Tim J. BourgeoisKen BramlettPete CraigShell Deepwater Production, Inc.New Orleans, Louisiana, USA

Darrel CannonKyel HodenfieldJohn LovellSugar Land, Texas, USA

Ray HarkinsIan Pigram ARCO British LimitedGuildford, Surrey, England

For help in preparation of this article, thanks to Dave Bergt, Schlumberger Oilfield Services, Sugar Land,Texas, USA; Ted Bornemann, Bill Carpenter, Frank Shrayand Rachel Strickland, Anadrill, Sugar Land, Texas; Joseph Chiaramonte and Darwin Ellis, Schlumberger-DollResearch, Ridgefield, Connecticut, USA; Craig Kienitz,Anadrill, The Hague, The Netherlands; Martin Lüling,Schlumberger Riboud Product Center, Clamart, France;Dave Maggs, Anadrill, New Orleans, Louisiana, USA; andDavid Robertson, Forest Oil, Denver, Colorado, USA.ADN (Azimuthal Density Neutron), ARC5 (Array ResistivityCompensated), ARC675, CDR (Compensated DualResistivity), ELAN (Elemental Log Analysis), FMI (FullboreFormation MicroImager), GeoVISION675, PowerPulse, RAB(Restivity-at-the-Bit), TLC (Tough Logging Conditions),VISION475, VISION675, VISION First Look and VISION TelemetryProtocol are marks of Schlumberger.

Logging-while-drilling (LWD) technology becameavailable a mere ten years ago. At that time, thetools fulfilled the primary purpose of theirdesign, which was to aid in correlation. Within acouple of years, the industry had found six mainapplications for these tools—applications thatremain key today:

Formation evaluation—Real-time correlationand evaluation allow coring and casing pointselection. Logging before extensive invasionoccurs may reveal hydrocarbon zones that can besaturated with borehole fluid by the time wire-line logs are run.

Multiple-pass logging—Comparison logsmade at different times can help distinguish payfrom water zones, locate fluid contacts and iden-tify true formation resistivity (Rt). Permeablezones may be identified from time-lapse filtratemovement.

Insurance logging—Logs obtained whiledrilling provide contingency data in case the wellis lost or when conditions create boreholes thatyield poor-quality wireline logs.

Cost reduction—Running wireline tools inhigh-deviation wells requires conveyance bydrillpipe. In some cases, these wells can belogged with LWD tools, either while or immedi-ately after drilling, saving rig time offshore or inwells otherwise needing the TLC Tough LoggingConditions system.

Enhancing drilling safety and efficiency—Measurements while drilling provide real-timedata on drillstring mechanics, fluid dynamics andpetrophysics for assessing pore pressure andwellbore stability and for drilling program andcompletion strategies (see “Using DownholeAnnular Pressure Measurements to ImproveDrilling Performance,” page 40 ).

Geosteering—By comparing real-time logresponses to an expected model, the wellboretrajectory is modified, thereby placing the well inthe most productive portion of a pay zone.

As the real-time nature of LWD informationbegan to be fully exploited, the early emphasison correlation in the late 1980s gave way to dom-inance by geosteering and well-placement appli-cations. Availability of LWD data permitted safeand efficient drilling of exotic trajectories andextended-reach and multilateral wells that wereunimaginable ten years ago (see “Key Issues inMultilateral Technology,” page 14 ). These wellsfrequently make headlines in industry journalswhen technological advances contribute tobreaking existing directional drilling records.1

Pushing the Limits of FormationEvaluation While Drilling

Through a few case studies this article demonstrates how new

logging-while-drilling measurements are being used to open

frontiers and evaluate formations as soon as they are encountered.

29Winter 1998

1. Allen F, Tooms P, Conran G, Lesso B and Van de Slijke P:“Extended-Reach Drilling: Breaking the 10-km Barrier,”Oilfield Review 9, no. 4 (Winter 1997): 32-47.

Many of the LWD innovations that havehelped directional drillers master the art and sci-ence of geosteering are also advancing the causeof assessing reservoir quality while drilling.Forward modeling routines have been developedthat allow real-time comparison between pre-dicted and observed logs, helping drillers stay inthe pay.2 This modeling capability also lets inter-preters evaluate LWD data for petrophysical andfluid properties and for geologic structure.

Oil company interpreters are becoming morefamiliar with while-drilling measurements,understanding their departure from wireline-style logs, and trusting them. Operators are alsodemanding more measurements for more holesizes, and as a result, a broader range of servicesis being offered. The more comprehensive reser-voir assessment that is now possible makes LWDformation evaluation results valuable not only forwellsite decisions, but also for longer term reser-voir planning and development—as wireline log-ging results have been all along.

Ten years ago, the available LWD measure-ments were gamma ray, neutron porosity, litho-density, photoelectric effect and phase-shift andattenuation resistivities.3 In the interim, techno-logical advancements have vastly improved andenhanced these basic measurements and evenadded new formation evaluation measurementsnot previously available in the logging industry.First came azimuthal or quadrant measurementssuch as the quadrant density and photoelectricfactor (Pe) on the ADN Azimuthal Density Neutrontool and the quadrant gamma ray and real-timeresistivities on the RAB Resistivity-at-the-Bit tool.Then came quantitative images with multiple-depth resistivity images from the RAB tool anddensity images from the VISION475 system. Theaddition of multiple depths of investigation to theazimuthal data has created new opportunities tocomplete the formation evaluation picture.

The comprehensive Schlumberger VISION475system (the nominal outer diameter of the tool is4.75 inches) encompasses the enhanced technol-

ogy to provide formation evaluation and drillingmeasurements in 53⁄4- to 63⁄4-in. holes. In additionto direction, inclination and toolface, theVISION475 tool makes a neutron porosity mea-surement, azimuthal readings of lithodensity, Peand gamma ray, and records 2-MHz phase-shiftand attenuation resistivities at up to ten depthsof investigation.

Deciphering phase-shift measurements withmultiple depths of investigation for resistivityinterpretation has become common practice inthe industry. However, the inclusion of attenua-tion resistivity measurements with multipledepths of investigation has brought additionalvalue to the petrophysicist. Although acquiredwith the same transmitter-receiver spacing, theattenuation measurement has a greater depth ofinvestigation than the corresponding phase-shiftmeasurement. These complementary measure-ments offer an opportunity to understand moreabout the fluid and resistivity characteristics ofthe formation. For example, comparison of atten-

> Possible oil-water contacton phase-shift resistivity. The gamma ray (GR) in track 1 shows sand from7740 to 8020 ft, and thephase-shift resistivity intrack 2 indicates the zoneabove 7920 has high resistivity—a possible payzone above the oil-watercontact. Attenuation resistivity in track 3 showsthe possible oil layer to be toa zone of resistive invasion,and not worth completing.

30 Oilfield Review

>

uation and phase-shift resistivities provides adiagnostic method for differentiating betweenborehole fluid invasion and formation anisotropy,a technique discussed later in this article

In one case from a Forest Oil well in the Gulfof Mexico, the while-drilling gamma ray (GR)indicated a sand from 7740 to 8020 ft and thephase-shift resistivities identified a possible oil-water contact at 7920 ft (previous page). The fivephase-shift resistivities, each with a differentdepth of investigation, have very little separa-

tion, which indicates little to no invasion. A sim-ple resistivity index calculation yields 38% watersaturation, making the potential oil layer a candi-date for testing.

However, the attenuation resistivities that aresimultaneously recorded by the VISION475 toolappear to contain evidence to the contrary. Thesedeeper reading resistivities show significant sep-aration, with the deepest measurement—anapproximately 30-in. [75-cm] depth of investiga-tion from the 34-in. receiver-transmitter spac-ing—recording the lowest resistivity of about 0.4ohm-m. This profile indicates resistive invasion,which might be expected for wells drilled withoil-base mud, but this was water-base mud witha resistivity, Rm, of 0.1 ohm-m. However, forma-tion water resistivity, Rw, in this zone is approxi-mately 0.03 ohm-m, causing the resistiveinvasion profile. When formation resistivity iscomputed by inversion processing that takesinvasion into account, the zone shows 100%

water saturation. If this zone had been com-pleted, a significant investment would have pro-duced only water. The extra information broughtby the deeper reading attenuation measurementsavoided the cost of an unnecessary completion.

Extracting meaningful information from thetwo-receiver, five-transmitter tool configuration toprobe five depths of investigation each for phase-shift and attenuation resistivity requires carefulborehole compensation and borehole correctionof the measurements. Without borehole correc-tion, washouts together with conductive mud canmasquerade as invaded or anisotropic zones.Borehole rugosity can cause spikes, or resistivityhorns that may be misinterpreted as laminatedformations (above). Borehole compensation isnecessary because it significantly reduces theeffects of borehole rugosity and precisely cancelsmeasurement errors caused by gain and phase-shift differences in the receivers’ electronics,which typically vary with temperature.

> Borehole compensation for accurate VISION475 multidepth measurements. Without borehole compensation and correction (top), spikes andseparations in the curves of the phase-shift resistivity measurements cannot be interpreted reliably. With correction (bottom), high-resolutiondata and curve separations can be identified and interpreted.

31Winter 1998

2. Bonner S, Burgess T, Clark B, Decker D, Orban J,Prevedel B, Lüling M and White J: “Measurements at theBit: A New Generation of MWD Tools,” Oilfield Review 5,no. 2/3 (April/July 1993): 44-54.Bonner S, Fredette M, Lovell J, Montaron B, Rosthal R,Tabanou J, Wu P, Clark B, Mills R and Williams R:“Resistivity While Drilling—Images from the String,”Oilfield Review 8, no. 1 (Spring 1996): 4-19.Allen D, Dennis B, Edwards J, Franklin S, Livingson J,Kirkwood A, White J, Lehtonen L, Lyon B, Prilliman J andSimms G: “Modeling Logs for Horizontal Well Planningand Evaluation,” Oilfield Review 7, no. 4 (Winter 1995): 47-63.

3. Bonner S, Clark B, Holenka J, Voisin B, Dusang J,Hansen R, White J and Walsgrove T: “Logging WhileDrilling: A Three-Year Perspective,” Oilfield Review 4, no. 3 (July 1992): 4-21.

A series of logs from a Shell deep-water pro-ject in the Gulf of Mexico demonstrates theimpact of adding still more LWD measurementsto the interpretation. In this highly deviated well,the standard GR, rate of penetration (ROP),phase-shift resistivity and average density andneutron data indicate a homogeneous formationin this potential pay zone (above). The well wasdrilled with high-salinity drill-in fluid, and phase-

shift curves exhibit a conductive invasion profilewith the deepest spacing at 34-in. measuring thehighest resistivity, about 4 ohm-m. Resistivityprocessing to compensate for the invasioneffects would correct Rt to above 4 ohm-m.

This is the limit of information available froma conventional “triple combo” be it LWD or awireline system, and it appears to give arespectable interpretation of the reservoir, butthe VISION475 system provides more informationand sheds new light on the reservoir interval.

If this were truly a conductive invasion profile,as the phase-shift measurement indicates, thedeepest attenuation curves would show higherresistivity than the deepest phase-shift curves.However, all the attenuation outputs read alower resistivity than even the shallowest phase-shift curve. This is an example of resistivityanisotropy—a difference in resistivity valuedepending on the direction in which the mea-

> Formation evaluation while drilling in the Gulf of Mexico. In this highly deviated well the GR and ROP in track 1, phase-shift and attenuation resistivities intracks 2, 3 and 4, and average density and neutron data in tracks 3 and 5 indicate a homogeneous pay formation. Phase-shift resistivity curve separationsuggests conductive invasion, but this is not confirmed by attenuation resistivities; resistivity anisotropy is responsible. Quadrant displays of density, on top,bottom, left and right of the borehole, are in tracks 6 and 7.

32 Oilfield Review

surement is made (see the first case study from 7700 to 7740 ft for anotherexample of anisotropy, page 30 ).4

In vertical wells penetrating horizontal layers with no invasion, 2-MHztools measure horizontal resistivity, Rh. This is taken as equivalent to Rt, theresistivity input to most formulae derived to predict fluid saturation, and soserves as the reference, or threshold, by which formations are judged to con-tain pay or not. At other angles, for example, in highly deviated and horizon-tal wells passing through horizontal layers, 2-MHz tools respond to somecombination of vertical and horizontal resistivities. Vertical resistivity (Rv), orresistivity perpendicular to bedding, is always at least as much as, and usu-ally more than, horizontal resistivity—sometimes reaching a 10 to 1 ratio (see“Anisotropy and Invasion,” next page ).

In the case at hand, the phase-shift curves are each reading a differentcombination of horizontal and vertical resistivity, depending on the transmit-ter-receiver spacing. Formation resistivity, Rt, is not greater than 4 ohm-m, aswould have been calculated by a radial-invasion resistivity inversion program.An anisotropy inversion program can be used to calculate Rh and Rv, and thenRh is used in water saturation calculations to derive Sw.

Density curves from the VISION475 log also provide more information thanprevious-generation LWD density tools, which combine weighted averages ofdensity from all around the borehole. The density and Pe measurements ofthe VISION475 tool are recorded in 16 oriented sectors. These can be displayedeither as an image, or presented as four quadrants—top, bottom, right andleft—as the drillstring rotates (below).

For a first view, the bottom and average densities can be compared forconsistency (previous page). This log was recorded in a highly deviated well,so the bottom-quadrant density, in closer contact with the borehole, shouldgive the best quantitative data. In this interval, not only does the bottom-quadrant density disagree with the average density, but it also occasionallymeasures a lower bulk density. This appears strange because assuming thebottom of the tool is in contact with the formation also implies that the topof the tool is not. When that occurs, the mud density, which here is less thanthat of the formation, should influence the top-quadrant reading and as aresult, the average density would tend to be lower.

Taking the next step in evaluating this well, all four quadrant densities arepresented with photoelectric factor and bulk density correction for each quad-rant (left). The right and left density quadrants agree well throughout theentire interval. The top and bottom quadrant densities not only disagree, butcross each other. A threaded borehole, borehole breakout, or a combinationof heavy mud and hole conditions could explain this unusual response. Theresponse could also be due to a position change of the borehole assembly in

< What quadrant densities mean. Density mea-surements are gathered into four quadrants—top, bottom, right and left—as the drillstringrotates. If the VISION475 tool cuts across a layerboundary with a near full-gauge stabilizer (left),all quadrant measurements are quantitative. The view is along the tool axis, which in thiscase is also in the plane of the layer boundary.Quadrant density measurements readily definethis boundary, providing a distinct bulk density in each layer. Without a stabilizer (right) the tooltends to lie on the bottom side of the hole, givingbetter borehole contact and so better quality to the bottom quadrant measurement. The topquadrant measurement would be qualitative due to the standoff distance.

33Winter 1998

4. Anderson B, Bryant I, Lüling M, Spies B and Helbig K: “Oilfield Anisotropy: Its Origins andElectrical Characteristics,” Oilfield Review 6, no. 4 (October 1994): 48-56.Lüling MG, Rosthal RA and Shray F: “Processing and Modeling 2-MHz Resistivity Tools inDipping, Laminated, Anisotropic Formations,” Transactions of the SPWLA 35th AnnualLogging Symposium, Tulsa, Oklahoma, USA, June 19-22, 1994, paper QQ.

>

Many resistivity logs exhibit a combination ofanisotropy, invasion and shoulder (adjacent)bed effects, and each effect must be taken intoaccount to deduce true formation resistivity.

Resistivity anisotropy can be caused by layer-ing, lithology or fluid content. It is typicallyexpressed as the ratio of vertical to horizontalresistivity, Rv/Rh, and its effects on tool responsecan be understood by modeling. The standardresponse of 2-MHz tools in vertical wells pene-trating horizontal layers with no invasion istaken as the reference, and tool response to lay-ers at other relative angles can be computed(below left).

In this model the formation consists of a sandinterbedded with an equal amount of shale foran anisotropy ratio Rv/Rh of 6.7. As the relativeangle increases, the apparent resistivity mea-sured by both phase shift and attenuationincreases. Above 45 degrees, the effect isgreater on the longer spacings; for example, thephase-shift 34-in. curve measures a higherapparent resistivity than the phase-shift 28-in.curve. The curve order resembles a conductiveinvasion profile, and therefore may be misinter-preted. Anisotropy can cause resistivities mea-sured in high-angle wells to be deceptively high.

Two keys are used to distinguish conductiveinvasion from anisotropy. The first is comparisonof phase-shift to attenuation measurements:although phase-shift resistivities indicate a con-ductive invasion profile, corresponding attenua-tion outputs measure lower resistivity. Ifconductive invasion were causing the phase-shiftcurve separation, the deeper reading attenuationoutputs would measure a higher apparent resis-tivity than the phase-shift curves. This is animportant use of attenuation measurements. Thesecond key is revealed in the modeled exam-ple—the curves are uniformly separated whenviewed on a logarithmic scale. Uniform separa-tion is less common with invasion.

If anisotropy can be identified in sands, it isusually an indication of hydrocarbons. However,in anisotropic formations that are hydrocarbon-bearing, deep invasion could hide theanisotropic response if Rmf and Rw are similar.1

To understand the effect of invasion on ananisotropic formation, a state-of-the-art 3Dfinite-difference code was developed that com-putes phase-shift and attenuation responseswith increasing diameter of invasion.2

Resistivity responses were modeled for ananisotropic formaton whose anisotropy changeswith invasion (below right). The virgin forma-tion anisotropy ratio, Rvt/Rht, is 6.8, but onceinvaded, the anisotropy ratio falls to 1.25—nearly isotropic. This change will have differenteffects on the phase-shift and attenuation resis-tivity responses, depending on their depth ofinvestigation. For invasion diameters less than15 in., the anisotropy effect dominates.Anisotropy is recognizable by separated phase-shift curves reading higher than attenuationcurves. At invasion diameters greater than 50inches, the effects of invasion rule curve separa-tion. Measuring phase-shift and attenuationresistivities before extensive invasion is there-fore crucial when anisotropy is present.

Anisotropy and Invasion

> Effects of anisotropy on phase-shift and attenuation resistivities. Anisotropybecomes evident as the relative angle between bedding and tool axis increases.The curves resemble those seen in a conductive invasion profile except that withanisotropy, phase-shift curves read more resistive than attenuation.

> Effect of invasion on an anisotropic formation. In this formation, which isanisotropic before invasion, but less so after, modeling shows that for invasiondiameters less than 15 in., the anisotropy is still interpretable from phase-shiftand attenuation resistivity curves. After invasion diameters surpass 50 in., theeffects of invasion mask the anisotropy of the virgin formation.

34 Oilfield Review

1. Klein JD, Martin PR and Allen DF: “The Petrophysics ofElectrically Anisotropic Reservoirs,” Transactions of theSPWLA 36th Annual Logging Symposium, Paris, France,June 26-29, 1995, paper HH.

2. Anderson B, Druskin V, Habashy T, Lee P, Lüling M, Barber T, Grove G, Lovell J, Rosthal R, Tabanou J,Kennedy D and Shen L: “New Dimensions in ModelingResistivity,” Oilfield Review 9, no. 1 (Spring 1997): 40-56.

the wellbore caused by changes in wellbore incli-nation. But no interpretation guesses are neces-sary because the VISION475 system clearlyprovides the answer with density image data.

The density images reveal the detail of reser-voir configuration—a series of thin sands andshales dipping at varying angles relative to theborehole (below left). These VISION475 imagesprovide an easy and efficient means of interpret-ing complex data. The first track image is color-scaled to represent measured quantitativedensity variations, while in the second track thevariations have been enhanced by changing thecolor scale to bring out detail.

Throughout this interval, the azimuthal densityimaging was the only measurement to flag thesubtle sand-shale layering. The lessons learnedare twofold: first, the standard suite of GR, resis-

tivity, neutron and average density measurementsmay not always be sufficient for complete forma-tion evaluation. In this case, all the standard mea-surements pointed to a homogeneous zone.Clearly the revelation of a laminated sand-shalesequence can have an impact on the appraisal ofreservoir quality and its subsequent drainage.Second, techniques that assume maximum den-sity to be the correct density would greatly under-estimate porosity and distort the true reservoircharacter. This new information is valuable notonly to drillers in real time, but also to well plan-ners who may need to change future drilling tra-jectories, to completion engineers for effectingmore efficient completions, to reservoir engineersfor modeling and simulating production and togeologists for calculating structural dip.5

Invasion, Dip and GasThe previous examples show how LWD logsimprove formation evaluation in deviated oilwells with invaded zones, anisotropic layers andthin dipping beds. Determining accurate valuesof porosity and water saturation in gas wellsunder these conditions, however, has been aspecial problem that only recently is seeingsome resolution.

In vertical wells, depth of invasion of mud fil-trate into a formation depends on many factors,including mud properties and lithology, porosityand absolute and relative permeability of the for-mation. In the simplest case of a vertical hole ina homogeneous permeable formation, the inva-sion profile is radially symmetric. But whenimpermeable or dipping layers, or both, areencountered, the volume invaded by boreholefluid takes on a new shape (below right).

The invasion front becomes even more dis-torted in a gas zone, because the borehole fluid isso much heavier than the formation gas. Invasionbegins radially, but with time the heavier phase

g/cm3

ROP

RHOB Image

Gamma RayAPI

Image Orientation

g/cm3

Image OrientationU R B L U degrees0 90200 0

ft/hr

1:240 ApparentDip

TrueDip

0 90degrees

U R B L U

0 100

Dynamic RHOB Image

DEVI Deviation

Mea

sure

dDe

pth,

ft

2.05 2.45

XX300

XX250

XX200

XX350

> Density image from the VISION475 system. Measured densities appear in track 1, andare redisplayed to highlight detail in track 2. Structural dips (green dots in track 3) arehand picked using the same process as for wireline FMI Fullbore Formation Micro-Imager data and relative dip to the borehole is calculated (blue dots). These imagesare also useful for calculating a sand count or determining the net-to-gross sand ratio.

Filtrate

Impermeablelayer

Impermeablelayer

Slumpedfiltrate

Wellbore

Wellbore

> Invasion and slumping filtrate. A radial invasion front becomes distorted in the presence of a horizontal layer (top) or dippingimpermeable layer (bottom).

35Winter 1998

5. Bornemann E, Bourgeois T, Bramlett K, Hodenfield K andMaggs D: “The Application and Accuracy of GeologicalInformation from a Logging-While-Drilling Density Tool,”Transactions of the SPWLA 39th Annual LoggingSymposium, Keystone, Colorado, USA, May 26-29, 1998,paper L.

slumps in the down-dip direction. The rate ofslumping depends on the vertical permeability ofeach zone: the higher the vertical permeability,the more rapid the slumping. In addition, perme-ability anisotropy will distort slump geometry. Informations with permeability on the order of onedarcy, strong azimuthal variations in invasionhave been observed less than an hour after thebit penetrates the formation.

In gas zones, such variations can make quan-titative porosity interpretation from nuclear toolsan even greater challenge than usual. Porosity

determination from nuclear tools requires thatthe effects of gas be removed. For this, knowl-edge of the gas volume and both radial andazimuthal location is needed. Further complicat-ing the problem, the neutron and density mea-surements respond differently to the radial andazimuthal location of gas. The neutron tool readsdeeper and is sensitive to any gas near the bore-hole, relatively independent of the azimuthallocation of the gas. The density tool reads shal-lower and is sensitive only to the gas in front ofits detectors.

What is needed is a way to quantify the vol-ume of gas radially and azimuthally over thesame region of formation that the density andneutron logs investigate. Then those volumes areused to apply appropriate gas corrections to thenuclear tools for final computation of porosity.

Radial and azimuthal gas quantification isaccomplished by analyzing while-drilling quad-rant resistivity data acquired as the RAB toolrotates in the borehole. The RAB tool investi-gates a region similar to that probed by nucleartools, and has five depths of investigation. By

Dept

h, ft

X050

X100

X150

X200

ohm-m

Porosity

p.u.

RAB Resistivity

RABResistivity

VISIONDensity

Top

Botto

m

Top

Botto

m

Top

0.1 100 40 0

Shallow Button Down

Shallow Button Up

Deep Button Up

CDR

Density Up

Density Down

Neutron

> Images of invasion slump. Density and RAB images show filtrate slumping, but not always down. Track 1 displays resistivities:three from the RAB tool and attenuation resistivity from the CDR Compensated Dual Resistivity tool. Additional filtrate detected bythe resistivity in the down direction compared to the up direction is shaded. Similarly for the porosity curves in track 2, the leftcurves are from the up and down quadrants of the density tool and the right curve is from the neutron tool. Track 3 contains the RABimage, with white most resistive, and track 4 shows the density image with dark as the most dense, and lighter colors as less dense.

36 Oilfield Review

using readings from all around the borehole,three of those depths of measurement can bepartitioned azimuthally into 56 segments. Fromthese three measurements, three quantities canbe solved for—the diameter of invasion, DI;invaded zone resistivity, Rxo ; and true formationresistivity, Rt —in any or all of the 56 azimuthalsegments. Rxo and Rt are assumed to be constantaround the hole; only DI varies. The determina-tion of Rt is most robust from the direction withminimum DI, and Rxo is most robust from thedirection with maximum DI.

Correcting the LWD density and neutron toolsrequires an appropriate radial response functionand appropriate DI; both are different for eachtool. The qualitative response of density and neu-tron tools has long been understood.7 The radial

response function of the density tool has beenquantified and is relatively independent of thefluids involved. The neutron radial response func-tion has been elusive, but Ellis and Chiaramonteof Schlumberger-Doll Research, Ridgefield,Connecticut, USA have recently completed amodeling code to allow the response to be calcu-lated under all conditions. Their modeling showsthat the neutron responds to the gas closest tothe borehole. Therefore the DI needed to correctthe neutron is the minimum DI computed aroundthe wellbore. In the typical slumping-filtratecase, the closest gas usually is at the top of thehole or possibly on the sides, but definitely not atthe bottom. The DI for the density correction isthe one computed in the direction the densitysensor is pointing.

Finally Rxo, Rt, invasion factors for density andneutron, bulk density, neutron porosity, plusappropriate parameters are entered in ELANElemental Log Analysis software to solve forporosity and water saturation.

This method was tested by partners ARCOand Enterprise on a North Sea gas well deviatedabout 40° encountering formations with 70°apparent dip. Images from the ADN and RABtools plot the location of the higher density,lower resistivity mud filtrate, which did notalways slump straight down (previous page). Thediameter of invasion is plotted, and the com-puted porosity displayed for comparison withcore measurements (above).

in. 300

in. 300

Dept

h, ft

X050

X100

X150

X200

Porosity

p.u. 040

Depth of Invasion

RAB Down

RAB Up

Neutron

Quadrant Corrected Effective Porosity

Density Down

Density Up

Core porosity

> Porosity computed from corrected neutron- and density-while-drilling data. Track 1 displays the diameter of invasion, DI,used to calculate corrections. The area between DI calculated from the up- and down-RAB measurements is shaded. Track 2contains porosities from up- and down-quadrant density measurements, neutron measurements and core (red dots). The effective porosity computed after corrections (orange curve) compares favorably with core measurements.

7. Sherman H and Locke S: “Depth of Investigation ofNeutron and Density Sondes for 35-Percent-Porosity Sand,”Transactions of the SPWLA 16th Annual LoggingSymposium, New Orleans, Louisiana, USA, 1975, paper Q.

37Winter 1998

Getting a First LookHigh-quality images provide valuable input to theinterpretation process. Geological information,such as laminations, location of the wellborewith respect to bed boundaries and the apparentdip magnitude and direction of bedding planes, isessential for interpreting log responses in highlydeviated wells. Images quickly reveal whetherthe bit is drilling down into or up through beddingplanes—critical for geosteering a well and refin-ing geological interpretations.

The VISION First Look display is a wellsiteanswer product that combines images with thecomplete VISION dataset to provide a format for

quickly interpreting logs in highly deviated wellsand making drilling decisions. In this real-timeinterpretation for the Shell deep-water RamPowell field in the Gulf of Mexico, the horizontalwellbore is drilled in a clean sand sheet depositnearly parallel to bedding. During drilling, severaltight or hard features were encountered unex-pectedly (above). Rate of penetration droppedsignificantly, and resistivity and bulk densityincreased while neutron porosity approachedzero p.u. These tight streaks were a surprise,because two vertical wells drilled in this areahad not encountered such a feature.

These tight streaks were of concern, as theymight influence production, perhaps necessitat-ing a change in wellbore trajectory. They firstwere assumed to be depositional features lyingparallel to bedding. However, careful examina-tion of the signature of the events on the densityand neutron curves reveals that these are verticalinterfaces. If the hard streaks were parallel tobedding planes, the tool would encounter theboundary more gradually and the measurementtransition from reservoir to tight streak wouldoccur over some distance. These transitions arequite abrupt, indicating a high-angle boundary.

100

Gamma Ray

API0

100 0ft/hr

ROP5

Dept

h, ft

XX500

XX550

XX600

1 100ohm-m

Density, Bottom

g/cm31.65 2.65

Neutron Porosity

0.6 0ft3/ft3

22-in. Phase-ShiftResistivity

Real-time data revealing tight treaks.While drilling in the Ram Powell field,Gulf of Mexico, tight streaks wereencountered unexpectedly. Theseevents show up as high-density, low-porosity interfaces in track 3. Rate ofpenetration (track 1) dropped signifi-cantly each time, and resistivityincreased (track 2).

38 Oilfield Review

>

This interpretation of vertical boundariesraised new concerns for the operator: Was thereservoir compartmentalized? Were thesestreaks mineralized fault planes? What is thevertical extent of these features? Should thewellbore trajectory be changed? The VISION FirstLook log, played back with recorded mode data,was able to answer these questions (above).

The density images displayed on the VISION

First Look log reveal the true nature of the“tight streaks.” The boundaries of the featuresare not planar, but rather calcite-cemented nod-ules. The features are not continuous verticalplanar events and will not have a large-scaleimpact on production.

The deeper reading attenuation resistivitymeasurement confirms this interpretation. Theattenuation measurements are not influenced bythe high-resistivity hard streaks to the degreethat phase-shift measurements are, and thereare no polarization horns, which indicates thatthese events do not extend far from the wellbore.

The Future VisionThe ability to achieve better reservoir qualityassessments in real time has satisfied some, butnot all, formation evaluation while-drilling needs.Already operators are asking for these LWD mea-surements in more hole sizes, and this demand isbeing met with the imminent introduction of theVISION675 and GeoVISION675 systems for 8- to 12-in. holes. The VISION675 system will encompassthe ARC675 Array Resistivity Compensated mea-surement, an enhanced PowerPulse MWD toolwith the new VISION Telemetry Protocol systemand a new 6.75-in. VISION675 density-neutrontool. The VISION675 density-neutron tool extendsthe capabilities of the existing 6.75-in. ADN toolby adding multisector density, Pe and caliperimages for both oil-base and water-base mud. Arelated tool for geological imaging while drillingwill appear in the GeoVISION675 system, whichwill contain a new-generation laterolog imagingtool in place of the ARC675 module.

Other measurements are making their way tothe 4.75-in. format, including downhole annularpressure and bit inclination for precision trajec-tory control.

To keep pace with the introduction of newmeasurements, interpretation experts are devis-ing new techniques for getting the most from thenew data. Programs for interpreting measure-ments in layers that are anisotropic, invaded, thin,dipping, or all of the above, are finding new chal-lenges when applied to time-lapse LWD data—LWD logs acquired before and after bit changesor other delays in drilling. Researchers are devel-oping methods for faster modeling and inversionof tool responses in more complex geometriesand more realistic formations. These efforts willenhance our ability to perform formation evalua-tion while drilling, and also will improve all otherLWD applications. —LS

PhaseShift

TVD

API

ResistivityTime After Bit

Gamma Ray

Rw-corrected Bulk Volumes

Effective Porosity

MatrixBound Water

0.02 200ohm-m

Attenuation Resistivities

Phase Shift-Resistivities

0.2 2000ohm-m

Bottom Density

Neutron Porosity

Attenuation

Mea

sure

d De

pth,

ft

Clay

XX200

XX250

ft3/ft30 60

1.65 2.65g/cm3

BHA sliding

Image orientationU R B L U0 hours 10

0 0 0.5

0 150

ft ohm-m100

Azimuth andDeviation

0 90

34 in.

10 in.16 in.22 in.28 in.

g/cm3

RHOB Image

2.05 2.45

> VISION First Look wellsite images. Bulk volume analysis (track 2) and an Rw curve (track 1) corrected forclay volume and type are computed and displayed. [Adapted from Cannon D: “Shales: An Alternate Source forWater Resistivities,” Transactions of the SPWLA 36th Annual Logging Symposium, June 26-29, 1995, Paris, France,paper LLL.] Phase-shift and attenuation resistivity curves are displayed in track 3, densities are in track 4. Density images (track 5) of tight streaks show that these features are clearly not planar.

39Winter 1998