft], starting from a depth of 279 m [915 ft]. They rigged up a hand-oper-ated winch that lowered into the hole three insulated wires—cables of the type used for lighting fixtures—tied together with friction tape. The longest of the wires injected current into the well and formation; the return was at the surface. The other two wires, shorter and of slightly different lengths, measured the resulting potential field and provided the resistivity readings. Measurements were made at 1-m [3 ft] intervals; the entire operation took five hours. The result was a resistivity log that distinguished between the many layers of sand and shale pierced by the borehole.

Resistivity logging continued throughout 1928 at Pechelbronn, and the resulting correlations of resistivity from one well to the next revolutionized the understanding of the stratigraphy of the field. Soon, the Pechelbronn Oil Company was able to raise the annual capacity of its new refinery to 100,000 metric tons. By 1929, the new logging technique had gone global—Schlumberger logging crews were engaged by Shell for their explorations in Venezuela, the US and the Dutch East Indies, and by the Soviet Union for the oil fields of Grozny, Chechnya, and Baku, Azerbaijan.

Looking Back on Wireline

Meanwhile, Conrad was playing with new ideas for his well log-ging. Realizing that oil was infinitely resistive to electricity, he postulated that if a zone was oil bearing and reasonably thick, then increasing the spacing between the wires’ measuring potential would allow the logging tool to see deeper into the formation and record a higher resistivity; the phenomenon was not observed in Pechelbronn because the oil zones were too thin. In May 1930, he asked his field engineers to try the idea and report back. Within a month, Marcel Jabiol, the company’s solitary engineer in northern Sumatra, sent a telegram back confirming Conrad’s hypothesis. Electrical logging could locate oil from the borehole.

1. Mau M and Edmundson H: Groundbreakers: The Story of Oilfield Technology and the People Who Made it Happen. Peterborough, England: Fast-Print Publishing, 2015.

Pulley used in Pechelbronn in 1928. (Illustration courtesy of and adapted from Mau and Edmundson, reference 1.)

Harnessing the PotentialCombining another measurement technique with electrical logging, however, was necessary to fully optimize the electrical logging method—namely, spontaneous potential (SP). Natural electrical potentials in the subsurface had been discovered about a century earlier in Cornwall, England, by the British geologist, natural philoso-pher and inventor Robert Were Fox. Potentials in the borehole are caused by electrochemical interactions between the borehole fluid and adjacent sand and shale formations.

Conrad Schlumberger secured a French patent on the SP tech-nique in 1929, claiming it could be used to locate permeable strata, but found no practical application. A year later, Doll observed natural potentials while logging in the Oklahoma Seminole oil field. While the battery was disconnected, he noticed the potentiometer needle vibrating back and forth as the electrodes were lowered into the well. Experiments followed at Pechelbronn, and by 1930, they concluded that the SP method could differentiate permeable beds such as sand and limestone from impermeable formations such as shale. The combination of SP and resistivity curves turned out to be of much greater value than the resistivity log alone in detecting pro-duction possibilities. Electrical logging established a reliable method of achieving stratigraphic correlation, providing a way to distinguish between shale and porous rock and between hydrocarbon- and water-bearing rock.

In the decades that followed, electrical logging became increas-ingly sophisticated and paved the way for numerous other well log-ging techniques. It became an industry on its own, taking the key position among formation evaluation techniques.

Grow your Knowledge. For experienced professionals, newcomers and those simply interested in learning more about our industry, the Defining Series provides summaries of a wide range of industry topics, efficiently communicating basic principles and underlying science.

Added to the series and appearing in this issue are subsea infrastructure and geophysics. See the entire series online to grow your knowledge.

http://www.slb.com/oilfieldreview

Oilfield ReviewAuthoritative. Relevant. Informative.

Oilfield ReviewAuthoritative. Relevant. Informative.