2 Oilfield Review

Characterizing Permeability with Formation Testers

Cosan AyanAberdeen, Scotland

Hafez HafezAbu Dhabi Company for OnshoreOperations (ADCO)Abu Dhabi, United Arab Emirates (UAE)

Sharon HurstPhillips PetroleumBeijing, China

Fikri KuchukDubai, UAE

Aubrey O’CallaghanPuerto La Cruz, Venezuela

John PefferAnadarkoHassi Messaoud, Algeria

Julian PopSugar Land, Texas, USA

Murat ZeybekAl-Khobar, Saudi Arabia

For help in preparation of this article, thanks to Mahmood Akbar, Abu Dhabi, UAE.AIT (Array Induction Imager Tool), CQG (Crystal QuartzGauge), FMI (Fullbore Formation MicroImager), MDT(Modular Formation Dynamics Tester), OFA (Optical Fluid Analyzer) and RFT (Repeat Formation Tester) aremarks of Schlumberger. RDT (Reservoir Description Tool) is a mark of Halliburton.

We never seem to know enough about permeability. We measure it at small scales

through laboratory tests on cores. We infer it at large scales from well tests and pro-

duction data. But to manage the development of a reservoir, we also need to quantify

features at intermediate scales. This is where the versatility of wireline formation

testers comes into play.

1. In direct measurements of fluid flow in rocks, the quan-tity measured is the mobility (permeability/viscosity).According to Darcy’s law, all fluid effects are accountedfor by the viscosity term, and permeability is independentof fluid. In practice, this is not exactly true, even withoutchemical interactions between rock and fluid. Absolutepermeability is also known as intrinsic permeability.

2. The term radial permeability, kr, describes radial flowinto a wellbore. In vertical wells, radial permeability isthe same as horizontal permeability. Vertical permeabilityis written both as kv and kz. Spherical permeability iswritten as ks.

50387schD02R1.p2.ps 12/10/01 3:48 PM Page 2

Autumn 2001 3

Modern wireline formation testers bring specialknowledge about reservoir dynamics that noother tool can acquire. Through multiple pres-sure-transient tests, they can evaluate vertical aswell as horizontal permeability. By measuring ata length scale between cores and well tests, theycan quantify the effect of thin layers that are notseen by other techniques. These layers play avital role in reservoir drainage, controlling gas-and waterflood performance, and leading tounwanted gas and water entries. Modern wire-line formation testers can also be a cost-effec-tive, environmentally friendly alternative toregular drillstem and pressure-transient tests.This article shows how permeability measure-ments derived from wireline formation testersare contributing to reservoir understanding andmaking an impact on reservoir development.

Which Permeability?Permeability determines reservoir and well per-formance, but the term can refer to many types ofmeasurements. For example, permeability can beabsolute or effective, horizontal or vertical.Permeability is defined as a formation property,independent of the fluid. When a single fluidflows through the formation, we can measure anabsolute permeability that is more or less inde-pendent of the fluid.1 However, when two or morefluids are present, each reduces the ability of theother to flow. The effective permeability is thepermeability of each fluid in the presence of theothers, and the relative permeability is the ratio ofeffective to absolute permeability. In a producingreservoir, we are most interested in effective per-meability, initially of oil or gas in the presence ofirreducible water, and later of oil, gas and waterat different saturations. To further complicatematters, effective and absolute permeabilitiescan be significantly different (see “ConventionalPermeability Measurements,” page 6).

Formations are usually anisotropic, meaningtheir properties depend on the direction in whichthey are measured. For fluid-flow properties, weusually consider transversely isotropic forma-tions, meaning formations in which the two hori-zontal permeabilities are the same and equal tokh, while the vertical permeability, kv, is different.Although more complicated formations exist,there are typically not enough measurements to quantify more than these two quantities.Permeability anisotropy can be defined as kv/kh,kh/kv, or the ratio of the highest to the lowest per-meability. In this article we will use kh/kv, a quan-tity that is most often greater than 1.2

The next complication is related to spatial dis-tribution. Reservoir management would be muchsimpler if permeability were distributed uniformly,but, in practice, formations are complex and het-erogeneous—that is, they have a range of valuesabout two or more local averages. The number ofmeasurements needed for a full description of aheterogeneous rock is impossibly high; moreover,the result of each measurement depends on itsscale. For example, for an idealized reservoir com-prising isotropic sand with randomly distributedisotropic shales, there are three scales to con-sider—megascopic (the overall reservoir), macro-scopic (the grid squares used in reservoirsimulation), and mesoscopic (individual facies)(above). The megascopic anisotropy is veryhigh—between 103 and 105. However, areas Aand B are isotropic, while the grid squares are intermediate, showing that the large-scaleanisotropy is in fact caused by local heterogene-ity. Measurements at different scales and in different locations will find different values forboth kh and kv and hence different anisotropy.

Which permeability to choose? In a single-phase, homogeneous reservoir, the question isirrelevant—but such reservoirs do not exist.Almost all reservoirs, and particularly carbon-ates, are highly stratified. For some formations,flow properties also vary laterally. For instance,in deltaic sandstone deposits, the world’s mostprolific reservoirs, flow properties vary laterallybecause of the sorting of sediments according tosize and weight during transport and deposition.Whether in sandstone or carbonate, as hetero-geneity increases, the distribution of permeabil-ity becomes as important as its average value.

Early in the life of a reservoir, the main concernis the average horizontal effective permeability tooil or gas, since this controls the productivity andcompletion design of individual wells. Later on,vertical permeability becomes important becauseof its effect on gas and water coning, as well asthe productivity of horizontal and multilateralwells. The distribution of both horizontal and ver-tical permeability strongly affects reservoir perfor-mance and the amount of hydrocarbon recovery,while also determining the viability of secondary-and tertiary-recovery processes.

A B

100

Dept

h, ft

Horizontal distance, ft200 300 400 500 600 700 800 900 1000

100

200

300

400

500

0

0

Grid square

> A cross section of an idealized reservoir that exhibits large-scale anisotropy caused by localheterogeneity. A sandstone reservoir (yellow) contains randomly distributed shales (gray). Thevertical permeability for the whole reservoir is about 104 times less than the horizontal perme-ability—a very large anisotropy. However, the small areas A and B are in isotropic sand andshale, respectively. The grid square, which might represent a reservoir-simulation block, hasintermediate permeability anisotropy. Vertical permeability is close to the harmonic average ofsand and shale permeabilities, while the horizontal permeability is the arithmetic average.[Adapted from Lake LW: “The Origins of Anisotropy,” Journal of Petroleum Technology 40, no. 4(April 1988): 395–396.]

50387schD02R1.p3.ps 12/10/01 3:49 PM Page 3

The magnitude of permeability contrastbecomes increasingly important with prolongedproduction. Thin layers, faults and fractures canhave a dramatic effect on the movement of a gascap, aquifer, and injected gas and water. Forexample, a low-permeability layer, or baffle, willimpede the movement of gas downwards. Ahigh-permeability layer, or conduit, will quicklybring unwanted water to a production well. Bothcan significantly affect the sweep efficiency andrequire a change in completion practices. Soundreservoir management depends on knowing notonly the average horizontal permeability but alsothe permeability distribution laterally and verti-cally, and the conductivity of baffles and conduits(left). As has been known for a long time, reser-voir heterogeneity is one of the major reasonswhy enhanced oil recovery is so difficult.Permeability heterogeneity, unexpected bafflesand insufficiently detailed reservoir evaluationare often the reasons that these projects fail tobe economical.3

In normal reservoir-engineering practice, themain sources of average effective permeabilityare pressure-transient well testing and produc-tion tests. These are usually good indicators ofoverall well performance. Cores and logs areused, but often after some matching, or scalingup, to well-test results. Once a reservoir has beenon production, conventional history matchinggives information on average permeability, butcannot resolve its distribution. The presence ofhigh- or low-permeability streaks and their distri-butions are inferred from cores and logs, but thisinformation is qualitative rather than quantitative.Wireline formation testers (WFTs) have steppedinto this gap, providing various measurements ofpermeability from simple drawdowns with a sin-gle probe to multilayer analyses with multipleprobes. The latter were originally used mainly todetermine anisotropy.4 With recently developedanalytical techniques and further experience,multilayer analyses now provide quantitativeinformation about permeability distribution.

Wireline Formation TestersEarly wireline formation testers were designedprimarily to collect fluid samples. Pressures wererecorded, so that the pressure buildups at the endof sampling could be analyzed to determine per-meability and formation pressure. In spite of thelimited gauge resolution and the few data pointsavailable, the results were often an importantinput to formation evaluation. Now, buildupsacquired after sampling are still analyzed to obtainan estimate of permeability at little extra cost.

The Schlumberger RFT Repeat FormationTester tool introduced the pretest, a short test

4 Oilfield Review

3. Weber AG and Simpson RE: “Gasfield Development—Reservoir and Production Operations Planning,” Journalof Petroleum Technology 38, no. 2 (February 1986): 217-226.

4. Ayan C, Colley N, Cowan G, Ezekwe E, Wannel M, Goode P,Halford F, Joseph J, Mongini A, Obondoko G andPop J: “Measuring Permeability Anisotropy: The LatestApproach,” Oilfield Review 6, no. 4 (October 1994): 24-35.

5. The so-called drawdown permeability is calculated as kd = C qµ /∆pss in units of mD, where q is the flow rate incm3/s, µ is the fluid viscocity in cp, and ∆pss is the mea-sured drawdown pressure in psi (and therefore includesany pressure drop due to mechanical skin). C, the flow-shape factor, depends on the effective radius of theprobe, and equals 5660 for the standard RFT and MDTModular Formation Dynamics Tester probes and theunits given.

Baffles Conduits

Sealing fault Nonsealing fault

Gig

a

Healed fractures Open fractures

Low-permeability genetic units High-permeability genetic units

Meg

a an

d M

acro

Low-permeability stylolite High-permeability stylolite

Tight laminations Small fractures

Shale lenses Vugs

Mes

o

Low-permeability recrystallizationfeature

High-permeability solution channel

> Permeability baffles and conduits at different length scales. In each case, reser-voir management can be improved by quantifying the effects of these features.

6. Dussan EB and Sharma Y: “Analysis of the PressureResponse of a Single-Probe Formation Tester,” SPEFormation Evaluation 7, no. 2 (June 1992): 151-156.

7. Jensen CL and Mayson HJ: “Evaluation of PermeabilitiesDetermined from Repeat Formation TesterMeasurements Made in the Prudhoe Bay Field,” paperSPE 14400, presented at the SPE Annual TechnicalConference and Exhibition, Las Vegas, Nevada, USA,September 22-25, 1985.

8. Goode PA and Thambynayagam RKM: “Influence of anInvaded Zone on a Multiple Probe Formation Tester,”paper SPE 23030, presented at the SPE Asia PacificConference, Perth, Western Australia, Australia,November 4-7, 1991. We might expect the buildup permeability to be higherthan kd since, by reading farther into the formation, itshould read closer to the effective permeability of theformation to oil or gas. However, in general experience,the buildup permeability reads lower.

50387schD02R1 11/29/01 4:59 AM Page 4

Autumn 2001 5

initially designed to determine whether a pointwas worth sampling. To the surprise of many,pretest pressure turned out to be representativeof reservoir pressure. As a result, pressure mea-surements became the main WFT application.Permeability could be estimated from both thedrawdown and the buildup during a pretest.Since a reliable pressure gradient requiredpretests at several depths, much more perme-ability data became available. With tens of testpoints in a single well, it became easier to estab-lish a permeability profile and compare resultswith core and other sources.

Pretests continue to be an important featureof modern tools, although the reliability of thepermeability estimate varies. Since pretestssample a small volume, typically 5 to 20 cm3

[0.3 to 1.2 in.3], the drawdown permeability, kd,can be overly influenced by formation damageand other near-wellbore features.5 Detailed anal-ysis shows that kd is closest to kh, although it isinfluenced by kv.6 The volume of investigation issignificantly larger than that of a core plug, but ofthe same order of magnitude. However, kd is typ-ically the effective permeability to mud filtrate inthe invaded zone rather than the absolute per-meability as obtained from core. Although somegood correlations between the two have beenfound, kd is generally considered to be the minimum likely permeability.7 Nevertheless, itcan be computed automatically at the wellsite,and is still used regularly as a qualitative indi-cator of productivity.

Pretest buildups investigate farther into theformation than drawdowns, several feet if the

gauge resolution is sufficiently high and thebuildup is recorded long enough. Except in low-permeability formations, buildup time is short, sothat the tool may be measuring the permeabilityof either the invaded zone, the noninvaded zone,or some combination of the two.8 As in the inter-pretation of any pressure-transient data, flowregimes are identified by looking for characteris-tic gradients in the rate of change of pressurewith time. For pretest buildups in which the flowregimes are spherical and occasionally radial,consistent gradients often prove hard to find, andeven then may be affected by small changes in

the pretest sampling volume. For reliable results,each pretest must be analyzed—a time-consum-ing process. Today, the analysis of short pretestbuildups for permeability is rare, mainly becausethere are much better ways to obtain permeabil-ity with modern tools.

Modular Wireline Formation TestersThe third-generation WFT is the modular tester.This tool can be configured with different mod-ules to satisfy different applications, or to handlevarying conditions of well and formation (below).

8 ft

2.3 ft~3 ft

6.6 ft

ks

A B C D E F G H

Inputport

Usually

Sometimes kh

kh,kv kh,kv kh,kv,φCt kh,kv,φCt ks and/or kh kh,kv kh,kv

φCt φCt

> Typical MDT tool configurations for permeability measurements: single probe with sample chamber and flow-controlmodule (A); a sink, normally the bottom probe, with one (B) or two (C) vertical observation probes; dual-probe modulewith one (D) or two (E) vertical probes; mini-DST configuration with dual-packer and pumpout module (F); dual-packermodule with one (G) or two (H) vertical probes. The flow-control module, sample chamber and pumpout module can beadded to any configuration. When only one pressure transient is recorded, as in (A) and (F), permeability determinationdepends on identifying particular flow regimes, type-curve matching or parameter estimation using a forward model.With one or more vertical probes, as in the other configurations, it is possible to perform a local interference test, alsoknown as an interval pressure-transient test (IPTT). With these tests, interpreters can determine kv and kh for a limitednumber of layers near the tool. Storativity, øCt, can be determined with the dual-probe module, and sometimes whenthree vertical transients are available, as in (C) and (H). With other configurations, it must be determined from otherdata. Pretest drawdown and buildup permeabilities can be determined with the dual-packer module and each probe in all configurations.

50387schD02R1.p5.ps 12/10/01 4:43 PM Page 5

6 Oilfield Review

Pressure-transient analysis, production tests, his-tory data, cores and logs are all used to estimatepermeability. Each measurement has differentcharacteristics, advantages and disadvantages.

Core data—Routine core measurements giveabsolute, or intrinsic, permeability. In shalyreservoirs with high water saturation or in oil-wet reservoirs, the effective permeability can besignificantly lower than the absolute permeabil-ity (right). Core data are taken on samples thathave been moved to surface and cleaned, so thatmeasurement conditions are not the same asthose made in situ. Some of these conditions,such as downhole stress, can be simulated onsurface. Others, such as clay alteration andstress-relief cracks, may not be reversible.

To be useful for reservoir characterization,there should be enough core samples to capturesufficiently the reservoir heterogeneity—variousstatistical rules exist to determine how manysamples are required. But it is not always possi-ble to capture a statistically valid range of sam-ples even in one well. Highly porous samplesmay fall out of the core barrel, while cuttingplugs from very tight intervals is difficult. Someanalysts prefer permeameter measurementsbecause more samples can be taken.1 Averaging,or scaling up, is another tricky issue. For lay-ered flow, the arithmetic average, kav =[∑ki hi/∑hi], is the most appropriate for the horizontalpermeability. For random two-dimensional flow,it is the geometric average, kav =[∏ki

hi / ∑hi],while for the vertical permeability, the harmonicaverage, kav =[∑ki

-1 hi/ ∑hi]-1, is important.2

Log data—Logs measure porosity and otherquantities that are related to pore size, forexample irreducible water saturation andnuclear magnetic resonance parameters.3

Permeability can be estimated from these mea-surements using a suitable empirical relation-ship. This relationship normally must becalibrated for each reservoir or area to moredirect measurements, usually cores, but some-times, after scaling up, to pressure-transientresults. The main use of log-derived permeabilityis to provide continuous estimates in all wells.On the economic side, cores and logs have manyapplications, so that the extra cost of obtainingpermeability from them is relatively small.

Well tests—Pressure-transient analysis of welltests measures the average in-situ, effective permeability of the reservoir. However, theresults have to be interpreted from the changeof pressure with time. Interpreters use severaltechniques, including the analysis of specificflow regimes, and matching the transient totype curves or a formation model. In conven-tional tests, the well is produced long enough tosample up to the reservoir boundaries. Impulsetests produce for a short time and are useful forwells that do not flow to surface. In both cases,but especially for impulse tests, there is notnecessarily any unique solution for permeability.

In most conventional tests, the goal is to mea-sure the transmissivity (khh/µ) during radialflow. The reservoir thickness, h, can be esti-mated at the borehole, but is it the same tensand hundreds of feet into the reservoir wherethe pressure changes are taking place? In prac-tice, other information—geological models andseismic data—helps improve results. With con-ventional well tests, the degree of heterogeneitycan be detected, but the permeability distribu-tion cannot be determined and there is no vertical resolution.

Conventional Permeability Measurements

> Typical relative-permeability curves for oil and water in a water-wetreservoir (top) and an oil-wet reservoir (bottom). Effective permeabilitiesare relative permeabilities multiplied by the absolute permeability. PointsA and A’ represent the typical situation for a wireline formation testerdrawdown measurement in water-base mud. In a water-wet reservoir, thefiltrate flows in the presence of 20% residual oil and has a relative perme-ability of 0.3. Points B and B’ represent the typical situation for pressure-transient analysis in an oil reservoir. In a water-wet reservoir, the oil flowsin the presence of 20% irreducible water and has a relative permeability of0.9. Points A, A’, B and B’ are also known as endpoint permeabilities. Someengineers refer relative permeabilities to the effective permeability to oilrather than the absolute permeability, as shown here.

0

0.2

0.4

0.6

0.8

1.0

0.20 0.4 0.6 0.8 1.0Sw

Rela

tive

perm

eabi

lity

0

0.2

0.4

0.6

0.8

1.0

0.20 0.4 0.6 0.8 1.0Sw

B ARe

lativ

e pe

rmea

bilit

y

Water-wet

B’ A’

Oil-wet

krw

kro

krw

kro

50387schD02R1.p6.ps 11/17/01 6:01 PM Page 6

7

Some of these modules are particularly relevantfor permeability measurements. The descriptionsof the modules below refer to the SchlumbergerMDT Modular Formation Dynamics Tester tool,unless otherwise specified.

The single-probe module—This module pro-vides communication between the reservoir andthe tool. It consists of the probe assembly,pretest chamber, strain and quartz pressuregauges, and resistivity and temperature sensors.The probe assembly has a small packer, whichcontains the actual probe. When a tool is set,telescoping backup pistons press the packerassembly against the borehole wall. The probe ispressed farther through the mudcake into contactwith the formation. Special probe-assemblydesigns are available for difficult conditions.9

Communication is established with the formationby a short pretest, after which the module canwithdraw fluids for sampling or act as a passivemonitor of pressure changes.

The dual-probe module—This module con-sists of two probe assemblies mounted in fixedpositions on the same mandrel. In the HalliburtonRDT Reservoir Description Tool, the probes aremounted above one another, separated by a fewinches and facing the same way.10 One probe,known as the sink probe, withdraws fluids, whilethe other monitors the pressure transient. In theMDT tool, the two probe assemblies aremounted diametrically opposite each other onthe mandrel.11 One probe is a sink while the other,known as the horizontal probe, is solely a moni-tor with no sampling capability. The main pur-pose of the dual-probe module is to combine witha vertical probe to determine kh, kv and storativ-ity (øCt) through a local interference test or, touse a more specific name, the interval pressure-transient test (IPTT).12 By withdrawing fluidthrough the sink, three pressure transients canbe recorded at three different locations along thewellbore, two of which are from monitor probesand are not contaminated by the effects of toolstorage, skin and cleanup.13

The dual-packer module—This module hastwo packer elements that are inflated to isolate aborehole interval of about 1 m [3.3 ft]. Once theseare inflated, fluid is withdrawn, first from the iso-lated interval, and then from the formation. Sincea large section of the borehole wall is now opento the formation, the fluid-flow area is severalthousand times larger than that of conventionalprobes. This offers important advantages in bothlow- and high-permeability formations, and inother situations.

• Probes are sometimes ineffective when set inlaminated, shaly, fractured, vuggy, unconsoli-dated or low-permeability formations. The dualpacker allows pressure measurements andsampling in these conditions.

• Used alone, the dual packer makes a small ver-sion of a standard drillstem test (DST) that isknown as a mini-drillstem test, or mini-DST.Since the mini-DST opens up only 1 meter offormation, it acts as a limited-entry test fromwhich both kv and kh may be determined underfavorable conditions. Used in combination withone or more vertical probes, the dual packercan record an IPTT.

• The pressure drop during drawdown is typi-cally much smaller than that obtained with aprobe. Thus, it is easier to ensure that oil isproduced above its bubblepoint, and thatunconsolidated sands do not collapse. Also,with a smaller pressure drop, fluids can bepumped at a higher rate, so that for the sametime period, a larger volume of formation fluidcan be withdrawn and a deeper-reading pres-sure pulse created.

Economically, well tests are expensive fromthe point of view of both equipment and rigtime. Well tests are also undertaken to obtain afluid sample so that the incremental cost ofdetermining permeability may be small.However, obtaining high-quality permeabilitydata often requires long shut-in times and extraequipment such as downhole valves, gauges and flowmeters.4

Production tests and production history—An average effective permeability can beobtained from the flow rate and pressure duringsteady-state production, preferably from specifictests at different flow rates. Skin and othernear-wellbore effects have to be known orassumed. An average permeability can also bedetermined from production-history data byadjusting the permeability until the correct his-tory of production is obtained. However, in bothcases, the permeability distribution cannot beobtained reliably. In the presence of layering orheterogeneity, this is a highly nonlinear inverseproblem, for which there can be more thanone solution.

In the absence of other data, permeability isoften related to porosity. In theory, the relationis weak—there are porous media that havebeen leached to give high porosity with zeropermeability, and others that have been frac-tured to give the opposite. However, in practice,there do exist well-sorted sandstone reservoirswith a consistent porosity-permeability relation.Other reservoirs are less simple. For carbonaterocks in particular, microporosity and fracturesmake it almost impossible to relate porosity andlithofacies to permeability.

1. Zheng S-Y, Corbett PWM, Ryseth A and Stewart G:“Uncertainty in Well Test and Core Permeabilty Analysis:A Case Study in Fluvial Channel Reservoirs, NorthernNorth Sea, Norway,” AAPG Bulletin 84, no. 12 (December2000): 1929–1954.

2. Pickup GE, Ringrose PS, Corbett PWM, Jensen JL andSorbie KS: “Geology, Geometry, and Effective Flow,”paper SPE 28374, presented at the SPE Annual TechnicalConference and Exhibition, New Orleans, Louisiana, USA,September 25-28, 1994.

3. Herron MM, Johnson DL and Schwartz LM: “A RobustPermeability Estimator for Siliclastics,” paper SPE 49301,presented at the SPE Annual Technical Conference andExhibition, New Orleans, Louisiana, USA, September 27-30, 1998.

4. Modern Reservoir Testing. SMP-7055, Houston, Texas,USA: Schlumberger Wireline & Testing, 1994.

9. For the MDT tool these include: large-area packers fortight formations; large-diameter probes for unconsoli-dated as well as tight formations; long-nosed probes forunconsolidated formations and thick mudcakes; andgravel-pack probes and a large-area filter similar to anautomobile oil filter for extremely unconsolidated sands(the Martineau probe).

10. Proett MA, Wilson CC and Batakrishna M: “AdvancedPermeability and Anisotropy Measurements WhileTesting and Sampling in Real-Time Using a Dual ProbeFormation Tester,” paper SPE 62919, presented at theSPE Annual Technical Conference and Exhibition, Dallas,Texas, USA, October 1-4, 2000.

11. Zimmerman T, MacInnes J, Hoppe J, Pop J and Long T:“Applications of Emerging Wireline Formation TestingTechnologies,” paper OSEA 90105, presented at the 8th Offshore Southeast Asia Conference, Singapore,December 4-7, 1990.

12. The term vertical interference test (VIT) is also used forvertical wells. The terms local interference test andinterval pressure-transient test are appropriate for devi-ated or horizontal wells.Storativity is the product of porosity, ø, and total rockcompressibility, Ct, which is the sum of the solid com-pressibility, Cr, and the fluid compressibility, Cf . When notmeasured by an IPTT, Cf must be estimated from fluidproperties and Cr from knowledge of the solid frameworkbased on acoustic logs, porosity and other data. If thereis more than one fluid, the saturation of each fluid is esti-mated from logs or sample volumes.

13. Skin is defined as the extra pressure drop caused bynear-wellbore damage (mechanical skin), flow conver-gence in a partially penetrated bed, and viscoinertialflow effects (usually ignored). The flow-convergencefactor can be calculated from knowledge of bed thick-ness and test interval. Tool storage is due to the compressibility of the fluid inthe tool, and causes the measured flow rate to be differ-ent from the actual flow rate at the formation surface, or sandface. Cleanup refers to the increase in flow rateas the flow of fluids removes formation damage near the borehole.

Autumn 2001

50387schD02R1.p7.ps 12/10/01 4:43 PM Page 7

The pumpout module—This module pumpsfluid from the formation into the mud column, andfrom one part of the tool to another. Pumping intothe mud column allows much larger volumes offluid to be withdrawn than when sampling intofixed-volume sample chambers. The module canalso pump fluid from one part of the tool toanother; from the mud column into the tool, forexample to inflate the packer elements; or intothe interval between the packers to initiate asmall hydraulic fracture. For permeability mea-surements, the pumpout module is capable ofsustaining a constant, measured flow rate duringdrawdown, thereby simplifying considerably theinterpretation of pressure transients. The flowrate though the pump depends on the pressuredifferential, increasing at low differential to amaximum of 45 cm3/s [0.7 gal/min]. At very highdifferential, such as in a tight rock, the pump maynot be able to maintain a constant rate.

The flow-control module—This module with-draws up to 1000 cm3 [0.26 gal] of fluid from theformation while controlling and measuring theflow rate. The fluid withdrawn is either sent to asample chamber or pumped into the borehole.The module works in various modes such as constant flow rate, constant pressure andramped pressure, and can also draw repeatedpulses of fluid from the formation. The time forpulses to arrive at a vertical probe is an impor-tant input in the determination of kv. Since theflow-control module can control flow rate pre-cisely, it can regulate the withdrawal of sensitiveformation fluids into small-volume pressure-vol-ume-temperature (PVT) sample bottles. This isimportant for the sampling of condensate reser-voirs. (For more on sampling, see “QuantifyingContamination Using Color of Crude andCondensate,” page 24).

All these features provide many ways to mea-sure permeability, ranging from simple pretestdrawdown to multiple probes and dual packers(right). For the most reliable in-situ determinationof permeability and anisotropy, experience hasshown that interference tests should be per-formed with multiple pressure transients. Resultsfrom other methods will always be more ambigu-ous, but can still be useful, and even good, esti-mates in the right conditions. One such techniqueis the mini-DST.

8 Oilfield Review

Three probe(sink, horizontaland vertical)

• Analysis can be done without sink drawdown

Second verticalprobe

• Best configuration for layered reservoirs, faults and fractures

• Gives kh and kvDual packer+ probe ortandem probes

• Smaller vertical investigation than other IPTT configurations (sometimes an advantage)

• Longer tool

• Need to have a good idea of φCt

Probe • Simplest method of establishing communication with formation• Multiple probes can be added in one tool string

Dual packer • Easier to test fractured, vuggy and tight formations

• Difficult to get good tests in fractured, vuggy and tight formations (difficult to withdraw fluids, seal failures)• High drawdowns in low k/µ formations may release gas, complicating analysis

Drawdown • Automatic computation, available during acquisition

Buildup • Deeper radius of investigation than drawdown

• Fear (usually unjustified) of sticking or of releasing gas slug into borehole

• Data available while samplingDual-packermini-DSTor extendeddrawdown andbuildup withprobe

• Need a particular combination of formation properties and thickness to get both kv and kh

• At same flow rate as probe, less drawdown helps avoid gas and sanding• For same time period as probe, more fluid is withdrawn, creating deeper pulse

• Low drawdown may give insignificant signals at vertical probes in high k/µ formations

Flow Source

• Many (tens) of pretests often recorded for pressure, allowing qualitative comparisons

Probe Pretest

• Small volume of investigation (inches)• Measures effective permeabiliby to mud filtrate

• Many (tens) of pretests often recorded for pressure, allowing qualitative comparisons

• Small sampling volume, cleanup and tool storage can make analysis difficult• Measures effective permeability to mud filtrate, formation fluid or a mixture of the two

Single-Transient Analysis

• Gives ks and/or kh and can avoid costly DST

• Need to know φCt to get ks, and need to know h to get kh• Tool storage, skin, free gas and continuous cleanup can complicate analysis (especially with probe)

Dual-Transient IPTT

• The simplest configuration for an IPTT • Sink drawdown and early buildup affected by tool storage, skin, free gas and cleanup

Multiple-Transient IPTT

• Gives φCt as well as kh and kv

• Analysis can be done without sink drawdown

Advantages Limitations

> Features of the flow sources and methods used to derive permeability from the MDT tool.

50387schD02R1.p8.ps 01/10/2002 03:52 PM Page 8

14. In one recent job, the pumpout module was run continu-ously for 36 hours. In another job, the dual-packer mod-ule was in the hole for 11 days.

15. Ayan C and Nicolle G: “Reservoir Fluid Identification andTesting with a Modular Formation Tester in an AgingField,” paper SPE 49528, presented at the 8th Abu DhabiInternational Petroleum Exhibition and Conference, AbuDhabi, UAE, October 11-14, 1998.

16. Tool storage includes the compressibility of the fluidbetween the packers. A common model is to relate thesandface flow rate, qsf, to the measured flow rate, q, andthe rate of change of pressure by a constant, C: qsf =q+24Cdp/dt. The very early part of a buildup is dominatedby wellbore storage, also called afterflow. C can be esti-mated from the rate of change of pressure at this time.

Autumn 2001 9

Mini-DSTsIn a standard DST, drillers isolate an interval ofthe borehole and induce formation fluids to flowto surface, where they measure flow volumesbefore burning or sending the fluids to a disposaltank. For safety reasons, many DSTs require thewell to be cased, cemented and perforatedbeforehand. The MDT tool, in particular the dual-packer module, provides similar functions to aDST but on wireline and at a smaller scale.

The advantages of the mini-DST are less costand no fluids to surface. Cost benefits come fromcheaper downhole equipment, shorter operatingtime and the avoidance of any surface-handlingequipment. On offshore appraisal wells, cost sav-ings can be more than $5 million. With no fluidsflowing to surface, there are no problems of fluiddisposal, no surface safety issues and no prob-lems with local environmental regulations. Mini-DSTs are much easier to plan and can test multiplestations on the same trip—usually a sufficientnumber to sample the entire reservoir interval.

The mini-DST has disadvantages: it investi-gates a smaller volume of formation due to thesmaller packed-off interval (3 ft versus tens offeet), and withdraws a smaller amount of fluid ata lower flow rate. In theory, we may be able toextend the tests and withdraw large amounts offluid, but in practice, there may be a limit to howlong the tool can safely be left in the hole.14 Theactual depth of investigation of a wireline testerdepends on formation permeability and other fac-tors, but is of the order of tens of feet, rather thanthe hundreds of feet seen by a normal DST.

The smaller volume of investigation is notnecessarily a disadvantage. A full DST reveals theaverage reservoir characteristics and assessesthe initial producibility of a well. Permeabilityvariations will be averaged, and although theycontribute to the average, they are neitherlocated nor quantified. With the help of logs, thesmaller volume mini-DST can evaluate key inter-vals. The procedure for interpreting pressure tran-sients from mini-DSTs is the same as for full DSTsand the same software can be used for both.

TotalFinaElf used mini-DSTs in the Arab reser-voir of an aging Middle East field to look for zoneswith moveable oil and to calibrate the permeabil-ity anisotropy used in a simulation model.15 Sincethe packed-off interval rarely covers the wholereservoir, a mini-DST is a limited-entry, or partiallypenetrating, well test. To determine formationparameters, interpreters need to identify flowregimes in the buildup. In a homogeneous layer,there are three flow regimes: early radial flowaround the packed-off interval, pseudosphericalflow until the pressure pulse reaches a boundary,

and finally total radial flow between upper andlower no-flow boundaries. Rarely are all threeseen because tool storage effects can mask theearly radial flow, while the distance to the near-est barrier determines whether or not the otherregimes are developed during the test period.16

However, it has been common to observe a pseu-dospherical flow regime, and occasionally totalradial flow in buildup tests (below). On a log-logplot of the pressure derivative versus a particularfunction of time, spherical flow is identified by a slope of –0.5, and radial flow by a stabilized horizontal line.

Spherical permeability, ks=3√(k2hkv) can be

estimated from a pressure-derivative plot duringspherical flow or from a separate specialized

plot.17 Horizontal permeability, kh, can be esti-mated from a pressure-derivative plot duringradial flow, or from a specialized plot of pressureversus Horner time, provided the thickness of theinterval is known.18 In this case, the thicknesswas obtained from openhole logs, particularlyimages from the Schlumberger FMI FullboreFormation MicroImager tool. When both spheri-cal- and radial-flow regimes occurred, the inter-preters could estimate vertical permeability, kv,from kh and ks. These initial estimates were com-bined with the geological data to build a model offormation properties. Different analytical tech-niques, such as type-curve matching, were thenused to match the full pressure transient andimprove the permeability estimates.

0.1 1 10 100 1000Time since end of drawdown, sec

0.01

0.1

1

10

100

1000

Pres

sure

diff

eren

ce, p

sia,

and

der

ivat

ive

Sphericalflow

Radialflow

Type-curve parameters:kh = 39 mDkv = 24 mDµ = 1 cpThickness of zone = 8 mMechanical skin = 1.3

Measured pressure differenceMeasured derivativeModel pressure differenceModel derivative

> Pressure difference and the derivative of pressure withrespect to a function of time for the buildup at the end of a typi-cal mini-DST. The pressure difference is between the measuredpressure and a reference taken near the end of the drawdownperiod. The derivative is calculated from d∆p/dln[(tp+∆t)/∆t]where tp is the producing time and ∆t is the time since the end of the drawdown. We identify spherical flow by the slope of –0.5 on the log-log derivative, and radial flow by the slope of 0 (horizontal). The solid lines are the results of a type curve, ormodel, computed with the parameters in the table.

17. On a specialized spherical plot, the slope, msp duringspherical flow is given by: msp = 2453qµ(√µøCt)/ks

3/2 inoilfield units, where ø is usually taken from logs, and q,the flow rate, is measured or estimated. The viscocity, µ,is determined from the PVT properties of the mobile flu-ids. If there is more than one mobile fluid, their satura-tions are estimated from logs or sample volumes.

18. Horner time is [(tp+∆t)/∆t] where tp is the drawdowntime, and ∆t is the time since the end of the drawdown.The slope, mr , during radial flow is given by mr =162qµ/khh, where h is the thickness of the formationinterval, and the other terms are defined in reference 17.

50387schD02R1 11/29/01 4:24 AM Page 9

TotalFinaElf recorded ten tests in two wells,one of which was cored. Both kv and kh were sub-sequently measured on core plugs sampled every0.25 or 0.5 m [9.8 or 19.6 in.], and compared withthe mini-DST results (below). Care was taken toscale up the core data to the mini-DST intervaland to convert from absolute to effective perme-ability. For some of the tests, pressure-transientdata were also available from two probes in theMDT tool string, making it possible to comparemini-DST results with results from a full IPTT as

well as from core samples. The IPTTs measurelarger volumes of formation, yet the results gen-erally agree with the mini-DST, especially for thenear probe. The fact that the different measure-ments agree suggests that the formations may berelatively homogeneous, or that the scaling up ofthe core data was appropriate. While this goodagreement validates the use of a mini-DST inthese conditions, it is inadvisable to assume thesame degree of homogeneity in other formations.

Cased-Hole Mini-DSTsPhillips Petroleum, operating in the Peng Lai fieldoffshore China, found that cased-hole mini-DSTswere a valuable complement to full DSTs andopenhole WFTs in evaluating their reservoir.19

Like many operators, they initially ran mini-DSTsto obtain high-quality PVT samples, but thenfound that the pressure-transient data containedvaluable information. Peng Lai field consists of aseries of stacked, unconsolidated sandstonereservoirs with heavy oil—11° to 21° API—oflow gas/oil ratio (GOR), whose properties varywidely with depth. Testing each reservoir in eachwell with full DSTs was proving expensive, andwas not always successful. Among other factors,the handling of the heavy oil at surface causedeach DST to last between five and seven days.Large drawdowns, which were sometimesneeded to lift the oil to surface, caused the for-mation to collapse and the near-wellbore pres-sure to drop below the bubblepoint. As a result,mini-DSTs were an attractive alternative for allbut the largest zones.

With a probe, the drawdowns were too high,while unstable boreholes and high pressure dif-ferentials made openhole wireline testing with adual-packer module risky. Phillips’ answer was torun the dual packer in cased holes. By the end of2000, they had performed 27 cased-hole mini-DSTs in seven wells. In one typical test, theyidentified a 3-ft low-resistivity zone that was iso-lated from the main reservoir at the well by thinshales above and below (next page, left). Aftercement isolation was checked, a 1-ft [30-cm]interval was perforated, and the MDT dual pack-ers were set across it. Communication wasestablished, and the formation fluid was pumpedinto the borehole until the oil fraction stabilized(next page, top right). Two oil samples weretaken, and after an additional drawdown, a pres-sure buildup was recorded over 2 hours. The totaltesting time of 16 hours would normally be con-sidered excessive and risky in openhole condi-tions, but presented no problem in cased hole.

The pressure derivative during buildup showsa short period of probable spherical flow fol-lowed by a period of radial flow (next page, bottom right). With initial values of ks and kh fromflow-regime identification, the buildup data werematched with a limited-entry model, assuming aformation thickness of 3 ft with no outer bound-aries. The match is excellent. The high horizontalpermeability (2390 mD) and the low vertical per-meability (6 mD) were not surprising for thiszone. Overall, a zone that looked doubtful on logsproved not only to be oil-bearing but also to haveexcellent producibility.

10 Oilfield Review

0

0

400

Perm

eabi

lity,

mD

600

500

300

200

100

0

Test number

Horizontal Permeability

0

5

10

15

20

25

30

35

Perm

eabi

lity,

mD

Test number

Vertical Permeability

1 2 3 4 5

1 2 3 4 5

Verticalpermeability

IPTT (V2)

Mini-DSTCoreIPTT (V1)

IPTT (V2)

Mini-DSTCoreIPTT (V1)

> Comparison of the horizontal (top) and vertical (bottom) permeabili-ties measured by mini-DSTs, cores and IPTTs. The core data wereaveraged over each mini-DST test interval and converted to effectivepermeability using relative-permeability curves. Arithmetic averagingwas used for horizontal permeabilities, and harmonic averaging forvertical permeabilities. The IPTT data are from the same tests as themini-DSTs, but using two probes: V1 at 2 m [6.6 ft] and V2 at 4.45 m[14.6 ft] above the packer interval. The intervals tested are thereforedifferent. In this case, the agreement between the different measure-ments is generally good.

50387schD02R1.p10.ps 11/17/01 6:02 PM Page 10

Autumn 2001 11

Mini-DST LimitationsIn spite of these good results, the permeabilitymeasurements have some limitations. The lack ofan observation probe means that the only pres-sure transient comes from the pressure sink,which is affected by skin and tool storage. Bothskin and storage influence the early part of thebuildup and make identification of flow regimesand interpretation more difficult. Later in thebuildup there needs to be the right combinationof formation properties and bed thickness for sig-nificant periods of both spherical and radial flowto be observed. The radial-flow interpretationdepends directly on identifying bed boundaries,while spherical-flow interpretation depends onknowing the storativity. Thus, it is difficult todetermine both kv and kh simultaneously.

Finally, several factors can make a singletransient hard to interpret. These include gasevolution near the wellbore, pressure and flow-rate variations due to continuous cleanup, andnoisy drawdown pressures from pump strokes.Pressure measurements at observation probesare not usually affected by these phenomena.Since these probes are higher up the string, they also increase the volume investigated.

19. Hurst SM, McCoy TF and Hows MP: “Using the CasedHole Formation Tester for Pressure Transient Analysis,”paper SPE 63078, presented at the SPE Annual TechnicalConference and Exhibition, Dallas, Texas, USA, October1-4, 2000.

Gamma Ray Resistivity Porosityohm-mAPI0 150

SP-100 0mV

1 1000 45 0p.u.

Dept

h, ft

X00

X10

X20

X30

X40

X50

X60

Perforations

> Gamma ray, resistivity and porosity logs acrossa low-resistivity reservoir in the Peng Lai field,offshore China. The mini-DST was performed in athin 3-ft zone that is isolated above and below bythin shale beds (gray) within a larger reservoir.Any oil found in this zone was expected to beabout 13º API with high viscosity.

Initial buildup

Pres

sure

, psi

a

1700

1600

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

300-rpm constant pump rate

Time, hr

Oilbreakthrough

SamplingBuildup

0

300

600

1800

Pum

p ra

te, r

pm

> Pressure and pump rate during the cased-hole mini-DST from PengLai field. After communication was established with the formation,the pump withdrew invasion fluids until oil broke through. Once theoil fraction had stabilized (as measured by the OFA Optical Fluid Analyzer tool, not shown), two samples were taken. After one additionaldrawdown, a 2-hr buildup was recorded. Minimum drawdown pressurewas 164 psi [1130 kPa], at or above the expected bubblepoint pressure,thereby avoiding free gas. The solid pressure line is the result predictedby the limited-entry model.

Pressuredifference

Radialflow

Sphericalflow

Pressurederivative

Pres

sure

diff

eren

ce, p

sia,

and

der

ivat

ive

1

10

100

1000

0.0001Time since end of drawdown, hr

0.001 0.01 0.1 1 10

Model parameters:kh = 2390 mDkv = 6 mDµ = 300 cpThickness of zone = 3 ftSkin = + 5.5Depth of investigation = 80 ft

> Pressure difference and derivative for the buildup at the end of thePeng Lai test. Spherical flow is identified by the slope of –0.5 on thederivative and radial flow by the slope of zero. The solid lines are thepredictions of a limited-entry model using the parameters in the table.

50387schD02R1.p11.ps 11/17/01 6:02 PM Page 11

IPTTs have proved to be an effective means fordetermining permeability distribution near thewellbore; in fact, they are the preferred methodfor layered systems. Mini-DSTs are usually runwhen the main objective is to recover a fluidsample, or to measure reservoir pressure, partic-ularly in tight or heterogeneous formations.Permeability is an additional parameter withwhich to judge the producibility of the interval.

Interval Pressure-Transient TestAn IPTT run in a carbonate reservoir in the UnitedArab Emirates (UAE) illustrates the sequence ofoperations and methods employed in a full anal-ysis.20 This reservoir has distinct, contrasting lay-ers that appear to extend over large areas.Reservoir management and the design of sec-ondary-recovery schemes depend strongly onknowing the vertical and horizontal permeabili-ties and the communication between layers. Inparticular, the implementation of an injectionscheme depends on the permeability of severallow-porosity, stylolitic intervals. Will the stylo-lites act as baffles to injected fluid and severelyaffect sweep efficiency?

The stylolitic intervals may be thinner than1 ft, but can be observed on logs and cores (left).However, their effectiveness as barriers is notclear. They can be correlated between wells, buttheir lateral continuity and permeability areuncertain. Cores could not be recovered from

12 Oilfield Review

0

Porosity, p.u.05 10 15 20 25 30 35

X180

X190

X200

X210

X220

X230

X240

X250

X260

X270

X280

X290

X300

X100

X110

X120

X130

X140

X150

X160

X170

Dept

h, ft

UAE Carbonate Porosity

UAE Carbonate Permeability

kh (Core)mD

kh (Layered Model)

kv (Layered Model)

0.1 1000

0.1 1000mD

mD

0.1 1000

No CorePermeability

Layer No.

31

30

2928

27

262524

23

2221

20

19

1817

16

1514

13

12

11

1098 7

6

4

5

321

< Log porosity in a layered carbonate (left). Thelow-porosity streaks are stylolites. The positionsof the packer and the probes at each test loca-tion were chosen to straddle the stylolites. Theright track shows the layered model used tointerpret the IPTTs, with kv and kh from the modeland kh from core. Core permeability is generallytoo high and is either absent from the stylolites or fails to reflect the large contrasts seen by theIPTT. The FMI image (left) shows two low-poros-ity streaks (white) separated by a dark interval.The top streak is particularly patchy. The layeredmodel used to match the IPTT showed that thetop streak had higher kv than kh, while the centerinterval had very high permeability.

20. Kuchuk FJ, Halford F, Hafez H and Zeybek M: "The Use ofVertical Interference Testing to Improve ReservoirCharacterization," paper ADIPEC 0903, presented at the9th Abu Dhabi International Petroleum Conference andExhibition, Abu Dhabi, UAE, October 15-18, 2000.

50387schD02R1.p12.ps 11/17/01 6:25 PM Page 12

Autumn 2001 13

many of these intervals, and, in any case, give avery local value of the permeability. The operatordecided to investigate the stylolites with a seriesof IPTTs in a new well. These could be recordedon a single trip in the hole, allowing the completereservoir section to be tested efficiently.

An IPTT needs a minimum of one verticalobservation probe and a sink, either a dual-probeor a dual-packer module. In this case, in order tosample more layers, the MDT tool was equippedwith two vertical observation probes at 6.4 ft and14.4 ft [1.95 and 4.4 m] above the center of the

packer interval. The dual-packer module waschosen so as to generate a sufficiently largepressure change at the far probe. The pumpoutmodule was used to withdraw formation fluidsfrom each tested interval. Pressures were mea-sured by quartz-crystal and strain gauges at bothprobes and packer.

Sequence of operations—Using openholelogs, the operator selected six test locations,with the depths chosen so that the stylolites laybetween the dual packer and near probe. At eachtest location, the operator followed the samesequence of events: set the packers and probes,

pretest probes and packer interval, drawdown,buildup, and retract packers and probes (above).The pretests measured formation pressure andestablished communication with the formation.Once communication was established, formationfluids were withdrawn through the packer inter-val at an almost constant rate for between 30and 60 minutes. The rate was slightly differentfor each test, but remained between 15 and 21B/D [2.4 and 3.3 m3/d]. After each drawdown, theinterval was shut in for another 30 to 60 minutes.

Pack

er p

ress

ure,

psi

3600

3800

4000

4200

0 40001000 2000 3000

Time, sec

20

15

10

5

0

Flow

rate

, B/D

Flow rate

Pressure

Tool setting Pretest Drawdown Buildup

Toolretraction

Pack

er p

ress

ure,

psi

3800

3600

0 40001000 2000 3000

Time, sec

3880

3890

3900

3910

Prob

e pr

essu

re, p

si

4000

4200

3920

3930

Packer

Probe 1

Probe 2

> The sequence of events in a typical IPTT, as shown by the pressureand the flow rate recorded in the dual-packer interval (top). After toolsetting, the pretest establishes communication with the reservoir bywithdrawing up to 1000 cm3 [60 in.3] through the packer and 20 cm3

[1.2 in.3] through each probe. During drawdown, the flow rate is con-stant since it is controlled by the pumpout module. During the buildupperiod, the pressure is recorded for a sufficiently long time, approxi-mately the same as the drawdown period, to ensure good pressure-transient data. At the end of the buildup period, the probes and packerare retracted. Packer and probe pressures were recorded with CQGCrystal Quartz Gauge pressure gauges during the IPTT (bottom). Notethe much more sensitive scale for the probe pressures. Their finalbuildup pressure is lower because they are higher in the well. Notealso the distinct delay in the start of the buildup on Probe 2, due to thelow vertical permeability. The delay on Probe 1 cannot be seen at thistime scale. The packer pressure is slightly noisy due to pump movement.

50387schD02R1 12/21/2001 02:57 PM Page 13

In this test, packer pressure dropped sharplyby approximately 300 psi [2070 kPa], while near-probe pressure dropped more slowly by 10 psi [69 kPa] and far probe by 2 psi [14 kPa]. Theseresponses give a first idea of permeability. Thefact that there is a response at the verticalprobes showed that there was communicationacross the stylolite.

Analysis—Interpretation starts with a look ateach test independently. As with mini-DSTs, thefirst step is to analyze flow regimes. Buildups arepreferred to drawdowns because they are less

affected by near-wellbore factors, such ascleanup and pressure fluctuations caused by thepumpout piston. The interpreter examined eachof the three pressure transients from the sixtests, and established some initial estimates ofpermeability. Because of the highly stratifiednature of this carbonate formation, these esti-mates were rough averages of the permeabilitynear each station.

The heart of the interpretation is a realisticmodel, layered in this case, with permeabilities,porosities and thicknesses for 31 layers (above).Initial layer boundaries and thicknesses aredetermined from the logs, actually from high-res-

olution images since layers as thin as 0.5 ft [15 cm] may play an important role. Porosity androck-framework compressibility are based on logdata; fluid compressibility and viscosity comefrom fluid saturations and PVT analysis. Initialhorizontal and vertical permeabilities are takenfrom the flow-regime analyses and other avail-able sources—cores, logs and pretests. Initialestimates are also needed for tool storage andskin around the packer.21 Finally, the flow rateduring drawdown is an important input; in thiscase, it was measured and was taken to beessentially constant during most tests.

14 Oilfield Review

mDmDmDftNumberConfidence CommentsPorosity

kvkhCore khThicknessLayer

low0.2165989771

moderate dense zone0.150.0210.1_22

high high permeability0.27610610_63

moderate0.2635687874

low0.28162633105

low0.2848676186

low0.1839534627

low0.152832190.58

moderate patchy stylolite0.1411.10.9_0.59

high superpermeability0.277251350_410

moderate0.283175811211

low0.26142430812

low patchy stylolite0.233.89.92.7214

high0.295.415.616515

high0.312.911.318716

high dense zone0.111.31.49.3217

high0.292.36.713718

high0.283.569.4619

high0.37.87.412.3820

high0.253.53.312.1321

high dense zone0.191.11.3_222

high0.23.23.2_823

high0.286.47.98.6424

high patchy stylolite0.23.819.819.1125

high0.282.35.416626

high0.294.611.410527

high0.283.16.811728

high dense zone0.190.890.1_129

high0.2814.211.32230

high dense zone0.10.450.91.41431

26 low0.26468-60913

31-Layer Model

> Model with 31 layers used for interpreting pressure transients. Each layer is assigned a thickness,vertical and horizontal permeability, porosity, and level of confidence.

50387schD02R1.p14.ps 11/17/01 6:02 PM Page 14

Autumn 2001 15

With these initial estimates, the expectedpressure transients at the packer and the twoprobes are computed and compared with themeasured transients during drawdown andbuildup (above). An automatic optimization proce-dure adjusts the model parameters to minimizethe differences over all transients. The main goalis to obtain the best kv and kh for the layers nearthe station. Bed boundaries are changed manu-ally if necessary, while, in this case, øCt was

known well enough to be fixed. Permeabilities oflayers away from the station may affect results tosome extent but are not allowed to change signif-icantly. Flow rate is held closely to the measuredrate, but is still computed so as to allow for toolstorage and the effect of small flow-rate changeson the transients.

When the results are not satisfactory, the geo-logical model is reexamined with the geologist,redefining some layers and changing some initialestimates. Different weights can be applied to

different time periods and different transients. Forexample, the packer drawdown period mightreceive less weight because, unlike observation-probe pressures, it is affected by the noise asso-ciated with production and variable cleanup.

The interpreter applied the model to each testin turn. However, this was not the end, sincesome tests were conducted close enough to eachother that changing the parameters in the vicinityof one may have altered the results from another.

21. Since the flow rate into the probe is negligible, the skinand tool storage at the probe can be ignored.

Probe

Packer

P

Probe

Packer

t

P

t

Flow-regimeidentificationand analysis

Modeldefinition

Skin, storage constants,formation pressures,flow rates

Single-layermodel

Multilayermodel

Other data

MDT-measured data

Probe

Packer

Probe

Packer

Computetransientsfrom model

Measured data

Adjust model tominimize differencebetween computedand measured data

Probe

Packer

Probe

Packer

Probe

Packer

Computed data

khkvφCt

khkvφCt

khkvφCt

khkvφCt

Pretestanalysis

PressuretransientFlow rate

• Formation pressures• Drawdown permeabilities

Initial average• ks, if spherical flow• kh, if radial flow• kv,kh, if both

log∆t

∆P,∆P’

∆P,∆P’

log∆t

Fluid analysis: µ,Cf

Openholelogs: φ,Sw,Cr

Openhole logs,images: layers

> A typical workflow for the interpretation of an IPTT, with dual packer and one vertical probe. Each job is different, and the actual path taken depends on a trade-off between speed, complexity of problem and accuracy of results. Quickest, but least accurate results come from analyzing individual transients.Next may be analysis of all transients from one test with a single-layer model, then with a multilayer model. Adjusting the model to best match all the available data may require several iterations.

50387schD02R1.p15.ps 11/17/01 6:25 PM Page 15

Therefore, the optimized model was reapplied toeach test so as to achieve a good match betweenall measured and computed transients (left).Some layers were better defined than othersbecause there were more pressure transients intheir vicinity. For this reason, the confidence fac-tor for the bottom 15 layers, for which there werefour tests, was higher than for the top 15, inwhich there were only two tests.

Results—Overall, the interpreter performed atype of history matching in which the reservoirmodel was iteratively adjusted to match 18 pres-sure transients distributed along the wellbore.The estimated permeabilities differed consider-ably from core permeability, being generallylower and varying by several orders of magnitude,from almost 0.02 mD to 1350 mD. No core-derivedpermeability measurements were available fromintervals having these extreme values. On theother hand, the porosity varied little, exceptwithin stylolitic zones. As for most carbonate for-mations throughout the Middle East, porosity isnot a good indicator of permeability. Of the sixlow-porosity intervals on the logs, only two hadpermeabilities below 1 mD. Two others werepatchy with significant permeability, one with kv > kh at X151 ft. In this particular test, the smallpressure response at the probes (less than 0.5 psi[3.5 kPa]) could be explained only by a superper-meability layer between packer and probe. Thissurprising result was supported by an FMI imageof the stylolite, which showed a conductive layerbetween two dense streaks, one of which hadgaps in it (figure, page 12). None of this wasapparent from the core data.

16 Oilfield Review

Time, sec0 500 1000 1500 2000 2500

Time, sec0 500 1000 1500 2000 2500

Time, sec0 500 1000 1500 2000 2500

400

350

300

250

200

150

100

50

0

Pres

sure

diff

eren

ce, p

si

4

0

Pres

sure

diff

eren

ce, p

si

12

10

8

6

4

2

0

Pres

sure

diff

eren

ce, p

si

1

2

3

Probe 2

Probe 1

Packer

ComputedMeasured

ComputedMeasured

ComputedMeasured

Probe 2 (for reference)

> A comparison between the measured pressure-transient responseat the packer (bottom) and the two probes (top and middle), and theresponse computed from the layered model after nonlinear optimiza-tion of the parameters. The good agreement validates the parametersin the model. Other solutions may be possible, but were ruled out onthe basis of other data.

50387schD02R1.p16.ps 11/17/01 6:03 PM Page 16

Autumn 2001 17

The final model suggested that the layersshould communicate over time. Pressure commu-nication was confirmed by the formation pressuregradient from MDT pretests (left). The relativelyuniform gradient showed that the stylolites didnot act as pressure barriers. However, good pres-sure communication does not necessarily meanthat fluids will flow uniformly through the reser-voir. As the model showed, at least two high-permeability layers can act as conduits forinjected water. This information has been used in the full-field reservoir simulator, and to examine unexpected water breakthroughs in production wells.

Mapping StylolitesCarbonate rocks typically form in shallow, tropi-cal marine environments. In some cases, a for-mation can extend for hundreds of miles.Carbonate sediments contain significantamounts of the metastable minerals aragoniteand magnesium calcite; calcite itself is readilydissolved and reprecipitated by percolating porefluids. Carbonate rocks are, therefore, likely to

undergo dissolution, mineralogical replacementand recrystallization. These effects vary accord-ing to temperature, pore-fluid chemistry andpressure. Carbonate diagenesis commonlybegins with marine cementation and boring byorganisms at the sediment-water interface priorto burial. It continues through shallow burial withcementation, dissolution and recrystallization,and then deeper burial where dissolution pro-cesses, known as pressure solution, may formsuch features as stylolites and vugs (below).

The resulting diagenetically altered zones,whether of lower or higher permeability than thesurrounding formation, are frequently extensiveand affect large sections of a potential reservoir.For this reason, such features detected by bore-hole measurements often can be extrapolatedsome distance from the well.

The first IPTT example showed how the per-meability of stylolites could be determined in asingle well. The next question is how far the lay-ers extend across the field. The depth of investi-gation of an IPTT depends on transmissivity(khh/µ) and storativity, and varies with each test.

8300

Dept

h, ft

8250

8200

8150

8100

3840 3860 3880 3900 3920Pressure, psi

0.34 psi/ft

> Pressure profile from MDT pretestsacross the reservoir. The pretests weretaken at the packer and probes as part ofeach IPTT. The reservoir has been on pro-duction for nearly 20 years. After thismuch production, any barriers to pressurecommunication should cause the pres-sure gradient to be much less uniform.However, the lack of pressure barriersdoes not necessarily mean that fluids willflow vertically with ease.

> Large dissolution cavity. Although carbonates can have large dissolution cavities, they arenot always as large as this.

50387schD02R1 11/29/01 4:43 AM Page 17

In the previous example, the depth of investiga-tion ranged from about 20 to 30 feet [6 to 9 m].The next example, from another field in the UAE,examines the lateral extent of barriers by runningIPTTs in several adjacent wells (right).22 The low-porosity, dense stylolites can be correlated easilybetween wells, but their actual density varies, soit is quite possible that their permeability alsovaries. The size and number of stylolites areobserved to increase towards the flanks andtoward one side of the field.

A total of 23 IPTTs was recorded in sevenwells in two areas in which pilot gas-injectionschemes were to be implemented. The mainobjective was to determine the vertical perme-ability of four stylolites—Y2, Y2A, Y3 and Y4.

In this case, the MDT tool was configuredwith four probes (next page, top). A sink probe Screates the transient, which is measured by ahorizontal observation probe H at the same depthbut diametrically opposite the sink, and twoobservation probes V1 and V2 vertically displacedfrom the sink by 2.3 ft and 14.3 ft, [0.7 and 4.4 m].With this configuration, the storativity, øCt, neednot be assumed in the permeability analysis,since it can be determined directly from the tran-sients. An FMI image, recorded after the tests,clearly showed the imprint left by the probeassemblies on the borehole wall. The tool can beseen straddling two stylolites. In some tests, theflow-control module was used to give a constantflow rate. In others, formation fluids were with-drawn using the pumpout module for a longertest. Thus, as in the last example, a measuredflow rate was generally available for each test.

In some tests, the sink probe could not with-draw fluids as it was set against a highly local-ized tight spot. In these cases, the operation waschanged to withdraw fluids from the V1 probe,using S and V2 as the observation probes. Morerecently, interval tests in carbonates have beenperformed with the dual packer because its pro-duction interval is several thousand times that ofa sink probe. Fluid withdrawal is then possibleeven with a high degree of heterogeneity and inrelatively low permeabilities.

18 Oilfield Review

22. Badaam H, Al-Matroushi S, Young N, Ayan C, Mihcakan Mand Kuchuk FJ: “Estimation of Formation PropertiesUsing Multiprobe Formation Tester in LayeredReservoirs,” paper SPE 49141, presented at the SPEAnnual Technical Conference and Exhibition, NewOrleans, Louisiana, USA, September 27-30, 1998.

A

B

C

D

EF

G

North pattern

South pattern

XI

XII

XIIIA

XIIIB

XIV

XV

XVI

Y2

Y2A

Y3

Y4

Y5

Y1

Stylolites analyzed

> Field with two pilot gas-injection schemes planned, one in the north, andthe other in the south. The design depended heavily on the properties of thestylolites, Y1 through Y5. These zones could be easily identified on densitylogs and could also be correlated fairly easily across the reservoir. However,their properties varied, and it was not clear how effective they were as barri-ers to flow. IPTTs were recorded in seven wells (A through G) to quantify andmap their properties correctly.

A GB C D E F

Y2

Y2A

Y3

Y4

0 0 - 0.3 0.3 -1 1-3 3 -10 >10 Nottested

Permeability, mD

Styl

olite

Well

> Vertical permeability of the four main stylolitic intervals as found by 23 IPTTs run in seven wells.

50387schD02R1.p18.ps 11/17/01 6:03 PM Page 18

Autumn 2001 19

The interpretation began, as before, by flow-regime identification and analysis. Because of thelarge volume of data, each test was initially inter-preted assuming a single homogeneous butanisotropic layer. This interpretation is quickerand gives an average kh and kh/kv over some inter-val of reservoir rock containing the stylolite. Later,a more complete study was undertaken using amultilayer model as in the previous example.

The results showed considerable variationbetween the wells (previous page, bottom). In general, the stylolites were not absolute barri-ers to flow. For example, the Y2 stylolite wasfound to be a barrier in the south of the area, inWells F and G, but very conductive in Well E. TheY2A stylolite was also very conductive in Well E.FMI images showed that the stylolite and its adja-cent layers had a significant number of vugs, afeature not captured by the cores. Cores generallyfound a higher kh than did the IPTT but missed the vuggy intervals entirely (below). The IPTTquantified the degree of hydraulic communi-cation and allowed better planning of the pilot gasflood scheme.

Stylolite

Stylolite

X125

X150

X175

X200

Discontinuousstylolite

Dept

h, ft

Porouslimestone

Volume, %100 0p. u.

Unmoved Oil

Limestone

Anhydrite

Dolomite

Moved Oil

Clay

Water

V2

V1

SH

< Volumetric analysis (left) and the four-probeMDT tool (middle) set across the Y3 styloliticinterval in Well F. The FMI image (right) was runafter the tests and shows clearly the imprint (cir-cled in green) of the four probe assemblies attwo different tool locations.

MDT

laye

r per

mea

bilit

y, m

D

0.1

1

100

0 10 1001

10

Core-plug permeability, mD

Well E - Y4Well G - Y2AWell E - Y2AWell D - Y2Well E - Y2 Layers with

vugs in Well E

> Comparison of kh from core plugs with kh from thecorresponding layers of the IPTT interpretation. Thecore values were obtained by arithmetic averagingof the samples within the IPTT interval and by con-verting from absolute to effective permeability. In aperfect match, points would lie on the dotted line.Core-derived kh is generally higher. The core data donot capture effectively the vuggy layers of Well E.

50387schD02R1 11/29/01 4:43 AM Page 19

Anisotropy in SandstonesSandstones also pose questions about verticalpermeability and barriers to flow. AnadarkoAlgeria’s plans for the development of HassiBerkine South field called for injection of bothmiscible gas/water and possibly water-alternat-ing-gas (WAG) in the future (left). They needed toknow the permeability anisotropy in the field toimprove confidence in the vertical sweep effi-ciency, and in the recovery values being pre-dicted from numerical models. This informationwas required early in the appraisal-drilling pro-gram as it affected decisions on facilities andinfrastructure. The reservoir is in the TriassicArgilo-Gréseux Inferior (TAGI) sandstone.23 TheTAGI is fluvial in origin, with sands that are 5 to15 m [15 to 50 ft] thick. The area of interest hastwo major rock types: a fine- to very fine-grainedsand with interspersed shale laminae, and a fine-to medium-grained braided-stream deposit withdiscrete claystone layers (next page).

Upon reinjection, gas and water will be takenmainly by the high-permeability layers. It wasimportant to determine the degree of gravity seg-regation expected in the TAGI, and the corre-sponding influence on vertical sweep, oilrecovery and future production performance.

20 Oilfield Review

km 500

miles 300

ALGERIA

TUNISIA

LIBYA

HassiBerkineSouth

ALGERIA

> The Hassi Berkine South field in Algeria operated by Anadarko.

2200.00

2198.95

2200.00

2199.852193

2200

2200

2184

2200

2198

2200

2040

Probe Probe Probe

Packer Packer Packer

Time, sec0 500 1000

Time, sec0 500 1000

Time, sec0 500 1000

Time, sec0 500 1000

Time, sec0 500 1000

Time, sec0 500 1000

kh = 10

kh = 10

kh = 100

kh = 100

kh = 1000

kh = 1000

Pres

sure

, psi

Pres

sure

, psi

Pres

sure

, psi

Pres

sure

, psi

Pres

sure

, psi

Pres

sure

, psi

kh/kv = 100kh/kv = 10kh/kv = 1

kh/kv = 100kh/kv = 10kh/kv = 1

> The pressure response at a dual packer and a vertical probe 6.6 ft [2 m] higher during a drawdown followed by a buildupmodeled for different horizontal permeabilities and anisotropies, but the same flow rate. Note the expanding pressure scale foreach plot from low kh on the left to high kh on the right. Higher kh reduces the signal (causes a smaller pressure drop) at bothpacker and probe. Higher kh/kv reduces the signal at the probe but increases it at the packer. The response is complex andsometimes paradoxical. For example, at the end of a very long flow period, the pressure drop at the vertical probe depends onlyon kh, while the drop at the dual packer depends on both kh and anisotropy. Also, no signal at the vertical probe can mean thatthere is a layer of either zero or infinite permeability between it and the dual packer. These paradoxes partly explain why simpleanalytical solutions are not reliable.

50387schD02R1.p20.ps 11/17/01 6:03 PM Page 20

Autumn 2001 21

For the reservoir engineers simulating the gasinjection, the most critical parameter was theanisotropy, kh/kv. They were not confident in theanisotropy from cores, around 10, as this valuewas unexpectedly low for such a depositionalenvironment. The claystone layers were a partic-ular worry since they seemed to extend acrossthe field. An IPTT offered an attractive solution. Itwould test the anisotropy on a much larger scalethan cores, and would provide permeability val-ues at nearly the same vertical scale as the gridblocks used in the numerical simulation.

Four stations were planned—two in the fine-grained, lower resistivity layer; two in themedium-grained layer, one of which wasdesigned to straddle a thin claystone.

Permeabilities are high, so as part of the pretestplanning it was important to check that sufficientpressure changes would be seen at the monitorprobe. Using expected values for permeabilityand other parameters, simulations showed that ifthe flow-control and pumpout modules wereused as flow-rate sources, the resulting pressurepulse at the monitor probe would be barely mea-surable (previous page, bottom). A higher flowrate, and hence a larger pressure response, couldbe obtained by flowing directly to a samplechamber. This is clearly desirable unless it drawsgas out of solution or causes sanding. After fur-ther modeling and checking experience else-where, the operator ran tests with the dualpacker connected directly to the sample chamber.

The interpreters analyzed each test with asingle-layer model, treating the entire 15-msandstone as one layer. With no flow-rate mea-surement available, a special approach to theanalysis had to be taken. In this approach, theprobe pressure transient is used to estimate kv

and kh, while the packer transient is used to esti-mate the flow rate and packer-interval skin. Sincethe estimates are interdependent, it is necessaryto iterate between the formation parameters atthe probe and the flow rate and skin at the packeruntil the results converge.

23. Peffer J, O’Callaghan A and Pop J: “In-Situ Determinationof Permeability Anisotropy and its Vertical Distribution—A Case Study,” paper SPE 38942, presented at the SPEAnnual Technical Conference and Exhibition, SanAntonio, Texas, USA, October 5-8, 1997.

Dept

h, m

XX30

XX40

XX50

MDT

Core

1 100

1 100

Gamma Ray

Caliper

in.4 20

API0 140

Probe Pressure (Quartz)psi 51505110

Drawdown Mobility

1 3000mD/cp

AIT Resistivityohm-m1 3000

Anisotropykh/kv

Volumetric Analysis0 1vol/vol

Water

Oil

Sandstone

Bound Water

Clay

Layer 1

Layer 2

0.1 mm

0.1 mm

1 3000

Horizontal Mobilityfrom IPTT, mD/cp

> The two layers of the15-m TAGI sandstone. Layer 1 is fine-grained with shale laminations; Layer 2 is a medium-grained massive sandstone withthin claystone beds. The two IPTTs in Layer 1 both give horizontal mobilities below 100 mD/cp and moderate anisotropy. In Layer 2, both tests showhigh horizontal mobility, but the top test has low anisotropy, while the bottom test has high anisotropy, most likely due to the thin clay (green highlightin Track 4) at XX40.2 m between packer and probe. The average core anisotropy is similar, but slightly higher.

50387schD02R1 11/29/01 4:43 AM Page 21

The resulting permeabilities reflected the aver-age properties of the formation near each station.The results near the top two stations were similar,with horizontal mobility (permeability/viscosity)near 50 mD/cp and anisotropy near 10. The bot-tom two stations lay in the medium-grained layer.They both showed high horizontal mobility, butwhile the third station was nearly isotropic, thefourth station showed a much higher kh/kv.Assuming that the third station defines the prop-erties of the clean sandstone, it seems likely thatthe fourth station is affected by the thin clay atXX40.2 m, which lies between probe and packer.Assuming also that the clay acts as an imperme-able disk lying around the wellbore, we can esti-mate its radius as 2 m [6.6 ft].24 By this estimate, itis quite limited in extent.

The entire TAGI interval in this well wascored, with horizontal permeability measure-ments made on plugs every 15 to 30 cm [6 to12 in.], and vertical permeabilities about everymeter. When the core permeabilities were aver-aged over the 2-m interval of each MDT station,they compared well with MDT results, both indi-cating anisotropy less than 100.25 When shalelaminae or claystone beds are absent, theanisotropy is less than 10. These results were

further supported by five whole-core samplesfrom other wells in the field.

The MDT data were analyzed further with atwo-layer model, the only multilayer model avail-able at the time. The results were similar. Ideally,a model with at least five layers is needed to sim-ulate the whole formation. However, in this caseof relatively homogeneous formations, the opera-tor obtained answers that were sufficiently fit forthe purpose with the simpler single-layer model.

The MDT results increased confidence in theanisotropy values that reservoir engineers wereusing for numerical modeling, and thus also inthe predicted performance of the planned injec-tion scheme. In fact, the MDT-measured valueswere used directly in the simulator. The field hasbeen on production since early 1998, producingin excess of 70,000,000 barrels [11,123,000 m3].The MDT-derived anisotropy values continue tobe used in the simulator, since the history matchbetween actual field performance and predic-tions from the simulator have been excellent.Although in this case the core anisotropy dataproved to be broadly correct, the confirmation ona much larger scale was a key piece of informa-tion gathered during the appraisal of the field.

Horizontal WellsOperators rarely acquire permeability data in hor-izontal wells for reservoir description. However,horizontal wells often fail to live up to expecta-tion. Some of the many causes are related toreservoir heterogeneities. In one horizontal well,6 IPTTs and 19 pretests were run to investigatewhy neighboring wells had performed below par(above).26 Two major features were observed thatcould cause poor production—the variation inreservoir pressure, dropping by as much as100 psi [689 kPa] in the middle of the well; andthe variation in permeability from 5 to 50 mD forfairly constant porosity. Clearly, the middle inter-val has been more depleted and received lesssupport from water injection into the reservoir.Upon completion, the middle interval is predictedto clean up less easily, while injection water willprobably break through first at the toe, or end, ofthe well. For these reasons, it was recommendedto complete the well with a casing.

IPTTs are particularly useful for evaluating theconductivity of faults and fractures in horizontalwells. Interpreting conventional well tests is dif-ficult due to strong crossflow from pressure andpermeability variations. Borehole images candetermine the location of faults and fractures,

22 Oilfield Review

Horizontal displacement, ft0 2000 4000 6000 7000

6800

TVD,

ft

6900

Rese

rvoi

r pre

ssur

e, p

si

4200

4150

4100

4050

4000

6820

6840

6860

6880Pe

rmea

bilit

y, m

D

10

1

1000 3000 5000

0.4

Poro

sity

0.3

0.2

0.1

0

Frac

ture

test

Well trajectoryPorosityPretest (khkv)1/2

Interval test (khkv)1/2

Pressure

> Reservoir pressure and permeability from the MDT tool in a horizon-tal well. Permeability is measured by both pretest drawdowns andinterval pressure-transient tests, the latter being generally an order of magnitude higher. The pretest permeability may be low due to for-mation damage or because it is measuring the effective permeabilityto filtrate in a water-wet reservoir. Porosity is from openhole logs.Between 1765 and 5266 ft horizontal displacement, the pressure is significantly lower than elsewhere, indicating higher depletion andpoorer pressure support from water injection in the reservoir.

50387schD02R1.p22.ps 11/17/01 6:03 PM Page 22

Autumn 2001 23

and whether or not they are mineralized. In thiswell in a carbonate reservoir, images showedmany vertical fractures but could not determinetheir hydraulic conductivities. Pressure differ-ences indicated that while some were closed,others may have been open. Open fractures couldharm production by quickly drawing water upinto the well.

To test the fractures, the MDT tool was setwith a dual-packer module straddling a set offractures seen at 2983 ft (below). The logarithmicderivative with respect to Horner time for thebuildup test at the packer location indicates atool storage-dominated period that ends with ashort slope of –1.0 at 0.015 hr. Following thestorage period, the derivative exhibits a –0.5slope spherical-flow regime until 0.15 hr, afterwhich the derivative goes downward, indicatinga higher permeability region. The probe buildupderivative also exhibits a short spherical-flowregime, though its value is lower than that of thepacker test. The fact that the probe derivative islower but ends at the same time at both packer

and probe indicates a conductive fracture to theleft of the probe. The fracture or fractures musteither be short or have a finite conductivitybecause the derivative decreases only gradually.In addition, the best match to the transients wasachieved with a positive skin—another indica-tion that the fractures opposite the packer werenot open.

All the major fracture intervals were analyzedin this manner. The combination of fracture analy-sis, permeability and pressure data is of great usenot just for predicting the performance of a partic-ular well, but also for analyzing how the reservoiris responding to water injection and decidingwhether to drill horizontal or vertical wells.

ConclusionOperators are expanding their use of modernwireline formation testers to determine perme-ability and help make important well-completionand reservoir-management decisions. Comparedwith conventional cores and well tests, thesetesters provide cost-effective information at a

scale intermediate between the two. This infor-mation is critical for evaluating the effect ofreservoir heterogeneities, baffles and conduits.

Wireline formation testers measure perme-ability in different ways, depending on the hard-ware configuration. The mini-DST is particularlyuseful for evaluating small intervals at a fractionof the cost of a full well test. The interval pres-sure-transient test provides the most reliable andextensive permeability information from thesetools. With recent developments in software andinterpretation techniques, interval tests can nowevaluate highly layered formations, horizontalwells, and even gas reservoirs.27 The latter haveoften been considered too challenging becauseof the high compressibility and mobility of thefluid. In addition, the risk of sticking the tool—the fear of many operators—has been reducedthrough the use of risk-assessment software.28

Currently, engineers are seeking to improveresults in formations with high mobilities, heavyoil or unconcolidated sands—all difficult but notimpossible cases. Work continues on the peren-nial problem of scaling up from cores to tests,and of integrating interval-test results with otherdata. Attempts are being made to measure in situthe variation of effective permeability with watersaturation, using the fluid fractions measuredwhile sampling in combination with openholelogs and interval-test data. As long as reservoirscontinue to be heterogeneous and permeabilitydistribution remains an issue—both virtual cer-tainties—wireline formation testers will beneeded to evaluate them, and improvements willcontinue to be made. —JS/LS

24. Goode PA, Pop JJ and Murphy WF III: “Multiple-ProbeFormation Testing and Vertical Reservoir Continuity,”paper SPE 22738, presented at the SPE Annual TechnicalConference and Exhibition, Dallas, Texas, USA, October6-9, 1991.

25. A thickness-weighted arithmetic average was used forthe horizontal permeability, and a thickness-weightedharmonic average for the vertical permeability.

26. Kuchuk FJ: “Interval Pressure Transient Testing withMDT Packer-Probe Module in Horizontal Wells,” paperSPE 39523, presented at the SPE India Oil and GasConference and Exhibition, New Delhi, India, February17-19, 1998.

27. Ayan C, Donovan M and Pitts AS: “Permeability andAnisotropy Determination in a Retrograde Gas Field toAssess Horizontal Well Performance,” paper SPE 71811,presented at the Offshore Europe Conference,Aberdeen, Scotland, September 4-7, 2001.

28. Underhill WB, Moore L and Meeten GH: “Model-BasedSticking Risk Assessment for Wireline Formation TestingTools in the U.S. Gulf Coast,” paper SPE 48963, presentedat the SPE Annual Technical Conference and Exhibition,New Orleans, Louisiana, USA, September 27-30, 1998.

Packer

Probe

Pres

sure

der

ivat

ive

0.1

1

10

100

0.001 0.01 0.1 1Time since end of drawdown, hr

ProbeDual-packer

module

Slope = 1

Slope = 1/2 Slope = 1/2

> Pressure derivatives from probe and packer transients (bottom) for the analysis of fractures in a horizontal well. Theengineer set the dual-packer (top) astride a set of fracturesthat had been interpreted on FMI images (at 2983 ft, see figureprevious page), and performed an IPTT. The probe derivativeis less than the packer derivative, but spherical flow ends atthe same time on both transients. These observations alongwith the positive skin are best explained if the fracturesbetween the packers are not hydraulically conductive, and if there is a conductive fracture to the left of the probe.

50387schD02R1.p23.ps 11/17/01 6:26 PM Page 23