Oilfield Review May 2016

Marine Seismic Imaging Microseismic Fracture Mapping Corrosion Review Slide Drilling—Farther and Faster 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 Review Authoritative. Relevant. Informative.

16-OR-0002 EDITORIAL

Schlumberger and Cameron: A Meeting of Minds and Technologies

In the early 1920s, Jim Abercrombie and Harry Cameron as a consequence, they are recognized as innovators in designed and built the world’s first blowout preventer their respective industry sectors. (BOP) in their shop in Houston. In France, brothers Conrad Cameron is adopting wireless technologies to track assets and Marcel Schlumberger were developing methods to and monitor equipment condition and is investigating the explore the Earth’s subsurface using electronic sensors. advantages of nanotechnology and composite materials. The Cameron Iron Works Company was established in And both companies will combine existing research pro- Houston in 1920, just six years before the Schlumberger grams such as the use of 3D printing to print metal parts brothers registered their company—Société de Prospection using high-end metals such as Inconel† alloy. This tech- Électrique—in France. nology has the potential to allow engineers to design highly Although the two companies are of similar vintage, they complex structures that would otherwise be impossible pursued different paths to success and now are one. to manufacture. Cameron built its reputation predominantly on the surface, providing the equipment necessary for operators to manage The integration of Schlumberger and drilling and production operations. Schlumberger success is anchored by decades developing and refining methods to Cameron is a story of combining expertise remotely explore, understand and produce from reservoirs and innovative technologies. below the surface. As is often the case when two companies join forces, the greatest benefit arises from a unique com- Schlumberger state-of-the-art mechanical, electronics bination of similarities and differences. and sensing technologies, combined with the rigorous Cameron has constantly defined and refined the surface application of materials science, have led to the develop- technologies and services essential for operators to drill, ment of high-performance downhole drilling systems. complete and produce hydrocarbons. Schlumberger tech- Joining these systems with the Cameron drilling equip- nologies help operators find, access, quantify and optimally ment portfolio and automation capabilities will create a recover hydrocarbons. Both companies have forged much fully integrated land drilling system to create a step of their reputations in the most challenging surface and change in operational efficiency. subsurface environments. The integration of Schlumberger and Cameron is a story Cameron provides the means and methods to confidently of combining expertise and innovative technologies. The manage and control flow in all environments, pressure union creates a company that can function as a partnership regimes and temperature ranges throughout the world. internally and with its clients. In the larger view, the inte- Schlumberger uses technology to help operators understand gration of surface and subsurface experts can enhance the and optimize their assets through seismic surveying, down- performance of drilling and production systems and serve hole logging, testing, laboratory analysis and software plat- as a model of the collaboration and commercially beneficial forms. For Cameron and Schlumberger, the meeting place alignment between operator and service company. is the wellhead. Managed pressure drilling (MPD) is one illustration of Justin Rounce Vice President Marketing & Technology, Cameron Group the Schlumberger and Cameron fit. Today, a small but Houston, Texas, USA growing group of engineers is coming to view MPD not as a specialty method reserved for wells that are otherwise Justin Rounce returns to Schlumberger as the Vice President Marketing & technically impossible to drill but as a conventional drill- Technology, Cameron Group from the position of Cameron vice president of marketing and chief technology officer. Prior to joining Cameron, Justin was ing technique. Schlumberger will leverage its research, vice president of marketing and technology for OneSubsea, a joint venture software and engineering analytics capabilities and down- with Schlumberger and vice president and director, mergers and acquisitions hole expertise together with Cameron manufacturing, flow 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 control and automation technology to enable MPD to fulfill various roles within the company, including field operations, operations man- its potential as a standard well construction technique. agement, product management and new technology development. In 2003, Schlumberger and Cameron also share a culture of excel- he joined the Schlumberger Information Solution Segment as vice president, software products. In 2007, he was promoted to vice president, software lence. They employ award-winning engineering fellows, governance, and was responsible for all internal and commercial software in holders of PhDs and recognized industry experts. Both Schlumberger. In 2009, he became vice president, marketing and technology companies invest substantially in research and development; for the Schlumberger Production Group.

† Inconel is a registered trademark of Special Metals Corporation.

1 Oilfield Review

May 2016 Volume 28, Number 2 Articles ISSN 0923-1730 www.slb.com/oilfieldreview 4 Executive Editor Marine Imaging in Three Dimensions: Charlie Cosad Viewing Complex Structures Senior Editors Seismic imaging may fail to resolve potential Tony Smithson exploration targets beneath shallow rock Matt Varhaug layers. New developments in marine seismic Rick von Flatern acquisition and imaging are helping to reduce Editors this uncertainty. Irene Færgestad Richard Nolen-Hoeksema

Contributing Editors David Allen 16 H. David Leslie Hydraulic Fracturing Insights from Ginger Oppenheimer Microseismic Monitoring Microseismic monitoring is used for evaluat- Design/Production Herring Design ing hydraulic stimulation operations in unconventional reservoirs. Advanced tech- Illustration niques and new technologies are allowing opera- Herring Design Departments Pennebaker tors to use microseismic data for effective well George Stewart placement and manage stimulations in real time. 1 Printing Editorial RR Donnelley—Wetmore Plant Schlumberger and Cameron: A Meeting of 34 Minds and Technologies Advisory Panel Corrosion—The Longest War Hani Elshahawi Oil and gas operations often provide ideal envi- 57 Shell Exploration and Production ronments for corrosion to develop and grow. In Houston, Texas, USA Looking Back spite of the challenges created by corrosion, anti- Birth of La Pros: The 90th Anniversary of the Gretchen M. Gillis corrosion practices in use today help operators Aramco Services Company first Schlumberger Company Houston, Texas maintain equipment integrity and safely produce Roland Hamp hydrocarbons. 60 Woodside Energy Ltd. The Defining Series Perth, Australia Selected from the Defining Series online: Dilip M. Kale 50 Subsea Infrastructure and Geophysics ONGC Energy Centre Slide Drilling—Farther and Faster Delhi, India A surface-mounted torque-oscillation system 64 George King provides consistent weight transfer to the bit, Looking Back Apache Corporation Houston, Texas improving rate of penetration and directional Origins of the Technique of Wireline Logging control during sliding operations. Michael Oristaglio Yale Climate & Energy Institute New Haven, Connecticut, USA John Thorogood Drilling Global Consultant LLP Aberdeenshire, UK On the cover: Cameron specialists with a seven-cavity TL* offshore ram-type blowout preventer (BOP) Publishing stack at the Cameron facility in Berwick, Louisiana, USA. Cameron became part of Oilfield Review is published and Schlumberger in March 2016. Although the company is most well-known for BOPs, the printed in the USA. 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 © 2016 Schlumberger. as valves; OneSubsea provides subsea production and processing systems. All rights reserved. An asterisk (*) denotes a mark of Schlumberger.

2 Article Summaries

Marine Imaging in Three Hydraulic Fracturing Insights Corrosion—The Longest War Slide Drilling—Farther Dimensions: Viewing Complex from Microseismic Monitoring Corrosion has brought down bridges, and Faster Structures Horizontal drilling and hydraulic downed aircraft, leveled chemical Directional wells allow operators to Prospective reservoirs are typically fracturing revolutionized the exploi- plants, parted drillpipe and ruptured efficiently access reservoirs and located beneath complex rock layers tation of tight and unconventional pipelines. Given sufficient time, this maximize wellbore exposure to pro- that act as reflective barriers to oil and gas reservoirs. Engineers phenomenon has the potential to ductive zones. In many such wells, seismic signals and thus distort can track the progress of hydraulic degrade any material. In certain directional drillers use steerable images of the subsurface. The chal- fractures through a formation using environments, the unchecked effects mud motors to kick off the well, lenge to geophysicists is to peer microseismic monitoring. of corrosion can come swiftly, and build angle, drill tangent sections through the overburden to the reser- Armed with new techniques and the consequences of failure to man- and maintain trajectory necessary to voir beneath it. technologies, operators are able to age corrosion can be costly. hit target zones. To do so, geoscientists use seismic develop unconventional reservoirs One North Sea operator reported When using mud motors, drillers survey data, which undergo advanced with reduced risk and greater under- that outlays for corrosion prevention alternate between rotating and sliding processing, imaging, inversion and standing and certainty than ever and control averaged about 8% of its modes of drilling. In rotating mode, in interpretation workflows. The results before. By mapping the spatial and total project capital expenditures. addition to the downhole motor, the enable exploration teams to make temporal patterns produced by micro- Direct costs associated with those drilling rig’s rotary table or topdrive immediate decisions about a pros- seismic events, they can refine reser- expenses include replacement of cor- rotates the entire drillstring to trans- pect’s value. voir models and create field roded equipment and lost production mit power to the bit. During slide In challenging offshore frontier development strategies. and contamination; indirect costs drilling, the drillstring does not rotate, environments, operators are using a Microseismic monitoring, intro- arise from health, safety and environ- and the bit is turned by the mud recently developed seismic technol- duced in the 1980s, evolved as the ment concerns. motor alone. As a consequence, less ogy and inversion process to provide application of hydraulic fracturing of The environments that host oil and weight is transferred to the bit, and 3D, full-bandwidth imaging of fine- horizontal wells increased since that gas operations often provide ideal slide drilling is less efficient than scale structures. Full waveform time. Using new methods and tech- conditions for corrosion. Ongoing rotary drilling. inversion results in a model of seis- niques, geoscientists are improving research and advances in coatings, An automated torque control sys- mic velocities that can be used with the answers that can be derived from cathodic protection, nondestructive tem alternates torque direction to the seismic data to form an image of microseismic data. Engineers can testing, corrosion analysis and inhibi- rock the drillstring, which increases the geology from the surface to the even adjust stimulation programs on tors allow operators to safely produce ROP through better transfer of weight targets of interest. Page 4. the fly using real-time data to respond oil and gas in these corrosive environ- to the bit while in sliding mode. This to downhole events as they occur. ments. Page 34. weight transfer also helps control tool- Modeling software can predict antici- face orientation. In addition, it helps pated downhole responses and then minimize the number of downhole update the production predictions motor stalls and increases bit life by based on the data they acquire from preventing weight from being trans- downhole. Page 16. ferred to the bit suddenly. Page 50.

About Oilfield Review Oilfield Review online Oilfield Review app Correspondence Oilfield Review is the Schlumberger Visit www.slb.com/oilfieldreview for the Download the free app by visiting the Oilfield Review flagship journal of technology, innovation current and all previous editions, videos Apple or Google Play online store and 5599 San Felipe and the science of E&P. Contributors to associated with certain articles, the searching for Schlumberger Oilfield Houston, TX 77056 articles are industry professionals and defining series, as well as links to the Review. Previous issues and video United States experts from around the world. Those apps and Oilfield Glossary. content are available. (1) 713-513-3760 listed with only geographic locations are E-mail: [email protected] employees of Schlumberger. www.slb.com/oilfieldreview

Reproductions without permission are strictly prohibited. 3 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 Aseev Hydrocarbon exploration requires that geoscien- interpretation workflows. These workflows pro- Moscow, Russia tists understand the geology of prospective reser- vide vital inputs for geomechanical, reservoir and voirs often located beneath complex rock layers. basin models. Sandeep Kumar Chandola From the geophysicist’s perspective, the overbur- IsoMetrix marine isometric seismic technology Low Cheng Foo den acts as a defective lens, distorting seismic and full waveform inversion processing are PETRONAS Carigali Sdn Bhd images of deeper geologic structures. As a result, enabling imaging of complex structures in frontier Kuala Lumpur, Malaysia targets appear indistinct, distorted, out of place or, areas. The IsoMetrix technology allows for full- in extreme cases, completely obscured. The geo- bandwidth imaging of fine-scale structures in the Chris Cunnell Malcolm Francis physicist’s challenge has been to devise methods subsurface in all directions—inline, crossline and Shruti Gupta for peering through the overburden and bringing vertical—for detailed imaging from seabed to Peter Watterson the underlying geology into focus. reservoir. Full waveform inversion results in a Gatwick, England Make-or-break decisions on project viability model of seismic velocities, which is used with the often hinge on how well prospective reservoirs can seismic data to form an image of the geology from Michelle Tham be imaged, a key factor determining exploration the surface to the targets of interest. Kuala Lumpur, Malaysia risk. Operators need accurate images of reservoirs This article describes surveys acquired using to help them place exploration wells where they IsoMetrix technology in offshore Malaysia and Oilfield Review 28, no. 2 (May 2016). Copyright © 2016 Schlumberger. effectively test the prospect, conduct field planning the North Sea. The survey results demonstrate For help in preparation of this article, thanks to Thomas and place development wells. In addition to imaging the benefits of IsoMetrix technology for over- Ajewole, M. Nabil El Kady, M. Faizal Idris, Satyabrata Nayak reservoirs, geophysicists must correctly image the coming a challenging acquisition environment and M. Iqbal Supardy, PETRONAS Carigali Sdn Bhd, Kuala Lumpur, Malaysia; and Richard Coates, Houston, Texas, USA. overburden—the layers above the reservoir—to and increasing spatial bandwidth and of applying Dynel 2D, IsoMetrix and Q-Marine are marks of reduce drilling risks from operational challenges full waveform inversion for determining over- Schlumberger. such as maintaining a stable wellbore and burden and reservoir properties, specifically seis- 1. Christie P, Nichols D, Özbek A, Curtis T, Larsen L, Strudley A, Davis R and Svendsen M: “Raising the controlling formation pressure. mic velocities. Standards of Seismic Data Quality,” Oilfield Review 13, The value that seismic data adds to the no. 2 (Summer 2001): 16–31. Improving Data and Image Quality 2. Robertsson JOA, Moore I, Vassallo M, Özdemir K, exploration process depends on the quality of van Manen D-J and Özbek A: “On the Use of the image produced and the cost incurred in Good seismic imaging requires a chain of factors: Multicomponent Streamer Recordings for Reconstruction of Pressure Wavefields in the Crossline Direction,” acquiring such data. Cost-effective seismic a good acquisition system, optimal survey geom- Geophysics 73, no. 5 (September–October 2008): acquisition requires surveying large areas etry and accurate processing algorithms and A45–A49. quickly without compromising data quality and workflows. More than 15 years ago, Schlumberger 3. For more on full-azimuth seismic surveying and imaging: Brice T, Buia M, Cooke A, Hill D, Palmer E, Khaled N, while minimizing operational and environmental geophysicists embarked on a program to move Tchikanha S, Zamboni E, Kotochigov E and Moldoveanu N: exposure. Fast acquisition helps shorten the from conventional seismic acquisition toward dis- “Developments in Full Azimuth Marine Seismic Imaging,” Oilfield Review 25, no. 1 (Spring 2013): 42–55. time frame between the decision to evaluate a crete sensor technology. The technology includes 4. Inline is in the direction the seismic vessel travels and play and the decision to drill. improvements in receiver sensitivity and position- acquires data; crossline is in the direction perpendicular ing accuracy, steerable streamers, increased to vessel travel. High-quality data enable exploration teams to 5. For more on the Q-Marine system: Christie et al, reference 1. attain a clear understanding of the geology from source control and point-receiver acquisition, For more on the 3C seismic MEMS accelerometer unit: the seafloor to the target prospect and then to which records traces from individual receivers to Paulson H, Husom VA and Goujon N: “A MEMS provide consistently repeatable high-quality data.1 Accelerometer for Multicomponent Streamers,” paper decide whether to test and appraise the prospect. We P6 06, presented at the 77th European Association of The acquired data must also be suitable for use in These capabilities are evolving. New measure- Geoscientists and Engineers Conference and Exhibition, Madrid, Spain, June 1–4, 2015. advanced processing, imaging, inversion and ments of the crossline and vertical gradients—

4 Oilfield Review 10 mm Single Channel Interpolation

x

P Crossline direction y z Multichannel Reconstruction

Crossline direction

Figure 1. Streamer element. An element of the IsoMetrix streamer system the wavefield (blue) can be measured at each streamer location (black (left) combines a hydrophone (inset) that measures pressure (P) with a dots). Therefore, the reconstructed wavefield (red) between streamers calibrated, triaxial microelectromechanical (MEMS) accelerometer that is aliased and incorrect. Using multisensor streamers (bottom right), the measures the axial, or inline (x), radial, or crossline (y), and vertical (z) wavefield amplitude and gradient (cyan) can be measured at each streamer accelerations. The IsoMetrix technology facilitates interpolation between location. Consequently, using both attributes of the wavefield, geophysicists streamers. Using hydrophone streamers (top right), only the amplitude of can reconstruct the wavefield accurately between streamers. variations with distance—of the pressure wave- engineers who develop static and dynamic crossline direction may be aliased, or inade- field enable the signals received from a marine models of the reservoir. These models are based quately sampled. seismic shot to be processed as a full 3D wave- on the seismic results—images, velocities and Often, the focus of marine seismic imaging is to field rather than as a collection of 2D profiles.2 In horizons—that are integrated with well data. thoroughly sample the wavefield in the reservoir. addition, a newly developed, calibrated, broad- Before drilling, explorationists use the models to However, good sampling of the wavefield in the band marine seismic source provides improved predict the petroleum systems present within the overburden is also important because these depths low-frequency signal content; no source notches, seismically imaged volume, define plays and must be imaged correctly to enable the geophysi- or missing frequencies, below 150 Hz for all direc- locate prospects for drilling. Reservoir engineers cist to see clearly into the reservoir. Sampling the tions within a 20° cone from the vertical; and use refinements of these models to plan field seabed or other interfaces that generate multiple cancellation of the source ghost—a delayed development and, later, manage hydrocarbon reflections is important because such reflections reflection of the source from the sea surface. recovery operations. interfere with primary reflections. Shallow depths These acquisition improvements have been are important because of possible seabed and complemented by innovations in marine Imaging Between Streamers shallow subsurface hazards to drilling. surveying geometries—for example, multivessel The purpose of IsoMetrix technology is to pro- Typical marine seismic receivers are hydro- shooting and full-azimuth source-receiver vide a densely sampled representation of the phones that record the pressure wavefield only. configurations. Together, these technologies wavefield in all directions. An idealized seismic Reconstruction of the pressure field between make it possible to illuminate targets of interest acquisition system would be able to record the streamers requires interpolation between previously obscured by folded or faulted seismic signals from everywhere below the sur- known pressures at each streamer location and sediment, overlying salt layers or other complex face. This capability would maximize the oppor- results in crossline pressure fields becoming geologic bodies.3 tunities for separating the signal from unwanted aliased and incorrect. Seismic acquisition and survey geometry are noise and imaging the reflectors in the subsur- The IsoMetrix technology is based on the only the starting points for seismic imaging. face. However, conventional seismic data are Q-Marine point-receiver marine seismic system Accompanied by onboard processing capabilities, recorded along only a small number of long and combines hydrophones for measuring the seis- data reliability has vastly improved. In addition, streamers towed behind a vessel. Thus, although mic wavefield pressure with a three-component the application of robust seismic inversion and conventional seismic data are well sampled in (3C) microelectromechanical systems (MEMS) imaging techniques, such as full waveform inver- recording time and along the streamer (inline), unit.5 The 3C MEMS unit contains three orthogo- sion and reverse time migration, allow geophysi- they are not recorded between the streamers nal accelerometers for measuring the full 3D vec- cists to deliver sharper images and estimate rock (crossline), which may be separated by large dis- torial motion—magnitude and direction—of the properties for explorationists and reservoir tances of 50, 75 or 100 m [164, 246 and 328 ft].4 recorded wavefield (Figure 1). As a result, any waves propagating in the

May 2016 5 Survey vessel 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- Streamers 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 Seismic dataset 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, Crosslin e Inline any energy arriving from the crossline direction, which had been spatially aliased previously in Figure 2. Marine seismic acquisition via conventional versus IsoMetrix technology. Conventional conventional datasets, can now be sampled surveys (left) are acquired using streamers of closely spaced hydrophones. The resultant seismic appropriately using GMP spatially and tempo- 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 rally by taking advantage of the crossline and interpolation between streamers, and the resultant dataset is a true 3D grid. vertical gradient measurements. The GMP process is data driven and has proved that it can interpolate the pressure wavefield accurately in the crossline direction, By adding 3C accelerometers, the marine seismic data using streamers spaced farther even in adverse situations in which the results receivers record the variation of acceleration, apart than those in conventional surveys and to from conventional processing would be highly which is proportional to the pressure gradient, or reconstruct the 3D wavefield on a dense grid at aliased. The output from the GMP process is a the spatial derivative of pressure with direction. points between streamers (Figure 2). For exam- grid of data channels spaced 6.25 m [20.5 ft] In an acoustic material such as water, hydro- ple, if the actual recordings were accomplished apart in the inline direction along virtual phones measure the pressure (P) fluctuations using eight streamers spaced 75 m apart, provid- streamers, which are nominally separated by caused by the seismic wave. Three-component ing a streamer spread that is 525 m [1,720 ft] 6.25 m in the crossline direction. accelerometers measure the accelerations in wide, the wavefield may be reconstructed as if it The ability to image in 3D enables geophysi- three orthogonal directions (ax, ay and az). were recorded using virtual streamers spaced a cists to consider seismic survey acquisition Newton’s Second Law specifies the force that tenth of the distance—7.5 m [24.6 ft] apart. designs that depart from common practice, as results from a difference in pressure; the force is When wide streamer spacing is used, areas of one operator learned when faced with data directed from high to low pressure. The relation- exploration can be surveyed faster and more effi- acquisition challenges. ship between the difference in pressure with ciently using fewer sail lines, thereby reducing direction—the spatial derivative of P—and the survey duration, acquisition cost, operational Challenging Acquisition Conditions acceleration, for example in the x direction, is complexity and exposure to adverse environmen- To clearly define prospects in the South China

ρ × ax = −∂P/∂x, where ρ is the material density, tal conditions. Sea, geophysicists at PETRONAS Carigali Sdn and the direction of force is opposite, or negative Recording the vertical wavefield component Bhd acquired a broadband 3D seismic survey off- to, that of the pressure gradient. This type of rela- improves the geophysicist’s ability to remove shore Malaysia. The survey area is an elongated tionship holds for each spatial direction (x, y and z) noise, particularly ghost reflections, which are rectangle oriented NW–SE. A major N–S striking and allows the calculation of the spatial derivative always present in marine seismic survey record- fault crosses the survey area, and structural dips of pressure directly from the acceleration mea- ings. Ghosts are generated when the upward trav- 6. For more on generalized matching pursuit: Özbek A, surement. Consequently, knowing the pressure eling primary signal is reflected downward by the Vassallo M, Özdemir K, van Manen D-J and gradients, geophysicists can reconstruct the una- sea-air interface. This downward traveling ghost is Eggenberger K: “Crossline Wavefield Reconstruction from Multicomponent Streamer Data: Part 2—Joint liased pressure field in all directions. Therefore, detected by the seismic receivers and, if uncor- Interpolation and 3D Up/Down Separation by Generalized geophysicists can estimate the 3D wavefield rected, causes a frequency dependent blurring of Matching Pursuit,” Geophysics 75, no. 6 (November– around the streamers using the same spacing in all the final image. Using the vertical acceleration December 2010): WB69–WB85. 7. Chandola SK, Foo LC, El Kaldy MN, Ajewole TO, Nayak S, directions—inline, crossline and vertical. measurements, the geophysicist can separate the Idris MF, Supardy MI, Tham M, Bayly M, Hydal S, upgoing and downgoing components of the wave- Seymour N and Chowdhury B: “Dip or Strike?— Complementing Geophysical Sampling Requirements and Reconstructing the Wavefield field, thereby facilitating removal of ghost reflec- Acquisition Efficiency,” Expanded Abstracts, 85th SEG The ability to measure the crossline wavefield tions. The ability to remove the ghosts also allows Annual International Meeting and Exhibition, New Orleans (October 18–23, 2015): 110–114. gradient enables geophysicists to acquire marine IsoMetrix streamers to be towed deeper than

6 Oilfield Review as high as 50° occur in the area along a W–E trend (Figure 3). The area is bounded on the west Average dip ~ by a no-access zone that survey vessels are not 7˚ N 7 permitted to enter. Fault Typically, optimal seismic acquisition geometry Average dip for conventional 3D surveys requires shooting ~40˚ to 50˚ parallel to the predominant structural dip No-access area direction. This inline direction facilitates close- 60˚ spaced sampling of the seismic wavefield in the dip direction, in this case W–E, in which the geology 30˚ has the most variation. In addition, the typical Average dip conventional seismic bin, or survey subdivision, ~25˚ into which geophysicists sort seismic traces, is 0˚ asymmetric and elongated in the structural strike direction, which is the crossline direction. Structural dip, E–W direction Geologic dip of surface of interest Major faults, N–S direction The no-access zone prohibited the vessel from obtaining full subsurface coverage at the western Figure 3. Geologic structure. In the time structure map of the horizon of interest (left), the contour edge of the survey and presented an acquisition interval is 100 ms two-way traveltime. The black area is a major fault surface that dips to the east. challenge to geophysicists, who considered two 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 options (Figure 4). In the first scenario, they the right—the surface area at the prospective reservoir level—shows structural dips that have been could acquire most of the survey by shooting estimated from legacy seismic data. The dips are aligned along a W–E trend. 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 Scenario 1 Scenario 2 Main Survey Area Entire Survey Area parallel to strike adjacent to the no-access zone E–W shooting N–S shooting boundary; the close line spacing of these strike- 10 streamers, 8,000 m long, 10 streamers, 8,000 m long, parallel lines ensured adequate sampling of the towed 100 m apart towed 100 m apart Patch Survey Area During seismic processing, the structural dip. Alternatively, they could acquire N–S shooting streamer data are processed the entire survey using exclusively strike-parallel 10 streamers, 8,000 m long, and output to a 6.25-m × 6.25-m towed 50 m apart grid that is 8,000 m long and sail lines. 950 m wide. The first option was inefficient because of the two acquisition directions, which required nonpro- ˜ 16 km ˜ 16 km Strike ductive time during the many turns and while the Strike ˜ 5 km ˜ 5 km streamers were repositioned for close spacing. The Dip Dip second option was more efficient for acquiring data but risked degrading the seismic information

˜ if acquired using conventional technology. 50 km ˜ Main Fault Zone 50 km According to conventional wisdom, the strike-par- allel survey direction, which had typical line spac- Patch Survey Are No-access area ing and sampling of the seismic wavefield in the No-access area Entire Survey Area dip direction, was not ideal for imaging the subsur- face and meeting the objectives of company geolo- gists and geophysicists. a a The company used IsoMetrix technology, Main Survey Are 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- Estimated Survey Duration: pany acquired a smaller swath of data in the Scenario 1 = 2 × Scenario 2. direction of the dominant structural dip, which would allow comparison and validation of the Figure 4. Acquisition options. The survey was restricted by a no-access area on its western integrity of survey shooting in strike. boundary. The company geophysicists considered two options for acquiring the seismic data. In the The data were acquired using ten 8-km [5-mi] first option (left), the acquisition vessel would sail the main survey area in the dip direction and then long streamers spaced 100 m apart. The stream- 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 ers were towed at a water depth of 18 m [59 ft] to in the direction of geologic strike and would parallel the no-access boundary. The company chose minimize noise from variable currents and the second option and elected to use IsoMetrix technology, which allows for reconstruction of the inclement weather during the survey campaign. wavefield sampled equally in both inline and crossline directions, to acquire the data.

May 2016 7 Bathymetry from Multibeam Echo Sounder Bathymetry from IsoMetrix Technology

5 km 5 km

10 × 100 m spread 5 × 5 m sampling 6.25 × 6.25 m binning

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.

After acquisition, the data were preprocessed The data proved to be of high quality. For direction, geophysicists judged the datasets to be and then the full 3D wavefield was calculated example, a map of the seafloor surface showed similar (Figure 6). The fine spatial sampling of using simultaneous interpolation and deghosting sand banks similar to those observed in bathym- the wavefield in the inline and crossline direc- by means of the GMP method. Next, the upgoing etry data obtained using a high-resolution multi- tions obtained with IsoMetrix technology enabled pressure wavefield (P-wave) was output on a beam echo sounder (Figure 5). the company to accomplish its geologic and geo- 6.25‑m by 6.25‑m grid for each shot record for Upon comparing the dataset acquired in the physical objectives and achieve acquisition oper- further processing and imaging. strike direction with that acquired in the dip ational efficiency. In addition to freeing up constraints on seismic Control Swath Production Volume survey acquisition design, uniform inline and crossline data wavefield estimation facilitates the increase in spatial resolution and bandwidth 1 s 1 s 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 3D Oil 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 5 s 5 s 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 Dip shooting direction, inline stack Strike shooting direction, crossline stack west of the Loppa High, a roughly 150 km [90 mi] long and 100 km [60 mi] wide tilted fault block 18 km 18 km that has been affected by a series of events in the North Atlantic Ocean that include: 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 • Paleozoic rifting the control swath acquired in the dip, or crossline, direction and stacked in the strike, or inline, • Mesozoic opening of the North Atlantic Ocean direction. The section on the right was from the production volume and acquired in the strike direction and of the Greenland and Norwegian seas but stacked in the dip direction. Except for subtle differences, the sections show similar results • Quaternary glaciation. 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.

8 Oilfield Review W E 0.5 s Bjørnøyrenna Hoop fault complex fault complex

Bjørnøyrenna fault complex

A A

Hoop fault complex Loppa High

5.0 s

N S 0.5 s

Asterias fault complex

Asterias fault complex B B

Variance

LowHigh 5.0 s Reflection amplitude

–0+

Figure 7. Fault system. This seismic time slice (left) at 1,100 ms is through Loppa High graben and the Bjørnøyrenna fault complex separates the the Loppa High; the seismic attribute is displayed to emphasize the variance Loppa High from the Bjørnøya basin (not shown) on the west. Sections A in seismic reflectivity—areas of high variance values are colored from and B (right top and bottom) display grabens associated with the fault black to red and yellow. Three major fault systems, which show up as areas systems. The northern portion of the Loppa structure is in the center of of high variance, affected the Loppa High. The W–E striking Asterias fault Section A. Section B shows graben structures in the north associated with complex crosses the Loppa structure in the south; the southern portion of the Hoop fault complex and in the south associated with the Asterias fault the SW–NE striking Hoop fault complex cuts across and forms the narrow complex, which separates the Loppa High from the Hammerfest basin.

The WesternGeco seismic vessel Western spatial bandwidth in all directions. In the spatial were preprocessed and then simultaneously spa- Trident acquired the East Loppa Ridge survey in domain, the wavenumber (k) is the spatial tially dealiased and receiver-deghosted in 3D by 2014. The survey covered 4,777 km2 [1,844 mi2] frequency, or the number of wavelengths—wave- means of the GMP method. and is part of the Schlumberger Multiclient cycle lengths—λ per unit distance. The The tectonic, stratigraphic and petroleum Barents Sea program. The program used wavenumber is analogous to the temporal systems geology of the southwest Barents Sea IsoMetrix technology to record wide spatial frequency (f) or the number of wave periods— region is complex.9 The structural setting resulted bandwidth data—the recorded wavefield con- wave-cycle times—T per unit time. Wavenumber from several tectonic events that established a tains the fine-scale detail necessary to represent in the space domain and frequency in the time dense mosaic of fault systems (Figure 7). The subsurface geology accurately. domain are related through the phase velocity 8. In this context, phase refers to a wave of a single In conventional 3D seismic surveys, a com- (vp), which is equivalent to wavelength divided frequency in a wave train. The phase could be of a mon objective is to acquire broadband surveys of by period (vp = λ/T), frequency divided by compressional (P) wave, shear (S) wave, other waves or their associated reflections and refractions; the wave’s high temporal—traveltime—bandwidth and wavenumber (vp = f/k) or wavelength times fre- velocity is the phase velocity. 8 resolution. The ideal broadband survey has a quency (vp = λ × f ). Consequently, for 3D seis- 9. Henriksen E, Ryseth AE, Larssen GB, Heide T, Rønning K, wide band, or range, of frequencies and is mic imaging of geology, the notion of broadband Sollid K and Stoupakova AV: “Tectonostratigraphy of the Greater Barents Sea: Implications for Petroleum acquired at a high sample rate. The objective for must be expanded to include 3D spatial band- Systems,” in Spencer AM, Embry AF, Gautier DL, maximizing temporal bandwidth is primarily to width and resolution. Stoupakova AV and Sørensen K (eds): Arctic Petroleum Geology. London: The Geological Society, Memoir 35 maximize resolution in depth—to image thin The East Loppa Ridge survey was acquired (August 9, 2011): 163–195. beds and small faults. using 12 streamers that were 7 km [4.3 mi] long, Gernigon L, Brönner M, Roberts D, Olesen O, Nasuti A and Yamasaki T: “Crustal and Basin Evolution of the Geology is best understood by observations in spaced 75 m apart and towed at a constant depth Southwestern Barents Sea: From Caledonian Orogeny to three dimensions, which requires maximizing of 25 m [82 ft]. After acquisition, the datasets Continental Breakup,” Tectonics 33, no. 4 (April 2014): 347–373.

May 2016 9 W E Loppa High area contains three major fault com- 0.5 s 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

5 km petroleum systems. 5.0 s The Gohta and Alta oil discoveries were in Reflection amplitude reservoirs located in carbonates of the Gipsdalen Group, which were deposited in warm, shallow –0+ marine environments during the Late Carboniferous to Permian periods and, since BCU Seafloor then, have been altered by dolomitization and karstification (Figure 8).11 Additional petroleum Tertiary systems elements in the Loppa High area include Upper Triassic-Jurassic reservoir prospects in Triassic sandstones, source rocks in Carboniferous synrift and postrift sedi- Lower-Middle Triassic ments and in Permian and Triassic sediments and seals formed by Triassic and Cretaceous Late Carboniferous to 12 Permian carbonates shales. The broadband East Loppa Ridge seis- mic dataset offers an opportunity for detailed Evaporites Carboniferous postrift interpretation of the complex geology in the Loppa High area. Carboniferous synrift The upper Paleozoic carbonates have been Pre-Carboniferous the most promising stratigraphic level for Loppa basement High exploration (Figure 9). Broadband seismic data facilitate detailed mapping, analysis and 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, interpretation of the carbonate morphology, which was modeled using the Dynel 2D restoration and forward modeling tool. This section shows which has polygonal ridges characteristic of mod- extensional rifting and synrift and postrift sediment deposition during the Carboniferous period. ern carbonate platforms. 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 Oil has been discovered in the upper Triassic of the carbonates and sedimentation of shales occurred. The upper Triassic and Jurassic periods Snadd Formation but at locations with low res- were characterized by high clastic sedimentation rates and floodplain development from rivers and ervoir quality. Within the Snadd Formation, the deltas. Rifting occurred again during the upper Jurassic to lower Cretaceous; the base Cretaceous broadband seismic data reveal the fluvial sys- unconformity (BCU) defines the transition from synrift to postrift sedimentation. Finally, Tertiary sediments occur above the BCU to the seafloor. 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 12. For more on petroleum systems: Al-Hajeri MM, International Meeting and Exhibition, Las Vegas, Riis F: “Structural Elements of the Norwegian Al Saeed M, Derks J, Fuchs T, Hantschel T, Kauerauf A, Nevada, USA (November 4–9, 2012). Continental Shelf. Part I: The Barents Sea Region,” Neumaier M, Schenk O, Swientek O, Tessen N, Welte D, Houbiers M, Mispel J, Knudsen BE and Amundsen L: Stavanger, Norway: Norwegian Petroleum Directorate, Wygrala B, Kornpihl D and Peters K: “Basin and “FWI with OBC Data from the Mariner Field, UK—The NPD Bulletin no. 6, May 1990. Petroleum System Modeling,” Oilfield Review 21, no. 2 Impact on Mapping Sands at Reservoir Level,” paper 11. For more on the Gipsdalen Group: Larssen GB, (Summer 2009): 14–29. We 11 05, presented at the 75th European Association of Elvebakk G, Henriksen LB, Kristensen S-E, Nilsson I, 13. For more on the Snadd Formation: Klausen TG, Geoscientists and Engineers Conference and Exhibition, Samuelsberg TJ, Svånå TA, Stemmerik L and Worsley D: Ryseth AE, Helland-Hansen W, Gawthorpe R and London, June 10–13, 2013. Upper Paleozoic Lithostratigraphy of the Southern Laursen I: “Regional Development and Sequence Østmo S, McFadzean P, Silcock S, Spjuth C, Sundvor E, Norwegian Barents Sea. Stavanger, Norway: Norwegian Stratigraphy of the Middle to Late Triassic Snadd Letki LP and Clark D: “Improved Reservoir Petroleum Directorate (2002). Formation, Norwegian Barents Sea,” Marine and Characterisation by Multisensor Towed Streamer “Barents Sea—Carboniferous to Permian Plays,” Petroleum Geology 62 (April 2015): 102–122. Seismic Data at the Mariner Field,” paper We P03 12, Norwegian Petroleum Directorate, http://www.npd.no/ 14. Houbiers M, Wiarda E, Mispel J, Nikolenko D, Vigh D, presented at the 76th European Association of en/Topics/Geology/Geological-plays/Barents-Sea/ Knudsen B-E, Thompson M and Hill D: “3D Full- Geoscientists and Engineers Conference and Exhibition, Carboniferous-to-Permian/ (accessed August 29, 2015). Waveform Inversion at Mariner—A Shallow North Sea Amsterdam, June 16–19, 2014. Reservoir,” Expanded Abstracts, 82nd SEG Annual

10 Oilfield Review 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). fluvial and floodplain geology and reveal that results as seismic waves travel through the Earth the channel system is associated with flood- and encounter changing properties in the plain development (Figure 10).13 The data subsurface geology. The starting point for FWI is reveal a variety of fluvial features, including an approximate model of velocities. Geophysicists point-bar systems, clustered channel fill com- use this velocity model to simulate the recorded Variance plexes and ribbon-channel sandstone bodies; wavefield. They then subtract the simulated Low High the ribbon channels were at depths greater wavefield from the observed wavefield to obtain than 1,000 m [3,280 ft] and estimated to be less the residual wavefield. The residual wavefield Figure 10. Floodplain channels. This seismic time slice at 1,100 ms is at the depth of the upper than 100 m wide. is then backward propagated—extrapolated Triassic Snadd Formation. The time slice shows The East Loppa Ridge survey demonstrates downward in space or backward in time—through the variance in reflectivity. The dark linear and the imaging power of acquiring true 3D, broad- the velocity model to obtain a dataset of velocity curvilinear features are faults. The lighter gray, band seismic data. High spatial resolution in all gradients. These gradients inform where to sinuous and interweaving features are networks of fluvial channels crisscrossing a floodplain. directions facilitates and improves imaging of increase or decrease velocities but not by how complicated 3D geology such as fault networks, much. To calculate a velocity model update, the anastomosing fluvial channel complexes and car- gradients are multiplied by a step length, which Formation, composed predominantly of shale. The bonate platform deposition and karstification. scales the gradients. The velocity updates are deeper reservoir in the Maureen sandstone The increased detail offered by broadband images added to the current velocity model to create a member contains heavy oil of 14.2 API gravity and promotes improved understanding of petroleum new velocity model, and the process is repeated. is at the base of the Early Paleocene Våle system geology and better discrimination of lithol- The iterations continue until the residual Formation at depths of 1,400 to 1,500 m [4,590 to ogies and their rock properties. wavefield is acceptably small, meaning that the 4,920 ft] below sea level. modeled wavefield closely approximates the The Mariner field presents various challenges Full Waveform Inversion observed wavefield. The final model of seismic for seismic imaging.14 The shallow overburden Geophysicists use full waveform inversion (FWI) velocities can be used as an input to migration to above the reservoirs contains channel sands that for calculating horizontal and vertical seismic produce an image that better represents subsur- have higher seismic velocities than those of sur- wave velocities of geology from the surface to tar- face rock characteristics or may be used directly to rounding geologic units. These sands can be gets of interest. The result is a velocity image in interpret rock and fluid properties. mapped easily, but their presence causes distor- depth that reveals the sought-after structural This technique was used in Mariner field, dis- tions in the images of the reservoir zones beneath and depositional information. covered in 1981 and located about 150 km [93 mi] them. For example, shallow, high-velocity chan- Traditional migration produces an image of the east of the Shetland Islands on the UK Continental nel sands cause pull-ups of, or apparent struc- subsurface by attempting to reposition, or migrate, Shelf in the North Sea. The field is under develop- tural high spots in, underlying reflectors. The seismic data reflection points to their correct loca- ment by operator Statoil UK Limited with partners Heimdal reservoir sands consist of complex chan- tions in 3D space. A velocity model is almost JX Nippon Exploration and Production (UK) nel sands as well as sand injectites, or sand intru- always an input to migration; and a refined veloc- Limited and Dyas UK Limited. The field consists of sions; these sands are difficult to image because ity model may be a byproduct of migration. two reservoirs. The shallow reservoir contains of their low impedance contrast with the shales Unlike conventional migration, FWI is a heavy oil of 12.1 API gravity and is about 1,200 m that host them. The Maureen sandstone contains method for building a velocity model by attempting [3,940 ft] below sea level in sands of the Heimdal small-scale faults and calcite layers that are to match the complete recorded wavefield that member of the Middle to Late Paleocene Lista important for developing production from the

May 2016 11 3 km [1.9 mi] long, spaced 75 m apart and towed The geophysicists wanted to know whether the Build initial velocity model at a constant depth of 18 m. After acquisition, the results of FWI would isolate the velocities in data were preconditioned and then simultane- shallow channel sands within the overburden. As a ously interpolated and deghosted using the GMP test, one of the known channels delineated from Control test for FWI: Insert known channel into velocity model method. The upgoing pressure wavefield was legacy 3D seismic data was inserted into the initial then output on a 6.25‑m by 6.25­m grid for subse- velocity model and given a higher velocity than its quent processing and imaging.15 host units. If successful, the FWI method would One iteration of common image point (CIP) Initial inspection of the dataset showed it to sharpen the velocities within this control channel tomography to smooth velocity model be richer in high frequencies than in two conven- but also pick out other channels in the area. tional 3D seismic datasets and richer in low fre- To compensate for velocity imprecisions quencies than in an earlier ocean bottom cable introduced by interpolation, the geophysicists Low-frequency FWI iterations using frequency band, 1.5 to 7 Hz (OBC) survey. Both qualities are important for applied one iteration of common image point peak frequency, 2.5 Hz resolving subsurface geology and velocities (CIP) tomography to the interpolated velocity through inversion of seismic data. High frequen- model. Common image point tomography is an cies enable resolution of relative velocities iterative method of inverting for seismic veloci- Intermediate-frequency FWI iterations using between small stratigraphic and structural ties using seismic reflections. During an itera- frequency band, 1.5 to 13 Hz details. Low frequencies facilitate determination tion, the amount of residual moveout—depth peak frequency, 5 Hz of absolute velocities, which are calibrated variation—along reflections in prestack depth- against borehole data. migrated (PSDM) CIP gathers is used to deter- The data underwent fast-track processing, mine adjustments in the velocity model to bring One iteration of CIP tomography to refine velocities in deepest intervals of the model using prestack time migration, which demon- the subsequent version of the PSDM image into strated the Heimdal member sands could be better focus.20 After one iteration of CIP tomogra- imaged more reliably using the broadband data phy, the velocity model was smoothed and ready Multiparameter FWI to refine than the earlier data.16 The operator’s geoscien- for input to the FWI process. anisotropic parameters tists were able to establish the relationship Next, the geophysicists started the FWI pro- between seismic reflectors and geologic horizons cess, which, beginning with the initial earth with improved confidence.17 Encouraged by these model of velocities, iteratively models the Final high-frequency FWI iterations to enhance the resolution of velocities results, WesternGeco geophysicists applied FWI to observed seismic wavefield and adjusts the veloc- the broadband dataset.18 ities in the earth model until there is an accept- Figure 11. Workflow for full waveform inversion. The starting point for FWI is a velocity model able match between the modeled wavefield and (Figure 11). The geophysicists began with a simple the recorded wavefield.21 The observed wavefield model, using seismic velocities interpreted from was the upgoing P-wave wavefield that had been sandstone but are below the detection sonic logs from wells in the area of the Mariner isolated at an early stage of processing from the capabilities of traditional seismic techniques. field, which were then interpolated laterally broadband dataset. The criterion for conver- The imaging challenges presented by the between the wells along layers bounded by known gence to an acceptable match between synthetic reservoirs may be mitigated by full waveform geologic horizons. Based on previous processing and observed wavefields is to minimize a misfit processing techniques that enable removal of the studies, the overburden formations were assumed function that quantifies the difference between distortions caused by the high-velocity channel to be anisotropic; the P-wave anisotropy parame- the modeled and measured data. To ensure that sands in the shallow overburden. ters epsilon (ε) and delta (δ) were initially defined the FWI process converges on the global, or true, In 2012, the operator acquired a broadband as linearly increasing from the seafloor to the base minimum rather than a localized minimum, the seismic survey at the Mariner field using the Cretaceous unconformity but were subsequently geophysicists conduct FWI in stages. First, they WesternGeco IsoMetrix technology. The survey updated using a multiparameter inversion step in find an acceptable fit of the low-frequency data were acquired using eight streamers, each the FWI workflow.19 wavefield. They then add and fit to successively

15. Özbek et al, reference 5. Using Multimeasurement Towed Streamer Data: North of the difference between the horizontal and vertical 16. Migration is a seismic processing step in which Sea Case Study,” Expanded Abstracts, 85th SEG Annual P-wave velocities squared divided by the vertical reflections in seismic data are moved to their correct International Meeting and Exhibition, New Orleans P-wave velocity squared. Delta is describes near- locations. Time migration locates reflections in two-way (October 18–23, 2015): 1049–1053. vertical P-wave velocity anisotropy and the difference traveltime—from the surface to the reflector and back 19. The base Cretaceous uniformity is the term applied to a between the vertical and small-offset moveout velocity as measured along the image ray. Depth migration strong seismic reflection surface that is mappable over of P-waves. For more on seismic anisotropy parameters: locates reflectors in depth. Mathematically, migration is much of the continental shelf in the North Sea. The Thomsen L: “Weak Elastic Anisotropy,” Geophysics 51, performed by various solutions to the wave equation reflector is an unconformity that is located close to the no. 10 (October 1986): 1954–1966. that describe the passage of seismic waves through bottom of Cretaceous-age rocks and separates 20. For more on CIP tomography: Woodward M, Nichols D, rock. Kirchhoff migration is a ray-based approximation sediments deposited before rifting of the North Sea from Zdraveva O, Whitfield P and Johns T: “A Decade of founded on the integral solution to the wave equation sediments deposited after rifting. Tomography,” Geophysics 73, no. 5 (September– derived by 19th-century German physicist Anisotropy is the variation of a physical property, October 2008): VE5–VE11. Gustav Kirchhoff. such as P- or S-wave velocity, with the direction 21. Vigh D, Starr EW and Kapoor J: “Developing Earth For more on migration and imaging: Albertin U, Kapoor J, of its measurement. For more on elastic anisotropy: Models with Full Waveform Inversion,” The Leading Randall R, Smith M, Brown G, Soufleris C, Whitfield P, Armstrong P, Ireson D, Chmela B, Dodds K, Esmeroy C, Edge 28, no. 4 (April 2009): 432–435. Dewey F, Farnsworth J, Grubitz G and Kemme M: Miller D, Hornby B, Sayers C, Schoenberg M, Leaney S 22. For more on multiscale inversion: Bunks C, Saleck FM, “The Time for Depth Imaging,” Oilfield Review 14, no. 1 and Lynn H: “The Promise of Elastic Anisotropy,” Zaleski S and Chavent G: “Multiscale Seismic (Spring 2002): 2–15. Oilfield Review 6, no. 4 (October 1994): 36–47. Waveform Inversion,” Geophysics 60, no. 5 (September– 17. Østmo et al, reference 14. Epsilon (ε) and delta (δ) are P-wave parameters that October 1995): 1457–1473. 18. Gupta S, Cunnell C, Cooke A and Zarkhidze A: describe vertical transverse isotropy. Epsilon is the 23. Reference 16. “High-Resolution Model Building and Imaging Workflow P-wave anisotropy parameter and equal to half the ratio

12 Oilfield Review higher frequency bands until there is an dataset acquired using IsoMetrix technology can be high-frequency reverse time migration (RTM) per- acceptable fit of the full-frequency wavefield. inverted for a geologically relevant seismic velocity formed directly in the natural shot domain after This sequential FWI procedure stabilizes the model that is capable of sharpening the focus of GMP.23 The velocity model from FWI sharpened the inversion algorithm and ensures that the process seismic images. After FWI processing, the velocity image of the control channel embedded into the converges to a global minimum.22 model was input into two prestack depth migration overburden of the initial velocity model and Application of FWI to the broadband dataset algorithms: a Kirchhoff depth migration (KDM) to highlighted additional channels (Figure 12). collected at Mariner field showed that a seismic compare directly against legacy data volumes and a

Before FWI Processing

2,250 m/s Depth slice, 158 m

1,700 m/s

2,025 m/s Depth slice, 278 m

1,625 m/s

2,225 m/s Depth slice, 844 m

2,500 m/s

1,500 m/s 2,025 m/s

After FWI Processing

2,250 m/s Depth slice, 158 m

1,700 m/s

2,025 m/s Depth slice, 278 m

1,625 m/s

2,225 m/s Depth slice, 844 m

2,500 m/s

1,500 m/s 2,025 m/s

Figure 12. Comparing models before and after full waveform inversion (FWI). section resulted after using the velocities output after completion of FWI. Both seismic sections (left top and bottom) show the same geology to a The control channel is in better focus, and the velocities of other channels depth of 1,200 m [3,940 ft] below sea level. The depth sections are the result are evident. The velocities of the overburden units have become more of Kirchhoff depth migration (KDM); the sections are overlain by the velocity defined. The images on the right are depth slices at 158, 278 and 844 m model (colors) that was used as input to KDM. The top section resulted [518, 912 and 2,770 ft] below sea level. Compared with the before FWI from KDM using the initial velocity model. The control channel is in the top processing results, geologic features (yellow arrows) have become better center and was given a higher velocity than its surroundings. The bottom defined after FWI processing.

May 2016 13 Legacy KDM Model 1,500 Isolated depositional sandstones 1,500 Sandstone intrusions crosscut the in the Frigg Formation mudstones Balder and Frigg Formation mudstones 1,600 Onlap 1,600 Onlap 1,700 1,700 W Crestal intrusion fringe ing-like intrusion 1,800 1,800

1,900 1,900

wo-way traveltime, ms 2,000 wo-way traveltime, ms 2,000 T T

2,100 2,100 Heimdal sandstone 1 km Heimdal sandstone 1 km

Concept 1: isolated depositional sandstones Concept 2: sandstone intrusions crosscut mudstones

Heimdal sandstone Top of the Frigg sandstone Top of the Balder sandstone Top of the Sele Formation Top of the Heimdal sandstone KDM using FWI Model Reflection amplitude

– 0 +

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 sandstone Heimdal injectites against the background Lista shales (Figure 13).24 The IsoMetrix marine isometric seismic tech- nology and full waveform imaging are enabling and RTM using FWI Model 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 Heimdal sandstone distinctly. Measurements of the pressure wavefield and its gradients using IsoMetrix technology represent a significant development 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 24. For more on injectites: Braccini E, de Boer W, Hurst A, high-resolution reverse time migration (RTM) also using the FWI model (bottom). The progression Huuse M, Vigorito M and Templeton G: “Sand Injectites,” demonstrates improved imaging of the steep dips and signal-to-noise characteristics in the reservoir Oilfield Review 20, no. 2 (Summer 2008): 34–49. section, discriminating the Heimdal injectite, or intrusion, features (circled) from the background Lista Huuse M, Cartwright J, Hurst A and Steinsland N: shales. The inset shows conceptual sketches of how the sandstone bodies or intrusions might have “Seismic Characterization of Large-Scale Sandstone Intrusions,” in Hurst A and Cartwright J (eds): Sand become incorporated into the Lista shales above the Heimdal formation. (Inset adapted from Huuse Injectites: Implications for Hydrocarbon Exploration and et al, reference 24.) Production, AAPG Memoir 87. Tulsa: AAPG (2007): 21–35.

14 Oilfield Review Ocean Bottom Cable Survey IsoMetrix Survey 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 1.7 km 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.

Contributors

Anatoly Aseev, based in Moscow, was a Seismic worked in technical, service and marketing managerial Shruti Gupta is an Area Geophysicist for Schlumberger Interpreter for Schlumberger Multiclient seismic positions in the UK, US and Egypt. Before joining in Gatwick, England, where she provides technical projects from 2014 to 2016, with focus on the the team for IsoMetrix technology, Chris was based support for time and depth processing of marine and Norwegian Continental Shelf area. He began his career in Cairo and managed advanced imaging services, ocean bottom cable (OBC) seismic data. Shruti, who in 2006 as a geologist with Rosneft in Krasnodar, including full waveform inversion and Seismic has more than seven years of experience in the oil and Russia, and worked on exploration projects in Guided Drilling* service, across the Middle East gas industry, started her career with Schlumberger the Ciscaucasia basin. He joined Schlumberger and North Africa. He received an MBA degree from as a field geophysicist on a land and transition zone PetroTechnical Services (PTS) in 2011 and served the Rotterdam School of Management at Erasmus seismic acquisition crew in Egypt. She then worked as a geologist and then senior geologist working on University, the Netherlands. with the depth imaging group in Houston. She has an exploration projects in the Timan-Pechora, West Low Cheng Foo is Custodian of Geophysical Acquisition MSc degree in applied geology from the IIT Kharagpur, Siberia, Barents Sea and West Black Sea basins. for PETRONAS Carigali Sdn Bhd in Kuala Lumpur, West Bengal. Anatoly holds an MSc degree in petroleum geology where he is involved with new technology projects Michelle Tham is the Technical Support Manager from the North-Caucasus Federal University, Stavropol, such as broadband, multicomponent, multiazimuth for WesternGeco in the Asia Pacific region as well Russia. He is pursuing a PhD degree in regional and full azimuth seismic data acquisition. He has 35 as the Petrotechnical Expertise Discipline Career geology from Lomonosov Moscow State University. years of experience with the company. Previously, he Manager for the Schlumberger Asia region; she is Sandeep Kumar Chandola is a Custodian of was head of acquisition after serving as an acquisition based in Kuala Lumpur. She began her career with Geophysics with PETRONAS Carigali Sdn Bhd in and processing geophysicist. He has been involved in Schlumberger in Calgary and has worked in the US, Kuala Lumpur. He served with Oil and Natural Gas land, marine and transition-zone seismic acquisition Myanmar, Indonesia, Australia, Nigeria, UAE and Corporation, the Indian national oil company, for programs in various countries in Southeast Asia, the Malaysia. Before her current position, she served as a more than 20 years before joining Petronas Carigali Middle East, Suriname and Cuba. Low earned a BSc seismic data processing geophysicist, data processing in 2005. His work has supported the design of 3D (Hons) degree in physics, majoring in geophysics, from supervisor, staff geophysicist, area geophysicist, acquisition geometries and the introduction of new the University of Science Malaysia in Penang. seismic survey design and modeling manager and geophysical technologies to the company. He has a Malcolm Francis is a Schlumberger Advisor and the geophysics global discipline career manager. Michelle master’s degree in physics from Hemvati Nandan Eastern Hemisphere Exploration Services Manager holds a BS degree in geophysics from the University Bahuguna Garhwal University, Sringar, Uttarakhand, for WesternGeco in Gatwick. Before his current role, of Calgary. India, and a specialized diploma in petroleum he held technical and management positions as the Peter Watterson is the Manager of the Marine geophysics from the Indian Institute of Technology Eastern Hemisphere multiclient chief geophysicist, Geosolutions Technology Commercialization group (IIT) Roorkee, Uttarakhand. He is a member of the global manager of geology and interpretation for WesternGeco in Gatwick. His focus is on research, SEG, the European Association of Geoscientists and and senior manager E&P solutions. Earlier in his engineering and marketing of various marine seismic Engineers and the Society of Petroleum Geophysicists career, Malcolm managed the special processing acquisition and processing technologies. Pete (India), an SEG Honorary Lecturer and an adjunct and interpretation departments. He began in the began his career in the geophysics industry with lecturer at Universiti Teknologi PETRONAS, Malaysia. industry in 1980 with Western Geophysical, where he Western Geophysical in 1991 in London. He has held Sandeep has authored more than 50 publications and undertook collaborative research with Saudi Aramco. positions in seismic data processing and technology is a recipient of the National Petroleum Management He obtained a bachelor’s degree in geology from the management in the UK, Venezuela, US, Trinidad and Programme Award for Excellence from the government University of Manchester, England, and MSc and PhD Brazil and worked for several years as the regional of India. degrees in geophysics from Imperial College London. geophysicist for WesternGeco for South America. He Chris Cunnell leads Technical Sales and Marketing Malcolm is a member of the European Association received a BSc degree in physics from the University of IsoMetrix* for WesternGeco in Gatwick, England. of Geoscientists and Engineers, the SPE, SEG and of Leeds, England. Chris, who has more than 20 years of geophysics Petroleum Exploration Society of Great Britain and is experience, joined Schlumberger in 1997 and has a Fellow of the Geological Society of London. An asterisk (*) denotes a mark of Schlumberger.

May 2016 15 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 Calvez Operators producing from unconventional reser- hydraulic fractures as they advance through and Raj Malpani voir plays face many challenges. Fluid flow alter a formation. Jian Xu through unconventional reservoir rocks is lim- Engineers may employ several techniques to Houston, Texas, USA ited by matrix permeability, which is generally determine the effectiveness of hydraulic stimula- several orders of magnitude smaller than that of tion operations.2 For instance, during stimulation Jerry Stokes conventional reservoir rocks. Preexisting faults operations, microseismic (MS) monitoring and Mid-Continent Geological, Inc. and fracture networks often provide pathways tiltmeter measurements can indicate mechanical Fort Worth, Texas for the flow of hydrocarbons and play an impor- changes in the subsurface that occur over a wide 3 Michael Williams tant role in increasing reservoir drainage vol- area centered on the treatment well. Afterward, Cambridge, England umes. Hydraulic fracture stimulation treatments engineers have used radioactive and chemical can often connect the wellbore to existing natu- tracers, temperature tools and production logs to Oilfield Review 28, no. 2 (May 2016). ral fracture networks; however, effective stimu- provide complementary indications of changes in Copyright © 2016 Schlumberger. lation requires knowledge of the distribution of fluid pathways resulting from the stimulation. For help in preparation of this article, thanks to Julian Drew, Perth, Western Australia, Australia; Tony Probert and those networks. Geophysical service companies often acquire Ian Bradford, Cambridge, England; and Nancy Zakhour, Well completion engineers use geomechani- MS data, which they interpret and integrate with Callon Petroleum, Houston. cal and fracture models to plan where to initiate other measurements to provide oil and gas opera- CMM, ECLIPSE, Mangrove, MS Recon, NetMod, Petrel, ThruBit, UFM, VISAGE, VSI and VSI-40 are marks of hydraulic fractures and predict their propagation tors with an understanding of hydraulically Schlumberger. through the reservoir. These models require cali- induced fracture systems. The primary data used 1. Seismic waves convey energy by means of the particle bration and validation. Microseismic monitoring for evaluating MS events are waveform measure- motion of solid materials. 2. For more on fracture diagnostic techniques: Bennett L, has proved to be a viable means for calibrating ments acquired from a network of receivers Le Calvez J, Sarver DR, Tanner K, Birk WS, Waters G, the models and for providing empirical data placed either downhole or at the surface. Drew J, Michaud G, Primiero P, Eisner L, Jones R, Leslie D, Williams MJ, Govenlock J, Klem RC and about the effectiveness of stimulation operations. Geoscientists use these data to map the extent Tezuka K: “The Source for Hydraulic Fracture Microseismic monitoring is a technique that and evolution of MS events. These maps provide Characterization,” Oilfield Review 17, no. 4 (Winter 2005): 42–57. records and locates microseismic events—col- valuable information related to strain and stress 3. A tiltmeter measures minute rotations—changes of lectively referred to as microseismicity—which variations in the reservoir and surrounding for- inclination—of the ground in which it is embedded. are small bursts of seismic wave energy gener- mations and are used to guide stimulation deci- ated by minute rock movements in response to sions during job execution. If MS events indicate changes of the in situ stresses and rock volume undesired fracture growth or fault activation, such as those that occur during fracture stimula- operators may choose to terminate stage pump- tion operations.1 During these operations, frac- ing early, use diverter technology or skip stimula- tures are created by injecting fluid at high tion stages. pressure. These fractures propagate and are then Microseismic monitoring also provides infor- held open using a solid proppant. Mapping the mation about the nature of the physical pro- spatial and temporal pattern of these events has cesses—induced fracturing of the rock or slippage proved successful for monitoring the progress of on preexisting fractures—that occur at the

16 Oilfield Review May 2016 17 Treatment well Monitoring well 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 Microseismic event Sensors during the stimulation. To describe the magnitude and direction of the rock movements at each

Reservoir source location, geophysicists process MS wave- form recordings, account for propagation effects, determine the radiation pattern of the acoustic

Stimulated volume 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 Anchoring from Texas, USA, illustrate how MS monitoring arm has added value to stimulation operations by helping geoscientists identify fault interactions, fracture growth and stage-to-stage variability of stimulation responses.

Typical Monitoring Systems Microseismic monitoring (MSM) is the detec- Shaker Coupling tion of signals generated by small seismic—or contacts microseismic—events. Engineers began using Three y this technique during hydraulic fracturing in oil Shaker component x 5 accelerometers z z and gas operations as early as the 1980s. The Cotton Valley Consortium—a research group Isolation y studying hydraulic fracturing of the Cotton Valley x spring 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 Figure 1. Hydraulic fracture monitoring from a vertical well. Multicomponent sensors in a vertical monitoring borehole record microseismic events caused of the fracture network development during stim- by hydraulic fracturing (top). Event locations determined from data processing ulations, avoid geohazards and enhance produc- allow engineers to monitor the progress of stimulation operations. To tion.7 These early MSM operations incorporated acquire high-fidelity seismic data, the VSI versatile seismic imager (bottom) uses three-axis (x, y and z) geophone accelerometers (inset) that are arrays of three-component (3C) geophones or acoustically isolated from the tool body by isolation springs. The VSI service accelerometers deployed near reservoir depth in is mechanically coupled to the casing or formation by a hydraulically a nearby vertical monitor well (Figure 1).8 powered anchoring arm. TheOilfield acquisition Review engineer can test the coupling quality by activating an internalMAY shaker16 before operations begin. The VSI-40 40-shuttle versatile seismicMicroseismic imager allows Fig up 1to 40 sensor packages, or shuttles, to be linked together;ORMAY however, 16 MCSMC 12 shuttles 1 are typically used in hydraulic fracture monitoring operations.

18 Oilfield Review Extensive hydraulic fracturing of horizontal fidelity—accuracy for measuring signal magni- of interest to the receiver array. Geophysicists wells began after 1997 as a result of Mitchell tude and direction—because accurate waveform calibrate the velocity model using perforation Energy’s successful application of the method in polarization information is crucial for determin- shot, string shot, checkshot or VSP survey data.15 the Barnett Shale; MSM from adjacent horizontal ing the direction to each event hypocenter.14 In Analysts originally performed event detection and boreholes soon followed. The use of sensor arrays addition, seismic tools must record incoming MS localization processing using P-wave and S-wave in horizontal wells next led to the evaluation of signals with the same spectral fidelity—accu- arrival time picks and polarization from 3C MS MSM performance. racy for measuring frequency content—within waveforms.16 Today, event localization algorithms, The effectiveness of the sensor array geometry the typical signal bandwidth of 10 to 1,000 Hz such as the CMM coalescence microseismic depends on the layout of the monitoring and used in these operations. When monitored from mapping procedure, use an automatic scanning treatment wells. Monitoring from vertical wells single wells during a multistage stimulation, and grid search algorithm that correlates signal close to treatment stages results in improved some stages may be too distant from the sensors traveltimes and waveform polarizations to locate location accuracy for in-zone and out-of-zone for reliable event detection or characterization. hypocenters.17 Multicomponent waveforms are microseismicity. Monitoring from nearby horizon- Sensor geometries along a single linear array are processed to assess how well the observed tim- tal wells often provides coverage along the length insufficient to determine the source mecha- ing and polarization of arrival phases across of a stimulated lateral well; similar coverage may nisms—size, direction, orientation and duration the receiver array match the modeled values be unavailable from surface arrays or vertical of 3D rock movements—associated with MS associated with potential hypocenter locations monitor wells. The recording geometry may events; thus, microseismic engineers seek to in the volume of interest. Arrival time picking require cost trade-offs between monitoring the record seismic waveforms from multiple observa- can also be performed automatically and then entire treatment well and detecting MS events tion points and azimuths. refined manually. that may occur outside the target interval.9 Valid interpretation of MS data also requires Analysts interpret MS locations to show Microseismic events that are detected outside careful signal analysis. The processing of MS data induced fracture extent—length, height and the targeted interval can indicate unintended is preceded by the construction and calibration of azimuth. However, stimulations in tight and consequences of the stimulation program such a model of seismic P-wave and S-wave velocities unconventional reservoirs often produce com- as breaching the reservoir seal or activating exist- extending from the planned stimulated volume plex, nonplanar hydraulic fracture geometries. ing faults.10 4. The radiation pattern is a description in 3D space of 12. To predict sensor network performance, Schlumberger Knowledge of MS measurement accuracy is the amplitude and sense of initial motion of P and S engineers used the NetMod microseismic survey design crucial for understanding the validity of MS data wavefronts as they propagate away from the initiation and evaluation software. For more on microseismic 11 position of a microseismic event. For more on seismic survey design: Raymer DG and Leslie HD: “Microseismic interpretations. Survey designers have devel- sources and their radiation patterns: Lay T and Network Design—Estimating Event Detection,” oped modeling software that predicts the mini- Wallace TC: Modern Global Seismology. San Diego, presented at the 73rd EAGE Conference and Exhibition, California, USA: Academic Press, 1995. Vienna, Austria, May 23–26, 2011. mum detectable event magnitude with respect to 5. For more on the context for microseismic monitoring: 13. The hypocenter, or focus, is the point within the Earth the distance of the monitoring array from source Maxwell SC, Rutledge J, Jones R and Fehler M: at which rupture starts during an earthquake or “Petroleum Reservoir Characterization Using Downhole microseismic event. The point directly above it on locations. The software also outputs estimates of Microseismic Monitoring,” Geophysics 75, no. 5 the Earth’s surface is the epicenter. the associated uncertainty for locating and char- (September–October 2010): 75A129–75A137. 14. For more on vector fidelity: Berg EW, Rykkelid, Woje G acterizing MS events.12 Accuracy of the estimated 6. The Cotton Valley formation is a Cretaceous-age tight and Svendsen Ø: “Vector Fidelity in Ocean Bottom sandstone that stretches from Texas to northern Florida, Seismic Systems,” paper OTC 14114, presented at event hypocenters—3D locations using easting USA. The main play produces mostly natural gas and is the Offshore Technology Conference, Houston, and northing geographic Cartesian coordinates located in north Louisiana and northeast Texas. For more May 6–9, 2002. on the Cotton Valley Consortium project: Rutledge JT, 15. In a checkshot survey, seismic specialists measure the along with the depth of event initiation points— Phillips WS and Mayerhofer MJ: “Faulting Induced by traveltime of seismic waves, usually P-waves, from the is affected by the monitoring geometry and the Forced Fluid Injection and Fluid Flow Forced by Faulting: surface to known receiver depths. A vertical seismic An Interpretation of Hydraulic-Fracture Microseismicity, profile (VSP) is a more extensive survey in which accuracy of the velocity model that is used to Carthage Cotton Valley Gas Field, Texas,” Bulletin of geophones are placed at regular, closely spaced transform waveform arrival times at recording the Seismological Society of America 94, no. 5 positions in the borehole. Both surveys use a seismic (October 2004): 1817–1830. source positioned on the surface. Perforation shots and instruments to distances of the instruments from 7. For more on the use of microseismic monitoring in the explosive string shots serve as seismic sources in the the events. Barnett Shale: Maxwell S: “Microseismic: Growth treatment well, and traveltimes are measured downhole Born from Success,” The Leading Edge 29, no. 3 or at the surface. In all cases, the source and receiver The precision of hypocenter estimates depends (March 2010): 338–343. locations are known and, from the observed traveltimes, on the geophone array geometry and data errors, 8. Seismic data acquired from three-component (3C) velocity may be calculated. which influence the determination of event arrival geophones use three orthogonally oriented geophones 16. The P-waves used for seismic processing are elastic 13 or accelerometers. The early Schlumberger body waves, or sound waves, in which particles time and the direction of arrivals at the receivers. microseismic acquisition system typically included a oscillate in the direction the wave propagates. The During stimulation operations, extraneous, high- VSI tool with eight 3C geophones. S-waves are elastic waves in which particles oscillate 9. For more on the accuracy of hypocenter estimates: perpendicular to the direction in which the wave amplitude noise sources are numerous. As a conse- Maxwell S and Le Calvez J: “Horizontal vs. Vertical propagates. quence, the low signal-to-noise ratio (S/N) is one Borehole-Based Microseismic Monitoring: Which 17. Drew J, Bennett L, Le Calvez J and Neilson K: is Better?,” paper SPE 131780, presented at the “Challenges in Acoustic Emission Detection and of the greatest challenges in the acquisition and SPE Unconventional Gas Conference, Pittsburgh, Analysis for Hydraulic Fracture Monitoring,” paper processing of MS data. Pennsylvania, USA, February 23–25, 2010. presented at the 17th International Acoustic Emission Early MSM from a single monitoring well pro- 10. Induced seismicity refers to earthquakes that are Symposium, Kyoto, Japan, November 9–12, 2004. attributable to human activities, which may alter the Drew J, Leslie D, Armstrong P and Michaud G: vided valuable information, although it had local stresses and strains in the Earth’s crust and cause “Automated Microseismic Event Detection and Location shortcomings. Microseismic monitoring from rock movements that generate earthquakes. by Continuous Spatial Mapping,” paper SPE 95513, 11. Maxwell, reference 7. presented at the SPE Annual Technical Conference single wellbores imposes the requirement that all and Exhibition, Dallas, October 9–12, 2005. multicomponent sensors have the same vector

May 2016 19 7,200 17:00

18:00

19:00 Top of Barnett Shale

20:00 7,600 Time, hr:min 21:00

22:00 Z-axis Depth, ft

8,000

Bottom of Barnett Shale 8,400 7,120,000 7,130,000 South Y-axis S–N, ft North

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.)

Consequently, geophysicists compute the effec- Monitoring from surface and near-surface Monitoring from multiple wells provides tive stimulated volume (ESV) as a measure of MS positions offers a potentially larger field of view observations of the source position from multiple activity. The size, shape and extent of the ESV is than that from monitoring wells alone, and it directions and enables more-complete source based on a distribution of event locations and eliminates the need for providing dedicated deep characterization compared with that from sin- their uncertainties (Figure 2).18 The ESV provides monitoring wells.22 Surface monitoring enables gle-well monitoring. Downhole monitoring information on the complexity of the hydraulic long treatment laterals to be monitored along requires correct velocity models to reduce event fracture network. A long, narrow ESV is probably their entire lengths. However, because the S/N is localization uncertainty, and it requires precise dominated by a single planar through-going frac- often low, locating and characterizing MS events well deviation surveys to determine exact ture, whereas a short, wide ESV probably consists using data recorded at the surface may be diffi- receiver positions. The models must also contain of a complex, multibranching fracture network. cult. To overcome low S/N and detection uncer- accurate values for Qp and Qs, the quality factors tainty, survey designers use receiver arrays related to P-wave and S-wave attenuation during Taking the Broad View containing many hundreds to thousands of sen- propagation. These factors are used to deter- Engineers may conduct MSM from a single well, sors. Data from multiple points can be processed mine wave amplitudes at the MS hypocenters multiple wells, grids of shallow wells, surface to reduce noise and accentuate the true signal. and reduce uncertainty in the inversion for the arrays or networks of surface sensor patches. To Aided by recent improvements in signal process- source mechanism.23 meet acquisition goals, they may also combine ing, geophysicists can use these monitoring arrays Modern signal processors use nonlinear a variety of designs (Figure 3).19 Typically, ana- to map microseismic events more completely over mathematical methods for the detection and Oilfield Review lysts employ numerical simulation techniques extended MAYstimulations 16 than is possible from an localization of MS events. In combination with that account for signal frequency content and array placedMicroseismic in a single monitoring Fig 2 well. CMM processing, these mathematical methods attenuation and that make use of source, geol- When ORMAYnearby observation 16 MCSMC wells 2 are available, have the potential to detect and locate weak MS ogy and noise models. They may use statistical downhole monitoring offers proximity to treat- events automatically without prior knowledge of analysis to predict the number of detectable ment well stages and ensures higher S/N than that the source mechanism and its radiation pattern.24 events for given monitoring geometries. Analysts offered by surface monitoring. Broad bandwidth Analysts have also extended event localization also recognize the importance of accounting signals recorded by downhole arrays often retain algorithms to use the full MS waveforms. Early for anisotropy in the velocity models. Seismic more high-frequency content than do surface event localization methods used traveltimes and velocity and attenuation tomography based on arrays. This high-frequency content is useful for polarizations of the P-wave and S-wave direct crosswell surveys can be used to constrain these MS event characterization. Recording P-wave and arrivals only. However, geoscientists can use models.20 During perforating, data acquired S-wave arrivals downhole using 3C sensors also waveform synthetics to model the source time from surface sensors can provide calibrated improves localization accuracy compared with functions, the principal features of the waveform P-wave traveltimes.21 that from surface recordings of P-waves alone. time series recorded by each instrument in the

20 Oilfield Review 6

7

4 3

5 1

2

Figure 3. Sensor network deployment options. Sensors for MS monitoring event compression (red ellipsoids) and dilation (blue ellipsoids) radiation of hydraulic fracturing may be deployed in vertical (1), horizontal (2) or patterns (5) and estimate source mechanisms. Schlumberger engineers deviated monitoring boreholes. Survey engineers may use a grid of shallow use the MS Recon high-fidelity microseismic surface acquisition system wells (3) containing arrays of multicomponent sensors. On the surface, they to acquire MS data at the surface. The system incorporates proprietary may deploy single component or multicomponent geophones in 2D patches geophone accelerometers (6) (inset), ultralow-noise electronics and or in extensive linear arrays (4). Sensor networks that record MS waveform a nodal-based wireless acquisition technology (7). (Adapted from data over a broad area provide data that can be used to characterize MS Le Calvez et al, reference 19.) sensor arrays. Analysts then extract arrival times 18. Effective stimulated volume (ESV), also referred to as 23. For more on uncertainty in microseismic monitoring: of direct, refracted and reflected P-wave and stimulated reservoir volume, is an estimate of the total Eisner L, Thornton M and Griffin J: “Challenges for rock volume affected by the hydraulic fracture Microseismic Monitoring,” Expanded Abstracts, S-wave arrivals. An extension of the CMM method stimulation. 81st SEG Annual International Meeting and Exhibition, uses these additional arrivals to identify their 19. For more on survey design: Le Calvez J, Underhill B, San Antonio, Texas, USA (September 18–23, 2011): 1519–1523. 25 Raymer D and Guerra K: “Designing Microseismic energy for event detection and characterization. Surface, Grid, Shallow and Downhole Surveys,” 24. For more on nonlinear processing methods: Özbek A, Expanded Abstracts, 85th SEG Annual International Probert T, Raymer D and Drew J: “Nonlinear Processing Meeting and Exposition, New Orleans (October 18–23, Methods for Detection and Location of Microseismic Tracking Microseismicity to the Surface 2015): 2645–2649. Events,” paper Tu 06 06, presented at the 75th EAGE In 2011, Schlumberger and an independent 20. For more on the use of crosswell surveys: Le Calvez J, Conference and Exhibition, London, June 10–13, 2013. operator acquired a comprehensive MS dataset Marion B, Hogarth L, Kolb C, Hanson-Hedgecock S, 25. For more on full MS waveform processing: Williams MJ, Puckett M and Bryans B: “Integration of Multi-Scale, Le Calvez JH and Gendrin A: “Using Surface and while monitoring hydraulic fracturing operations Multi-Domain Datasets to Enhance Microseismic Data Downhole Data to Drive Developments in Event Detection in the Fayetteville Shale in Arkansas. Completion Processing and Evaluation,” paper BG08, presented at Algorithms,” Extended Abstracts, 76th EAGE Conference the 3rd EAGEOilfield Workshop Review on Borehole Geophysics, and Exhibition, Amsterdam, June 16–19, 2014. engineers stimulated two horizontal wells using Athens, AprilMAY 19–22, 16 2015. 26. Zipper fracture, or “simul-frac,” is a technique in which a zipper fracture method, in which hydraulic 21. For more on the use of perforation shots to build velocity two or more parallel wells are drilled, perforated and models: ProbertMicroseismic T, Raymer D and Fig Bradford 3 I: “Comparing stimulated via an alternating sequence of stages. This fracturing is conducted sequentially in side-by- Near-SurfaceORMAY and Deep-Well 16 MCSMC Microseismic 3 Data stimulation method results in a high-density network of side wells.26 Concurrently, MS survey engineers and Methods for Hydraulic Fracture Monitoring,” fractures between the wells that increases production paper PS07, presented at the 4th EAGE Passive Seismic in both wells. conducted a test to assess and quantify the event Workshop, Amsterdam, March 17–20, 2013. For more on advances in fracturing technology: detection capabilities, accuracy and resolution 22. For more on surface microseismic monitoring: Rafiee M, Soliman MY and Pirayesh E: “Hydraulic of surface, near-surface and downhole acquisi- Duncan PM and Eisner L: “Reservoir Characterization Fracturing Design and Optimization: A Modification to Using Surface Microseismic Monitoring,” Geophysics 75, Zipper Frac,” paper SPE 159786, presented at the tion systems. Sixteen stimulation stages were no. 5 (September–October 2010): 75A139–75A146. SPE Eastern Regional Meeting, Lexington, Kentucky, USA, October 3–5, 2012.

May 2016 21 Line 3

1,000 ft

Line 4 Line 2 Line 7

Downhole array

Well M Well 1H

Well 3 Well 4 Line 5 Well 5 Line 6 Line 1

Well 2 Well 1

N Shallow well Well 2H 2D patch

Treatment well

1,000 ft Well M

Depth, ft Well 1H Well 2H N

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.) monitored across a reservoir interval at 3,600 ft 8,000 ft [915, 1,520 and 2,440 m] from the treat- ple directions but cannot cover the same [1,100 m] TVD. ment wellhead. Time synchronization between distances as can linear arrays.29 Sensor arrays in Survey engineers deployed a wide-aperture all recording systems ensured that the same MS shallow wells are less sensitive to noise propagat- borehole seismic array that extended from the events could be identified on all monitoring sys- ing along the surface, but signal processing that reservoir to the surface (Figure 4). This array tems (Figure 5).27 discriminates against noise is hampered by the acquired an MS dataset at the reservoir level as Data from this comprehensive test allowed small number of sensors available in these well as data that revealed how signals propagated analysts to compare the effectiveness of near- arrays.30 Surface and near-surface array designs and how noise levels varied between the reser- surface and Oilfielddownhole Review hydraulic fracture moni- may be adapted to known noise conditions but MAY 16 voir and the surface. Engineers acquired addi- toring. AnalystsMicroseismic observed that Fig surface4 array line are constrained by land access, environmental tional MS data from a deep horizontal well and segments canORMAY mitigate 16 surface-wave MCSMC 4 noise from a effects and cost concerns. from five shallow vertical wells, each containing a known source, such as treatment wellhead The results of the Fayetteville survey showed seismic array. They also recorded MS data using pumps, but are less effective against distributed that a downhole array could detect MS events an extensive surface seismic array that consisted or moving sources of noise, which may be domi- from nearby stage treatments better than from of five radial lines fanning and emanating from nant in areas covered by the array.28 Surface other array geometries. But the downhole array the treatment wellhead, two parallel lines that patches—2D arrays of closely spaced sensors— suffered reduced sensitivity and increased loca- crossed the radial lines and three 2D surface can effectively remove noise coming from multi- tion uncertainty for distant stage treatments and patches that were located about 3,000, 5,000 and events. Surface and near-surface monitoring,

22 Oilfield Review MSM surface lines S N

MSM wells

Vertical array

Horizontal array

Vertical Seismic Array MSM Wells

Horizontal Seismic Array MSM Stacked Data

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.)

although less sensitive to recording deep signals 27. For more on the Fayetteville Shale MS test: Maxwell SC, 29. For more on signal processing approaches applied to than downhole monitoring, offers more uniform Raymer D, Williams M and Primiero P: “Tracking surface patch data: Petrochilos N and Drew J: “Noise MicroseismicOilfield Signals Review from the Reservoir to Surface,” Reduction on Microseismic Data Acquired Using a sensitivity over a wider area. Surface patches are The LeadingMAY Edge 31,16 no. 11 (November 2012): 1300–1308. Patch Monitoring Configuration: A Fayetteville Formation easily deployed over wide areas and may ulti- Peyret O, DrewMicroseismic J, Mack M, Brook Fig 5K, Maxwell S and Example,” Expanded Abstracts, SEG 84th Annual Cipolla C: “Subsurface to Surface Microseismic International Meeting and Exposition, Denver mately become the preferred surface monitoring Monitoring ORMAYfor Hydraulic 16 Fracturing,” MCSMC paper5 SPE 159670, (October 26–31, 2014): 2314–2318. configuration. However, their successful use will presented at the SPE Annual Technical Conference and 30. The amplitude of seismic surface waves decreases as Exhibition, San Antonio, Texas, October 8–10, 2012. its position away from the surface increases. For more require the recording of sufficient signal and the 28. For more on surface array performance: Schilke S, on noise sources in the Fayetteville Shale test: Drew J, effective application of noise attenuation meth- Probert T, Bradford I, Özbek A and Robertsson JOA: Primiero P, Brook K, Raymer D, Probert T, Kim A and Leslie D: “Microseismic Monitoring Field Test Using ods. The lessons learned in the test provide the “Use of Surface Seismic Patches for Hydraulic Fracture Monitoring,” paper We E103 04, presented at the Surface, Shallow Grid and Downhole Arrays,” paper planners of future MSM systems with a clearer 76th EAGE Conference and Exhibition, Amsterdam, SEG 2012 0910, presented at the 82nd SEG Annual June 16–19, 2014. International Meeting and Exposition, Las Vegas, understanding of the trade-offs involved when Nevada, USA, November 4–9, 2012. specifying acquisition equipment layouts.

May 2016 23 estimates of the event density and the actual the active volume of microseismicity may be an

Mo μ D A. extent of the ESV. overestimation of the hydraulically connected vol-

Es = 0.5 (Δσs / μ) Mo. When determining ESV, geoscientists must ume. To reduce this uncertainty, some analysts take into account the spatial density of intercon- consider how many events occur in the neighbor- Mw 2/3 (log Mo) – 6.06. nected fractures within the stimulated volume hood of each event location when computing ESV. and their surface area in contact with the For MSM that has limited array coverage, distant Figure 6. Seismic moment, energy and magnitude reservoir. The set of MS event locations—the low-amplitude events may not be detected. This equations. The seismic moment, Mo, of an earthquake source is defined as the product microseismic cloud—may include stress-induced phenomenon, referred to as monitoring bias, of the shear modulus (µ) of the host rock that events in nonhydraulically connected areas. Thus, may also reduce the computed ESV. 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 18 to the seismic moment. Seismologists have 16 also related seismic moment to the seismic 14 energy (Es), which is radiated when a fault slips, 12 resulting in a change in the static shear stress R2 = 0.80 (Δσs) along the fault. The moment magnitude 10 (M w) is computed from Mo and is a logarithmic 8 measure of the energy released during a seismic Production, % 6 event. In this equation, Mo is expressed in units of N.m. When expressed in units of dyne.cm, 4 10.73 is used instead of 6.06; for units of lbf.ft, the 2 constant is 5.97. 0 400 500 600 700 800 900 1,000 ESV, 1,000 ft3 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 879,000 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 878,000 intensity is related to the seismic moment, MO, Production which can be determined by measuring the data amplitudes of seismic waves generated during the event (FigureOilfield 6). 32Review Y-axis, ft In 1977, JapaneseMAY 16 seismologist Hiroo Kanamori 877,000 used the relationshipMicroseismic between Fig 6seismic moment ORMAY 16 MCSMC 6 Monitoring well 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 876,000 is routinely used to characterize the size of MS events. Individual source dimensions such as incremental fracture surface areas and lengths 1,538,000 1,539,000 1,540,000 1,541,000 can be estimated for MS events from their X-axis, ft waveform spectra and source models. Figure 7. Eagle Ford Shale effective stimulation. The map view (bottom) The distribution of MS source locations pro- shows MS events (colored dots), effective stimulated volume (ESVs, colored vides an indication of the rock volume affected by opaque envelopes) and production log results (dashed red lines) for the the hydraulic stimulation. Reservoir engineers 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 initially related, with some success, well produc- events for each perforated interval. The length of the bisecting red lines from tivity to MS activity using the ESV as a measure of the production data are related to the contribution of individual perforation the volumetric extent of reservoir stimulation clusters to the total flow of hydrocarbons. Engineers observed a definite (Figure 7).33 Pinpointing the location of MS correlation (top) between production contribution from individual perforated intervals (red circles) and the ESV derived from the hydraulic fracture model events, however, is sometimes insufficient for analysis. For the plotted data, R2 is a linear regression measurement related accurately predicting reservoir performance. to the quality of curve fitting. A value of 0.80 indicates a good fit. (Adapted This insufficiency may result from inaccurate from Inamdar et al, reference 33.) Oilfield Review MAY 16 Microseismic Fig 7 ORMAY 16 MCSMC 7 24 Oilfield Review 2.0 60 60 8,400 1.3 55 55 1.8 Pressure 8,200 1.2 50 50 1.6 8,000 1.1 45 45 7,800 1.0 1.4 6 40 40 0.9 7,600 1.2 35 Pump rate 35 0.8 7,400

1.0 30 30 0.7 7,200 Proppant concentration 0.6 25 25

0.8 Pump rate, bbl/min 7,000 reatment pressure, psi T Cumulative moment 0.5 20 20 6,800 Cumulative moment, N.m × 10 0.6 0.4 Event rate, accepted events per 2 min

Proppant concentration measured, lbm/galUS 15 6,600 15 0.3 0.4 Event rate 10 6,400 10 0.2 0.2 5 6,200 5 0.1

0 0 6,000 0 0 1 am 2 am 3 am 4 am 5 am Time of day 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.)

Microseismic events are produced when rapid Engineers may compare time series of cumu- into fracture behavior during stimulations; such deformation occurs within the reservoir or sur- lative moment and stimulation treatment data to insights can be used to calibrate complex hydrau- rounding formations in response to stress better understand the stimulation process lic fracture models. changes arising from increased pressure during (Figure 8). An increase in cumulative moment Borrowing concepts from earthquake seis- fracture stimulation operations. The deformation over time indicates progressive deformation. mology, analysts use statistical measures such as consists of slip of unknown length along failure Maps of final values of cumulative moment indi- b-values and D-values to further describe groups planes of unknown area and orientation.34 cate the spatial distribution of seismic deforma- of detected MS events.37 The relative frequency of Analysts estimate the seismic moment of indi- tion observed during the stimulation. Because occurrence of earthquakes over a range of magni- vidual events using the amplitudes and frequency large, detectable events contribute significantly tudes is described by b-values. Many more small content of received seismic waveforms.35 more moment than numerous, small, undetect- magnitude events tend to occur than do large Geophysicists can use the seismic moment to able events, cumulative moment provides a mea- ones, and the b-value quantifies this tendency. enhance the interpretation of MS data.36 By sum- sure of stimulation response that is less sensitive The statistics of the distances separating earth- ming these moment values over time for all to monitoring bias than is an ESV based on event quake hypocenters is described by D-values. events within distinct spatial volumes or grid locations alone. Using 3D mapping of seismic Populations of events occurring on the same frac- cells, analysts obtain the cumulative moment as a moment or cumulative moment provides insight ture and fault planes tend to present characteris- function of time and space. Oilfield Review tic distributions of spatial separation. MAY 16 31. For more on source parameters: Shearer PM: the SPE Tight Gas Completions Conference, San Antonio, 36. Analysts use source models, formation material Introduction to Seismology, 2nd ed. Cambridge, England: Texas, NovemberMicroseismic 2–3, 2010. Fig 8 properties and measured frequency spectra to estimate Cambridge University Press, 2009. 34. Downie R,ORMAY Xu J, Grant 16 D, MCSMC Malpani R 8and Viswanathan A: the fracture surface area of individual events. Slip 32. Seismic moments range from 105 N.m [105 lbf.ft] in the “Utilization of Microseismic Event Source Parameters lengths can then be inferred from seismic moment case of the smallest detectable microearthquakes to for the Calibration of Complex Hydraulic Fracture estimates. 1023 N.m [1023 lbf.ft] in the case of great earthquakes. Models,” paper SPE 163873, presented at the 37. For more on b-values and D-values: Grob M and 33. For more on ESV versus reservoir production: SPE Hydraulic Fracturing Technology Conference, van der Baan M: “Inferring In-Situ Stress Changes Inamdar A, Malpani R, Atwood K, Brook K, Erwemi A, The Woodlands, Texas, February 4–6, 2013. by Statistical Analysis of Microseismic Event Ogundare T and Purcell D: “Evaluation of Stimulation 35. Seismic sources are formally treated as “displacement Characteristics,” The Leading Edge 30, no. 11 Techniques Using Microseismic Mapping in the discontinuities” to describe the difference in motion of (November 2011): 1296–1301. Eagle Ford Shale,” paper SPE 136873, presented at material on opposing faces of fracture surfaces. This motion need not be parallel to the surfaces.

May 2016 25 1.0 rate.38 Global seismology data often show the Gutenberg-Richter relationship slope b to be about 1 for tectonic earthquakes. In log10 N = a – b × M. some geologic settings, MS interpreters use c

M b-values to distinguish failure along naturally Events occurring faults from that along hydraulically 0.1 induced fractures. Determination of b-values may also provide completion engineers with an indication of stress changes over the course of multistage stimula-

Cumulative frequency tions. During hydraulic fracturing treatments, Magnitude of completeness, b-values greater than 1 are typically observed, 0.01 whereas b at about 1 has been observed dur- ing MS episodes dominated by movement along faults.39 Scientists have observed relationships –2.8 –2.6 –2.4 –2.2 –2.0 –1.8 –1.6 –1.4 –1.2 –1.0 40 Moment magnitude between the b-value and local stress conditions. Some studies of MS data have shown variations in Figure 9. Plot of b-values. In seismology, the Gutenberg-Richter empirical b-values within regional shale formations being law relates the magnitude (M) of earthquakes to their frequency of 41 occurrence, N, where N is the number of events of magnitude M or greater, stimulated. Other studies have shown that and a and b are constants. Events observed during a stimulation stage in the b-values may be time dependent and vary as stress Barnett Shale in Denton County, Texas (red dots), show that the cumulative changes throughout the stimulation process.42 frequency is equal to N divided by 336, the total number of detected events. Seismologists use D-values to convey the spa- A statistical method has been used to estimate b, which accounts for the limited ability to detect low-magnitude events. Because data from small tial statistics of earthquake hypocenter occur- magnitude events cannot be recorded reliably, events that have magnitudes rence. Computed from event locations, D-values smaller than the magnitude of completeness, Mc (dashed line), are not used may be used to summarize interevent distance in the calculation because the S/N is too low. (Adapted from Williams et al, reference 61.) 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, 43 In seismology, the Gutenberg-Richter empiri- intercept, respectively, of the log10 N versus Mw 2 if planar and 3 if dispersed. The distribution of cal law relates the magnitude of earthquakes, M, relationship. Some evidence suggests a possible MS hypocenters has the potential to reveal the to their frequency of occurrence, N (Figure 9). In relationship between the value of a—the num- location of interconnected fracture surfaces, and microseismic studies, geoscientists have substi- ber of events at the intercept at Mw equaling 1— analysts have developed a variety of techniques tuted moment magnitude, Mw, in this relation and the pump rates used in hydraulic fracturing, to extract linear and planar features from the and have explored how hydraulic fracture param- for example, when the cumulative volume of microseismic cloud.44 eters affect the values of a and b, the slope and pumped fluid influences the microseismic event In one method, D-values were computed sepa- rately around each detected MS event location (Figure 10). Closely spaced events are more likely

–2.0 to show some random scatter due to processing effects (in which D equals approximately 3). –2.5 Events occurring on the same fracture and fault plane will tend to align (in which D equals approx- –3.0 log10 (C) Oilfield D log10 (l) Review imately 2). Interpreters identified linear and pla- nar structures in the data by selecting only events (C (l)) –3.5 MAY 16 10 Microseismic Fig 9 for which the D-value is less than or equal to 2. log –4.0 ORMAY 16 MCSMC 9 Analysts have used changes in both b-values and D-values to infer stress changes in the reservoir. –4.5 Applying additional concepts from earth- quake seismology, geophysicists measure P-wave –5.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 and S-wave amplitudes across broad receiver net-

log10 (l) works to determine radiation patterns and then Figure 10. Plot of D-values. The D-value is estimated from the slope of the invert them to estimate seismic moment tensors, line of event locations during a stimulation in the Barnett Shale in Denton which describe the orientation, magnitude and County, Texas. The cumulative distribution of interevent distances, C(l), is the slip of individual MS events.45 For MSM, geophysi- total number of MS events (purple circles) that have hypocenters separated cists use moment tensor inversion (MTI), which 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 is an advanced seismic processing technique, to corresponds to D = 1.86. This D-value indicates a near-planar distribution provide information about the mechanism of fail- of event locations for a range of intermediate distances around the event. ure at fracture sites.46 They then decompose each The closer and more distant events trend away from straight-line behavior. moment tensor into its constituents to estimate (Adapted from Williams et al, reference 61.)

26 Oilfield Review

Oilfield Review MAY 16 Microseismic Fig 10 ORMAY 16 MCSMC 10 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 Perforation Although the deformation represented by MS cluster signals constitutes a small fraction of the total deformation and fracture volume created dur- Well 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 Expansion Opening Slip 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 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, Integrating MS Data and Geomechanical opening and slip components and displayed as glyphs. For reference, the well (magenta) is shown Modeling 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 Engineers integrate MS data with geologic mod- if blue. The thickness of the disks represents opening, and their relative displacement represents the els, mechanical earth models (MEMs), formation degree of slip. The glyph’s central plane, which is parallel to the disks’ planar surfaces, is oriented imaging logs and production logs to characterize with respect to the strike and dip of the fracture plane activated or caused by the event. (Adapted from reservoirs and aid their understanding of micro- Leaney et al, reference 47.) seismicity. In the past, engineers predicted future reservoir production based on correlations between poststimulation production and ESV. 38. Shapiro SA, Dinske C, Langenbruch C and Wenzel F: 45. By analyzing the amplitudes of waveforms received “Seismogenic Index and Magnitude Probability of at an array of recording sensors, geophysicists can Analysts now use geomechanical modeling to Earthquakes Induced During Reservoir Fluid determine the location, shape, size and orientation of enhance the forecasting. Stimulations,” The Leading Edge 29, no. 3 (March 2010): the motions of the causative event. Geophysicists then 304–309. use the amplitude data to invert for the moment tensor, Geomechanics can aid in the design of 39. Cipolla C, Maxwell S and Mack M: “Engineering Guide a system of point-force couples, which is the best-fit hydraulic stimulations to maximize the hydraulic to the Application of Microseismic Interpretations,” seismic radiation pattern equivalent to that observed from seismic event displacement discontinuities. fracture surface area exposed to the reservoir paper SPE 152165, presented at the SPE Hydraulic Fracturing Technology Conference, The Woodlands, For more on radiation patterns and moment tensors: and to the system of natural fractures within. For Texas, February 6–8, 2012. Lay and Wallace, reference 4. planning wells and determining in situ stress 40. Schorlemmer D, Wiemer S and Wyss M: “Variations in 46. Moment tensor inversion (MTI) is now integrated with Earthquake-Size Distribution Across Different Stress the Petrel software platform and Mangrove reservoir- states, engineers perform stress simulations Regimes,” Nature 437 (September 22, 2005): 539–542. centric stimulation design software workflows. using static and time-lapsed 3D MEMs that are Downie RC, Kronenberger E and Maxwell SC: “Using 47. For more on the theory, decomposition and display Oilfield Review of moment tensors: Leaney S, Chapman C and Yu X: integrated with results from reservoir simulation Microseismic Source Parameters to Evaluate the Influence of Faults on Fracture Treatments—AMAY 16 “Anisotropic Moment Tensor Inversion, Decomposition models.49 Geomechanical modeling can provide Geophysical Approach to Interpretation,”Microseismic paper Fig 11and Visualization,” Expanded Abstracts, 84th SEG Annual International Meeting and Exposition, Denver insight into the extent of fracture-to-fracture SPE 134772, presented at the SPE AnnualORMAY Technical 16 MCSMC 11 Conference and Exhibition, Florence, Italy, (October 26–31, 2014): 2250–2255. interference between fracture stages in a treat- September 19–22, 2010. 48. For more on deriving DFN from MS data: Yu X, ment well, nearby treated wells and the natural 41. Boroumand N: “Hydraulic Fracture b-Value from Rutledge JT, Leaney SW and Maxwell S: “Discrete- Microseismic Events in Different Regions,” presented Fracture-Network Generation from Microseismic Data fracture system. at the GeoConvention 2014, Calgary, May 12–16, 2014. by Use of Moment-Tensor- and Event-Location- Constrained Hough Transforms,” paper SPE 168582, Knowledge of the regional stress state and the 42. Zorn EV, Hammack R and Harbert W: “Time Dependent presented at the SPE Hydraulic Fracturing Technology b and D-values, Scalar Hydraulic Diffusivity, and characteristics and distribution of natural frac- Conference, The Woodlands, Texas, February 4–6, 2014. Seismic Energy From Microseismic Analysis in the tures in the reservoir is important for predicting Marcellus Shale: Connection to Pumping Behavior 49. Schlumberger reservoir engineers couple ECLIPSE 3D simulations with the VISAGE finite-element the effectiveness of reservoir stimulations. During Hydraulic Fracturing,” paper SPE 168647, presented at the SPE Hydraulic Fracturing Technology geomechanics simulator to create dynamic, time-lapse During stimulation operations, hydraulic frac- Conference, The Woodlands, Texas, February 4–6, 2014. models of stress and production history of single and multiple wells and fields. For more on integrated 43. Grob and van der Baan, reference 37. tures interact with preexisting natural fractures. modeling: Alexander T, Baihly J, Boyer C, Clark B, Slip along natural fractures generally increases 44. Williams MJ, Khadhraoui B and Bradford I: “Quantitative Waters G, Jochen V, Le Calvez J, Lewis R, Miller CK, Interpretation of Major Planes from Microseismic Event Thaeler J and Toelle BE: “Shale Gas Revolution,” the permeability of the stimulated fractures. The Locations with Application in Production Prediction,” Oilfield Review 23, no. 3 (Autumn 2011): 40–55. density and orientation of the natural fracture Expanded Abstracts, 80th SEG Annual International Meeting and Exposition, Denver (October 17–22, 2010): population are significant factors that influence 2085–2089.

May 2016 27 Data Product Reconciling the UFM modeling results with Real-time Treatment the MS event patterns requires the evaluation of Seismic Microseismic Monitoring Modification multiple DFN realizations and may not result in •Structure •Event processing •Modify pump rate an exact match but rather a statistically probable •Natural fractures and faults •Interpretation •Modify proppant schedule match to the MS pattern. After the UFM simula- •Rock properties •Visualization •Continue or abort job •Stress variations tions are calibrated, the fluid flow characteristics predicted by the UFM technique can be incorpo- Model rated in a reservoir simulation. In this process, Logs, Cores Earth Models the hydraulic fractures are explicitly gridded in and Petrophysics Hydraulic Fracture Model the reservoir model to honor the 3D hydraulic •Natural fractures •Velocity model •Fracture geometry and fracture geometry and proppant distribution. •Rock properties •1D or 3D MEM conductivity distribution •Stress profile and anisotropy •Geologic model In addition to the reservoir grid, a fluid model, •Reservoir properties •Reservoir model a set of relative permeabilities, stress-dependent •DFN hydraulic fracture conductivity profiles, histori- cal production rates and bottomhole pressure are Completion and Completion and Stimulation Stimulation Hydraulic Fracture Models input into the reservoir simulator. After the res- •Modify completion ervoir models are calibrated, analysts can use •Perforation locations •Fracture model calibration strategy •Stage sequence them to forecast hydrocarbon recovery and per- •Treatment data form sensitivity analyses. The analysts can vary completion parameters, including the number of Reservoir Simulation Models Reservoir Simulation stimulation stages, the number of perforation Flow clusters per stage and reservoir parameters such •Reservoir model calibration •Production profile as permeability, porosity and saturations to maxi- •Production logs •Hydrocarbon recovery mize future stimulation operations. •Production data •Drainage architecture 50. For more on fracture network development: Johri M Figure 12. Workflow for completion and stimulation design and field development applications and Zoback MD: “The Evolution of Stimulated Reservoir of microseismic mapping. Using seismic data, well logs and core data along with treatment and Volume During Hydraulic Stimulation of Shale Gas production data (left), engineers build a series of earth models and discrete fracture network (DFN) Formations,” paper SPE 168701/URTeC 1575434, models (middle). They use these models to generate hydraulic fracture predictions for the fracture presented at the Unconventional Resources Technology geometry and conductivity distribution resulting from stimulation operations. They also use MS event Conference, Denver, August 12–14, 2013. locations and deformation to calibrate the earth and fracture models. Reservoir engineers are then 51. Maxwell SC and Cipolla C: “What Does Microseismic Tell Us About Hydraulic Fracturing?,“ paper SPE 146932, able to predict fracture network geometry and conductivity and generate reservoir performance presented at the SPE Annual Technical Conference and simulations (right). (Adapted from Cipolla et al, reference 39.) Exhibition, Denver, October 30–November 2, 2011. 52. For more on an early, computationally efficient geomechanical modeling approach: Xu W, Le Calvez J the stimulated fracture network development the UFM unconventional fracture model can then and Thiercelin M: “Characterization of Hydraulically- 50 Induced Fracture Network Using Treatment and and control reservoir productivity. be used to predict the fracture geometries that Microseismic Data on a Tight-Gas Formation: A Hydraulic fracturing creates a tensile frac- result from the stimulation.54 Geomechanical Approach,” paper SPE 125237, presented at the SPE Tight Gas Completions Conference, ture that opens slowly, and most of the rock Modelers can use treatment data such as pump San Antonio, Texas, June 15–17, 2009. deformation occurs aseismically at much lower pressure, fluid volumes and proppant loadings as 53. For more on the construction of DFNs: Will R, Archer R frequencies than the typical MS signal band.51 inputs to the numerical calculations and then use and Dershowitz B: “Integration of Seismic Anisotropy and Reservoir Performance Data for Characterization of Shear deformation occurs in the process zone MS data to constrain the results. These UFM simu- Naturally Fractured Reservoirs Using Discrete Feature around the fracture tip, in the vicinity of the frac- lations yield predictions of stimulated fracture Network Models,” paper SPE 84412, presented at the SPE Annual Technical Conference and Exhibition, ture face as a result of leakoff into preexisting geometry and conductivity distributions. Analysts Denver, October 5–8, 2003. natural fractures and at doglegs and other geo- iteratively calibrate the model by adjusting the Offenberger R, Ball N, Kanneganti K and Oussoltsev D: “Integration of Natural and Hydraulic Fracture Network metric deflections. In contrast to the aseismic input parameters to the UFM simulation to achieve Modeling with Reservoir Simulation for an Eagle Ford nature of tensile dilation, shear deformation a match between fracture geometry predictions Well,” paper SPE 168683/URTeC 1563066, presented at the Unconventional Resources Technology Conference, often emits sudden, high-frequency, audible seis- and observed MS event locations and deformation, Oilfield Review Denver, August 12–14, 2013. mic energy. MAY 16 or seismic moments (Figure 14). Adjustable 54. UFM processing is embedded in the Mangrove software Reservoir engineers have developed a variety parameters include horizontal stresses and prop- platform and uses the output of the VISAGE simulator. Microseismic Fig 12 For more on UFM processing: Weng X, Kresse O, of approaches to characterize poststimulationORMAY 16 MCSMCerties of the12 DFN and fracturing fluid. Analysts use Cohen C, Wu R and Gu H: “Modeling of Hydraulic- fracture networks (Figure 12).52 In one approach, volumes of fluids pumped and stresses to constrain Fracture-Network Propagation in a Naturally Fractured Formation,” SPE Production and Operations 26, no. 4 analysts use information from seismic reflection the fracture model to reduce the set of possible (November 2011): 368–380. surveys, well logs and cores to build a DFN, which solutions. They conduct sensitivity analyses to For more on UFM modeling: Cipolla C, Weng X, Mack M, Ganguly U, Gu H, Kresse O and Cohen C: “Integrating they combine with a set of earth models to determine how much the answer changes as input Microseismic Mapping and Complex Fracture Modeling describe the reservoir and surrounding forma- parameters are varied and to ensure that non­ to Characterize Fracture Complexity,” paper SPE 140185, 53 presented at the SPE Hydraulic Fracturing Technology tions. Fracture information may often be unique multiple solutions give results that are Conference and Exhibition, The Woodlands, Texas, derived from resistivity or ultrasonic image logs reasonable in terms of predicted production. January 24–26, 2011. (Figure 13). Hydraulic fracture models such as

28 Oilfield Review Primary fracture set n = 775 Orientation: Start 0° approximately N54°E 330° 30°

300° 60° Observations: Logs, seismic data and natural fractures

DFN Realization

270° 90°

Fracture treatment data and mechanical earth model (MEM)

UFM Prediction 240° 120°

210° 150°

180° Secondary fracture set Microseismic Events Comparison Composite Fracture Set n = 175 Orientation: approximately No N60°W Match?

Yes

Output fracture model

Figure 14. Flowchart for calibrating UFM processing and DFN simulations.

1,000 ft 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, Figure 13. Construction of a discrete fracture network. For a fracture adjusting UFM processing inputs for multiple DFN realizations (bottom stimulation in the Eagle Ford Shale, analysts processed logging data right) to arrive at the best overall agreement between the modeled fracture obtained from a well drilled with oil-base mud using an imaging tool and geometries and the observed deformations. (Adapted from Cipolla et al, were able to detect fractures and determine their orientations from a reference 54.) 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 Oilfield Review inside to the outside edge of the rose plot, the dip varies from 0° to 90°. MAY 16 The black lines represent fracture strike orientation and the length is Microseismic Fig 15 related to the abundanceOilfield along Review a given direction. Analysts identified a ORMAY 16 MCSMC 15 primary and a secondaryMAY 16 fracture set. The intensity of fracture occurrence along the wellboreMicroseismic correlated with Fig formation 13 curvature determined using reflection seismic ORMAYsurveys. 16Co-kriged—a MCSMC 13 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.)

May 2016 29 Monitor Well C Geophones Stimulation stage

12345678910 11 12 13 14 15 16 Well plug

Treatment Wells A and B

Austin Chalk

Eagle Ford Shale

Buda Limestone

Figure 15. Vertical section view of MS events recorded during the of the stimulation operation. Microseismic events were largely confined to stimulation of two horizontal wells in the Eagle Ford Shale. Microseismic the Eagle Ford Formation, bounded above by the Austin Chalk and below by event hypocenters (spheres, color-coded by stage number) and the the Buda Limestone. During stimulation Stage 8 (dark blue) of Well A, the trajectories of Wells A (orange), B (blue) and C (yellow) are shown in MS events aligned with the adjacent major fault and propagated downward conjunction with the formation tops (horizontal tan and light blue) and faults into the Buda Limestone. This observation suggested the fault affected (vertical gray and green). Treatment Wells A and B were stimulated in a fracture growth during the stimulation. A decision was made to abort 21-stage zipper fracture operation. Engineers used a 12-level geophone Stages 9 (teal) and 10 (crimson) and proceed with Stage 11 (orange) on the array positioned in the vertical portion of Well C to monitor 16 of the 21 stages other side of the fault. (Adapted from Zakhour et al, reference 58.)

6,000 Engineered Completions The Eagle Ford Formation, an upper Cretaceous marl in South Texas, is a target for oil and gas 6,400 development. Because the formation is highly laminated and has ultralow permeability, effec- 6,800 tive completion designs are required to maximize production. In September 2013, an operator 7,200 tested MSM as a method for guiding hydraulic Major fault fracture operations and evaluating fault interac- 7,600 tions during stimulation. Depth, ft The operator drilled two horizontal lateral 8,000 wells in the gas-condensate window of the Eagle Oilfield Review Ford Shale trend in Karnes County, Texas. These 8,400 MAY 16 wells, which were drilled parallel to each other Microseismic Fig 16 8,800 ORMAY 16 MCSMC 16 and approximately 330 ft [100 m] apart, crossed a major fault. After the operator drilled the wells, 3H 1H measurements of petrophysical and geomechani- 400 ft cal properties were acquired using the ThruBit 2H through-the-bit logging services.55 Using the N Mangrove engineered stimulation design module Figure 16. Trajectories of three wells to be stimulated in the Barnett Shale in relation to a fault system in the Petrel platform, engineers developed a mapped from 3D surface seismic data. Proximity of faults can influence the local stress field, affecting completion design for one of the two laterals.56 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.)

30 Oilfield Review

Oilfield Review MAY 16 Microseismic Fig 17 ORMAY 16 MCSMC 17 Stimulation design engineers used reservoir Well 1H and completion quality parameters derived from Stage 1 Figure 17. Microseismic events the Mangrove software to aid in optimizing stage Stage 2 detected during stimulation of intervals and the placement of perforations along Stage 3 57 three horizontal wells in the the lateral. Flow measurements in some previ- Stage 4 Barnett Shale. Fault traces ous wells that had geometric—evenly spaced— Well 2H (cyan) are mapped at the Stage 1 depth of laterals 1H (red), 2H completions had shown unequal contribution to 3H 1H Stage 2 (green) and 3H (yellow). The production across perforations. The engineered 2H Stage 3 zipper fracturing performed on stages grouped perforation clusters in regions of Stage 4 Wells 1H and 3H was monitored the lateral that had similar horizontal stress. from Well 2H. Engineers also Well 3H Completion engineers anticipated that all perfo- monitored four stimulation Stage 1 stages performed on Well 2H ration clusters in a stage would break down and Stage 2 using sensors in Well 3H. initiate fractures simultaneously because each Stage 3 Color coding and symbols cluster had similar stress characteristics. Stage 4 are used to represent stages Measurements made during stimulation con- Stage 5 and wells. Microseismicity for stages closer to the fault firmed that lower than average treating pres- system tended to be compact,

sures were required for the engineered Y-axis, ft an observation that was completion than had been used for the lateral explained later by fracture that was stimulated using a geometric model. modeling as stimulation-fault interaction. Longer hydraulic Survey engineers acquired MS data during fracture wings occurred in the the stimulation using a 12-receiver downhole stages that were executed array in a nearby vertical well. Using these MS closer to the toe of the well than data, the operator monitored hydraulic fracture those executed close to the heel. Microseismicity overlap development near a fault system, identified fault observed between successive interaction and adjusted the completion design stages indicates insufficient 400 ft to avoid the fault. Engineers later studied the fracture isolation. (Adapted from mechanisms of fault interaction through postjob N Le Calvez et al, reference 59.) 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 400 ft X-axis, ft treatment fluid and proppant to thief zones along the fault. In this project, the completion design excluded stimulation stages that were within Maximizing Recovery varied from 20 to 100 ft [7 to 30 m]. Engineers 250 ft [76 m] of the identified major fault. In 2011, engineers and geoscientists with Teleo used a zipper fracture stimulation on the central Monitoring revealed that MS events were Operating, LLC and Eagleridge Energy, LLC con- Well 1H and eastern-most lateral 3H. The west- generally well bounded within the target Eagle ducted a multistage, multilateral stimulation in ern-most lateral 2H was stimulated later. Ford Formation and the overlying Austin Chalk the Barnett Shale in Denton County, Texas. The Microseismic survey engineers monitored (Figure 15). However, for some stages, MS activ- operators drilled three parallel horizontal wells stimulations on Wells 1H and 3H using a ity and treating pressure records indicated unex- into the lower Barnett Shale. Well trajectories 3C-accelerometer array that was placed in hori- pected fracture bridging and potential proppant were about 500 ft [150 m] apart; the central well zontal Well 2H using a tractor and repositioned screenout. Analysis of MS event clusters and was landed about 80 ft [25 m] shallower than the according to the well and stage to be monitored. b-values alerted the engineers that the hydrau- outside laterals. Because of lease boundary con- Operations on Well 2H were later monitored lic fractures had encountered the nearby fault, straints, the wells had to be placed in the vicinity using an array deployed in a vertical section which blocked and limited fracture develop- of several large faults. The lateral sections of the below the 3H wellhead. During monitoring, engi- ment and led to premature stage terminations. wells were drilled awayOilfield from Review the major fault and neers observed MS activity across all pumped Real-time interpretation allowed modification of through a smaller faultMAY (Figure 16 16). Fault throws stages (Figure 17). the completion strategy, and several stimulation Microseismic Fig 18 55. For more on ThruBitORMAY services: Aivalis 16 MCSMC J, Meszaros 18 T, Chadwick C, Reischman R and Wigger E: “Eagle Ford stages planned near the fault were abandoned. Porter R, Reischman R, Ridley R, Wells P, Crouch BW, Completion Optimization Using Horizontal Log Data,” Recommendations were made for future comple- Reid TL and Simpson GA: “Logging Through the Bit,” paper SPE 166242, presented at the SPE Annual Oilfield Review 24, no. 2 (Summer 2012): 44–53. Technical Conference and Exhibition, New Orleans, tion designs to increase buffer zones from 250 ft 56. For more on the Mangrove service: Ajayi B, Aso II, September 30–October 2, 2013. to about 400 ft [120 m] on either side of major Terry IJ Jr, Walker K, Wutherich K, Caplan J, 58. For more on horizontal completion optimization across faults to minimize the risk of pumping nonpro- Gerdom DW, Clark BD, Ganguly U, Li X, Xu Y, Yang H, a major fault: Zakhour N, Sunwall M, Benavidez R, Liu H, Luo Y and Waters G: “Stimulation Design for Hogarth L and Xu J: “Real-Time Use of Microseismic ductive stages.58 Unconventional Resources,” Oilfield Review 25, no. 2 Monitoring for Horizontal Completion Optimization (Summer 2013): 34–46. Across a Major Fault in the Eagle Ford Formation,” 57. For more on determining reservoir and completion paper SPE 173353, presented at the SPE Hydraulic quality: Slocombe R, Acock A, Fisher K, Viswanathan A, Fracturing Technology Conference, The Woodlands, Texas, February 3–5, 2015.

May 2016 31 Fault using a range of analytical techniques and incor- 100 porated summary statistics of MS event attri- 90 butes within the workflow. Event attributes

80 included seismic moment and moment magni- tude. From these parameters, analysts deter- 70 mined summary b-value statistics and inferred 60 relative stress magnitudes for each stimulation 50 stage.61 They used D-value estimates to extract 3H 40 fracture planes from the clouds of microseismic

Effective stress, psi 1H events—D-values near two indicated planar 30 alignments—and then used the planes for DFN 20 2H construction. Monitoring geometries that can 10 provide full MTI offer an additional means to con- 0 strain DFN construction and modeling but were not available in this study. Additional input to the analysis included earth model parameters such 100 Fault as the rock mechanical properties of layers and 90 natural fracture geometries along with stimula- 80 tion data such as well geometry, flow rates, fluid 70 types and surface pumping pressures.

60 Engineers combined the Mangrove, UFM and VISAGE software to model the hydraulic fractur- 50 ing process, the interaction of induced and natu- 40 ral fractures and the stress field.62 These models Effective stress, psi 30 predicted the chronological development of the 3H 20 interconnected fracture network and were cali- brated iteratively using the observed evolution of 10 1H MS activity. They tested fracture propagation sce- 0 2H narios by matching time-distance relationships within the microseismicity pattern and then Figure 18. Geomechanical modeling using finite element analysis. The stimulation of the Barnett iteratively improved the interpretation by updat- Shale Well 3H, Stage 5, was near a major fault. The fault plane projections (olive green) are shown ing the location and properties of the natural intersecting Wells 1H (blue), 2H (green) and 3H (yellow). Simulations were run for materials near the fractures. Engineers constrained the simulation fault that had equivalent high stiffness (top) and low stiffness (bottom). The fault properties affect the using material balance, which reconciled 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 fracture volume opened during stimulation with microseismicity are expected to be limited in the region near the fault. The colored volumes show only the volumes of pumped fluid and proppant and the simulation elements where the stresses are perturbed toward failure by the stimulation. The colors the volume of fluid estimated to have leaked off correspond to the minimum in situ effective stress; purple is low compression, showing the region into the formation. The simulation provided a where tensile failure is most likely to occur, and red is high compression. (Adapted from Williams et al, reference 61.) description of proppant placement together with a prediction of which natural fractures might During zipper fracturing of Wells 1H and 3H, of the target depth interval. Observations 59. Le Calvez J, Xu W, Williams M, Stokes J, Moros H, interpretation of MS event location trends indi- revealed downward fracture growth and alerted Maxwell S and Conners S: “Unconventional Approaches cated the hydraulic fracture azimuth, N50–55°E, the operator to stop pumping to avoid fracturing for an Unconventional Faulted Reservoir—From Target Selection to Post-Stimulation Analysis,” paper P336, was consistent with the expected maximum hori- into the water-bearing Viola Limestone below the presented at the 73rd EAGE Conference and Exhibition, zontal stress direction; however, the extent of the zone of interest. During another stage, engineers Vienna, Austria (May 23–26, 2011). microseismicity varied from stage to stage. The recognized that the planar alignment of MS 60. Williams MJ, Le Calvez JH and Stokes J: “Towards Self-Consistent Microseismic-Based Interpretation of MS locations for the early stages, near the toe of events indicated slip along a fault and were able Hydraulic Stimulation,” paper Th 01 15, presented at the wells, extended farther from the Oilfieldwellbore Review to stop pumping for that stage and bypass a the 75th EAGE Conference and Exhibition, London, June 10–13, 2013. than those observed during later stages,MAY toward 16 faulted zone before resuming stimulation. Microseismic Fig 19 61. Williams MJ, Le Calvez JH, Conners S, Xu W: the heel of the wells. These later stages were Microseismic monitoring allowed the operator to “Integrated Microseismic and Geomechanical Study in ORMAY 16 MCSMC 19 the Barnett Shale Formation,” Geophysics 81, no. 3 closer to the main fault and displayed a more modify the stimulation program during the ongo- (May–June 2016): 1–13. compact microseismicity pattern and shorter ing job operation.59 62. In the Barnett Shale case study, Schlumberger fracture wings away from the wellbore. Postsurvey modeling and data integration engineers used Mangrove software with the UFM complex fracture simulator to model fracture interaction The MS locations observed in these later provided a more complete explanation of the and the VISAGE simulator to model stress. stages overlapped with those observed in some stimulation and MS responses.60 Analysts con- 63. For more on modeling of faulted rock masses: Pande GN, Beer G and Williams JR: Numerical Methods in Rock early stages. Several later stages near the heels of structed a complete history of the treatment Mechanics. Chichester, New York, USA: John Wiley the wells displayed microseismicity that was out and Sons Ltd., 1990.

32 Oilfield Review open and which ones behaved as barriers that large values of fault-related stiffness. By simulat- Advances in acquisition, processing, interpre- promoted vertical or asymmetric fracture growth. ing all stages and varying fault-related stiffness, tation and integration of MS data are providing Modeling for this Barnett Shale stimulation modelers demonstrated that the interaction with unique insights into and increased understand- helped analysts understand the downward growth the fault was consistent with microseismicity ing of stimulated reservoir behavior. Models help of microseismicity into the Viola Limestone. The that was more compact in the stages toward the engineers interpret and constrain MS data, and complex fracture simulator reproduced observa- heel than in those toward the toe. The under- geomechanical earth models help them charac- tions of hydraulic fracture interactions with standing of fault interaction gained from this terize the variability of reservoir properties. natural fractures, which acted as barriers to prop- case study should benefit reservoir engineers Fracture network modeling facilitates predic- agation in the Barnett Shale and in the underlying planning future refracturing operations in these tions of the interactions between hydraulic frac- Viola Limestone. By tuning the fault properties wells or stimulation treatments in nearby wells. tures and rock fabric. Reservoir simulations within the finite-element geomechanical model, assist in predicting field drainage patterns, and analysts were able to match results with the Challenges and the Future productivity may be validated through produc- observed distribution of MS events. As operators develop unconventional resources, tion history matching. Monitoring microseismic- Analysts used finite-element geomechanical determining the optimum well spacing and com- ity offers valuable data for validating these modeling to study how fault properties influ- pletion strategies that maximize ultimate recov- models and simulations. enced the zone that a hydraulic fracture perturbs ery is critical. To help operators achieve these Methods such as MSM give operators insight toward failure (Figure 18). In this type of model- objectives, MSM provides key data for constrain- into reservoir dynamics that far exceeds what was ing, the rock mass, including its embedded faults ing and calibrating models used to help geoscien- possible even a few years ago. Success in develop- and fractures, is described as an equivalent tists with data interpretation and integration. ing unconventional reservoirs owes much to the medium that has standard mechanical properties Microseismicity is induced as the reservoir and pioneers working in plays such as the Barnett such as elastic modulus and strength.63 The pres- adjoining formations respond to stimulation Shale. Recent advances in tools and technolo- ence of an active fault may alter the regional treatments. Models used to predict how these for- gies are allowing operators to develop unconven- stress field in its vicinity, pushing outward, nor- mations should respond to stimulations must tional reservoirs with greater certainty, reduced mal to its plane, and increasing the stress on frac- simulate and faithfully reproduce the observed risk and deeper understanding of the nature of ture interfaces. The simulations revealed that microseismicity. Challenges for geoscientists are these formations. —HDL the compact distribution of MS events in stages the accurate measurement of MS events and pumped close to the fault was consistent with extracting of maximal information from them.

Contributors

Joël Le Calvez is a Schlumberger Geophysics Advisor include hydraulic fracture treatment design and evalu- and geomechanical modeling. He joined Schlumberger and Microseismic Domain Expert in Houston. He works ation, production data analysis, reservoir simulation, GeoQuest in 1997 as a commercialization software on development and commercialization of microseismic geomechanics, microseismic monitoring, restimulation, engineer and worked as project leader and team leader and borehole seismic products while heading a team multiwell pad development and weak interface model- in Abingdon, England. In 2002, he was a team leader in of geophysicists, geologists and stimulation engineers ing. Raj holds a BTech degree in petrochemical engi- Sugar Land, Texas, developing the first hydraulic frac- working on various plays around the world. He has man- neering from Dr. Babasaheb Ambedkar Technological ture monitoring software to support the interpretation aged the Microseismic Services Answer Product Center University, Lonere, Maharashtra, India, and an of microseismic information in the context of fracture and the borehole seismic processing and crosswell seis- MS degree in petroleum engineering from Texas A&M stimulation. He joined SGR as a senior research scien- mic groups in Houston since 2014. His main responsibili- University, College Station. tist in 2004, where he worked in applied reservoir engi- ties are the processing and interpretation of data for Jerry Stokes is the President and Owner of Mid- neering, fluid measurements (as program manager) and geologic, geophysical and geomechanical applications. Continent Geological, Inc. in Fort Worth, Texas. He has well test interpretation. Michael received a BS degree He also works with product centers defining and testing been a certified petroleum geologist with the AAPG for in physics and an MS degree in geophysics, both from software programs and with research centers on defin- more than 35 years. Since 1987, he has been involved Imperial College of Science, Technology and Medicine, ing and testing of algorithms. Joël joined Schlumberger in oil and gas exploration, geologic consulting and sales University of London. He also has a PhD degree in phys- in 2001, and after several years in the field acquiring and marketing of geologic projects throughout Texas ics from the University of Wales, Aberystwyth. and processing seismic data, he led the microseismic and nearby states. As a geologist for Panhandle Eastern Jian Xu is a Senior Microseismic Services Engineer in processing and interpretation team in Dallas from 2008 Pipeline, Jerry was responsible for the early develop- Houston. He focuses on microseismic data interpreta- until 2011. He then moved to Houston to manage the ment of underground gas storage fields in Kansas, tion, hydraulic fracture monitoring and stimulation North America microseismic processing and interpreta- Louisiana, Illinois and Michigan, USA. He then worked program evaluation in unconventional plays. He joined tion center. He earned a BSc degree in mathematics and for Rust Oil Corporation as the exploration manager Schlumberger in 2008 as a field engineer in Bryan, physics and an MSc degree in geology and geophysics, for the Permian basin. He is a member of the Society of Texas. He held various positions, including access both from the Université de Nice Sophia Antipolis, Independent Professional Earth Scientists and the Fort field engineer, production stimulation engineer and France; a Diplôme d’Etudes Approfondies in tectono- Worth Wildcatters. Jerry has a BSc degree in geology microseismic services engineer working on several physics from the Université Pierre et Marie Curie, Paris; and geophysics from Texas Tech University, Lubbock. unconventional plays in the US, all while located in and a PhD degree in geology from The University of Houston. Before his current assignment, Jian was Texas at Austin, USA. Michael Williams is a Principal Reservoir Engineer in Geophysics at Schlumberger Gould Research (SGR), a senior production stimulation engineer at the Raj Malpani is a Senior Completions and Production Cambridge, England. Since 2008, he has worked in the Production Technology Integration Center in Houston. Engineer with Schlumberger Technology Corporation area of interpretation of microseismicity, specifically He obtained BS and MS degrees in electrical engineer- in Houston. For the past 10 years, he has been a part the accurate recovery of statistical information from ing from Tianjin University, China, and a PhD degree of integrated teams that address technical challenges detection-limited microseismic data, and the application in petroleum engineering from Texas A&M University, pertaining to unconventional reservoirs. His interests of microseismic interpretation to reservoir simulation College Station.

May 2016 33 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 Asrar Corrosion validates the universal law of entropy; Corrosion has brought down bridges, downed Bruce MacKay everything trends toward a state of greater chaos aircraft, leveled chemical plants, parted drill- Sugar Land, Texas, USA and disorder. The flecks of rust on an iron bar or pipe and ruptured pipelines. Given sufficient the green patina on a copper fixture are evidence of time, this adversary has the potential to degrade Øystein Birketveit the insidious effects of corrosion. These examples any material. In certain environments, the Marko Stipanicˇev may be regarded as an annoyance, but taken to the unchecked effects of corrosion can come swiftly, Bergen, Norway extreme, the results of corrosion can lead to cata- and the consequences of failure to manage corro- strophic outcomes. sion can be costly. Joshua E. Jackson G2MT Laboratories, LLC Houston, Texas

Alyn Jenkins Aberdeen, Scotland

Denis Mélot Total Paris, France

Jan Scheie Stavanger, Norway

Jean Vittonato Total Pau, 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.

34 Oilfield Review

Oilfield Review MAY 16 Corrosion Fig Opener ORMAY 16 CRSSN Opener According to the US Federal Highway Admin- istration, the approximate annual direct cost of C H I N A NEP corrosion for the US in 2015 was an estimated P A K I S T A N AL US$ 500 billion, representing around 3.1% of New Delhi 1 the nation’s gross domestic product. This figure BANGLADESH amounts to six times the average annual cost of weather-related disasters for the US, which was I N D I A about US$ 87 billion in 2011.2 Unlike weather events, corrosion can be controlled or at least Mumbai managed; scientists estimate that 25% to 30% of corrosion costs could be avoided if good corrosion management practices and preventive strategies 3 were employed. 0 500 km SRI LANKA Throughout the ages, and despite an early 0 miles 500 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- Figure 1. Delhi pillar. This iron pillar is located in the Qutub Complex in New Delhi, Delhi, India (inset). ing—fouling of underwater surfaces by organisms It is about 9.1 m [30 ft] tall and weighs approximately 6,000 kg [13,200 lbm]. Erected in 400 CE, the such as barnacles and algae—and increase speed 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 accelerated the corrosion of nails that held the attributed to the high concentration of phosphorus in the iron. ships together.5

1. Koch GH, Brongers MPH, Thompson NG, Virmani YP and Payer JH: “Corrosion Costs and Preventive Strategies in Michael Faraday was one of the most impor- nisms of corrosion and search for methods to the United States,” Washington, DC: US Department of tant contributors to the early understanding of manage and control it. Transportation Federal Highway Administration, Publication FHWA-RD-01-156, March 2002. corrosion; in the early 1800s, he established a Combating corrosion is a significant source Jackson JE: “Corrosion Will Cost the US Economy over quantitative relationship between the chemi- of expenditures for the oil and gas industry $1 Trillion in 2015,” G2MT Laboratories, http:// 6 www.g2mtlabs.com/corrosion/cost-of-corrosion/ cal action of corrosion and electric current. (Figure 2). British Petroleum (BP) conducted a (accessed January 6, 2016). Although much more is known about the subject study of its operations in the North Sea in 1995.7 Papavinasam S: Corrosion Control in the Oil and Gas today, scientists continue to study the mecha- The company found that outlays for corrosion Industry. Waltham, Massachusetts, USA: Gulf Professional Publishing, 2014. 2. The US$ 87 billion cost of weather-related disasters in US Oil and Gas Corrosion Expenditures, US$ billion/year 2011 was the highest on record. The average annual cost has been closer to US$ 10 billion in recent years. For Oilfield Review more on the cost of weather-related disasters: Smith AB and Katz RW: “U.S. Billion-Dollar Weather and Climate MAY 16 Refining Distribution Disasters: Data Sources, Trends, Accuracy and Biases,” Corrosion Fig 1 3.7 5.0 Natural Hazards 67, no. 2 (June 2013): 387–410. ORMAY 16 CRSSN 1 3. Chillingar GV, Mourhatch R and Al-Qahtani GD: The Production Fundamentals of Corrosion and Scaling for Petroleum 1.4 and Environmental Engineers. Houston: Gulf Publishing Storage Company, 2008. 7.0 4. Kumar AVR and Balasubramaniam R: “Corrosion Product Analysis of Corrosion Resistant Ancient Indian Iron,” Pipelines Corrosion Science 40, no. 7 (July 1, 1998): 1169–1178. Tankers 7.0 Balasubramaniam R: Story of the Delhi Iron Pillar. Delhi, 2.7 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. Figure 2. Corrosion expenditures. Corrosion expenditures in the US 7. Kermani MB and Harrop D: “The Impact of Corrosion on the Oil and Gas Industry,” SPE Production & Facilities 11, oil and gas industry are about US$ 26.8 billion/year. The downstream no. 3 (August 1996). 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.)

May 2016 35

Oilfield Review MAY 16 Corrosion Fig 2 ORMAY 16 CRSSN 2 Iron can also react with CO2 to form iron carbon- ate [FeCO3] and with H2S to form iron sulfides Anodic reaction Cathodic reaction [FexSx]. In the absence of O2 but the presence of 0 2+ – H O + 2e– 0.5 O + 2OH– Fe Fe + 2e 2 2 CO and H S, the cathodic reaction can generate Fe2+ + 2OH– Fe(OH) 2 2 2 hydrogen gas. These reactions can occur rapidly, but if the reaction rate can be reduced, the overall corro- OH– Water OH– sion rate will also be reduced. Many factors influ- – OH ence the reaction rate. These include the type and quality of metal, electrolyte compositions, Fe2+ Fe(OH) 2 pH, temperature, pressure, presence of dissolved Fe(OH)2 gases, liquid velocity, water salinity, applica- Fe0 Anode Fe0 Cathode tion of cathodic protection and the presence of 11 Electron flow microbes. To manage corrosion and corrosion rate, knowledge of the metallurgy of the mate- Fe0 Fe0 rials to be used and the environments in which Steel they will operate is important. If CO2 comes into contact with water in the producing or transportation system of an oil and Figure 3. Corrosion cell. When steel in water rusts, several reactions take gas operation, areas typically affected include 0 place simultaneously. At the anode, steel [Fe ] goes readily into solution to well internals, gathering lines and pipelines. In form ferrous iron [Fe2+] and ferric iron [Fe3+] (not shown) ions, and electrons CO corrosion of iron, the products of reaction move to the cathode. Electrons at the cathode react with water [H2O] to 2 – – form oxygen [O2] and hydroxyl [OH ] ions. The OH ions combine with the are carbonic acid, iron carbonate [FeCO3] and solubilized Fe2+ to form iron hydroxide [Fe(OH) ]. 12 2 hydrogen gas [H2]. 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, prevention and control averaged about 8% of Some forms of metal corrosion are related operators commonly use organic films that act as the total capital expenditure for its projects. to stability; for example, galvanic corrosion is barriers and inhibitors that neutralize the acidity On the UK Continental Shelf, 25% to 30% of BP’s an electrochemical process associated with the of the carbonic acid generated in the corrosion operating costs were related to the control and movement of electrons between areas that have process. Operators may also use corrosion resis- management of corrosion. Costs associated with different electrochemical potentials. The corro- tant alloys (CRAs), which are resistant to general replacing corroded equipment, lost production sion cell schematically describes oxidizing corro- and localized corrosion, in environments that are and corrosion-related contamination contributed sion, which is analogous to a battery in which two corrosive to carbon and low-alloy steels. to overall expenditures. In addition to the direct dissimilar metals are connected by an electrolyte Hydrogen sulfide is often found in produced 9 13 costs, the company found that corrosion had a (Figure 3). A metal that has a higher corrosion fluids or as a result of MIC. Although H2S is not significant indirect cost on health, safety and rate—more unstable—represents the negative corrosive, it becomes corrosive in the presence 14 environmental concerns. part of the cell and acts as the anode; a second of water. Sour corrosion from H2S can affect This article focuses on descriptions of cor- metal that has a lower corrosion rate—more any part of the producing system, including well rosion, management techniques and advances stable—acts as the positive part of the cell, internals and oil and gas gathering lines. Oilfield in corrosion abatement technologies. Field the cathode.10 fluids are considered sour if the produced gas con- 3 examples from Gabon, deepwater NigeriaOilfield and Review During the galvanic corrosion process, metal tains more than 5.7 mg of H2S per m [4 parts per the North Sea illustrate the ongoing battleMAY waged 16 oxides are formed as electrons flow from the million (ppm)] of natural gas or produced water 15 against corrosion by oil and gas operators.Corrosion Figanode 3 to the cathode through the electrolyte— has greater than 5 ppm H2S. At the anode, the ORMAY 16 theCRSSN fluid 3 in contact with the anode and cathode. H2S reacts with the iron to form several vari- The Corrosion Process A simplified version of iron oxidation can be used ants of iron sulfide [FexS] such as mackinawite Scientists and engineers today have a better to illustrate the galvanic corrosion process—the [(Fe,Ni)(1 + x)S], pyrrhotite [Fe(1 - x)S] and troilite understanding of corrosion processes than did actual process is more complex. The presence of [FeS].16 These iron sulfide species precipitate and 0 the ancient Romans and Chinese. Fighting cor- water [H2O] on the surface of the iron [Fe or Fe ] can form localized microgalvanic corrosion cells. rosion requires an understanding of the principal releases electrons to form ferrous iron [Fe+2] and The corrosion cells formed during sour corro- elements that cause and contribute to the corro- ferric iron [Fe+3] ions, which act as the anode in sion cause pitting, sulfide stress cracking (SSC) sion. There are several categories of corrosion; for our battery analogy. The liberated electrons flow and hydrogen embrittlement.17 Stress corrosion the oil and gas industry, common types include to the cathode, where, in the presence of oxygen cracking is a result of tensile stress combined exposure to carbon dioxide [CO2, sweet corro- [O2], ferrous oxide [FeO] and ferric oxide [Fe2O3] with a wet environment and often causes shal- sion], hydrogen sulfide [H2S, sour corrosion], form as scales of rust or precipitates. A byproduct low, round pits that have etched bottoms accom- oxygen [O2] and corrosion causing microbes, of the reaction at the cathode is hydroxyl ions panied by branching cracks that can lead to rapid referred to as microbiologically influenced cor- [OH–] from the reduction of oxygenated water. failure. Hydrogen embrittlement occurs when 8 rosion (MIC). H2S and H2 diffuse into metal, recombine with

36 Oilfield Review other molecules and create pressure within the 30 metal matrix; byproducts of cathodic protection, galvanic corrosion and other mechanisms may lead to hydrogen embrittlement. 25 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 20 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- 15 ging, nitrate treatments and biocides that reduce O2 the growth of microbes that cause MIC.18 Oxygen-related corrosion in oil and gas 10 Corrosion rate of carbon steel, mp y CO producing environments is often much more 2 aggressive than corrosion caused by CO2 or H2S (Figure 4).19 Corrosion by oxygen is directly pro- H2S portional to the concentration of the dissolved 5 gas. If chlorides, CO2 or H2S are present, the cor- rosion rate can increase significantly. Oxygen has the ability to induce corrosion 0 throughout producing systems. Inhibition of oxy- 12345 6789 gen corrosion is difficult, and corrosion reduction Oxygen efforts for production and water handling facili- 50 100 150 200 250 300 350 400 450 ties have usually been directed toward exclusion Carbon dioxide of oxygen from the system and the use of oxy- 100 200 300 400 500 600 700 800 900 gen scavengers. Typical oxygen scavengers are Hydrogen sulfide Gas concentration in water phase, parts per million ammonium bisulfite [NH4HSO3], sodium sulfite 20 [Na2SO3] and sodium bisulfite [NaHSO3]. In Figure 4. Corrosion rates. The relative rates of corrosion in milli-inches/year (mpy) of carbon steel addition to scavenger stripping, vacuum deaera- show pronounced differences when the steel is exposed to varying concentrations of O2, CO2 and H2S. tors are sometimes used to control the corrosive 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 effects of oxygen on metals. metal surfaces. 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 8. Popoola LT, Grema AS, Latinwo GK, Gutti B and 15. Stewart M and Arnold K: Gas Sweetening and into contact with oxygen in the mud system. Balogun AS: “Corrosion Problems During Oil and Gas Processing Field Manual. Waltham, Massachusetts. Both instances can induce corrosion. The usual Production and Its Mitigation,” International Journal of Gulf Professional Publishing, 2011. Industrial Chemistry 4, no. 1 (2013). 16. Ning J, Zheng Y, Young D, Brown B and Nesic S: expression of oxygen-related corrosion is pitting. Chillingar et al, reference 3. “A Thermodynamic Study of Hydrogen Sulfide Corrosion Pitting can even develop under mud left on and 9. Stansbury EE and Buchanan RA: Fundamentals of of Mild Steel,” paper NACE 2462, presented at the NACE Corrosion 2013 Conference and Exhibition, inside drillpipe, where pipe storage racks contact Electrochemical Corrosion. Materials Park, Ohio, USA: ASM International, 2000. Orlando, Florida, USA, March 17–21, 2013. the pipe and at crevices. Deep corrosion pits in Brondel D, Edwards R, Hayman A, Hill D, Mehta S 17. Kvarekval J: “Morphology of Localized Corrosion Attacks drillpipe can lead to the onset of fatigue failure. and Semerad T: “Corrosion in the Oil Industry,” in Sour Environments,” paper NACE 07659, presented at Oilfield Review 6, no. 2 (April 1994): 4–18. the NACE Corrosion 2007 Conference and Exposition, Drillpipe may be coated with epoxies or resins Nashville, Tennessee, USA, March 11–15, 2007. 10. Although the battery analogy is acceptable for to stop corrosion, but the harsh downhole envi- explaining corrosion involving two dissimilar metals, 18. Pipeline operators send mechanical devices called pigs Oilfield Review through pipelines to clean the inner surface. This can be ronment often quickly removes these protective corrosion processes also take place on single metals. In single metals, the mechanism for corrosionMAY consists 16 done without halting flow, and the flow stream pushes coatings. Pipe dope, lubricating grease applied of small crystals with slightly different compositions.Corrosion Fig 4 the pig through the piping. to threaded connections, may help prevent cor- The anode and the cathode are located on differentORMAY 16 CRSSN19. Popoola 4 et al, reference 8. areas of the metal surface and, depending on the Chillingar et al, reference 3. rosion of these connections. conditions, may be close to each other or far apart. 20. Chillingar et al, reference 3. 11. Heidersbach R: Metallurgy and Corrosion Control in Care must be taken when using NH HSO as an oxygen Oil and Gas Production. Hoboken, New Jersey, USA: 4 3 scavenger. This compound is corrosive in itself and can Corrosion Form and Appearance John Wiley & Sons, Inc., 2011. also act as a food source for bacteria, thereby The word corrode comes from the Latin corrodere 12. At elevated temperatures, magnetite [Fe3O4] may potentially encouraging MIC. also form. meaning to gnaw; it can carry the additional 21. Davis JR: Corrosion—Understanding the Basics. 21 13. In MIC, the H2S is produced as a byproduct of the Materials Park, Ohio: ASM International, 2000. meaning of eat or wear away gradually. activities of sulfate reducing bacteria (SRB). Corrosion typically leaves a visible signature that 14. Chillingar et al, reference 3. is characteristic of the agent and mechanism

May 2016 37 Water Flow

Anode

Steel Cathode

General or Uniform Galvanic Erosion or Flow Induced Crevice

Water

Steel Force

Stress Pitting Intergranular Corrosion Cracking Corrosion Fatigue

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.

that caused it. Although not an exclusive list, cor- the potential for rapid growth. Localized corro- aged by cleaning internal piping surfaces, for rosion usually falls into one or more of the fol- sion, of which even CRAs such as stainless steels example, with the use of pipeline pigs. lowing categories: general or uniform, localized, are susceptible, can be subdivided into pitting, Galvanic corrosion can be a problem when galvanic, erosion or flow induced, crevice, pitting, crevice and under deposit corrosion. Pitting ulti- two dissimilar metals are in contact. The metal under deposit, cavitation, intergranular, stress mately can cause holes in metal components and that has the least resistance to corrosion acts as cracking and corrosion fatigue (Figure 5). Other is one of the primary causes of failure in oilfield the anode and the more resistant metal serves as types of corrosion include environmental, top-of- hardware, including tubing, casing, sucker rods the cathode. The anode typically corrodes pref- line and microbial. Based on the observed char- and surface equipment. erentially. This form of corrosion is frequently acteristics of the corrosion, engineers can adopt Crevice corrosion occurs in constricted areas, observed in offshore platforms and pipelines. appropriate preventive and mitigation measures. wherein the metal at the crevice becomes anodic The galvanic series, which orders metals accord- Uniform corrosion is typical of low-alloy and the rest of the metal serves as the cathode. ing to their anodic or cathodic tendencies, is a steels and may be observed over an entire The crevice can form where two dissimilar metals good predictor of corrosion severity (Figure 6). exposed area. Initial evidence of uniform corro- come into contact or be created by microgalvanic Galvanic corrosion is controlled and mitigated by sion is surface roughness. The metal becomes cells that may occur in certain steel alloys. use of the following: thinner as the corrosion progresses, and it will Pitting corrosion rates are often much higher • good engineering design—to ensure that cor- eventually fail from internal pressure or external than those of other types of corrosion. Inhibitors rosively active components present larger sur- forces. Because this type of corrosion is linked to may be appliedOilfield to the Review surface to prevent initia- face area than do less active components MAY 16 surface exposure, it may be prevented by prop- tion, but once Corrosiona pit has formed Fig 5 the inhibitors are • material selection—to avoid metals far apart erly protecting the surface. Uniform corrosion often unable toORMAY slow its 16 growth. CRSSN 5 in the galvanic series may occur in equipment used for oilfield opera- Under deposit corrosion occurs when sand, • isolation—to provide pipelines coming from the tions such as hydraulic stimulation and acidizing. corrosives or porous solids adhere to the metal sea with sacrificial anodes and protect those Localized corrosion occurs at specific sites surface. Although the area underneath the going into land with impressed current systems rather than over a generalized area and may be deposit is resistant to inhibitors and can corrode • inhibitors and coatings—to control initiation more dangerous than some other types of cor- quickly, this type of corrosion can often be man- of corrosion, although this method may be inef- rosion because of its unpredictable nature and fective once corrosion forms.

38 Oilfield Review Flow-induced corrosion occurs when liquid Pipe 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 Wet gas 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. Condensate

Monoethylene glycol Anodic Magnesium Zinc Figure 7. Top-of-line pipeline corrosion. Top-of-line corrosion can Cadmium result from the stratified multiphase flow of wet gas in horizontal Aluminum pipelines. Liquids—including condensate and inhibitors such as monoethylene glycol—settle to the bottom of the pipe. Wet gas fills Steel the pipe above the liquid line. If either CO2 or H2S are present in the Chromium steel gas, along with water, corrosive byproducts form at the top of the Stainless steel pipe and may not be controlled if the inhibitor remains at the bottom Lead of the pipe. Tin Nickel Inconel Cavitation is caused by collapsing bubbles • hydrogen embrittlement—hydrogen enters Hastelloy that occur when the pressure changes rapidly in the metal matrix and weakens it Brasses flowing liquids. Over time, cavitation may cause • stress corrosion cracking—cracks form after Copper deep pits to form in areas of turbulent flow, espe- corrosion has attacked a surface Bronzes cially in pump impellers. Low-carbon steels are • sulfide stress cracking—a failure of the metal Monel susceptible; stainless steels are more resilient.22 Chromium steel caused by H2S. Silver Intergranular corrosion results from cor- Material selection—opting for materials that are Titanium rosive attacks at metal grain boundaries in the resistant to hydrogen embrittlement and sulfide Graphite form of cracks. The grain boundaries can become cracking—is the primary avoidance technique. Gold anodic with reference to the cathodic surround- Low-stress design practices and stress relief by Platinum ing surface, typically due to formation of chro- heat treatment are also commonly used, and pre- mium carbides or nitrides. Metal impurities can venting corrosion in components subject to stress Cathodic increase the effect, as can precipitates in the is another method. metal that form during heat treatments. When Pipelines are subject to top-of-line corrosion Oilfield Review chromium combines with nitrogen orMAY carbon, 16 (Figure 7). Water condenses at the top of the less free chrome is available locally forCorrosion corro- Figpipe 7 as the fluid inside cools. The corrosion rate sion protection, and cracks can form alongORMAY the 16 CRSSNdepends 7 on the condensation rate and concen- grain boundaries. Quenching—the rapid cool- tration of organic acids. Generally, this type of ing after heat treatments—may be effective in corrosion is controlled with inhibitors and pipe- Figure 6. Galvanic series. Metals (not all shown) reducing or eliminating intergranular corrosion. line insulation that reduces condensation. can be described by their anodic or cathodic Material selection—avoiding metals that are 22. Port RD: “Flow Accelerated Corrosion,” paper NACE 721, tendencies arranged in a galvanic series. When susceptible to this condition—is the most reli- dissimilar metals are connected electrically presented at the NACE Corrosion 98 Annual Conference, San Diego, California, USA, March 22–27, 1998. and submerged in an electrolyte, the anodic able method to preclude intergranular corrosion. 23. ASTM International: “Standard Practices for Detecting metal, rather than the cathodic metal, will Tests such as ASTM A262 can be used to evaluate Susceptibility to Intergranular Attack in Austenitic preferentially corrode. The rate of corrosion is susceptibility of materials to this mechanism.23 Stainless Steels,” West Conshohocken, Pennsylvania, a function of the separation between the paired Environmental cracking occurs when cor­ USA: ASTM International A262-15, 2015. metals in the galvanic series. The series shown here is for seawater; the order may change rosion coincides with tensile stress. It may be based on the electrolyte. manifested as the following:

May 2016 Oilfield Review 39 MAY 16 Corrosion Fig 6 ORMAY 16 CRSSN 6 sion by taking a role in the formation of cathodic and conductive corrosion products on the metal surface (Figure 9). Sulfate producing prokaryotes Separator Tank (SPPs) are the chief offenders. Prokaryotes are microbes that have no cell nucleus or membrane- bound organelles. The most prominent group of Emulsion Sidestream SPPs are sulfate reducing bacteria (SRB) and sulfate reducing archaea (SRA). They contribute to corrosion by various means, for example, by Water outlet Oil outlet 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 Water Oil layer also lead to sour corrosion. Corrosion Biofilms may develop local concentration inhibitor film cells that are created by oxygen depletion or may Iron sulfide particle attach to a metal surface. Microbes can contrib- ute to corrosion by the direct effects of meta- Figure 8. Microbiologically influenced corrosion products. Softscale bolic waste products such as organic acids that corrosion, referred to as schmoo (right), can form in production systems are capable of altering the local pH and forming if microbes are not controlled. The photograph shows a mixture of iron sulfide [FeS], asphaltenes and biomass that was collected at the pH cells. Some microbes are anaerobic and can sidestream outlet of a separator tank (top left). Corrosion inhibitors form tolerate extremes of pressure, temperature, pH protective films around iron sulfide particles (bottom left) inside the and fluid salinity. These include methanogens— separator and prevent softscale formation in produced waters. microbes that produce methane as a metabolic Microbiologically Influenced Corrosion embedded in or on an attaching surface.25 Wher- byproduct in anoxic conditions. From the moment of immersion in a nonster- ever biofilms are found in producing systems, Regardless of the source of MIC, prevention ile fluid that supports microbial development, MIC can occur, including inside production tub- measures in most cases attack planktonic and microorganisms begin to attach to the material ing, gravity and hydrocyclone separators, storage sessile populations.26 Methods include biocides surface (Figure 8). Planktonic microorganisms tanks, pipelines and water injection systems. to kill the microorganisms, coatings to inhibit become fixed at the fluid boundary, typically a Depending on the microbial species, the corro- biofilm formation, removal of nutrients from the pipe wall, or onto porous media such as preexist- sion mechanisms can take various forms. flow stream to control microbial populations and ing corrosion.24 Biofilms can trap ions and create localized mechanical removal of an established biomass Attachment of microorganisms leads to the electrochemical potentials analogous to a gal- via pigging. formation of biofilms—microbial communities vanic corrosion cell or may contribute to corro- Corrosion Control Methods Metallurgical 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 Oilfield Review the world to attempt such a task. Aluminum is MAY 16 a corrosion-resistant metal used in many oilfield Corrosion Fig 9 applications; however, it is unsuitable for high- ORMAY 16 CRSSN 9 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 Figure 9. Microbiologically influenced corrosion from byproducts. Biofilm on meet both strength and cost requirements. the surface of this metal piece produced H2S that damaged the piece and led to premature failure of the equipment.

40 Oilfield Review

Oilfield Review MAY 16 Corrosion Fig 8 ORMAY 16 CRSSN 8 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. Galvanic DC current 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 milliamp 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 Electrolyte include epoxies, phenolic resins, polyurethanes, Anode 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 Cathode Backfill 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. Figure 10. Cathodic protection circuit. Cathodic protection methods may Cathodic protection (CP) consists of two pri- use naturally occurring galvanic current or employ a direct current (DC) mary forms: passive and active (Figure 10). In source (impressed current) when the electrolyte is resistive. The protected either form, it relies on a movement of electrons element—a pipeline is shown—is the cathode. The sacrificial element, located some distance from the cathode, serves as the anode. The DC (current) from an external anode to the equip- source may be batteries or solar panels in remote pipeline applications. 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- nails—led to the corrosion. Davy and his assis- Because the direct current (DC) is exter- tures, piping, casing and equipment exposed to tant carried out a number of experiments on cor- nally applied, this type of corrosion manage- the elements. It may also be installed inside or rosion prevention techniques; that assistant was ment is referred to as impressed cathodic outside tanks and pressure vessels. Michael Faraday, who would later establish the protection. It is most frequently used for cases Operators often use sacrificial anodes with relationship between the chemical action of cor- in which the electrolyte resistance is high, such CP to protect structures in areas where electrical rosion and electric current. as in soil or freshwater, and where a constant power sources are not readily available such as In the oil field, CP was first applied to land- 24. Stipanicˇev M: “Improved Decision Support Within in remote operations or on offshore structures. If based pipelines, and the first documented use Biocorrosion Management for Oil and Gas Water the structure can be made to serve as the cathode was by Robert J. Kuhn in 1928.29 He established Injection Systems,” PhD thesis, Institut National Polytechnique de Toulouse, France (2013). in relation to an anode, the disposable sacrificial a negative 850 mV potential between the steel 25. Madigan MT, Martinko JM, Bender KS, Buckley DH, anode will corrode while the cathode remains pipe of a pipeline and a copper-sulfate electrode. Stahl DA and Brock T: Brock Biology of Microorganisms, 14th ed. San Francisco: Benjamin Cummings, 2014. unscathed. This type of CP has been referred to This example became the foundation ofOilfield modern Review 27 26. Sessile refers to fixed or immobile organisms. as fighting corrosion with corrosion. CP technology, although for many years theMAY effec 16- 27. Lehmann JA: “Cathodic Protection of Offshore The first use of CP is attributed to Sir tiveness was met with scientific skepticism.Corrosion Fig 10Structures,” paper OTC 1041, presented at the ORMAY 16 CRSSNFirst Annual 10 Offshore Technology Conference, Houston, Humphry Davy, who described the process in a Today, CP uses sacrificial elements made May 18–21, 1969. series of articles to the Royal Society of London from aluminum, zinc and magnesium to protect 28. Davy H: “On the Corrosion of Copper Sheeting by Sea in 1824.28 The technique was used in an attempt the steel of large structures and piping. These Water, and on Methods of Preventing This Effect; And on Their Application to Ships of War and Other Ships,” to prevent the corrosion of nails used in wooden dissimilar metals create the galvanic coupling Philosophical Transactions of the Royal Society of oceangoing vessels. Accelerated corrosion of the that establishes a current path between the London 114 (January 1, 1824): 151–158. 29. von Baeckmann W: “The History of Corrosion nails occurred when copper cladding—used to anode and the cathode, and, over time, the Protection,” in von Baeckmann W, Schwenk W and prevent biofouling—was applied to the outside sacrificial anode rather than the protected Prinz W (eds): Handbook of Cathodic Corrosion Protection—Theory and Practice of Electrochemical of vessels. Davy found that sacrificial anodes structure experiences metal loss. Appropriate Protection Processes, 3rd ed. Houston: Gulf Professional protected the iron nails. The actual processes placement and distribution of the anodes is cru- Publishing (1997): 1–26. were not well understood at that time, but it is cial to ensure that all parts of the structure are 30. Amani M and Hjeij D: “A Comprehensive Review of 30 Corrosion and Its Inhibition in the Oil and Gas Industry,” recognized today that the contact of dissimilar sufficiently protected. paper SPE 175337, presented at the SPE Kuwait Oil metals—the copper cladding and the iron of the and Gas Show and Conference, Mishref, Kuwait, October 11–14, 2015.

May 2016 41 Reactive inhibition operates at the cathode Water Water level for the corrosion cell. The cations of the inhibitor react with the cathodic anions to form Oil Oil insoluble films, which adhere to the surface of the metals and prevent O from coming into CH3 2 contact with the metal. These films also prevent

the evolution of H2, a byproduct of the corrosion Alkyl cell. Examples are forms of calcium carbonate, chain CnH2n magnesium carbonate and iron oxides. Reactive inhibitors can also serve as poisons to the corro- N+ Polar head sion cell process by interfering with the forma- group tion of H2 and reducing the reaction rates at both Metal surface the cathode and anode. Vapor phase inhibitors are primarily used for

Figure 11. Film formers. Although they vary in composition and avenue of combating CO2 corrosion. These inhibitors neu- protection, film formers create barriers between corrosive elements (water tralize CO2 and block the formation of carbonic 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 acid [H2CO3]. They are transported via vapor metal (polar head group, bottom). When molecules of the polar head group phase in wet gas lines. To protect against future of film formers attach to the surface of the metal, a portion of the molecule corrosion, they may also be used during hydro- extends into the fluid. This usually oil-soluble tail is hydrophobic, repelling static testing of components with water, espe- water away from the metal surface. cially when the components are to be stored after fitness testing. Examples of these types of inhibi- tors include morpholine and ethylenediamine. source of current is readily available. The use is to interrupt the electrochemical process by of solar panels in remote locations has greatly which the corrosion cell forms between the metal Film Formers increased the potential applications of impressed and the liquids in and around the equipment. Film formers are the most widely used corrosion cathodic protection. Inhibitors can be a flexible and cost-effective inhibitors in the oil and gas industry. They cre- In the impressed CP technique, current of method of fighting corrosion, and the inhibi- ate a continuous layer between the metal and the several amps from a low-voltage rectifier passes, tor application can be altered when conditions reactive fluids, thus reducing the attack of corro- or is impressed, from an inert anode (for exam- change. Although acquiring and delivering the sive elements (Figure 11). They may also attach ple, graphite or iron) to the structure being pro- inhibitor incur an ongoing cost, the lower costs to the surface of corroded metal, altering it and tected, which acts as the cathode. The anode associated with using less corrosion resistant reducing the corrosion rate. Although they are is attached to the positive terminal of the DC low-carbon steels usually more than make up effective in reducing CO2 and H2S corrosion, film source, and the cathode is attached to the nega- the difference. formers are not effective against O2 corrosion. tive terminal. The anode and cathode are often Inhibitors fall into four main categories: scav- Film formers are available in oil-soluble, some distance from each other, separated by engers, reactive agents, vapor phase and film water-soluble and oil soluble–water dispersible an electrolyte. formers. Oxygen scavengers are frequently used forms. Oil-soluble inhibitors are used to treat oil- To counteract corrosion, sufficient current in operations in which oxygen poses a corrosive and gas-producing wells. Water-soluble inhibi- density must be supplied to all parts of the pro- threat. These agents not only reduce oxidizing tors are used in high water-cut flow streams, tected structure and the current density must corrosion, but also control the growth of microbes including those found in producing wells, trans- always exceed what would be the measured cor- that require oxygen to thrive. Examples of oxygen mission lines and separators. Oil soluble–water rosion rate under the same conditions.Oilfield If the Reviewscavengers used in the oil and gas industry are dispersible inhibitors are used in oil and gas corrosion rate increases, the impressed MAYcurrent 16 sodium sulfite, sulfur dioxide, sodium bisulfite, wells that are also producing water. density must be increased.31 Although theCorrosion initial Figsodium 11 metabisulfite and ammonium bisulfite. Film-forming inhibitors take various chemical equipment cost may be higher for impressedORMAY CP 16 AmmoniumCRSSN 11 bisulfite and sodium bisulfite are forms but are typically composed of long carbon than it is for sacrificial protection, this technique commonly used in seawater injection systems. To chains with nitrogen, phosphate esters or anhy- may be less expensive over the long term because speed reaction rates, a catalyst may be included drides. Inhibitors may adhere to or be adsorbed sacrificial anodes do not need to be replaced. in the chemical. on the metal surface, which prevents the corro- Impressed CP also has the advantage of provid- Hydrogen sulfide scavengers reduce the sives from attacking the metal. The most effec- ing information to the operator about the extent level of H2S in the flow stream. Examples of H2S tive film-forming inhibitors create a molecular of corrosion over time. scavengers are amines, aldehydes and zinc car- bond at the metal surface in a process of charge boxylates. Common forms of amines are mono- sharing or charge transfer. For effective inhibi- Corrosion Inhibitors ethanolamine (MEA) and monomethylamine tion, the surface of the metal being protected Another line of defense against corrosion is (MMA) triazine. In some situations, operators must be fully covered; injection of the proper inhibitors, of which there are a variety of types may be able to regenerate MEA and MMA for concentrations of the inhibitor are crucial. After and applications. The primary goal of inhibitors reinjection and reuse. they interact with the corrosive elements, some

42 Oilfield Review inhibitors are gradually removed from the metal Use of corrosion surface and must be continuously replenished inhibitor recommended 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. Predict corrosion rate Amides and imidazolines are examples of organic from field data film-forming inhibitors that are effective over a wide range of conditions, especially in sweet

(CO2) and sour (H2S) gas corrosion environments. Develop and execute They can be water or oil soluble. Amines, which testing program 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 Rotating Cylinder Kettle Test Autoclave Test surface of the material to be protected, they dis- Electrode Test rupt the normal corrosion cell charge. However, at least one study indicated that quaternary Recommend and ammonium inhibitors may actually increase the implement corrosion corrosion rate of sweet corrosion in the presence inhibitor addition 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, Conduct field trial ester quats, dimer and trimer acids and alkyl pyr- idine quaternary compounds. Most film-forming applications include multiple inhibitors; labora- Sidestream Test tory testing is used to establish optimum concen- Analyze field trations, fluid tolerances, stability, effectiveness trial results and persistency of the film. Inhibitor selection can be a complicated process and typically must be adjusted over time to meet the demands of Make final changing fluid conditions. recommendation

Inhibitor Selection Figure 12. Laboratory testing of corrosion inhibitors. Operators usually Laboratory evaluation is the key to developing develop corrosion control plans and then test inhibitors using conditions an effective program in inhibitor selection for expected from the field. This flowchart follows a testing sequence. Three corrosion control. Technicians begin the process common testing methods are the rotating cylinder electrode, kettle and autoclave tests (middle). Even after laboratory testing, field trials should be using fluid samples that replicate field condi- conducted to validate the effectiveness of the program. A sidestream test tions—actual produced fluids are best if avail- (lower left) acquires samples for analysis. If the proposed method provides able. Simulated and synthetic fluids are used acceptable results, the method is adopted, although the corrosion inhibition program must be reevaluated during the life of the well. 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- 31. Schweitzer PA: Corrosion of Linings and Coatings: Carbon Steel under Water and Its Inhibition by a ing corrosion inhibitors include the following Cathodic and Inhibitor Protection and Corrosion Quaternary Ammonium Salt,” paper NACE 05307, Monitoring. Boca Raton, Florida: CRC Press, 2006. presented at the NACE Corrosion 2005 Conference and tests: wheel, kettle (also called linear polariza- 32. Binks BP, Fletcher PDI, Hicks JT, Durnie WH and Exhibition, Houston, April 3–7, 2005. tion resistance (LPR) tests), rotating cylinder Horsup DI: “Comparison of the Effects of Air, Carbon 33. Binks et al, reference 32. Dioxide and Hydrogen Sulphide on CorrosionOilfield of a Low Review MAY 16 Corrosion Fig 12 ORMAY 16 CRSSN 12

May 2016 43 electrode, autoclave, jet impingement and flow Rack 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 Metal coupon container field conditions. The samples are tested with and without inhibitor and the results are compared. High-temperature autoclave The kettle test, or LPR test, measures corro- Pressure source sion rates electrochemically. Metal electrodes (hydraulic pump) 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 Figure 13. Autoclave corrosion testing. A high-temperature autoclave is sources, and high agitation replicates turbulent used to test the effectiveness of corrosion inhibitors on metal coupons (inset). Hydrostatic pressure and temperature can be applied to simulate fluid flow that has dispersed hydrocarbons. To downhole conditions. 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

Thermometer increased inhibitor volumes. Linear polarization is performed by controlling the voltage potential and measuring the current then controlling the Gas sparge ltage potential, mV Vo Voltage Corrosion rate Steel electrode Current

Heating mantel Current, mA

Stir bar

Figure 14. Kettle test. To perform kettle tests, or linear polarization resistance and measuring the current then controlling the current and measuring tests, technicians use a test fixture (left) and control the pressure and the voltage. The electrolyte can be agitated using the stir bar. Gas can be temperature. They submerge electrodes inside the fixture into the fluids injected into the test fixture, a process referred to as sparging. From the expected downhole and then measure electrical properties of the slope of the polarization resistance curve (right), the corrosion rate can electrodes. The tests are performed by controlling the voltage potential be computed.

44 Oilfield Review Cape Lopez Cathodic protection station Section 1 Section 2 Section 3

Tchéngue Cathodic protection station

ATLANTIC OCEAN GABON

Batanga Cathodic protection stations Cathodic protection station powered by two solar cells

Input temperature ~ 60˚ C Rabi Field Cathodic protection station

Figure 15. Corrosion in a pipeline from the Rabi field to Cape Lopez. A of the protective outer covering of the pipeline. Engineers concluded three-section, 18-in. pipeline carries oil from the inland Rabi field in Gabon that corrosion observed in Section 1 resulted from a combination of the to Cape Lopez on the coast. Cathodic protection stations are located along disbonding of the protective covering and ineffective cathodic protection. the pipeline. Because the incoming oil is hot (around 60°C), Section 1 of the Although the pipeline’s safety was not compromised, the operator pipeline (red and dark blue) is exposed to a higher temperature than is the implemented new procedures to prevent the corrosion from recurring. remainder of the pipeline. The elevated temperature led to the disbondment current and measuring the voltage potential. • thermal stability 100 km [62 mi] and then Section 3 from Tchengué The data are plotted, and the slope of the line • emulsification tendency to Cape Lopez, 29 km [18 mi] (Figure 15). is the polarization resistance, which is inversely • foaming tendency The inlet pressure at Rabi was about 40 bar proportional to the corrosion rate (Figure 14). • metal compatibility [580 psi], and the flowing temperature was 60°C This technique provides corrosion rate evalua- • elastomer compatibility [140°F] at the inlet. Beyond the inlet, the line oper- tion from external measurements, whereas other • compatibility with other chemicals used in the ates at about 35°C [95°F]. Impressed cathodic pro- methods require technicians to physically mea- same stream. tection is used for the pipeline, which has sections sure and evaluate corrosion. Application methods should be evaluated as well. that have solar cells to provide current. The pipeline The effectiveness of inhibitors is dependent Injection may be continuous, batch or squeeze. was coated with three-layer polyethylene; each joint on fluid velocity. For fluids containing little or The rate of film removal is a key concern when was brush cleaned and wrapped with heat-shrink no solid particles, high flow rates can lead to determining the optimal application mode. 34. NACE Task Group T-1D-34 on Laboratory Corrosion flow-accelerated corrosion. If the flow stream Inhibitor Test Parameters: “Laboratory Test Methods for contains solid particles, the accelerated corro- Corrosion in the Oil Field Evaluating Oilfield Corrosion Inhibitors,” Houston, NACE International, NACE Publication 1D196, December 1996. sion is termed erosion corrosion. Several test A recent example of pipeline corrosion from 35. NACE Task Group T-1D-34 on Laboratory Corrosion methods have been developed to model corrosion Gabon illustrates the need for thorough testing Inhibitor Test Parameters, reference 34. in high-flow conditions and determine a film’s and understanding of the corrosion process.37 A 36. Efird KD: “Jet Impingement Testing for Flow Accelerated Corrosion,” paper NACE 00052, presented at NACE persistence, especially where turbulent flow is pipeline transports oil from the Rabi field to Cape Corrosion 2000 Conference and Exhibition, Orlando, present.36 Test methods include jet impingement, Lopez—a distance of approximately 234 km Florida, March 26–31, 2000. 37. Melot D, Paugam G and Roche M: “Disbondments of rotating cylinder electrodes and flow loop testing. [145 mi]. The 18-in. pipeline comprises three Pipeline Coatings and Their Effects on Corrosion Risks,” The testing of inhibitors should determine sections: Section 1 from Rabi to Batanga, 105 km Journal of Protective Coatings and Linings 26, no. 9 the following: [65 mi]; Section 2 from Batanga to Tchengué, (September 2009): 67–76.

May 2016 45 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 Deepwater Deepwater projects can pose unique challenges for corrosion control because the completions are usually located at the seafloor and flowlines Figure 16. An example of pipeline corrosion. After the protective coating must come to the surface or back to shore. A disbonded on a pipeline in Gabon, corrosion formed as pitting (inset). deepwater field located in the southern Niger delta demonstrates the use of inhibitors to com- 39 sleeves and hot-melt adhesive that overlapped the disbondment occurred in the sleeves. The main bat CO2-induced corrosion (Figure 17). polyethylene.38 The pipe was buried in wet, com- difference between the sections that had differ- The production path for deepwater wells pacted sand that had a pH of approximately 5.4. ing corrosion levels was that the temperature in passes through cold water, which can subject the Problems began to develop in the first 15 km the more corroded sections was higher. Further originally hot fluids in the flow stream to rapid [9.3 mi] of pipe. Routine inspections found dis- testing of pipe Sections 2 and 3 found no evidence cooling. Conversely, inhibitor injection is often bondment at the sleeves where they overlapped of similar levels of disbondment or corrosion. through long umbilicals that are subjected to the polyethylene coating in the Rabi section. After a thorough examination, engineers recom- temperature contrasts between the surface and Disbondment allowed water under the protective mended abrasive blast cleaning prior to applying subsea wellheads. Injection of inhibitors is fur- coating, which negated the cathodic protection heat-shrink sleeves for future installations rather ther complicated by the normally high flowing and allowed corrosion to develop (Figure 16). than the standard brush cleaning of connections. pressures associated with deepwater production. The remainder of the pipeline did not experience Another possible solution was liquid polyurethane Temperature extremes, pressure extremes the same level of corrosion, although significant or epoxy applied at the joints. The disbonded coating and long umbilicals combine to affect inhibitor

BURKINA FASO FPSO vessel BENIN NIGERIA COTE GHANA D’IVOIRE CAMEROON Niger Delta field

GABON Gulf of Guinea

Subsea wellheads Flowlines and umbilicals

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.

46 Oilfield Review stability, performance and properties. Thorough Inhibitor Dose Rate, ppm Uninhibited Inhibited Protection, % testing of the inhibitors is required to ensure Corrosion Rate, mpy Corrosion Rate, mpy corrosion controlling properties are maintained, DS-1617 inhibitor 10 173.01 4.18 97.58 that the injected chemicals remain stable and that the inhibitors can be reliably delivered via DS-1617 inhibitor 20 156.43 0.98 99.37 the umbilicals into the flow stream. Inhibitor Dose Rate, ppm Corrosion Rate, mpy Protection, % Another risk in deepwater production is the None — 71.04 — formation of hydrates—ice-like solids of water and gas that form above the normal freezing point DS-1617 inhibitor 20 1.16 98.37 of water—that can plug flowlines. To ensure cor- Figure 18. Corrosion testing of the DS-1617 inhibitor. Technicians conducted kettle tests with fluids rosion controlling properties are maintained, representative of field conditions to evaluate the effectiveness of the DS-1617 inhibitor (top). They inhibitors must be thoroughly tested to confirm also performed autoclave testing at high temperature (bottom). The corrosion rate is in milli-inches of that the injected chemicals remain stable and penetration/year (mpy). that the inhibitors can be reliably delivered via umbilicals into the flow stream. Engineers developed the DS-1617 deepwater and to pressurized CO2 heated to 120°C [248°F]. These conditions were faced by an operator corrosion inhibitor to meet the challenges of The results indicated a 98% reduction in the of a deepwater production platform in Nigeria. this facility. corrosion rate.41 The 20 ppm concentration The platform served nine wells drilled in water To qualify this inhibitor, they tested the yielded corrosion rates of about 0.00016 in./year depth of 1,030 m [3,380 ft]. The operator used chemicals in accordance with the API TR 17TR6 [0.004 mm/year]. For corrosion rate, the stan- subsea completions that included five manifolds standard, which requires replicating the temper- dard industry units are milli-inches/year, or mpy. and eight production flowlines and risers. The atures and pressures experienced by the inhibitor For this test, the corrosion rate was equivalent to flowlines were connected to a floating produc- during deployment through the umbilicals.40 The 0.16 mpy. Test technicians reported no foaming tion, storage, and offloading (FPSO) vessel that evaluation included high-pressure flow-loop sta- problems associated with the inhibitor. had 320,000 m3 [2 million bbl] of onsite storage bility tests. The engineers conducted additional The operator adopted the use of the DS-1617 capacity. Produced oil flowed to the FPSO for tests to look at resistance to hydrate formation, inhibitor and monitored corrosion at six loca- transfer to tankers. Produced gas was directed to thermal aging and compatibility with seawater. tions on the FPSO. No corrosion monitoring shore via pipelines. Because the operator was concerned about foam- was installed on the deepwater flowlines. The The pipelines used to transport the oil and ing in the glycol regeneration unit, the inhibitor DS-1617 inhibitor was injected at a 100-ppm gas were constructed of carbon steel. The flow- was tested for foaming tendency. rate, which is a lower rate than the initial inhibi- ing pressure from the wells averaged 80 bar Laboratory technicians performed kettle tor that was deemed insufficient. Criteria for [1,160 psi], and the average temperature was tests using the DS-1617 inhibitor at 20 ppm, corrosion protection established by the opera- 85°C [185°F]. The water cut was 45% and the which is a relatively low dosage; the corrosion tor was a rate below 0.05 mpy [0.0013 mm/year]. natural gas contained about 1.4% CO2. The com- rate was reduced by 99% (Figure 18). They also Testing at all six locations indicated corrosion bination of produced water (brine) and CO2 performed high-temperature autoclave testing rates below the target rate (Figure 19). Based on presented a high corrosion-rate potential for on carbon steel coupons. The samples were sub- the testing, the operator implemented the use of the low-carbon steel. In wet gas pipelines such jected to test fluids that had corrosion inhibitor the DS-1617 inhibitor. as these, produced water has a tendency to con- 6.0 dense at the top of the pipe, allowing top-of-line Low-pressure separator A corrosion; the presence of both water and CO Low-pressure separator B 2 5.0 accelerates corrosion. Bulk oil treater Engineers installed chemical umbilicals of Target corrosion rate 1 to 20 km [0.6 to 12.4 mi] to inject corrosion 4.0 inhibitor into the deepwater production flow- lines. As the project progressed, engineers at 3.0 M-I SWACO, a Schlumberger company, reevalu- ated the initial inhibitor used in the project, Corrosion rate, mp y 2.0 which also protected the topside piping and storage vessels, and deemed it to be insufficient. 1.0 38. Roche M: “The Problematic of Disbonding of Coatings and Corrosion with Buried Pipelines Cathodically Protected,” presented at the 10th European Federation 0.0 of Corrosion, Nice, France, September 12–16, 2004. 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 6.0 39. Jenkins A: “Corrosion Mitigation in a Deepwater Time, hr Oilfield Case Study,” paper IBP1194_15, presented at the Rio Pipeline Conference and Exposition, Rio de Janeiro, Figure 19. Corrosion monitoring at a production facility. The operator injected DS-1617 inhibitor into September 22–24, 2015. the flowlines of producing wells using underwater umbilicals. The corrosion rate of the flowlines was 40. API: “Attributes of Production Chemicals in Subsea monitored at the low-pressure separator A (blue), low-pressure separator B (red) and the bulk oil Production Systems,” Washington, DC: API, API Technical Report 17TR6, 2012. 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 41. Jenkins, reference 39. test sites for the duration of the testing period.

May 2016 47 New Developments in Corrosion Control Anode Water Depth, m Weight Loss, % Controlling corrosion has been an ongoing 1 13 13 battle between humans and nature for millen- 2 73 31 nia. Since scientists such as Sir Humphry Davy 3 116 25 and Michael Faraday discovered some of the 4 116 39 underlying physics that explained corrosion, various methodologies have been adopted and Figure 20. Anode corrosion after eight years of service in the North Sea. adapted. Modern scientific understanding and new technologies are combining to improve the tools available to fight the unending battle with corrosion. North Sea Cathodic Protection determined changes in physical dimensions and One area of emerging materials science is North Sea production platforms routinely use measured electrical properties. Four anodes nanoparticles and nanostructures.43 Having sur- cathodic protection. On one platform, the opera- were analyzed for the study. The reduction of face thickness of 1 to 100 nm, these coating mate- tor installed 10 sacrificial anodes below the sur- anodes that had been placed in deeper water was rials have unique properties that may make them face of the water and left them in place for eight greater than that of those placed in shallower almost impervious to corrosion. Nanoparticles years. The anodes were composed of zinc, silver water. Some of the anodes were so corroded that and nanostructures may be deposited on metal and silver chloride and were located at various visual inspection was difficult (Figure 21). surfaces as films, similar to film-forming tech- depths and locations on the plaform.42 The system The original 20-year design projected that at niques, but because of nanoparticles’ greater was designed to protect the structure for a mini- eight years, the anodes should be reduced by 40%; persistence, reapplying them is unnecessary. The mum of 20 years. Engineers monitored the output however, the average weight loss of the anodes surfaces also become super-slick—exhibiting current from three of the anodes over the period. was only 24%. The engineers concluded that the low friction coefficients—which reduces wear The anodes were removed and inspected at the original design, although conservative, would and increases durability. Such surfaces are also end of eight years. protect the structure for at least 20 years. Based less likely to experience biofouling.44 After retrieval, the sacrificial anodes were on the results of the study, a model was estab- The battle against corrosion will never cleaned and weighed (Figure 20). Technicians lished for periodic inspections to be performed. 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

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.

42. Roche M: “Offshore Cathodic Protection: The Lessons of 44. Tesler AB, Kim P, Kolle S, Howell C, Ahanotu O and Long-Term Experience,” paper OMC-2005-020, presented Aizenberg J: “Extremely Durable Biofouling-Resistant at the 7th Offshore Mediterranean Conference and Surfaces Based on Electrodeposited Nanoporous Exhibition, Ravenna, Italy, March 16–18, 2005. Tungstite Films on Steel,” Nature Communications 6, 43. El-Meligi AA: “Nanostructure of Materials and Corrosion no. 8649 (October 20, 2015). Resistance,” in Aliofkhazraei M (ed): Developments in Corrosion Protection. Rijeka, Croatia: InTech (2014): 3–23.

48 Oilfield Review Contributors

Nausha Asrar is the Manager for Materials Support Bruce MacKay is Client Support Manager for the Marko Stipanicˇev is Corrosion Discipline Lead for and Failure Analysis at the Schlumberger Houston Schlumberger North America fracturing and cement- Schlumberger Production Technologies in Bergen, Pressure and Sampling and Formation Evaluation ing operations in Sugar Land, Texas. He has worked Norway. Upon graduation from the University of Centers in Sugar Land, Texas, USA. He began his as a chemical problem solver in various capacities Zagreb, Croatia, he worked as an external consultant career with Schlumberger in 2005 as a senior materi- for Schlumberger for 10 years, spanning the R&D on industry related projects at the Faculty of Chemical als scientist. He previously worked for Shell Global spectrum from research to product development to Engineering and Technology, in Croatia. Beginning in Solutions in the US, the Saudi Basic Industries technology implementation. He has authored 12 peer- 2010, he worked as a research engineer for Det Norske Corporation Technology Center and Saline Water reviewed scientific journal articles and five SPE papers Veritas in Bergen, investigating corrosion-based failures Conversion Corporation, both in Saudi Arabia, and and has been granted several patents on chemical and performing root cause analysis studies. He joined as principal corrosion engineer at the Research and technologies related to oilfield applications. He has M-I SWACO in 2013 as a corrosion specialist, working Development Center for Iron and Steel for the Steel been a speaker on the importance of chemistry in oil- in Bergen, and in 2015, he was named the corrosion Authority of India, Ltd. A NACE certified material field development to a variety of audiences, including discipline lead. Marko is responsible for Schlumberger selection and design specialist, Nausha is a member the US National Academy of Sciences, the American corrosion products, which include inhibitors, biocides, of NACE, ASM and SPE as well as a life member of the Chemical Society and the National Aeronautics and scavengers and nutrients. He has authored and coau- Indian Institute of Metals; he is the author of more Space Administration Jet Propulsion Laboratory. thored numerous papers and publications related to than 60 technical papers and reviews on corrosion, Bruce was a Natural Sciences and Engineering corrosion and corrosion management. He holds an MSc phase diagrams, composite materials and failure cases. Research Council of Canada postdoctoral research degree in chemical engineering and technology from He received an MS degree in chemistry from Aligarh scholar at the California Institute of Technology, the University of Zagreb and a PhD degree in environ- Muslim University, Uttar Pradesh, India, and a PhD Pasadena, USA. He earned a BS degree in chemistry mental process and biocorrosion management from the degree in materials science and engineering from the and a PhD degree in inorganic chemistry from the Université de Toulouse, France. Moscow State University. University of British Columbia, Vancouver, Canada. Jean Vittonato, is Head of the Total E&P Technology Øystein Birketveit is Technical Manager for Production Denis Mélot is a Nonmetallic Materials Expert with Division corrosion department in Pau, France. He Technologies for M-I SWACO, a Schlumberger company, the technology department of Total Upstream, in Paris. is responsible for the headquarters’ corrosion team in Bergen, Norway. For the past 18 years, he has Using his foundation of studies in polymer science, and provides technical assistance to projects and specialized in the field of corrosion. Prior to joining his focus is on nonmetallic materials and corrosion. operating subsidiaries worldwide. He started work in M-I SWACO, Øystein worked for Statoil and for Det Prior to beginning work with Total in 2003, he was a 1999 focusing on cathodic protection with COREXCO, Norske Veritas. He earned his MSc degree in materials researcher in the R&D department of Elf Atochem, an engineering cathodic protection company, where and electrochemistry from the Norwegian University of which is now Arkema, in Serquigny, France. He then he was in charge of designing cathodic protection Science and Technology, Trondheim. spent six years as the technical manager for pipe systems for both onshore and offshore and for installa- Joshua E. Jackson is the CEO of G2MT LLC as well as coating products with the company. Denis chaired tion, monitoring and maintenance follow-up. In 2006, the cofounder of G2MT Laboratories, LLC in Houston. the ISO 12736 working group on wet thermal insula- he joined Total as a corrosion specialist and was in G2MT Labs is a metallurgical consulting and analysis tion systems, was a member of pipeline coatings work charge of cathodic protection activities. He provided company that performs nondestructive materials char- group ISO 21809 and holds certifications from the support for projects for both Total E&P and operating acterization to evaluate residual stress mechanical Association pour la Certification et la Qualification subsidiaries and was in charge of research projects properties and other critical parameters including the en Peinture Anticorrosion and Faglig Råd for related to cathodic protection. He spent three years effects of corrosion. His scientific focus areas include Opplæring og Sertifisering av Inspektører innen in Republic of the Congo as the head of the Total cor- corrosion analysis, high-temperature materials, hydro- Overflatebehandling . He holds numerous patents in rosion department, where he supervised all projects gen absorption effects, failure analysis and statistics. his field and has coauthored several papers on the related to corrosion. Jean is a certified Cathodic Joshua is the coauthor of numerous papers in the field subject of coatings and corrosion. Denis has a degree Protection Specialist with the National Association of materials science covering subjects including non- in materials science from the École Universitaire of Corrosion Engineers and with the Centre Français destructive testing, metallurgy, welding, corrosion and D’Ingénieurs de Lille, France, and received his PhD de la Protection Cathodique and is chair of the hydrogen. He obtained BS degrees in both mathemat- degree in polymer science from Université de Lille. ISO TC 67 SC2 GW11 working group on cathodic ics and physics from the Massachusetts Institute of Jan Scheie is a Project Leader and Account Manager protection of pipelines. He obtained an engineer- Technology, Cambridge, USA, and MS and PhD degrees for Production Technologies (PT) in Schlumberger ing degree from Institut National Polytechnique de in metallurgical and materials engineering from the Norge A/S in Stavanger, where he serves customers in Grenoble, France. Colorado School of Mines, Golden, USA. Scandinavia. He has also been an account manager Alyn Jenkins, based in Aberdeen, serves as the Global for production technologies, an international sales Asset Integrity Manager for Schlumberger Production manager and an area manager for production chemi- Technologies. He manages asset integrity product cals in Stavanger. He has worked for M-I SWACO in market development for the eastern hemisphere, as lines that include corrosion inhibitors, biocides, H2S scavengers and oxygen scavengers and is responsible technical manager in the Middle East and CIS, as sales for research and development projects related to manager in South Asia and as principal engineer for corrosion. He began his career in 1998 with Clariant developing sales strategy in mainland Europe. He is a Oil Services in Aberdeen and then worked for Baker member of TEKNA, the Norwegian Society of Graduate Hughes in Liverpool, England. Alyn joined M-I SWACO Technical and Scientific Professionals, the SPE and in 2005 as a corrosion specialist in Stavanger and then the National Association of Corrosion Engineers. He served as lead integrity management specialist. Alyn received an MSc degree in chemical engineering from holds BS and MS degrees, both in chemistry, from the the Norwegian University of Science and Technology, University of Wales, Bangor. Trondheim, Norway, and an MBA degree from Thunderbird School of Global Management, Glendale, Arizona, USA.

May 2016 49 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 Duplantis Directional wells have been a boon to oil and This slight bend is sufficient for pointing the bit Houston, Texas, USA gas production, particularly in unconventional in a given direction yet is small enough to permit plays, where horizontal and extended-reach rotation of the entire mud motor assembly during Oilfield Review 28, no. 2 (May 2016). wells maximize wellbore exposure through pro- rotary drilling. This seemingly minor deflection Copyright © 2016 Schlumberger. For help in preparation of this article, thanks to ductive zones. In many of these wells, steerable determines the rate at which the motor builds Edina Halilagic and Brandon Mills, Houston. mud motors have been crucial to achieving the angle to establish a new wellbore trajectory. By Slider is a mark of Schlumberger. well trajectory necessary to hit operators’ target orienting that bend in a specific direction, called 1. Maidla E and Haci M: “Understanding Torque: The Key to zones. Directional drillers use a downhole mud its toolface angle, the driller can change the Slide-Drilling Directional Wells,” paper IADC/SPE 87162, presented at the IADC/SPE Drilling Conference, Dallas, motor when they kick off the well, build angle, inclination and azimuth of the well path. March 2–4, 2004. drill tangent sections and maintain trajectory. To maintain the orientation of that bend and A bend in the motor bearing housing is key thus change wellbore trajectory, the drillstring to steering the bit toward its target. The surface- must not be allowed to rotate, and this is where adjustable bend can be set between 0° and 3°. the mud motor comes into play. A mud motor is

Rotation

Mud flow

Wear pad 2.7 7 3.0 0 Protective 2.89 housing

Effective

bend 2

2

.7

.3

7

8

3

2

2

.0 .8

.6 Stator Rotor

0

9 0

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.

50 Oilfield Review 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 Kickoff 1 power to turn a drive shaft that causes the bit to rotate. Build section, 3°/100 ft 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 Tangent section 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 Marker bed straight drilling path (Figure 2). In most opera- Kickoff 2 tions today, measurement-while-drilling (MWD) tools provide real-time inclination and azimuth Build section, 3°/300 ft Lateral section measurements that alert the driller to any devia- Reservoir

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- Bent housing 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, Kickoff prompting the driller to initiate a second slide section to land the well horizontally within the reservoir.

tions from the intended course. To correct for Slide Drilling Challenges those deviations or to alter the wellbore trajec- To initiate a slide, the driller must first orient the tory, the driller switches from rotating to sliding bit to drill in alignment with the trajectory pro- mode. In sliding mode, the drillstring does not posed in the well plan. This requires the driller to Build section rotate; instead, the downhole motor turns the bit stop drilling, pull the bit off-bottom and recipro- and the hole is drilled in the direction the bit is cate the drillpipe to release any torque that has pointing, which is controlled by toolface orienta- built up within the drillstring. The driller then tion. Upon correcting course and reestablishing orients the downhole mud motor using real-time the wellbore trajectory needed to hit the target, MWD toolface measurements to ensure the speci- Rotary Sliding the driller may then switch back to rotating fied wellbore deviation is obtained. Following mode mode mode. (Figure 3). this time-consuming orientation process, the Of the two modes, slide drilling is less effi- driller sets the topdrive brake to prevent further cient; lateral reach usually comes at the expense rotation from the surface. The slide begins as the of penetration rate. The rate of penetration (ROP) driller eases off the drawworks brake to control Increased diameter achieved using conventional sliding methods the hook load, which, in turn, affects the mag- caused by outward tilt of bit typically averages 10% to 25% of that attained in nitude of weight imposed at the bit. Minor right rotating mode.1 Conversely, by rotating the entire and left torque adjustments are applied manually drillstring, drillers gain a substantial advantage to steer the bit as needed to keep the trajectory Figure 2. Drilling using a bent housing. In rotating in ROP. This article describes an automated sys- on course. mode, the bit carves a straight path parallel to the axis of the drillstring, which is also rotating. tem that helps drillers achieve significant gains As the depth or lateral reach increases, the Because the bent housing forces the bit to tilt in horizontal reach with noticeably faster rates of drillstring is subjected to greater friction and drag. outward by a few degrees, the bit drills a hole that penetration. Field experience in Colorado, USA, These forces, in turn, affect the driller’s ability to is slightly larger than the diameter of the bit. When illustrates how a torque-oscillation system can transfer weight to the bit and control toolface ori- the driller switches to sliding mode, only the bit rotates. The resulting hole is in gauge and follows help operators exploit unconventional plays. entation while sliding, making it difficult to attain the axis of the BHA below the bent housing.

May 2016 51 sufficient ROP and maintain trajectory to the tar- accumulate on the low side of the hole in cut- WOB. When weight is applied to the bit, torque at get. Such problems frequently result in increased tings beds that increase friction on the drillpipe, the bit increases. Torque is transmitted downhole drilling time, which may adversely impact project making it difficult to maintain constant weight on through the drillstring, which turns to the right, economics and ultimately limit the length of a lat- bit (WOB). in a clockwise direction. As weight is applied to eral section. Differences in frictional forces between the the bit, reactive torque, acting in the opposite The capability to transfer weight to the bit drillpipe inside of casing versus that in open hole direction, also develops. This left-hand torque affects several aspects of directional drilling. The can cause weight to be released suddenly, as can is transferred upward from the bit to the lower driller transfers weight to the bit by easing, or hang-ups caused by key seats and ledges. A sud- part of the drillstring. Reactive torque builds as slacking off, the brake; this transfers some of the den transfer of weight to the bit that exceeds weight is increased, reaching its maximum value hook load, or drillstring weight, to the bit.2 The the downhole motor’s capacity may cause bit when the motor stalls. This reactive torque also difference between the weight imposed at the bit rotation to abruptly halt and the motor to stall. affects the orientation of the motor. and the amount of weight made available by eas- Frequent stalling can damage the stator compo- Reactive torque must be taken into account ing the brake at the surface is primarily caused nent of the motor, depending on the amount of as the driller tries to orient the drilling motor by drag. As the horizontal departure of a wellbore the weight transferred. The driller must operate from the surface. In practice, the driller can increases, so does the longitudinal drag of the the motor within a narrow load range to maintain make minor shifts in toolface orientation by drillpipe along the wellbore. an acceptable ROP without stalling.4 changing downhole WOB, which alters the reac- Controlling weight at the bit throughout At the driller’s console, an impending stall tive torque. To produce larger changes, the driller the sliding mode is made even more difficult might be indicated by an increase in WOB but can lift the bit off-bottom and reorient the tool- by drillstring elasticity, which permits the pipe with no corresponding upsurge in downhole pres- face. Even after the specified toolface orienta- to move nonproportionally. This elasticity can sure to signal that an increase in downhole WOB tion is achieved, maintaining that orientation cause one segment of drillstring to move while has actually occurred. At some point, the WOB can be challenging. Longitudinal drag increases other segments remain stationary or move at indicator will show an abrupt decrease, indicat- with lateral reach, and weight transfer to the different velocities.3 ing a sudden transfer of force from the drillstring bit becomes more erratic along the length of the Poor hole cleaning may also affect weight to the bit.5 horizontal section, thus allowing reactive torque transfer. In sliding mode, hole cleaning is less effi- Increases in drag impede a driller’s ability to build and consequently change the toolface cient because there is no pipe rotation to facilitate to remove torque downhole, making it more dif- angle.7 The effort and time spent on orienting the turbulent flow; this condition reduces the drilling ficult to set and maintain toolface orientation.6 toolface can adversely impact productive time on fluid’s ability to carry solids. Instead, the solids Toolface orientation is affected by torque and the rig.

Topdrive

Maximum Rotating surface-applied only torque

S l id in D g y n p a lu m s Maximum rocking depth Point of interference Bit i to c p fr d ic riv ti e on to o rqu Mud motor f r e ota tion Sliding only Sliding plus reactive torque Static friction Dynamic friction

Minimum surface-applied torque Minimum reactive torque Maximum reactive torque

Figure 4. Torque versus friction. Longitudinal drag along the drillstring and longitudinal drag across the bottom section of the drillstring up to a can be reduced from the surface down to a maximum rocking depth, at point of interference, where the torque is balanced by static friction. An which friction and imposed torque are in balance. By manipulating the intermediate zone remains unaffected by surface rocking torque or by surface torque oscillations, this point can be moved deep enough to reactive torque. By continuously monitoring torque, WOB and ROP while produce a significant reduction in drag. Similarly, reactive torque from sliding, the Slider system helps minimize the length of this intermediate zone the bit creates vibrations that propagate back uphole, breaking friction and thus reduces longitudinal drag.

52 Oilfield Review 0

2,000

4,000

6,000 No friction reduction

Depth of twist, ft 8,000

Reactive 10,000 torque

12,000

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 Surface applied torque, ft-lbf 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.)

The Slider System nonrotating pipe. Rocking can also help reduce not rotate. This section of drillstring, which has Manually correcting and maintaining toolface axial friction along the drillstring. However, this no tangential motion, moves by sliding only and is orientation can be a difficult process. Drilling motion is not necessarily transmitted all the way subject to static friction, which is greater than the efficiency is largely dependent on the driller’s to the bit—other processes are at work. dynamic friction of pipe in motion. ability to: Torque from the topdrive rotates the drill- The location of the point of interference • transfer weight to the bit without stalling the string from the surface down to a maximum varies with changes in the amount of reactive mud motor rocking depth, where friction against the side of torque. To efficiently minimize the sliding inter- • reduce longitudinal drag sufficiently to achieve the hole prevents the pipe from turning. At the val between the depth of rocking and the point of and maintain a desired toolface angle same time, as the mud motor turns the bit, it gen- interference while keeping the maximum rocking • attain acceptable ROP. erates a reactive torque in the opposite direction. depth relatively constant, an automated control The Slider automated surface rotation control This torque is transmitted a short distance up system must be used. system was developed to help operators regain the drillstring until it is overcome by friction at The amount of surface torque supplied by the some of the drilling performance of a conven- some point between the bottom of the wellbore topdrive dictates in large part how far downhole tionally rotating drillstring. The Slider interface and the BHA, referred to as the point of interfer- the rocking motion will be transmitted. This rela- interacts with the topdrive control system to ence (Figure 4). Throughout the interval between tionship between torque and rocking depth can rotate the drillstring back and forth. This torque the bit and the point of interference, the velocity be modeled using conventional torque and drag rocking technique reduces longitudinal drag component of reactive torque imposes a reduction programs (Figure 5). However, these programs along part of the drillstring while slide drilling. in longitudinal drag along the lower part of the are not needed when using the Slider system. Rocking back and forth subjects the upper drill- drillstring and possibly a change in toolface orien- Using inputs from surface hook load and stand- string to near-constant tangential motion, pro- tation. Between the depth where surface torque pipe pressure as well as downhole MWD toolface ducing a dynamic friction coefficient, which is is overcome by friction and the point where reac- angle, the Slider system automatically deter- lower than a static friction coefficient created by tive torque is overcome by friction, the pipe does mines the amount of surface torque needed to

2. The hook load includes more than simply the weight of 3. Maidla and Haci, reference 1. 5. Maidla E, Haci M, Jones S, Cluchey M, Alexander M and the drillstring in air; it is the total force pulling down on 4. Tello Kragjcek RH, Al-Dossary A, Kotb W and Al Gamal A: Warren T: “Field Proof of the New Sliding Technology for the hook as it hangs beneath the derrick traveling block. “Automated Technology Improved the Efficiency of Directional Drilling,” paper SPE/IADC 92558, presented This total force includes the weight of the drillstring, drill Directional Drilling in Extended Reach Wells in Saudi at the SPE/IADC Drilling Conference, Amsterdam, collars and ancillary equipment reduced by any forces Arabia,” paper SPE 149108, presented at the SPE/DGS February 23–25, 2005. that tend to decrease the weight. These forces might Saudi Arabia Section Technical Symposium and 6. Maidla and Haci, reference 1. include friction along the wellbore wall and buoyant Exhibition, Al-Khobar, Saudi Arabia, May 15–18, 2011. 7. Maidla and Haci, reference 1. forces on the drillstring caused by its immersion in drilling fluid.

May 2016 53 Rotary Torque transfer weight downhole to the bit, thus elimi- 0 lbf 20,000 nating the need to come off-bottom to make Weight on Bit Rotary Speed Toolface toolface corrections (Figure 6). This results in an 0 1,000 lbf 100 0 rpm 150 0 degree 360 efficient drilling operation and reduced wear on Differential Pressure Hook Load Rate of Penetration downhole equipment. 0 psi 800 0 1,000 lbm 300 0 ft/h 50

RotatingRotating MoModede System Hardware Slider 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 Manual Mode 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 Slider Mode 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 Figure 6. Comparison of rotating and sliding drilling parameters. Rate of penetration (ROP) and toolface to the right and the left. The transfer of weight is 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); controlled by varying surface torque to compen- however, toolface varies drastically, as there is no attempt to control it (Track 3). Hook load (Track 2) sate for changes in reactive torque. Corrections and weight on bit (WOB) remain fairly constant while differential pressure (Track 1) shows a slight in toolface angle are achieved through addi- increase as depth increases. To begin manual sliding (middle), the driller pulls off-bottom to release 8 trapped torque; during this time, WOB (Track 1) decreases while hook load (Track 2) increases. As tional torque pulses during the rocking cycles. drilling proceeds, inconsistencies in differential pressure—the difference between pressures when For every torque cycle to the left or right, a cor- the bit is on-bottom versus off-bottom—indicate poor transfer of weight to the bit (Track 1). Spikes responding differential pressure peak occurs, of rotary torque indicate the directional driller’s efforts to orient and maintain toolface orientation indicating that the weight is being transferred (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 to the bit. To adjust the toolface orientation, the quickly gained toolface orientation. When the WOB increased, differential pressure was consistent, driller can control the magnitude and frequency demonstrating good weight transfer (Track 1). Weight on bit during a Slider operation is lower than of torque pulses during a rocking cycle. 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).

54 Oilfield Review Figure 7. The Slider system graphical display. Downhole performance left-hand torque (orange). Up-and-down keys allow the driller to set values parameters are monitored and controlled via a notebook computer for left and right torque (upper left). Brief torque increases above set values interface between the topdrive and Slider system. The directional driller can be added for one oscillation cycle by bumping left or right (middle left). can configure this display to show various key parameters such as toolface The driller can immediately override the system by hitting the disable button (dial, center) and torque and differential pressure (chart, bottom right). (upper right). Torque curves show higher values for righ-hand torque (yellow) than for

Field Experience Slider technology has been instrumental in developing unconventional plays throughout North America. Wattenberg field, one of the more Wyoming prolific fields in the Denver-Julesburg basin, is located in Weld County, Colorado. There, a lead- Nebraska ing operator in the area used the Slider system to Denver-Julesburg basin drill horizontal wells in the Cretaceous Niobrara Wattenberg USA field gas play (Figure 8). One of those wells, spudded in February Colorado 2016, was drilled vertically to its kickoff point, Kansas then drilled in a westerly direction to its land-

8. “Slider: New Level of Efficiency to Directional Drilling,” Figure 8. Wattenberg field. The prolific Wattenberg field lies in north-central Colorado, USA, within the Drilling Contractor 11, no. 4 (July–August 2004): 28–31. Denver-Julesburg basin.

May 2016 55 Toolface Rotary Torque Pump 1 0 degree 360 0 lbf 34.39 0 strokes/min 200 Hook Load Rotary Speed Standpipe Pressure 0 1,000 lbm 300 0 rpm 80 0 psi 12,000 Depth, Time, Weight on Bit Rate of Penetration Differential Pressure ft h:min 0 1,000 lbm 100 0 ft/h 300 0 psi 1,000 07:12

07:14

07:16

07:18

07:20

07:22

X,X25 07:24

07:26

07:28

07:30

07:32

X,X35 07:34

07:36

07:38

07:40

07:42

07:44

X,X47 07:46

07:48

07:50

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.

ing point in the Niobrara pay zone. Beginning Faster and Farther rocking system facilitates a longer horizontal sec- at X,X25 ft, the directional driller manually con- By sensing the amount of surface torque required tion, which has less tortuosity, ultimately leading trolled the slide drilling while averaging 38 ft/h to transfer weight to the bit and by eliminating to increased production. —MV [11.6 m/h] ROP. Upon engaging the Slider auto- the need to pull off-bottom to make toolface cor- mated system, the driller reported an average rections, the Slider automated surface rotation ROP of 51 ft/h [15.5 m/h], for an improvement control system enables substantial increases Contributor of 34% in ROP compared with that of manually in ROP and lateral reach for directional wells. Steven Duplantis is an Operations Manager for controlled sliding (Figure 9). The Slider system Rocking, or oscillating, the drillstring back and Schlumberger in Houston, where he oversees day-to-day was engaged several times during the course of forth helps the driller overcome friction and thus operations, sales and development for the Slider auto- drilling this well. Each time the trajectory began reduce drag on the drillstring. Along with reduc- mated surface rotation control system. Steven began to drift beyond specified tolerances, the direc- ing drag, operators can decrease the amount of his oilfield career in 1994 and has held positions with MD Totco, Noble Engineering and Epoch/Optidrill. tional driller switched from rotating to sliding mud additives normally used for lubrication. The Since 2006, he has been focused on the Slider system, modes to bring the wellbore back on course. Slider automated system typically applies less working as a field coordinator and engineer involved Comparisons between manually controlled slid- WOB to maintain toolface control and has mark- with Slider system testing and field trials along with ing and automated sliding with the Slider system edly fewer motor stalls than are experienced implementation and ongoing development of advanced consistently showed significant gains in ROP over while manually slide drilling. By achieving con- features for the Slider product line. the manual approach. sistent toolface control, this automated torque

56 Oilfield Review LOOKING BACK

Birth of La Pros: The 90th Anniversary

Philippe Theys Houston, Texas, USA

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- But Paul Schlumberger, Conrad and Marcel’s visionary father, who had cal resistivity mapping of the lawn of the family property at Val-Richer in another four children—Jean, Daniel, Maurice and Pauline—with his wife Normandy, France, in the summer of 1912—and the official registration of Marguerite de Witt, redirected Conrad from factory life and tied him to his Société de Prospection Électrique—or SPE, and soon nicknamed La Pros— brother’s future. in July 1926, a long gestation period occurred. Paul wrote, “The scientific interest in research must take precedence over The primary reason for the delay in the company’s full formation was financial interests. Marcel will bring to Conrad his remarkable competence World War I. Engineers, scientists and inventors in Europe did not have the as an engineer and his common sense. Conrad will be the wise physicist. I freedom to choose not to spend time in the army of fighting in the war. The will support them.” He offered a sum of 500,000 francs for the endeavor.2 Schlumberger brothers were no exception. Conrad Schlumberger joined the In 1919, 500,000 francs represented 160 kg [5,100 troy oz] of gold. Unfortu- French artillery as an officer and his younger brother Marcel joined the nately, the franc value was cut by a factor of five from 1914 to 1926, and La Pros, cavalry. Consequently, the Schlumberger startup history was put on hold which had engineers on four continents, soon had to learn how to manage in 1914. currency variations. The business invested in trucks and equipment and sent Conrad was intimately involved with and horrified by the war. He began to people to other countries, some of which were not hospitable. These early question his future and considered giving up his work to become a pacifist. processes meant the company had essential cash flow needs. While in the trenches at Verdun, he said of the family textile business: “I’m The survival of the young company was based on three components: seed ashamed of belonging to a family of textile manufacturers whose employees money, innovation and sound management of intellectual property and got up at dawn, walked for miles to reach the factory, and worked twelve or people. Success came about because the company built an invincible team fourteen hours a day. Not one of those workers had a life fit to live.”1 and instilled exceptional values.

Oilfield Review 28, no. 2 (May 2016). 1. Schlumberger AG: The Schlumberger Adventure: Two Brothers Who Pioneered in Copyright © 2016 Schlumberger. Petroleum Technology. New York City: Arco Publishing, Inc., 1982. 2. Schlumberger, reference 1. Microsoft is a registered mark of the Microsoft Corporation. Xerox is a trademark of the Xerox Corporation.

May 2016 57 LOOKING BACK (continued)

Figure 1. Patents filed in France, Brazil, Mexico and Australia. The first work in these countries until 1936 and 1945, respectively. The patent at the patent was filed in France on September 27, 1912 (top left); the illustration bottom was filed in Australia. (Documents courtesy of the Schlumberger (top right) is from the first patent. Conrad Schlumberger filed patents in, museum in Crevecoeur, France, and Collections École Polytechnique, among other countries, Mexico and Brazil (middle) although SPE did not Palaiseau, France.)

58 Oilfield Review 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 responsibility and trust that are difficult to imagine in our era of instant Innovation is the backbone of startups, but new techniques and technolo- texting and emailing. gies often fall in the public domain and quickly become available to com- All the right components were finally available; Société de Prospection petitors and to clients who may take advantage of this availability without Électrique was thus incorporated in Paris on July 1, 1926, with 200,000 rewarding the original inventor. Chester Carlson, for example, working in francs in capital divided in 2,000 shares. SPE was now the organization the 1930s and 1940s, wisely registered every development of his research for overseeing all operations, personnel and premises that were previously held the company that would eventually become Xerox Corporation. loosely by the brothers. Conrad was president. Maurice and Albert Doll, who Conrad Schlumberger filed his first patent, Procédé pour la détermina- was Pauline’s husband and Henri-George’s uncle, were scrutateurs, or trea- tion du sous-sol au moyen de l’électricité (Subsurface characterization by surers. Jean, cofounder of La Nouvelle Revue Française or NRF, and electrical methods) on September 27, 1912, in France. He then took care to Maurice—cofounder of the bank Neuflize Schlumberger Mallet—were file patents in what seemed like unusual locations, including Mexico, named minority shareholders. The headquarters, at 30 rue Fabert in Paris, Czechoslovakia, Brazil and the Belgian Congo. By 1926, he had filed patents consisted of a small, five-room apartment close to Place des Invalides. in 18 countries, which would become the core of the intellectual property of In September 1926, Conrad left France for the US to further develop that SPE (Figure 1). 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, The Team Marcel and Henri-Georges—who contributed to the success of the mission Reinforced by their father’s monetary donation, Conrad and Marcel consti- Paul had dreamed of—joined the family patriarch in the cemetery of the vil- tuted an unbreakable partnership. “Question the one without the other and lage of Saint-Ouen-le-Pin, France, in 1936, 1953 and 1991, respectively. the answer would be: ‘I’ll talk about it with my brother Marcel… .’ [and] ‘I’ll Of the many Schlumberger companies, SPE is the oldest. In the US, the ask my brother Conrad.’” 3 Schlumberger Electrical Prospecting Method, defunct as the result of the Eugène Léonardon was the first full-time engineer hired by the brothers. Great Depression, was replaced by Schlumberger Well Services Corporation in He began working with Conrad in 1913 but then served in the French artil- 1934 and by Schlumberger Technology Corporation in 1984. In 2016, SPE still lery during World War I; he was rehired in 1919. Additional field hands were exists and recruits young engineers to send to the far corners of the globe. cautiously recruited. The ads in Le Journal des Mines called for ingénieurs Schlumberger currently operates in more than 85 countries and has sportifs, athletic engineers. After Léonardon came a series of intrepid engi- about 100,000 employees. Wherever the drill bit is turning in the search neers (Figure 2). Jules Carré had worked with Conrad as early as 1913. for hydrocarbons, the Schlumberger name can be found. The startup com- Gilbert Deschâtre and Raymond Sauvage participated in the first survey for pany formed by a few industrious and adventurous engineers has become Shell in 1928. Jacques Gallois and Pierre Baron constituted the first pros- one of the most recognized names in the oil and gas industry. Could Paul, pecting crew in Freeport, Texas, USA. Felicien Mailly joined in 1925 to man- Conrad or Marcel have ever imagined the enterprise that today bears their age the electrical research division. name? Schlumberger stresses its guiding values of people, technology and Not all of these new recruits were graduate engineers. Roger Jost was profits. Although they differ from the original principles, they echo the the nephew of the régisseur, or stage manager, for events at Val-Richer. The “seed money, innovation and team” concepts that formed the company brothers also looked for help within the family. Nephew Marc Schlumberger that would become Schlumberger; together, these ideas continue to guide was sent to the US to scout for business. Henri-Georges Doll, who was the company into the future. 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 Philippe Theys was hired by Société de Prospection Électrique in 1972, 46 years 1921. Swiss geologist Edouard Poldini joined in 1922. By 1926, the company 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, had a staff of 17 field engineers working in Romania, Serbia, Canada, South Western and Eastern Australia and Taiwan as well as in Louisiana and Texas in the Africa and the Belgian Congo. Because of the distances and lack of com- US before a 26-year career in marketing, research and engineering and quality. He munication, these people were given high levels of independence, initiative, 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 3. Schlumberger, reference 1. and now consults for oil and gas companies.

May 2016 59 THE DEFINING SERIES

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 FPSO 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.

Production risers Umbilical riser Subsea tree Boosting pump station Manifold

PLET Jumper

Electric power leads Injection fluid jumper Electric power distribution unit UTA Hydraulic fluid jumper

Subsea Infrastructure Matt Varhaug Senior Editor

To replace dwindling reserves from onshore fields or from offshore wells Seabed Equipment drilled in shallow waters, many E&P companies are turning to their deepwa- The subsea wellhead, installed at the beginning of the drilling phase, pro- ter prospects. Exploration or production in deep and ultradeep waters is vides the structural foundation for the well. The wellhead is also where the carried out at water depths of 300 to 3,050 m [1,000 to 10,000 ft] or greater. subsea tree is mounted. In some configurations, the tree contains the pro- These depths dictate that most wells be completed subsea, with wellheads, duction tubing hanger and accommodates hydraulic and electrical lines pressure-control equipment and production equipment placed at the seafloor. used for managing downhole safety valves, completion valves and pressure From deepwater and ultradeepwater completions, produced fluids are or temperature sensors. The function of the subsea tree is to control and sent to a processing facility by way of a subsea production system. A sub- manage pressure and flow over the life of the well and enable any necessary sea production system consists of the subsea infrastructure used to pro- intervention. The tree is the primary mechanism for shutting in the well at duce oil and gas from offshore reservoirs. It encompasses one or more the seabed and serves as the interface for well reentry operations. A subsea subsea wells and the subsystems necessary to deliver hydrocarbons to a control module (SCM) attached to the tree contains the instrumentation, fixed, floating, subsea or onshore processing facility. These subsystems can electronics and hydraulics connections needed for safe operation of the be divided into subsea trees, production controls, manifolds, jumpers, flow- subsea tree valves, chokes and downhole valves. lines, risers, umbilicals and processing components. Injection of water or Sections of pipe, known as jumpers, run between subsea structures to gas back into subsea wells is also a function of the subsea production sys- serve as links through which fluids are transmitted. These pipe segments tem. Generally, oil, gas and water produced from the reservoir will flow range in length from a few meters to hundreds of meters. A jumper is often from a wellbore to a subsea tree and through a jumper to a manifold and installed to carry production downstream from the tree. The produced fluid subsea flowline. Today, many operators route the flowline to a booster may be routed through a multiphase flowmeter to measure production pump to energize the flow as it travels between the seafloor and a riser that rates and volumes. carries it to the surface for processing. Where multiple wells produce in a subsea development, flowline jumpers from individual wells send produced fluids to a subsea production manifold Oilfield Review 28, no. 2 (May 2016). (Figure 1). By routing produced fluids from multiple wells to the production Copyright © 2016 Schlumberger. manifold, the operator can reduce the number of flowlines that must be accommodated at the next step in the production chain. Upon reaching the

60 Oilfield Review 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 High-voltage cable injected water, gas and chemicals to one or more subsea wells. Electric cable Some subsea installations have a pipeline end manifold (PLEM), which Chemical injection connects a flowline with another subsea structure or joins a main pipeline fluid line Fiber-optic cable with a branch pipeline. The PLEM can incorporate tie-in points with other Outer plastic sheath components such as isolation valves, diverter valves and sensor arrays. Some PLEM designs incorporate facilities for launching pipeline pigs— Hydraulic fluid line Wire cable armor 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 Figure 2. Cross section of an umbilical. Umbilicals supply electric power, installed. Boosting pumps function as a seafloor artificial lift system, increas- hydraulic fluid, chemicals and fiber-optic communications to the subsea production system. Separate hoses, cables, injection lines and other ing both flow rate and recovery by reducing backpressure on the reservoir. conduits are bundled together and enclosed within an armored outer ring Other recent advances in subsea processing are used to enhance field designed to withstand harsh subsea conditions. 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 point that gas hydrates start to form. The change in fluid temperature gas compression, can improve viability of certain marginal developments. 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 Lifeline The surface processing facility provides power, control, communication and chemical injection services back to the subsea production system. These 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 services are transmitted through a subsea distribution system using umbil- a flowline running to a production platform. 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 Flow to the Surface conduits require sophisticated materials and manufacturing techniques to Flowlines tie one or more fields back to a production facility—a shore- withstand deep-ocean currents, pressures and temperatures. Power con- based processing facility or fixed production platform in shallower waters— ductors provide electricity for subsea equipment and system sensors. but in deeper waters, a semisubmersible, spar and floating production, Hydraulic lines are used to open and close subsea valves. Fiber-optic lines storage, and offloading vessel (FPSO) is used. The flowlines do not necessar- instantly relay sensor information and control commands between the sea- ily trace a straight course from wellhead to platform but may bend to avoid floor and the surface. Some umbilical lines pump chemicals into the produc- obstacles such as existing subsea infrastructure or natural obstructions tion stream. Umbilicals directly or indirectly service nearly every component such as underwater seamounts or canyons. As it follows the topography of in the subsea production system and are critical to operating the field. The the seafloor, the flowline climbs gradually from the colder, deeper reaches lines typically run from the surface processing facility to an umbilical ter- of the field upward through relatively warmer, shallower waters before mination assembly (UTA) on the seafloor, from which services are distrib- reaching the production facility. uted throughout the field. Water depth affects temperature, which can adversely impact flow Upon reaching the surface, the produced fluids are separated and between the subsea tree and the production facility. Upon exiting the well- treated by the processing facility. From there, an export pipeline transports head, warm produced fluids may encounter deepwater temperatures the product to a storage and offloading installation or to an onshore refinery approaching 2°C [36°F] at the seafloor. Heat transfer between the produced for further processing and distribution. fluid in the pipeline and the surrounding seawater can cool the fluid to the

May 2016 61 THE DEFINING SERIES

Unifying Characteristics Geophysics The challenges usually encountered in geophysics are posed in the form of Richard Coates an inverse problem, such that a set of measurements and known physical Research Manager and Scientific Advisor 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- Geophysics is the study of the physics of the Earth, the propagation of elas- sible solution that satisfies the measurements. Geophysicists attempt to tic waves within it as well as its electrical, gravitational and magnetic fields. resolve this ambiguity by integrating complementary data acquired from Although the origins of geophysics can be traced to ancient times, it was not dissimilar methods or by adding supplemental knowledge such as wellbore until the early 20th century that scientists began applying geophysical con- measurements to determine which solution is correct. cepts and techniques to the search for hydrocarbons and minerals and to In addition to nonuniqueness, all geophysical methods exhibit a evaluate geothermal energy resources. Now, geophysics plays a critical role decrease in the resolving power with distance from the measuring equip- in the petroleum industry because geophysical data are used by exploration ment. This concept is analogous to the difficulty of distinguishing objects and development personnel to make predictions about the presence, nature by sight at increasing distances. The characteristic is more pronounced and size of subsurface hydrocarbon accumulations. 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 vessel Airgun array Sea surface Seismic Surveying Streamer with hydrophone sensor arrays, 6 to 12 m deep In 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 Seabed by sensors such as hydrophones or geophones. The distances between source and sensor can exceed 15 km [9.3 mi]. Sedimentary layers 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 Figure 1. Marine seismic acquisition. An airgun array (top) produces pulses extracted from seismic data. For example, by studying the size, or ampli- of seismic energy (green) that penetrate the subsurface and are reflected tude, of the reflections and how the amplitude changes with the angle at back (red) from the seabed and interfaces between rock layers. These which the seismic waves hit the reflectors, geophysicists may be able to reflections are detected by hydrophone arrays towed behind the seismic vessel. Geophysicists invert the recorded data to construct a 3D image of determine whether the pores within the rocks contain gas, oil or water. This the subterranean layers. Modern seismic acquisition methods (bottom) step, known as amplitude versus offset (AVO), often has a higher level of illuminate 3D swaths of the Earth from a variety of angles. In one common uncertainty than does structural imaging. 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 Although most seismic work uses active sources designed to create seis- distance behind. The outermost ships also tow streamers (black lines) mic waves, the detection of weak seismic waves generated during hydraulic typically 10 km [6 mi] in length, which record the reflections from below the fracturing is of increasing interest. These faint signals are used to deter- seabed and within a rock volume (tan) beneath and between the two sets mine the locations of microseismic events, which can indicate the position of streamers. and extent of the hydraulic fractures. Oilfield Review 28, no. 2 (May 2016). Copyright © 2016 Schlumberger.

62 Oilfield Review Electromagnetic Methods To reduce the interpretational uncertainty remaining after seismic survey- Air (resistive) Natural-source ing, geophysicists can choose from several techniques. The most common magnetotelluric fields are electromagnetic (EM) methods, which leverage the fact that some important subterranean formations have strong EM signatures. For exam- CSEM transmitter Seawater (conductive) 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 Electric and magnetic field recorders resistivity. Their high velocity makes seismic imaging beneath them prob- Oil and gas (resistive) Seafloor (variable conductivity) Salt, carbonates lematic, but their high electrical resistivity makes them easy to detect using and volcanics EM surveys. (resistive) 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 Distance 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 10 near, the surface. The first technique is calledcontrolled-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- Resistivity, ohm.m Resistivity, nique, magnetotellurics (MT), was introduced in the 1950s. Some modern 1 systems can acquire CSEM as well as MT data when the controlled source is Depth 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 Joint imaging base salt 1 km more appropriate for detailed reservoir-scale targets and high-resistivity Seismic base salt 1 km anomalies. Consequently, the CSEM method is typically used to investigate potential hydrocarbon reservoirs previously suggested by seismic images. Figure 2. Marine magnetotellurics (MT) and controlled-source EM (CSEM) acquisition. For marine MT studies (top), electric and magnetic recorders on Magnetic Surveying the seafloor make time series measurements of the Earth’s varying magnetic Magnetic surveying is another type of subsurface prospecting. Unlike EM field and the induced electric field, which can be interpreted to infer deep methods, which rely on fields that fluctuate rapidly in time, magnetic survey- geologic structures. Towing a CSEM transmitter, which has receivers behind it, close to the seafloor allows geophysicists to map shallow structures, ing depends on the permanent magnetic properties of rocks, whose strength including thin, resistive features such as hydrocarbon reservoirs. Joint and orientation are fixed at the time of their deposition and may be in con- evaluation of multiple geophysical measurements (bottom) enables trast with those of the surrounding rock. Measuring these subtle anomalies geophysicists to obtain a consistent interpretation of the base of salt. The can help geophysicists map subsurface formations over large areas. 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 The advantage of magnetic surveying is that data can be collected from MT resistivity data (colors) provides significant new information. Combining aircraft or satellites as well as from land or by ship. Consequently, magnetic seismic and MT data improves the previous interpretations of the base of salt surveys can inexpensively cover large geographic areas as well as sites that and gives interpreters greater confidence in their result (yellow dashed line). are otherwise difficult to access. Because the strongest anomalies are pro- duced by volcanic or metamorphic formations, magnetic surveys are widely frequently acquired using aircraft and satellites; taking measurements by used for mineral exploration. ship is also common.

Gravimetry Surveying Variations and Value Gravity measurements have been applied in the oil field since the 1920s. Geophysical methods are applied in various ways. For example, seismic The technique is based on recording spatial variations in the Earth’s gravi- receivers are sometimes deployed in boreholes to generate detailed images tational field, caused by differences in the density of rocks below the survey of small portions of the Earth. In addition, certain niche techniques, such location. The size of these variations is typically less than 1/100,000th of as the hyperspectral imaging, spontaneous potential and electrokinetic— Earth’s gravitational field’s nominal value of about 9.81 m/s2 [32.2 ft/s2]. seismoelectric and electroseismic—methods, are available but are not Detecting such small variations requires extremely sensitive instru- widely used. Of all geophysical techniques, seismic surveying is by far the ments and the application of multiple corrections. For example, the most widespread. Because of this dominance, “seismics” and “geophysics” Bouguer correction accounts for variations in gravity caused by local are often used interchangeably in the oil industry, although for purists, this topography and corrects for the influence of latitude and measurement is wrong. Nevertheless, the integrated use of complementary geophysical altitude that might otherwise mask the signal. Because the low density of methods provides critical information about the subsurface. This informa- salt generates a large gravity anomaly, the most common oilfield applica- tion is used by exploration and development personnel to make decisions tion of gravity surveying is to help delineate salt domes. Gravity data are about where and how to drill.

May 2016 63 LOOKING BACK

Looking Back on Wireline

On the occasion of 90 years following the July 1926 official Meanwhile, Conrad was playing with new ideas for his well log- registration of Société de Prospection Électrique (abbreviated ging. Realizing that oil was infinitely resistive to electricity, he to SPE and soon nicknamed La Pros), this Looking Back article postulated that if a zone was oil bearing and reasonably thick, then examines the origins of the first logs and a contributed article increasing the spacing between the wires’ measuring potential would by Philippe Theys (“Birth of La Pros,” page 57) discusses those allow the logging tool to see deeper into the formation and record a first engineers who established an industry. higher resistivity; the phenomenon was not observed in Pechelbronn because the oil zones were too thin. In May 1930, he asked his field In the late 1920s, an alternative to the time-consuming process of engineers to try the idea and report back. Within a month, Marcel obtaining and analyzing rock cores was born in the wine region of Jabiol, the company’s solitary engineer in northern Sumatra, sent Alsace, France. This new technique became known as “carottage a telegram back confirming Conrad’s hypothesis. Electrical logging électrique,” or electrical coring, and eventually developed into the could locate oil from the borehole. 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. Pulley used in Pechelbronn in 1928. The Pechelbronn Oil Company had opened a new refinery that could (Illustration courtesy of and handle 80,000 metric tons, approximately 11,000 bbl, of oil annually and adapted from Mau and Edmundson, reference 1.) 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. Harnessing the Potential Marcel Schlumberger had already tested the technique in 1921, Combining another measurement technique with electrical logging, taking resistivity measurements over about a meter [several feet] at however, was necessary to fully optimize the electrical logging the bottom of a 760-m [2,500-ft] hole in Molières-sur-Cèze in south- method—namely, spontaneous potential (SP). Natural electrical ern France. The results were inconclusive, but the feasibility of a potentials in the subsurface had been discovered about a century downhole resistivity measurement had been proved. The geophysical earlier in Cornwall, England, by the British geologist, natural philoso- community remained skeptical, however. pher and inventor Robert Were Fox. Potentials in the borehole are On September 5, 1927, Henri-Georges Doll, Conrad’s son-in-law, and caused by electrochemical interactions between the borehole fluid two colleagues, Roger Jost and Charles Scheibli, conducted the first and adjacent sand and shale formations. electrical logging operation in a 500-m [1,600-ft] Pechelbronn well Conrad Schlumberger secured a French patent on the SP tech- named Diefenbach 2905. The team logged an interval of 140 m [460 ft], nique in 1929, claiming it could be used to locate permeable strata, starting from a depth of 279 m [915 ft]. They rigged up a hand-oper- but found no practical application. A year later, Doll observed ated winch that lowered into the hole three insulated wires—cables natural potentials while logging in the Oklahoma Seminole oil field. of the type used for lighting fixtures—tied together with friction tape. While the battery was disconnected, he noticed the potentiometer The longest of the wires injected current into the well and formation; needle vibrating back and forth as the electrodes were lowered into the return was at the surface. The other two wires, shorter and of the well. Experiments followed at Pechelbronn, and by 1930, they slightly different lengths, measured the resulting potential field and concluded that the SP method could differentiate permeable beds provided the resistivity readings. Measurements were made at 1-m such as sand and limestone from impermeable formations such as [3 ft] intervals; the entire operation took five hours. The result was a shale. The combination of SP and resistivity curves turned out to be resistivity log that distinguished between the many layers of sand and of much greater value than the resistivity log alone in detecting pro- shale pierced by the borehole. duction possibilities. Electrical logging established a reliable method Resistivity logging continued throughout 1928 at Pechelbronn, of achieving stratigraphic correlation, providing a way to distinguish and the resulting correlations of resistivity from one well to the next between shale and porous rock and between hydrocarbon- and revolutionized the understanding of the stratigraphy of the field. water-bearing rock. Soon, the Pechelbronn Oil Company was able to raise the annual In the decades that followed, electrical logging became increas- capacity of its new refinery to 100,000 metric tons. By 1929, the new ingly sophisticated and paved the way for numerous other well log- logging technique had gone global—Schlumberger logging crews ging techniques. It became an industry on its own, taking the key were engaged by Shell for their explorations in Venezuela, the US position among formation evaluation techniques. and the Dutch East Indies, and by the Soviet Union for the oil fields of Grozny, Chechnya, and Baku, Azerbaijan. 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.

64 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. Oilfield Review http://www.slb.com/oilfieldreview Authoritative. Relevant. Informative. Oilfield Review Authoritative. Relevant. Informative.