oilfield review autumn 2007

SCHLUMBERGER OILFIELD REVIEW

AUTUMN

2007VOLUM

E 19 NUM

BER 3

Autumn 2007

Tsunami Science

Borehole Seismic Surveys

Geomechanics

Fluid-Property Measurements

Oilfield Review

60336schD1R1.qxp:60336schD1R1 12/4/07 4:31 PM

07-OR-004-0

60336schD1R1.qxp:60336schD1R1 12/4/07 4:31 PM

Petroleum geomechanics is playing an increasingly impor-tant role in hydrocarbon appraisal and development. Inmany regions, hydrocarbon reservoirs and the overburdenthrough which wells must be drilled pose significant chal-lenges from either extreme of burial depth: from deeplyburied reservoirs with high temperature and pressure; orshallow, heavy-oil deposits with low pressure and relativelycold temperature affecting fluid viscosity. Substantial vol-umes of hydrocarbons are currently produced from com-plex formations and unconventional reservoir rocks thathave only recently been conquered through technologicaladvances. Geomechanics data are crucial for developingthese difficult, complex and unconventional reservoirs.

A second reason for the increased emphasis on geome-chanics is that it makes good economic sense. Experiencefrom the recent past—such as the compaction and subsi-dence in the Ekofisk and Valhall North Sea fields—hasshown that remediation of problems resulting from incom-plete evaluation of geomechanical processes can be extremelyexpensive. The cost includes not only well or infrastructurereplacement, but also the greater cost of lost production.

New technologies have greatly enhanced the role of geo-mechanics. With the expansion of geomechanics applica-tions, the industry has recognized the need to make rock-mechanical property characterizations under appropriatereservoir conditions. This need is being met, in part, throughtesting facilities such as the TerraTek* Geomechanics Lab-oratory Center of Excellence with its high-pressure wellboresimulator and triaxial-stress frames.

Other innovative approaches have paved the way toextracting more information from wireline logging data.Cluster analysis applies multidimensional log analysis andpattern-recognition techniques to discriminate rock unitsfor core targeting and testing, and for relating rock proper-ties from core data to cored and nearby uncored wells (see“Rocks Matter: Ground Truth in Geomechanics,” page 36).My company, BP, is successfully extending this approach inthe deepwater Gulf of Mexico to identify potentially sand-prone clastic formation intervals in wells that do not havecore recovery. Advances in sonic wave velocity measurement,such as those provided by the Sonic Scanner* acousticscanning platform, permit measurement of stress-depen-dent properties of rocks near the wellbore. Recently, radialand axial sonic data analysis of Sonic Scanner logs allowedprediction of the magnitude and orientation of in-situ stressesin BP’s Wamsutter, Wyoming (USA) field, and was used tohelp optimize hydraulic fracture stimulation treatments.

Petroleum Geomechanics: From Six Miles Deep to Just Below Your Feet

The availability of robust geomechanical data from theoverburden and reservoir allows more effective use of 3Dearth models. The integration of time-lapse seismic dataextends geomechanical earth models into the fourthdimension of time, to better characterize reservoirresponses to production and injection. This life-of-field geomechanics approach has been effectively integratedusing the VISAGE* coupled geomechanical reservoir simu-lator in the Valhall field.

Yet challenges remain. Although wellbore stability predic-tion has become commonplace in many drilling provinces,nonproductive drilling time still costs the industry hundredsof millions of dollars per year. The cause may not be inade-quate predictions, but rather that predrill estimates of for-mation properties do not match the in-situ properties. Newdevelopments in MWD and LWD capabilities—such as thoseoffered by the Scope* family of tools—together with betterintegration of rigsite and real-time office-based support permit a more comprehensive focus on a “No Drilling Sur-prises” approach to optimize wellbore construction. Withtoday’s high drilling-rig rates and the quest to improve effi-ciency, it is this area where the application of geomechanicstechniques will help improve near-term drilling performance.For example, BP staff members in Houston have used theInterACT* real-time monitoring and data delivery system toassess hole quality and wellbore stability in wells drilled inthe Chirag field, offshore Azerbaijan.

Clearly, whether the focus is six miles deep or closer tosurface, geomechanics has a rock-solid role to play in sus-taining and increasing hydrocarbon production now, andinto the future.

Stephen WillsonRock Mechanics AdvisorBP America Inc.Houston, Texas, USA

Stephen Willson is the Rock Mechanics Advisor in the BP Drilling and Completions Technology Unit, with more than 20 years of experience in petro-leum geomechanics. His current focus is on wellbore stability, salt mechanics,and compaction and subsidence, including geomechanical well-integrity chal-lenges facing BP developments in the deepwater Gulf of Mexico. Since joiningBP in 1988, he has held various research and technology development posi-tions in both Sunbury, England, and Houston. He also served as completions manager for TerraTek, Inc., in Salt Lake City, Utah, USA, from 1992 to 1995.Stephen is a civil engineer with a PhD degree in soil mechanics from the University of Manchester, England.

1

An asterisk (*) is used to denote a mark of Schlumberger.

60336schD2R1.qxp:ORSPR04_Editorial_01 12/4/07 7:23 PM Page 1

Schlumberger

Oilfield Review4 The Science of Tsunamis

When the Sumatra-Andaman earthquake rattled the easternIndian Ocean, it generated a tsunami that alerted the world tothe destructive nature of these waves. Scientists use a varietyof tools to understand the tectonics of the region that generatedthe tsunami and to establish an early-warning system for theIndian Ocean.

20 Borehole Seismic Surveys: Beyond the Vertical Profile

Borehole seismic surveys provide much more than time-depthcorrelation and simple images for comparison with surfaceseismic sections. With examples from deepwater Gulf ofMexico, onshore Brazil, north Texas and the North Sea, thisarticle describes a number of innovative applications, includinglook-ahead imaging while drilling, 3D borehole seismic surveys,optimization of hydraulic fracture stimulation, assessment ofperforating-gun performance, and high-pressure, high-temper-ature acquisition.

Executive EditorMark A. Andersen

Advisory EditorLisa Stewart

EditorsMatt VarhaugRick von FlaternVladislav GlyanchenkoTony Smithson

Contributing EditorsRana RottenbergJudy JonesErik Nelson

Design/ProductionHerring DesignSteve Freeman

IllustrationTom McNeffMike MessingerGeorge Stewart

PrintingWetmore Printing CompanyCurtis Weeks

Address editorial correspondence to:Oilfield Review1325 S. Dairy Ashford Houston, Texas 77077 USA(1) 281-285-7847Fax: (1) 281-285-1537E-mail: [email protected]

Address distribution inquiries to:Tony SmithsonOilfield Review12149 Lakeview Manor Dr.Northport, Alabama 35475 USA(1) 832-886-5217Fax: (1) 281-285-0065E-mail: [email protected]

Useful links:

Schlumbergerwww.slb.com

Oilfield Review Archivewww.slb.com/oilfieldreview

Oilfield Glossarywww.glossary.oilfield.slb.com

On the cover:

A specialist at the Schlumberger rockmechanics laboratory in Salt Lake City,Utah, USA, prepares a rock sample fortriaxial compression testing. The inset at right displays the result of a coupledgeomechanics-reservoir flow simulationshowing permeability within layers of afractured reservoir. A seismic section,inset at left, shows faulting in the area ofa large earthquake, resulting in a tsunamithat occurred on December 26, 2004.

2

60336schD3R1.qxp:60336schD3R1 12/4/07 7:06 PM Page 2

Autumn 2007Volume 19Number 3

71 Contributors

75 New Books and Coming in Oilfield Review

3

σHσH

σh

σh

Wellbore

36 Rocks Matter: Ground Truth in Geomechanics

Drilling and production activities cause stress changes in reser-voir and overburden rock. These rocks vary across a range oflithologies, and each type reacts differently to stress. E&P com-panies require an understanding of the interactions between rockfabric, in-situ pressures, temperature and other conditions—and geomechanics is key to this understanding. This articledescribes recent approaches to measuring, modeling andmonitoring geomechanics rock properties and their behaviorthat are helping E&P companies find and produce reservesmore efficiently.

56 Advancing Fluid-Property Measurements

Following the discovery of an oil or gas accumulation, chemistsand engineers analyze reservoir-fluid samples to determine theoptimal reservoir-management strategy. They identify chemical,rheological or depositional problems that might inhibit orinterrupt production. This article surveys the roles of chem-istry, geology and thermodynamics in downhole fluid analysis,fluid sampling, laboratory testing and completion design. Alsoincluded are case studies that demonstrate how these activi-ties benefit reservoir development, well-completion design andproduction operations.

Abdulla I. Al-KubaisySaudi AramcoRas Tanura, Saudi Arabia

Dilip M. KaleONGC Energy CentreNew Delhi, India

Roland HampWoodside Energy, Ltd.Perth, Australia

George KingBPHouston, Texas

Eteng A. SalamPERTAMINAJakarta, Indonesia

Richard WoodhouseIndependent consultantSurrey, England

Advisory Panel

Oilfield Review subscriptions are available from:Oilfield Review ServicesBarbour Square, High StreetTattenhall, Chester CH3 9RF England(44) 1829-770569Fax: (44) 1829-771354E-mail: [email protected] subscriptions, including postage,are 200.00 US dollars, subject toexchange-rate fluctuations.

Oilfield Review is published quarterly bySchlumberger to communicate technicaladvances in finding and producing hydro-carbons to oilfield professionals. OilfieldReview is distributed by Schlumberger toits employees and clients. Oilfield Reviewis printed in the USA.

Contributors listed with only geographiclocation are employees of Schlumbergeror its affiliates.

© 2007 Schlumberger. All rights reserved.No part of this publication may be repro-duced, stored in a retrieval system ortransmitted in any form or by any means,electronic, mechanical, photocopying,recording or otherwise without the priorwritten permission of the publisher.


4 Oilfield Review

The Science of Tsunamis

Tim BuntingKuala Lumpur, Malaysia

Chris ChapmanPhil ChristieCambridge, England

Satish C. SinghUniversity of CambridgeCambridge, England

Jim SledzikGatwick, England

For help in preparation of this article, thanks to Eric Geist,United States Geological Survey, Menlo Park, California,USA; and Robert Stewart, Texas A&M University, CollegeStation, USA.Q-Marine is a mark of Schlumberger. DART is a registeredtrademark of the US National Oceanic and AtmosphericAdministration (NOAA).

The Sumatra-Andaman earthquake of 2004 produced the deadliest tsunami on record,

alerting the world to the destructive power of this phenomenon. In studying this

tsunami, scientists are using new tools that provide unprecedented insight into the

causes and effects of these events. The knowledge gained from their work will help

improve early-warning systems, mitigating the consequences of future occurrences.

On December 26, 2004, the Sumatra-Andamanearthquake, with an estimated magnitude of 9.3on the Richter scale, was one of the largest everrecorded using modern seismographic equip - ment. As it shook the west coast of Sumatra,Indonesia, and proceeded along a fault line atthe eastern edge of the Indian Ocean, theearthquake generated a tsunami that focused theworld’s attention on the devastating power of thisnatural phenomenon. With estimates of morethan 232,000 deaths and 2,000,000 peopledisplaced in 12 countries in South Asia and East Africa, the impact of the tsunami was truly global.1

In addition to being one of the worst naturaldisasters in human history, the tsunami wasunique in other aspects. It was the first globaltsunami to occur since modern sea-levelmonitoring networks were established and thefirst to be continuously tracked and recorded bya satellite. No other seismic event of thismagnitude has occurred with so many data-gathering sources available. From a scientificperspective, the event provided a wealth ofinformation for analysis. These data will be used to better understand and prepare for future incidents.

The earthquake and tsunami exacted anobservable physical toll—on houses, bridges andbusinesses—that can be seen by comparingbefore and after photographs (next page, bottom).

These images reveal the damage that emanatedfrom events that began below the surface.However, a full understanding of the earth-quake and the sub se quent tsunami requires amultifaceted approach.

To develop an appreciation for the magnitudeof this event—the energy released temporarilyaltered the Earth’s rotation—we present a basicreview of the theory of plate tectonics as itrelates to the earthquake.2 A discussion of thephysics of ocean waves and tsunamis follows. Wealso examine some of the tools used—such asseismic and ocean monitoring networks, land-based global positioning systems (GPS) andtsunami modeling software—to better com -prehend the scope of this event. Details of theWesternGeco tsunami seismic survey will beincluded, along with some preliminary findings.This article also reviews the status of ongoingefforts to develop an integrated monitoring andearly-warning system in the Indian Ocean region.

Tectonic Foundations for a TsunamiOn a geological time-scale, the surface of theEarth is constantly changing—oceans form anddisappear, continents collide with one another,and mountains rise and fall or erode away. Toexplain the processes that shaped and continueto shape the surface of the Earth, the theory ofplate tectonics was proposed.3 It states that theEarth’s lithosphere, the outermost layer, isbroken into rigid plates that are moving relative

1. “Indian Ocean Earthquake & Tsunami Emergency Update December 29, 2005,” Center of Excellence inDisaster Management & Humanitarian Assistance,http://www.coe-dmha.org/Tsunami/Tsu122905.htm(accessed September 27, 2007).

2. Nirupama N, Murty TS, Nistor I and Rao AD: “Energeticsof the Tsunami of 26 December 2004 in the Indian Ocean:A Brief Review,” Marine Geodesy 29, no. 1 (January2006): 39–47.

3. The term plate tectonics was coined by Bryan Isacks,Jack Oliver and Lynn Sykes in a 1968 research paper.Isacks B, Oliver J and Sykes L: “Seismology and the NewGlobal Tectonics,” Journal of Geophysical Research 73(September 15, 1968): 5855–5899.


Autumn 2007 5

> Courtesy of US Geological Survey (USGS).

Sri Lanka

Andaman Islands

Sumatra

> High-resolution imaging satellite photographs of Banda Aceh, Indonesia, before and after the tsunami. Banda Aceh is located at the northern tip ofSumatra. With a population of 260,000, it was the closest major city to the epicenter of the Sumatra-Andaman earthquake. (Photographs courtesy ofDigitalGlobe.)

Before After


to one another, “floating” on the asthenosphere,a hotter, denser, more mobile layer. Below theasthenosphere are the upper mantle, the mantle,the outer core and, at the center of the Earth, theinner core. The major plates have been identifiedand, by plotting seismic activity, their boundarieshave been defined (above).4

Tectonic plates are constantly diverging,converging or transforming. In divergent zones,the plates move away from each other, allowingbasaltic magma to ooze to the seafloor and createthe dense oceanic crust at midocean rift zones.The magma cools as it meets seawater and forms a series of underwater mountain ridgesthat are carried away from the rift by thediverging plates.

Landmasses above sea level form thecontinental crust, which is usually thicker andmuch less dense than oceanic crust. The denseoceanic plate slides beneath the overriding platein what is termed a subduction zone. Eventually,the subducting plate melts and returns to theasthenosphere. As the subducting materialdewaters, the fluid migrates upward, mixing withthe material of the overriding plate, reducing itsmelting point. This produces magmatic melts,rich in dissolved gases, that exert enormousupward pressure on the overriding plate; thesecan erupt if a weakness in the crust develops(next page, top).5

Along boundaries where crust is neithercreated nor destroyed, changes still occur,transforming the surface of the Earth. Over time,as landmasses collide, an ocean that separatedthe masses may disappear, while the previousocean bottom is lifted above sea level. Plates maydeform along their borders into mountain ranges.Landmasses that make up the continental crustmay slide horizontally, creating earthquakes asplates stick and slip.

The Indo-Australian plate, which played a keyrole in the Sumatra-Andaman earthquake,comprises both continental and oceanic crust. Thelandmasses of India and Australia make up themajority of the continental portion, while theoceanic segment lies beneath the Indian Ocean.According to theory (and data), 100 million yearsago, India was an island off the east coast of Africa,south of the equator, and it has been making arelentless journey northward, creating theHimalaya Mountain system along the way. Today,India is penetrating the Eurasian plate at a rate of 45 mm/yr [1.8 in./yr] while slowly rotatingcounterclockwise.6 Mount Everest, the tallest of theHimalayan chain, grows 4 mm [0.1576 in.] per yearbecause of this movement (left).7 The oceaniccrustal portion of the plate is subducting under theBurma microplate and the Eurasian plate.

To the west of Sumatra, the Sunda (or Java)trench marks the edge of the subduction zone.

6 Oilfield Review

> India in motion. India was an island off theeast coast of Africa 100 million years ago. It is part of the Indo-Australian plate and hasbeen advancing into the Eurasian plate as itjourneys northward. During this movement, theHimalaya Mountains were formed along India’snorthern border.

O c e a nI n d i a nLocation of India

70 millionyears ago

Bangladesh

Indiatoday

E U R A S I A N P L A T E

Sri Lanka

Equator

Himalayas

> Plate boundaries defined by seismic activity. The mapping of medium to large seismic events (red) helps identify crustal plate boundaries (yellow). The area known as the Pacific Ring of Fire is the most active region on the planet, with 90% ofrecorded seismic events. By comparison, the Indian Ocean is most active along the eastern edge—especially in the vicinity of the December 2004 Sumatra-Andaman earthquake. [Adapted from an image courtesy of the US National Oceanic andAtmospheric Administration (NOAA).]

Crustal plate boundaries Earthquake epicenters, MW >5, 1980 to 1990

IndianOcean

PacificRing of Fire

Sumatra-Andamanearthquake, 2004


Autumn 2007 7

The trench extends some 3,000 km [1,865 mi],from the Andaman Islands in the northwest tothe Lesser Sunda Islands in the southeast, andhas a depth in excess of 7,700 m [4.8 mi].8 TheBurma microplate is wedged between the Indo-Australian and the Eurasian plates (right). Asthe Indo-Australian plate subducts beneath

4. Oreskes N (ed): Plate Tectonics: An Insider’s History ofthe Modern Theory of the Earth. Boulder, Colorado, USA:Westview Press, 2001.

5. Volcanoes result from these upward flows, creatingconduits through the overriding plate for molten magmato reach the surface.

6. Bilham R: “Earthquakes in India and the Himalaya:Tectonics, Geodesy and History,” Annals of Geophysics 47,no. 2 (2004): 839–858.

7. http://www.nationalgeographic.com/features/99/everest/roof_content.html (accessed October 14, 2007).

8. The Sunda trench was once thought to be the deepestpoint in the Indian Ocean until the 8,000-m [26,250-ft]Diamantina Deep was discovered in 1961.

> The ever-changing face of our planet. According to the theory of plate tectonics, the lithosphere is composed of variously sized rigid plates, which arediverging, converging or transforming along boundaries. At rift zones, plates move away from each other, leaving spaces that are filled with dense basalticmagma rising from the asthenosphere. At convergent plate boundaries, subduction takes place as dense oceanic crust dives beneath the more buoyantcontinental crust, eventually returning to the asthenosphere. Earthquakes occur along these boundaries as stress created by friction between plates is released, often catastrophically. The sudden movements of submerged plates play an important role in the generation of tsunamis. Bathymetry data(inset) from a section of the December 2004 earthquake zone shows the Indo-Australian plate subducting beneath the Burma microplate. A trench forms at their boundaries.

Shield volcanoOceanic spreading ridge

Divergent boundary(rift zone)

Continentalrift zone

Convergent plateboundary

Trench

LithosphereOceanic crust

AsthenosphereSubducting plate

Continental crust

Burma

microplate

Convergent

plate boundary

Indo-Australianplate

Tren

ch

Depth indication, m

1,0004,000

> Tectonics of the Sumatra-Andaman earthquake. The eastern edge of the Indo-Australian plate issubducting beneath the Eurasian plate and Burma microplate at a rate of 52 mm/yr [2.05 in. /yr]. TheIndo-Australian plate is moving northward while slowly rotating counterclockwise. The December2004 Sumatra-Andaman earthquake began at the epicenter (star) and continued north for 1,200 km[745 mi] along the fault line (blue), terminating at the Andaman Islands. Boundaries of plates (triangles)and microplates (gray lines) are indicated.

Eurasianplate

Burmamicroplate

Indo-Australianplate

December 26, 2004

Andaman Islands

Sumatra

km

miles 1,000

0

0

1,000


these plates, stresses build when the platesbecome stuck. Because the plates continue tomove, the time between major earthquakes andthe extent of the area where their relativemotions are constrained determine the potentialearthquake severity.

Although the Indian Ocean has its share ofearthquakes in seismically active zones, theboundaries of the Pacific Ocean are actually themost active in the world, with 90% of allearthquakes—80% of the major ones—occurringwithin the Pacific basin. The primary mechanismfor this seismic activity is the movement of thesubducting plate described above.9

Because the Pacific basin is so seismicallyactive, an extensive network of sensors has beenestablished for earthquake and tsunamidetection. Although there were plans to developa system modeled after the one used in thePacific, at the time of the tsunami, there was nosuch network for the Indian Ocean. Largetsunamigenic events were infrequent, with onlyone major tsunami occurring there during theprevious century and only four reported in the1800s. The tsunami created by the well-knowneruption of Krakatoa in 1883, and by its ensuingcollapse, was one of those four. Historical datacombined with the high level of seismic activitysuggested a likelihood of tsunamis occurring in the region, but nothing on the scale of thetsunami of 2004 was anticipated.10

Making Waves Ocean waves—tsunamis being one category—are classified as gravity waves. Although themechanisms that generate them are different,the physics that describe gravity waves areapplicable to those in a pond, on the open oceanor after a significant impact such as the Sumatra-Andaman earthquake. To understand tsunamis,it is essential to recognize how they aregenerated and how they differ from wind-generated waves.

Most ocean waves are primarily generated bywind turbulence creating friction along thesurface of the water. Turbulence produces ripplesthat are capillary waves—waves that travelbetween two fluids. Gravity and surface tensionpull the peaks of the ripples back towardequilibrium, but the ripples overshoot theoriginal level of the water, causing the surface tooscillate. Should the wind stop, the oscillationswill die out due to friction. Once the oscillationshave a wavelength greater than 2 cm [0.8 in.],wind-induced ripples can become gravity waves.This occurs at the point where the effects ofgravity are greater than the effects of surfacetension. Dispersion from gravity cancelsdispersion caused by surface tension of thewater, resulting in a radiating wave that has thepotential for traveling great distances. As windcontinues providing energy to the waves, theperiod, wavelength and speed increase, and the

resulting waves can even travel faster than thewind that generated them.

Waves can travel great distances, oftengaining strength and speed by combining withother waves or by the addition of more windenergy. A wave in Hawaii might have begunduring a storm in Alaska, arriving on the beachwith little loss of speed or energy. Although thewave began many miles away, the molecules ofwater were not displaced any great distance untiljust before the wave reached the shore.

In deep water, if the wavelength is muchshorter than the water depth, the motion of thewater can be described as circular during thetrough-peak-trough cycle. In shallow water, orwhen the wavelength is greater than the waterdepth, the motion is more elliptical, with the ratioof the horizontal to vertical motions proportionalto the ratio of wavelength to depth. For a tsunami,because of its long wavelength, this occurs evenin the deep ocean, and the horizontal motion canbe much greater than the vertical motion. At theshore, the elliptical motion transforms into for -ward motion, and the water molecules advancewith the wave (below).

In the ocean, with all its variability, wavemotion is more complex. Gravity, tides, cross -winds, submarine and shoreline features, waterdepth and wave arrivals from various angles willact upon the wave to affect wave height, speedand direction. Because of the long distances

8 Oilfield Review

> Wave basics. Wind-generated swells move across the surface of the ocean. The water molecules generally have acircular motion that becomes more elliptical as the wave approaches the shore. The velocity of a wave slows as itapproaches the shore, forcing the water upward. The tip of the wave continues moving faster than the base until itreaches the surf zone, where the peak of the wave breaks over due to gravity.

Wave height increases Surf zone

Orbital path ofwater molecules

Elliptical path


Autumn 2007 9

open to wave travel in the oceans, the simplewave train can develop into swells, which arelong-wavelength waves. As the swells reachshallow-water depths, they rise higher than theywere when over deep water and form peaks.These peaks will eventually break over becauseof the steepness of the wavefront, the pull ofgravity and the peak moving faster than the baseof the wave.11

Whether the water movement is created bythe wind, the sudden movement of the seafloorduring an earthquake, the downward force froma landslide or even the impact from an asteroid,these forces all generate oscillatory motion thattranslates into gravity waves. A tsunami differsfrom waves produced by the wind in that it is animpact-generated wave, deriving its speed andpower from the event that created it. Largeimpact-generated waves also have extremelylong wavelengths. Tsunamis can have wave -lengths in excess of 100 km [62 mi], whereaswind-generated swells have wavelengths on theorder of 150 m [500 ft].

Wavelength is a useful characteristic forclassifying wave types. A shallow-water gravitywave is characterized by the fact that the ratiobetween the water depth and the wavelength isquite small. These waves travel at a speed that isequal to the square root of the product of the

acceleration due to gravity (9.8 m/s2) [32 ft/s2]and the water depth. Because of a tsunami’s longwavelength, it acts like a shallow-water waveeven in deep water, and its speed can beapproximated if the water depth is known. With awater depth of 7,700 m, the Sunda trench was aperfect incubator for a fast-moving tsunami,which attained speeds of more than 900 km/h[560 mi/h], rivaling the speed of a moderncommercial jetliner.

Not only do tsunamis travel at high rates ofspeed, they maintain their wave height, oramplitude, for great distances. The amplitudes ofwater waves decay as they propagate for threereasons: the waves spread out over the surface ofthe water; the waves disperse because longerwavelengths travel faster; and energy isattenuated by viscous damping in the water. Fora large tsunami, all three effects are minimal.Since the energy for initiation occurs along anextended fault, the waves spread out linearlyrather than cylindrically, resulting in littlespreading. For extremely long wavelengths, thewaves are not highly dispersive because thevelocity is proportional to the square root of thewater depth, resulting in little dispersion in theopen ocean. Attenuation loss is inversely relatedto the wavelength, and thus there is littleattenuation. As a result, a tsunami propagates athigh speeds and travels great distances withlimited energy loss.

As a wave moves into shallow water, thepropagation speed developed in deeper watercannot be maintained. For a tsunami thatoriginally traveled at 900 km/h in deep water, themaximum sustainable velocity would be less than50 km/h [31 mi/h] in a water depth of 10 meters[33 ft]. Energy continues pushing the waveforward, leaving only one direction for the waterto go—upward. Wave height on shore, or run-up,of 35 m [115 ft] was reported on the island ofSumatra (above).

Ironically, the tsunami would have beenhardly noticed near the epicenter of the quake. Arise in ocean levels would have felt like a largerthan average swell. For example, theWesternGeco survey vessel Geco Topaz wasacquiring seismic data off the coast of India1,500 km [930 mi] from the epicenter. Thetsunami passed under the vessel 2 to 3 hoursafter the initial earthquake and was only a fewtens of centimeters in height—in the open waterof the Indian Ocean.

9. Volcanic activity around the subduction zones has resultedin the area being known as the Pacific Ring of Fire.

10. For an in-depth review of plate tectonics, see theSchlumberger SEED Web site: http://www.seed.slb.com/en/scictr/watch/living_planet/index.htm (accessedAugust 18, 2007).

11. Stewart RH: Introduction to Physical Oceanography.College Station, Texas: Texas A&M University, 2005.http://oceanworld.tamu.edu/resources/ocng_textbook/(accessed September 17, 2007).

> A tsunami approaching the shoreline. When the tsunami arrives at the shore, its velocity decreasesrapidly and its height increases and rises well above the average sea level. The original long-wavelengthwave becomes somewhat shorter at the coastline. The distance the wave travels inland—inundation—and the height of the wave at the shoreline—run-up—are determined by coastal geometry and thecharacteristics of the individual tsunami. Contrary to popular belief, a tsunami rarely has a break-over,rising much like a fast-moving tide. After the wave inundates the low-lying coastal regions, the out-rush of water returning to the ocean carries debris from inland. Since the tsunami is actually a seriesof waves, subsequent surges return the debris, acting like battering rams along the coastline.

WavelengthMean sea levelCrest

Trough

Waveamplitude

Run-up

Wavelength


A Wakeup CallAt approximately 8 a.m. local time onDecember 26, 2004, the Sumatra-Andamanmegathrust fault earthquake began. The largestrecorded earthquakes have been along thrustfaults, where subducting and overriding platessuddenly shift to relieve built-up stresses. Over aneight-minute period, the rupture traveled fromthe epicenter off the coast of Sumatra, northwardalong the fault plane for about 1,200 km [745 mi]as the Indo-Australian plate slipped beneath theBurma microplate. This long section of lockedplates broke apart and the overriding plate, nolonger constrained, heaved upward.

Not all earthquakes produce tsunamis; itrequires the right set of circumstances. In thiscase, the fault plane of the earthquake extendedfrom 30 km [19 mi] below Sumatra to theseafloor of the Indian Ocean. From a surfacedamage standpoint, an earthquake centered in

the ocean might seem fortuitous. However, thislocation facilitated direct transfer of energy from the plate movement to the water. With a1,200-km long fault plane, a subduction zonethickness of 500 m [1,640 ft], and a verticaldisplacement of 5 to 15 m [16 to 50 ft], the upliftof the overriding plate and downdrop of thesubducting plate sent water oscillations travelingaway from the source of the energy, initiating atremendous tsunami (below).

Within 15 minutes of the quake, the tsunamiarrived along the Sumatra shoreline. There waslittle warning of its approach, although it is likelythat because of its proximity, the earthquakewould have been felt by those living in the region.The first indication of an approaching tsunamiwas probably a forerunner, a swell ahead of thelarger waves.12 Preceding the forerunner wouldbe a sudden out-rush of water, exposing largesections of the nearshore seabed. Based on

eyewitness accounts, this oddity drew people outalong the exposed seafloor, placing them in thepath of the approaching wave.13 Several minutespassed, and depending on the distance from thesource and speed of the tsunami, the first waveinundated the exposed beach and rushed inlandto flood the low-lying coastlands. The dangerdoes not end with the first wave, since the thirdto eighth waves generally are even larger. In SriLanka, arrival of the surges came at approxi mately40-minute intervals, indicating a wave length inthe hundreds of kilometers.14

A Measure of Perspective For the general public, earthquakes are oftenclassified using a magnitude based on the well-known Richter scale. Seismologists use moremeaningful measures such as the momentmagnitude scale. Richter and moment magni -tude are logarithmic measures of the amplitudeobserved on seismograms and are related to theenergy released in an earthquake.

Dr. Charles F. Richter developed his scale toquantify earthquake magnitude, and it isdesignated ML, with the L referring to local. Bycomparing the seismic data for numerousCalifornia earthquakes as measured by shearwaves recorded on a Wood-Anderson seismom e -ter, Richter correlated the amplitude of themeasured signal to the size of the earthquake.The Richter magnitude is the logarithm of thepeak amplitude of the seismic record, with adistance correction applied. Since it is a loga -rithmic scale, each whole number on theWood-Anderson seismometer represents anamplitude 10 times greater than the lesser wholenumber. Because the energy is proportional tothe square of the amplitude, and larger earth -quakes radiate more low-frequency energy notrecorded by the Wood-Anderson seismom e ter,each whole number in the magnitude scaleactually represents about a 30-fold increase inenergy for very large earthquakes.

Moment magnitude, MW, more accuratelydescribes the physical attributes of an earth -quake and is used by modern seismologists,especially when ranking large earthquakes.Moment is a function of the total energy releasedand is a physical quantity proportional to the slipdistance and the average slip area along the faultsurface. Seismic data are used to estimate themoment and then converted, using a standardformula, into a number representative of otherearthquake measurements, such as the Richtermagnitude.15 Depending on the source quoted,the Sumatra-Andaman earthquake received a 9.0to 9.3 MW rating.

10 Oilfield Review

> A tsunami-generating earthquake. The Indo-Australian plate is slidingbeneath the Burma microplate along a subduction zone, developingstresses between the plates (top). The overriding plate became stuck andbuckled upward. The rupture relieved the stress created by the lockedplates and upward buckling (dashed line) and caused the overriding plateto move upward and outward (middle). It heaved an estimated 5 to 15 m,raising the overlying water, and initiating the tsunami (bottom). The rupturezone was more than 1,200 km in length.

Subducting plate

Overridingplate

Area ofplate sticking

Stuck arearuptures

Tsunami begins

Tsunami wavesspread


Autumn 2007 11

Within minutes after the earthquake, reportswere issued from seismic monitoring stationsaround the globe. The first magnitude estimatewas 6.2 MW, using the arrivals of early body wavesmeasured at the reporting station in Hawaii. Body-wave magnitudes are known to under estimatevery large earthquakes. A preliminary-magnitudereport (8.5 ML) was issued by the United StatesGeological Survey (USGS) and Pacific TsunamiWarning Center (PTWC) one hour and 15 minutesafter the event, which was as soon as sufficientsurface wave data were available. Estimates werelater increased to 9.1 MW, which is the estimatepublished by the USGS.16 Post-earthquake analysishas put the figure as high as 9.3 MW, but there isno figure for which a consensus has beenreached.17 Much of the difficulty is due to relatingthe seismic information to the volume of earththat moved.

The magnitude of an earthquake is crucialbecause the strength of the initiating event is acritical component of the modeling programsused to predict tsunami generation. A 6.2 MW

earthquake would not have generated a tsunamibulletin. The PTWC’s report was upgraded assoon as information became available, but thediscrepancy underscores the difficulty inherentin an early-warning system.

Data are available from sources other thanseismic monitoring stations, and an earthquakeof this magnitude has never been scrutinizedwith such an array of scientific tools. With anetwork of approximately 60 GPS monitoringstations in the vicinity of the earthquake,accurate ground movement could be quantified.The GPS network was part of an ongoingcollaborative project, Southeast Asia: MasteringEnvironmental Research Using Geodetic SpaceTechniques (SEAMERGES), with additional GPSdata coming from monitoring stations with theInternational GPS Service. The GPS dataprovided the actual earth displacementinformation, which was then used to estimateenergy released in the earthquake—but thiscould not be accomplished in real time.Reconciling the data from the seismicmonitoring and the GPS stations resulted inassigning a magnitude of 9.3 MW to the Sumatra-Andaman earthquake.18

Looking DeeperWithin days following the earthquake, humani -tarian relief poured into the region surroundingthe Indian Ocean. Individuals and organizationsaround the world offered help in the form ofdonations and services. Schlumberger made athreefold promise of funding, volunteers andtechnology. The funding and volunteers cameimmediately, addressing human aspects of thetragedy. On the technology front, one projectquickly emerged: a deep seismic survey along thefault line to improve the understanding of thecomplex tectonics in the region of the earthquake.Previous surveys, using academic researchvessels, could not image structures at 30 km, thedepth inferred from historical seismic activity.Understanding the distribution and geometry ofthe faults that control seafloor displacement iscritical in determining the mechanisms thatgenerated the tsunami.19

This is not the first time Schlumberger hasbeen an active participant in earthquake-relatedscientific studies. The San Andreas Fault

Observatory at Depth (SAFOD) Project incor -porated many oilfield technologies in theassessment of the seismically active San AndreasFault.20 The ability to deploy, acquire and analyzedata using tools developed for oil and gasexploration has been invaluable in understandingthe mechanisms that generate seismic events in regions such as the Sumatra-Andamanearthquake zone.

WesternGeco committed resources to acquireand process the data for the Sumatra EarthquakeDeep Seismic Reflection survey, or “the tsunamisurvey.” The vessel Geco Searcher was used forthe acquisition of seismic data (above). Inconjunction with Schlumberger CambridgeResearch in England and Institut de Physique duGlobe de Paris in France, WesternGeco donatedits services, including logistical and technicalsupport. The survey was conducted cooperativelywith the Indonesian Agency for the Assessmentand Application of Technology, which retains the rights to the data. In the future, WesternGeco plans to make its data available to

12. A forerunner is a series of oscillations of the water levelpreceding the arrival of the main tsunami waves.

13. Barber B: Tsunami Relief. US Agency for InternationalDevelopment, Bureau for Legislative and Public Affairs(April 2005): 4. http://www.reliefweb.int/library/documents/2005/usaid-tsunami-30apr.pdf (accessedOctober 31, 2007).

14. Cyranoski D: “Get Off the Beach—Now!,” Nature 433,no. 7024 (2005): 354–354.

and Ambrosius BAC: “Insight into the 2004 Sumatra–Andaman Earthquake from GPS Measurements inSoutheast Asia,” Nature 436, no. 7048 (2005): 201–206.

19. Singh S: “Seismic Investigation of the Great Sumatra-Andaman Earthquake,” First Break 24, no. 12 (December 2006): 37–40.

20. Coates R, Haldorsen JBU, Miller D, Malin P, Shalev E,Taylor ST, Stolte C and Verliac M: “Oilfield Technologiesfor Earthquake Science,” Oilfield Review 18, no. 2(Summer 2006): 24–33.

15. Hanks T and Kanamori H: “A Moment Magnitude Scale,”Journal of Geophysical Research 84, no. B5 (1979): 2348–2350.

16. http://earthquake.usgs.gov/eqcenter/eqinthenews/2004/usslav/#summary (accessed August 22, 2007).

17. Ishii M, Shearer PM, Houston H and Vidale JE: “Extent,Duration and Speed of the 2004 Sumatra–AndamanEarthquake Imaged by the Hi-Net Array,” Nature 435,no. 7044 (2005): 933–936.

18. Vigny C, Simons WJF, Abu S, Bamphenyu R, Satirapod C,Choosakul N, Subarya C, Socquet A, Omar K, Abidin HZ

> The Geco Searcher in action. The WesternGeco vessel Geco Searcher acquired the data for theSumatra Earthquake Deep Seismic Reflection survey. The data will be made available for futureacademic research.


the global academic community for additionalscientific analysis.

The survey was part of a larger initiative, theSumatra-Andaman Great Earthquake Research(SAGER) project, which included high-resolutionsea-bottom bathymetry and an ocean-bottomseismometer (OBS) refraction survey deployedby the French research vessel MarionDufresne.21 The Institut Polaire Français madethe Marion Dufresne available for the survey andprovided technical support. OBS sensors wereplaced on the seabed to record seismic activity(below left).

In July 2006, the Geco Searcher acquiredthree seismic lines, totaling 926 surface km[575 mi] of deep seismic profiling (below right).The seismic survey had several objectives:• image active faults along the subduction zone• quantify the volume of water that penetrated

along these faults• provide information to optimize the location

of a future borehole for the Integrated OceanDrilling Program.22

Providing an image of faults at a depth of30 km required long offsets.23 In the oil and gasindustry such depths would not be consideredbecause they are beyond the reach of any drillingoperation. The Geco Searcher used the Q-Marinesingle-sensor marine seismic system to providethe technology needed to acquire 12-km [7.5-mi]

offsets with a single-vessel operation. The sourceand streamer depths were maximized for theacquisition of low-frequency data, and aftermodeling and analysis, the decision was made totow sources and streamers at a depth of15 meters. An additional shorter streamer wastowed at 7.5 m [25 ft] to provide high-resolutionimages for defining features nearer the surface.Compared with surveys used in oil and gasexploration, this survey design was elaborate andextensive: tripled streamer depth, tripledstreamer length, tripled energy source andtripled recording time (next page, top).

Concurrent with the seismic survey, theFrench research vessel Marion Dufresnedeployed 56 ocean-bottom seismometers alongthe route of two of the seismic lines. The widelyspaced OBS sensors recorded naturally occurring

seismic activity but were also able to acquireseismic data during the WesternGeco acquisi -tion. Using 5- to 20-km [3- to 12-mi] spacing, thesensors recorded the shots from the survey andthe reflections from the subducting layer. Theseismic reflection data from the WesternGecooperations and the refraction data from the OBSsensors are complementary because the reflec -tion data provide high-resolution images of thecrust, and the OBS refraction data providedeeper images of the crust and upper mantle.24

The volume of data acquired is massive.Preliminary processing and analysis were carriedout by WesternGeco staff aboard the GecoSearcher, and later on shore in Indonesia, butmore analysis will be required to identify thesignificant features and fully utilize the data(next page, bottom).

12 Oilfield Review

> Deploying an OBS. The research vessel Marion Dufresnedeployed 56 ocean-bottom seismometers along the path ofthe WesternGeco seismic survey. Intended for monitoringseismic activity at the seafloor, the OBSs were used torecord reflections from the sources used by WesternGeco.(Photograph courtesy of First Break, reference 19.)

> The survey area. In the vicinity of the Sumatra-Andaman earthquake, three seismic lines(WG1, WG2 and WG3), totaling 926 surface kilometers, were acquired. Preliminaryprocessing has provided high-resolution imaging to depths greater than 30 km. The mapalso contains bathymetry data for the area that was under study.

Depth, m 0

442

972

1,354

1,680

1,985

2,616

2,290

2,998

3,528

5,216

Sumatra fault

Sumatra

Simeulueplateau

December 26, 2004epicenter

Simeulue

52 mm/yr

WG1

Deformationfront

Indo-Australian

plate

WG2

West Andaman fault

Burmamicroplate

Aceh basin

Nicobar

1000 miles

0 100km

WG3


Autumn 2007 13

The seismic data, along with SAGERbathymetry and refraction data, are being usedto understand the features that control platemovement. Preliminary analysis of the dataconfirmed that a fault plane, from theearthquake epicenter at 33 km, extends to theseabed. The seismic images validated the

21. Bathymetry is the surveying or mapping of harbors,inlets or deepwater locations. Echo sounder techniquesare used in the measurement and study of water depthsto create bathymetric maps or charts of seafloor relieffor navigation purposes.

22. For more on the Integrated Ocean Drilling Program:Brewer T, Endo T, Kamata M, Fox PJ, Goldberg D,Myers G, Kawamura Y, Kuramoto S, Kittredge S,Mrozewski S and Rack F: “Scientific Deep-OceanDrilling: Revealing the Earth’s Secrets,” Oilfield Review16, no. 4 (Winter 2004/2005): 24–37.

23. Offsets are the distance between the airgun array andthe sensors.

24. Singh, reference 19.

> Seismic images from two streamer depths. The image from the 7.5-m streamer (top) shows finer details nearer the surface. The image from the15-m streamer (bottom) uses deeper penetrating seismic energy. Features deeper than 30 km can be studied using these data.

Tim

e, s

1.5

2.0

2.5

3.0

Tim

e, s

1.5

2.0

2.5

3.0

> Preliminary results. From the WG1 seismic line, preliminary interpretation reveals faulting anddeep boundaries. The main thrust fault can be seen on this image, as well as other reflectors.The Moho, short for the Mohorovicic discontinuity, is the boundary between the Earth’s crust andthe mantle, and can be identified here.

Simeulue fore-arc basinSimeulue plateau

Accretionary wedgeSW

Tim

e, s

16

14

12

10

8

6

4

2

0

Active main thrust fault

Active frontal thrust

Oceanic Moho

West Andaman fault

Backthrust

ContinentalMoho

NEThrust reflectors

WG1

0 miles

0 km 25

25


premise that a large upheaval of the seabedcontributed to the strength of the tsunami(above). Early analysis has also identified a verywide locked zone, greater than 135 km [85 mi],whose rupture contributed to the magnitude ofthe earthquake.25

On September 12, 2007, an 8.4 MW earth -quake occurred on the December 2004 fault line,but produced relatively little tsunami energy(above right). Scientists can use the seismicimages and information acquired during bothearthquakes to better understand the mecha -nisms that initiated the earthquakes andproduced (or failed to produce) a large tsunami.Ultimately, the information can be integrated intomodeling programs to improve tsunami forecasts.

Subduction zones typified by the area thatcreated the Sumatra-Andaman earthquake existin other places around the world. Technologysuch as the Q-Marine system can be appliedelsewhere to better understand seismically activeregions. Collaboration between the academicworld and companies like Schlumberger willequip scientists and researchers with advancedtools to prepare at-risk locations.

Moving Towards Early WarningThe following is a timeline of the early eventsthat occurred December 25, 2004, at the NationalOceanic and Atmospheric Administration(NOAA) Pacific Tsunami Warning Center(PTWC) in Honolulu, Hawaii:

• 2:59 p.m. local time, the Sumatra-Andamanearthquake begins

• 3:07 p.m., first seismic arrivals detected at thePTWC

• 3:10 p.m., PTWC issues an alert that a 8.0 MW

earth quake has occurred near Sumatra,Indonesia

• 3:14 p.m., PTWC issues bulletin 1—no tsunamithreat to Pacific Ocean basin. There was noestablished protocol to contact other regions.

• 3:15 p.m., first tsunami wave strikes Sumatra.As per standard operating procedure, a text

message was distributed to participants of theTsunami Warning System (TWS) in the Pacific,and e-mail notification was sent to 25,000interested parties. Alerts were issued bytelephone to various agencies, including theHawaii Civil Defense and the InternationalTsunami Information Center.26

With 80% of major earthquakes occurringaround the Pacific Ocean, it is critical to have aneffective tsunami early-warning system thatoperates as described above. The PTWC is justone part of a cooperative network coordinated bythe Intergovernmental Oceanographic Commission(IOC), functioning under the United NationsEducational, Scientific, and Cultural Organi -zation (UNESCO).27 The Pacific TWS compriseshundreds of seismic monitoring stationsworldwide, sophisticated tsunameters monitor -ing wave heights in the open ocean andstrategically placed tidal gauges (next page, top).

Various organizations representing 26 countriesfrom that region collaborate to alert the publicwhenever the danger of a tsunami is present.

By their very nature, warning networks suchas the Pacific TWS are expensive, having tocontend with vast stretches of open water,expensive monitoring equipment on land and inthe oceans, and the need for continuous staffingof monitoring stations with qualified personnel.The events of December 2004 demonstrate justhow costly the lack of an early-warning systemcan be. The Pacific Tsunami Warning Center iswell established and is the model for the IndianOcean Tsunami Warning Center (IOTWC). ThePTWC relies on four primary tools: seismicmonitoring, ocean monitoring, fast modelingsoftware and communication.

Listening to the EarthThree key earthquake parameters can bedetermined from seismic waveform data topredict an earthquake’s tsunamigenic potential:• location—whether the earthquake is located

under or near the sea• depth—whether the earthquake is located

near enough to the Earth’s surface to createsignificant displacement

• magnitude—whether the size of the earth-quake is sufficient to produce a tsunami.

14 Oilfield Review

25. Singh, reference 19.26. http://www.noaanews.noaa.gov/stories2004/s2358.htm

(accessed August 18, 2007).27. http://ioc3.unesco.org/itic/ (accessed September 27, 2007).

28. ICG/IOTWS-II, Communications Plan for the InterimTsunami Advisory Information Service for the IndianOcean Region, ver. 15, January 2006. http://ioc3.unesco.org/indotsunami/documents/IOTWS_CommunicationPlan_15Jan06.pdf (accessed October 25, 2007).

> Detailed interpretation of the seismic data. The epicenter of the December 26, 2004 earthquake wasbeneath the Simeulue plateau, located west of Sumatra. The earthquake occurred when the continentalplate broke free of the oceanic plate along the subduction zone (red line). The zone extends more than150 km [93 mi] from the epicenter to the ocean floor. (Adapted from Singh, reference 19.)

Dept

h, k

m

50

0

5

10

15

20

25

30

35

40

45 500 miles

0 50km

Oceanic mantle

Oceanic Moho

Sediments

Upperseismogenic zone

Mantle wedge

Continental Moho

December 26, 2004

Backthrust

Subducting oceanic crust

Accretionarywedge Simeulue plateau

West Andaman faultSimeulue fore-arc basinIndo-Australian plate

Frontal thrust fault

Crustal-scalethrust fault

Main thrust faultIndicates motion

into pageIndicates motionout of page

> Two major earthquakes, with very differentresults. The epicenter of a September 12, 2007earthquake, 8.4 MW, was in the vicinity of theDecember 2004 Sumatra-Andaman earthquake,9.3 MW. Although the 2007 earthquake waspowerful enough to generate a tsunami, therupture did not extend from the epicenter as itdid in the 2004 earthquake (red). The smalltsunami produced during the 2007 earthquakehad little effect on the region.

Eurasianplate

Burmamicroplate

Indo-AustralianPlate

December 26, 2004

Andaman Islands

Sumatra

km

miles 1,000

0

0

1,000

September 12, 2007


Autumn 2007 15

Seismic monitoring is primarily accomplishedusing monitoring stations supported by variousgovernmental agencies and educationalinstitutions. The Global Seismographic Network(GSN) is a primary source of data. It comprises225 monitoring stations in more than 80 coun -tries. In addition, the PTWC and the IOTWCreceive data from other seismic monitoringnetworks such as the International MonitoringSystem (part of the Comprehensive Nuclear TestBan Treaty Organization) and those coordinatedby the Incorporated Research Institutions forSeismology (IRIS).

The warning centers receive seismic dataover the Internet. However, because reliabletransmission of data from the Internet is notguaranteed, especially in the case ofinfrastructure damage during and after a majorearthquake, additional sources for data areavailable. The Matsushiro Seismic Array Systemof Matsushiro Seismological Observatory(Nagano, Japan) and the Large Aperture Arraycomprising Japanese seismological observation

networks—are examples of the warning centers’contingent data sources.

When a seismic event occurs, data areprocessed at the warning centers to evaluate thepotential for a tsunami. The warning centers use

an established criterion, based on the magnitudeof the earthquake, to decide which type of bulletinto issue (above). A reliable location can bedetermined using the least-squares method, withP-wave arrival times and various reflected phasesused to provide epicenter depth estimations.28

> Global Seismographic Network (GSN). With a large number of seismic monitoring stations, the GSN comprises a multinational, multidisciplinary networkof cooperating research seismometer stations, including those affiliated with the Incorporated Research Institutions for Seismology (IRIS). The network, asof April 2007, includes the following stations: 86 operated by the United States Geological Survey (USGS), 39 operated by International Deployment ofAccelerometers (IDA), a global network of broadband and very long period seismometers, and other affiliated stations. The University of California SanDiego (UCSD), CU in the legend, is a major participant in the network, with funding from the National Science Foundation. For more on GSN, IRIS, UCSDand IDA: http://www.iris.edu/. (Modified from Global Seismic Network, http://www.iris.edu/about/GSN/map_family.html.)

BBSRDWPF

WVT

CCMWCI SSPA

HRV

FFC

ANMO

TEIGSUG

SDV

RSSDCOR

TUCHKT

JTS

PFOPASCMB

COLA

KDAKADK

H2O

KIPPOHA

MIDWSLBS

JOHN

XMASKANT

AFI

RAP

RAOPTON

RPN

PAYG OTAV

NNALPAZ

LCO

LVC

PTGA

SAML

CPUP

TRQA

BDFB

RCBR

SACV

CMLA

MACI

KOWADBIC

ASCN

MSKU

BGCAMBAR

SHELTSUM

LSZ

PLCA

EFI

PMSA HOPE

TRIS

LBTB

SURBOSA

QSPAVNDA

SBA

CASY

ABPO

KMBO

FURI

DGAR

MSEYPALK

COCO

NWAO

MBWA

BTDF

TAU SNZO

WRAB

KAPI

PMG

CTAOMSVF

HNRFUNA

TARA

KWAJDAV

QIZ

GUMO

WAKE

TATO

ERM

MAJO

SSEENH

XAN

INCNBJT

MDJ YSS

PET

BILL

YAKMA2

HIATLYULNWMQ

NILLSA

KMI

CHTO

UAERAYN

ABKT

AAKMAKZ

KURK

GNIKIV

BRVKARU

NRILTIXI

OBNKIEV

ANTO

GRFO

PAB

BFODPC

LVZKEV

KGNO

ESKBORG

KBS

SEJD

ALE

Installed Planned

IRIS/USGS stationsIRIS/IDA stationsUSGS/CU stationsAffiliated GSN stations

> Tsunami bulletin criteria. The Tsunami Warning Centers use magnitude, location (under sea orunder land) and depth of the earthquake to determine the potential for a tsunami and issue bulletinsbased on those criteria. (Source of data is reference 28.)

EarthquakeDepth

EarthquakeLocation

EarthquakeMagnitude, Mw

Description of Tsunami PotentialBulletin

Type

Very small potential for adestructive tsunami

Potential for a destructivelocal tsunami

Potential for a destructiveregional tsunami

Potential for a destructiveocean-wide tsunami

No tsunami potential

No tsunami potential

Tsunamiinformation

Local tsunamiwatch

Regionaltsunami watch

Ocean-widetsunami watch

Tsunamiinformation

Tsunamiinformation

Under or verynear the sea

< 100 km

≥ 100 km

Inland

All locations

≥ 7.9

≥ 6.5

≥ 6.5

7.6 to 7.8

7.1 to 7.5

6.5 to 7.0


The seismic data are the first piece of thepuzzle. If an earthquake is sufficiently large,takes place in a shallow portion of the Earth’scrust, and occurs in a location under or close tothe sea, it has the potential for generating atsunami. Whether or not a tsunami has actuallybeen created can be determined only at theocean’s surface.

The Ocean’s PulseIdentifying the formation of a tsunami andaccurately forecasting its arrival times and waveamplitudes depends on precise ocean-levelmonitoring. This is accomplished using twoprimary sources—NOAA’s DART Deep-OceanAssess ment and Reporting of Tsunamis buoys indeep water and tidal gauges near coastlines.Although DART buoys have been deployedglobally, the Pacific Ocean has the majority, with28 DART buoys in place, and four more to bedeployed by the end of 2008 (right). The DARTbuoy consists of an anchored seafloor bottom-pressure recorder (BPR) and a tethered surfacebuoy that provides real-time communications(next page). An acoustic link transmits

16 Oilfield Review

> Buoy network for monitoring ocean activity. The Pacific Ocean is encircled by DART Deep-OceanAssessment and Reporting of Tsunamis monitoring buoys, with more planned. The network suppliesinformation to the Pacific Tsunami Warning System. NOAA operates the majority of the buoys,although a few are maintained by other agencies. As of October 2007, two DART buoys are active inthe Indian Ocean. (Adapted from NOAA, http://www.ndbc.noaa.gov/dart.shtml.)

DART Locations

4 PlannedNOAA34

Other3

> Global Sea Level Observing System (GLOSS). With more than 290 sea-level monitoring stations, GLOSS is at work around the world monitoring long-termclimate change and oceanographic sea-level variations. In the event of a tsunami, these data are incorporated into modeling software to refine forecastsand inundation estimations.

GLOSS Tidal-Gauge Locations


Autumn 2007 17

temperature and pressure data from the BPR tothe surface, which are converted to an estimatedsea-surface height. The accuracy of themeasurement is ± 1 mm in 6,000 m [20,000 ft] ofwater depth. These data are transmitted to anIridium commercial satellite that relays theinformation to moni toring stations. Turnaroundtime for data is less than three minutes, frombuoy to warning center.29

Tidal gauges record coastal sea-levelvariations using an international network ofmonitors. The Global Sea Level Observing System(GLOSS) is a network of more than 290 sea-levelmonitoring stations coordinated under theauspices of the Joint Technical Commission forOceanography and Marine Meteorology(JCOMM) of the World MeteorologicalOrganization (WMO) and the IntergovernmentalOceanographic Commission (IOC). GLOSSprovides high-quality global and regional sea-level data for application to climate,oceanographic and coastal sea-level research(previous page, bottom).30

A Model ForecastWhen a seismic or other event of sufficientmagnitude triggers the need for tsunami modeling,various software programs may be used to estimatea tsunami’s potential severity. The seismicinformation is the initial source, but real-time sea-level data are incorporated into the model as theybecome available. These modeling programsprovide estimated wave-arrival time, and wave-height and inundation patterns. It is critical that asimulation model be able to provide accurateforecasts as rapidly as possible. The elapsed timeof 15 minutes between the earthquake and thefirst wave arrival in Sumatra underscores the needfor speed in model predictions.

The United States National Oceanic andAtmospheric Administration (NOAA) has devel -oped a cutting-edge modeling program, known asMethod of Splitting Tsunami (MOST).31 The MOSTprogram uses a suite of numerical simula tioncodes to compute predetermined wave behaviorfor three stages of a tsunami—generation,propagation and run-up. The program can providecoarse grids in deep water, where the wavelengthis long and fewer node points are needed. Inshallow water, the tsunami wavelength shortensand the amplitude rises. To better model thewave, the program narrows its focus to high-resolution grids.

The early-warning system issues alerts andnotifications to potential at-risk areas based onthe MOST outputs. The MOST program is first runin a research mode to create scenarios using

predetermined inputs—such as earthquakemagnitude, directionality and location. Thesesimulations can take hours to run, which wouldbe inappropriate for an early-warning system. To

speed the process, when an earthquake isdetected, the software attempts to match thereal-time data to a preexisting scenario topredict the likelihood and potential of a tsunami.

29. http://nctr.pmel.noaa.gov/Dart/dart_home.html (accessedOctober 1, 2007).

30. http://www.gloss-sealevel.org/ (accessed October 18,2007).

> NOAA’s DART II system. Anchored to the ocean floor, the tsunameter monitors temperature andpressure. These data are passed to a separate surface buoy by means of acoustic pulses. The buoycommunicates with the tsunami warning centers using a commercial Iridium satellite link. First-generation DART systems featured an automatic detection and reporting algorithm triggered by athreshold wave-height value. Today's design permits two-way communications, enabling datatransmission on demand, independent of the automatic triggering. This ensures the measurement andreporting of tsunamis with amplitudes below predetermined threshold limits. When a seismic eventoccurs, the tsunami warning centers use predictive software to model tsunami magnitude andseverity, but until empirical data, such as wave height from DART buoys, become available, thecenters can only forecast the likelihood of a tsunami. DART system information is used to confirm and refine tsunami characteristics. With these data, more accurate reporting is possible, improvingwatches, warnings or evacuation bulletins. (Adapted from NOAA, http://nctr.pmel.noaa.gov/Dart/.)

Iridium andGPS antennas

Electronic systemsand batteries

Acoustic transducers(2 each)

Surface buoy,2.5-m diameter,

4,000-kg displacement

Bidirectionalcommunication

and control

Iridium satellite

Tsunamiwarningcenter

Tsunameter

Signalflag

Glass ballflotation

Acoustictransducer

Anchor, 325 kgAnchors, 3,100 kg

1,000 to 6,000 m

~ 75 m Bidrec

tiona

l aco

ustic

telem

etry

31. Titov VV and Synolakis CE: “Numerical Modeling of TidalWave Run-Up,” Journal of Waterway, Port, Coastal and Ocean Engineering 124, no. 4 (July/August 1998):157–171. For more on tsunami modeling: http://nctr.pmel.noaa.gov/model.html (accessed August 10, 2007).


As additional information, such as DART andtidal-gauge data, becomes available, the model isadjusted (above).

Another tool used in analyzing the 2004tsunami, the Jason-1 earth-imaging satellite,provided modelers with accurate wave-heightdata for the duration of the tsunami. Measuredfrom space, the resolution was in the centimeterrange. Rather than having only point-to-pointmeasurements, such as from tidal gauges orDARTs, the waves could be measured continu -ously with satellite data. Unfortunately, the lagtime is much too great and the coverage area toosparse to use satellite data in real time. However,satellite information can provide validation andimprovement for the current modeling programs.

Even with all the data at their disposal,experts were challenged to explain how theSumatra-Andaman earthquake produced atsunami whose magnitude exceeded initial wave-height predictions. NOAA’s tsunami forecast,running the model with seismic data alone,originally underestimated tsunami heights in theopen ocean by a factor of 10. Integration oftsunami amplitudes from tidal gauges improved

the results iteratively, but the results were notconsidered satisfactory. Analyses of the shock’sstrong seismic waves indicated that the initialfault break traveled northward from Sumatra at2.5 km/s [1.6 mi/s]. The analysis also pinpointedthe areas of greatest slip—and thus of thegreatest wave generation. The problem fortsunami modelers was that none of these seismicsolutions included enough overall fault motion toreproduce either the satellite observations ofwave heights in the open ocean or the severeflooding in Banda Aceh.

The critical piece of the puzzle came fromelevation and displacement data provided byland-based global positioning system (GPS)monitors, used to track ground movements. TheGPS sensors, recording at a much slower ratethan seismic monitors, revealed that the faultcontinued to move long after it stoppedemanating seismic energy. Although there is alimit to how slowly a fault can slip and stillgenerate a tsunami, this often overlookedphenomenon, called after-slip, accounted for theobserved tsunami wave heights. IncorporatingGPS readings into modeling programs will be an

important component in improving the accuracyof tsunami warning systems in the future.32

Another challenge is integrating the data in atimely manner.

A major drawback in developing and usingmodeling software is that there is so littleempirical data to compare with the model. OnSeptember 12, 2007, an 8.4 MW earthquakeoccurred in the vicinity of the December 2004earthquake. This was the first major event sincethe deployment of a DART buoy in the IndianOcean. The MOST program predicted a 2-cm[0.75-in.] rise in wave height at the location ofthe buoy with an arrival time of approximately 2 hours and 50 minutes. The observed wave heightsand arrival times matched MOST predictions(next page).33

Inundation models, estimates of how farinland a tsunami will travel, are another criticalcomponent. Scientists use measurementsrecorded near the coast from tidal gauges or post-event estimates from water damage to determinerun-up. Early programs calculated wave heightsat the shore’s edge but had difficulty projectingthe effects onto the shore. A 1992 Nicaraguantsunami gave scientists an opportunity to makecomprehensive measurements and compare themwith model predictions.34

Using large-scale laboratory experiments andfield measurements, investigators refined theirmodels until they could match the empiricaltsunami inundation measurements. Using high-resolution land imagery, accurate bathymetrydata, coastal and offshore topographical data,historical information from previous tsunamisand software to make rapid calculations, theydemonstrated that an early-warning systemcould provide reliable estimations.

Sounding the AlarmThree months prior to the December 2004tsunami, a working group for the SouthwestPacific and Indian Ocean Tsunami WarningSystem was established. Under the auspices ofthe International Tsunami Information Center(ITSU), a UNESCO organization, this group’scharter was to expand the Pacific warningsystem to include other regions with thepotential for tsunamis, including the IndianOcean. When the earthquake occurred, thePacific Tsunami Warning Center (PTWC)attempted to contact affected countries acrossthe Indian Ocean; unfortunately, it was Sundayas well as a holiday for many. Most offices wereclosed, and the warnings did not reach theinhabitants of the affected coastlines. One result

18 Oilfield Review

32. Geist EL, Titov VV and Synolakis CE: “Tsunami: WAVE of CHANGE,” Scientific American 294, no. 1 (January 2006): 56–63.

33. http://nctr.pmel.noaa.gov/sumatra20070912.html(accessed September 21, 2007).

34. Imamura F, Shuto N, Ide S, Yoshida Y and Abe K:“Estimate of the Tsunami Source of the 1992 Nicaraguan

Earthquake from Tsunami Data,” Geophysical ResearchLetters 20, no. 14 (1993): 1515-1518.

35. “Tsunami 2004: Waves of Death,” The History ChannelWeb site, http://www.history.com/shows.do?action=detail&episodeId=173117 (accessed September 27, 2007).

36. http://www.sciencedaily.com/releases/2006/07/060710085816.htm (accessed October 1, 2007).

> Model of the Sumatra-Andaman earthquake tsunami. Using NOAA’sMethod of Splitting Tsunami (MOST) program, the tsunami (arrow) wasmodeled as it traveled across the Indian Ocean. Shown here at approximately1 hour after initiation, the wave will take three more hours to reach theAfrican coastline. (Adapted from NOAA/PMEL/Center for TsunamiResearch, http://nctr.pmel.noaa.gov/model.html.)


Autumn 2007 19

of the Sumatra-Andaman earthquake was toaccelerate the pace of developing global early-warning networks.

Post-tsunami analysis confirmed thatcommunication within the region and links toother monitoring sites were lacking. A strikingexample of the importance of having anemergency management system in place isevidenced by comparing the tsunami mortalityrate in Kenya and Somalia. Kenya did not have atsunami warning system, but it did have achemical and oil spill-alerting system. Whenword of the approaching tsunami reachedKenyan officials (tsunami travel from Sumatra toKenya took four hours), they activated the spill-alerting system. Approximately 800,000 peoplewere warned to move inland or seek higherground. Four hours after the earthquake, thetsunami reached the shores of Kenya andSomalia. The death toll for Kenya was one. Inneighboring Somalia, where there was nowarning system, the death toll was 150.35

With modern Internet and satellite connec -tivity, communication over a wide area is almostinstantaneous, but communications can bechallenging in developing countries. Problemsalso arise when the alert must be communicatedto the general population. Planning for eventslike this must assume that infrastructures arelikely to be severely damaged. Satellite linksmake it possible to communicate in the absenceof land lines, but contingencies must also be inplace to alert the general population if localsystems are destroyed.

Effective warning systems for natural hazardsrequire public information and preparednesscomponents. Early warning is largely a socialissue, and technology alone will not solve theproblem. Early-warning systems may fail at timesof crisis if warnings are not received by thepeople at risk, or are not understood, or are notacted upon. An effective early-warning systemneeds to be people-centered in addition tohaving sound technical methods of communi -cation. Trained and experienced emergency

management personnel are critical to ensurethat warnings are clearly communicated, wellunderstood and rapidly implemented. Inaddition, regional coordination is important, asearthquakes and tsunamis do not restrictthemselves to territorial borders.

Even with the best data, the accuracy of themodels used to predict tsunamis is limited byerrors in bathymetry and uncertainties in thetriggering mechanism. Each earthquake isunique, and every tsunami has a uniquecombination of wavelengths, wave heights anddirectionality. From a warning perspective, thismakes the problem of forecasting tsunamis inreal time difficult. In the case of the December2004 tsunami, the areas north and south of theearthquake epicenter had little damagecompared with areas to the east and west. CocosIsland is 1,500 km to the south, and Sri Lanka is1,500 km to the west of the epicenter. Themaximum wave height on Cocos was 42 cm[16 in.], while sections along the Sri Lankancoast experienced run-up in excess of 8 meters[26 ft]. Warning center personnel understand theneed for a delicate balance between creatingundue panic and underestimating the severity,potentially causing an even greater tragedy.

The Way ForwardIn 2006, UNESCO Director-General KoïchiroMatsuura announced that, after much coopera -tive effort, 26 national tsunami informationcenters had been established around the IndianOcean.36 As part of the Indian Ocean TsunamiWarning System (IOTWS), this is the first stage in the development of an integrated organi -zation modeled after the Pacific TsunamiWarning System.

As of October 2007, seismographic reportingstations have been upgraded and two DART buoyshave been deployed in the Indian Ocean. Twenty-five additional monitoring stations will be addedand linked in real time to analysis centers.Information bulletins are being issued fromJapan and Hawaii, pending a decision on the finallocations of Indian Ocean regional centers. In thefuture, additional DART buoys and satellite linkswill be deployed. The work certainly is not over; ithas taken 40 years to develop the Pacific TsunamiWarning System, and developing a comparablesystem for the Indian Ocean will also take sometime. However, a basic system is now in place forwhen—not if—the next great Indian Oceantsunami occurs. —TS

> MOST predictions compared with tsunami data. Shown on the map of theIndian Ocean (top), the Thailand DART buoy (yellow circle) was installed inAugust 2007. On September 12, 2007, an 8.4 MW earthquake (red star)occurred with an epicenter just south of the Sumatra-Andaman earthquakeof 2004. A minimal tsunami was generated by the event. In a comparison ofwave heights (bottom), the MOST wave-height simulation (red curve) aftereight hours compares favorably with the data recorded by the ThailandDART buoy (blue curve) both in wave amplitude and arrival time. (Adaptedfrom data courtesy of NOAA/PMEL/Center for Tsunami Research.)

0 1 2 3 4 5 6 7 8 9 10 11 12

MOST modelDART data

Time after earthquake, h

Ampl

itude

, cm

0

2

4

–2

–4

Thailand DART buoy Epicenter of 2007 earthquake


20 Oilfield Review

Borehole Seismic Surveys: Beyond theVertical Profile

John BlackburnConocoPhillips U.K., Ltd.Aberdeen, Scotland

John DanielsOklahoma City, Oklahoma, USA

Scott DingwallAberdeen, Scotland

Geoffrey Hampden-SmithShell Exploration & ProductionAberdeen, Scotland

Scott LeaneyJoël Le CalvezLes NuttHouston, Texas, USA

Henry MenkitiLondon, England

Adrian SanchezVillahermosa, Tabasco, Mexico

Marco SchinelliPetrobrasRio de Janeiro, Brazil

For help in preparation of this article, thanks to ReginaldBurl, Ed Ferguson and William Phebus, Belle Chasse,Louisiana, USA; Allan Campbell, Mike Craven, Rogelio Rufinoand Bill Underhill, Houston; John Edwards, Muscat, Oman;Alan Fournier, St John’s, Newfoundland, Canada; KevinGalliano, Larose, Louisiana; John Graves, Hess Corporation,Houston; Caroline Kinghorn, Dave Milne, Gary Rogers andThilo Scharf, Aberdeen; and Colin Wilson, Fuchinobe, Japan.seismicVISION, SlimXtreme, StimMAP, VSI (Versatile SeismicImager) and Xtreme are marks of Schlumberger.

Today’s borehole seismic methods create new opportunities for investigating

formations penetrated by a borehole. From well construction and 3D subsalt imaging

to stimulation monitoring and high-pressure, high-temperature acquisition, borehole

seismic surveys reduce operator risk and help improve recovery.

3. Arroyo JL, Breton P, Dijkerman H, Dingwall S, Guerra R,Hope R, Hornby B, Williams M, Jimenez RR, Lastennet T,Tulett J, Leaney S, Lim T, Menkiti H, Puech J-C,Tcherkashnev S, Burg TT and Verliac M: “Superior Seismic Data from the Borehole,” Oilfield Review 15,no. 1 (Spring 2003): 2–23.

1. Point sources are implosive or explosive sources, suchas dynamite or airguns. Sweep sources are vibroseistrucks or other vibrating sources.

2. Marine vibrating sources have been attempted: Fischer PA: “Seismic Source Offerings Provide Optionsfor Operators,” World Oil 227, no. 6 (June 2006),http://www.worldoil.com/magazine/MAGAZINE_DETAIL.asp?ART_ID=2913&MONTH_YEAR=Jun-2006 (accessed October 8, 2007).


Autumn 2007 21

Borehole seismic surveys are now among the mostversatile of all downhole measurement techniquesused in the oil field. Historically, the main benefitderived from these surveys, also known as verticalseismic profiles (VSPs), has been to link time-based surface seismic images with depth-basedwell logs. However, today’s borehole seismicsurveys have expanded beyond a simple time-depth correlation. The wide spectrum of seismicenergy that is now recorded and the variousgeometries currently possible with boreholeseismic surveys combine to deliver results notpreviously available. From these data, E&Pcompanies derive important information aboutreservoir depth, extent and heterogeneity, as well

as fluid content, rock-mechanical properties, porepressure, enhanced oil-recovery progress, elasticanisotropy, induced-fracture geometry andnatural-fracture orientation and density.

Originally, VSPs consisted of receiversdeployed in a vertical borehole to record the mostbasic signals from a seismic source at the surface.The innovations delivered by modern VSPs havecome about by recording more information andexpanding survey geometries with improvedacquisition tools. This article describes the typesof waves that can be recorded in the borehole, andthe tools that record them. We then briefly catalogthe many types of surveys that can be acquired,along with the information they can provide. Wecontinue with case studies demonstratingadvances in borehole seismic surveys, including3D VSPs and VSPs acquired while drilling,optimizing hydraulic fractures, monitoring perfo ra tion operations, and VSP acquisition inhigh-pressure, high-temperature conditions.

Types of WavesThe main types of waves generated and recordedin borehole seismic surveys are body wavesemitted by point sources or frequency-sweepsources, and consist of compressional, or primary,

P-waves and shear, or secondary, S-waves.1 Thesewaves propagate from man-made sources nearthe surface to borehole receivers at depth. In thecase of marine VSPs, and where land VSPs deployairguns in a mud pit, typically only P-waves aregenerated. However, depending on the receivergeometry and formation properties, both P-wavesand S-waves may be recorded if S-waves havebeen generated by conversion from a reflecting P-wave (below left). For land VSPs with sourcescoupled directly to the earth, both P- and S-wavesare generated and may be recorded.2

The signals recorded by borehole receiversdepend on the incoming wave type, the surveygeometry and the type of receiver. Most moderndownhole hardware for recording VSPs consistsof clamped, calibrated three-component (3C)geophones, which are able to record all compo -nents of P- and S-wave motion, including SV- andSH-waves.

The Schlumberger borehole seismic tool, theVSI Versatile Seismic Imager, offers up to 40 three-component receivers, called shuttles, that can bespaced up to 150 ft [46 m] apart to form an array6,000 ft [1,830 m] long (below).3 The 40-shuttletool has been deployed several times for VSPacquisition in the Gulf of Mexico. The VSI tool

> Propagation and reflection of compressionaland shear waves. At normal incidence,compressional P-waves reflect and transmit onlyas P-waves. However, at incidence other thannormal, such as when the source is placed somedistance from the rig, an incident P-wave canreflect and transmit P-waves and shear S-waves(top). P-waves have particle motion along thedirection of propagation, and S-waves haveparticle motion orthogonal to the direction ofpropagation (bottom). SV-waves are polarized inthe vertical plane and SH-waves are polarized inthe horizontal plane. Incident SV- and SH-wavesare generated by shear-wave sources.

SP

PS

PIncident P-wave

Receiver

SV

SH

P

> VSI Versatile Seismic Imager. Each of the 40 VSI shuttles contains threeorthogonally oriented geophone accelerometers in an acoustically isolatedsensor package that can be clamped to the borehole wall.


can be run in open hole, cased hole or drillpipe,and is clamped into position to provide optimumcoupling. Options for conveyance include wire -line, downhole tractor or drillpipe.

An advantage that borehole seismic surveyshave over their surface seismic counterparts istheir capability to record direct signals in a low-noise environment. The direct signal travelsdownward to the receivers, and so is referred to asa downgoing signal. Waves that reflect at deeperinterfaces and then travel up to a boreholereceiver are recorded as upgoing signals (above).Upgoing signals contain reflection information,and are used to create seismic images of sub -surface reflectors. Both upgoing and downgoingsignals can contain multiples, or energy that hasreflected multiple times, which can interfere withthe desired signal. Signals without multiples arecalled primaries. Downgoing signals can be used todistinguish multiples from primary arrivals, and toenable more reliable processing of the surfaceseismic upgoing wavefield.

In conjunction with P- and S-waves, whichpropagate from a near-surface source to thereceiver, different types of source-generatednoise arise. Tube waves are formed when source-generated surface waves transfer energy to theborehole fluid. The resulting fluid-guided wavetravels down and up the borehole, forcing theborehole wall to flex radially. Receivers clampedto the borehole wall record tube-wave energy onhorizontal geophone components. Tube waves

are sensitive to changes in borehole dimension,which can cause them to reflect. Another form ofnoise that sometimes contaminates recordings iscasing ringing.

The majority of VSPs use compressional andshear waves from airguns, vibrating trucks ordynamite sources for imaging reflectors, butenergy from other sources can be recorded andprocessed to yield information about the sub -surface. For example, the drill bit can act as adownhole source, generating vibrations that aredetected by sensors deployed at surface or onmarine cables.4 These recordings require special -ized processing, but can provide critical answersin time for decisions to be made while drilling,such as changing mud weight or setting casing.

Hydraulically induced fractures emit energyin much the same way as natural earthquakes,and these microseisms can be recorded bysensors in neighboring boreholes. Similarly,production of fluids or injection of fluids forenhanced recovery or waste disposal all inducestress redistribution that in turn can causedetectable microseismicity. And finally, boreholesensors can be used to record natural seismicity.5

Types of SurveysBorehole seismic surveys are usually categorizedby survey geometry, which is determined bysource offset, borehole trajectory and receiver-array depth. The survey geometry determines thedip range of interfaces and the subsurfacevolume that can be imaged.

The simplest type of borehole seismic surveyis the zero-offset VSP. The basic zero-offset VSPfeatures a borehole seismic receiver array and anear-borehole seismic source (next page, top). Inmost cases (unless formation dips are very high),this survey acquires reflections from a narrowwindow around the borehole. The standardoutput from a zero-offset VSP is a corridor stack,created by summing the VSP signals thatimmediately follow the first arrivals into a singleseismic trace. That trace is duplicated severaltimes for clarity and comparison with surfaceseismic images. Processing yields velocities offormations at different depths, which can be tiedto well log properties and interpreted fordetection and prediction of overpressured zones.The velocity model can also be used to generatesynthetics to identify multiples in surfaceseismic processing.

Another type of zero-offset VSP is known as adeviated-well, walkabove, or vertical-incidenceVSP. It is designed to ensure that the source isalways directly above receivers deployed in adeviated or horizontal wellbore. This surveyacquires a 2D image of the region below theborehole. In addition to formation velocities andan image for correlation with surface seismicdata, benefits of a walkabove VSP are goodlateral coverage and fault and dip identificationbeneath the well.

Offset VSPs are acquired using a sourceplaced at a horizontal distance, or offset, fromthe wellbore, again producing a 2D image. Thereceiver arrays are deployed at a wide range ofdepths in the borehole. The offset increases thevolume of subsurface imaged and maps reflectorsat a distance from the borehole that is related tothe offset and subsurface velocities. The addedvolume of illumination enhances the usefulnessof the image for correlation with surface seismicimages, and for identification of faulting and diplaterally away from the borehole. In addition,because the conversion of P-waves to S-wavesincreases with offset, an offset VSP allows shear-wave, amplitude variation with offset (AVO) andanisotropy analysis. The degree to which P-wavesconvert to S-waves depends on offset and oninterface rock properties.

Walkaway VSPs are similar to offset VSPs inthat the source is offset from vertical incidence,but the acquisition geometry is somewhatreversed. The borehole receiver array remainsstationary while the source moves away from it,or “walks away” at a range of offsets. The range ofoffsets acquired in a walkaway VSP is particularlyuseful for studying shear-wave, AVO and

22 Oilfield Review

> Upgoing, downgoing, primary and multiple arrivals. Upgoing wavesreflect at interfaces below the receiver and then travel upward to berecorded (blue and green). Downgoing waves arrive at the receivers fromabove (red and orange). A wave that arrives at the receiver withoutreflecting is called the direct arrival (red). Waves that reflect only once arecalled primaries. The reflected upgoing primary (blue) is the arrival that isdesired for imaging reflections.

Time

Dept

h

1

2

3

4

5

• Downgoing direct arrival• Reflected upgoing primary• Downgoing multiple• Reflected upgoing multiple


Autumn 2007 23

anisotropy effects. And, because they canilluminate a large volume of subsurface, offsetand walkaway VSPs are useful elements in thedesign of surface seismic surveys.

The surveys described so far are all designedto provide information and images in one or twodimensions. To adequately illuminate 3D struc -tures requires 3D acquisition and processing. Inthe same way that surface seismic surveys haveprogressed from 1D and 2D to 3D, so have VSPs.

Three-dimensional VSPs can be acquired onland or offshore. Acquisition of 3D marine VSPsis similar to that of 3D marine surface seismicsurveys and can follow parallel lines orconcentric circles around a borehole (right). Onland, source positions typically are laid out in agrid. Three-dimensional VSPs deliver high-resolution subsurface imaging for explorationand development applications, and requiredetailed prejob modeling and planning. Inaddition to producing images at higherresolution than surface seismic methods, 3DVSPs can fill in areas that cannot be imaged bysurface seismic surveys because of interferingsurface infrastructure or difficult subsurfaceconditions, such as shallow gas, which disruptspropagation of P-waves.

4. Breton P, Crepin S, Perrin J-C, Esmersoy C, Hawthorn A,Meehan R, Underhill W, Frignet B, Haldorsen J, Harrold Tand Raikes S: “Well-Positioned Seismic Measurements,”Oilfield Review 14, no. 1 (Spring 2002): 32–45.

5. Coates R, Haldorsen JBU, Miller D, Malin P, Shalev E,Taylor ST, Stolte C and Verliac M: “Oilfield Technologiesfor Earthquake Science,” Oilfield Review 18, no. 2 (Summer 2006): 24–33.

> Variations on a theme of VSPs (left to right). The original acquisition geometry, with no offset between source and wellbore, creates a zero-offset VSP.Seismic waves travel essentially vertically down to a reflector and up to the receiver array. Another normal-incidence, or vertical-incidence, VSP isacquired in deviated wells with the source always vertically above each receiver shuttle. This is known as a deviated-well, or walkabove VSP. In an offsetVSP, an array of seismic receivers is clamped in the borehole and a seismic source is placed some distance away. The nonvertical incidence can give riseto P- to S-wave conversion. In walkaway VSPs, a seismic source is activated at numerous positions in a line on the surface. All these survey types may beacquired onshore or offshore.

Offset VSP

Receivers

Source

Deviated-Well VSP

Receivers

Sources

Walkaway VSP

Sources Receiver

Zero-Offset VSP

Source

Receivers

> Three-dimensional VSPs. Onshore and offshore, 3D VSPs tend to borrow surface seismicacquisition geometries. On land, source positions usually follow lines in a grid. Offshore, sourcepositions can be laid out in lines or in a spiral centered near the well (left). Ray-trace modeling prior toacquisition ensures proper coverage and illumination of the target. In this offshore example (right),source lines at the surface are shown in red. Green lines are rays traced from source to receiver.Wells are positioned at the light blue triangles at the surface. Blue surfaces are the top and bottom ofa salt body. The target horizon is the red surface at the bottom.

Receiver

3D VSP


VSPs have long been used to tie time-basedsurface seismic images to depth-based well logs.In many exploration areas, the nearest wells maybe quite distant, so VSPs are not available forcalibration before drilling begins on a new well.Without accurate time-depth correlation, depthestimates derived from surface seismic imagesmay contain large uncertainties, adding risk andthe cost of contingency planning to drillingprograms. One way to develop a time-depthcorrelation is to perform an intermediate VSP: torun a wireline VSP before reaching total depth(TD). These surveys provide reliable time-depthconversions, but add cost and inefficiency to thedrilling operation, and may come too late toforecast drilling trouble.

To help reduce uncertainty in time-depthcorrelation without having to stop the drillingprocess, geophysicists devised a seismic-while-drilling process (above left). This technology usesa conventional seismic source at the surface, anLWD tool containing seismic sensors in thedrillstring, and a high-speed mud-pulse telemetrysystem to transmit information to the surface.6

Availability of real-time seismic waveforms allowsoperators to look thousands of feet ahead of thebit to safely guide the well to TD. Because drillinggenerates noise that could jeopardize seismic

data quality, source activation and signalmeasurement must take place during quietperiods, when drilling has paused for otherreasons, such as making drillpipe connections. Alimitation of this method is that the seismic LWDreceivers, being part of the drillstring, are notclamped to the borehole wall, althoughformation-receiver coupling generally improveswith well deviation.

Several borehole seismic technologies areavailable for understanding fractures andfracture systems, both natural and hydraulicallyinduced. The walkaround VSP is designed to

characterize the direction and magnitude ofanisotropy that arises from aligned naturalfractures. In this survey, offset source locationsspan a large circular arc to probe the formationfrom a wide range of azimuths (above).7

Hydraulically induced fractures can also bemonitored using borehole seismic methods. Whilethe fracture is being created in the treatmentwell, a multicomponent receiver array in amonitor well records the microseismic activitygenerated by the fracturing process (below).Locating hydraulically induced microseismicevents requires an accurate velocity model.

24 Oilfield Review

6. Breton et al, reference 4.7. Horne S, Thompson C, Moran R, Walsh J, Hyde J and

Liu E: “Planning, Acquiring and Processing a WalkaroundVSP for Fracture Induced Anisotropy,” presented at the64th EAGE Conference and Exhibition, Florence, Italy, May 27–30, 2002.

8. The fluid injection under discussion here is for pressuresupport, not for hydraulic fracturing.

9. Hornby BE, Yu J, Sharp JA, Ray A, Quist Y and Regone C:“VSP: Beyond Time-to-Depth,” The Leading Edge 25,no. 4 (April 2006): 446–448, 450–452.

10. Leaney WS and Hornby BE: “Subsalt Elastic Velocity Prediction with a Look-Ahead AVA Walkaway,” paperOTC 17857, presented at the Offshore Technology Conference, Houston, May 1–4, 2006.

> A VSP while drilling. The seismicVISIONseismic-while-drilling tool positioned near thedrill bit receives signals generated by a seismicsource at the surface. Signals are transmitted tothe surface for real-time, time-depth information.

VSP While Drilling

Receiver > A walkaround VSP. With the offset source at several azimuths, this surveycan detect anisotropy caused by aligned natural fractures.

> Microseismic method of hydraulic fracture monitoring. Sensitivemulticomponent sensors in a monitoring borehole record microseismicevents, or acoustic emissions, caused by hydraulic fracturing. Dataprocessing determines event location, and visualization allows engineers to monitor the progress of stimulation operations.

Reservoir

Microseism

Hydraulic fracture

Treatment well Monitoring well


Autumn 2007 25

Mapping the extent of the fracture with time helpsmonitor the progress of stimulation treatmentsand allows comparison between actual andplanned fractures. Real-time information aboutfracture extent and orientation promises to helpstimulation engineers optimize treatments byallowing them to modify pumping rates andvolumes when observed fractures differ from plan.A drawback of the method is that nearly allapplications have required deploying the receiverarray in a monitoring well because it is believedthat the treatment well is too noisy. The cost ofdrilling a monitoring well could be saved if thetechnology could be applied in treatment wells.

Another borehole seismic technology, calledpassive seismic monitoring, characterizes frac -tures by recording microseismic signals generatedwhen fluid is produced from or injected into anaturally fractured reservoir. When fluid injectionand production modify the stress state enough tocause seismic events, the resulting acousticemissions can be recorded in nearby monitoringwells by arrays of multicomponent boreholereceivers.8 The technique is similar to monitoringhydraulic fractures, but the events are smaller inmagnitude. The microseismic events can beplotted in space and time to identify the fracturesthat are responding to the change in stress state.Because the timing of microseismic events cannotbe predicted, acquisition systems for passive

seismic monitoring must be different fromstandard VSP acquisition systems. Recordingsystems need to be active for long periods of time,waiting to be triggered by acoustic emissions. Insome cases, receiver arrays are installedpermanently to record for extended periods.

Propagating seismic signals between wellscreates yet another type of borehole seismicprofile, known as a crosswell seismic survey(left). In these surveys, downhole seismicsources, such as downhole vibrators, aredeployed at selected depths in one borehole,shooting to a receiver array in another borehole.Because the direction from source to receiver issubparallel to layer boundaries, most raypathspropagate without reflecting. Recorded data areprocessed to extract information about thevelocities in the interwell region. Since crosswelldata do not contain much information aboutreflectors, layer boundaries in the initial velocitymodel used to process the crosswell datatypically come from sonic logs or standard VSPs.A limitation of the crosswell method is the maxi -mum allowable distance between boreholes—afew thousand feet is typical—which varies withrock type, attenuation, and source strength andfrequency content.

Many of the borehole seismic surveysmentioned above can be acquired at differentstages in the life of a reservoir. Offset VSPs,walkaways, 3D VSPs and crosswell surveys canalso be acquired in time-lapse fashion, beforeand after production. Time-lapse surveys canreveal changes in the position of fluid contacts,changes in fluid content, and other variations,such as pore pressure, stress and temperature.As with time-lapse surface seismic surveys, caremust be taken to repeat acquisition conditionsand processing as closely as possible so thatdifferences between baseline and monitoringsurveys may be interpreted as changes inreservoir properties.

The VSP method has evolved from its humblebeginnings as a time-depth tie for surface seismicdata to encompass a wide range of solutions toexploration and production problems.9 Theremainder of this article is devoted to casestudies that highlight the versatility of today’sborehole seismic surveys, starting with VSPsacquired while drilling.

Reducing Uncertainty in Well ConstructionBorehole seismic surveys are best known fortheir ability to tie time-based seismic sections todepth-based information such as well logs anddrilling depths. These correlations are possiblebecause the depth of each borehole seismic

sensor is known, and the time it takes for aseismic wave to arrive at the sensor is known.However, these correlations contain uncertain -ties when the well has yet to attain the depthsthat need to be correlated. In such situations, itis necessary to look ahead of the well’s TD andpredict formation properties ahead of the bit.

Two types of borehole seismic surveys—seismic-while-drilling imaging and intermediateVSPs—can provide look-ahead information. Inan example of the first, Devon Energy obtained aVSP image, in addition to time-depth and velocityinformation, while drilling a directional well inthe Gulf of Mexico. Waveforms acquired duringdrillpipe connection and transmitted to surfaceduring drilling operations were processed at aSchlumberger processing center and reported toDevon engineers at the rig site and in remoteoffices. An initial seismicVISION seismic-while-drilling image acquired 1,000 ft [305 m] abovethe target indicated that the well would notreach the target as planned (below). Devon teammembers in Houston decided to sidetrack thewell and used additional seismicVISION data toguide the well to the intended TD.

Intermediate VSPs also deliver informationbeyond TD. BP ran such a “look-ahead” walkawayVSP in a deepwater well in the Gulf of Mexico.10

> Crosswell seismic surveys, with sources inone borehole and receivers in another. Becauseraypaths are at large angles to any formationinterfaces, little energy is reflected; most energyrecorded by the receivers comes from directarrivals. These data reveal information aboutformation velocities in the interwell volume. Therepeatable survey geometry makes crosswellseismic surveys useful for time-lapse monitoringof steam injection, for example.

Sources

Receivers

Crosswell VSP

> Imaging while drilling. Two seismic imagesacquired while drilling (red and blue) aresuperimposed on preexisting surface seismicdata (black and white). The first seismic image(left of vertical black line), acquired in the originalwell (green) indicated to Devon interpreters thatthe well would not reach the target as planned.The well was sidetracked (yellow), and anotherseismic image acquired while drilling (right ofvertical black line) indicated that the well wouldreach the target.

Image acquiredfrom original well

Image acquiredfrom sidetrack

Original well

Sidetrack


The well was to penetrate a salt structure to tapsubsalt sediments. Drilling deepwater wellsthrough salt is expensive and risky. The saltobscures seismic signals from underlyingformations, making it difficult to image themproperly, and also forms such a strong seal thatpore pressure below salt can be abnormally high.

Estimates of pore pressure can be made fromthe ratio of seismic velocities derived fromprocessing surface seismic data, but these

velocities often have large uncertainties.11

Borehole seismic surveys can help reduce the riskof drilling into subsalt sediments by obtainingmore accurate seismic velocity ratios before thewellbore exits the salt.

In the BP survey, a 12-level borehole seismictool acquired walkaway data while clamped in thesalt near the base salt interface (above). In thiswalkaway configuration, 800 surface shots werefired in a line extending approximately 25,000 ft[7,600 m] on both sides of the well. Compressionalwaves generated by the source reflect back asboth P-waves, called P-p arrivals, and as S-waves,called P-s arrivals. With the tool clamped as closeas possible to the base of salt, the seismic energyreflecting at varying angles near the base of thesalt can be analyzed for amplitude variation withangle (AVA) of incidence. Analysis of AVA—analogous to well-known amplitude variation withoffset (AVO)—reveals elastic properties of thematerials at the reflecting interface.12

In this case, geophysicists expected tomeasure P- and S-wave velocities of the subsaltlayers, along with quantified uncertainties, to beused in estimates of pore pressure and safe mudweight.13 If results of the survey were to be usefulfor salt-exit drilling, time from last shot to mud-weight prediction had to be short, within two days.

Amplitude variation with angle depends on thedensity and compressional and shear velocities ofthe material on either side of the reflecting

interface. Measured AVA properties for P-p and P-s arrivals were compared with modeled values,and the inversion process iteratively modified themodel to achieve a best fit with the data (belowleft). Inverting for subsalt compressional andshear velocities is possible because the densityand velocities within the salt are known with ahigh degree of certainty. Noise in the data makesit difficult to invert for subsalt density, so anexpected value is assumed.

The inversion predicted the ratio of P- and S-wave velocities with lower uncertainties thanpredrill estimates. A dipole sonic log recordedbelow and through the salt provided a post-drillmeasure of subsalt velocities, which were withinthe uncertainties predicted by the look-aheadwalkaway VSP (below).

Double-Well 3D VSPIn the Riacho de Barra field, a mature asset inthe Recôncavo basin of northeast Brazil,Petrobras sought to reduce risks in an infill-drilling campaign. Conventional 3D surfaceseismic data over the field had not satisfactorilyresolved structural and stratigraphic traps: ahigh-velocity conglomerate formation in theoverburden attenuated seismic signals andreduced bandwidth, deteriorating resolution andmaking it difficult for interpreters to definereservoir boundaries (next page, top).14

To improve the seismic image, geophysicistsexamined the feasibility of conducting a 3D VSPin existing wells. The main goal of the survey wasto resolve erosional truncations of the upperreservoir and delineate a deeper target that had

26 Oilfield Review

> Acquisition of an amplitude variation with angle (AVA) walkaway VSP atthe salt base. Processing assumes that the raypaths through the salt areequivalent for the direct ray and the ray that reflects off the base of the salt.

Salt

> Comparison of AVA data and modeled results.P-p (red) and P-s (green) reflected amplitudescan be corrected with a 6° shift in the angle ofthe interface, corresponding to the dip of the saltbase (blue for corrected P-p, black for correctedP-s.) The best-fit model curves are shown inpurple for P-p and orange for P-s. (Modified fromLeaney and Hornby, reference 10.)

Refle

ctio

n co

effic

ient

–0.25

0

0.25

0.50

–0.50

–75 –50 –25 0 25 50 75Incidence angle, °

> Comparing predictions of compressional (Vp)and shear (Vs) velocities and uncertainty rangeswith measured values. The look-ahead walkawayVSP prediction of Vp and its uncertainty range(green) span the values later obtained by loggingin the same well (black). Similarly, the predictedVs and its uncertainty range (blue cloud)accurately estimated the subsequently loggedshear velocities (red curve). Also shown is thepredicted Vp /Vs ratio (red cloud) and the ratio oflogging results (blue curve). (Modified fromLeaney and Hornby, reference 10.)

Monte Carlopredictions

Vs Vp Vp/Vs

Dept

h

Velocity Vp/Vs ratio


Autumn 2007 27

been poorly defined by surface seismic imaging.An initial velocity model was constructed fromthe 3D surface seismic data and calibrated by logdata from more than 30 wells in the area. Ray-tracing through the model helped select thesurvey design that would maximize coverage atthe targeted interfaces.

The 3D VSP design comprised 2,700 shotpoints over a 13-km2 [5-mi2] area, to be recordedfrom two neighboring wells simultaneously(right). To optimize acquisition logistics, aPetrobras seismic crew performed essentialsurvey operations, such as location of shot points

11. Bryant I, Malinverno A, Prange M, Gonfalini M, Moffat J,Swager D, Theys P and Verga F: “Understanding Uncertainty,” Oilfield Review 14, no. 3 (Autumn 2002): 2–15.

12. Leaney WS, Hornby BE, Campbell A, Viceer S, Albertin Mand Malinverno A: “Sub-Salt Velocity Prediction with aLook-Ahead AVO Walkaway VSP,” Expanded Abstracts,74th SEG Annual International Meeting and Exposition,Denver (October 10–15, 2004): 2369–2372.Chiburis E, Franck C, Leaney S, McHugo S andSkidmore C: “Hydrocarbon Detection with AVO,” Oilfield Review 5, no. 1 (January 1993): 42–50.

13. Dutta NC, Borland WH, Leaney WS, Meehan R andNutt WL: “Pore Pressure Ahead of the Bit: An IntegratedApproach,” in Huffman A and Bowers G (eds): PressureRegimes in Sedimentary Basins and Their Prediction,AAPG Memoir 76. Tulsa: AAPG (2001): 165–169.

14. Sanchez A and Schinelli M: “Successful 3D-VSP on LandUsing Two Wells Simultaneously,” Expanded Abstracts,77th SEG Annual International Meeting and Exposition,San Antonio, Texas (September 23–28, 2007): 3074–3078.

Tucano basin

Riacho deBarra field

Recôncavo basin

km 250

miles 250

S O U T H A M E R I C A

B r a z i l

> Double-well 3D VSP acquisition design. More than 2,700 shot points were planned on lines over a13-km2 area. The area covered joins two circles centered on two wells (right). Shot locations arecolor-coded from low elevation (blue) to high elevation (red). A velocity model from existing 3Dsurface seismic data (left) was useful in planning the 3D VSP. In the velocity model, low velocities areblue and high velocities are red. (Modified from Sanchez and Schinelli, reference 14.)

N

> The Riacho de Barra field, onshore Brazil. A cross sectioninterpreted from well logs (top) shows the main reservoir (yellow)and the lower target (orange). Both are truncated at their uppersurfaces by erosion and are overlain by a conglomerate thatobscures seismic signals. After a 3D surface seismic survey failed toadequately image the erosional truncation, Petrobras acquired a 3DVSP to better delineate the limits of the reservoir. (Modified fromSanchez and Schinelli, reference 14.)

60527schD5R1.qxp:60527schD5R1 1/8/08 9:19 AM Page 27

and drilling the 4-m [13-ft] shot holes for deploy -ment of the dynamite sources. Rugged topog raphyand a forested landscape added difficulty to theacquisition campaign. No rig was available ateither well location, so a crane was mobilized todeploy the long receiver tools.

Because data recording requires good couplingbetween receiver and formation, the two wellswere evaluated for cement bond quality. A well-intervention team performed cement squeezes inboth wells to guarantee transmission of signalsfrom the formation through the cement and casingto the accelerometer receivers in the well.

Before acquisition of the 3D VSPs, a 115-levelconventional VSP was acquired in each well. Thequality of recorded data helped optimize thedepth location of the VSI arrays for the 3Dacquisition, and the velocity information fromeach well was used to facilitate processing of the3D VSP.

To reduce complexity of data processing, the3D VSPs from each well were handled separately,and then merged before the final stage ofmigration. The imaging results show an increasein resolution over that of the 3D surface seismicdata (above). Interpreters are currently workingwith the new 3D VSP data to define the limits ofthe reservoir.

28 Oilfield Review

15. “Thunder Horse: No Ordinary Project,” http://www.bp.com/genericarticle.do?categoryId=9004519&contentId=7009088 (accessed October 8, 2007).

16. Camara Alfaro J, Corcoran C, Davies K, Gonzalez Pineda F,Hill D, Hampson G, Howard M, Kapoor J, Moldoveanu Nand Kragh N: “Reducing Exploration Risk,” OilfieldReview 19, no. 1 (Spring 2007): 26–43.

17. Ray A, Hornby B and Van Gestel J-P: “Largest 3D VSP in the Deep Water of the Gulf of Mexico to ProvideImproved Imaging in the Thunder Horse South Field,”Expanded Abstracts, 73rd SEG Annual International

Meeting and Exposition, Dallas (October 26–31, 2003):422–425.Jilek P, Hornby B and Ray A: “Inversion of 3D VSP P-WaveData for Local Anisotropy: A Case Study,” ExpandedAbstracts, 73rd SEG Annual International Meeting andExposition, Dallas (October 26–31, 2003): 1322–1325.Pfau G, Chen R, Ray A, Kapoor J, Koechener B andAlbertin U: “Imaging at Thunder Horse,” ExpandedAbstracts, 72nd SEG Annual International Meeting andExposition, Salt Lake City, Utah, USA (October 6–12,2002): 432–435.

Inline, Crossline and Time Slice

3D VSP Cube

700 m

Surface Seismic Section 3D VSP

> Thunder Horse field in the Mississippi Canyon, Gulf of Mexico (left). BP ran several 3D VSPs in this area, which has numerous salt intrusions that reduce the effectiveness of surface seismicsurveys. Three-dimensional VSPs can be designed so that many raypaths avoid propagation throughthe salt (right).

New Orleans

km

miles

150

150

0

0

Gul f of Mex ico

ThunderHorse

Salt

VSI tool

Target

> Petrobras 3D VSP results. The borehole survey produced high-resolution results that can be interpreted using software designed for3D surface seismic data interpretation, including cube displays (topleft), and inline, crossline and time-slice displays (bottom left). Theresolution of the 3D VSP data was superior to that of the surfaceseismic data over the same area (right).


Autumn 2007 29

Gulf of Mexico 3D VSPsAn example of a marine VSP comes from the BP-operated Thunder Horse field in the south-central Mississippi Canyon, Gulf of Mexico. Thefield is in water depth of approximately 6,300 ft,[1,920 m], and is home to the largest mooredsemisubmersible rig in the world.15

Seismic imaging in the area is extremelycomplicated because of the abundance of over -lying salt bodies. Resolving structural complexityand stratigraphic detail is necessary for success,but difficult with 3D seismic data because the salt obscures major subsalt targets. Three-dimensional surface seismic data suffer fromwater-bottom and salt-sediment multiples, andfrom attenuation at the deeper reservoir levels.

Three-dimensional VSPs can be designed toreduce wave propagation through the salt(previous page, bottom). Avoiding raypathsthrough the salt eliminates some of thechallenges inherent in conventional surfaceseismic surveys. And with VSPs, the reflectedenergy travels a shorter path, reducingattenuation and improving resolution. The true3D geometry also produces data from a widerange of azimuths, a feature that improvesillumination in surface seismic surveys.16

Day rates for deepwater drilling rigs are high,and 3D VSP acquisition can take several days toa few weeks, so the operation must be efficient.At the time of the first 3D VSP in Thunder Horse,a VSI tool with 12 three-component shuttles wasavailable, the most that could be run. Standardpressures and temperatures were expected:17,400 psi [120 MPa] and 275°F [135°C].17

The first 3D VSP was completed in February2002 in the Mississippi Canyon 822-3 Well. The12-shuttle VSI tool was positioned at threeconsecutive depths to produce an effective 36-level VSP. A spiral source pattern was selectedfor efficiency, and repeated for each receiver-array depth, firing approximately 30,000 shots,and generating more than one million traces(above right). The image was found to be muchsuperior to the available surface seismic data,with markedly higher resolution, less noise andfewer artifacts (right).

Using the multilevel VSI tool enabledefficient and cost-effective acquisition of 3D VSPdata around targeted wells. High-resolutionimages from these VSPs can be used to guideplacement of development wells, and imagesfrom multiple wells can be combined to give amore comprehensive image of the subsurface.

> Spiral 3D VSP. A spiral shooting pattern included operation of a dual-source array and flip-flop shooting, with the source vessel first firing asource on the port side (blue dots), then a source on the starboard side(green dots). The spiral was repeated for each receiver-array depth.(Modified from Ray et al, reference 17.)

Y-Of

fset

, km

6

4

2

0

–2

–4

–6

–6 –4 –2 0 2 4 6

X-Offset, km

> Comparison of 3D VSP results with a 3D surface seismic line. The 3D VSPdata (left) show higher resolution everywhere compared with surfaceseismic data (right). (Modified from Ray et al, reference 17.)

15,250

Dept

h, ft

17,750

20,250

22,750

25,250

27,750

5,000

Distance, ft0–5,000 10,000

3D VSP Surface Seismic Section


Marine 3D VSPs can even be run without adrilling rig. One example comes from the GreenCanyon area of the Gulf of Mexico, where acomplex salt body overhanging the Mad Dog fieldcreated a shadow zone that made it difficult toobtain a clear image from surface seismic data.18

After casing was set at TD, a well in the field wastemporarily abandoned, and the rig was moved todrill another well from the same deck. To acquirea 3D VSP in the first well, a wireline winch,capstan and acquisition unit were installed onthe aft end of the semisubmersible’s main deck.Through this opening, a 20-level VSI array with100-ft [30-m] spacing between shuttles was runinto 4,500 ft [1,370 m] of open water and thencaught and guided into the subsea wellhead by aremotely operated vehicle (ROV). A video feed ofthe operation allowed the winch man and loggingengineers to coordinate tool deployment with theROV operator.

Once the receiver array was in place, dataacquisition continued efficiently, with nononproductive time. The source vessel, theWesternGeco Snapper, towed a three-gun arrayand shot two walkaway lines, and then shot thespiral survey geometry. The VSI system acquiredthe 32,000-shot 3D VSP in six days. BP realizedsubstantial savings by not using rig time for the acquisition.

Results from the Mad Dog 3D VSP helpedproduce an improved image in an area wheresurface seismic data had been affected by over -hanging salt (above left). Interpreters delineateda fault of approximately 1,640-ft [500-m] throwthat had caused an early well to completely missthe pay interval. Of three wells drilled into thestructure before the availability of the VSP, onehit the target in the right place, and logs from allthe wells corroborated the fault location and dipinterpreted from the borehole seismic data. BPdetermined that the cost of drilling two of thesidetracks could potentially have been saved if the 3D VSP had been acquired before drillingthe first well.

Optimizing Hydraulic Fractures in Real TimeBorehole seismic tools have been used since the1980s to detect seismic energy generated byhydraulic fracture treatments.19 The goal is to useknowledge of the fracture geometry and spatialdevelopment to help improve fracture operations.20

The ability to make decisions that can optimizestimulation treatments relies on two mainrequirements: receiving accurate informationabout fracture propagation in time to changeongoing operations, and having the technology toeffect the desired change.

30 Oilfield Review

> Rigless 3D VSP in the Gulf of Mexico. While the rig was being used todrill one well, a 3D VSP was performed in another well by running a 20-levelVSI tool through a false rotary on the aft end of the semisubmersible deck(left). In an image from the VSP data, a large-throw fault (purple) explainswhy some wells drilled into the structure did not hit the pay zone (red). Well 1 encountered the fault but failed to reach the reservoir. Well 2intersected a small portion of the pay zone, and Well 3 hit the pay in thecorrect location. Fault location and dip information from dipmeter logs (blue)confirm the fault interpretation on the VSP image. (Modified with permissionfrom Hornby et al, reference 18.)

500 m

3

2

1

> Estimated stimulated fracture networks and a horizontal well in theBarnett Shale formation. Vertical wells (circles) penetrating the BarnettShale produce from stimulated areas approximated by the shaded areas(left). The operator drilled a horizontal well (black line) to tap undrainedareas. The wellbore trajectory (right) dipped low at the heel of the well thenrose 30 ft [9 m] over the 2,000 ft [610 m] between heel and toe. The fiveperforation clusters in the toe section of the well (red and green) are theentry points for Stage 1 of the hydraulic fracture treatments. Blue dots arethe entry points for Stage 2.

7,420

7,410

7,430

7,440

7,450

7,460

7,470

7,480

True

ver

tical

dep

th, f

t

9,500 9,000 8,500 8,000 7,500

Measured depth, ft

Horizontal Wellbore TrajectoryBarnett Shale Production Areas


Autumn 2007 31

To address the first requirement,Schlumberger has developed an innovativehydraulic fracture monitoring technique thatprovides stimulation engineers with real-timeinformation pertaining to the geometry anddevelopment of hydraulically induced fracturenetworks. Real-time results allow operatingcompanies to make timely decisions to alter thefinal geometry of fractures and reduce or preventsuch undesirable situations as water production,overlap with previous treatments, fluid loss anduneconomic pumping.

The ability to change the outcome of astimulation treatment depends on the problemat hand. If the fracture is developing out of itsplanned zone, a decision can be made to end thejob. If the treatment is not reaching the desiredintervals, pumped fluids can be adjusted to sealcompeting zones. Diversion technology caneffectively bridge fracture systems and createadditional complex fractures.

One operator used StimMAP hydraulicfracture stimulation diagnostics to track theprogress of a multistage fracturing operation in ahorizontal well in the Barnett Shale. Thisformation in the Fort Worth basin of north-central Texas is the most active gas play in theUnited States. The Barnett Shale formation is a naturally densely fractured, ultralow-permeability reservoir that requires a largehydraulic fracture surface to be effectivelystimulated, and hence be economic.

The horizontal infill well was drilled in thedirection of minimum principal stress tofacilitate creation of transverse hydraulicfractures. The estimated stimulated fracturenetworks of several nearby hydraulicallyfractured vertical wells intersected the heel

section of the well (previous page, bottom).These regions of low stress caused by previousstimulation treatments will tend to attractpropagating fractures, making it potentiallydifficult to stimulate the toe of the well.

The treatment was designed to comprise twostages, with the first stage targeting fiveperforation clusters nearest the toe of the well.From the microseismic events localized inStage 1a, it is clear that the fracture developedaway from the higher stress interval near the toe,and extended toward the lower stress interval inthe heel, leaving the toe section understimulated(right). A diversion stage was pumped to try todivert the next treatment to the far perforations.Monitoring the seismic activity during Stage 1bindicated that again the toe section of the wellwas not fracturing, and again, stages of diversionfluid were pumped to try to divert fluid from thecompeting zones.

Inspection of the microseismicity maprevealed that seismic events were occurring nearthe first two perforation clusters, but not beyond.Coiled tubing was run to see if some type ofobstruction was preventing a fracture frominitiating between the second and thirdperforation clusters. Engineers determined thata sand plug was prohibiting stimulation in thatsection of the well.

After the sand plug was removed, Stage 1csuccessfully stimulated the toe section.Immediately, microseismic events were detectedin the previously unstimulated sections of thetoe. With additional diversion stages pumpedwhenever the real-time microseismicity ceasedto grow, the operator was able to stimulate the900-ft [274-m] toe section of the lateral withoutusing numerous time-consuming bridge plugsand perforating steps. A subsequent stagetreated the heel of the well, which was alsomapped by microseismic activity.

18. Hornby BE, Sharp JA, Farrelly J, Hall S and Sugianto H:“3D VSP in the Deep Water Gulf of Mexico Fills in Subsalt‘Shadow Zone’,” First Break 25 (June 2007): 83–88.

19. Albright JN and Pearson CF: “Acoustic Emissions as aTool for Hydraulic Fracture Location: Experience at theFenton Hill Hot Dry Rock Site,” SPE Journal 22, no. 4(August 1982): 523–530.

20. Fisher MK, Heinze JR, Harris CD Davidson BM, Wright CAand Dunn KP: “Optimizing Horizontal Completion Techniques in the Barnett Shale Using MicroseismicFracture Mapping,” paper SPE 90051, presented at theSPE Annual Technical Conference and Exhibition, Houston, September 26–29, 2004.Ketter AA, Daniels JL, Heinze JR and Waters G: “A FieldStudy Optimizing Completion Strategies for Fracture Initiation in Barnett Shale Horizontal Wells,” paper SPE 103232, presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, September 24–27, 2006.Le Calvez JH, Klem RC, Bennett L, Erwemi L, Craven Mand Palacio JC: “Real-Time Microseismic Monitoring ofHydraulic Fracture Treatment: A Tool to Improve Comple-tion and Reservoir Management,” paper SPE 106159,presented at the SPE Hydraulic Fracturing TechnologyConference, College Station, Texas, January 29–31, 2007.

> Microseismic events mapped during progres-sion of hydraulic fracture treatments. Stage 1a(top) stimulated the region near the heel of thewell, but left the toe mostly unfractured. Diver-sion fluid was introduced to divert the nexttreatment to the perforation clusters at the toe.Stage 1b (second) also failed to stimulate the toe,and indicated an obstruction in the well betweenthe second and third perforation clusters. Follow-ing removal of a sand plug, Stage 1c (third)successfully stimulated the remaining 900-ft toesection. When all stages are plotted together(bottom), it can be seen that Stage 2 stimulatedthe heel section of the well (dark blue dots).

500 ft

Stage 1a

Obstruction

Stage 1b

Stage 1c

Stages 1 and 2


Monitoring Perforating OperationsShell Exploration & Production was constructingproduction wells in the Cormorant field, UKNorth Sea. The wells were to be perforated withtubing-conveyed perforating (TCP) guns. Shellhad considered several methods of verifying TCPoperations, and decided to try monitoring theshots with a borehole seismic tool. In wireline-conveyed perforating, changes in cable tensioncan indicate that the guns have fired, and thiscan be confirmed when the guns are retrievedand inspected on the surface. In tubing-conveyedperforation, the guns may be left in the well andnever returned to the surface. Without positiveindications that the guns have fired, the onlyproof of the operation’s success is pulling thetubing and retrieving the guns, at great expenseto the operator.

Although the VSI tool is designed to recordborehole seismic surveys, the receivers are alsoable to detect signals generated by disturbancesin the vicinity of the borehole. The tool wouldundoubtedly be able to detect signals from asource as powerful as the shaped charges usedfor perforating if it were run in the same well.Unlike other borehole seismic tools, the VSI toolcan be used to acquire records of any timeduration. In typical deployment for logging

borehole seismic surveys, the recording length isset to approximately 5,000 ms and starts at theactivation of the controlled seismic source.However, for monitoring perforation shots, therecording system was set to begin recording oncethe tool had been anchored in position, and tocontinue recording until switched off by theseismic field engineer.

The wells were to be multilaterals with amain bore and one lateral bore. Typically, afterthe main bore was drilled and cased, more than 3,000 ft [910 m] of TCP guns were run to the reservoir interval and left in place, to be detonated by a trigger-delay system. Awhipstock—for exiting the casing to drill thelateral bore—was then set in the main wellboreabove the interval to be perforated. A VSI shuttlewas anchored 100 ft [33 m] above the whipstockto monitor the detonation of the perforating guns(above). After the firing of the guns and drilling,completing, perforating and cleaning up thelateral bore, the whipstock was perforated toallow the reservoir penetrated by the main boreto flow.

The VSI tool detected the sharp onset of signalfrom the perforation shots (next page, top). Thetool was close to the whipstock, and the largemagnitude of the signal saturated the dynamicrange of the recording system. Although the

amplitude cannot be read from the recording, anincrease in frequency of the signal can be detectedfor several seconds after the onset. Signal levelreturned to the level of background noiseapproximately 8 seconds after signal onset. Theseismic signals confirmed the successful firing ofperforation guns.

The primary purpose fulfilled, Shell engineersexamined the seismic data for additionalinformation. The guns had fired, and the emptyguns had filled with fluid. The return of theseismic signal level to background noise levelsindicated that fluids were no longer moving in thisportion of the borehole. The total duration ofsignal on the seismic record was interpreted torepresent the time it took the empty gun volumeto fill, and could be related to the inflow perfor -mance of the well. Given that the borehole belowthe whipstock is a closed system, and knowing thevolume of the perforating guns, effectively achamber at atmospheric pressure, Shell engineersincluded the time required to fill the guns in acalculation to obtain a rough estimate of absoluteopen-flow potential. With this additionalinformation from the seismic monitoring ofperforation shots, Shell engineers gainedunderstanding of reservoir behavior.

32 Oilfield Review

> Monitoring TCP operations with a borehole seismic receiver. Perforating guns were conveyed by coiled tubing, left on the bottom of the hole and set tofire with a long delay. After a whipstock was set, a VSI tool was deployed through drillpipe and anchored 100 ft above the whipstock. Detonation of theguns created seismic signals recorded by the sensors.

VSI tool

Whipstock Perforating guns


Autumn 2007 33

High-Pressure, High-Temperature SurveysWhile the VSI tool can log borehole seismicsurveys in most wells, high-pressure, high-temperature (HPHT) wells have specialrequirements. The seismic acquisition tooldeveloped for the SlimXtreme slimhole high-pressure, high-temperature well logging platformcombines high-performance packaging withanalog recording, minimizing the use of fragileelectronics (right). This 33⁄8-in. tool, like the othertools in the Xtreme family, was engineered tooperate in conditions up to 30,000 psi [207 MPa]and 500°F [260°C]. The short, lightweight sondewas designed with a single three-component setof receivers to handle checkshot surveys, but isnow also being used to acquire full VSP images inHPHT wells.

ConocoPhillips (U.K.) Limited had severalreasons for running the slim analog seismic toolin a challenging HPHT well drilled in the centralNorth Sea. The first was to generate an accuratetime-depth correlation between well data andthe time-based 3D marine seismic data over thetarget. While the reflection at the base of thechalk was clearly interpretable in seismicsections, the deeper reflection at the top of thereservoir was not as easy to pick. Correlationbetween VSP, well log and surface seismic datawould increase confidence in interpreting theshape and extent of the reservoir.

> Seismic recording of perforation shots andother events. This display is a continuous record,starting at the top, with the second line a continu-ation of the first, and so on. For each line, thevertical axis is signal amplitude. The signal fromthe perforation shots appears with a sharp onsetat 04:44:22. The signal saturates the dynamicrange of the recording system for several sec-onds. The recording returns to background noiselevels at 04:44:30, but some isolated noise burstsoccur earlier and later.

Tim

e

04:44:03

04:44:15

04:44:27

04:44:39

04:44:51

0 0.5 1.0 1.5 2.0 2.5 3.0Time, s

Onset of signal at04:44:22, October 25

Signal heavily saturated as dynamic range of recordingsystem is unable to cope with magnitude of event.Frequency of the event increasing with time.

Signal saturates completelyover this short duration ofrecording time Isolated noise burst about 6.7 s

after main energy burstReturn to background

noise levels at04:44:30, October 25

Random noise bursts observed aftercessation of main event. Magnitudeis down in the noise level.

> Borehole seismic acquisition tool for extreme conditions. The SlimXtreme slimhole high-pressure,high-temperature logging platform operates in conditions up to 30,000 psi and 500°F. Operatingcompanies have used the tool in conditions up to 238°C [460°F].

StandardSchlumberger

wireline unit

High-strengthwireline cable

High-strength wirelinedual-drum capstan

Airgun


ConocoPhillips also wanted to acquire adepth-based VSP image of the reservoir intervaland layers below the well TD. In the surfaceseismic data, the dipping reservoir layers arepartially disrupted by noise from multiples,which appear as horizontal reflections thatinterfere with the signals from the reservoir.Because a VSP records both downgoing andupgoing waves and features multicomponentprocessing, a VSP image may contain fewermultiples and give a more accurate picture of thereservoir structure. And by extending the imagebelow the well, it would be possible to correlatehorizons beneath the reservoir with reflectionsseen on surface seismic data.

The third reason for acquiring VSP data wasto obtain better estimates of formation velocitiesfor improved reprocessing of the 3D marineseismic data. Reducing uncertainties in thevelocities of the chalk and underlying formationswould produce more accurate 3D images,potentially leading to reduced risk in futuredrilling in the area.

The slim analog seismic tool was the onlyoption for acquiring a VSP in the expectedpressure and temperature conditions. With TDbelow 15,000 ft [4,600 m], temperatures could beas high as 380°F [193°C]. The well trajectory was

deviated above the chalk, and then sidetrackedout of the plane of deviation as depth increased.

In spite of the extreme conditions, loggingproceeded smoothly. The tool acquired data atreceiver stations every 50 ft [15 m] spanning adepth interval from the reservoir up through thechalk, and also at more widely spaced intervalshigher in the section. At the deepest of the73 stations, temperature reached 380°F. Theseismic source comprised three 150-in.3 airgunsand was deployed at the rig in a zero-offset survey configuration.

Processing the three-component data todetermine where the reflections originatedincluded standard steps as well as a specialcorrection for the 3D nature of the boreholetrajectory. This would allow the VSP data to bemigrated using a 2D algorithm. The 3D trajectoryof the borehole was projected onto a verticalplane aligned with the shallow portion of the well(left). Reflection times, locations and amplitudeswere calculated assuming the VSP signals wereconfined to this plane, but in reality, somereflections occurred out of the plane. To takeaccount of this, raypaths and traveltimes for eachtrace were calculated using the 3D velocitymodel derived from initial surface seismicprocessing, and compared with raypaths andtraveltimes calculated from a 2D modelextracted from the 3D volume in the dominantvertical section chosen for processing. Thedifference between the two sets of computedtraveltime residuals was added as a staticcorrection to each trace prior to migration.

The differences in the velocity models alsoindicated that the VSP detected higher velocitiesin the chalk layer and lower velocities below thechalk than were seen in the surface seismicvelocity model. These differences translate intomis-ties observed between the VSP image andthe surface seismic image below the chalkinterval (below).

The depths of reflectors in the VSP image alsomatched those of a synthetic trace generatedfrom sonic and density well logs, confirming theaccurate depths of the VSP image in spite of theconflict between the 3D nature of the acquisitionobjective and the 2D approach to solve it (nextpage, top left). ConocoPhillips (U.K.) Limited isusing velocities obtained from the boreholeseismic survey to aid the reprocessing of existingsurface seismic data, and plans to use the slimanalog seismic tool in future HPHT wells.

Waves of the FutureBorehole seismic surveys have advanced farbeyond their origins as methods for convertingtime to depth for well-to-seismic correlations,although they are still used primarily for time-depth ties. As seen in this article, VSPs cansatisfy a wide variety of needs, providing 3Dimages of the subsurface, contributing tooptimized hydraulic fractures, verifying perfo -rating operations and obtaining high-quality datain HPHT conditions.

34 Oilfield Review

> Trajectory of the ConocoPhillips North SeaHPHT well. In this plan view, the source positionis a blue sphere, the receivers in the boreholeare green dots, and the reflection points on thetarget are in shades of blue and white. Theupper portion of the well follows an azimuth ofN61E, then veers to the northwest with depth.The source-receiver geometry and traveltimeswere projected onto a vertical section alongN61E to define a single azimuth with which tomigrate the data.

61°

> Comparison of VSP results with surface seismic data. The surface seismic image produced usingchalk velocities that are too low (left) fails to tie with the VSP (right). (The VSP is a small region withhigher amplitudes and higher resolution than the surface seismic image, and narrows upward.) Themismatch can be seen at several intervals.

X.250

X.500

X.750

Y.000

Y.250

Two-

way

tim

e, s

Chalk

Well TD

21. Hornby et al, reference 9.22. Djikpesse H, Haldorsen J, Miller D and Dong S: “Mirror

Imaging: A Simple and Fast Alternative to InterferometricMigration of Free-Surface Multiples with Vertical Seismic Profiling,” submitted to Geophysics, 2007.


Autumn 2007 35

The future of VSPs will undoubtedly take manydirections. Hardware innovations will include newdownhole tools to withstand demanding condi -tions and new sources to enable even moreefficient acquisition. Permanent instal lation,allowing long-term reservoir monitoring, has beentested by some operators.21 Permanently installedtools could be used to conduct time-lapse surveysor to detect seismicity induced by production orinjection, even when deployed in the producing orinjection wells.

Other advances will come in processing toproduce better images from acquired data. Mostprocessing for creating images from VSP data hasborrowed from surface seismic methods. Butborehole seismic surveys, with their particulargeometries, offer opportunities that have notbeen fully explored.

One promising area is called interferometry,which is the interference of two or more waves toproduce an output wave that is different from theinput waves. Researchers are investigating waysto use interferometry to transform signalspreviously considered as noise into valuableinformation. For instance, in typical VSP dataimaging workflows, only primary reflections aremigrated. Free-surface multiple reflections areusually regarded as noise and thus eliminatedbefore migrating the recorded data. Whilebenefiting from reduced attenuation andimproved velocity control with respect tomigrated surface seismic data, the resultingmigrated VSP images are restricted to arelatively narrow zone of illumination lying belowthe borehole receivers. However, free-surface-related multiples contain valuable information

about shallower subsurface structures, and ifproperly migrated, they can provide widerillumination, and better vertical resolution of thesubsurface properties than when imaging usingprimaries alone (above).22

The early goal of VSPs was to reduce risk byenabling accurate time-to-depth correlationbetween surface seismic data and well logs.Current and future capabilities of boreholeseismic surveys still include risk reduction, butalso extend to improving recovery. –LS

>Matching reflector depths in a VSP image and a log-derived synthetictrace. One test of properly depth-correlated seismic data is matching witha synthetic trace generated from sonic and density well logs. In this case,the synthetic trace is plotted in yellow for visibility, and only positiveamplitudes are plotted, so as not to obscure the seismic data. Throughoutmost of the well, the positive amplitudes in the synthetic trace correlatewith those in the VSP, giving confidence in the projection assumptions madeduring processing. The VSP image extends beyond the bottom of the well.

X.500

X.750

Two-

way

tim

e, s

Y.000

> Mirror imaging, an example of interferometry.The free surface and the area above it arereplaced by a mirror image of a medium with thesame elastic properties as the mediumcontaining the borehole and receivers. Receiversin the new material are the mirror image of theoriginal receivers. Whereas the original boreholeseismic experiment had a zone of illuminationrestricted to below the receivers, the mirroredexperiment has a zone of illumination thatextends to the former free surface.

Free surface

Mirroredreceivers

Downholereceivers

Source


36 Oilfield Review

Rocks Matter: Ground Truth in Geomechanics

John CookCambridge, England

René A. FrederiksenKlaus HasboHess Denmark ApSCopenhagen, Denmark

Sidney GreenArnis JudzisJ. Wesley Martin Roberto Suarez-RiveraSalt Lake City, Utah, USA

Jorg HerwangerPatrick HooymanDon LeeSheila NoethColin SayersHouston, Texas, USA

Nick KoutsabeloulisRobert MarsdenBracknell, England

Morten G. Stage DONG EnergyHørsholm, Denmark

Chee Phuat TanKuala Lumpur, Malaysia

For help in preparation of this article, thanks to Ben Elbel,Dallas; Ian Walton, Rosharon, Texas; and Smaine Zeroug,Clamart, France. Thanks also to Hess Denmark ApS, DONGExploration and Production A/S, Noreco ASA, and Danoil forcontributing their North Sea case study.ECLIPSE, Petrel, TerraTek, UBI (Ultrasonic Borehole Imager)and VISAGE are marks of Schlumberger.

Stress and pressure act upon every reservoir, wellbore and completion. Drilling,

production and injection processes modify these stresses and pressures, sometimes

to the operator’s detriment. Through advances in geomechanical measurements,

modeling and monitoring, E&P companies are now able to predict and mitigate the

effects of stress and pressure as they change throughout the life of their fields—

from appraisal to abandonment.

Change the stress on a rock and it deforms,altering its volume and geometry, as well as thepaths of fluid flow within. The stress regime of aformation can be impacted by multiple factors,including rock type, depositional settings, regionaltectonics, episodes of erosion or uplift, localseismic disturbances and even tidal variations.The influence of such stressors is furthercomplicated by differences in rock fabric.

The manner in which formations react tochanging stress is becoming a focus of increasinginterest to E&P companies. In-situ reservoirstresses, having reached equilibrium overgeologic time, are altered by the process ofdrilling, production and injection. If drilling- orproduction-induced changes in stress are notanticipated, the challenges and costs of managinga prospect can far exceed an operator’s initialexpectations. To characterize stress, strain anddeformation in their reservoirs, E&P companiesemploy the discipline of geomechanics. Thiswide-ranging field applies solid and fluidmechanics, engineering, geology and physics todetermine how rocks and the fluids they containrespond to force or to changes in stress, pressureand temperature caused by drilling, completionand production.

In the past, most drilling and productiondepartments were not particularly attuned toformation stresses and geomechanics; manyreservoirs were deemed technically straight -forward and had undergone only limited

depletion. But declining resource volumes andfavorable oil prices are prompting operators todrill deeper, more intricate well trajectories, atthe same time that new technologies areextending the lives of mature fields. Operatorstherefore are becoming more mindful ofgeomechanics as they assess drilling andproduction difficulties—especially those whoendeavor to protect their investments inexpensive completions installed in high-pressure, high-temperature, tectonically activeor ultradeepwater prospects.

Failure to appreciate the importance of geo -mechanics may have severe consequences.Excessive mud loss, wellbore instability, casingcompression or shearing, reservoir compaction,surface subsidence, sand production, faultreacti vation and loss of reservoir seal may all be manifestations of stress changes within a formation.

Some operators are forced to react to changesin stress or rock fabric as they drill and producetheir wells. Others are more proactive. Throughcore testing and geomechanical model ing of rockstrength, deformation and stress behavior, theyare engineering better wells and fields. Theseefforts have recently been aided by newlyestablished centers of excellence forgeomechanics in Bracknell, England, and inHouston and Salt Lake City, Utah, USA.

This article describes advances in geo -mechanics laboratory testing techniques,stress-dependent reservoir simulations and


Autumn 2007 37

monitoring. Field studies performed at theSchlumberger Geomechanics Laboratory Centerof Excellence and the Schlumberger ReservoirGeomechanics Center of Excellence show howthis science is helping E&P companies optimizedrilling and production in increasinglychallenging reservoirs.

Stress in the SubsurfaceThe stresses acting on a formation can vary inorigin, magnitude and direction. Natural, in-situvertical stresses stem primarily from the weightof overburden. Horizontal stresses also have agravitational component that may be enhancedby tectonics, thermal effects and geologicalstructure. However, other factors such as litho -

logy, pore pressure and temperature influencestress magnitude and orientation as well as thedegree to which rock responds to stress.

Stress, a measure of force acting on a givenarea, is made up of normal and shearcomponents. Normal stress (σ) is that which isapplied perpendicular to a plane or rock surface.Shear stress (τ) is applied along the face of the

Majoreffectivestress σ1

Tensile strength Minor effective stress σ3

Uniaxial compressive strength


plane. Mathematically, there is one orientation oforthogonal axes defining the stress directions forwhich the shear stresses are zero. Thatorientation defines the principal axes of stress,wherein applied stresses are strictly normal.

In situ, these orthogonal principal axes areoften assumed to be oriented vertically andhorizontally (left); however, this condition isoften not the case. The magnitude andorientation of stresses in the earth change withthe structural dip of the formation, which canrotate the orientation of principal stresses fromthe vertical and horizontal orientations, as canthe presence of faults, salt diapirs, mountains orother complex structures.1

In the earth, where deformation is restricted,the three stress components are linked, and anychange of stress in one direction is accompaniedby changes in stress along the orthogonal axes.For example, when continued deposition bringsabout greater burial depths, the resultingincrease in overburden vertical stress cangenerate changes in horizontal stress, dependingon the degree to which the formations are able tospread out laterally. This response is generallyconstrained by the presence of adjacentformations that confine the rock deformation.Differences in formation properties also imposecontrasts in stresses between adjacent litho -logies. Furthermore, formation anisotropy canresult in greater lateral stress in one directionthan in another.

A body of rock responds to applied stressthrough various modes of strain, or deformation,causing changes in volume and shape, oftenaccompanied by changes in rock properties(left). The spectrum of deformation ranges fromreversible, or elastic deformation, to permanent,or plastic defor mation, before eventually endingin failure of the rock. Deformation caused bycompression, tension or shear can result incompaction, extension, translation or rotation,eventually ending in failure by shearing,fracturing or faulting. In addition to themagnitude of stress applied, a rock’s response tostress depends largely on rock type, cementation,porosity and burial depth. In sandstones, thesize, shape and area of contact points betweenindividual rock grains influence deformation. Inlimestones, the shape and strength of theskeletal rock framework influence deformation.2

Small increases in stress generally cause asmall strain from which the rock may recover.

38 Oilfield Review

> In-situ stresses and principal stresses. Stresses on a cube of materialburied in the earth are given the designation σV, σH and σh, where V indicatesvertical, H indicates the direction of the larger horizontal stress, and h that ofthe smaller horizontal stress. For simplicity, it is often assumed that these arethe principal stress directions, but the principal directions of stress can berotated significantly from these three axes. The principal stresses aregenerally indicated as σ1, σ2 and σ3, in decreasing order of magnitude. Whenthe principal stress directions do not coincide with the vertical and horizontaldirections, there will also be shear stresses on the cube faces in theorientation shown.

σV

σV

σHσH

σh

> Stress-strain diagram. Rocks that undergo elastic deformation store strainenergy as their volume changes. When the applied boundary stresses areremoved, the rock returns to its original state of deformation while the strainenergy returns to its original value. With application of greater stress, rocksundergo inelastic deformation as nonrecoverable, internal structural changesoccur (starting at the yield point), such as tensile microcracking, grain crushingor slippage at grain boundaries. These changes result in permanent volumetricdeformation, often referred to as plastic deformation. Higher stresseseventually cause the rock to fail (fracture point), as exemplified by crushing orfracturing of constituent grains and cement or by mineral dissolution.

Stre

ss

Strain

Yield point

Fracture point

Ductile field

Elastic field


Autumn 2007 39

Beyond a certain point, it will deform plasticallyor fail. The mode of deformation and failure isdictated by the relationship between changes inmaximum and minimum stresses (right). Thisrelationship is called a stress path.3 In petroleumgeomechanics, the stress path (K) isconventionally the ratio of change in effectiveminimum horizontal stress to the change ineffective vertical, or overburden, stress frominitial reservoir conditions during fluid pressure drawdown, more simply expressed as K = Δσ3/Δσ1. This can also be expressed in termsof changes of shear stress (Q) and changes inmean stress (P'), as shown on the P'-Q diagram.4

A sufficiently low stress-path value impliesthat the rock will fail in shear, generating a shearplane. Shear strength increases as lateralconfining stress on the rock increases. Wherelarger stress-path values are seen, the rockundergoes compaction or reduction in porosity.This is most common in soft, high-porosity rockssuch as chalk, porous sands and diatomite.5

When subjected to differential stresses, otherrocks, such as salts, will tend to flow over time toreduce shear stresses and move towardshydrostatic stress states.

To manage reservoirs, oil and gas companiesmust contend with a variety of downholestressors—not all of which are caused by over -burden or tectonics. Pore pressure, tempera turedifferences and chemical interactions can alsoaffect localized perturbations in stress orienta -tion and magnitude.

Stress and pore pressure are intrinsicallylinked.6 In formation pore spaces, stress istransmitted to liquids or gases in the form ofpressure. The magnitude of pressure applied inany one direction is the same for all directions. Ifa fluid is compressed, it reacts by exerting anequal and opposite pressure outwards. Underpressure, pore fluids often take up some of thestress imposed on a formation. Thus pore

pressure is an important component of the netstress applied to a body of rock.

Temperature is yet another contributor to theoverall stress regime. Temperature differencesbetween drilling fluids and downhole formationswill result in heat transfer between the twomedia. Given the low thermal conductivity ofmost rocks, these temperature variationsgenerate large strain gradients that may causesevere fracturing and stress realignments. Sincethermal expansion of water in the pore space ismuch higher than that in the rock matrix, theheat transferred into a formation by drilling fluid will generate a larger volume expansion ofthe pore fluid and a corresponding increase inpore pressure.7

Thermal expansion of the rock matrix underconstrained conditions will generate further

stress. A reduction in effective mud support isoften associated with an increase in porepressure. This reduction, together with thethermal matrix expansion, will lead to a lessstable wellbore condition. Conversely, coolingthe formation may result in a more stablecondition because of decreased pore pressureand tangential stress. The reduction of tangen -tial stress may also lead to a lower hydraulicfracture gradient, and, in extreme cases, thetangential stress will become negative andinitiate hydraulic fracture.

Stress and pore pressure can also be affectedby interactions between rock and drilling fluid.Shales, which account for the majority of drilledsections in most wells, are particularly sensitiveto drilling fluids. Somewhat porous and usuallysaturated with formation water, these rocks maybe susceptible to chemical reactions with certain

1. Addis MA: “The Stress-Depletion Response ofReservoirs,” paper SPE 38720, presented at the SPEAnnual Technical Conference and Exhibition, SanAntonio, Texas, October 5–8, 1997.

2. Geertsma J: “Land Subsidence Above Compacting Oiland Gas Reservoirs,” paper SPE 3730, presented at SPE-AIME European Spring Meeting, Amsterdam, May 16–18, 1972.

3. For more on stress paths: Crawford BR and Yale DP:“Constitutive Modeling of Deformation and Permeability:Relationships between Critical State andMicromechanics,” paper SPE/ISRM 78189, presented atthe SPE/ISRM Rock Mechanics Conference, Irving,Texas, October 20–23, 2002.Rhett DW and Teufel LW: “Effect of Reservoir Stress Path on Compressibility and Permeability ofSandstones,” paper SPE 24756, presented at the SPEAnnual Technical Conference and Exhibition,Washington, DC, October 4–7, 1992.

Scott TE: “The Effects of Stress Paths on AcousticVelocities and 4D Seismic Imaging,” The Leading Edge 26,no. 5 (May 2007): 602–608.Teufel LW, Rhett DW and Farrell HE: “Effect of ReservoirDepletion and Pore Pressure Drawdown on In-SituStress and Deformation in the Ekofisk Field, North Sea,”Proceedings of the 32nd US Rock MechanicsSymposium. Rotterdam, The Netherlands: A.A. Balkema(1991): 63–72.

4. A relationship exists between the stress path, shearstress and mean stress. While the stress path (K) can be expressed as K = Δσ3/Δσ1, shear stress (Q) isexpressed as (Q = σ1-σ3), and effective mean stress (P')is [P' = (σ1+σ2+σ3)/3]. In laboratory uniaxial stress tests,in which the minimum and intermediate principal stressesare considered equal (σ2 = σ3), the slope η in the P'-Qplane corresponding to the stress path K is given by thisequation, following Crawford and Yale (reference 3):

5. Doornhof D, Kristiansen TG, Nagel NB, Pattillo PD andSayers C: “Compaction and Subsidence,” OilfieldReview 18, no. 3 (Autumn 2006): 50–68.

6. Addis, reference 1.7. Choi SK and Tan CP: “Modeling of Effects of Drilling Fluid

Temperature on Wellbore Stability,” Proceedings,SPE/ISRM Rock Mechanics in Petroleum EngineeringSymposium, Trondheim, Norway (July 8–10, 1998): 471–477.Li X, Cui L and Roegiers J: “Thermoporoelastic Analysisfor Inclined Borehole Stability,” Proceedings, SPE/ISRMRock Mechanics in Petroleum Engineering Symposium,Trondheim, Norway (July 8–10, 1998): 443–452.

> Distortion and failure. Distinct modes of distortion and failure can be plottedas a function of shear stress (Q) and mean effective stress (P'). At relativelylow P' and high Q, rock failure typically occurs as localized shear along aplane oriented at an angle to the principal stress axes. At relatively high P'and low Q, rocks may undergo compaction or pore collapse. (Adapted fromScott, reference 3.)

Dilation

CompactionNear-elastic region

Mean effective stress (P'): (σ1 + σ2 + σ3) / 3

Shea

r stre

ss (Q

): σ 1 –

σ3

Impossiblestates

Critical

state

line

Compaction surfaceShear failure surface

Tens

ile

failu

re su

rface

Ductile failure surface


drilling fluids. When a formation is drilled withan incompatible fluid, the invading filtrate maycause the shale to swell, which can lead toweakening of the rock and wellbore instability.Shales may also be susceptible to time-dependent changes in effective mud supportcaused by differences between the mud pressureand pore-fluid pressure, or between drilling fluidsalinity and formation salinity.8 Furthermore,volume changes in shales arising from inter -actions between shale and drilling fluid canlocally disturb the stress orientation andmagnitude in a borehole.

Thus, while local and regional tectonicstresses play a major role in rock deformation,other downhole factors, such as pore pressure,mud weight and downhole pressure fluctuations,temperature and chemistry must also beconsidered for their distinctive contributions tothe local stress-deformation continuum. Theireffects may also be tempered by textural proper -ties unique to the local lithology, such as the sizeand distribution of framework grains and pores,mineralogy and the composition of diageneticcements. Given the variety of reactions to stress,it is crucial for an operator to know as much aspossible about the rock surrounding a wellboreand the conditions to which it will be subjected.

Changes in Stress Drilling and production activities affect localstress regimes. Problems encountered duringdrilling may portend difficulties encounteredsubse quently during the production phase.Changes in stress may result in rock failure thatcauses wellbore instability during drilling. Thesechanges may later lead to sand production oncethe well has been completed. Other activitiesduring the life of a field can cause pore pressureand temperature changes, which can modifystresses acting farther from the wellbore. Stresschanges affect not only the reservoir but alsoadjacent formations.

Drilling activity perturbs the initial equilib -rium of stresses in the near-wellbore region. Asa cylindrical volume of rock is excavatedthrough drilling, the stresses formerly exertedon that volume must instead be transferred tothe surrounding formation. This process createstangential, or hoop stresses, which must beborne by the rock surrounding the borehole.These wellbore stresses are a function of mudweight, wellbore inclination, formation dipangle and azimuth, and the magnitude andorientation of far-field stresses (σV, σH and σh).Hoop stress varies strongly as a function of

borehole radius and azimuth.9 Furthermore, itcan far exceed σH (left).

In most conventional drilling operations,drillers use hydraulic pressure from drilling fluidas a substitute for the mechanical support that islost through the cylindrical volume of rockexcavated while drilling a wellbore. They essen -tially replace a cylinder of rock with a cylinder ofdrilling fluid. However, mud pressure is uniformin all directions, and cannot balance againstoriented shear stresses in a formation. As stressis redistributed around the wall of the wellbore,shear stresses can exceed rock strength. When thishappens, the wellbore will deform or fail entirely.

Typical examples of geomechanics-relateddrilling problems include wellbore instability andfracturing of the formation. Ramifications includefinancial loss resulting from lost circulation,kicks, stuck pipe, additional casing strings,sidetracks and even abandonment. To sustainwellbore stability, operators must develop drillingand well construction plans that consider stressmagnitude and direction, mud weight, trajectoryand pore pressure before, during and after a wellis drilled.

Drillers manage pressures imposed by mudweight to avoid wellbore stability problems.Their control of wellbore hydraulics reflects a petroleum engineering approach to ageomechanical problem. During drilling, well-bores can be compromised through a variety ofmud-induced modes of failure:10

• Tensile failure occurs by increasing mud pres-sure until it causes the wellbore wall to go intotension and eventually to exceed the rock’stensile strength. This fractures the rock alonga plane perpendicular to the direction of mini-mum stress, often resulting in lost circulation.

• Compressive failure may be caused by mudweight that is too low or too high. In eithercase, the formation caves in or spalls off, pro-ducing borehole damage and breakouts (nextpage, top). Unless the wellbore is properlycleaned out, the accumulation of breakoutdebris can lead to stuck pipe as the boreholepacks off or collapses.

• Shear displacement takes place when the mudpressure is high enough to reopen existingfractures that the wellbore has intersected. Asa fracture is opened, stresses along the open-ing are temporarily relieved, allowing oppos ingfaces of the fracture to shear, creating a smallbut potentially dangerous dislocation along the wellbore.

Wellbore stability is further affected bystructural factors, such as the interplay betweenwellbore inclination, formation dip and

40 Oilfield Review

> Plan view of hoop stresses surrounding a vertical wellbore. In this model,pore pressure and wellbore pressure are equal, while maximum and minimumeffective stresses within the formation equal 2,000 psi and 3,000 psi [13.8 and20.7 MPa], respectively. However, hoop stress, which varies as a function ofradius and azimuth, is strongly compressive along the azimuth aligned withminimum horizontal stress (σh) (red shading above and below the wellbore),where it reaches almost 7,000 psi [48.3 MPa]. Wellbore failure will be morelikely to occur along this axis. (Adapted from Sayers et al, reference 9.)

σH = 3,000 psiσH = 3,000 psi

σh = 2,000 psi

σh = 2,000 psi

2,000 3,000 4,000 5,000

Hoop stress, psi

6,000

Wellbore

7,000


Autumn 2007 41

directional variations in strength between andalong formation bedding planes (below right). Itis not unusual for some degree of wellbore failureto occur in vertical wells that encounter steeplydipping shales, or inclined wells that intersectshale bedding planes at low angles. Such failuresare initiated by low shear and tensile strengthalong planes of weakness in shales.11

The issue of strength, or a rock’s capacity towithstand stress, points to an important under -lying influence on deformation and failure: thatof rock fabric.12 Rock fabric can dictate whethera given amount of stress will cause a rock todeform or to completely fail, and can influencethe extent and orientation of fractures orbreakouts in a wellbore. Thus, although boreholebreakout is typically assumed to be orientedalong the axis of least stress, the bedding,cementation, mineralogy and grain size of a rockmay actually redirect the course of a breakoutalong the rock’s weakest points.

For help in anticipating and circumventingproblems such as those described above, some

8. Gazaniol D, Forsans T, Boisson MJF and Piau JM:“Wellbore Failure Mechanisms in Shales: Prediction and Prevention,” paper SPE 28851, presented at the SPE European Petroleum Conference, London, October 25–27, 1994.Mody FK and Hale AH: “A Borehole Stability Model toCouple the Mechanics and Chemistry of Drilling FluidInteraction,” in Proceedings, SPE/IADC DrillingConference, Amsterdam (February 22–25, 1993): 473–490.Tan CP, Rahman SS, Richards BG and Mody FK:“Integrated Approach to Drilling Fluid Optimization forEfficient Shale Instability Management,” paper SPE48875, presented at the SPE International Oil and GasConference and Exhibition, Beijing, November 2–6, 1998.van Oort E, Hale AH and Mody FK: “Manipulation ofCoupled Osmotic Flows for Stabilization of ShalesExposed to Water-Based Drilling Fluids,” paper SPE30499, presented at the SPE Annual TechnicalConference and Exhibition, Dallas, October 22–25, 1995.

9. Sayers CM, Kisra S, Tagbor K, Dahi Taleghani A andAdachi J: “Calibrating the Mechanical Properties and In-Situ Stresses Using Acoustic Radial Profiles,” paperSPE 110089-PP, presented at the SPE Annual TechnicalConference and Exhibition, Anaheim, California, USA,November 11–14, 2007.

10. For more on wellbore stability problems: Addis T, Last N,Boulter D, Roca-Ramisa L and Plumb D: “The Quest forBorehole Stability in the Cusiana Field, Colombia,”Oilfield Review 5, no. 2 & 3 (April/July 1993): 33–43.

11. Aoki T, Tan CP and Bamford WE: “Stability Analysis ofInclined Wellbores in Saturated Anisotropic Shales,” in Siriwardane HJ and Zaman MM (eds): ComputerMethods and Advances in Geomechanics: Proceedingsof the Eighth International Conference on ComputerMethods and Advances in Geomechanics, Morgantown,West Virginia, USA, May 22–28, 1994. Rotterdam, TheNetherlands: A.A. Balkema (1994): 2025–2030. Yamamoto K, Shioya Y, Matsunaga TY, Kikuchi S andTantawi I: “A Mechanical Model of Shale InstabilityProblems Offshore Abu Dhabi,” paper SPE 78494,presented at the 10th Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, UAE, October 13–16, 2002.

12. Rock fabric is a term that loosely encompasses themineral content, size, shape, orientation andcementation of component grains within a rock,including their overall arrangement into microscopiclaminations or larger beds.

> Formation effects on wellbore stability. Structural and stratigraphic factorscan combine to cause well damage. Here, incompetent beds overlie a strongerformation near the crest of a structure; relative movement results in damagedcement and collapsed casing.

> Borehole breakout. Results from a UBI Ultrasonic Borehole Imager loggingtool show the extent of stress-related damage in a wellbore. In isotropic ortransversely isotropic rock, where rock properties do not change along theplane of the wellbore, such damage is generally aligned along a plane ofleast horizontal stress.

5,321

5,322

5,323

5,324

–50

5–5

0

5

Radius, in.

Dept

h, ft

5,325

5,320


operators are turning to geomechanics experts atthe Schlumberger Center of Excellence for PorePressure Prediction and Wellbore StabilityAnalysis. Located in Houston, the geomechanicsexperts in this group have a global reach, andsupport operators around the world. Thisinterdisci plinary team is actively involved inhelping clients mitigate risk in drilling,completing and producing wells in difficultgeomechanical environments, such as deepwaterexploration, subsalt drilling, unconventional gasand unconsolidated reservoirs.

Beyond the WellboreGeomechanical influences can extend past theborehole, into the reservoir and beyond—thoughtheir extent may not be recognized until areservoir is produced. The pressure sink createdby a well to induce production will result in lowerwellbore pressures than the pore pressure of thesurrounding formation, and this difference canincrease the risk of rock failure.13

With the withdrawal of reservoir fluids duringproduction, the overburden load borne by porefluids must be transferred to the rock frameworksurrounding the pore space. Resulting changes inpore pressure will prompt adjustments in totalstress and effective stress. Within the rock,increased loading will cause varying degrees ofdeformation or failure, evidenced by grain sliding

and rotation, plastic deformation, cementbreakage at grain contacts, or activation ofexisting fractures.14

On a larger scale, production-induced stresschanges on the rock framework can lead to porecollapse and compaction of the reservoir.15

(Compaction is not always a problem, however—compaction drive has helped to pressurize oil insome reservoirs, thereby increasing productionrates and improving ultimate recovery.)16 As aresult, operators have had to contend withsurface subsidence problems, deformation orshearing of wellbore tubulars and buckling ofcompletion components. Other effects range from reduction of porosity and permeability tofault reactivation, formation fracturing, sandproduction or loss of reservoir seal.

The effects of geomechanics are especiallypronounced in gas storage operations, where thecyclic process of injecting and withdrawing gas toor from a reservoir provokes changes in fluidpressures within reservoir pore spaces. Thesepressures cushion the stresses acting on the rockmass, but the pressures increase or decreasewith injection and withdrawal. The loads actingon the rock matrix thereby decrease andincrease in response to these cycles. Althoughtotal overburden stress may remain constantthroughout these cycles, the total horizontalstresses acting throughout the reservoir can vary

with pressure, generally decreasing as the gas iswithdrawn. If induced stresses exceed the elasticlimits of the rock, porosity and permeability maybe permanently reduced, along with reductionsin overall storage capacity. Furthermore, as thesurrounding rock adjusts to the isostaticimbalance caused by pressure cycling and stresschanges, nearby faults may be reactivated.17

Production-induced changes can also affectthe rock beyond the productive areas of areservoir. Even in producing formations, reser -voir attributes such as porosity and permeabilitycan vary, giving rise to nonuniform drainage anddepletion. As a reservoir is produced, the rockmay eventually compact, leaving surrounding,undrained areas of the formation to compensatefor changes in pressure and displacement of theadjacent rock. Above the productive formation,compaction will lead to changes in the over -burden, as described later in this article.

Changes in stress imposed on a producinghorizon can put the rock out of equilibrium withits surroundings. The result is a correspondingtransfer of stress between the depletingreservoir or injection interval and the rockimmediately surrounding the reservoir. Ensuingrock deformations may compromise theintegrity of existing completions within thereservoir and overburden (above left). Thesignificance of production-induced stresschanges and their potential to adverselyinfluence field operations, production andeconomics will depend on mechanicalproperties of the rocks, natural fractures andfaults.18 To understand and anticipate suchchanges in the wellbore and beyond, operatorsare increasingly turning to advanced geome -chanical testing and modeling techniques.

42 Oilfield Review

13. Cook J, Fuller J and Marsden JR: “GeomechanicsChallenges in Gas Storage and Production,” presentedat the United Nations Economic and Social Council:Economic Commission for Europe: Working Party onGas: Proceedings of 3rd Workshop on Geodynamic and Environmental Safety in the Development, Storage and Transport of Gas, St. Petersburg, Russia,June 27–29, 2001.

14. Sayers CM and Schutjens PMTM: “An Introduction toReservoir Geomechanics,” The Leading Edge 26, no. 5(May 2007): 597–601.

15. Doornhof et al, reference 5. Sayers C, den Boer L, Lee D, Hooyman P andLawrence R: “Predicting Reservoir Compaction andCasing Deformation in Deepwater Turbidites Using a 3DMechanical Earth Model,” paper SPE 103926, presentedat the First International Oil Conference and Exhibition,Cancun, Mexico, August 31–September 2, 2006.

16. Andersen MA: Petroleum Research in North Sea Chalk,Joint Chalk Research Monograph, RF-RogalandResearch, Stavanger, 1995.

17. Cook et al, reference 13.18. Marsden R: “Geomechanics for Reservoir

Management,” in Sonatrach-Schlumberger WellEvaluation Conference – Algeria 2007. Houston:Schlumberger (2007): 4.86–4.91.

> Stress changes induced by production. As a field depletes, the magnitudeof stresses may alter drastically. Under such conditions, a completion orperforation originally oriented in the most stable direction at the onset ofproduction may subsequently become unstable and fail as productionproceeds. Here, the horizontal perforation will permit the greatest safedrawdown (blue curve) and solids-free production. However, as the fielddepletes and stresses change, this previously stable perforation willcollapse and the vertical perforation will assume a greater role inproduction, though safe drawdown pressure has decreased (red curve).(Adapted from Marsden, reference 18.)

Safedrawdown

Depletion

Wel

l pre

ssur

e, p

si

Reservoir pressure, psi

0 3,000 6,000 9,000 12,000 15,000

15,000

12,000

9,000

6,000

3,000

0


Autumn 2007 43

Measuring Ground TruthDespite years of geomechanical analysis, manyE&P companies continue to experience drilling-or production-induced problems. However, thefield of geomechanics involves much more thananalysis of stress. Though changing stress fieldscan wreak havoc on drilling and productionplans, the orientation or magnitude of stressesand strains have little significance withoutframing such measurements in the context of therock itself. And rocks are highly variable. Otherproblems are caused, in part, by oversimplifiedcharacterization of rock behavior, and by limitedmodeling and analysis capabilities compoundedby a lack of comprehensive rock property data.

These issues are being addressed by theTerraTek Geomechanics Laboratory Center ofExcellence in Salt Lake City, Utah. TerraTek, Inc.was acquired by Schlumberger in July 2006 (see“Geomechanics Laboratory: Testing UnderExtreme Conditions,” page 44). The modernhigh-pressure testing systems and techniquesdeveloped at the TerraTek facility evolved froman effort to characterize and predict groundmotion and crater development in response tonuclear tests. Evaluation of these tests could notbe performed without rock property measure -ments obtained under high pressure. Measuringthese properties was very difficult, and spawned a number of technical breakthroughs by TerraTek.

Highly accurate load-deformation measure -ments were essential, requiring measurementsinside test vessels under extreme pressures.TerraTek scientists conducted research tomeasure rock properties to pressures of150,000 psi [1,034 MPa]. The TerraTek high-pressure rock property data enabled analysis ofthe magnitude of ground motions caused by anuclear event.

TerraTek researchers carried out tens ofthousands of tests on rocks under high pressure.Their testing capabilities were subsequentlyapplied to other geomechanics investigations,including geother mal energy recovery, coalmining, deep geologic nuclear waste storage,underground energy storage, as well as oil andgas recovery. Today, the TerraTek GeomechanicsLaboratory Center of Excellence regularlyconducts rock tests for deep wells, achievingpressures of 30,000 psi [207 MPa], or to higherpressures of 50,000 to 60,000 psi [345 to 414 MPa]when required for drilling-rock destruction orperforation analysis. In addition to high-pressuregeomechanics testing capabilities, the TerraTekfacility conducts large-scale drilling andcompletions performance testing.

Specialized geomechanics laboratory testingprovides crucial data for wellbore and comple -tion design and for reservoir management. Thiswas not always the case. Traditional engineeringanalysis of reservoir potential and productivitytended to overlook the heterogeneity of reservoirrock. Although heterogeneity may have beencaptured in mud logs and core photographs, orinferred from logs of various petrophysicalproperties, these characteristics were notreflected in simplified homogeneous systemscreated for reservoir and geomechanical models.

Properties related to reservoir rockmechanics were often characterized as uniformthroughout all locations and for all orientationswithin a particular geological unit. This approachinevitably led to underestimations of the role ofmaterial properties in geomechanics. Theindustry, however, is coming to realize that therock matters, and that its varying propertiescannot be ignored in geomechanical analysis.

Further complicating the evaluation processis the fact that each stage of reservoir analysis—from predrill geological studies, throughexploration, to reservoir modeling andproduction—tends to be evaluated in isolation,

and without a common reference for scale. Untilrecently, there has been no framework to makethe process consistent for every stage. However,the development of continuous property profilingand multidimensional cluster analysis of welllogs now provides a uniform scale of referencefor incorporating heterogeneity during allaspects of reservoir analysis and evaluation.

Continuous Profiling—Scratch testing,known formally as continuous profiling ofunconfined compressive strength, provides aquantitative means of evaluating variability instrength, texture and composition of coresamples. By association, this variability may berelated to other rock properties. Scratch testinghas become critical in correctly defining faciesand heterogeneities that would be difficult orimpossible to observe by geologic description orlog characteristics alone. Digital photographs ofthe core, in conjunction with scratch testing,allow visualization of textural heterogeneity andassociated strength heterogeneity (below).

When continuous-strength profiling iscombined with cluster analysis of well logs, it

> Overlay of a core photograph and scratch test results. A scratch test uses a sharp point that is pulledalong the core with a fixed force to press it into the core’s surface. The depth of the scratch, as anindicator of rock strength (red curves), can be correlated to mechanical properties of the rock. Coredintervals exhibiting visually similar properties (same shades of gray, points A and A’) may have differentstrengths, while other intervals exhibiting different visual properties (lighter and darker shades of gray,points B and B’) have equal strengths. Variability in mechanical strength along the length of the core ishigh, ranging from 8,000 psi to 23,000 psi [55 to 159 MPa] within just 8 contiguous feet [3.6 m] of core.

0 ft 0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

ft 1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

ft

AA'

B B'

0 psi

50,000 psi

0 psi

50,000 psi

0 psi

50,000 psi

0 psi

50,000 psi

(continued on page 48)


44 Oilfield Review

The TerraTek facility in Salt Lake City, desig-nated as the Schlumberger GeomechanicsLaboratory Center of Excellence, investigatesthe impact of geomechanics on a wide rangeof exploration and production applications.The range of applications also provides insightinto the kinds of problems that operatorsmust try to circumvent: • Well construction and completion: evaluate

wellbore stability and the potential for sandproduction and perforation collapse; analyzemultilateral junctions and evaluate stabilityof conventional and expandable liners.

• Completion and stimulation design: deter-mine optimal completion alternatives basedon rock mechanical and physical properties;investigate options for delayed gravel pack-ing and oriented perforating; optimizestimulation treatment design.

• Long-term production behavior: investigatestress regimes contributing to reservoir compaction during production; predict surface subsidence and subsequent loss ofreservoir permeability; analyze fines gener-ated during the compaction process, alongwith associated skin damage; evaluatepotential for casing collapse.

• Overburden rocks: test for compatibilitybetween drilling fluids and shales; optimizeselection of drilling fluids; evaluate poten-tial for delayed shale failure caused bymud-shale interactions; analyze thermaleffects that cause delayed shale failure.

• Exploration and frontier drilling operations:develop field and laboratory correlations forpredicting mechanical properties and in-situstress prior to and concurrent withexploratory drilling activity.Testing is conducted in different specialized

laboratories, depending on available test mate-rial, client specifications and research efforts.Many large-scale tests are carried out in thecompletions laboratory. One of the moreprominent features of this facility is its large-block polyaxial stress frame. The stress frameprovides a controlled environment for monitor-ing rock responses during pseudostatic testing.

In this setting, researchers can measure defor-mation parameters while simultaneouslymeasuring dynamic responses of rock samplesto different load rates and magnitudes. Thelarge-block stress frame can be configured tosimulate a variety of downhole pressures andconditions. Large-block testing applicationsrange from wellbore stability analysis to evalu-ating sanding potential, liner and screenloading, perforating effectiveness and hydraulicfracturing simulations.

Located inside a pit, the exterior of thestress frame is formed by a series of steelrings. These rings are stacked to encase an

internal chamber that can accommodateblocks of rock measuring up to 30 x 30 x 36 in.[76 x 76 x 91 cm]. The chamber is sealed withsteel platens that are bolted over 12 large tierods (above).

Pairs of bladder-like devices, called flatjacks,are placed on opposite sides of the sample toapply independent triaxial loading in each ofthe three principal stress directions. The threepairs of flatjacks are internally pressurized,with one surface of the flatjack reacting againstthe face of the rock, and the other surfacereacting against the wall of the internal cham-ber of the stress frame, or its platen.

Geomechanics Laboratory: Testing Under Extreme Conditions

> Large-block polyaxial stress frame for simulating downhole conditions.Here, a worker lowers a steel platen while preparing to seal the test chamber.


Autumn 2007 45

A maximum stress of 8,000 psi [55 MPa] can beapplied in all three directions, with a maximumdifference of 2,000 psi [13.8 MPa] between thetwo horizontal stresses. Each stress can be con-trolled independently.

The stress frame also has the capability tocontrol pore pressure within a sample. In suchtests, the rock sample is encased in a thinsteel canister. Thick rubber sheets are placedat the top and bottom surfaces of the rock toact as pore-pressure fluid seals. A porousproppant pack placed around the block estab-lishes a constant pressure boundarycondition. Custom software controls each ofthe three principal stresses, along with porepressure and wellbore pressure. The softwarecan be programmed to keep a constant effec-tive stress on the sample block at all times.

Some experiments require a simulated permeable zone bounded above and below byimpermeable formations. In these kinds of

tests, a servo-controlled injector is used tosupply the fluid at either a constant rate or aconstant pressure. The injected fluids canrange from brine to drilling mud to variouscompletion fluids. The injection can simulatea scaled or actual-sized wellbore.

For smaller samples, a medium-sizedpolyaxial stress frame is used (above left).This device is often used for studying acidfracturing and other stimulation techniques,providing a wide range of testing capabilities.

Another unique testing facility is the rockmechanics laboratory, where 14 stress framesare used to test cylindrical samples with diam-

eters ranging from 0.5 inches [12.7 mm] to6 inches [152.4 mm]. Testing on a smallerscale can also provide valuable insights intorock characteristics.1 A special triaxial testframe has been designed to measure rockstrain as well as its effects on seismic veloci-ties (above right). Ultrasonic velocities,

1. Laboratory capabilities include an extensive variety of tests—unconfined compression, uniaxial-straincompression, triaxial compression, multistage triaxialcompression, controlled constant stress path, thick-walled cylinder (with and without radial fluid flow andmeasurements of produced sand), and tensile strengthtests, as well as testing with concurrent ultrasonicvelocity and acoustic emissions measurements—along with many customized test programs andresearch efforts.

> Polyaxial stress frame. This device canaccommodate rock samples measuring up to12 x 12 x 16 in. [30 x 30 x 41 cm].

> Instrumented sample for triaxial testing. This test-frame assembly is used to measure radialand axial strains, along with compressional and shear wave velocities. In this configuration, bothpseudostatic and dynamic elastic properties are determined concurrently under simulated in-situstress conditions. Here, a core consisting of alternating light and dark layers of siltstone andmudstone is subjected to ultrasonic pulses to test seismic responses in the rock. The sample issealed with a clear polyurethane jacket that prevents fluid communication between the confiningfluid pressure and the pore pressure. These test frames can also perform uniaxial straincompaction testing, thick-walled cylinder testing and other specialized stress paths up totemperatures of 200°C [392°F]. Axial force up to 1.5 x 106 lbf [6.7 MN] can be applied to samplesup to 6 in. [15 cm] in diameter. Confining pressure and pore pressure are monitored withconventional pressure transducers with pressure limits of 30,000 psi [207 MPa]. Another systemin this laboratory can attain 60,000 psi [414 MPa].


46 Oilfield Review

obtained in combination with deformationmeasurements of axial and radial strain, pro-vide information on static and dynamicmechanical properties that can be correlatedto well-log data.

The triaxial test frame holds a core samplebetween polished, hardened-steel end-caps. The sample, measuring 1 in. [2.5 cm] in diame-ter by 2 in. [5 cm] in length, is jacketed by animpermeable membrane. Axial and radial can-tilever-beam sets are mounted to the sample tomeasure displacements when the sample issubjected to stress and pressure. The axialstrain cantilever set is attached to the upperend-cap and measures axial displacementthrough deflection on the base cone attachedto the bottom end-cap. The radial-strain can-tilever set consists of a ring with fourstrain-gauge arms, which measure radial dis-placement at four points, forming twoperpendicular directions at the midpoint of thesample. The bottom end-cap rests on an inter-nal load cell, and the axial stress is calculatedfrom measurements of force on the internalload cell. During testing, data are corrected forelastic distortion of the end-caps and forstrains associated with the jacketing material.

The end-caps also contain ultrasonic trans-ducers. Ultrasonic velocity measurements areperformed with piezoelectric transducers thattransform electrical pulses into mechanicalpulses and vice-versa. Compressional andshear pulses are generated by a pulse genera-tor that applies a high-voltage, short-durationelectrical pulse at an ultrasonic frequency toone of the piezoelectric transducers. Thispulse is transmitted through the rock samplein the form of an elastic wave. The receivingtransducer at the opposite end of the rocksample transforms this elastic wave into anelectric signal, which is captured on a digitaloscilloscope. The P-wave and S-wave velocitiesare calculated on the basis of the time requiredfor the compressional or shear pulses to travelthrough the specimen.

The instrumented test sample is nextplaced inside a pressure vessel. The pressurevessel is then filled with either mineral spir-its or oil to apply confining pressure. Axialstress, axial strain, radial strain and confiningpressure are all measured and controlled

during each test. Depending on testing objec-tives, these tests may be performed with porefluids drained to atmospheric pressure, orwith pore fluids undrained. Temperatures canalso be increased to better approximateactual in-situ conditions.

The triaxial test frame permits measure-ments to be taken at different orientationswith respect to bedding planes. Using thesemeasurements, the failure envelope of therock sample can be defined as a function ofstress orientation to bedding; in addition,anisotropic properties of the rock can bedefined. This information is essential for pre-dicting wellbore stability, evaluating in-situstress and designing hydraulic fracture pro-grams for strongly anisotropic formationssuch as those found in unconventional tightgas shales.

Ultrasonic velocities, obtained in combina-tion with deformation measurements of axialand radial strain, provide information onstatic and dynamic mechanical propertiesthat can be correlated to well-log data. Ultra-sonic wave velocities in sandstones, particu-larly those that are poorly consolidated, arestrongly dependent on stress; thus, stresschanges can be calibrated to seismic velocitymeasurements. Other, more consolidatedrocks, such as tight sands and tight shales,exhibit an entirely different behavior. Wavevelocities in these rocks are virtually inde-pendent of stress, so changes in measuredseismic velocities can be attributed to otherphenomena such as anisotropy.

Early knowledge of rock behavior was basedon testing of homogeneous and isotropicmaterials; early models reflected this simplic-ity. New opportunities, such as unconventionalhydrocarbon plays, are emerging, and callattention to the true nature of the rocks inwhich they are based. Platforms such as thetriaxial test frame provide data that are fun-damental for developing new models to honorthe heterogeneous, anisotropic nature of com-plex formations.

The TerraTek facility is also called upon totest new drilling, completion and stimulationtechnologies, including evaluation of drillingfluids and bits at high pressures. Althoughcapabilities exist for measuring individual rock

properties or fluid properties at extreme tem-peratures and pressures, determining themanner in which complex rock cutting andbreakage mechanisms interact in the presenceof drilling fluids at great depth is much moredifficult. To accommodate large-scale geome-chanics testing, the drilling laboratory isequipped with a wellbore simulator capable ofreproducing pressure conditions at reservoirdepth while also accommodating the flow ratestypically required to drill in extreme environ-ments (above).

> TerraTek wellbore simulator. The full-scaledrilling rig and wellbore simulator can beconfigured to test the performance, wear,deviation and dynamics of full-size drill bits inoverbalanced or underbalanced conditionsand at simulated depths. A triplex mud pump,equipped with a special high-pressure fluidmanifold, can produce wellbore pressures upto 11,000 psi [75.8 MPa] to simulate high-pressure drilling conditions. Effects of variousfluids on drilling performance, bit balling,formation damage, coring and core invasionare also investigated here.


Autumn 2007 47

The TerraTek wellbore simulator was cen-tral to a recent high-pressure drilling studysponsored by the US Department of Energy(DOE) joint industry program, called DeepTrek. The facility was contracted to providefull-scale laboratory tests of drill bits and pro-totype drilling fluids at 10,000-psi [68.9-MPa]borehole pressure—substantially higher pres-sures than any previously studied. Resultsfrom these tests may influence the economicsof deep drilling.

The study demonstrated that drilling ratesof penetration (ROPs) can be increased indeep-well applications using advanced bit anddrilling fluid designs. Although previous stud-ies have shown that ROP usually falls withincreasing borehole pressure, these earlierstudies did not account for certain mecha-nisms that affect ROP at great depth, such astype of drilling fluid, weighting material andspurt loss.2

Another common wellbore stability probleminvolves borehole breakouts. Although break-outs often occur during drilling, they can alsoaffect the completion process. In one break-out investigation, TerraTek engineers drilledan 81⁄2-inch [21.6-cm] borehole in a large sandstone core. The core was subjected toincreasing rates of confining pressure in thelaboratory. The resulting borehole breakoutwas similar to that produced in actual well-bores when drilling fluid weights are too low(above left).

The sample was subsequently used for anexpandable sand screen (ESS) mechanicalintegrity test. The screen and basepipe assem-bly was compliantly expanded to the boreholewall and partially into the breakout zone.Results from this test showed how far thescreen could be expanded into the boreholebreakout, in addition to determining the col-lapse load resistance of the ESS product.

Other problems that adversely impactdrilling performance, such as vibration orborehole spiraling, are identified throughexamination of drilling patterns (left).Through the aid of the borehole simulator,researchers have an opportunity to closelystudy bottomhole patterns that would other-wise not be accessible.

2. Spurt loss is an instantaneous loss of a volume of the liquid component of drilling fluid as it passesthrough the borehole wall prior to deposition of competent filtercake.For more on ROP testing: Judzis A, Bland R, Curry D,Black A, Robertson H, Meiners M and Grant T:

“Optimization of Deep Drilling Performance;Benchmark Testing Drives ROP Improvements for Bits and Drilling Fluids,” paper SPE/IADC 105885,presented at the SPE/IADC Drilling Conference,Amsterdam, February 20–22, 2007.

> Breakout simulation. With no drilling mud used to drill thissandstone subjected to increasing confining pressure, thissimulated wellbore progressively broke down, producing a classicborehole breakout pattern.

> Bottomhole drillbit patterns. Bottomhole impressions trackperformance of a bit as it drills a borehole through high-strengthsandstone. In this case, a polycrystalline diamond compact bit wasdrilling with a 16-lbm/gal (ppg) [1.9-g/cm3] oil-base mud at 10,000-psi[68.9-MPa] wellbore pressure. The patterns on the bottom weresubsequently studied to determine how various drilling conditionsaffected drilling performance. As depth of the rings decreases, sodoes the cutting efficiency of the bit, and hence the ROPdecreases. With different drilling fluids, the patterns sometimesdisappear altogether.


provides fundamental relationships for upscalingor downscaling, and so is a powerful tool for core-log integration.

Cluster Analysis—Cluster analysis defineslog-scale heterogeneity, based on multidimen -sional analysis of log responses (above left). Thistechnique uses detailed algorithms to distinguishsimilar and dissimilar patterns of log responses.Because this technique interprets the combined

effect on all measurements, it is able torecognize small but consistent variations incombined log responses. As applied to heteroge -neous distributions of material properties,cluster analysis also provides a relevant scale formanipulating property variability in subsequentevaluation steps throughout a project.

Cluster Tagging—The application of clusteranalysis can be extended to multiple wells,providing comparisons between the cored, orreference, well and other wells in a field. Detailsobtained through cluster analysis of one well canbe used to recognize similar traits in adjacentwells through a process known as cluster tagging.

Cluster tagging begins with log-responseclusters defined over discrete cored intervals in areference well, then compares these clusterswith log responses from a noncored well. Usingdefinitions established from core-log responses

in the reference well, the technique assignsclusters to logs from the noncored well and thenoutputs an error curve to help evaluatecompliance between two correlative zones.Clusters showing poor compliance, where errorexceeds 40%, indicate a log response that is notrepresented in the defined clusters, and thus anew facies. These clusters are candidates fordetailed core sampling to provide new clusterdefinitions and further characterize the range offacies in a prospect (above right).

Cluster analysis is also used for optimalselection of core samples. In reservoir studies,both the strongest and weakest core samplesmeasured by continuous profiling must be testedin proportion to their relative abundance in areservoir. Improper sampling in heterogeneousor thinly interbedded formation cores can resultin biased representation of the formation.Cluster analysis can help operators tie log

48 Oilfield Review

> Cluster analysis of well logs. A multidimensionalstatistical algorithm is applied to the well-logmeasurements to identify similar and dissimilarcombined log responses, enabling users toidentify rock units with similar and dissimilarmaterial properties. The output is displayed as acolor-coded representation of clusters for visualinterpretation of rock units with distinct propertiesalong the interval of interest (Track 4).

Caliperin.5 15

Resistivityohm.m0 1,000

Neutron PorosityCluster Tag

vol/vol0.45 –0.15

Bulk Densityg/cm32 3

PEbarn/e-1 6

Gamma RaygAPI0 150

> Cluster tagging between two wells. Color-coding of log responses from each well, combined withanalysis of compliance in the Error track, is useful in identifying changes in thickness and location ofpreviously defined cluster units between wells. Here, the red-blue-yellow sequences are significantlyhigher and thicker in Well 1 than in Well 2. Three excursions above 40% error (red line) indicatecandidate zones for further sampling to better describe the range of facies encountered.

Y,500

Y,400

Y,300

Y,200

Y,100

Depth, ft

Y,000

X,900

X,800

X,700

X,600

X,500Depth, ft

Y,500

0 50Percent

100

Y,400

Y,300

Y,200

Y,100

Y,000

X,900

X,800

X,700

X,600

X,500

Y,500

Y,400

Y,300

Y,200

Y,100

Y,000

X,900

X,800

X,700

X,600

X,500Depth, ft

Well 1 Well 2 Error


Autumn 2007 49

properties to core properties throughout thereservoir, and thereby recognize which parts of acore merit additional plug-sample analysis(below). With cluster-analysis measurements oflog-scale heterogeneity and core-scale hetero -geneity measurements obtained through scratchtesting, the operator can determine the locationand number of samples required to adequatelycharacterize the core.

Cluster-Level Property Predictions—Sincemodels are traditionally built around thestructure and stratigraphic layout of a basin, thediscontinuous and heterogeneous distribution ofreservoir and nonreservoir lithological unitswithin a single stratigraphic section is oftenpoorly represented across the basin. Cluster

analysis identifies units by their materialproperties and maps their distribution along thelength of a well. By relating laboratory measure -ments of these units to their combined logresponses, core-log relationships are developedfor each cluster. Since the method is unaffectedby variability in thickness or stacking arrange -ments of the various cluster units, it allowsprediction of properties along the length of thelogged section in a well.

Multiwell Analysis—For basin-wide analysis,cluster tags of multiple wells are tied to a singlereference model containing definitions ofmaterial properties across the basin. The resultscan be used for 3D visualization of lateralvariability in reservoir and nonreservoir units.

Cluster tag analysis was instrumental ingenerating a regional study for a client who waspursuing an unconventional gas play. The goalwas to model the vertical and lateral discon -tinuity of principal reservoir units in a tightgas-shale reservoir. These reservoirs are highlyheterogeneous both vertically and laterally, withlocalized diagenetic alterations that create greatvariability in material properties. As a result,reservoir and mechanical properties changesignificantly from location to location betweenwells, and production performance often varies,even between wells drilled in close proximity toeach other.

The client ordered a study to understand thevariability in permeability, gas-filled porosity and

> Using rock heterogeneity to select laboratory samples. Log-scale heterogeneity, indicated by cluster colors (left), is compared against core-scaleheterogeneity data obtained through scratch testing (red curves) superimposed onto core photographs (middle). In the log-scale heterogeneity plot, color isused to differentiate between zones of similar or dissimilar material properties as a function of unconfined compressive strength measurements. Here theyellow clusters are the weakest units and brown clusters are strongest. Progressing from region 1 (yellow cluster), region 2 (yellow cluster transitioning todark blue), region 3 (dark blue transitioning to brown), and region 4 (brown cluster), the rock strength varies by more than 400%. Core photographs (middle)show a corresponding transition in unconfined compressive strength from 10,000 psi [68.9 MPa] in the argillaceous mudstone (core section 1) to 40,000 psi[275.8 MPa] in the basal carbonate (core section 4) within this 40-ft [12-m] interval. Sample plugs (right) are taken from the whole core for detailed analysisand testing. This methodology helps operators ensure that their 2-in. sample plugs account for the variability present in the whole core.

Core-Scale Heterogeneity Sample-Scale

Heterogeneity

Log-Scale

Heterogeneity 10 k

50 k

10 k

50 k

10 k

50 k

0 ft

1 ft

2 ft

2 in

.

0 ft

1 ft

2 ft

0 ft

10 k 40 k

1 ft

2 ft

0 ft

1 ft

2 ft

10 k

50 k

1

1 2

3 4

2

3

4

1

4

2

3


total organic content as they relate to reservoirquality. It was also important to understand thevariability in the conditions of hydraulic fracturecontainment along the various wells containingunits with best reservoir quality. For optimal wellproductivity, reservoir quality must be coupledwith completion quality. In this field, reservoirquality alone, without successful fracturing andfracture-height containment, would result inpoor well productivity. By mapping locationsthroughout the field where both conditions ofreservoir quality and completion quality existsimultaneously, the client could identify sweetspots in the reservoir (above left). The results ofthis field study would also help improvevisualization of production distribution acrossthe basin.

TerraTek geoscientists used cluster analysisand cluster tagging to evaluate the field.Understanding the vertical stacking patterns ofcluster units on a well helped the client definethe location and thickness of clusters with thebest reservoir-quality properties.

Once these parameters were defined, theclient could select the best geometry ofhorizontal wells and the best locations forperforating. Understanding the properties ofcluster units immediately above and below thebest reservoir-quality units also helped identifymechanical properties and conditions forhydraulic fracture containment.

Modeling Geomechanical PropertiesThe interaction between geology, wellboreorientation and stress changes caused by drillingor production is a complex 3D process. Thisinteraction continually changes over time,adding yet another dimension of complexity.Over the life of any productive field, innumerableevents take place that alter the geomechanicalframework between the reservoir and thesurface. Exploration wells are drilled and tested;additional wells are drilled and produced; somemay be turned into injectors, some are workedover, while others are plugged and abandoned.Each activity causes changes in stress—someephemeral, others more enduring. And thesechanges can be costly, with potential to affectformation integrity, porosity and permeability;reservoir compaction and subsidence; and welland completion integrity.

The movement to understand such changeswas spurred in part by recognition thatsubsidence in certain fields was directly relatedto production. Basic mathematical models weredeveloped by the early 1950s to understand andpredict subsidence in Wilmington field,California.19 Later, subsidence of the North SeaEkofisk field, discovered in the early 1980s,prompted development of more comprehensivecomputer models, based on finite-elementanalysis. These models linked hydrocarbonproduction to changes in reservoir properties anddeformation and, in turn, to seabed movementand faulting in the overburden.

E&P companies became interested inlearning how stress evolves as reservoirs become

50 Oilfield Review

19. McCann GD and Wilts CH: “A Mathematical Analysis ofthe Subsidence in the Long Beach-San Pedro Area,”technical report, California Institute of Technology,Pasadena (November 1951), in Geertsma, reference 2.

20. Ali AHA, Brown T, Delgado R, Lee D, Plumb D,Smirnov N, Marsden R, Prado-Velarde E, Ramsey L,Spooner D, Stone T and Stouffer T: “Watching RocksChange—Mechanical Earth Modeling,” OilfieldReview 15, no. 2 (Summer 2003): 22–39.

> Basin-wide multiwell cluster analysis. This presentation uses Petrel seismic-to-simulation softwareto help operators visualize the cluster-analysis results and track reservoir quality throughout the field.Different cluster units are associated with distinct reservoir qualities. They are also associated withdifferent values of fracture containment potential. Once the reservoir quality and fracture containmentpotential are identified in detail by laboratory testing, they can be tracked laterally across the basin.Surfaces identifying the intervals of best reservoir quality have been delineated. Cluster analysis in this case identifies the heterogeneity inherent in any of these units that otherwise might be considered homogeneous.

1 2 3 4 5 6 7 8 9 10 11 12

Cluster tag

> Array of input parameters for a mechanicalearth model.

• Regional tectonic framework• Structure depth maps• Lithostratigraphic column• Regional compaction trends• Basin analysis• Earthquake fault-plane solutions• Tiltmeter surveys• Core tests and descriptions – Rock composition and texture – Core-log integration – Heterogeneity and anisotropy – Petrophysical and mechanical characterization

Geologic Data

• 3D seismic cube• 2D seismic profiles• Tomographic velocity• Vertical seismic profiles and checkshot data• P-wave velocity profiles

Seismic Data

p

• Daily drilling reports• End of well reports• Mud weight profile• Leakoff tests, extended leakoff tests, formation integrity tests, minifrac tests• Directional surveys• Mud logs

Drilling Data

• Laboratory measurements on cores• In-situ stress measurements from hydro- fracturing tests• Observed breakouts and stress-induced features• Field and production observations

Calibration Data

• Wireline and LWD logs – Gamma ray, resistivity, density, sonic, caliper – Acoustic scanning tool – Borehole imaging• Well test and production pressure measurements – Formation tests and drillstem tests

Formation Evaluation Data


Autumn 2007 51

depleted. If stress changes could be modeledover the life of a field, operators could predictproblems during the life of a well or anticipatethe need for infill drilling. With a steady growthof computational capacity, geomechanicsprograms acquired increasingly sophisticatedmodeling capabilities. Rock mechanical modelsdeveloped to analyze stress changes in reservoirsincluded the VISAGE stress analysis simulator.This advanced geomechanics modeling systememerged from waterflood directionality studiesin the North Sea and elsewhere.

Developed in 1993 by V.I.P.S. (VectorInternational Processing Systems) of Bracknell,England, VISAGE geomechanics software solvesequations that relate rock stress and porepressure to deformation and reservoir proper ties.By integrating geomechanics and rock mechanicswith reservoir engineering, V.I.P.S. developed theworld’s first coupled geome chanics stress-dependent reservoir simulator. With acquisitionof V.I.P.S. by Schlumberger in April 2007, theBracknell facility was designated as the ReservoirGeomechanics Center of Excellence.

Finite-element modeling is widely used forstress analysis in both conventional engineeringand geomechanics. Finite-difference modeling isused to analyze fluid flow. The advantage of theVISAGE simulator is its capability to describeand simulate the coupled nature of geomechanicalstresses and fluid flow as they change over timeby linking these two analyses. This capability iskey to development of 3D and time-sequenced 4Dmechanical earth models.

Unlike reservoir production models,mechanical earth models (MEMs) must take intoaccount not only the reservoir, but also theoverburden, seabed, the underburden, or rockbeneath the reservoir, and sideburden, oradjacent rock, which often provides stressboundary conditions.20 The MEMs are usuallymuch larger than ordinary reservoir models. Assuch, they have substantial data requirementsthat may be difficult to satisfy.

Complex rock behavior, varying rockproperties and large-scale simulations requirebetter software and better data, especially with

regard to cores. Basic models of the past enabledthe industry to opt for simplified assumptions,using homogeneous formation propertiesthroughout their models. Today’s sophisticatednumerical simulators inevitably dictate a widerarray of data. The MEM is built to honor this widearray of data (previous page, top right).

A geomechanical simulation might begin withconstruction of a 3D structural model. Next, themodel is populated with mechanical propertiesof each formation and fault. The properties arederived from seismic data, logs, cores, geostatis -tical projections and inversion of breakout anddrilling data for individual wells. Boundaryconditions, simulating the expected stressprofiles at the sides of the model, are then added.This populated model is imported into theVISAGE system to calculate the evolution ofstresses throughout the model (above).

The driving mechanism of the modeling ismainly pressure changes induced by fluidextraction from the reservoir, or by injection into

>Workflow for 4D coupled reservoir geomechanics modeling. Formation and structural data form the framework for the initial reservoir model, thencharacteristics from surrounding rock bodies are added. Stress and strain are modeled throughout the reservoir and adjacent rock to understand changesover time.

Importing from ECLIPSEor Petrel software, or both

Importing fault surfaces Embedding in overburden,underburden and sideburden

Population with properties andassign behavioral models

Initialization and coupledsimulation (parallelization)

Data and results utilizedin engineering designs

and planning

VISAGEsimulation

ECLIPSEsimulation

Δp, ΔT

Δkij, ΔVpore


the reservoir. Fluid flow is modeled using areservoir simulator, such as the ECLIPSEreservoir simulation package. By accounting forthese pressure changes in the stress calculationsusing VISAGE software, it is possible to accuratelypredict subsurface deformations and stresschanges, and evaluate their influence on materialproperties such as permeability and porosity.

The resulting model can be used as a sourceof stress data for several key stages:• well planning—wellbore stability and optimum

drilling azimuth• well completions—sand management• formation stimulation—hydraulic fracture

orientation• field management—pressure maintenance

and injection

• well integrity—well design to accommodatecompac tion and subsidence as the well is produced.

This coupled approach was recently used in aNorth Sea field study. The South Arne field,located in the Danish sector of the North Sea,produces from the Maastrichtian Tor and DanianEkofisk chalk formations. Oil production fromthe low-permeability chalk is driven both bywater injection and by compaction of the chalk.

In 2006, a field study of the South Arne fieldwas conducted to quantify the effects ofproduction from 1999 to 2005, and to assessoutcomes of a proposed development plan. Thefield study was carried out using a history-matched ECLIPSE model and the VISAGEgeomechanical simulator. The geomechanicalstudy comprised four phases.

The goal of the first phase was to enhance anexisting reservoir model by adding more rocklayers and structural detail. First, the reservoirmodel was extended up to the seafloor, adding 20new layers and eight horizons for optimaldescription of the overburden sequence. Tenlayers were added below the reservoir layer toserve as underburden, and eight cells were addedon each of the four vertical boundaries to serveas sideburden. Next, 45 faults and two differentfracture sets were incorporated into theembedded model. The mechanical propertieswere determined based on laboratory tests, corecalibration and literature reviews. A 1D stresscalibration was determined from density logintegration, leakoff tests and pore-pressuremodeling based on wireline log data.

The second phase sought to characterize thestress state prior to production operations. Aninitial effective stress state was computed, basedon properties determined in the first phase. Thestress-state computations accounted for contrastsin deformation and strength properties betweendifferent rock layers, and also considereddiscontinuity within the rock layers themselves(left). The computed initial stress state waschecked to verify agreement with field data andgeological features relating to stress orientations,stress magnitudes and fault orientation.

The goal of the third phase was to determinethe state of present-day stresses. The approachcalled for both flow and stress modeling, startingwith the change in pressure predicted by theECLIPSE reservoir simulator. The changes instress and strain induced by production andinjection operations were then assessed usingthe VISAGE geomechanical simulator. Thecomputed compaction at the top of the reservoirwas in good agreement with the estimated valuefrom 3D seismic inversion.

It was also important to assess the risk of wellfailure. The coupled simulations demonstratedthat pore collapse within the reservoir layerswould cause compaction and subsidence, and thatdifferential pore collapse could result in localizedwell failure (next page).

In the last phase, fluid-flow and stresssimulation was performed in which permeabilitychanged in accord with stress and strainchanges. After history-matching of productionand injection data, the geomechanical modelagreed with production history.

52 Oilfield Review

> Three-dimensional view of a reservoir. The uppermost horizon of an anticlinalreservoir is intersected by numerous faults (inclined planes of semitransparentpurple, red, green and blue). The axis of the anticline is aligned with the longaxis of this figure. Colors on the reservoir surface represent the computed stateof initial maximum principal stress acting on this horizon. In regions remote fromand unaffected by the presence of faults, the maximum principal stresses(green) correspond closely to the magnitude of the vertical or overburden stress,meaning that the principal stresses are near-horizontal and near-vertical. Theregions of reduced stress (blue) are the result of stress-arching in areas wherethe structural geometry and the stiffness of overburden layers create anincomplete transmission of overburden weight onto the underlying reservoir. Thehigh maximum stress concentrations (yellows and reds) near the faults coincidewith inclined principal stresses, causing the magnitudes of the maximumprincipal stresses to exceed lithostatic or overburden stresses generated bygravity and the weight of the overlying rock mass. The black box in the upperquadrant represents the area of study shown in the following figure (next page).

0 Maximum

Stress


Autumn 2007 53

Monitoring: Geomechanics and 4D Seismic DataOnce a field model is developed, it should beperiodically updated with data obtained throughmonitoring. A variety of techniques have beendevised for monitoring field-scale geomechanicaleffects. For example, global positioning systems,bathymetry and borehole tiltmeter surveys havebeen used to measure surface subsidence.Reservoir compaction can be detected by moni -toring casing collar movement, although thismethod is not precise. Microseismic tech niqueshave been used to detect regions of movementand rock failure during depletion, and areparticularly useful for identifying faultmovements and monitoring fracture creationduring injection and thermal recovery processes.21

Time-lapse, or 4D, seismic surveys are also beingused for geomechanical monitoring.22

Both seismic compressional and shear wavesare influenced by production-induced stresschanges inside and around a reservoir. Time-lapse seismic surveys, which predominantly usecompressional waves, have long been used tomonitor reservoir changes caused by production.Repeatedly surveying a reservoir over timeenables geophysicists to compare differences inseismic attributes, such as reflection amplitudesand traveltimes, between the initial baselinesurvey and subsequent monitor surveys. Thesedifferences are particularly useful in detectingmovements of gas/liquid contacts that occur asreservoirs are produced. In recent years, 4Dseismic techniques have also been used tomonitor production-induced changes in reservoirgeomechanical properties.

Among the differences between baseline andmonitor surveys, geophysicists sometimesobserved shifts in seismic traveltimes to aspecific horizon. Initially, these discrepancieswere attributed to logistical problems associatedwith repeating surveys over a reservoir: namely,the difficulty in placing seismic sources andreceivers in exactly the same position for everysurvey. The slightest mispositioning of sources orreceivers could lead to modified raypaths thattraveled through slightly different parts of thesubsurface, generating perturbations in observedtraveltimes. In the past, discrepancies in seismic

> Production-induced compaction. These figures correspond to the boxed area shown in the previous figure (page 52). Production-induced time shifts seenfrom the 4D seismic response (left) closely match the pattern of computed plastic strains obtained through coupled numerical simulation (right). Maximumcompaction (red) follows the NW trend of horizontal wellbores (dark blue lines) in the upper part of this figure. As expected, the area of greatest compactioncorresponds to that part of the reservoir experiencing the greatest production and consequently the greatest depletion. The computed maximum compactionof 1.45 m [4.76 ft] at the top of the reservoir was in good agreement with the estimated value of 1.4 m [4.59 ft] from 3D seismic inversion. The absence of 4Dseismic data (white zone) is caused by a gas cloud. Close agreement between the 4D seismic data and the numerical model reinforces confidence in modelresults over the area where seismic data were not available.

0 Maximum

Compaction

21. For more on microseismic applications: Bennett L, Le Calvez J, Sarver DR, Tanner K, Birk WS, Waters G,Drew J, Michaud G, Primiero P, Eisner L, Jones R,Leslie D, Williams MJ, Govenlock J, Klem RC andTezuko K: “The Source for Hydraulic FractureCharacterization,” Oilfield Review 17, no. 4 (Winter2005/2006): 42–57.

22. Doornhof et al, reference 5.


traveltimes were frequently attributed todifferences in acquisition geometry or toprocessing artifacts.

However, seismic acquisition and processingtechnology has steadily improved, so that sourcesand receivers can now be repeatedly positioned

with high accuracy, allowing reliable measures oftraveltime changes as small as 1 millisecond.With this level of accuracy, geophysicists are ableto use time-lapse seismic techniques for observ -ing depletion-induced traveltime changes for a

growing number of fields. In the North Sea’sEkofisk and Valhall fields, combined observationsby reservoir engineers, geophysicists and geome -chanics specialists have led them to concludethat the soft chalk of the reservoir rock wasundergoing substantial reservoir compaction,

54 Oilfield Review

> Changing seismic characteristics. Both change in geometry (top left) and change in seismic velocity (bottom left) influence seismic reflection traveltimes.The seismic two-way traveltime (TWT) (right) gradually increases toward the top of the reservoir due to overburden stretching and associated velocitydecrease. The largest time shifts are observed around the producing wells. Inside the reservoir, the seismic velocity increases because of increased stress, so the time shifts become smaller.

Dept

h, m

1,500

2,000

2,500

3,000

3,500–4,000 –3,000 –2,000 –1,000 0

Distance, m1,000 2,000 3,000

Dept

h, m

Decr

ease

in T

WT

Incr

ease

in T

WT1,500

2,000

2,500

3,000

3,500–4,000 –3,000 –2,000 –1,000 0

Distance, m1,000 2,000 3,000

Dept

h, m

1,500

2,000

2,500

3,000

3,500–4,000 –3,000 –2,000 –1,000 0

Distance, m

Vertical displacement, Δz

Time-lapse time shift for vertical P-waves, Δt

Change in vertical P-velocity, ΔVp

Δz, m

Δt, ms

ΔVp, m/s

1,000 2,000 3,000

–2

0

2

4

6

8

–0.05

0.00

0.05

0.10

0.15

0.20

0.25

23. Barkved O, Heavey P, Kleppan T and Kristiansen TG:“Valhall Field—Still on Plateau After 20 Years ofProduction,” paper SPE 83957, presented at SPEOffshore Europe, Aberdeen, September 2–5, 2003.

24. Guilbot J and Smith B: “4-D Constrained DepthConversion for Reservoir Compaction Estimation:Application to Ekofisk Field,” The Leading Edge 21, no. 3(March 2002): 302–308.Nickel M, Schlaf J and Sønneland L: “New Tools for 4DSeismic Analysis in Compacting Reservoirs,” PetroleumGeoscience 9, no. 1 (2003): 53–59.Hall SA, MacBeth C, Barkved OI and Wild P: “Time-Lapse Seismic Monitoring of Compaction andSubsidence at Valhall Through Cross-Matching andInterpreted Warping of 3D Streamer and OBC Data,”presented at the 72nd SEG International Exposition andAnnual Meeting, Salt Lake City, Utah, October 6–11, 2002.

25. Hatchell PJ, van den Beukel A, Molenaar MM,Maron KP, Kenter CJ, Stammeijer JGF, van der Velde JJand Sayers CM:“Whole Earth 4D: Monitoring

Herwanger JV and Horne SA: “Linking Geomechanicsand Seismics: Stress Effects on Time-Lapse Multi-Component Seismic Data,” presented at the 67th EAGE Conference and Exhibition, Madrid, Spain, June 13–16, 2005.Sayers CM: “Asymmetry in the Time-Lapse SeismicResponse to Injection and Depletion,” GeophysicalProspecting 55 (September 2007): 699–705.Sayers CM: “Sensitivity of Time-Lapse Seismic toReservoir Stress Path,” Geophysical Prospecting 54(September 2006): 369–380.Sayers CM: “Sensitivity of Elastic Wave Velocities toReservoir Stress Changes Caused By Production,” paper ARMA/USRMS 06-1048, presented at the 41st USSymposium on Rock Mechanics, Golden, Colorado,June 17–21, 2006.Sayers CM: “Sensitivity of Elastic-Wave Velocities toStress Changes in Sandstones,” The Leading Edge 24,no. 12 (December 2005): 1262–1267.

Geomechanics,” Expanded Abstracts, 73rd SEG AnnualInternational Meeting, Dallas (October 26–31, 2003):1330–1333.Hatchell P and Bourne S: “Rocks Under Strain: Strain-Induced Time-Lapse Time-Shifts Are Observed forDepleting Reservoirs,” The Leading Edge 24, no. 12(December 2005): 1222–1225.

26. Hatchell et al, reference 25.Hatchell and Bourne, reference 25.Herwanger JV, Palmer E and Schiøtt CR: “FieldObservations and Modeling Production-Induced Time-Shifts in 4D Seismic Data at South Arne, DanishNorth Sea,” presented at the 69th EAGE Conference and Exhibition, London, June 11–14, 2007.

27. Herwanger et al, reference 26.Sayers C: “Monitoring Production Induced Stress-Changes Using Seismic Waves,” presented at the SEGInternational Exposition and 74th Annual Meeting,Denver, October 10–14, 2004.


Autumn 2007 55

accompanied by another significantphenomenon—that of overburden stretching.23

Resulting traveltime changes are significant andof a magnitude that could not be explained bynonrepeatability of survey acquisition geometry.24

Seismic data confirmed that the reservoirrock did not deform uniformly, and deformationin the reservoir rock caused the surrounding rockto deform. In this case, the differential deforma -tion associated with reservoir compac tion and anarching effect in the overburden resulted incompressive stress relaxation and correspondingstretching in the overburden. Similar overburdentime shifts were subsequently reported abovehigh-pressure, high-temperature reservoirs andcertain deepwater-turbidite fields.25

The geomechanical implications of time-lapse time shifts are evaluated with reservoirgeomechanical models to characterizeproduction-induced subsurface deformation andto predict associated stress changes. Establishedworkflows allow geophysicists to compareobserved time-lapse time shifts against timeshifts predicted by the reservoir geomechanicalmodels.26 Both subsurface deformation and stresschanges influence the seismic traveltime, eitherby changing the length of the path that a seismicwave must travel or by altering the propagationvelocity of the seismic wave, respectively(previous page). The workflows allow predictionsof traveltime changes to any point in a three-dimensional subsurface model.

Traveltime changes can also be observed from4D seismic field experiments (left). Theprediction and observation of 4D traveltimechanges may be used to validate and calibratereservoir geomechanical models and therebyimprove their capability to predict stresschanges for a variety of projected productionscenarios. Furthermore, laboratory measure -ments conducted on rock cores are helping E&Pcompanies learn more about changes inultrasonic velocities under various stressconditions and saturation states. This allowsoperators to better manage reservoir stress andoptimize the trade-off between compaction driveof hydrocarbon production and unwantedcompaction problems such as wellbore failureand reduced permeability.

At present, the observation of changes invertical traveltime is a common practice formonitoring geomechanical changes such asvertical stress and strain. This techniqueprovides useful information, and allows geophysi -cists to identify compacting and noncompacting

reservoir compartments. However, to understandand predict other geomechanical factors, such aswellbore stability or rock failure, the triaxialstress state must be known. Recognizing thisneed, Schlumberger and WesternGeco scientistsare exploring the use of surface-seismic 4Dmeasurements to characterize the change intensor-stress over time.27

Future DevelopmentsThe industry is striving to develop furthercapabilities for integrating rock fabric intogeomechanics analysis, with the vision ofenabling operators to extrapolate informationfrom rock microstructure observations to thecore-sample scale, through the well-log scale andeventually up to the seismic scale. This capabilitywill let operators track reservoir characteristicsalong the extent of a play and beyond, tolocations where no well control exists. In doingso, geomechanics may change not only the waythat fields are drilled and produced, but also theway in which they are explored. To this end,researchers at Schlumberger are activelyinvestigating new laboratory measurementtechniques, wellbore logging methods, seismicmeasurements and modeling programs. Indeed,computational capabilities already exist; it is theactual rock, its fabric, and the relation of fabricto rock behavior that must be furthercharacterized. —MV

>Monitoring compaction over time. Acomparison of traces using the same source andreceiver position between the baseline (green)and monitor surveys (blue) shows the effect ofoverburden stretching on the arrival time of theseismic signal. Note the consistent shift towardlater arrival times of the monitor surveycompared with the baseline survey.

Top reservoir reflectionshifts toward later arrivaltime and brightens

Bottom reservoir reflectionshifts toward later arrivaltime and dims

Baseline surveyMonitor survey


56 Oilfield Review

Advancing Fluid-Property Measurements

Soraya BetancourtCambridge, Massachusetts, USA

Tara DaviesRay KennedyEdmonton, Alberta, Canada

Chengli DongSugar Land, Texas, USA

Hani ElshahawiShell International Exploration and ProductionHouston, Texas

Oliver C. MullinsJohn NighswanderHouston, Texas

Michael O’KeefeHobart, Tasmania, Australia

For help in preparation of this article, thanks to Gretchen Gillisand Don Williamson, Sugar Land, Texas; and Lisa Stewart,Cambridge, Massachusetts.CHDT (Cased Hole Dynamics Tester), Fluid Profiling, LFA (LiveFluid Analyzer), MDT (Modular Formation Dynamics Tester),Oilphase-DBR, PVT Express, Quicksilver Probe and RealVieware marks of SchlumbergerOLGA is a trademark of Scandpower AS.

Reservoir-fluid properties play a key role in designing and optimizing well completions

and surface production facilities to manage reservoirs efficiently. Therefore, accurate

fluid characterization is a vital part of any oil or gas production project. Advanced

fluid-analysis techniques provide the high-quality data required to develop appropriate

production strategies.

Discovery of an oil or gas accumulation imme di -ately prompts questions about its economicviability. Operators want to learn about theextent of the reservoir, the types of fluids thatwould be produced, expected production ratesand how long production might be sustained.Fluid analysis is a critical part of the process bywhich engineers perform reservoir characteri -zation, determine the reservoir architecture anddecide whether an oil or gas accumulation isworth developing. High-quality samples areessential, because erroneous data could leadengineers to misinterpret production parameterssuch as drainage volume, flow rates, reserves andfacilities design and completion. Clearly, poor ormisleading fluid data can have a severe negativefinancial impact.

If the reservoir analysis is positive, engineersbegin designing a production system that will

efficiently transport the reservoir fluids from theformation through wells, flowlines, productionfacilities and beyond. During this journey,reservoir fluids experience temperatures andpressures far different from their initial in-situconditions. These variations may inducephysical-state changes that would inhibit orinterrupt production if not understood prior todesigning tubulars and facilities. Therefore, todetermine how the fluids will respond toproduction conditions, engineers may want tocollect and analyze fluid samples from eachpotentially productive layer in the reservoir.

Traditionally, fluid samples are collected andsent to offsite laboratories for testing, a processthat delays data access and impedes an opera -tor’s ability to make time-sensitive developmentdecisions. Today, sophisticated formationsampling and testing tools allow data collection

1. Ratulowski J, Amin A, Hammami A, Muhammed M andRiding M: “Flow Assurance and Subsea Productivity:Closing the Loop with Connectivity and Measurements,”paper SPE 90244, presented at the SPE Annual Technical Conference and Exhibition, Houston,September 26–29, 2004.

2. For more on scale-removal techniques:Crabtree M, Eslinger D, Fletcher P, Miller M, Johnson Aand King G: “Fighting Scale—Removal and Prevention,”Oilfield Review 11, no. 3 (Autumn 1999): 30–45.

> Typical deepwater Gulf of Mexico oil-phase diagram. During the journey fromthe reservoir to the flowline, the oil temperature and pressure decline, and maycross phase boundaries at which asphaltenes (purple), waxes (blue) andhydrates (green) will tend to separate and form solid deposits. Gas begins toseparate from the oil as it passes through the bubblepoint boundary (red).

Wax

Hydrate

Bubblepoint

Reservoir

16,000

14,000

12,000

10,000

8,000

6,000

4,000

2,000

00 50 100 150

Temperature, °F

Pres

sure

, psi

200 250 300

Asphaltene

Flowline

60336schD7R1.qxp:Format Layout V7.2 12/4/07 7:11 PM Page 56

Autumn 2007 57

much earlier in the exploration process,including the ability to conduct real-time down -hole Fluid Profiling characterization ofreservoir-fluid properties and quantification oftheir variation. This technology gives operatorsthe ability to evaluate the reservoir while thesampling tool is in the borehole, and acquireadditional data if the reservoir is more complexthan previously thought. In addition, engineersand fluid-property specialists can betterdetermine where and when to sample, and howmany samples to collect. As a result, the qualityof fluid samples brought to surface issubstantially improved.

In the laboratory, chemists determine fluidcompositions, the temperatures and pressures atwhich phase transitions occur, and how eachphase behaves as a function of temperature andpressure. Accurate fluid characterization andknowledge of pressure-volume-temperature(PVT) behavior are critical for makingappropriate, cost-effective decisions about wellplanning, well construction, production and

monitoring. When initial PVT screening andthermodynamic modeling identify nonstandardphase behavior (such as emulsions, wax orasphaltene precipitation, hydrates and scales),specialized testing is often performed to betterunderstand reservoir-fluid behavior. All theseactivities fall under a general umbrella calledflow assurance.

When confronted with potential flow-assurance problems, engineers have several waysto mitigate or prevent difficulties.1 Thesemethods include thermal management (hot-fluidcirculation, electrical heating and insulation),pressure management (pumping, boosting andblowdown) and chemical treatments. Thesetechniques adjust the pressure-temperature paththat hydrocarbons experience during productionor, in the case of chemical treatments, alter thefluid composition to prevent phase changes ordisperse solid particles when precipitationoccurs. In addition, there are physicalremediation techniques such as pigging, jettingand cutting.2

As E&P companies venture into increasinglyremote production environments, particularlydeep water, flow assurance is criticallyimportant. Deepwater reservoir fluids follow atortuous PVT path from the formation to theproduction facility, increasing the probability offlow-assurance difficulties (previous page). Flowassurance may also be a concern in arcticenvironments where thermal differencesbetween the reservoir and surface facilities canbe extreme. Accurate knowledge of PVT behavioris vital because reservoir-fluid problems in theseremote locations could threaten the economicviability of a project.

This article describes the roles of chemistry,geology and thermodynamics during reservoir-characterization and flow-assurance studies atthe wellsite and the laboratory. Also presentedare two offshore-field case studies that demon -strate how these activities benefit well-completiondesign and operation.

Condensate

Median-weight oil

Water

Hydrocarbons

45

40

50

60

70

80

90

403530

454035302520151050

25201510

5

54

32

10

0


Reservoir-Fluid SamplingScientists and engineers of various disciplines relyon fluid-sample data when making decisions aboutreservoir development. For example, reservoirengineers use the data to determine reservoirarchitecture, estimate reserves, perform material-balance calculations and analyze fluid flow inporous media. Geologists need accurate informa -tion to perform reservoir correlations andgeochemical studies. Refining and marketingpersonnel make decisions about product yield andvalue. If erroneous data are used, unanticipatedand expensive consequences could result during production.3

A wide range of fluid behaviors can influencea sampling and analysis program. A reservoir-fluid system can be roughly categorized by itsvapor-liquid phase behavior; the classificationsrange from dry gas, wet gas and retrograde gas tovolatile oil, black oil or heavy oil (left).4 Anotherconsideration is hydrocarbon solid-phasebehavior. Wax and hydrate formation ispredominantly induced by a temperaturedecline, and pressure reductions or fluidcommingling generally cause asphaltenes to fallout of solution (next page, top).5

Engineers collect formation-water samples todetermine whether calcite, barite or halitescales will form within flowlines. Corrosive andtoxic substances such as carbon dioxide [CO2]and hydrogen sulfide [H2S] must be detected andmeasured because they influence tubular-alloyselection and the design of safety andenvironmental systems from the wellhead to thesurface production facility. Water pH is also animportant parameter governing scale andcorrosion, and may be measured downhole toavoid uncertainties.6

58 Oilfield Review

3. Nagarajan NR, Honarpour MM and Sampath K:“Reservoir-Fluid Sampling and Characterization—Key toEfficient Reservoir Management,” Journal of PetroleumTechnology 59, no. 8 (August 2007): 80–91.

4. McCain WD Jr: “The Five Reservoir Fluids,” in TheProperties of Reservoir Fluids (2nd Edition). Tulsa:PennWell Books (1990): 147–164.

5. Akbarzadeh K, Hammami, A, Kharrat A, Zhang D,Allenson S, Creek J, Kabir S, Jamaluddin A, Marshall AG,Rodgers RP, Mullins OC and Solbakken T: “Asphaltenes–

Problematic but Rich in Potential,” Oilfield Review 19,no. 2 (Summer 2007): 22–43.

6. Raghuraman B, Gustavson G, Mullins OC and Rabbito P:“Spectroscopic pH Measurement for High Temperatures,Pressures and Ionic Strength,” AIChE Journal 52, no. 9(2006): 3257–3265.Xian C, Raghuraman B, Carnegie A, Goiran P-O andBerrim A: “Downhole pH as a Novel Measurement Toolin Formation Evaluation and Reservoir Monitoring,”

Transactions of the 48th SPWLA Annual LoggingSymposium, Austin, Texas, June 3–6, 2007, paper JJ.

7. Riemens WG, Schulte AM and de Jong LNG: “Birba FieldPVT Variations Along the Hydrocarbon Column andConfirmatory Field Tests,” Journal of PetroleumTechnology 40, no. 1 (January 1988): 83–88.

8. Ruiz-Morales Y, Wu X and Mullins O: “ElectronicAbsorption Edge of Crude Oils and Asphaltenes Analyzedby Molecular Orbital Calculations with OpticalSpectroscopy,” Energy & Fuels 21, no. 2 (2007): 944–952.

> Generalized pressure-temperature (PT) diagram for reservoir fluids. Thediagram contains two principal regions: single-phase (green to orange) andtwo-phase (beige). The boundary between these regions is called thesaturation envelope; it has three principal features. The bubblepoint locus isthe portion at which gas begins to separate from liquid. The dewpoint locus isthe segment at which liquid begins to condense from gas. The critical point isthe location where the loci meet. The cricondentherm is the highesttemperature on the saturation envelope, and the cricondenbar is the highestpressure on the saturation envelope. Reservoir fluids are classified accordingto their in-situ reservoir and production behavior in the PT scheme. Dry gasdoes not enter the two-phase region at any point during the production PTpath. Wet gas remains a single-phase system in the reservoir regardless ofpressure depletion; however, during production, it crosses the dewpoint locusand forms a liquid phase. Retrograde gas resides in the single-phase region attemperatures between the critical point and the cricondentherm. Duringpressure depletion at reservoir temperature, liquid forms within the reservoiritself, and persists throughout the production PT path. Volatile oil resides in thesingle-phase region just to the left of the critical point. Gas liberation occurs asthe fluid crosses the bubblepoint locus during production. Black oil exists in thesingle-phase region at reservoir temperatures far lower than the critical point.Gas evolves during production, but the relative proportion of gas is smallcompared with that of volatile oil. Heavy oil is a subset of black oil that containsvery low quantities of gas, and the liquid phase is predominantly composed ofhigh-molecular-weight components.

Oil

Gas

50%

Criticalpoint

40%

Cricondenbar

Cric

onde

nthe

rm

20%

10%

5%

Temperature

Gas Field

with Retrograde

Condensation

Gas Field with

No Retrograde

Condensation

Oil Field

with Dissolved Gas

0%

Dewp

oint

locu

s

Pres

sure

Bubblepoint lo

cus

Oil and gas

> Crude-oil samples from a single column of oil in a reservoir. The continuous color change is a vivid illustration of compositional grading. (Photograph iscourtesy of Shell.)

Increasing depth


Autumn 2007 59

Another concern is the variability ofreservoir-fluid composition within a field orformation. Petroleum reservoirs may consist ofcompartments that are isolated from oneanother. Independent flow elements may have anenormous impact on an operator’s ability to drainthe reservoir. As an analogy, consider a singlereservoir compartment to be a sponge. Like asponge with its open-cell structure, the entirecontents can be drained by a single hole or well.Carrying the analogy further, independentreservoir compartments are similar to a spool ofplastic bubble wrap—a closed-cell systemthrough which the contents of one bubble cannotflow to another. If a single hole is punchedthrough the spool, drainage occurs only fromcells that are penetrated. The bubble-wrap spoolis thus highly compartmentalized.

An additional consideration is the existenceof large compositional variations of hydrocarbonsvertically and laterally within a compartment.Compositional grading is often caused by gravity,or nonequilibrium forces of biodegradation,temperature gradients, current charging, chargehistory or incompetent sealing shales.7 Themagnitude of compositional variation can varygreatly, depending on the geological andgeochemical history of the reservoir (previouspage, bottom).8

A technically robust fluid-sampling programis vitally important when reservoir compart men -talization, compositional grading or both exist.Reservoir-formation properties influence theability to collect representative fluid samples.Sample collection requires fluid flow into theborehole, which occurs only when the wellboreflowing pressure is lower than formationpressure. However, if the flowing pressure fallsbelow the fluid saturation pressure, a gas phase(in the case of volatile or black oil) or a liquidphase (in the case of retrograde gas) will form(right). The relative mobility of each fluid phaseis different; because of unequal flow, thecomposition of the fluid exiting the formationwill not be the same as that in the reservoir. Thiseffect can be minimized or eliminated bysampling at flow rates and pressure differentialsthat are as low as feasible.

Finally, accurate reservoir-temperaturemeasure ments are vital. Errors of just a fewdegrees during PVT testing could result inmisinter pretation. For example, what is conden -sate in the formation may behave like a volatileoil at an incorrect temperature in the laboratory.This error could result in costly production-design errors.

> Common deposits that form in tubulars during hydrocarbon production.Wax and hydrate deposition mainly result from a temperature decrease,while asphaltene precipitation may be triggered by changes in pressure,temperature and composition. Inorganic scales arise from changes inpressure, temperature and composition of aqueous fluids that accompanyhydrocarbon production. (With kind permission of Springer Science andBusiness Media.)

Asphaltene Wax

Inorganic scale Hydrate

> Pressure-decline effects during reservoir-fluid sampling. If the reservoirpressure falls below the bubblepoint while sampling oil, gas separation willoccur, creating a two-phase system (top). Similarly, if the reservoir containsretrograde gas, liquid will form if the reservoir pressure falls below thedewpoint. When phase changes occur in the reservoir, the high-mobility phaseflows preferentially because of relative permeability effects, and the sample isnonrepresentative. Maintaining the reservoir pressure above the bubblepoint ordewpoint during sampling preserves single-phase behavior and ensurescollection of a representative sample (bottom).

Pres

sure

Reservoir pressure

Originalreservoirfluid

Originalreservoirfluid

BubblepointWellbore flowing pressure

Free gas

One-phaserepresentativesample fluid

Pres

sure


There are two principal hydrocarbon-sampleacquisition methods—bottomhole and surfacesampling. Bottomhole sampling involves convey -ing a sampling tool on a drillstem-test (DST)string, wireline or slickline to the producing zoneor zones. In an open hole, sampling can be

performed by the MDT Modular FormationDynamics Tester and the Quicksilver Probe tool forfocused extraction of pure reservoir fluid. Cased-hole sampling devices include the CHDT CasedHole Dynamics Tester, single-phase reservoirsampler (SRS) and SCAR sampling tool.9 Surfacesampling, which is most frequently performed atthe separator under stable flow conditions,involves collecting gas and liquid samples.Engineers may acquire surface samples duringexploration if downhole methods are unavail able,

and may continue to do so throughout a well’slifetime to monitor fluid-property evolution.10

Bottomhole fluid samples must be extractedfrom locations that will yield the most relevantinformation for making decisions. To aid in thisendeavor, today’s sampling and testing toolsinclude an array of instruments that can performdownhole fluid analysis (DFA). DFA tools providereal-time fluid-property measurements atreservoir conditions, allowing engineers toanalyze samples before they are gathered.

60 Oilfield Review

9. For more on formation-fluid sampling devices:MDT tool: Colley N, Ireland T, Reignier P, Richardson Sand Joseph J: “The MDT Tool: A Wireline TestingBreakthrough,” Oilfield Review 4, no. 2 (April 1992): 58–65.Andrews RJ, Beck G, Castelijns K, Chen A, Cribbs ME,Fadnes FH, Irvine-Fortescue J, Williams S, Hashem M,Jamaluddin A, Kurkjian A, Sass B, Mullins OC, RylanderE and Van Dusen A: “Quantifying Contamination UsingColor of Crude and Condensate,” Oilfield Review 13, no.3 (Autumn 2001): 24–43.Quicksilver Probe: Akkurt, R, Bowcock M, Davies J, Del Campo C, Hill B, Joshi S, Kundu D, Kumar S, O’Keefe M, Samir M, Tarvin J, Weinheber P, Williams Sand Zeybek M: “Focusing on Downhole Fluid Sampling and Analysis,” Oilfield Review 18, no. 4 (Winter2006/2007): 4–19.CHDT tool: Burgess K, Fields T, Harrigan E, Golich GM,Reeves R, Smith S, Thornsberry K, Ritchie B, Rivero Rand Siegfried R.: “Formation Testing and SamplingThrough Casing,” Oilfield Review 14, no. 1 (Spring 2002):46–57.SRS and SCAR tools: Aghar, H, Carie M, Elshahawi H,Gomez JR, Saeedi J, Young C, Pinguet B, Swainson K,

> Schematic diagram of a MDT ModularFormation Dynamics Tester, using the QuicksilverProbe tool for focused extraction of reservoirfluid. The focused sampling probe is set againstthe borehole wall to withdraw formation fluids forFluid Profiling characterization and samplecollection. The downhole LFA Live Fluid Analyzersprovide real-time quantitative measurements ofdensity, viscosity, GOR, hydrocarbon compositionand formation-water pH.

Sampleflow

Power cartridge

Sample-bottlemodule

Pump module(sample)

LFA fluid analyzer(sample)

Hydraulic module

Focused samplingprobe

LFA fluid analyzer(guard)

Pump module(guard)

Guardflow

> Reservoir section (left) and real-time DFA GOR measurements (right) ofreservoir fluids in a deepwater exploration well. There is good agreementbetween the GOR values and the reservoir structure. The FCA analysissuggested locations for fluid sampling (blue dots, right). The reservoir fluidsvary significantly from dry gas (Fluids A and B) and condensate gases (Fluid C)at the top, to black oils with different GORs (Fluids D through J) at the bottom.At the bottom of the oil column (Fluids H, I and J), GOR variations indicate agentle fluid-composition gradient. On the other hand, a GOR inversion isevident between Fluids E and F; Fluid F is deeper than Fluid E, but has a higherGOR. A similar inversion occurs between Fluids G and J, suggesting a complexreservoir structure with a potential flow barrier at sampling station J.

Seal

Seal

Reservoir section

Water Rich gas and condensateOil

0

B

CDE

F GJH

I

500

100 1,000

Depth, ft

10,000 100,000

DFA GOR, ft3/bbl

1,000

1,500

2,000

2,500

3,000

3,500

Dry gas

A

Takla E and Theuveny B: “The Expanding Scope of WellTesting,” Oilfield Review 19, no. 1 (Spring 2007): 44–59.

10. For more on surface sampling: Aghar et al, reference 9.11. For more on optical DFA methods:

Betancourt S, Fujisawa G, Mullins OC, Carnegie A,Dong C, Kurkjian A, Eriksen KO, Haggag M, Jaramillo ARand Terabayashi H: “Analyzing Hydrocarbons in theBorehole,” Oilfield Review 15, no. 3 (Autumn 2003): 54–61.Crombie A, Halford F, Hashem M, McNeil R, Thomas EC,Melbourne G and Mullins OC: “Innovations in WirelineFluid Sampling,” Oilfield Review 10, no. 3 (Autumn 1998):26–41.Dong C, Hegeman PS, Carnegie A and Elshahawi H:“Downhole Measurement of Methane Content and GORin Formation Fluid Samples,” SPE Reservoir Evaluation &Engineering 9, no. 1 (February 2006): 7–14.

12. Betancourt SS, Fujisawa G, Mullins OC, Eriksen KO,Dong C, Pop J and Carnegie A: “Exploration Applicationsof Downhole Measurement of Crude Oil Composition andFluorescence,” paper SPE 87011, presented at the SPEAsia Pacific Technical Conference on IntegratedModeling for Asset Management, Kuala Lumpur,March 29–30, 2004.

13. Dong CM, O’Keefe M, Elshahawi H, Hashem M,Williams S, Stensland D, Hegeman P, Vasques R,Terabayashi T, Mullins O and Donzier E: “New DownholeFluid Analyzer Tool for Improved ReservoirCharacterization,” paper SPE 108566, presented at theSPE Offshore Europe Oil and Gas Conference andExhibition, Aberdeen, September 4–7, 2007.

14. Venkataramanan L, Weinheber P, Mullins OC,Andrews AB and Gustavson G: “Pressure Gradients and Fluid Analysis as an Aid to Determining ReservoirCompartmentalization,” Transactions of the 47th SPWLAAnnual Logging Symposium, Vera Cruz, Mexico, June 4–7, 2006, paper S.

15. Dong C, Elshahawi H, Mullins OC, Venkataramanan L,Hows M, McKinney D, Flannery M and Hashem M:“Improved Interpretation of Reservoir Architecture andFluid Contacts through the Integration of Downhole FluidAnalysis with Geochemical and Mud Gas Analyses,”paper SPE 109683, presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition, Jakarta, October 30–November 1, 2007.


Autumn 2007 61

DFA methods include in-situ opticalabsorption spectroscopy, optical reflectance,fluorescence and some nonoptical measure mentsincluding density, viscosity and pH. Thespectrometer operates in the visible to near-infrared range—at wavelengths between 400 and2,100 nm. Spectra are recorded in real time,revealing the proportions of methane [C1], ethaneto pentane [C2–5], hexane-plus [C6+] and CO2

fractions, as well as a gas/oil ratio (GOR) estimate.Additionally, differences between the reservoir-fluid and drilling-fluid spectra indicate the level ofsample contamination.11 Downhole fluorescencemeasurements provide fluid-phase informationthat is especially useful for retro gradecondensates and volatile oils.12 Fluorescence isalso sensitive to liquid formation in a condensategas when the flowing pressure falls below thedewpoint, allowing sampling engineers to monitorfluid-phase separation in real time, and ensurethat representative single-phase samples arecollected (previous page, top left).13

Fluid Profiling reservoir-fluid characteri -zation by DFA can diagnose compositionalgrading and help identify reservoir compart -ments. For example, abrupt fluid-composition orGOR changes between zones in a single well orbetween neighboring wells may indicate compart -mentalization. To confirm that perceivedfluid-property differences are truly significant,engineers must first consider measurementuncertainties. A recent method to evaluateuncertainties is the fluid comparison algorithm(FCA).14 FCA uses parametric models to estimateGOR and coloration uncertainties as a functionof optical-density (σε) and mud-contamination(ση) measurement variability. The algorithmcompares measurements acquired from twofluids and calculates the probability that differ -ences are statistically significant. When the FCAresult indicates that the fluids are different,sample acquisition for detailed surface analysisis justified. The following case study demon -strates how engineers employ DFA and FCA to characterize a reservoir and determinesampling locations.

Sampling and Reservoir Characterization in aDeepwater AccumulationIn a deepwater exploration well, Shell employedreal-time DFA and FCA to characterize thereservoir architecture and decide where tocollect fluid samples.15 As the MDT assemblytraveled down the well, the spectrometersmeasured the GOR at several locations (previouspage, top right). Based on FCA analysis, fluidsamples were collected at 10 different depths.

The analysis revealed wide variations inreservoir-fluid composition, ranging from dry gasand condensate gases at the top to black oils withdifferent GORs at the bottom. At the bottom ofthe oil column, the GOR varied gradually withdepth in the bottom sand, indicating a fluid-composition gradient. GOR inversions were alsodetected between the top and bottom sands,suggesting the presence of flow barriers and acomplex reservoir structure.

DFA and FCA analysis showed that fluidsabove and below the inversion had a greater than99% probability of existing in differentcompartments. A pressure discontinuity betweenthe compartments confirmed the absence ofhydraulic communication.

Shell and Schlumberger fluid specialistsexplored the oil column in detail, comparing theGOR information with petrophysical, formation-pressure and mud-gas logs, and performing FCAanalysis (below). The gamma ray and pressure

> Expanded log presentation and FCA analysis of an oil column at the bottom of a deepwaterreservoir. Correlation of the GOR log (top left) with the gamma ray log (green, top center) reveals thatFluids F and G come from one sand zone, and Fluids J, H and I reside in another. In the upper sand,Fluids F and G have the same GOR, while in the lower sand, Fluids J, H and I display a GOR decreasewith depth. The fluid densities derived from pressure gradients (blue) reveal two principal features—afluid-density inversion between Fluids G and J, and a gradual density increase from Fluid J to Fluid I—suggesting no vertical communication between the top and bottom sands. The mud-gas log (top right)offers further supporting evidence. At the depth of Fluid J, the δ13C value (red) falls abruptly andincreases gradually with depth. The relative methane concentration (blue) also fluctuates sharply atFluid J, then decreases gradually with depth—a feature consistent with the fluid-density increase.Final confirmation resulted from applying the FCA technique (bottom). The FCA algorithm generates anumber representing the probability that two fluids are statistically different. As measurementuncertainties (σ) decrease, confidence in data quality increases. Therefore, low σ values indicate ahigh probability that perceived fluid-property differences are real. The contour plot indicates that Fluid G has a 99% probability of being different from Fluid J, and a 95% probability of being differentfrom Fluid H. Discovery of reservoir compartmentalization led Shell to reevaluate the field and makesignificant strategic adjustments in reservoir development.

FG

H

I

J

Dept

h, ft

2,750

2,800

2,850

2,900

2,950

3,000

3,050

3,100

3,150

5,500 Pressure, psi 6,000

0 Gamma Ray, API 150

Relative Methane Concentration

Higher concentration ofC1 corresponds to lowerdensity of Fluid J

Increasingcontribution

of thermogeniccharge

0 2 4

–55 δ13C methane, parts per thousand –50

900 GOR, ft3/bbl 1,500

Top sand

Bottom sand

JH

I

GF

Oil-b

ase

mud

con

tam

inat

ion,

ση

Optical density, σε

0.10

0.09

0.07

0.08

0.06

0.05

0.04

0.03

0.02

0.01 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10

Fluids G and H

P = 0.80

P = 0.99

P = 0.95

P = 0.90Fluids G and J


logs showed that the top portion of the oil residesin one sand lobe with a relatively constant GORand pressure gradient. The rest of the oil is in alower sand lobe, where the GOR decreases withdepth. The pressure log was consistent, indi -cating a significant fluid-density differencebetween the upper and lower lobes.

Recent advances in mud-gas logging havegiven engineers another tool to perform real-time detection of seals and permeability barriers,lithological variations and fluid contacts.16 Gases

gathered at the surface during drilling orseparated from fluid samples can be analyzed forisotopic content. The isotopic signature, δ13C, isthe value of the 13C/12C methane-isotope ratio ina sample relative to a standard, expressed inparts per thousand. When plotted alongside astandard mud-gas log, larger δ13C values mayindicate higher concentrations of biogenic gas inthe reservoir. Trending δ13C values may indicatenonequliibrium methane distributions, and aclear break in the methane-isotope signature in

mud gas may imply the presence of a seal.17

Engineers noted a δ13C discontinuity at around2,950 ft [899 m], suggesting yet another flowbarrier. FCA analysis provided confirmation,calculating a 95 to 99% probability that fluidsabove and below the discontinuity weredissimilar and from different compartments withno communication.

The discovery of the fluid compartments inthis deepwater field led Shell engineers to adjusttheir reservoir models and development plan -ning. Reservoir compartmentalization increasesdesign complexity and cost because engineersmust treat each zone independently. Productionforecasts, reserve calculations and enhancedrecovery schemes became proportionately more complex.

Improved regional understanding ofsubsurface architecture impacted short-termdecisions on sidetrack objectives. From thelocation of the borehole, Shell also reasoned thatcompositional grading in the lower reservoirlikely extended downward from the penetratedzone; as a result, production-facility plans werealtered to anticipate a GOR reduction with time.

Access to real-time fluid analyses allowedShell to make decisions much earlier in the field-development process, and accelerated theproject by at least six months. At today’s oilprices, the time saved was worth hundreds ofmillions of dollars.

Laboratory Fluid Preparation and Sample Chain of CustodyThe preceding case study demonstrates theconsiderable effort and care that engineers apply during the sampling process. Nevertheless,the harsh downhole environment and the nature of well operations may render DFAequipment, sampling devices and subsequentanalysis suscep tible to fouling, failure and other inaccuracies.

Schlumberger engineers addressed thisproblem by implementing a chain-of-custodyprocedure, a concept borrowed from forensicscience.18 Evidence must make the journey froma crime scene to the courtroom in a validatedand secure manner; otherwise, it may not beadmissible in court. Similarly, chemists at aremote testing laboratory should be able todetermine whether the chemical composition ofa field sample has been preserved. DFA providesa convenient way to establish a chain of custodyfor fluid samples, because chemists have theopportunity to compare analytical data acquireddownhole with those from the correspondingsamples that reach the laboratory.

62 Oilfield Review

> Chain-of-custody investigation of a valid crude-oil sample. Chemistscompare visible near-IR spectra from DFA and the laboratory sample. Bothspectra were measured at downhole conditions—15,000 psi [103 MPa] and250°F [121°C]. The spectral analysis (top) compares the downhole discretedata (red circles) with the continuous spectrum from the laboratory sample(blue). Data regression (bottom) of the LFA optical densities versus thosefrom the laboratory shows excellent agreement, evidenced by the near-perfect overlay of the regression line (red) over the X = Y line (blue). The fluidsample is well-preserved and suitable for further laboratory studies.

0.9

1.0

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

-0.11,000 1,100 1,200 1,300 1,400 1,500

Wavelength, nm

Laboratory optical density

1,600 1,700 1,800 1,900 2,000

Optic

al d

ensi

tyLF

A op

tical

den

sity

1.5

1.0

0.5

0

0 0.5 1.0 1.5

Laboratory sampleLFA tool

X = YData regression

1,6001,920 1,445

1,6711,290

1,725

1,070


Autumn 2007 63

In the field, after tool retrieval, engineersremove the reservoir-fluid samples. At this point,a PVT Express onsite well fluid analysis systemmay be available to conduct preliminarymeasurements that determine whether thecollected-sample properties agree with thosemeasured by DFA.19 If onsite analysis isunavailable or more sophisticated testing isrequired, the samples are shipped to a remotetesting laboratory in the original sample bottlesor transferred to an approved shipping container.When fluid samples arrive at the Schlumbergerlaboratory, chemists restore the fluid inside thesample container to the original reservoirtemperature and pressure, and allow the fluid toequilibrate by agitating it continuously for up tofive days. The restoration process is intended toredissolve precipitated asphaltene and waxparticles, ensure a homogeneous fluid through -out the sample cylinder and provide asingle-phase representative fluid for testing.

The visible near-infrared (IR) spectrum is aneffective hydrocarbon fingerprint.20 The DFAspectrometer performs sample analysis as thecrude-oil sample is acquired, providing a directfluid-property measurement under downholeconditions. In the laboratory, chemists performthe same measurement with a research-gradespectrometer at the downhole temperature andpressure. Differences between the DFA andlaboratory spectra may indicate that thelaboratory sample has been compromised. Forexample, if the methane concentration is lowerin the laboratory spectrum, then sample-bottleleakage or a fluid-transfer error may haveoccurred during sampling or transport to thelaboratory. The examples presented belowillustrate the chain-of-custody technique.

The first example involves a fluid sampleacquired from an offshore oil field. There isexcellent agreement between the downhole and laboratory spectra (previous page). Thisindicates that the sample is well-preserved andsuitable for further laboratory studies.

The second example involves another samplefrom an offshore oil field (right). The downholeand laboratory spectra have good overallagreement, but there are subtle differences inthe wavelength region above about 1,600 nm.Chemists investigated the spectral differencesfurther by employing algorithms to calculateGORs and probabilities of similarity between thetwo samples.21 The calculations showed that theGOR from the downhole spectrum, 580 ft3/bbl[103.3 m3/m3], was significantly higher than thatfrom the laboratory spectrum, 320 ft3/bbl[57.0 m3/m3]. This difference corresponded to a

16. Venkataramanan L, Elshahawi H, McKinney D,Flannery M, Hashem M and Mullins OC: “Downhole Fluid Analysis and Fluid Composition Algorithm as an Aid to Reservoir Characterization,” paper SPE 100937,presented at the SPE Asia Pacific Oil and GasConference and Exhibition, Adelaide, Australia,September 11–13, 2006.

17. Berkman T, Ellis L and Grass D: “Integration of Mud GasIsotope Data with Field Appraisal at Horn MountainField, Deepwater Gulf of Mexico,” AAPG Bulletin 86,no. 13 (2002): supplement.

18. Betancourt SS, Bracey J, Gustavson G, Mathews SG andMullins O: “Chain of Custody for Samples of Live CrudeOil Using Visible Near-Infrared Spectroscopy,” AppliedSpectroscopy 60, no. 12 (December 2006): 1482–1487.

19. Aghar et al, reference 9.20. For more on visible and near-IR spectroscopy:

Crombie et al, reference 11.21. Mullins OC, Beck G, Cribbs MY, Terabayashi T and

Kagasawa K: “Downhole Determination of GOR onSingle-phase Fluids by Optical Spectroscopy,”Transactions of the 42nd SPWLA Annual LoggingSymposium, Houston, June 17–20, 2001, paper M.Venkataramanan L, Fujisawa G, Mullins OC, Vasques RRand Valero H-P: “Uncertainty Analysis of Near-InfraredData of Hydrocarbons,” Applied Spectroscopy 60, no. 6(June 2006): 653–662.

> Chain-of-custody investigation of a compromised crude-oil sample. Bothspectra were measured at downhole conditions—20,000 psi [138 MPa] and 200°F[93°C]. The visible near-IR spectral analysis (top) shows subtle OD differences inthe region above about 1,600 nm. At 1,671 nm, indicating methane, thelaboratory-sample OD is lower than that measured by the LFA tool. At 1,725 nm,the methylene “oil peak,” the laboratory sample OD is higher than that measuredby the LFA tool. These differences are also visible in the linear regression plot(bottom). The ratio between the methane and oil peaks can be used to calculatethe GOR. In this case, the algorithms indicated that the laboratory-sample GORwas significantly lower than that of the field sample. Therefore, further analysiswould be necessary before this sample could be trusted.

1,000 1,200

0 0.5 1.0 1.5

1,400 1,600 1,800 2,000

Wavelength, nm

Laboratory optical density

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

1.5

0.5

0

1.0

0

Optic

al d

ensi

tyLF

A op

tical

den

sity

X = YData regression

Laboratory sampleLFA tool

1,600

1,671

1,820

1,280

1,445

1,725

1,070


93% likelihood that either one (or both) of thespectra were in error, or the laboratory samplewas compromised. Further investigation ofsurface procedures and DFA-tool performancewould be required before this sample could beused with confidence in the laboratory.

Laboratory Techniques for Flow AssuranceIn the laboratory, chemists determine fluidcompositions and measure fluid propertiesrelated to flow assurance. For compositionalanalysis, an accurately measured volume of fluidis isobarically and isothermally transferred to apycnometer to measure mass and density. Then,the pycnometer is connected to an apparatus inwhich the oil sample is cooled to ambienttemperature and decompressed. The volume ofgas liberated by this procedure allowscalculation of the GOR.

Chemists employ gas chromatography todetermine the vapor-phase composition up toC15+ and the liquid-phase composition up to C36+.The bulk crude-oil composition is calculated bysumming the individual contributions from eachphase (above left). This procedure ensures that aconsistent reservoir-fluid composition is avail ablefor subsequent fluid-property characterizationand reservoir-production simulation.

The bubblepoint pressure is determined byperforming a constant composition expansion(CCE) test. Technicians place a known volume ofequilibrated fluid in a PVT cell at reservoirtemperature and pressure (left). The fluid isinitially single phase, and testing begins byreducing the pressure isothermally andmonitoring the fluid-volume change. Eventually,the fluid separates into two phases. Pressurereductions continue in increments ranging from100 to 500 psi [0.69 to 3.45 MPa], the vapor andliquid phases are equilibrated at each step, andphase volumes are measured and plotted as afunction of pressure. For a black oil, theintersection of the single- and two-phase lines inthe PV plot defines the bubblepoint.

The flow-assurance testing protocol dependson the nature of the crude oil. For example, whenasphaltenes are of concern, then saturate,aromatic, resin and asphaltene (SARA) analysisand paraffinic solvent titration with dead oil areprincipal screening techniques.22 It is alsocommon to measure the asphaltene-precipitationpressure on a live-oil sample. If testing identifiesan asphaltene-precipitation problem, additionalstudies are conducted to map out the asphaltenephase diagram and evaluate the effectiveness ofchemicals or coatings as prevention strategies.23

64 Oilfield Review

> Typical crude-oil compositional analysis, determined by gas chromatography. The plot shows thehydrocarbon contribution from C1 to C30+, as well as the CO2 concentration.

100.00

10.00

1.00

Components

CO2

H2S N2 C1 C2 C3

0.10

0.01

I-C4

N-C

4I-C

5N

-C5 C6

MCY

C-C5

Benz

ene

CYCL

-C6 C7

MCY

CL-C

6To

luen

e C8C2

-Ben

zene

M&

P-Xy

lene

O-Xy

lene C9 C1

0C1

1C1

2C1

3C1

4C1

5C1

6C1

7C1

8C1

9C2

0C2

1C2

2C2

3C2

4C2

5C2

6C2

7C2

8C2

9C3

0+

Wei

ght p

erce

nt

> Pressure-volume-temperature (PVT) cell and bubblepoint determination. The fully visual PVT cellallows direct confirmation of bubblepoints at various temperatures and pressures (top). A magneticmixer provides vigorous agitation to maintain phase equilibrium. A video-based cathetometermeasures fluid levels in the cell for phase-volume calculations. The oven heats the PVT cell to thetest temperature, and technicians monitor and plot phase volumes as a function of pressure. Theinflection point in the curves as pressure decreases defines the bubblepoint. In the case depictedhere, the bubblepoint is approximately 5,000 psi [34.5 MPa] (bottom).

High-pressure pumps

Oil

Solv

ent

Side view

Magnetic mixer

Cathetometer

16,030

30

2,030

4,030

6,030

10,030

12,030

14,030

8,030

Pres

sure

, psi

Volume, cm3

25 30 35 40 45 50 55 60 65 70 75

Tres = 176°FT = 120°FT = 75°F


Autumn 2007 65

Waxy crude oils pose different production andtransportation challenges. Wax deposition insidetubulars and pipelines reduces the effective flowarea, increasing the pressure drop and potentiallycausing complete blockage. Therefore, it isimportant to fully understand the oil’s behaviorthroughout the pressure and temperature pathfrom the formation to the production facilities.Temperature is the dominant parameter affectingwaxy crude-oil viscosity, gel strength, pour point,wax crystallization and deposition. Although deadoils such as stock-tank oil (STO) can be used togenerate preliminary data, it is important toinclude live oils in the testing program becausepressure and dissolved gases may stronglyinfluence wax solubility.

The first characterization step is to measurethe amount of wax that can precipitate anddeposit on a solid surface. Live-oil filtration andhigh-temperature gas chromatography (HTGC)are common methods to measure the waxcontent. HTGC is more valuable because itprovides the n-paraffin composition at highcarbon numbers (from C60 to C100)—informationchemists enter into thermodynamic models topredict wax behavior.

The wax-appearance temperature (WAT) isone of the most important flow-assurancemeasure ments, indicating the temperature atwhich wax crystals begin to form in a crude-oilsample. This measurement provides a preliminaryassessment of the likelihood of wax-relateddeposition problems. Laboratory workers placedead oil on the stage of a cross-polar microscope(CPM) and block light transmission by adjustingpolarized prisms at opposite ends of the sample.When illuminated by polarized light, crystallinematerials disturb the polarization plane;therefore, as the fluid sample cools, wax-crystalformation is clearly visible as bright spots appearagainst the black background (above right). Somelaboratories have high-pressure CPM instrumentsthat can measure the WAT in live oils. The samplecell operates at pressures up to 20,000 psi[138 MPa] and temperatures up to 392°F [200°C].

Formation and growth of wax crystals mayaffect the rheological behavior of the crude oil.Above the WAT, most hydrocarbon systemsbehave as Newtonian fluids; however, non-Newtonian behavior such as shear thinning maycommence as the fluids cool and phase changesoccur. This behavior must be properly quantifiedto allow engineers to design a suitableproduction system.24 Rheological properties ofinterest in the context of waxy crude are pourpoint, apparent viscosity and gel strength.

The pour point is the temperature belowwhich a fluid is no longer pourable because ofviscosification, gelation or solids formation.Pour-point testing with dead oils conforms toASTM Standard D97.25 For live oils, engineers usean apparatus containing a visual sapphire cellmounted on an automatic, vibration-free pivotingbracket that resides in a programmableconvection oven. The oil sample is heated to thereservoir temperature, and slowly cooled untilfluid movement in the cell ceases. Most waxycrude oils begin to display non-Newtonianbehavior near the pour point. If the pour pointexceeds anticipated oil-production temperatures,engineers may add pour-point-depressantchemicals to maintain flow.

The apparent viscosity of waxy crude oilsgenerally increases dramatically as thetemperature and shear rate decrease,particularly at temperatures near the pour point.Schlumberger engineers measure viscosity with arheometer that can operate at 6,000 psi[41.4 MPa] and 302°F [150°C], allowing workwith live oils.

Waxy crude oils tend to form gels attemperatures below the pour point. In the eventof a production shutdown, high pumping pres -sures may be required to break the gel andrestore flow. Therefore, yield-stress data fromwaxy crude oils are necessary to properly designflowlines and avoid production problems. The

22. Live oil contains the dissolved gases present in thereservoir. If the gases are released, or “flashed,” at thesurface or in the laboratory, the residual liquid is calleddead oil. Stock-tank oil (STO) is also a dead oil.

23. For a detailed discussion of asphaltene flow assurance:Akbarzadeh et al, reference 5.

24. At constant temperature and pressure, Newtonian fluidsdisplay constant viscosity at all shear rates. Theviscosity of non-Newtonian fluids is not constant at allshear rates. Apparent viscosity is the viscosity of a fluidat a given shear rate and temperature.

25. ASTM D97-06 Standard Test Method for Pour Point ofPetroleum Products. West Conshohocken, Pennsylvania,USA: ASTM International, 2006.

> Determination of wax-appearance temperature (WAT) by cross-polar microscopy (CPM). Themicroscope is equipped with a heated stage between two polarized prisms (top). At the beginning ofthe test, the prisms are adjusted to block light transmission. As the stage cools, wax-crystal formationchanges the light polarity, and the video camera sees the appearance of bright spots. In this example,no light is transmitted at 111°F [44°C] (bottom, right panel) because the fluid temperature is above theWAT. Spots begin to appear when the fluid cools to the WAT (center panel), and the image becomesbrighter as the fluid cools to 0°C (left panel). This method is accurate to ± 2°F [1.1°C].

0°C 44°CWAT = 42°C

10 50Temperature 48.6

IR filter

Polarizer

Analyzer

20

360°rotatable

stage

Coolinggas

Hot stagetop view

Chargedcoupleddevice

Hot stage


yield stress of live fluids can be determined by amodel pipeline test (MPT) (left).26

Another important measurement in thecontext of production-system design and flowassurance is the wax-deposition rate. Theprincipal controlling parameters are fluidtemperature, heat loss through the pipeline wall,wax content, shear rate and fluid viscosity. Thewax-deposition rate can be determined undersimulated pipe-flow conditions in a wax-deposition flow loop (WDFL) (below left).

The following case study illustrates howoperators use laboratory measurements to developoperational strategies that prevent, mitigate orremediate wax deposition and gel formation.

Optimizing Subsea System Design in West AfricaA West African field, located at a water depth lessthan 1,000 ft [300 m], has a reservoirtemperature and pressure of 170°F [76.7°C] and3,180 psi [21.9 MPa]. The seabed temperature is55°F [12.8°C]. The operator planned to produceoil through a looped 6-in. ID subsea tieback to ahost facility 2 miles [3.2 km] from the reservoir.With the large temperature difference betweenreservoir and seabed, and the long distance thatfluids would flow along the cold seafloor, theoperator needed assurance that solids depositionwould not impede flow. Schlumberger collecteddownhole fluid samples and sent them to theOilphase-DBR fluid sampling and analysislaboratory in Edmonton, Alberta, Canada for aflow-assurance study.27

Compositional analysis of the reservoir fluidrevealed a black oil with a GOR of 230 ft3/bbl[41.0 m3/m3] and an API gravity of 36.3.Compositional analysis indicated that thefraction with a carbon number above C30 was35.8%. Further characterization of the C30+

fraction and n-paraffin distribution revealed thatthe crude oil contained about 13.1 wt% C17+

n-paraffins.28 The relatively high n-paraffinconcentration was cause for concern that the

66 Oilfield Review

26. The yield stress, τy, is calculated by the following force-

balance equation: τy = where Py is the hydraulic pressure necessary to cause fluid movement, D is theinner diameter of the coil and L is the coil length.

27. Alboudwarej H, Huo Z and Kempton E: “Flow-AssuranceAspects of Subsea Systems Design for Production ofWaxy Crude Oils,” paper SPE 103242, presented at theSPE Annual Technical Conference and Exhibition, San Antonio, Texas, September 24–26, 2006.

28. Paraffin is a common name for a group of alkanehydrocarbons with the general formula CnH2n+2, where n is the number of carbon atoms. The simplest paraffinmolecule is methane, CH4, a gas at room temperature.Octane, C8H18, is liquid at room temperature. The solidforms of paraffin are heavier molecules from C20 to C40.Linear members of the series (those with no branches orcyclic structures) are called n-paraffins.

>Model pipeline test (MPT) apparatus. Stock-tank oil circulates through acoil of tubing immersed in a temperature-controlled bath. Flow through thecoil stops when the bath reaches the test temperature, allowing the fluid toage and form a gel structure. After the aging period, laboratory workersmeasure the nitrogen pressure necessary to initiate flow in the coil andcalculate the gel strength from a simple force-balance equation.

N2 pressure to break gel

System charging pump

Conv

ectio

n ov

en

High-pressurecirculation pump

Fluidsamplecylinder

Backpressureregulator

Heated lines

Temperature-controlled

bath

>Wax deposition flow loop (WDFL). The WDFL is a miniature flow loop that exposes stock-tank oils(STO) to a range of heat fluxes and shear rates that would be expected in the flowline during actualproduction. The deposition section of the flow loop is a 39-in. [1-m], 0.375-in. OD stainless steel tube. A 0.53-galUS [2.0-L] oil reservoir maintains the oil above the WAT throughout the test. The oil exits thereservoir and flows at a controllable rate through 50 ft [15 m] of coiled copper tubing in a temperingbath for temperature adjustment. The deposition loop is immersed in a cooling bath to simulate heatloss in the pipeline. As the oil flows through the deposition loop, a data-acquisition system monitors the wall temperature, oil and water temperatures, the pressure drop between the inlet and outlet of thecoil and the flow rate. Wax deposits constrict flow inside the deposition loop, increasing the pressurerequired to maintain flow. The increase in pressure allows calculation of the amount of deposited wax.

Thermocouples

Pressure transducer

Reservoir

Remelting bath Tempering bath

Pump

Flowmeter

Thermocouples

Deposition section

Cooling water

Cooling bath Thermocouples

PyD____4L


Autumn 2007 67

fluid might exhibit flow-assurance problemsrelated to wax deposition. Therefore, theOilphase-DBR team objective was to generatemeasurements that would provide guidanceconcerning ways to mitigate and remediate waxdeposition during steady-state and transientevents during crude-oil production.

Constant composition expansion (CCE)testing measured a bubblepoint pressure of 700 psi [4.8 MPa] at the reservoir temperature.WAT and pour-point tests were performed withboth live- and dead-oil samples (right). As thefluid pressure fell below the bubblepoint toambient conditions, dissolved gas escaped, theaverage sample composition shifted towardheavier hydrocarbons, and both the WAT andpour point increased approximately 20°F [11.1°C].

Rheological testing revealed gel-structureformation in live crude oil at low shear rateswhen the fluid temperature fell below the pourpoint and approached the seabed temperature(below). As the rheometer pressure decreased,the shear stress required to break the gelsincreased—behavior consistent with the loss oflighter hydrocarbons. Around the pour point,

>Wax-appearance temperature (WAT) and pour-point (PP) data measuredwith live and dead West African crude oil. The dead-oil measurements areshown at atmospheric pressure. As the live-oil fluid pressure decreased from700 psi to atmospheric pressure, both the WAT (red) and pour point (blue)increased approximately 20°F. The increases resulted from the loss ofdissolved gases in the fluid.

160

40

70

100

130

Tem

pera

ture

, °F

Pressure, psi0 200 400 600 800

WATPour point

1.0E+11

1.0E+10

1.0E+09

1.0E+08

1.0E+07

1.0E+06

1.0E+05

1.0E+04

1.0E+03

1.0E+02

1.0E+01

1.0E+00

Visc

osity

, mPa

-s

Shear stress, Pa

300 psi

10 100

1.0E+11

1.0E+10

1.0E+09

1.0E+08

1.0E+07

1.0E+06

1.0E+05

1.0E+04

1.0E+03

1.0E+02

1.0E+01

1.0E+00

Visc

osity

, mPa

-s

Shear stress, Pa

1.0E+11

1.0E+10

1.0E+09

1.0E+08

1.0E+07

1.0E+06

1.0E+05

1.0E+04

1.0E+03

1.0E+02

1.0E+01

1.0E+00

Visc

osity

, mPa

-s

Shear stress, Pa

100 psi

10 100

700 psi

10 100

Viscosity at 55° FViscosity at 65° FViscosity at 80° F

> Rheological behavior of a live West African crude oil. Shear-stressversus viscosity measurements were performed near and below the pourpoint at 100, 300 and 700 psi [0.69, 2.07 and 4.83 MPa]. At 700 psi (top left),elevated low-shear-rate plateau viscosities at 55° and 65°F [12.8° and18.3°C] indicated the presence of gel structures. Little gelation occurrednear the pour point at 80°F [26.7°C]. Similar behavior occurred at 300 psi(top right); however, higher shear stresses were required to break the gelsat 55° and 65°F—behavior consistent with the loss of lighter hydrocarbonsin the sample. At 100 psi (bottom left), the 80°F sample exhibited strongshear thinning behavior.


shear-thinning behavior with no gelationoccurred at all pressures. Further investigationof gel strength involved aging live- and dead-oilsamples for 12 hours at the seabed temperature.The live-oil gel strength between 100 and 700 psivaried from 38 to 42 Pa, and the dead-oil gelstrength at ambient pressure was more thanthree times greater—142 Pa.

Dead-oil wax-deposition tests wereconducted in the WDFL at temperaturesbetween 81° and 122°F [27.2°C and 50.0°C](top). The results revealed a deposition-ratediscontinuity at the higher end of the shear-raterange. For each shear rate, chemists entered themeasured deposition rates, the n-paraffindistribution, C30+ composition and viscosity

profile into a standard equation-of-state modelthat calculates an n-paraffin diffusioncoefficient. The model assumes that moleculardiffusion of wax molecules is the principaldriving force governing wax deposition.29

The next step involved entering the flow-assurance measurements, wax-deposition dataand diffusion coefficients into the OLGAsimulator—a commercial multiphase-flow fluid-transport model—to assess and predict crude-oilbehavior in various production scenarios.Schlumberger engineers applied OLGA simula -tions to two cases: producing at a steady stateand restarting production after a shutdown.

Because the WAT of the West African crude oilwas about 50°F [27.7°C] higher than the pourpoint, wax deposition and gel formation were theprincipal flow-assurance risks during steady-state production. The operator’s design goal wasto prevent wax deposition at production ratesabove 5,000 bbl [795 m3] of oil per day. Methodsto prevent or slow wax deposition in a pipelineinclude using insulated pipe, installing heatersand injecting paraffin inhibitors into the crude-oil stream. OLGA simulations determined thatpipe insulation was the most suitable flow-assurance method (next page). Having both live-and dead-oil data proved important. Knowingonly the dead-oil WAT would have led theoperator to believe that expensive pipe-in-pipeinsulation was necessary. Calculations using thelower live-oil WAT showed that more economicalwet insulation would be sufficient to prevent wax deposition.

Because the oil cools as it flows through thepipeline to the collection facility, it was essentialto predict the likelihood of gel formation.Assuming a production rate of 5,000 bbl/dthrough wet insulated pipe, the OLGA simulatorshowed that live oil could flow through thepipeline for about 20 hours before reaching its pour point—sufficient time to reach the stock tank.

For flow rates less than 5,000 bbl/d, the OLGAsimulator predicted rates at which wax deposi -tion would occur in the pipeline. Withoutexperimental WDFL wax-deposition data,engineers would have to use the standard modelto estimate the n-paraffin diffusion coefficient,predict a deposition rate, and schedule wax-removal operations. For this West African crudeoil, the standard model predicted that remedi -ation would be necessary every two weeks. WithWDFL data, the simulator predicted a muchlower deposition rate, increasing the timebetween remedial jobs to six weeks. Methods toremove wax deposits include pipeline pigging,

68 Oilfield Review

>Wax-deposition behavior of a dead West African crude oil. Depositionmeasurements at two shear rates, 170 and 511 s–1, revealed unusual behavior.At the lower shear rate (red), deposition slowed steadily with increasingtemperature. However, when the fluid temperature exceeded about 100°F[37.8°C], the deposition rates at the higher shear rate (blue) suddenly surpassedthose at the lower shear rate. Chemists repeated the measurements to verifythe reproducibility of this behavior. The deposition rates were entered intoequation-of-state and fluid-transport models that help engineers predict howthe crude oil will behave in various production scenarios.

40

10

15

20

25

30

35

Depo

sitio

n ra

te, m

g/m

2 s

Temperature, °F70 80 90 100 110 120 130

170 s-1

511 s-1

> OLGA calculation of flowline-restart pressures in 6-in. pipe. If the oil gels inthe flowline during a shutdown, pressure must be applied to overcome the gelstrength and initiate flow. The maximum pressure that can safely be applied is500 psi. The simulation shows that maintaining an internal-flowline pressuregreater than 100 psi would allow engineers to safely restart the flowline.

2,000

0

250

500

750

1,000

1,250

1,500

1,750

Rest

art p

ress

ure,

psi

Fluid pressure, psi

0 100 200 300 400 500 600 700 800

Maximum restart pressure


Autumn 2007 69

hot-oil circulation and solvent treatments withcoiled tubing.

A vital flow-assurance question concerns theability of the system to restart after a productionshutdown. According to the operator, 500 psi wasthe highest pressure that could be safely appliedto overcome gel strength and initiate flow(previous page, bottom). Using live-oil gel-strength data and fluid-transport parameters,the OLGA simulator calculated that fluid flowcould be restored as long as the pressure in thepipeline remained above about 100 psi.

This case study shows that reliance on dead-oil experimental data and standard wax-deposition calculations could lead operators tomake unduly conservative decisions whendesigning production systems. In this case, flow-assurance predictions based on live-oil dataallowed the operator to save millions of dollars in flowline costs and less frequent wax-removal operations.

Coming Advances in Waxy-Crude Flow AssuranceSignificant work is underway to continueimproving flow-assurance testing and fluid-property surveillance during a field’s productivelife. The West Africa case study demonstrated the benefits of performing flow-assurance experi ments with live crude oils. However, wax-deposition testing in devices like the WDFL hasbeen confined to dead oils. The WDFL could bemodified to perform high-pressure tests, butconsuming two liters of live oil would beprohibitively expensive.

Oilphase-DBR scientists overcame thisproblem by designing and introducing theRealView live solids deposition cell.30 Requiringonly 0.04 galUS [150 mL] of oil, the cell canoperate at pressures up to 15,000 psi [103.4 MPa],temperatures up to 392°F and Reynolds numbersup to 500,000.31 Unlike the WDFL, the oil resides in a cylindrical vessel. A rotating spindleat the center induces fluid movement. Thedevice can simulate production conditions oftemperature, pressure, composition, pipe-surface roughness and both laminar andturbulent flow. Turbulent-flow testing is useful

29. Hayduk W and Minhas BS: “Correlations for Predictionof Molecular Diffusivities in Liquids,” Canadian Journalof Chemical Engineering 60, no. 2 (April 1982): 295–299.

30. Zougari M, Hammami A, Broze G and Fuex N: “Live OilsNovel Organic Solid Deposition and Control Device: Wax Deposition Validation,” paper SPE 93558, presentedat the 14th SPE Middle East Oil and Gas Show andConference, Bahrain, March 12–15, 2005.

31. In fluid mechanics, the Reynolds number is adimensionless ratio of inertial forces to viscous forces.Turbulent-flow conditions exist when the Reynoldsnumber exceeds 3,000.

> OLGA model simulations of West African crude-oil behavior in a flowlineduring steady-state production. Engineers used the simulations as a designtool to choose pipe insulation and avoid wax deposition during production. The temperatures at which oil would arrive at the collection station are plottedagainst production rate (top). Wet insulation (red) is more economical than apipe-in-pipe (PIP) enclosure (blue), but it is four times less efficient. Having thelive-oil WAT was fortunate because it showed that wet insulation would besufficient at production rates greater than 5,000 bbl/d. The next graph (center)reveals the cooldown profile as oil travels through the flowline. Using wetinsulation (red), the cooldown time to gel formation would either be 12 hours to the dead-oil pour point or 20 hours to the live-oil pour point. The latter timeperiod is sufficient to displace oil from the wellhead to the collection station.The OLGA simulator also predicted two-week wax-deposition profiles alongthe flowline (bottom). The wax-deposit thickness predicted by the standardmodel (blue) is nearly three times higher than the value calculated fromlaboratory wax-deposition tests in the WDFL (red). As a result, the frequency of wax-removal treatments could be reduced significantly.

160

140

120

80

60

40

100

160

150

140

120

110

100

130

Tem

pera

ture

, °F

Time, h

0 5 10 15 20 25

Arriv

al te

mpe

ratu

re, °

F

Oil production rate, bbl/d

0 2,500 5,000 7,500 10,000 12,500 15,000 17,500 20,000

STO WAT

STO pour point

Live-oil pour point

Live-oil WAT

1.2

1.0

0.8

0.4

0.2

0

0.6

Depo

sit t

hick

ness

, mm

Flowline distance, mi

0 0.5

Wellhead Collectionfacility

1.0 1.5 2.0

Recommended wax-removal thickness

PIP insulationWet insulation

PIP insulationWet insulation

Standard n-paraffindiffusion coefficient

WDFL n-paraffindiffusion coefficient


because it simulates the shear environment atthe flowline wall (above).

Thanks to live-oil testing in the RealView cell,significant advancements in asphaltene flow

assurance have already been realized.32

Preliminary laboratory results show that live-oiltesting will lead to similar improvements for waxycrude oils. A recent laboratory study tested thebehavior of a black waxy crude oil in a hypothet -ical flow-assurance situation. Oilphase-DBRscientists assumed the oil entered a 6-in. ID, 2-mi long pipeline at a temperature and pressureof 170°F and 3,170 psi [21.9 MPa]. The simulatedoil-production rate was 5,000 bbl/d, and theseabed temperature was 65°F [18.3°C]. Withinthis scenario, engineers performed deposition-rate tests that compared the performance of STO and live oils in laminar and turbulent flow(above right). The results showed that waxdeposition is slower in turbulent flow, and live-oilwax-deposition rates are far lower than thoseobserved with STO.

The low wax-deposition rates observed withlive oils would significantly impact the hypo -thetical flowline design and the frequency ofwax-removal operations. OLGA simulationsdemon strated that the 2-mi flowline could beconstructed from foam-insulated pipe instead ofpipe-in-pipe, potentially saving US $4 million. Inthe context of flowline remediation, traditionalWDFL measurements with STO indicated thatwax removal would be necessary every two weeks.Simulations with live-oil data predicted that wax-removal would be necessary about once per year,providing significant operational savings.

As a field produces, the crude-oil propertiesoften change. For example, as gas condensatefalls below the saturation pressure duringdepletion, the condensate yield and the WAT mayfall. In a compositionally graded accumulation,

composition may change as fluids are producedfrom regions that were originally distant from theinitial sampling point. The impact on flowassurance may be significant.

Traditionally, engineers monitor fluid-property evolution by periodically sampling fromseparators or extracting live oil near perfora -tions, and performing flow-assurance tests in thelaboratory. In offshore and deepwater fields, thisapproach is costly.33 With the advent ofintelligent completions, equipped with sensorsthat transmit downhole temperature, pressureand flow rates in real time, production moni -toring can be performed remotely. Thesecompletions also incorporate remote-controlpumps and valves that engineers can use tomitigate flow-assurance problems.

Chemical sensors are being developed thatcan detect fluid-composition changes. Wheninstalled at strategic locations in the wellcompletion and along a pipeline, the sensors willprovide real-time data for monitoring of solidsdeposition, corrosion rates and rheologicalproperties. As a result, interventions forsampling or remediation will be performed onlywhen necessary.

Today, fluid sampling and analysis areprogressing to a point at which consistentstandards are applied along the continuum fromDFA and reservoir characterization, to samplingand laboratory analysis, and on to productionsurveillance. This integrated approach will beincreasingly valuable to operators makingreservoir exploration, development andproduction decisions, particularly in high-risk,remote locations. —EBN

70 Oilfield Review

32. Akbarzadeh et al, reference 5.33. Amin A, Smedstad E and Riding M: “Role of Surveillance

in Improving Subsea Productivity,” paper SPE 90209,presented at the SPE Annual Technical Conference andExhibition, Houston, September 26–29, 2004.

> Illustration and schematic diagram of theRealView live solids deposition cell. The cell canachieve turbulent flow and wall-shear conditionsthat reflect those found in flowlines (top). Theinner spindle inside the cell rotates to createfluid movement (middle). Wall temperature andsystem pressure can be independentlycontrolled. The deposition surface androughness can be changed by inserting specialsleeves. The wax deposit appears on thestationary-cylinder surface (bottom).

>Wax-deposition behavior of waxy crude oil in the WDFL and RealViewapparatus. Engineers performed tests with dead and live oils. In laminar flow,dead-oil wax-deposition rates were essentially the same in the WDFL andthe RealView cell. Turbulent flow in the RealView cell reduced the dead-oildeposition rate substantially. Live-oil deposition rates in the RealViewapparatus were even lower for both flow regimes.

30

25

20

15

10

5

0

Tota

l wax

-dep

ositi

on ra

te, m

g/m

2 s

STO,laminar

STO,laminar

STO,turbulent

Live oil,laminar

Live oil,turbulent

24.5 23.7

15.5

9.2

3.3

WDFL RealView cellMultipoint

thermocouples

Shear cell

Variable-speedDC motor

High-pressureflow-controlvalves

Mechanicalmountingstand

Innerrotatingcylinder

Outerstationarycylinder

Deposit

Coolant

Coolantout

Charge

Discharge

Coolantin

Oil

Electricalheatingcartridge


Soraya Betancourt is a Research Engineer with theFluid Analysis group at Schlumberger-Doll Research inCambridge, Massachusetts, USA. She joinedSchlumberger in 2000 as a reservoir engineer at theSchlumberger Reservoir Completions Center,Rosharon, Texas, USA. Soraya previously was employedby Petróleos de Venezuela SA (PDVSA) Oil & Gas inVenezuela as a production technologist. She has a BSdegree from Universidad del Zulia in Maracaibo,Venezuela, and an MS degree from the University ofTulsa, both in petroleum engineering.

John Blackburn joined ConocoPhillips six years agoafter obtaining his MS degree in geophysics from theUniversity of Leeds and a BS degree (Hons) in geo-physics from the University of Edinburgh, Scotland .Currently he is a Geophysicist working the J-Blockarea of the UK North Sea for ConocoPhillips UK Ltd.John has worked on normally pressured and high-pres-sure, high-temperature exploration and developmentopportunities throughout the UK continental slope.

Tim Bunting, Geophysics Manager for WesternGecoAsia and Australia, is based in Kuala Lumpur, wherehe manages the scientific and technical validity ofregional seismic solutions. Hired in 1990 as a field geo-physicist, he worked offshore for five years. Tim hasfilled various geophysical support roles in Europe,Africa, the Middle East and Russia. He was the globalacquisition operations support manager before assum-ing his current management role in 2005. He receivedhis BS degree in mining engineering fromStaffordshire University, England.

Chris Chapman, a Consultant in Seismology atSchlumberger Cambridge Research (SCR) in England,is also an honorary professor of theoretical seismologyat University of Cambridge. Following an extendedcareer as a physics and geophysics educator at theUniversity of Alberta and University of Toronto inCanada, and University of Cambridge in England, hebecame scientific advisor at SCR in 1991. Chrisauthored The Fundamentals of Seismic WavePropagation, a textbook published in 2004 by theCambridge University Press. An Addendum to the bookdescribes his firsthand experience during the tsunamion the coast of Sri Lanka. Chris holds a BA degree intheoretical physics and a PhD degree in geophysicsfrom University of Cambridge.

Phil Christie began his career with Schlumberger in 1972, as a wireline engineer in West Africa.Returning to the UK in 1975, he completed a PhDdegree and postdoctoral work in seismology at theUniversity of Cambridge, England, and rejoinedSchlumberger in 1981 to develop the borehole seis-mic business in Europe. During his career he hasestablished or led seismic departments atSchlumberger research and development centers inConnecticut, Paris, and Cambridge, England. From1996 to 1997, he was seconded to the BP AtlanticMargin group, where he jointly coordinated theseabed reservoir monitoring experiment in Foinaven.Phil then led a WesternGeco Reservoir Geophysicsgroup in Gatwick, where he also helped establish the

Schlumberger Geophysics Technical Community. In2000, Phil returned to Schlumberger CambridgeResearch as a Scientific Advisor, with interests inreservoir geophysics, time-lapse seismic and sub-basalt imaging. He is a coeditor of PetroleumGeoscience and is vice president of the EAGE.

John Cook is a member of the Drilling, Telemetryand Control department at Schlumberger CambridgeResearch in England, where he works on wellboreinstability control, sand management, perforatingstrategies and improvements to the drilling process.John is a graduate of the University of Cambridgewith a BA degree in materials science and a PhDdegree in physics.

John Daniels, Schlumberger DESC* Design andEvaluation Services for Clients Engineer, is involved inseveral projects for the Central Division of DevonEnergy, particularly development of the Barnett Shalegas reservoir. Based in Oklahoma City, Oklahoma, USA,he serves the Devon Oilfield Services (OFS) account byintegrating OFS technologies; he also specializes inreservoir optimization of unconventional gas plays uti-lizing logging tools and seismic, microseismic and frac-turing techniques. He joined Schlumberger in 2001 asWell Production Services (WPS) engineer and workedon fracturing and coiled tubing operations in SaudiArabia and Bahrain. He pumped the first fracturingjobs utilizing VDA* Viscoelastic Diverting Acid andClearFRAC* polymer-free frac fluid technologies. Hemoved to the position of WPS cell leader in 2003 andprovided logistical and technical support to districtoperations. He also served as the US Land East prod-uct champion for new multilayer efficient fracturingtechnologies, which he introduced in 2005. John has aBS degree in materials science and engineering fromthe University of Washington, Seattle, USA.

Tara Davies is Schlumberger Fluid Analysis ProductChampion at the Oilphase-DBR* Technology Center inEdmonton, Alberta, Canada. She is responsible forRealView* live solids deposition cell and flow assur-ance simulation services. Hired in 2001 as a softwaresupport engineer, she has worked as a product and ser-vice engineer in well testing at the Oilphase-DBR fluidsampling and analysis facility, where she focused onPVT equipment and software sales and support. Taraearned a BS degree in petroleum engineering from theUniversity of Alberta, Canada.

Scott Dingwall, Reservoir Evaluation Wireline DomainGeophysicist for the North Sea GeoMarket* region, isbased in Aberdeen. He came to this position in 2004from the Schlumberger Data & Consulting Services(DCS) group in Stavanger, where he was senior geo-physicist. Among his responsibilities are technical andgeophysical support for Schlumberger vertical seismicprofile (VSP) operations across the North Sea region.He began his career in offshore surface seismic pro-cessing before joining Schlumberger in 1996. Hestarted as a geophysicist in the borehole seismic pro-cessing group in Aberdeen, and later transferred tothe borehole seismic group in London. Scott is a grad-uate of Imperial College in London and holds an MSdegree in exploration geophysics.

Chengli Dong is a Principal Reservoir Engineer forSchlumberger in Sugar Land, Texas, where he workson formation testing and sampling, especially develop-ment of the MDT* Modular Formation DynamicsTester, LFA* Live Fluid Analyzer for MDT tool, CFA*Composition Fluid Analyzer and a new generation offluid analyzer interpretation algorithms. The author ofmany technical papers, Chengli obtained MS and PhDdegrees in petroleum engineering from The Universityof Texas at Austin.

Hani Elshahawi is Senior Staff Petrophysicist andGlobal Fluid Evaluation and Sampling Advisor withShell International Exploration and Production inHouston. His current focus is the planning, executionand analysis of global high-profile formation testingand downhole fluid-sampling operations. With morethan 20 years of experience in the oil industry, he hasworked in both service and operating companies inmore than 10 countries in Africa, Asia, the MiddleEast and North America and has held various posi-tions in interpretation, consulting, operations, mar-keting and product development. Hani has lecturedwidely in various areas of petrophysics, geosciencesand petroleum engineering, holds several patents, andhas written more than 50 technical papers. He has aBS degree in mechanical engineering and an MSdegree in petroleum engineering from The Universityof Texas at Austin.

René A. Frederiksen, Hess Production OptimizationTeam Leader, is based in Copenhagen, Denmark.Previously he was with Maersk Oil for eight years.Before joining Hess in 2006, he led a team of reservoirand petroleum engineers on a 100,000-bbl/d chalk-fielddevelopment. He has published articles and papersabout complex horizontal wells, completion stimula-tion effects on well productivity, scale managementand wellbore failures during injection and production.René received an MS degree in chemical engineeringfrom the Technical University of Denmark inCopenhagen, and earned an EBA (European BusinessAdministration) degree while working for Maersk Oil.

Sidney Green is Manager of Geomechanics BusinessDevelopment for the Schlumberger Data & ConsultingServices Group in Salt Lake City, Utah, USA. One ofthe founders and the retired President-Chairman-Chief Executive Officer of TerraTek, Inc., acquired bySchlumberger in 2006, he has worked in rock mechan-ics for the past four decades. He received many hon-ors including the Outstanding Engineer award for theState of Utah, Entrepreneur of the Year from theMountain West Venture Group, and the HonoraryAlumni Award from the University of Utah. He haspublished numerous papers and reports, holds a num-ber of patents and has given many presentations ongeomechanics. He is also a Research Professor at theUniversity of Utah in Salt Lake City, where he holds a

Autumn 2007 71

Contributors

60336schD8R1.qxp:ORSPR04_contribs_01 12/5/07 7:48 PM Page 71

dual appointment in mechanical engineering andcivil and environmental engineering. He has servedon a number of corporate boards, government com-mittees and university advisory boards and is the pastchairman of the National Academy of Sciences, USNational Committee on Rock Mechanics. Sidney has aBS degree from the University of Missouri at Rolla,USA, and an MS degree from the University ofPittsburgh, Pennsylvania, USA, both in mechanicalengineering. He also holds the Degree of Engineer inengineering mechanics from Stanford University inCalifornia, USA. He is a member of the US NationalAcademy of Engineers.

Geoffrey Hampden-Smith, based in Aberdeen, pro-vides a support role for Shell’s Drilling and WellServices project teams. He began his career withSchlumberger in 1971 as a wireline engineer workingmainly in Africa and Europe. In 1983, he helped man-age the introduction of tubing-conveyed perforating(TCP) operations in the North Sea. Five years later heset up his own company to develop project and risk-management software and provide project manage-ment support services to the oil industry. He has beeninvolved with Shell since 1998 and has supporteddiverse projects such as creating well services qualityplans, platform gas lift feasibility demonstrations,developing temporary pipework standards and down-hole gauge reliability studies. Geoffrey earned a BSdegree (Hons) in physics from Edinburgh Universityin Scotland, and an MS degree in risk and reliabilitymethods from Loughborough University in England.

Klaus Hasbo recently joined NORECO ASA in Holte,Denmark, where he works as Siri Fairway AssetManager (Southern North Sea). In 1997, after earningan MEng degree in petroleum engineering fromHeriot-Watt University in Edinburgh, Scotland, hejoined DONG Energy, Denmark, as a reservoir engi-neer working on the South Arne field. In 2002, Klauswas seconded to Hess Denmark ApS to serve as assetmanager for the South Arne field. From 2003 to 2005,he worked as subsurface manager for the Siri field.

Jorg Herwanger, Senior Geophysicist at theWesternGeco Houston Technology Center, specializesin the analysis of field seismic data for stress-inducedtime-lapse effects. His main focus is on closer inte-gration of time-lapse seismic imaging with reservoirmodeling and reservoir geomechanics. Jorg began hiscareer with WesternGeco in 2003 in Gatwick,England, as a Marie Curie Research Fellow. He wassenior geophysicist at Gatwick before moving toHouston in 2006. He received a PhD degree in geo-physics from Imperial College, London, and an MSdegree in geophysics from Technische UniversitätClausthal, Germany.

Patrick Hooyman is Houston Geomechanics Managerfor Schlumberger Data & Consulting Services, a posi-tion he assumed in 2002. He has more than 20 yearsof international experience in geophysics, integratedstudies and geomechanics. He began his career as ageophysicist with Amoco in Denver, where he partici-pated in the discovery of several significant US oiland gas fields including Whitney Canyon. Before join-ing Schlumberger in 1995, he held technicalmanagement positions in Compagnie Générale deGéophysique (CGG) as manager of advanced services,in GECO Geophysical in Houston as vice presidentand chief geophysicist, and with GeotraceTechnologies as vice president. He has taught acourse on the introduction to seismic interpretationfor the SEG for 12 years. Patrick has a BS degree inphysics from Benedictine University in Lisle, Illinois,USA, and a PhD degree in physics from the Universityof Wyoming at Laramie, USA. He is a licensed profes-sional geoscientist in Texas.

Arnis Judzis has been with TerraTek, Inc. in SaltLake City, since 1999. As Vice President for Data &Consulting Services and General Manager of opera-tions, he is responsible for business development andglobalization of the TerraTek* GeomechanicsLaboratory Center of Excellence. He joined TerraTekas vice president, Drilling and Completions, andbecame executive vice president. Before joining thecompany, Arnis spent 23 years in management posi-tions with BP in Dallas, London, Houston andAnchorage. In 1979, after receiving his PhD degree inchemical engineering from the University of Michigan,Ann Arbor, USA, he joined Standard Oil Company ofOhio, which was later absorbed by BP. He serves onthe SPE Research and Development AdvisoryCommittee to the Board of Directors, and is also onthe Board of Directors for the Offshore TechnologyConference Inc., serving as its Chairman for the pasttwo years. He holds an MS degree from the Universityof Michigan and a BS degree from Cornell University,Ithaca, New York, USA.

Ray Kennedy is Marketing Engineer for theSchlumberger Oilphase-DBR Technology Center atEdmonton, Alberta. He has worked in equipment salesand business development positions since joiningDBR in 1994. As project engineering team managerfor Schlumberger Oilphase-DBR fluid sampling andanalysis service, he coordinated a global group of 12individuals within the fluid sampling and analysisbusiness at laboratory facilities in Houston, Dubai andAberdeen. Ray also served as product champion forthe PVT ReCORD* data management and delivery sys-tem and was responsible for its implementation andcommercialization. He received his BS degree inchemical engineering from the University of Alberta.

Nick Koutsabeloulis is Vice President and GeneralManager of the Schlumberger ReservoirGeomechanics Center of Excellence at Bracknell,England. This facility integrates numerical geome-chanical solutions with those of reservoir engineering,seismic inversion, 4D seismic, microseismicity, well-bore integrity and completions design. From 1993 to2007, he was managing director of V.I.P.S. (VectorInternational Processing Systems) Limited atBracknell, where he developed the VISAGE* system, asuite of software products for coupled geomechanicalsolutions using the finite-element method and linking

industry standard reservoir simulators to state-of-the-art geomechanical solvers. Nick began his career in1985 as a geomechanics developer at D’AppoloniaS.p.A. in Genoa, Italy. From 1986 to 1993, he was asenior stress analyst at BP International, Sunbury-on-Thames, England, where he developed theGeomechanics Department for Mining and OilApplications. He has a PhD degree in civil engineeringfrom Manchester Metropolitan University, England,and a BS degree in civil engineering from AristotleUniversity of Thessaloniki, Greece.

Scott Leaney, based in Houston since 2002, isGeophysics Advisor for Schlumberger Data &Consulting Services. He specializes in the processingand inversion of three-component borehole seismicdata, anisotropy analysis and seismic integration.From 1988 to 1992, he was a geophysical softwaredeveloper at Schlumberger in Clamart, France. Hesubsequently transferred to Jakarta, where he was theborehole geophysicist for south and east Asia. From1998 to 2002, he was based in Gatwick, England,where he worked on developing borehole and surfaceseismic integrated processing techniques. Scott holdsdegrees in geophysics: a BS degree from theUniversity of Manitoba, Winnipeg, Canada, and an MSdegree from the University of British Columbia,Vancouver, Canada.

Joël Le Calvez, based in Houston, is a SchlumbergerSenior Geologist, recently promoted to lead the newlycreated microseismic cell in Dallas, effectiveDecember 2007. He will continue working on develop-ment and commercialization of the microseismic busi-ness and will lead a team of geophysicists andgeologists developing business in the Barnett Shaleand other unconventional plays. His main responsibil-ities are the processing and interpretation of data forgeological, geophysical and geomechanical applica-tions, and client presentations. He also works withproduct centers on defining programs and testingsoftware, and with research centers on defining andtesting of algorithms. Joël joined Schlumberger in2001 after acquiring his PhD degree in geology fromThe University of Texas at Austin. He has sinceworked on geological and seismic studies throughoutthe southwestern USA and offshore Angola. He holds aDiplome d’Etudes Approfondies in tectonophysics andcondensed matter from Université Pierre et MarieCurie in Paris, an MS degree in geology and geo-physics from the Université de Nice-Sophia Antipolis,and a BS degree in mathematics and physics from theUniversité de Nice, France.

Don Lee is a Principal Geoscientist withSchlumberger Data & Consulting Services in Houston.His work involves processing and interpreting infor-mation relating to formation mechanical properties,wellbore instability and real-time drilling data for pro-jects worldwide. After earning a BS degree in electri-cal engineering from Tennessee TechnologicalUniversity in Cookeville, USA, he joined Schlumbergerin 1980 as a field engineer in Texas. His subsequentpositions included special services engineer, log ana-lyst, senior log analyst, application development engi-neer, senior interpretation application engineer anddata center manager.

72 Oilfield Review


Rob Marsden, Schlumberger Reservoir GeomechanicsAdvisor, is based at the Reservoir GeomechanicsCenter of Excellence in Bracknell, England. Previouslyhe was geomechanics manager for Middle East andAsia-Pacific activities. He joined Schlumberger in2000, after spending 10 years as senior lecturer andhead of the Rock Mechanics Laboratories andWellbore Mechanics Research Group at ImperialCollege in London. Since receiving a degree in civilengineering from Sunderland Polytechnic in England,and MS and DIC degrees in engineering rock mechan-ics from Imperial College, Rob has had nearly 25 yearsof consulting, field, research and teaching experiencein petroleum rock mechanics. A chartered engineer,he has published numerous papers and has served onmany international and industry committees.

J. Wesley (Wes) Martin has more than 20 years ofrock mechanics laboratory experience in both com-mercial and government programs. He is DisciplineManager, Geomechanics, at the TerraTek facility inSalt Lake City, Utah, where he manages the RockMechanics Laboratory and works closely with thePetrology Group on all geomechanics-related pro-grams. He also manages a team of individuals withdiverse backgrounds in mining, electrical and mechan-ical engineering, geology and petrology. Wes began hiscareer with TerraTek Research in 1985 as a geologicaltechnician. From 2000 to 2005, as division head,Geomechanics Division, he managed all laboratoryprojects and personnel for the rock mechanics group.The following year he became vice president and divi-sion head, responsible for Geomechanics Division per-formance and profitability. Before joining TerraTek,Wes was a geologist with American InternationalResources, Inc. at Ely, Nevada, USA, and with theForestry Sciences Laboratory, Oregon State Universityat Corvallis, USA. He received his MBA degree fromthe University of Phoenix, Salt Lake City, and his BSdegree in geology from Humboldt State University,Arcata, California.

Henry Menkiti, Schlumberger Wireline (WL)Headquarters Domain Champion for Geophysics inLondon, oversees the WL global borehole seismic busi-ness. He spent five years as a field engineer inVenezuela, Nigeria, Canada and Brazil. In 1997, hetransferred to the Land Technical Support group inHannover, Germany, and later became a geophysicaland data processing instructor in Gatwick, England.He was then transferred to Oilfield Services (OFS) per-sonnel where he was responsible for OFS trainingwithin North and South America. He subsequentlybecame the Belle Chasse, Louisiana, wireline field ser-vice manager and then became the domain championfor the Gulf of Mexico. Before accepting his currentposition, Henry managed all Schlumberger WirelineSpecial Services Operations in the North SeaGeoMarket and Europe, Caspian and Africa area. Heearned an MS degree in petroleum geology fromImperial College, London.

Oliver C. Mullins is Schlumberger Scientific Advisorand Wireline Headquarters Reservoir DomainChampion in Houston. He is the originator ofDownhole Fluid Analysis (DFA), a significant new ser-vice in the oil industry, for which he was awardedthree gold medals, two from Schlumberger and onefrom the state of Connecticut. DFA acceptance by theindustry is reflected in his previously being namedSPWLA Distinguished Lecturer and SPE DistinguishedLecturer in 2007. The corresponding tool projects,based on near-infrared and fluorescence spectroscopy,are being used to uncover compartmentalization andhydrocarbon fluid complexities in subsurface forma-tions. Oliver also leads an active research group inasphaltene and petroleum science. He has coeditedthree books on asphaltenes, and coauthored nine bookchapters. He has won numerous awards and is coau-thor of one of the most cited papers in petroleum sci-ence. He has published 70 articles in refereedscientific journals, 35 in oilfield journals, and hascoauthored 40 allowed US patents. He has a BS degreein biology from Beloit College in Wisconsin, USA, andMS and PhD degrees in chemistry from Carnegie-Mellon University, Pittsburgh, Pennsylvania.

John Nighswander is Technical Manager andSchlumberger Advisor in Reservoir Fluid Sampling andAnalysis, based in Houston. He began his career in1989 with DB Robinson and Associates Ltd atEdmonton, Canada. Before joining Schlumberger in1998, he held various positions with the DB Robinsoncompanies including that of president, DBR FluidProperties, Inc. in Houston. During his career withSchlumberger, he has held several management posi-tions in Aberdeen and Houston. John received BS andPhD degrees in chemical engineering from theUniversity of Calgary.

Sheila Noeth is Principal Geomechanics Specialist inthe Geomechanics Group, Data & Consulting Services(DCS), Schlumberger Oilfield Services, Houston. As amember of the DCS geomechanics group, she isinvolved in 3D pore-pressure prediction studies andreal-time well monitoring. After completing her MSand PhD studies in geology at the Ruhr-UniversitätBochum, Germany, in 1991, she became a consultantat Institut für Sicherheit und Umweltschutz, an engi-neering geology company for environmental geology inDortmund, Germany. During her career she has alsobeen an assistant professor focusing on basin model-ing, petroleum geology and sedimentology at theDepartment of Geology, University of Cologne,Germany, and a visiting scholar and visiting assistantprofessor of geology and geophysics, at Texas A&MUniversity, College Station.

Les Nutt, who is based in Houston, has beenSchlumberger Borehole Seismic Operations Manager,North America, since 2004. He began his career as anarea geophysicist with Geophysical Services Inc. in theUK and Saudi Arabia. He joined Schlumberger in Parisin 1981 and then worked as a log analyst and geophysi-cist in the Far East and in Europe. In 1991, he movedto Norway as Wireline and Testing marketing manager.In 1995, he joined the Schlumberger InterpretationDevelopment team in Paris and Houston before trans-ferring to the Schlumberger Engineering Center inJapan as marketing manager. He moved to Houston in2002 as the geophysics domain manager. He obtained aBS degree (Hons) in pure and applied physics and aPhD degree in physics from Queen’s University Belfast,Northern Ireland.

Michael O’Keefe is Schlumberger Product Championfor Downhole Fluid Analysis based in Hobart,Tasmania, Australia, a position he has held since 2006.He joined Schlumberger in 1990 as a wireline fieldengineer in Austria. Since 1991, he has had assign-ments in Norway, Saudi Arabia and Scandinavia in hisprevious position as senior reservoir engineer. Authorof many technical papers, Michael is a recipient of theQuicksilver Probe* Gold Medal 2006 and a member ofthe Quicksilver Probe development team who also wonthe Hart’s Meritorious Engineering Award at theOffshore Technology Conference in 2006. He earned aBEng degree (Hons) in electronic engineering fromthe University of Tasmania, Australia.

Adrian Sanchez, based in Villahermosa, Tabasco,Mexico, is Borehole Seismic Domain Champion forMexico and Central America. He is in charge of intro-ducing new technologies, supporting operations anddeveloping the market in borehole seismic applica-tions. Adrian was senior geophysicist and domainchampion for Latin America-South overseeing market-ing and processing until August 2007, when he assumedhis current post. He began his career in 1994 as a sup-port geophysicist with Western Geophysical in seismicoperations and data processing. He was geophysicistand senior geophysicist for Schlumberger GeoQuestVenezuela Oilfield Services (OFS) from 1997 to 2002.Adrian holds a BS degree in geophysics from theUniversidad Simón Bolívar in Caracas.

Colin Sayers is a Scientific Advisor in theSchlumberger Geomechanics Group in Houston, pro-viding consultancy in pore-pressure prediction, well-bore-stability analysis, geomechanics, rock physics,geophysics and the properties of fractured reservoirs.He has a BA degree in physics from the University ofLancaster, England, and a PhD degree in theoreticalsolid-state physics from Imperial College, London. Heis a member of the AGU, ARMA, EAGE, SEG and SPE,and is a member of the Research Committee of theSEG and of the editorial boards of The Leading Edge,Geophysical Prospecting and International Journalof Rock Mechanics and Mining Sciences.

Autumn 2007 73


Marco Schinelli joined Petróleo Brasileiro S.A.(Petrobas) in 1980 after receiving his BS degree ingeology from Federal University of Bahia, Salvador,Brazil. After specialization courses and a short periodin seismic acquisition, he worked 14 years in land andmarine seismic processing. Since 1994, he has beenfocused on seismic interpretation, supporting assetteams at the Petrobras regional office in Bahia onseismic characterization of reservoirs. His main areasof interest are borehole and 4D seismic techniques,and advanced techniques of seismic interpretation.He earned a professional MS degree at Petrobas.

Satish C. Singh is a Professor at the Insitut dePhysique du Globe de Paris (IPGP) and PrincipalResearch Fellow at the University of Cambridge,England. He has written and cowritten more than 90papers in international journals, including seven inNature and Science. He created and directs LITHOS,a consortium of oil and service companies, and isDirector of the Laboratoire Geosciences Marine atIPGP. He led the creation of the French Ocean-Bottom Seismometer (OBS) Pool, which he heads,and is the project leader of the international Sumatra-Andaman Great Earthquake Research (SAGER) pro-ject, involving more than 50 scientists and 16international institutions. He is also the coordinatorof NERIES, the European Broadband SeismometerNetwork, and has been chief scientist aboard numer-ous scientific cruises, most recently the R/V MarionDufresne and WesternGeco Geco Searcher, both inSumatra. He is a graduate of the University of Toronto,Canada, with a PhD degree in theoretical seismology.

Jim Sledzik, Global Marketing Director forWesternGeco, based in London, is responsible for for-mulating and implementing the company’s strategicdirection. He began his career with WesternGeophysical in 1987 as assistant crew manager andaccountant in Nigeria and subsequently became crewmanager in the UAE, and field supervisor in Nigeriaand Argentina. He left the company for a year andreturned in 1996 to work in several management posi-tions in Venezuela, Nigeria, Mozambique andTurkmenistan until 2000, when he became countrymanager for WesternGeco in Argentina. Jim was thegeneral manager of Multiclient Services for WesternGeco, Houston, until 2003, when he was assigned tothe Schlumberger Corporate Strategy Initiative. Hewas Oilfield Services global account director atHouston until his move to England in 2006. Jim holdsan MBA degree in international business from theJoseph M. Katz Graduate School of Business,University of Pittsburgh, Pennsylvania, and has a BSdegree in geosciences from The Pennsylvania StateUniversity in University Park.

Morten G. Stage is Research and DevelopmentCoordinator for DONG Exploration and Production(E&P), Hørsholm, Denmark. Previously, he served asgeomechanics specialist and senior petrophysicist inDONG E&P. Before joining DONG in 2003, he hadbeen a senior project manager at the DanishGeotechnical Institute (GEO). Morten has more than10 years of industry experience and has publishedmany papers related to rock physics and geomechan-ics. He has an MS degree in physics from theUniversity of Odense, Denmark, and a PhD degree inrock physics from Chalmers Technical University,Göteborg, Sweden.

Roberto Suarez-Rivera is a Discipline Manager andHead of the Stimulation and Production Division atTerraTek in Salt Lake City. He is investigating theimpact of heterogeneity and material anisotropy onwellbore stability, productivity, in-situ stress and com-pletion design in tight gas-shale reservoirs. He haseight years of experience as a field service engineerfor Dowell Schlumberger. He has also worked as a sci-entist and consultant in petroleum-related rockmechanics with the Norwegian Institute of RockMechanics (IKU), and with the Lawrence BerkeleyNational Laboratory in California. Roberto received aPhD degree in rock mechanics from the University ofCalifornia, Berkeley.

Chee Phuat Tan is Schlumberger Middle East andAsia (MEA) Geomechanics Advisor. Based at theKuala Lumpur Deepwater Technology Hub, he pro-vides key technical input for geomechanics work-flow, software and solution development related toData & Consulting Services (DCS). He actively intro-duced new technologies such as the Sonic Scanner*acoustic scanning platform. His work crosses severalcompanies in Asia-Pacific, Middle East, Africa andthe Caspian. He joined Schlumberger DCS at Perth,Australia, in 2005 as geomechanics domain cham-pion and Asia-Pacific geomechanics coordinator ofthe Kuala Lumpur Technology Hub. Before joiningSchlumberger, from 1987 to 2005, he was withCSIRO Petroleum in Melbourne and Perth,Australia, where he held various positions includingbusiness opportunity manager and group leader.Chee earned a PhD degree in rock mechanics and aBS degree in civil engineering from MonashUniversity, Melbourne, Australia.

Stephen Willson, the Rock Mechanics Advisor in theBP Drilling and Completions Technology Unit inHouston, has more than 20 years of experience inpetroleum geomechanics. His current focus is on well-bore stability, salt mechanics, and compaction andsubsidence, including geomechanical well-integritychallenges facing BP developments in the deepwaterGulf of Mexico. Since joining BP in 1988, he has heldvarious research and technology development posi-tions in both Sunbury, England, and Houston. He alsoserved as completions manager for TerraTek, Inc., inSalt Lake City, from 1992 to 1995. Stephen is a civilengineer with a PhD degree in soil mechanics fromthe University of Manchester, England.

Oilfield Review74

An asterisk (*) is used to denote a mark of Schlumberger.


Coming in Oilfield Review

Intelligent Wells. The reliability ofwells that can be remotely moni-tored and controlled has improvedsteadily over the past decade. As aresult, the number of such wells isexpected to increase fivefold in thenext five years. This article dis-cusses the evolution of intelligentwells—from strategies to avoidhigh-cost, high-risk interventions tothe current, powerful reservoir-man-agement tools that can be applied inwells of all economic strata.

The Big Picture. The myriad datagenerated by a digital oil field canoverwhelm a production department.Some operators are now employingautomated workflows to make bestuse of their data. This articledescribes how real-timemeasurements, automated workflowsand analytical models can be broughttogether into a collaborativeenvironment to help operatorsdiagnose production problems andformulate timely responses.

Optimizing Asset Production.Oilfield processes reach new stagesof automation continually. However,each element of a field operationhas different software, which hasmade system optimization difficult.Now, innovative software enablescommunication and feedbackbetween disparate parts of the oper-ation, allowing integrated assetmodeling. This article describes howthe software facilitates communica-tion between the elements to opti-mize the whole asset, whether thesystem is constrained by formationproductivity or by facilities capacity.

NEW BOOKS

Why Beauty Is Truth: A History of SymmetryIan StewartPerseus Books Group1094 Flex DriveJackson, Tennessee 38301 USA2007. 290 pages. $26.95ISBN: 0-465-08236-X

Mathematician Ian Stewart discussesthe concept of symmetry, which is atthe heart of relativity theory, quantummechanics, string theory and much ofmodern cosmology. Stewart provides ahistorical view by introducing us toimportant people in science and mathematics. He also delves into thenumerology world of real mathematics,where particular numbers have uniqueand unpredictable properties related tosymmetry. Finally, Stewart goes beyondsuperstrings, to the “octonionic” symmetries that may explain the veryexistence of the universe.

Contents:

• The Scribes of Babylon

• The Household Name

• The Persian Pet

• The Gambling Scholar

• The Cunning Fox

• The Frustrated Doctor and the Sickly Genius

• The Luckless Revolutionary

• The Mediocre Engineer and theTranscendent Professor

• The Drunken Vandal

• The Would-Be Soldier and theWeakly Bookworm

• The Clerk from the Patent Office

• A Quantum Quintet

• The Five-Dimensional Man

• The Political Journalist

• A Muddle of Mathematicians

• Seekers after Truth and Beauty

• Further Reading, Index

75Autumn 2007

If it were just an authentic historyof mathematics, it would be creditable.If it were only for its lively informalstyle, its historical characters, itsintrigue…, its beautiful prose, it wouldbe praiseworthy. Yet, its realuniqueness — its power — is in what ituncovers. It brings us the heart of whymathematicians pursue mathematics.

Why Beauty Is Truth is a brilliantinterweaving of politics, history andintrigue, with characters living ordinarylives, described in the spirit of aRussian novel. With one storythreading into another, the book movesus forward.Mazur J: Nature 447, no. 7140 (May 2007): 38.

Through this historical section,Stewart skillfully interweaves the mathwith colorful sketches of the lives ofthe mathematicians involved.Gardner M: “Is Beauty Truth and Truth Beauty?”

http://www.sciam.com/article.cfm?chanID=

sa006&articleID=5B8E1AAE-E7F2-99DF-

31FF9E4F79068FBE&pageNumber=1&catID=2

(accessed May 4, 2007).

The Trouble With Physics: The Rise of String Theory, the Fall of a Science, and What Comes NextLee SmolinHoughton Mifflin222 Berkeley StreetBoston, Massachusetts 02116 USA2006. 416 pages. $26.00ISBN 0-61855-1050

Renowned theoretical physicist LeeSmolin argues that physics—the basisfor all other science—has lost its way.The problem is string theory, anambitious attempt to formulate a theoryof everything that explains all the forcesand particles of nature and how theuniverse came to be. Smolin believesthat physicists are making the mistakeof searching for a theory that is“beautiful” and “elegant” instead of onethat’s actually backed up byexperiments. He encourages physiciststo investigate new alternatives andhighlights several young theoristswhose work he finds promising.

Contents:

• The Unfinished Revolution: The Five Great Problems in TheoreticalPhysics; The Beauty Myth; TheWorld as Geometry; UnificationBecomes a Science; FromUnification to Superunification;Quantum Gravity: The Fork in the Road

• A Brief History of String Theory:Preparing for a Revolution; The FirstSuperstring Revolution; RevolutionNumber Two; A Theory ofAnything; The Anthropic Solution;What String Theory Explains

• Beyond String Theory: Surprisesfrom the Real World; Building onEinstein; Physics After StringTheory

• Learning from Experience: How DoYou Fight Sociology?; What IsScience?; Seers and Craftspeople;How Science Really Works; WhatWe Can Do for Science

• Notes, Index

Smolin begins with an excellentpresentation of the foundations offundamental physics, laying the groundfor an understanding of the roots ofboth the present aims of string theoryand its problems.

Smolin crystallizes what many inthe physics community feel about theseextravagances of string theory.Ellis G: Nature 443, no. 7111

(October 5, 2006): 507–508.

The real conflict is physicist vs.physicist. It is a human story, and anold one, involving hubris, courage, andthe inertia of communal thought.Doerr A: http://www.boston.com/ae/books/

articles/2006/09/17/resisting_the_supremacy_

of_string_theory/ (accessed October 12, 2006).


A Different Universe:Reinventing Physics from the Bottom DownRobert B. LaughlinBasic Books11 Cambridge CenterCambridge, Massachusetts 02142 USA2005. 254 pages. $26.00 hardcover;$16.00 paperbackISBN 0-465-03828-X

Written by a Nobel Prize-winningphysicist, this book proposes a changein thinking on the fundamental laws ofphysics. The author argues that insteadof looking for ultimate theories, weshould consider the world of emergentproperties, such as crystal hardness andshape, that result from the organizationof large numbers of atoms.

Contents:• Frontier Law• Living with Uncertainty• Mount Newton• Water, Ice, and Vapor• Schrödinger’s Cat• The Quantum Computer• Vin Klitzing• I Solved It at Dinner• The Nuclear Family• The Fabric of Space-Time• Carnival of the Baubles• The Dark Side of Protection• Principles of Life• Star Warriors• Picnic Table in the Sun• The Emergent Age• Notes, Index

In this fascinating book, inter -spersed with witty lines and anecdotes,physicist Laughlin (Stanford Univ.)treats readers to a variety of naturalphenomena whose full understandingcalls for a revolutionary mindset(paradigm shift) in science’s approachto the phenomenal world.

But Laughlin speaks with deepunderstanding and insight, andcontributes to, if not charts, new roadsin science’s exploration of the phenom -enal world. A very enlightening,revelatory, and entertaining book…Highly recommended.

Raman VV: Choice 43, no. 3 (November 2005): 525–526.

Theoretical Optics: An IntroductionHartmann RömerWiley-VCH Verlag GmbH & Co. KgaABoschstraße 12D-69469 Weinheim, Germany2005. 361 pages. $160.00ISBN 3-527-40429-5

Starting from basic electrodynamics, this book provides an introduction totheoretical optics, containing topics such as nonlinear optics, light-matterinteraction and modern topics inquantum optics, including entangle ment,cryptography and quantum computation.

Contents:• A Short Survey of the History of

Optics• The Electrodynamics of

Continuous Media• Linear Waves in Homogeneous

Media• Crystal Optics• Electro-, Magneto-, and

Elastooptical Phenomena• Foundations of Nonlinear Optics• Short-Wave Asymptotics• Geometrical Optics• Geometric Theory of Caustics• Diffraction Theory• Holography• Coherence Theory• Quantum States of the

Electromagnetic Field• Detection of Radiation Fields• Interaction of Radiation and Matter• Quantum Optics and Fundamental

Quantum Theory• References, Index

It is unusual for an introductorymonograph on an advanced topic suchas theoretical optics to be a pleasantread, but indeed this is.

The text gives an excellent intro -duction to a range of theoretical topicswell suited to a graduate level course.…Although a high level of mathematicsis used in many of the sections, theymaintain relevance to optical applica -tions and real objects. The text is alsoexceptionally readable owing to theobvious pleasure with which the authorpresents theoretical optics.

Navarre E: Applied Spectroscopy 60, no. 3 (March 2006): 77A–78A.

The Periodic Table: Its Storyand Its SignificanceEric M. ScerriOxford University Press198 Madison AvenueNew York, New York 10016 USA2006. 368 pages. $35.00ISBN 0-19-530573-6

The periodic table lies at the core ofchemistry and embodies the mostfundamental principles of the field. Thisbook begins with an overview of thesignificance of the periodic table and of the elements and then discusses early developments that led to theclassification of the elements, includingthe work of important scientists alongthe way. Scerri considers the impact ofphysics, including the discoveries ofradioactivity and isotopy and successivetheories of the electron. He discussesthe response to the new physicaltheories by chemists, and evaluates theextent to which modern quantummechanics explains the periodic system.

Contents:• Introduction• The Periodic System—An Overview • Quantitative Relationships Among

the Elements • Discoverers of the System • Mendeleev• Prediction and Accommodation• The Nucleus and the Periodic Table• The Electron and the Periodic Table

• Electronic Explanations by Chemists • Quantum Mechanics and the

Periodic Table• Astrophysics, Nucleosynthesis, and

More Chemistry• Notes, Index

…Scerri has misunderstood theepistemological status of Mendeleev’sabstract notion of the element. Far fromreviving a metaphysical notion [asdescribed in Scerri’s book], Mendeleevdid his best to promote a positive, if abstract, notion of the element.

…Scerri makes a plea for theautonomy of chemistry. Heconvincingly argues that the abstractnotion of the element was crucial torescuing the periodic system in the lightof the discovery of isotopes.

Bensaude-Vincent B: Nature 445, no. 7125 (January 18, 2007): 263–264.

76 Oilfield Review


07-OR-004-0


SCHLUMBERGER OILFIELD REVIEW

AUTUMN

2007VOLUM

E 19 NUM

BER 3

Autumn 2007

Tsunami Science

Borehole Seismic Surveys

Geomechanics

Fluid-Property Measurements

Oilfield Review


oilfield review autumn 2007

Documents